You are on page 1of 395

category is a section that treats the effects of instrumental and environmental or all of the manuscript.

l of the manuscript. Included are Professors R. R. Bessette, Southeastern


noise on the precision and accuracy of spectroscopic absorption measurements. Massachusetts University; M. F. Bryant, University of Georgia; J. F. Coetzee,
Also largely new is the development of general chromatographic theory from University of Pittsburgh; P. Dumas, Trenton State College; E. T. Gray, Jr.,
kinetic considerations. The chapter on optical instruments brings together University of Hartford; D. M. King, Western Washington University; P. F. Lott,
information that was formerly found in chapters on infrared, ultraviolet and University of Missouri, Kansas City; C. H. Lochmuller, Duke University;
visible, and atomic absorption spectroscopy; flame and emission spectroscopy; F. W. Smith, Youngstown State University; M. Thompson, University of
and fluorescence and Raman spectroscopy. Toronto; A. Timnick. Michigan State University; W. H. Smith, Texas Tech
Since the publication of the first edition, a host of modifications of well- University; J. E. Byrd, California State College-Stanislaus; A. M. Olivares.
established instrumental methods have emerged and are now included. Among Texas A & I University; and E. J. Billo. Boston College.
these new developments are Fourier transform nuclear magnetic resonance and We also wish to thank Dr. Natalie McClure, now of Syntex Corporation.
infrared methods, infrared photometers for pollutant measurements, laser for preparing problem sets and their solutions for Chapters 8, 14, and 17, and
modifications of spectroscopic measurements, nuclear magnetic resonance Professor Daniel C. Harris. of University of California-Davis. for preparing
measurem<:nts on nuclei other than hydrogen, electron spin resonance problem sets and solutions for Chapters 5. 9-13, 23. and 25.
spectroscOpy, photon counting, inductively coupled plasma sources for emission
spectroscopy, semiconductor detectors for X-rays and gamma rays, energy Stanford, California DoUGLAS A. SKOOG
dispersive: X-ray systems, the electron microprobe, X-ray photoelectric San Jose, California DoNALD M. WEST
spectrometry, chemical, field ionization, and spark sources for mass spectrom- September, 1979
etry, gas electrodes, potentiostatic controls for polarography, differential
pulsed and rapid scan polarography, cyclic and ac voltammetry, high
performance liquid chromatography, new packings for gas-liquid chroma-
tography. interfacing of chromatography with mass. infrared. and fluorescence
detectors, instruments for simultaneous. multielement analysis, and automated
instruments for spectroscopy, electroanalysis. titrimetry, and chromatography.
In addition to the foregoing, the new edition contains, as an appendix, a
brief treatment of the propagation of random errors in analytical measurements.
It is worth noting that selected portions of the material found in this text
have also appeared in the authors' two other titles.· The principal overlap of
the material in this book with that in the earlier two occurs in the chapters
on potentiometry, coulometry, and voltammetry. Duplication willbtl found to
a lesser extent in the chapters on ultraviolet and visible absorption'spectroscopy
and atomic emission spectroscopy.

The authors wish to acknowledge with thanks the considerable contributions


of Professor Alfred R. Armstrong, College of William and Mary. and Dr. James
LuValle, Stanford University, who have read the entire manuscript in detail
and offered many useful suggestions. We are also grateful to Professor R. deLevie
of Georgetown University for his comprehensive review and numerous sug-
gestions for Chapters 2 and 3, to Professor H. S. Mosher of Stanford University
for his helpful comments on the chapter on optical activity. and to Professors
S. R. Crouch of Michigan State University and J. D. Ingle. Jr.. of Oregon State
Universit~ for their detailed criticisms on and suggestions for the section on
precision of spectral measurj:ments. Finally, we offer our thanks to several
others who have taken the time to comment, often in some detail, on portions

'0. A. Skoog and D. M. Wes •• FundamemaL, o(Analrrical Chern;.."r. 3d cd .. 1976. and Ana1l'tical
Chern;.",)'. 3d cd .. 1979. Hok. Rinehan and W'inston: l'ew York. . .
CONTENTS

1 INTRODUCTION
Types of Analytical Methods
Instruments for Analysis

2 ELECTRICITY AND ELECTRIC CIRCUITS


Introduction
Reactance in Electrical Circuits
Simple Electrical Measurements

3 ELEMENTARY ELECTRONICS
Semiconductors
Amplifiers Employing Transistors
Power Supplies and Regulators Employing Transistors
Operational Amplifiers
Noise
Readout Devices
Interfacing of Instruments with Computers

ELECTROMAGNETIC RADIATION AND ITS INTERACTIONS WITH


MATTER 91 (
Properties of Electromagnetic Radiation 92
The Interaction of Radiation with Maller 101
Emission of Radiation 109

5 COMPONENTS OF INSTRUMENTS FOR OPTICAL


SPECTROSCOPY 113
Components and Configurations of Instruments for Optical Spectroscopy 114
Radiation Sources 115
Wavelength Selection; Monochromators 123
Wavelength Selection: Filters 136
Sample Containers 137
Radiation Detection 137
Signal PrOcessors and Readouts 145

6 AN INTRODUCTION TO ABSORPTION SPECTROSCOPY 148


Terms Employed in Absorption Spectroscopy 149
Quantitative Aspects of Absorption Measurements 151
/ APPLICATIONS OF ULTRAVIOLET AND VISIBLE ABSORPTION '1'f. MISCELLANEOUS OPTICAL METHODS 352
MEASUREMENTS 168 V'1fefractometry . .
169 353
Absorbing Species
180
,Polarimetry 357
Some Typical Instruments Optical Rotatory Dispersion and Circul~r Dichroism 370
Application of Absorption Measurement to Qualitative Analysis 187
Quantitative Analysis by Absorption Measurements 188 14 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY 376
Photometric Titrations 197 Theory of Nuclear Magnetic Resonance 377
Automatic Photometric and Spectrophotometric Analysis 199 Experimental Methods of NMR Spectroscopy 386
Photoacoustic Spectroscopy 202 Environmental Effects on Proton NMR Spectra '390
Applications of Proton NMR 404
INFRARED ABSORPTION SPECTROSCOPY 209 Application of Proton NMR to Quantitative Analysis 407
Theory of Infrared Absorption 210 Fourier Transform NMR '409
Infrared Instrument Components 219 Electron Spin Resonance Spectroscopy .412
Some Typical Instruments 221
Sample Handling Techniques 226 \/'15 X-RAY METHODS 426
Qualitative Applications of Infrared Absorption 230 Fundamental Principles ,427
Quantitative Applications 237 Instrument Components 434
Infrared Fourier Transform Spectroscopy 241 X-Ray Fluorescence Methods 444
Hadamard Transform Spectroscopy 254 X-Ray Diffraction Methods 450
The Electron Microprobe Method 452
X-Ray Photoelectron Spectroscopy ,453
RAMAN SPECTROSCOPY 262
Theory of Raman Spectroscopy 263 16 RADIOCHEMICAL METHODS i458
Instrumentation 268 The Radioactive Decay Proa:ss ,459
Applications of Raman Spectroscopy 270 Neutron Activation Analysis 467
Isotopic Dilution Methods 471
\/10 MOLECULAR FLUORESCENCE SPECTROSCOPY 279 Radiometric Methods 473
Theory of Fluorescence 280
Instruments for Fluorescence Analysis 289 vi7 MASS SPECTROMETRY 476
Applications of Fluoremetry 293 The Mass Spectrometer 477
Nephelometry and Turbidimetry 294 Mass Spectra 488
Qualitative Applications of Mass Spectrometry 491
Quantitative Applications of Mass Spectrometry 496
".11 ATOMIC SPECTROSCOPY 301
Theory of Flame Spectroscopy 303 18 AN INTRODUCTION TO ELECTROANALYTICAL CHEMISTRY 500
Flame Characteristics 308 Electrochemical Cells 501
Atomizers for Atomic Spectroscopy 312 Cell Potentials 507
Atomic Absorption Spectroscopy 315 Half-Cell or Electrode Potentials .512
Atomic Emission Spectroscopy 326 Calculation of Cell Potentials from Electrode Potentials ,524
Atomic Fluorescence Spectroscopy 332 Effect of Current on Cell Potentials 527
Reference Electrodes i531
12 EMISSION SPECTROSCOPY 335 Standard Weston Cells '534
Emission Spectra 337
19 POTENTIOMETRIC METHODS - 538
Arc and Spark Sources 337
Argon Plasma Sources Indicator Electrodes 539
340
Instruments for Emission Spectroscopy Instruments for Cell Potential Measurement 554
342
Applications of Emission Spectroscopy Direct Potentiometric Measurements 555
347
Potentiometric Titrations 559
1 580 vAPPENDIX 1 PROPAGATION OF UNCERTAINTIES IN
20 . ELECTROGRAVIMETRIC AND COULOMETRIC METHODS
581 PHYSICAL MEASUREMENTS
Current-Voltage Relationship During an Electrolysis
Electrogravimetric Methods of Analysis 587 APPENDIX 2 SOME STANDARD AND FORMAL ELECTRODE
Coulometric Methods of Analysis 590 POTENTIALS

21 VOLTAMMETRY AND POLAROGRAPHY 605


Polarography 606
Applications of Polarography 617
Voltammetry at Solid Electrodes 622
Polarography with Potentiostatic Control 625
Modified Voltammetric Methods 627
Amperometric Titrations 635
Amperometric Titrations with Two Polarized Microelectrodes 637

22 CONDUCTOMETRIC METHODS 643


Electrolytic Conductance 644
The Measurement of Conductance 647
{::onductometric Titrations 648
Applications of Direct Conductance Measurements 651
Oscillometry 652

THERMAL METHODS 654


Thermogravimetric Methods 655
Differential Thermal Analysis and Differential Scanning Calorimetry 657
Thermometric Titrations 662

AN INTRODUCTION TO CHROMATOGRAPHY 666


A General Description of Chromatography 667
The Rate Theory or Chromatography 671
Separations on Columns 676
Summary of Important Relationships for Chromatography 681
Qualitative and Quantitative Analysis by Chromatography 681

LIQUID CHROMATOGRAP.HY \689


Column Chromatography 690
Planar Chromatography 705
Electrophoresis and ElectrochrOnilatography 709

GAS-LIQUID CHROMATOGRAPHY 715


Principles or Gas-Liquid Chromatography 716
Apparatus 717
Applications or Gas-Liquid Chromatography 722
Gas-Solid Chromatography 731
Examples of Applications of Gas Chromatography 731
PRINCIPLES
OF
INSTRUMENTAL
ANALYSIS
A chemical analysis provides information ductance is to serve as the signal for the separation and resolution of closely related arriving at a decision as to which method to
about the composition of a sample of matter. analysis of one of these species in a sample compounds. Common separation methods in- choose, the chemist must take into account
The results of some analyses are qualitative that also contains the other. clude chromatography, distillation, extrac- the complexity of the materi~ to be
and yield useful clues from which the molecu- tion, ion exchange, fractional crystallization, analyzed, the concentration of the species of
lar or atomic species, the structural features, and selective precipitation. One' of the signals interest, the number of samples to be
or the functional groups in the sample can be TYPES OF listed in Table 1-1 is ordinarily used to com- analyzed, and the accuracy required. His
deduced. Other analyses are quantitative; ANALYTICAL METHODS plete the analysis following the separation choice will then depend upon a knowledge of
here, the results take the form of numerical step. Thus, for example, thermal conductivity, the basic principles underlying the various
data in units such as percent, parts per Table I-I lists most of the common signals volume, refractive index, and electrical con- methods available to him, and thus their
million, or millignlms per liter. In both types that are useful for analytical purposes. Note ductance have all been employed in conjunc- strengths and limitations. The development of
of analysis, the required information is ob- that the first six signals involve either the tion with various chromatographic methods. this type of knowledge represents a major
tained by measurU1g a physical property that emission of radiation or the interaction of This text deals with most of the instrumen- goal of this book.
is characteristically_ related to the component radiation with matter. The next three signals tal methods listed in Table 1-1, as well as with
or components of mterest. are electrical. Finally, five miscellaneous sig- many of the most widely employed separation INSTRUMENTS
It is convenient to describe properties nals are grouped together. Also listed in the procedures. Little space is devoted to the clas- FOR ANALYSIS
which are useful for determining chemical table are the names of analytical methods sical methods, since their use is normally
composition as muilytical signals; examples of which are based on the various signals. treated in elementary analytical courses. In the broadest sense, an instrument for
such signals include emitted or absorbed It is of interest to note that until pCrhaps Table I-I suggests that the chemist faced chemical analysis converts a signal that is
light, conductance, weight, volume, and re- 1920, most analyses were based on the last with an analytical problem may have a bewil- usually not directly detectable and under-
fractive index. None of these signals is unique two signals listed in Table I-I-namely, mass dering array of methods from which to standable by man to a form that is. Thus, an
to a Jiven species. Thus, for example, aU me- and volume. As a consequence, gravimetric choose. The amount of time spent on the instrument can be viewed as a communica-
tallicelements in a Sample will ordinarily emit and volumetric procedures have come to be analytical work and the quality of the results tion device between the system under study
ultraviolet and visible radiation when heated known as classical methods of analysis, in are critically dependent on this choice. In and the scientist.
to a sufficiently hiih temperature in an elec- contrast to the other procedures, which are
tric arc; all charged species conduct electric- known as instrumental methods.
ity; and all of the components in a mixture Beyond the chronology of their develop-
contribute to its refractive index, weight, and ment, few features clearly distinguish in-
volume. Therefore, all analyses require a strumental methods from classical ODes.Some TABLE 1-1 SOME ANALYTICAL SIGNALS
separation. In some instances, the separation instrumental techniques are more sensitive
step involvc:s physical isolation of the individ- than classical techniques, but many are not. Signal Analytical Methods Baed on Maisurement of Signal
ual chemical components in the sample With certain combinations of elements or Emission spectroscopy (X-ray, UV, visible, electron,
prior to signal generation; in others, a signal compounds, an instrumental method may be Augerl; flame phot~metry; fluorescence (X-ray, UV,
IS generated or observed for the entire sample,
more Specific; with others, a gravimetric or a visible ; radiochemical method~ .
following which the desired signal is isolated volumetric approach may be less subject to Spectrop otometry (X-ray, UV: VISible,IR); .
from the others. Some signals are susceptible interference. Generalizations on the basis of colorimetry· atomic absorptiOn, nuclear magnetic
to the latter treatment, while others are not. resonance, ~nd electron spin resonance spectroscopy
accuracy, convenience, or expenditure of time Turbidimetry; nephelometry, Raman spectroscopy
For example, when a sample is heated in an Scattering of radiation
are equally difficult to draw. Nor is it neces- Refraction of radiation Refractometry; interferometry
electric arc, the w~velength distribution for sarily true that instrumental procedures X-Ray and electron diffraction !Deth~s .
Diffraction of radiation
the radiation of each metallic species is employ more sophisticated or more costly Rotation of radiation Polarimetry; optical rotatory dISpersion; Circular
unique to that spj:Cies; separation of the apparatus; indeed, use of a modern automatic dichroism
wavelengths in a s~itable device (a spectro- Electrical potential Potentiometry; chronopotentiometry
balance in a gravimetric analysis involves Polarography; amperometry; coulometry
scope) thus makes possible the identification Electrical current
more complex and refined instrumentation Electrical resistance Conductimetry
of each component without physical separa- than is required for several of the methods Mass spectrometry
Mass-to-char,ge ratio
tion. On the other hand, no general method found in Table 1-1. Rate of reaction Kinetic methods
exists for distinguishing the conductance of Thermal properties Thermal conductivity and enthalpy methods
In addition to the methods listed in the
sodium ions from that of potassium ions. Mass Gravimetric analysis
second column of Table I-I, there exists Volumetric analysis
Here, a physical separation is required if con- Volume
another group of analytical methods for the
CornponenQ sipal to another. An example is a thermo- addition to being amplified,signals are often electronics, because the generation, transduc-
of Inatrulll8l1t8 couple, which converts a radiant heat signal multiplied by a constant smaller than unity tion, amplification,and display of a signal can
into an electrical voltage; the bellows in an (attenuated1 integrated, differentiated,added be rapidly and conveniently accomplished
Regardless of its complexity, an instrument aneroid barometer transduces a pressure or subtracted, or increased exponentially. with electronic circuitry. Numerous trans- .
generally contains nO more than four fun- sipal into a mechanical motion signal. Most Other operations may involve conversion ducers have been developed for the conver- .
damental components. As shown schema- of the transducers that we will encounter con- into an alternating current, rectification to sion of chemical signals to electrical fonn,
tically in Figure 1-1,these components are a vert analytical signals into an electrical give a direct current, comparison of the and enormous amplification of the resulting
signal generator, an input tr~ucer or voltage,current, or resistance, because of the transduced signal with one from a standard, electrical signals is possible.In turn, the eleo- .
detector, a signal processor, and an output ease with which electrical signals can be and transformation from a current to a volt- trical signals are readily presented on meters,
transducer or readout. A general description amplified and modified to drive readout de- age or the converse. on recorders, or in digital form.
of these components follows. vices.Note, however,that the transduced sig- R__ l>efices. Readout devices employ
SipaI Geaerators. Signal generators pro- Owing to the large amount of electronic
nals for the last two instruments in Table 1-2 an output transducer to convert the amplified circuitry in the laboratory, the modem
duce analytical signals from the ,components are nonelectrical. signal from the processor to a signal that chemist is faced with the question of how
of the sample. The generator 'may simply SipaI Processors. The signal processor can be read by the human eye. Readout much knowledge of electronics is needed to
be the sample itself. For example, the signal modifiesthe transduced signal in such a way devices take the form of meters, strip chart make most efficient use of the equipment
for an analytical balance is ~ mass of a as to make it more convenient for operation recorders, oscilloscopes, pointer scales, and available for analysis.We believethat it is not
component of the sample; for a pH meter, the or the readout device. Perhaps the most digital devices. only desirable but feasible for the chemist to
signal is the activity of hydrogen ions in a common modification is that of amplifica- possess a qualitative knowledgeof how elec-
solution. In many other instruments, tion-that is,mUltiplicationof the signal by a tronic circuits work. Thus, Chapters 2 and 3
however, the signal generator is more elabo-
.rate. For example, the signal generator of an
infrared spectrophotometer includes, in addi-
constant greater than unity. In a two-pan
analytical balance, the motion of the beam
is amplified by the pointer, whose displace-
-.d
Electronics
Instrumentation
are largely devoted to a presentation of this
topic; subsequent chapters are concerned
The growth of instrumental analysis has with the application of various instruments to
tion to the sample, a source of infrared radia- ment is significantly greater than that of closely paralleled developmentsin the fieldof the solution of chemical problems.
tion, a monochromator, a beam Chopperand the beam itself. Amplification by a photo-
splitter, a sample holder, and' a radiation graphic film is enormous; here, a single
attenuator. photon may produce as many as 1012 silver
The second column of Table 1-2lists some atoms. Electrical signals can, of course, be
examples of signal generators. readily amplifiedby. factor or 106 or more. TABLE '-2 SOME EXAMPLES OF INSTRUMENT COMPONENTS
•••• T........ ••. Detectors. A trllllS- A variety of other modifications are
ducer is a device that converts one type of commonly carried out on electricalsignals.In SIpaI A..tyticaI IIIpIIt T~ SipaI
1• ...-. GeRenator Signal Tnnsducer Signal ~ Readout

Photometer Tungsten lamp. Attenuated Photocell Electrical None Current meter


glass filter, light beam current
sample

IIT'JI!
s:...-
Atomic emission Flame, UVor Photomultiplier Electrical Amplifier, Chart
sp«:trometer monochromator, visible tube potential demodulator recorder
Electrical chopper, radiation
or sample
mechanical
Coulometer dc source, Cell current Electrodes Electrical Amplifier Chart

-"";';"
_~_;.u_.~_,.:.=;.,
~.-- pH meter

X-Ray powder
diffractometer
sample
Sample

X-Ray tube,
sample
Hydrogen ion
activity
Dilfra(.ted
radiation
Glass-calomel
electrodes
Photographic
film
current
Electrical
potential
Latent
image
Amplifier,
digitalizer
Chemical
developer
recorder
Digital unit

Black images
on film

/112.301
o 0
'I Oi~ital
un,t
Color
comparator
Sunlight,
sample
Color Eye Optic nerve
signal
Brain Visual color
response

Output transducer
or readout
This chapter provides a brief and elementary where Q is the charge. The ampere, A, the unit
review of the laws governing electrical cur- of current, corresponds to a rate of one cou-
rents, the properties of direct and alternating lomb per second.
current circuits, and selected methods for The electrical potential V between two
measuring electrical quantities.! points in space is the work required to move
an electrical charge from one of the points to
the other. When the work is given in joules
and the charge in coulombs, the unit for
potential is the VOII, V. That is, one volt is
An electrical current is the motion of a charge equal to one joule per coulomb.
through a medium. In metallic conductors, The olun, n. is the unit of resistance R to
only electrons are mobile; here, the current current flow and is the resistance through
involves motion of negative charges only. In which a potential of one volt will produce a
media such as ionic solutions and semicon- current of one ampere.
ductors, both negative and positive species Electrical conductance G is the reciprocal
are mobile and participate in the passage of of resistance and has the units of reciprocal
electricity. ' ohms, n- !, or siemens, S.
Electrical power P is the rate of electrical
Electrical Units work in joules per second or watts W. The
The unit of charge or quantity of electricity is power dissipated during passage of electrons
the coulomb,: C, which is the charge required is given by
to convert O;<lOUI800 g of silver ion to silver
metal. Another unit, the faraday, F, corre-
sponds to o;ne equivalent of charge, that is, Substitution of Equation 2-4 yields the power
6.02 x 1023 charged particles. The relation- loss in a resistance of R ohms; that is
ship between the two units is readily cal-
culated from their definitions:
coulombs 107.87 g Ag"/equivalent Ag"
faraday 0.00111800 g Ag"/coulomb
Electrical Currents
I equivalent Ag"
x 1 equivalent of charge If a conducting path exists between two
points with differing potentials, charge will
= 9.649 x 10· flow until the difference in potentials becomes
zero. For example, when the switch is closed
Electric' current 1 is the rate of flow of a
in the direct current circuit shown in Figure
charge. That is,
2-1, a conducting path is provided, through
which charge can flow from one terminal of
the battery to the other. Current will continue
until the potential difference between .4 and B
I Some generai references. which cover these topics. becomes zero-that is, until the battery is
include: A. J.' Diefenderfer. Principles on E1«rronic
Insrrumentation, 2d cd. Philadelphia:' Saunders, 1979; discharged.
J. J. Brophy. Basic Electronics for Scientists. 3d cd. By convention, the direction of a current is
New York: McGraw-Hil~ t977; H. V. Malmstadl, C. G. always from the positive terminal to the nega-
Enke. S. R. Crouch, and G. Hortick, Electronic Measure- tive, regardless of the type of particle that car-
ments for Scientists. Menlo Park, CA: Benjamin, t974; ries the current. Thus. in the external circuit
and R. J. Smith, Circuits, Devices, and Systems, 3d cd. shown in Figure 2-1, current takes the form of
New York: Wiley. 1976.
with sulfuric acid at the expense of the region A to A. The three voltages thuS oppose the volt-
around C. age of the battery V and must be given oppo-
Two electrochemical reactions make pos- site signs.
',1

1
sible the transition from electronic conduc- Substitution of Ohm's 'law into Equation
tion in the metal to conduction by ions in the 2-5 gives
solution. In the first of these two reactions,
hydrogen ions consume electrons from C; in ,.1
the second, water molecules give up electrons
Note that the total resistance R of a series
at D. Thus, three modes of passage of electric- V I" ,.1 circuit is equal to the sum of the resistances of
ity exist in this simple circuit: electronic, +
ionic, and electrochemical. the individual components. That is,
o
R = RI + R2 + R3 (2-7)
Laws of Electricity" '-',v,- '.v.-'.-'.
v- + + v. Applying Ohm's law to the part of the cir-
R- R, + R2 + R2
OIun's Law. Ohm's law takes the form cuit from points B to A gives
Dxygen formation Hydrogen forlTjotion FIGURE 2-2 Resistors in series; a volt-
H.O-to. + 2H' + 2e 2H' + 2e-H.- age divider. VI = IIR. = IRI
Conduction by electrochemicai reaction
where V is the potential in volts between two Dividing by Equation 2-6 yields
points2 in a circuit, R is the resistance be- VI .lR1
tween the two points in ohms, and 1 is the
resulting current in amperes.
Kirchhotl's current law to point D in the 17 = ./(RI + R2 + R3)
circuit, we write
Ohm's law applies to electronic and ionic
motion of electrons e through the metal don- conduction; it is not, however, applicable to
:Iuctors Be and AD. The direction of the Cur- conduction across interfaces such as the elec-
rent I, however, is said to be from A to D to trode surfaces in the cell shown in Figure 2-1.
C to B; that is, current is always treated as if Kircbhofl's Laws. KirchhoJls cu"ent law It is clear that we may also write
it were a ftow of positive charge. states that the algebraic sum of currents
Note that the current out of D must be oppo- Y: = VR2/R
The two platinum electrodes (C and D) around any point in a circuit is zero. Kirch-
hofflJ voltage law states that the algebraic sum site in sign to the input current. Similarly,
md the dilute sulfuric acid solution in which
of the voltages around a closed electrical loop application of the law to point C gives
:hey are immersed represent an electrochemi-
:al cell. The mechanism by which electricity is zero. [3 = [2 Vl = VR3/R
:lasSeSthrough this cell is entirely different The applications of Kirchhotl's and Ohm's Voltage Dividers. Series resistances are
'rom that in the metallic conductor. Here, the laws to simple dc circuits are considered in Thus, it is apparent that the current is the
widely used in electronic circuits to provide
tow involves migration of positive hydrogen the next section. same at all points in a series circuit or
potentials that are variable functions of an
ons toward electrode C, and of negative sul- [=[t=[2=[3=[4 input voltage. Devices of this type are called
ate and hydrogen sulfate ions toward elec- voltage dividers. As shown in Figure 2-38, one
rode D. Note that the excess negative ch~rge Series Circuits Application of the voltage law to the cir- type of voltage divider provides voltages in
hat tends to accumulate around D as a result Consider the simple electrical circuit in cuit yields the equation discrete increments; the second type (Figure
If the anionic migration is exactly offset; by Figure 2-2, which consists of a battery, a 2-3b), called a potentiometer,3 provides a
V-V3-V2-VI=O
he postively charged hydrogen ions switch, and three resistors in series. Applying potential that is continuously variable.
Iroduced by the electrode process; similarly,
he excess hydrogen ions that migrate to elec-
rode C are removed by the electrochemical
2 In Chapten 2 aud 3, the symbol V will be used to 3 The word potentiometer is also used in a different c0n-
,rocess. The consequence is that the solution Note that point D is positive with respect to
denote electrical potential in circuits. In Chapters 18 to text as the name for a complete instrument that emplo}S
emains homogeneous with respect to charge, point C, which in turn is positive with respect a linear voltage divider for the accurate measurement or
22, however, the electrochemical convention will be
'ut the region around electrode D is enriched followed in which electromotive force is designated as E. to point B; finally, B is positive with respect potentials.
B
,.
l, " t -'3'.t -
A

'3 -
when R is the net circuit resistance, then
1 1 1 1
loon

loon
VAS V -

+ +
R,
'. +

'.t
I~ +
R.

-'. +
R.
-=-
R R.
+- +-
R2 R3
In a parallel circuit, in contrast to a series
(2-11)
15VL--=:J1J.Jn
~

circuit, it is the conductances that are additive B


FIGURE 2-4 Resistors in parallel.
rather than the resistances. That is, since
I G = l/R, Here, we have the equivalent of two series
VAS 100 !! I
resistances, and
I
I a . proportionality constant. Similarly, G = G. + G2 + G3 (2-12)
I
I RAIt = kAB. Combination of these relation- Just as series resistances form a voltage R = RI + R2• 3 = 9.0 + 13.3 = 22.3 0
200
loon V= VAS X 500 ships with Equation 2-8, yields divider, parallel resistances create a current From Ohm's law, the current I is given by
divider. The fraction of the total current that
RAC AC is present in R. in Figure 2-4 is
VAC = VA/t- = VA/t- (2-9)
loon RAIt AB
I. VIR. l/R. G. Employing Equation 2·8, the voltage V.
A Parallel Circuits
T= VIR = l/R =G" across RI is
·V = IS x 9.0/(9.0 + 13.3) = 6.0 V
Figure 2-4 depicts a parallel dc circuit. Apply-
ing Kirehho/J's current law to point A, we Similarly, the voltage across resistors R2 and
obtain R3 is
10.0 B
C V2 = V3 = V2.] = IS x 13.3/22.3
EXAMPLE

\
For the accompanying circuit, calculate: = It9s =9.0 V
(a) the total resistance: (b) the current draVl:n
Note that the sum of the two voltag\..'Sis 15 V.
AC from the battery; (c) the current present 10
VAS-
VAC· VAS X Ai as demanded by Kirchho/J's voltage law.
Kirchhoff's voltage law gives three independ- each of the resistors; and (d) the potential
I The current in R. is given by
I ent equations. Thus, we may write, for the drop across each of the resistors.
I
loop that contains the battery and R. II = I =0.67 A
~AC
I
V-/.RI =0 The currents through R2 and R] are found
I
0.0 A I from Ohm's law. Thus.
V= I.R.
/2 =9.0/20 = 0.45 A
lbl For the loop containing V and R2,
FIGURE 2-3 Voltage dividers: (a) selector 1., = 9.0140 = 0.22 A
V = 12R2
type and (b) continuously variable type Note that the two currents add to give the nct
(potentiometer). For the loop containing V and R3, R2 and R] are parallel resistances. Thus; current. as required by Kirchho/J's law.
V=/3R3 the resistance R2•3 between points A and B
will be given by Equation 2-11. That is.
We could write additional equations for the Alternating Current
loop containing R. and R2 as well as the loop 1 1 1
In most potentiometers, such as the one --=- +-- The electrical output from transducers of
containing R2 and R3• However, these equa- R2.] 20 40
shown in Figure 2-3b, the resistance is analytical signals often fluctuate pcriodicall~.
tions are not independent of the foregoing
linear-that is, the resistance between one These fluctuations can be represented (as 10
three. Substitution of the three independent
end, A, and any point, C, is directly propor- Figure 2-5) by a plot of th~ insta~taneous
equations into Equation 2-10 yields R2•3 = 13.30
tional to the length, AC, of that portion of the current or potential as a funct.on of tlJ~e. The
resistor. Then RAC = kAC where AC is ex- V V V V We can now reduce the original circuit to the period, t, for the signal is the time reqUired for
pressed in convenient units of length and k is 1=- =- +- +_ following equivalent circuit. the completion of one cycle.
R R. R2 R3
'-- ---------.1
I
The rCc:ipr()!:&1of the period is the fre- The vector rotates at a rate of 2n radians I
quency,/. of the cycle. That is, in the period t; thus, the angular velocity is I
1== 1/'r (2-14)
given by --------.- I
I
I
2n I I
The unit of frequency is the hertz, Hz, which co == - = 2nl (2-15) " -A I
is defined as one cycle per second. t i i
Sille-Wan Clneals. The sinusoidal wave If the vector quantity .is current or voltage,
(Figure 2-5a) is the most frequently en- the instantaneous current i or instantaneous
countered type of periodic electrical signal. A voltage v at time t is given by4 (see Figure
common example is the alternating current 2-6)
produced by rotation of a coil in a magnetic
fickt (as in an electrical generator). Thus, if j = I, sin cot = I, sin 21ift (2-16)
the instantaneous current or voltage or alternatively
produced by a generator is plotted as a func-
tion of time, a sine wave results. v == V, sin cot = V, sin 2nft (2-17)
A pure sine wave is conveniently repre- where I, and v,.
the maximum or peak cur-
sented as a vector of length I, (or V,1 which rent and voltage, are called the amplitude A of
is rotating counterclockwise at a constant an- the sine wave.
gular velocity co.The relationship between the Figure 2-7 shows two sine waves having
vector representation and the sine-wave plot different amplitudes. The two waves are also
is shown in Figure 2-6. out of phose by 90 deg or n/2 radians. The
phase difference is called the plulSe angle.
i
which arises from one vector leading or lag- sin wt=-
ging a second by this amount. A more gener- '.
i" I" sin wI • Ip sin 2.fr
alized equation for a sine wave, then, is
FIGURE 2-6 Relationship between a sine wave of period t and an,tplitude 1"
j == I, sin(cot + 1/1) = I, sin(21ift + 1/1) (2-18)
and the corresponding vector of length 1" rotating at an angular velocity or OJ or
where 1/1 is the phase angle from some refer- a frequency of 1Hz.
ence sine wave. An analogous equation coukt
be written in terms of voltage.
The current or voltage associated with a
sinusoidal current can be expressed in several
ways. The simplest is the peak amplitude I,
(or v,.), which is the maximum instantaneous
current or voltage during a cycle; the peak-
to-peak value. which is 21, or 2V" is also
employed occasionally. The root mean square
or rms current in an ac circuit will produce
the same heating in a resistor as a direct cur-

• In treating currents that change witb time, it is useful to


symbolize the instantJmeoUl current, voltage. or charII'
Time---~·
witb the lower case Ietten I, v. and q. On tbe other hand,
FIGURE 2-5 Examples of periodic signals: capital Ietten are used for steady current, voltage. or FIGURE 2-7 Sine waves with different amplitudes (1" or V,,)
(a) sinusoidal, (b) square wave, (c) ramp, and charII'. or a specificaUydefined variable quantity such as and out of phase by 90 deg or n/2 radians.
(d) sawtooth. a peak voltage current; tbat is, V,and I,.
rent of the same magnitude. Thus, the rms
cunent is important in power calculations
(Equations 2-2 and 2-3). The rms current is Structurally. capacitors and inductors are
given by quite different. A typical capacitor consists of

1,••••=
m = 0.70711'
,(f
a pair of conductors separated by a thin layer
of a dielectric substanee-that is, by an elec-
trical insulator that contains essentially no
mobile, current-carryin& charged species. The
simplest capacitor consists of two sheets of
metal foil separated by a thin film of a dielec-
tric such as air, oil, plastic, mica, paper, cer-
amic, or metal oxide. Except for air and mica
REACTANCE IN capacitors. the two layers of foil plus the insu- ,,-v-;--
ELECTRICAL CIRCUITS lator are usually folded or rolled into a com- I
pact packase and sealed to prevent
Whenever the current in an electrical qircuit atmospheric deterioration.
is increased or decreased, energy is reciuired In contrast, an in4uctor is ordinarily a coil
to c:harse the electric and magnetic fields as- of insulated wire of relatively large diameter
sociated with the flow of charse. As a to minimize resistance. Some inductors are
consequence, there develops a counterforce or
reactanu which tends to counteract the
change. Two types of reactance can be
recognized. namely. capacitance and i,wuct-
ance. When the rate of change in current is
wound around a soft iron core to enhance
their reactance; otl1ers have an air core. The
latter type is sometimes called a choke.
Capacitors are significantly less bulky than
·inductors and, in addition, can be fabricated
i~Ol'>--
\....
low. the reactance of most of the compOnents as part of printed circuits. Thus. to the extent
in a circuit is sufficiently small to be ne- possible, 'capacitors are employed in prefer- +I I / VL
glected. With rapid changes, on the other ence to inductors in electronic circuit design.
hand, circuit elements such as switches, junc-
tions, and resistors may exhibit a detectable
At high frequencies,-however. the reactance of
capacitors may be too small, and inductors
V
reactance. Ordinarily. this type ofreaetance is are used preferentially.
undesirable. and every effort is made to di- Capacitance. In order to describe the
minish its magnitude. properties of a capacitor. it is useful to con-
FIGURE 2-8 Behavior of (a) a series RC circuit and
Capacitance and inductance are often sider the de circuit shown at the top of Figure
(b) a series RL circuit.
deliberately introduced into a circuit with 2-8•• which contains a battery JIj. a resistor R,
capacitors and inductors. These devices pro- and a capacitor C in series. The capacitor is
vide a means for accomplishing such useful symbolized by a pair of parallel lines. A cir-
functions as converting alternating current to cuit of this kind is frequently called a series
direct or the converse, discriminating among RC circuit.
across the plates and prevents the con- the two plates disappears; here. the capacitor
signals of different frequencies, or sepaiating When the switch S is closed to position I,
tinued flow of electrons. When the current is said to be discharged.
ac and de signals. . electrons flow from the negative terminal of
ceases. the capacitor is said to be charged. A useful property of a capacitor is its abil-
In the section that follows, we shalll con- the battery into the lower conductor or plate
If the switch now is moved from 1 to 2, ity to store an electrical charge for a period
sider the properties of inductors and ~paci- ofthe capacitor. Simultaneously. electrons are
electrons will flow from the negatively of time and then to give up the stored charge
tors. The behavior and uses of these two repelled from the upper plate and flow
charged lower plate of the capacitor through as needed. Thus, if S in Figure 2-8a is first
important circuit components frequently par- toward the positive terminal of the battery.
the resistor R to the positive upper plate. held at I until C is charged and is then
allel one another. Thus. it will be instructive This movement constitutes a momentary
Again, this movement constitutes a current moved to a position between I and 2. the ca-
to note their similarities and differences as the current. which decays to zero. however, owing pacitor will remain in a charged condition for
discussion develops. to the potential difference that builds up which decays to zero as the potential between
electric circuit tend to complement one Note that the product RC that appears in
an extended period. Upon moving S to 2, as a counter-potential v when the current first
another. the last three equations has the units of time;
discharge occurs in the same way as it would begins, and tends to oppOse the flow of elec-
if the change from 1 to 2 had been rapid. trons. On the other hand, when the current in
Rate or C..-ent •••• Poteatial Cbaages in since R = vR/i and C = q/vc,
The quantity of electricity, Q, required to a cOnductor ceases, the magnetic field col- an RC Circuit. The rate at which a capacitor
RC = volts coulombs
charge a capacitor fully depends upon the lapses; this process causes the development of is charged or discharged is ~te. Consider,
coulombs/seconds x volts
area of the plates, their shape, the spacing a momentary potential which acts to continue for example, the circuit shown in Figure 2-8a.
between them, and the dielectric constant for the current. From Kirchhoffs voltage law, we know that = seconds
the material that separates them. In addition, The magnitude of the potential v that at any instant after the switch is moved to
The term RC is called the time constant for
the charge, Q, is directly proportional to the develops during the increase or decrease in position I, the sum of the voltage across C
the circuit.
applied voltage. That is, current is found to be directly proportional to and R (vc and VR) must equal the input volt-
The following example illustrates the use
the rate of current change di/dt. That is, age~. Thus,
Q = CV (2-20) of the three equations that were just derived.
di Y; = Vt + VR (2-24)
When V is the applied potential in volts and v = -L- (2-23) EXAMPLE
dt Because Y; is constant; the increase in Vc that
Q is the quantity of charge in coulombs, the Values for the components in Figure 2-8a
proportionality constant C is the capacitance Here, the negative sign indicates that the accompanies the charging of the capacitor are ~ = 10 V, R = 1000 n.
C = 1.00 pF or
of a capacitor in farads, F. One farad, then, induced potential acts to oppose the change must be exactly offset "y a decrease in VR • 1.00 x 10-6 F. Calculate: (a) the time con-
corresponds to one coulomb of charge per in current. Substitution of Equations 2-4 and 2-20 stant for the circuit; and (b) i, vc, and VR after
applied volt. Most of the capacitors used in The proportionality constant L in Equa- into this equation gives, upon rearrangement, two-time constants (t = 2RC) have elapsed.
electronic circuitry have capacitances in the tion 2-23 is termed the inductance of the (a) Time constant =
RC =
1000 x 1.00 x
microfarad (10-6 F) to picofarad (10-12 F) conductor; it has the units of henrys, H. One Y;=~+iR 10-6 = 1.00 X 10-3 s or 1.00 ms
ranges. henry of inductance produces a counter- (b) Substituting t = 2.00 ens in Equation
Capacitance is important in ac circuits, potential of one volt when the rate of change Differentiating with ~pect to time t yields 2-27 reveals
particularly because a voltage that varies with of current is one ampere per second. Induc-
time gives rise to a time-varying charge-that tors employed in electronic circuits have 0= dq/~t + R ~ . (2-26) . 10.0 -200/100
C dt 1= lOOOe . .
is, a current. This behavior is seen by differen- inductances that range from a few pH
tiating Equation 2-20 to give (microhenry) to several H or more. Here again, we have used lower case letters to = 1.35 X 10-3 or 1.35 mA
Figure 2-8b shows a series RL circuit, represent instantaneous charge and current.
dq = Cdv (2-21) which contains a battery ~,a resistor R, and As noted earlier, dq/dt = i. Substituting - We find from Equation 2-28 that
dt dt an inductor L connected in series. The magni- this expression into Equation 2-26 yields, VR = 10.0e-2.00/1.00 = 1.35 V
By definition (Equation 2-1 ~ the current i is tude of L depends upon the number of turns upon rearrangement,
in the wire coil. and by substituting into Equation 2-29
the rate of change of charge; that is, dq/dt = i. di dt
Thus, As was pointed out earlier, a capacitor -= - RC
V(" = 10.0 (I - e-2.0011.00) = 8.65 V
dv stores energy as an electric field across a
i= C dielectric; an inductor, on the other hand, Integration between the limits of the initial
dt The center illustration in figure 2-8a
stores energy as a magnetic field surrounding current link and i gives
It is important to note that the current in a shows the changes in i, 11K, and Vc that occur
a conductor. The electric field in the capacitor
capacitor is zero when the voltage is time i= linite-·/Rc (2-27) during the charging cycle of an RC circuit.
is proportional to the applied voltage; the
independent-that is, for a direct current. Be- These plots were based upon the data given
magnetic field in the inductor is proportional In order to obtain a relationship between
cause a direct current refers to a steady state, in the example just considered. Note that VR
to the current that exists in the inductor. the instantaneous voltage across the resistor,
the initial transient current that charges the and i assume their maximum values the in-
The difference between a capacitor and an Ohm's law is emplo)led to replace i and lini'
condenser is of no significance in considering stant the switch in Figure 2-8a is moved to 1.
inductor can also be seen by comparing in Equation 2-27. Th~s,
the overall effect of a capacitor on a current At the same instant, on the other hand, the
Equations 2-22 and 2-23. The former indi-
that has a de component. cates that a change in potential across a capac' VR = Y;e-·/RC (2-28) voltage across the capacitor incr~ rapidly
Inductance. It is found experimentally that from zero and ultimately approaches a con-
itor results in a current; the latter shows that Substitution of this expression into Equation
a magnetic field surrounds any conductor as stant value. For practical purpOses, a capaci-
a change in current through an inductor 2-24 yields, upon rearrangement,
it carries an electric current. The work tor is considered to be fully charged after
causes the development of a potential. Thus,
required to establish this field manifests itself Vc = Y;(1 - e-·/RC) (2-29) 5RCs have elapsed. At this point the current
the functions of the two components in an
will have decayed to less than 1 % of its initial It is useful to compare the behavior of RC
value (e-5 = 0.0067 ~ 0.01). and RL circuits during a variation in signaL
When the switch in Figure 2-8a is moved The plots in Figure 2-8 illustrate that the two
to position 2, the battery is removed from the reactances are alike in the sense that both
circuit and the capacitor becomes a source of show a potential change which is out of phase
R "
VR
15kn ~ \, >
current. The flow of charge, however, will be
in the opposite direction from what it was
previously. Thus,
dqldt = -i
with the current. Note again the complemen-
tary behavior of the two; for an inductor, the
voltage leads the current, while for a capaci-
tor, the reverse is the case.
2OV"" v,

-+- Vc
C
O.OO8I'F
~
a
-
0

-0.5
\
1,\
0.75 kHz"
,
"
o E
~

The initial potential will be that of the Response of


battery. That is, Series RC and RL
VC=V; Circuits to Sinusoidal Inputs

Employing these equations and proceeding as In the sections that follow, the response of
series RC and RL circuits to a sinusoidal ac
in the earlier derivation, we find that for the
discharge cycle voltage signal will be considered. The input
signal is described by Equation 2-17; that is,
. Vc' -t/RC
1= -lie (2-30) v. = v" sin cot = V" sin 2nft (2-35)
Voltage, Current, and Phue Relationship
VR = - Vce-t/RC (2-31) for Series RC Circuit. Application of Kirch-
and holrs voltage law to the circuit shown in
Vc = Vce-t/RC
Figure 2-9a leads to the expression
(2-32)
v. = Vc + VR = V" sin cot
The bottom plot of Figure 2-8a shows how
these variables change with time. Substitution of Equations 2-20 and 2-4 yields
It is important to note that in each cycle, FIGURE 2-9 Current, voltage, and phase relationships for an RC circuit:
the change in voltage across the capacitor is v" sin cot = ~ + Ri (a) Circuit, (b) currents at 0.75 and 7.5 kHz, (c) voltages at 0.75 kHz, and
out of phDse with and lags behind that of the (d) voltages at 7.5 kHz.
current and the potential across the resistor. Differentiating and replacing dqldt with
Rate of Current and Potential Change gives
Across an RL Circuit. Equation 2-23 and
techniques similar to those in the previous cov" cos cot = Ci + Rdidi The peak current occurs whenever the cies and capacitance, on the other hand, ep
section can be used to derive a set of expres- sin(wt + ep) is unity. Thus, becomes negligible and, for all practical pur-
sions for the RL circuit shown in Figure 2-8b The solution of the differential equation can poses, the current and input voltage are in
that are analogous to Equations 2-27 through be shown to be5 V phase. Under these circumstances, (llwe) be-
2-32. For example, when the switch is closed
I - I' (2-38)
, - JR2 + (1/coC)2 comes negligible with respect to R and Equa-
to position 1, i= JR 2
+
~
l/coe)
2 sin(cot + ep) (2-36) tion 2-38 reverts to Ohm's law.
VR = V;(1 - e-·R1L) (2-33) A comparison of Equations 2-36 and 2-35
where the phase angle ep is given by reveals that the current in a series RC circuit
1 is often not in phase with the input voltage EXAMPLE
tan ep=-- signaL This condition is likely to prevail AssuUJe the components in the RC circuit
coRC
whenever the frequency (or the capacitance) shown in Figure 2-9a have values of
where LIR is the time constant for the circuit. is small; then, tan ep and thus ep (Equation V" = 20V,R = 1.5 x 104n,and C= 0.OO80pF.
These relationships for a typical RL circuit , J. J. Brophy, Basic Electronics/or Scientists. New York: 2-37) becomes significant with respect to cot (a) Calculate the peak current (1) if
are shown in Figure 2-8b. McGraw-Hil~ 1966. pp. 57-59. in Equation 2-36. At sufficiently high frequen- f = 0.15 kHz, (2) f = 75 kHz, and (3) with the
capacitor removed from the circuit. (b) Calcu- and at 75 kHz
late the phase angles between the voltage and
2.65 x W
the current at the two frequencies. arc tan q, = arc tan 1.5 x 104
(a) (1)
= arc tan 0.0177 = 1.0 deg R
1.1 1 «
E
we = 2~jC = 2n x 0.75 x 103 x 8.0 X 10-9
Figure 2-9b, c, and d shows the instanta- 20V'" v, r::

neous current, voltage, and phase relationships ~


= 2.65 X 104 n U
:>
when the circuit characteristics are the same L
Note that the quantity l/2njC has the dimen- as those employed in the example. The time
sions of ohms. That is, scales on the abscissas differ by a factor of
1 1 100' to accommodate the two frequencies.
jC=s-I·C·V-I From Figure 2-9b and the example, it is evi-
dent that at the lower frequency, the current
But the number of coulombs C is equal to the is significantly smalIer than at the higher be-
product of the current in amperes and the cause of the greater reactance of the capacitor
time in seconds. Thus, at the lower frequency. This larger reactance
is also reflected in the large phase difference
_1_ 1_!_n between the current and the input voltage
jC-s-I·A·s·V-1- A- >
shown in Figure 2-9c (q, = 60.5 deg). At
Employing Equation 2-38, we write 75 kHz, this difference becomes considerably f
-5
0
sm~ler (Figure 2-9d). >
1 20 Figures 2-9c and 2-9d show the relation-
"= J(1.5 x W)2 + (2.65 x 104)2 ship between the input voltages and the volt-
= 6.6 x 10-4 A age· drops across the capacitor and the
resistor. At the lower frequency, the drop ~ ~
(2) across the capacitor is large with respect to FIG UR E 2-10 Current, voltage, and phase relationships in an RL circuit:
that across the resistor. In contrast, at (a) Circuit, (b) currents at 20 and 200 kHz, (c) voltages at 200 kHz, and
1 1
75 kHz, lie is small, indicating little reaction (d) voltages at 20 kHz.
2njC = 2n x 75 x 103 x 8.0 X 10-9
between the current and the capacitor. At
= 2.65 X 102 n frequencies greater than 1 MHz, essentialIy
no interaction would occur, and the arrange-
ment would behave as a purely resistive
1- 20 circuit. in an inductor opposes the current; thus the The peak current occurs where R sin (rot + q,)
" - J(1.5 x W)Z + (2.65 x W? Note also the phase relationships among polarity of Ri and L di/dt must be the same. is equal to unity, or
= 1.33 x 10-3 A t:•• vc, and VR shown in Figure 2-9c. Integration of this equation yields6
I _ V"
(3) In the absence of C
Voltage, Current, and Phase Relationships
for Series RL Circuits. Applying Kirchhoff's
.=
I J R 2 +V" (roL) 2 SID A.)
. (rot + 'I' (2 - 39) ,,- JR 2 + (roL)2

20 -3
voltage law to the circuit in Figure 2-10 gives Figure 2-10 shows the relationships among
I" = 1.5 X 104 = 1.33 x 10 A currents, voltages, and phases for an RL
v. = VR + VL = V" sin rot
circuit. Comparison with Figure 2-9 again
(b) Employing Equation 2-37, we find at SUb~tituting Equations. 2-4 and 2-23 yields roL
tanq,=- reveals the complementary nature of the two
0.75 kHz. R types of reactive components. For inductors,
V, sin rot = Ri + L di/dt
2.65 x W reaction to the current occurs at high
tan q, = 1.5 X 104 = 1.77 The second term on the right-hand side of the frequencies but tends to disappear at low.
equation is positive because the minus sign in Here, the current lags the input potential
• J. J. Brophy. Basic Eleclronicsfor Scientists. New York:
q, = 60.5 deg Equation 2-23 means that the induced voltage McGraw-Hili. 1966. pp. 67-69. rather than leading it.
Capacitative .nd Inductive Reactance; Im- cuits containing capacitive and inductive ele- ----
pedance. An examination of Equations 2-41
and 2-38 shows a similarity to Ohm's law,
with the denominator terms being an expres-
sion of the impedance exerted by the circuit
to the flow of electricity. Note that at high
ments. A convenient way of visualizing these
effects is by means of vector diagrams.
Vector Diagrams for Reactive Circuits.
Because the voltage lags the current by
90 deg in a pure capacitance, it is convenient
XL

0 ~
z
R

Za .JR2+xL2
XL
:
I
I

frequencies, where l/wC ~ R, Equation 2-38 to let q, equal -90 deg for a capacitive reac- ~=ilfCtanR

reverts to Ohm's law; similarly, when tance. The phase angle for a pure inductive
wL ~ R, Equation 2-41 behaves in the same circuit is +90 deg .. For a pure resistive
way. circuit, q, will be equal to 0 deg. The relation-
The term l/wC in the denominator of ship ainong Xc, XL' and R can then be rep-
Equation 2-38 is termed the capacitive reac- resented vectorially as shown in Figure 2-11.
tance Xc, where As indicated in Figure 2-11c, the im-
pedance of a series circuit containing a
resistor, an inductor, and a capacitor is
Similarly, the inductive reactance XL is defined determined by the relationship
as (see Equation 2-41)
Z = JR2 + (XL - Xc>2 (2-46)
XL = wL = 27tjL (2-43) z· .JR' +(XL -Xci'
Because the outputs of the capacitor and (X, - Xci
o :lIE arc tan -R--
The impedance Z of the two circuits under inductor are 180 deg out of phase, their
discussion is a measure of the total effect of combined effect is determined from the differ-
resistance and reactance and is given by the ence of their reactances.
denominators of Equations 2-38 and 2-41. FIGURE 2-11 Vector diagrams for series circuits:
Thus, for the RC circuit EXAMPLE (a) RC circuit, (b) RL circuit, (c) series RLC circuit.
For the following circuit, calculate the
Z=JR2 +X~ peak current and voltage drops across the
and for the RL circuit three components. ring at different times. The sum .of the instan- Since the voltage drop across the resistor is in
Z = JR2 + xl (2-45) taneous voltages, however, would add up to phase with the current,
10V.
Substitution of Equations 2-42 and 2-44 High-Pass .nd Low-Pass Fikers. Series RC
1 = (VI').
, R
into Equation 2-38 yields and RL circuits are often used as filters to
attenuate high-frequency signals while pas- The ratio of the peak output to the peak
1 = V"
, Z sing low-frequency components (a low-pass input voltage is obtained by dividing the first
filter) or, alternatively, to reduce low- equation by the second and rearranging.
A similar result is obtained when Equations frequency components while passing the high Thus,
2-43 and 2-45 are substituted into Equation Z = J(SO)2 + (40 - 2W = 53.8 n (a high-pass filter). Figure 2-12 shows how
2-41.
series RC and RL circuits can be arranged to
In one sense, the capacitive and inductive 1" = 10 V/53.8 n = 0.186 A
give high- and low-pass filters. In each case.
reactances behave in a manner similar to that VR = 0.186 x SO = 9.3 V the input and output are indicated as the volt-
of a resistor in a circuit-that is, they tend to
Vc = 0.186 x 20 = 3.7 V ~es (Vp)i and (Vp) •. A plot of this ratio as a function of frequency
impede the flow of electrons. They differ in ; In order to employ an RC circuit as a
two major ways from resistance, however. is shown in Figure 2-13a; here, a resistance of
VL = 0.186 x 40 = 7.4 V high-pass filter, the' output voltage is taken
First, they are frequency dependent; second, 1.0 x 10· n and a capacitance of 0.10 pF was
across the resistor R. The peak current in this
they cause the current and voltage to differ in Note that the sum of the three voltages employed. Note that frequencies below
circuit is given by Equation 2-38. That is,
phase. As a consequence of the latter, the (20.4) in the example exceeds the source 50 Hz have been largely removed from the
phase angle must always be taken into voltage. This situation is possible because the I _ (Vp); input signal. It should also be noted that be-
account in considering the behavior of cir- voltages are out of phase, with peaks occur- I' - JR2 + (l/wC)2 cause the input and output voltages are out of
resistor -behavior just opposite of that of the energy stored in the magnetic field of the
RC circuit. Curves similar to Figure 2-13 are inductor during the other. Thus, during a
obtaincd for RL filters. one-half cycle, the energy of the capacitor
Low- and high-pass filters are of great im- forces the current through the inductor; in
portance in the design of electronic cirCUits. the other half cycle, the reverse is the case. In
~-----~ principle, a current induced in a closed reso-
Resonant Circuits nant circuit would continue indefinitely except
for the loss resulting from resistance in the
A resonant or RLC circuit consists of a
leads and inductor wire.
resistor, a capacitor, and an inductor ar-
The resonant frequency I. can be readily
ranged in series or parallel, as shown in
calculated from Equations 2-42 and 2-43.
Figure 2-14.
That is, whenf=f., XL = Xc, and
Series Resonant Filters. The impedance of
the resonant circuit shown in Figure 2-14a is 1
given by Equation 2-46. Note that the im- 2nf.C = 2nf.L
pedance of a series inductor and capacitor
is the difference between their reactances and Upon rearrangement, it is found that
that the net reactance of the combination is
1
zero when the two are identical. That is, the
1.= 21tJfC
FIGURE 2-12 Filter circuits: (a) High- The follQwing example will show some of the
pass RC filter, (b) low-pass RC filter, (c) high- 1.0 (bl properties of a series resonant circuit.
pass RL filter, and (d) low-pass RL filter. FIGURE 2-14 Resonant circuits employed i
O.B
as filters: (a) series circuit and (b) parallel EXAMPLE
~ 0.6
~ circuit. Assume the following values for the com-
~
phase, the peak voltages will occur at differ- :: 0." ponents of the circuit in Figure 2-14a: (Vp)i =
15.0 V (peak voltage), L= 100 mH, R = 20 n,
ent times. 0.2
and C = 0200 JLF. (a) Calculate the peak cur-
For the low-pass filter shown in Figure 0.0
only impedance in the circuit is that of the rent at the resonance frequency and at a
2-12b, we may write 1 100
resistor or, in its absence, the resistance of frequency 75 Hz greater than the resonance
Frequency. Hz
(Vp). = IpXc the inductor coil and ·the other wiring. This frequency. (b) Calculate the peak potentials
(al
behavior becomes understandable when it is across the resistor, inductor, and capacitor at
Substituting Equation 2-42 gives
recalled that in a series RC circuit, the poten- the resonant frequency.
(Vp). = Ip/(roC) tials across the capacitor and the inductor are (a) We obtain the resonant frequency with
O.B
180 deg out of phase (see the vector diagrams Equation 2-49. Thus,
Multiplying by Equation 2-38 and rearrang-
ing yields
~ 0.6 in Figure 2-11). Thus, if Xc and XL are alike,
"i
there is no net potential drop across the pair. 1
;0- 0.4
1.;= 2nJl00 x 10 x 0.2 x 10
(Vp)., 1 1 Electricity continues to flow, however; there- 3 6

(VP)i = roCJR2 + (l/roC)2 = j(roCR)2 + 1 fore, the net reactance for the combination is
= 1125 Hz
(2-48) zero. Also, when XL and Xc are not identical,
their net reactance is simply the difference be- By defidition, XL = Xc at!.; therefore, Equa-
Figure 2-13b was obtained with this equation. Frequency. Hz
tween the two. tion 2-46 becomes
As shown in Figure 2-12, RL circuits can (b) The condition for resonance is
also be employed as filters. Note, however, FIGURE 2-13 Frequency response of (a) the
that the high-pass filter employs the potential XL= Xc
high-pass filter shown in Figure 2-118 and (b) the
across the reactive element while the low-pass low-pass filter shown in Figure 2-12b; R = Here, the energy stored in the capacitor
filter employs the potential across the 1.0 X 104 nand C = 0.10 JLF. during one half cycle is exactly equal to the
To calculate I, at 1200 Hz, we substitute ing upon the relationship between the widtp
Equations 2-42 and 2-43 into Equation 2-46. of the pulse and the time constant for the
Thus, circuit. These effects are illustrated in Figure
Z = JR 2 + [27tjL - (1/27tfC)j2 2-17 where the input is a square wave having
a pulse width of 1",' seconds. The second
= J202 + [27t x 1200 x 100 x 10
column shows the variation in capacitor
- 1/(27t x 1200 x 0.200 x lO-ti)j2 potential as a function of time, while the third
column shows the change in resistor potential
= J400 + (754 - 663)2 = 93 0
at the same times. In the top set of plots
and I, = 15.0/93 = 0.16 A. (Figure 2-17a~ the time constant of the circuit
(b) At f. = 1125 Hz, we have found that is much greater than the input pulse width.
I,= 0.75 A. The potentials across the Under these circumstances, the capacitor can
resistor, capacitor, and inductor (V,)Il' (V,)L, become only partially charged during each
and (J-;,)c are pulse. It then discharges as the input potential
returns to zero; a sawtooth output results.
(V')ll = 0.75 x 20 = 15.0 V 1050 1100
The output of the resistor under these circum-
Frequency, Hz . Frequenc:y. Hz
(V,)L = I,XL FIGURE 2-15 Frequency response of a stances rises instantaneously to a maximum
FIGURE 2-16 Ratio of output to input value and then decreases essentially linearly
= 0.75 X 27t x 1125 x 100 x 10-3 series resonant circuit; R = 20Q,L = l00mH, voltage for the series resonant filter; L =
=
C 0.2 IlF, and (J-;,), = 15.0 V. 100 mH, C = 0.20 IlF, and (v,,)i = 15 V.
during the pulse lifetime.
= 530 V The bottom set of graphs (Figure 2-17c)
Since XL = Xc at resonance, illustrates the two outputs when the time con-
stant of the circuit is much shorter than the
f If the output of the circuit shown in Figure filter (Equation 2-46). Note that the im-
pulse width. Here, the charge on the capacitor
2-14a is taken across the inductor or capaci- pedance in the latter circuit is a minimum at
rises rapidly and approaches full charge near
Part (b) of this example shows that, at tor, a potential is obtained which is several resonance, when XL = Xc. In contrast, the the end of the pulse. As a consequence, the
resonance, the peak voltage across the re- times larger than the input potential (see part impedance for the parallel circuit at reso-
potential across the resistor rapidly decreases
sistor is the potential of the source; on the b of the example). Figure 2-16 shows a plot of nance is a maximum and in principle is infinite.
to zero after its initial rise. When V;goes to
other hand, the potentials across the Inductor the ratio of the peak voltage across the induc- Consequently, a maximum voltage drop
zero, the capacitor discharges immediately;
and capacitor are over 35 times greater than tor to the peak input voltage as a function of across (or a minimum current through) the the output across the resistor peaks in a nega-
the input potential. It must be realized, frequency. A similar plot is obtained if the parallel reactance is found at resonance. With
tive direction and then quickly approaches
however, that these peak potentials do not inductor potential is replaced by the capaci- both parallel and series circuits at resonance,
zero.
occur simultaneously. One lags the current by tor potential. a small initial signal causes resonance in These various output wave forms find
90 deg and the other leads it by a similar The upper curve in Figure 2-16 was ob- which electricity is carried back and forth be-
applications in electronic circuitry. The
amount. Thus, the instantaneous potentials tained by substituting a 10-0 resistor for the tween the capacitor and the inductor. But cur- sharply peaked voltage output shown in
across the capacitor and inductor cancel, and 20-0 resistor employed in the example. rent from the source through the reactance is Figure 2-17c is particularly important in
the potential drop across the resistor is then Parallel Resonant Filters. Figure 2-14b minimal. timing and trigger circuits.
equal to the input potential. Clearly, in a cir- shows a typical parallel resonant circuit. The parallel circuit, sometimes called a
cuit of; this type, the capacitor and inductor Again, the condition of resonance is that tank circuit, is widely used as for tuning radio
may have to withstand considerably larger Xc = XL and the resonant frequency is given or television circuits. Tuning is ordinarily ac- SIMPLE
potenti;als than the amplitude of the input by Equation 2-49. complished by adjusting a variable capacitor ELECTRICAL MEASUREMENTS
voltage would seem to indicate. The impedance of the parallel circuit is until resonance is achieved.
Figure 2-15, which was obtained for the given by This section describes selected methods for
circuit shown in the example, demonstrates Behavior of RC
that the output of a series resonant circuit is a Z = JR2 + (XC-XL
XLXC )2 (2-50) Circuits with Pulsed Inputs
the measurement of current, voltage, and
resistance. More sophisticated methods for
narrow band of frequencies, the maximum of measuring these and other electrical proper-
When a pulsed input is applied to an RC
which depends upon the values chosen for L It is of interest to compare this equation with ties will be considered in Chapter 3 as well as
circuit, the voltage outputs across the capaci-
and C. the equation for the impedance of the series in later chapters.
tor and resistor take various forms, depend-
Meter
The Ayrton Shunt. The Ayrton shunt,

f1o-------AJV:J shown in Figure 2-19, is often employed to


vary the range of a galvanometer. The
example that follows demonstrates how the
'"
:\, '-

resistors for a shunt can be chosen.


~
..!!..'Aft
Shunt
EXAMPLE R, R. ·R, R.
3
Assume that the meter in Figure 2-19 reg- 2 4 t '.
isters full scale with a current of SO JlA; its
internal resistance is 3000 n. Derive values
for the four resistors such that the meter
(b) ranges will be 0.1, 1, 10, and 100 mA.
RCO!!T. For each setting of the contact, the circuit
is equivalent to two resistances in parallel. FIGURE 2-19 A current meter with an
One resistance is made up of the meter and Ayrton shunt.
the resistors to the left of the contact in
(el T+l .JnLJnLJnL T+ !7l
J
:Tl iT1
LJ U L T+~' I -, I -, I
Figure 2-19; the other consists of the remain-

t ing resistor to the right of the contact. The


Vi 0 Vc 0
RC~T.
1.. 1.. _ 1.._
V" 0 -
potential across parallel resistances is, of We may also write, from Kirchholfs current
o course, the same. Thus, law
--Time- 0 --T~_ o --Time---"

IM(RM + Rd = IRRR
I", = 1M + IR
FIGURE 2-17 Output signals VR and Vc for pulsed input signal J-l; (a) time constant ~ pulse
width Tp; (b) time constant ~ pulse width; (c) time co~tant <l! pulse width. which can be combined with the first equa-
where RM is the resistance of the meter, R,. is tion to give, after rearrangement,
the resistance of the resistors to the left of the
contact, and RR is the resistance to the right. I", RM + RL + RR R", + Rr
1M = RR RR
G•• vanomet ••.• cause rotation of the coil and deflection of the
attached pointer; the deflection i~directly where Rr is the total resistance of the shunt
The classical methods for measuring direct proportional to the current present in the (RL + RR)'
current, voltage, and resistance employ a coil. Pennlnent With the contact in position 1, RL = 0 and
galvanometer. This device, which was in- Meters with D'Arsonval movements are magnet \ RR = Rr; here, the greatest fraction of the
vented nearly a century ago, still finds wide- generally called ammeters, milliammeters, or current will pass through the meter. There-
spread use in the laboratory. It is based upon microammeters, depending upon the current fore, setting 1 corresponds to the lowest cur-
the current-induced motion of a coil sus- range they have been designed to measure. rent range, where the full meter reading
pended in a fixed magnetic field. This arrange- Their accuracies vary from 0.5 to 3 % of full (/",= SO JlA) should correspond to 0.1 mA
ment is called a D'Arsonval movement or scale, depending upon the quality of the (Ix = 0.1 mAl. Substituting these values into
coil. meter. the derived equation yields
Figure 2-18 shows construction details of a More sensitive galvanometers are obtained 0.1 x 10- 3 A 3000 + 0 + RR
typical D'Arsonval meter. The current to be by suspending the coil between the poles of a SO X 10-6 A RR
measured passes through a rectangular coil ma~et by a fine vertical filament that has
wrapped about a cylindrical soft iron core very; low resistance to rotation. A small RR = Rr = 3000 n
that pivots between a pair of jeweled bear- mirror is attached to the filament; a light
ings; the coil is mounted in the air space be- beam, reflected ofT the mirror to a scale, Substituting this value for Rr in the work-
tween the pole pieces of a permanent magnet. detects and measures the extent of rotation
FIGURE 2-18 Current meter. ing equation gives
The current-dependent magnetic field that (Taken from R. J. Smith, Circuits, De-
during current passage. Galvanometers can
vices, and 51'stems. New York: Wiley, Ix R", + Rr 3000 + 3000
develops in the coil interacts with the per- be constructed that will detect currents as I", = RR RR
manent magnetic field in such a way as to small as 10-10 A. 1976. With permission.)
VAB across the resistor Rz is obtained by con- obtained. Here, it is desired to determine the resistance to total resistance of the remaining
. necting the meter at points A and B. Under current I in the resistor RAB when the switch components in the system .
RR = 6000 x IMII;" these circumstances, is thrown to position 1. Oearly, the error associated with this
From Ohm's law, this current is given by measurement can be minimized by choosing a
6 0.300 VAB = IMRM
= 6000 X 50 X 10- II" = -1-. Jr. Jr.s meter having the lowest possible resistance.
" . where RM is the resistance of the meter (and 1= S Alternatively, the meter can be replaced by a
any resistors placed in series with it) and 1M is R1 + Rz + Rs = Rr small precision resistor; the potential drop
the current measured by the meter. Since RM where Vs is the potential of the source, Rs is across this resistor can then be measured with
is a constant, the meter scale can be cal- its internal resistance, and Rr is the total re- a high resistance meter, and the current can
Setting I" ibrated in volts rather than amperes. sistance of the circuit. be calculated from Ohm's law.
Effect of Meter Resistance on Current In order to measure the current, however, Often, a current meter or a precision re-
1 0.1 X 10-3 3000 = (R, + R2 + R) + R.) Measurements. The coil of a meter or galva- sistor is made a permanent part of an instru-
2 1x 10- 3 300 = (R2 + R) + R.) the meter must be brought into the circuit by
nometer offers resistance to the flo~ of electric- moving the switch from position I to 2. ment; under these circumstances, a somewhat
3 10 x 10- 3 30= (R) + R.)
ity, and this resistance will reduce the Clearly, the measured current 1M will differ smaller current is obtained, but its magnitude
4 100 X 10-3 3 = R.
magnitude of the current being· measured; from I and will be given by is given directly by the meter.
thus, an error is introduced by t~e measure- Effect of Meter Resistauce on Voltage
Thus, R~ = 3 n, R3 = (30 - 3) = 27 n, Rz = ment process itself. This situation is not pecul- Vs Vs Meaurements. As shown in the following
III = ------- ---
(300 - 30) = 2700, and R1 = (3000 - 300) = iar to current measurements. In fact, it is a R1 + Rz + Rs + RM Rr +. RM example, the measurement of voltage by the
27000. simple example of a general limitation to any circuit in Figure 2-20b may also alter the
The relative error associated with the
Meaurement of Current and Voltage. Fi- physical measurement. That is, the process of system in such a way as to create a significant
presence of the meter is given by
gure 2-20 presents circuit diagrams showing measurement inevitably disturbs the system of error.
how a galvanometer can be employed to interest so that the quantity actually reI error = (IM - 1)/1
measure current and voltage in a circuit. Note measured differs from its value ~rior to the EXAMPLE
that the meter is placed in series with the re- measurement. This type of error cim never be Substitution of the relationships for I and 1M The voltage drop VAB across the resistor
mainder of the circuit for a current measure- completely eliminated; often, however, it can yields, upon rearrangement, Rz in Figure 2-20b is to be measured by
ment (Figure 2-208). In contrast, a voltage be reduced to insignificant proportions. means of a meter having a resistance of
measurement is obtained by placing the meter The circuit shown in Figure 2-208 can be
reI error = - RM I(Rr + RM)
50.0 1e0. (a) Calculate the relative error asso-
across or in parallel with the potential to be used to illustrate how a simple measurement The data in Table 2-1 illustrate the magnitude ciated with the measurement. (b) Repeat the
measured. Thus, in Figure 2-20b, the voltage can perturb a system and thus alter the result of the disturbance to the system by the mea- calculations assuming the meter resistance is
surement process for various ratios of meter still 50.0 kn, but that all other resistances

Vs
12.GV l00kn

IIII
Rs.,00n R,
A
TABLE 2-1 EFFECT OF METER RESISTANCE ON CURRENT
MEASUREMENTS·

Meter Resistance, Measured Current,

S 8
R", kn I". mA
2
meter absent 0.00 0.100 o
I om 0.099 -0.8
10 0.08 0.092 -7.6
w ~ 50 0.42 0.071 -29
100 0.83 0.055 -45
FIGURE 2-20 Circuits for measuring (a) current I and
(b) voltage VAS.
have values that are 1/10 that shown in the the second part of the example, however, the to 1 mA in the initial adjustment, we may
figure. resistance of the potential source plays an im- write
(a) With the switch S open so that the portant part in determining the meter resis- 1.5 = 11(RM + Rv) = 1.00 X 1O-3(RM + Rv)
meter is out of the system, the potential drop tance required for a given accuracy. That is, as Device
V.u can be obtained by use of Equation 2-8 the source resistance increases, so also must (RM + Rv) = 1500 n to give
known" V.
(p.9) -. the meter resistance if accuracy is to be main-
tained. This relationship becomes particularly FIGURE 2-22 Block diagram for a com-
V = 12.0 x 20.0 _ parison measurement system.
significant in the determination of pH by Rx= (~- 1)1500
AB 100 + 20.0 + 0.100 - 2.00 V
measurement of the potential of a cell con-
taining a glass electrode. Here, the resistance 3
When the switch is closed, a system of two =(1.00;110- -1)1500 As shown in Figure 2-22, a typical compar-
parallel resistors is created whose net resis- of t~ system may be 100 megohms or more.
Thus; a potential measuring device with an ison instrument consists of three components.
tance RAB can be found by substitution into
eno~ous internal resistance is required. Clearly, the relationship between the One of these is a null detector, which indi-
Equation 2-11 (p. 11). Thus,
Resistance Measurements. Figure 2-21 gives measured current and the resistance is cates the inequality or equality of the two sig-
1 1 1 I 1 a circuit for the measurement of resistance nonlinear. nals. The second component is a reference
-=-+-=-+-
RAB R R 20.0 50.0 standard signal against which the unknown
1 AI with; a D'Arsonval meter. The source is
usually a 1.5-V dry cell. To employ the signal is compared. Finally, the instrument
RAB = 14.3 ill
circuit, the switch is placed in position I and
Comparison M ••••• rementa
contains a device for continuously attenuat-
The potential drop VAl, sensed by the meter, Comparison or Null Measurements. Instru- ing one of the two signals in such a way that
the meter is adjusted to full scale by means of
is now ments based upon comparison of the system the percent reduction is exactly known. In
the v;uiable resistor Rv • From Ohm's law, we
may write of interest with a standard are widely Figure 2-22, attenuation is performed on the
v. = 12.0 x 14.3 _ employed for chemical measurements; such reference signal; often, however, it is the un-
AI 100 + 14.3 + 0.100 - 1.50 V 1.5 = 11(RM + Rv) devices are called null instruments. Null in- known signal that is varied. Furthermore,
Therefore, struments are generally more accurate and attenuation may be performed mechanically
With: the switch moved to position 2, the cur-
precise than direct-reading instruments, in rather than electrically. An example is a var-
1.50 - 2.00 rent decreases to 11 because of the added re-
part because a null measurement generally iable diaphragm that reduces the power of a
re I error = ---- x 100 = -25% sistance of the unknown Rx• Thus,
2.00 causes less disturbance to the system being beam of light used for photometric analysis.
1.5 = 12(RAt + Rv + Rx)- studied. Moreover, the readout devices Potentiometers. The potentiometer is a
(b) Proceeding as in (a~ we find
employed can often be considerably more null instrument that permits accurate measure-
Division of the two equations and rearrange-
V = 12.0 x 2.00 _ rugged and less subject to environmental ment of potentials while drawing a minimum
ment yields
AB 10.0 + 2.00 + 0.0100 - 2.00 V effects than a direct-reading device. current from the source under study. A typi-.

1 1 1 Rx=(~ -1)(RM+Rv)
Ii= 2.0 + 50.0
The value of (RAt + Rv) is determined by
the range of the meter. For example, if the TABLE 2-2 EFFECT OF METER RESISTANCE ON POTENTIAL
R = 1.92 kCl meter range is 0 to I mA and the meter is set MEASUREMENT-
V. = 12.0 x 1.92 _
M 10.0 + 1.92 + 0.0100 - 1.93 V Meter Resistance, Measured Voltage, Relative Error,

I 1.93 - 2.00 RM,kCl VM. V %


re error = 2.00 x 100 = -3.5%
10 0.75 -62
SO 1.50 -25
Table 2-i shows the effect of meter resis- 1.935
500 -32
tance on a potential measurement. Note that, 1 X 103 1.967 -1.6
in contrast to a current measurement the 1 X 104 1.997 -02
accuracy of a potential measuremen~ in- FIGURE 2-21 A circuit for measurement
creases with meter resistance. As shown by of resistance.
callaboratory potentiometer employs a linear V. R The slide wire shown in Figure 2-22 is current of 1 pA (10-6 A), we can readily ~_
voltage divider such as that shown in Figure ·+10--1- ~w.. readily calibrated to read directly in volts, culate from Ohm's law that the minimum
2-3b (p. 10) to attenuate a reference voltage B thus avoiding the necessity of calculating v,. distinguishable voltage difference will be
until a null point is reached. In its simplest each time. Calibration is accomplished by set- 10-6 X 1000 = 10-3 V, or 1 mY. By use of a
0.0 0.5 1.0 1.5 2.0
form, the divider consists of a uniform resis, I, I I I II • I I II I I I II , I .I
ting the slide wire to the voltage of the stand- galvanometer sensitive to 10 - 7 A, a difference
tance wire mounted on a meter stick. A sliding B ard cell. Then, with that cell in the circuit, of 0.1 mV will be detectable. A sensitivity of
contact permits variation of the output
Lineor vollllge the potential across AB is adjusted by means this order is found in an ordinary pointer-
divider
voltage. More conveniently, the divider is a of the variable resistor R until a null is type galvanometer; more refined instruments
precision wire wound resistor formed in a hel- achieved. The slide-wire setting will then give with sensitivities up to 10-10 A are not
ical coil. A continuously variable tap that can V" directly when C is adjusted to the null posi- uncommon.
be moved from one end of the helix to the tion with the unknown voltage. From this discussion, it is clear that the
other gives a variable voltage. Ordinarily, the The need for a working battery B may be sensitivity of a potentiometric measurement
divider is powered by dry cells or mercury questioned. I~ principle, there is nothing to decreases as the electrical resistance of the cell
batteries, which provide a potential somewhat prevent a direct measurement of v,. by replac- increases. As a matter of fact, potentials of
larger than that which is to be measured. ing B with die standard cell V,. It must be cells with resistances much greater than
Figure 2-23 is a sketch of a laboratory remembered, thowever, that current is being 1 Mn cannot be measured accurately with a
potentiometer employed for the measurement continuously drawn from B; a standard cell potentiometer employing a galvanometer as
FIGURE 2-23 Circuit diagram of a lab-
of an unknown potential v,.
by comparing it
oratory potentiometer.
would not maintain a constant potential for the current-sensing device.
with a standard potential V,. The latter is long under such usage. Current Measurements with a NuU In-
ordinarily a Weston standard cell which can The accur!lCY of a. voltage measurement strument. The null method is readily applied
be represented as follows: with a potentiometer equipped with a linear to the determination of current. A small
voltage dividcjr depends upon several factors. precision resistor is placed in series in the cir-
Cd(Hg)"ICdS04 • (8/3)HzO(sat'd~ from the battery B through C and the circuit Thus, it is nel:essary to assume that the volt- cuit and the potential drop across the resistor
containing G and V,. Electrons will flow in
Hg2S04(sat'd) I Hg the opposite direction, of course, if VAC is
age of the working battery B remains con- is measured with a potentiometer.
stant during the time required to balance the Resistance Measurements; the Wheatstone
One of the electrodes is a cadmium amalgam smaller than V,. The position of contact C is BrilIge. The Wheatstone bridge shown in
instrument against the standard cell and to
represented as Cd(Hg),,; the other is mercury. then varied until a null condition is achieved, measure the potential of the unknown cell. Figure 2-24 provides another example of a
The electrolyte is a solution saturated with as indicated by the galvanometer when K Ordinarily, this assumption does not lead to null device, applied here to the measurement
cadmium and mercury sulfates. The half reac- is closed.
appreciable error if B consists of one or two of resistance. The power source S provides an
tions as they occur at the two electrodes are The foregoing process is repeated with S in heavy-duty dry cells (or mercury cells) in
position 2 so that the unknown cell is in the good condition or a lead storage battery. The
Cd(Hg)" ...•Cd2+ + xHg + 2e circuit. Equation 2-9 can then be employed to instrument should be calibrated against the
HgZS04 + 2e ...•2Hg + SO~- calculate v,..
Thus, with S at position 1 standard cell before each voltage measure-
The potential of the Weston cell at 2SoC is ment to compensate for possible changes in
V. = Yo AC.
1.0183 V. • B AB B, however.
A potential measurement with the instru- The linearity of the resistance AB, as well
ment shown in Figure 2-23 requires two null- and at position 2 as the precision with which distances along its
point measurements. Switch P is first closed length can ~ estimated, both contribute to
V. = Yo AC" the accuracy of the potentiometer. Ordinarily,
to provide a current in the slide wire and a " B AB
potential drop across AC. The Weston cell is however, the, ultimate precision of a good-
next placed in the circuit by moving switch S Dividing these equations gives, upon quality instniment is determined by the sensi-
to position 1. When the tapping key K is rearrangement, tivity of the galvanometer relative to the
closed momentarily, a current will be in- resistance of the circuit. Suppose, for example,
that the electrical resistance of the galva-
dicated by the galvanometer G, unless the
output from the standard cell is identical with nometer plus the unknown cell is 1000 n, a
'"s
the potential drop VAC between A and C. If Thus, v,.
is obtained from the potential of the typical figure. Further, if we assume that the FIGURE 2-24 A Wheatstone bridge for
VAC is greater than V" electrons will flow standard cell and the two slide-wire settings. galvanometer is just capable of detecting a resistance measuremet1ts.
Be current at a potential of 6 to 10 V. The To measUre R", the position of C is ad- 4. A voltmeter was employed to measure the potential of a cell having
resistances RAC and Rc" are determined from justed to a minimum as indicated by the null an internal resistance of 750 O. What must the internal resistance of
the total resistances of the linear voltage detector. Application of Equation 2-9 to the the meter be if the relative error in the measurement is to be less than
divider, AB, and the position of C. The upper voltage divider ACB yields (a) -1.0~", (b) -0.10%?
right arm of the bridge contains a series of 5. For the following circuit, calculate
precision resistors that provide a choice of
AC
VAC= VAI/- (a) the potential drop across each of the resistors.
several resistance ranges. The unknown resis- AB (b) the magnitude of each of the currents shown.
tance R" is placed in the upper-left arm of the For the circuit ADB, we may write ·(c) the power dissipated by resistor R) .
bridge. (d) the potential drop between points 3 and 4.
A null detector N D is employed to indicate R"
VAD=VA"-R
,,+ R •
R, 2 3

-" '·1 -
an absence of current between D and C. The
detector often consists of a pair of ordinary
But at nul~ VAC= VAD.Thus, the two equa- +
= R, -loon
'3
earphones; an ac signal of about 1000 Hz is tions can be combined to give, upon 15.0V R. R3 R.· soon
employed, the human ear being sensitive in rearrangement, .!!... R3·200n
this frequency range. Alternatively, the detec-
~ R.-l000n

tor may be a cathode-ray tube or an ac 4 R. 5 6


microammeter.
6. For the circuit shown below, calculate
(a) the power dissipated between points 1 and 2.
1. It was desired to assemble the voltage divider shown below. Two of .(b) the current drawn from the source.
each of the following resistors were available: 50 0. 100 0, and 200 0. :(c) the potential drop across resistor RA•
[(d) the potential drop across resistor RD'
10.0 V
;(e) the potential drop between points 5 and 4.
+ 11-
V.
R, R. R3

R.-
RA- 1.0 kn
2.0 kn
Rc- 4.0 kn
Ro- 2.0 kn
V,-1.0V V.-4.0V V,
3 4 RE- 1.0 kn
2
-1111 +
(a) Describe a suitable combination of the resistors that would give 24 V
the indicated voltages.
(b) What would be the IR drop across R)? 7. Assume that the resistor shown between A and B in Figure 2-23
(c) What current would be drawn from the source? (p. 34) is a linear slide wire exactly 1.00 m in length. With the stand-
(d) What power is dissipated by the circuit? ard Weston cell (1.018 V) in the circuit, a null point was observed
2. Assume that for a circuit similar to that shown in Problem I, when contact C was moved to a position 84.3 cm from point A. When
Rj = 200 0, R2 = SOO0. R) = 1000 0. and V" = 15 V. - the Weston cell was replaced with an unknown voltage, null was
(a) Calculate the voltage V2• , observed at 44.3 cm. Calculate the potential of the unknown.
(b) What would be the power loss in resistor R2 ? 8. , A galvanometer with an internal resistance of 5000 0 shows a full
(c) What fraction of the total power lost by the circuit would be i scale reading with a current of 0.100 mA. Design an Ayrton shunt
dissipated in resistor R2 ? such that the gal\L8nometer will have a range of 0 to 10.0, 25.0, 50.0,
3. For a circuit similar to the one shown in Problem I, Rj = 1.00 ldl, and 100 mA.
R2 = 2.50 ko, R) = 4.00 ldl, and V" = 12.0 V. A voltmeter was 9. Assume that the slide wire in the Wheatstone bridge shown in Figure
placed across contacts 2 and 4. Calculate the relative error in the 2-24 has a resistance of SOOn and that the switch is in position 3.
voltage reading if the internal resistance of the voltmeter was (a) (a) What is the resistance of the unknown R" if a null is found when
5000 0. (b) 50 ill, and (c) 500 kO. contact C is at 62.6% of the full length of the slide wire?
(b) If the source potential S is 15 V and the resistance of the null n
18. A series RL circuit conSisted of a 12 Vsource, a 1000 resistor, and a
detector is 100 n, what is the current in the detector when C is set 2 H inductor.
at 62.7%? (a) Calculate the time constant for the circuit.
(c) If the switch is moved from 3 to 2, with Rx still in the circuit, (b) Calculate the current and the voltage drops across the resistor
where would C have to be moved in order to reach a new null? and inductor for the following times after the current was ini-
10. The current in a circuit is to be determined by measuring the poten- tiated: 0, 1,2, 4, 8, 10 IDS.
tial drop across a precision resistor in series with the circuit. 19. Calculate the capacitive reactance, the impedance, and the phase
(a) What should be the resistance of the resistor in ohms if 1.00 V is angle q, for the following series circuits.
to correspond to SO JlA?
Frequency, Hz R, 0 C,Ilf'
(b) What must be the resistance of the voltage measuring device if the
error in the current measurement is to be less than 1.0% relative? 1 20,000 0.033
11. An electrolysis at a nearly constant current can be performed with the 103 20,000 0.033
following arrangement: 106 20,000 0.033
I 200 0.033
103 200 0.033
106 200 0.033
1 2,000 0.33
103 2,000 0.33
106 2,000 0.33
20. Calculate the inductive reactance, the impedance, and the phase angle
q, for the following series circuits.
The 9O-V source consists of dry cells whose potential can be assumed Frequency, Hz R,O L,mH
to remain constant for short periods. During the electrolysis, the re-
3
sistance of the cell increases from 20 0 to 40 0 due to depletion of 10 2000 SO
ionic species. Calculate the percent change in the current assuming 10· 2000 SO
that the internal resistance of the batteries is zero. 105 2000 SO
12. Repeat the calculations in Problem II assuming that Ys = 9.0 V and 103 200 SO
R = 0.50 kO- 104 200 SO
105 200 SO
13. A 24-V dc potential was applied across a resistor and capacitor in
103 200 5
series. Calculate the current after 0.00, 0.010, 0.10, 1.0, and 10 sec if
104 200 5
the resistance was 10 MO and the capacitance 0.20 JlF.
IOS 200 5
14. How long would it take to discharge a oms JlF capacitor to I % of its
full charge through a resistance of (a) 10 MQ, (b) I MQ, (c) I kO? 21. Calculate the capacitive and inductive reactances and the impedance
IS. Calculate time constants for each of the RC circuits described in of the following series resonant circuits.
Problem 14. L,mH
Frequency, Hz R,O C,IlF
16. A series RC circuit consisted of a 2S-V dc source, a SO kO resistor,
102 2000 0.20 1.2
and a 0.035 IlF capacitor.
104 2000 0.20 1.2
(a) Calculate the time constant for the circuit.
106 2000 0.20 1.2
(b) Calculate the current and potential drops across the capacitor
102 20 0.20 1.2
and the resistor during a charging cycle; employ as times 0, 1,2,3,
104 20 020 1.2
4, 5, and 10 ms.
106 20 0.20 1.2 •
(c) Repeat the calculations in (b) for a discharge cycle.
17. Repeat the calculations in Problem 16 assuming that the potential 22. For the circuits described in Problem 21,
was 15 V, the resistance was 20 MQ, and the capacitance was (a) calculate the frequency at which a current maximum would
0.050 JlF. Employ as times 0, I, 2, 3, 4, 5, and 10 s. appear.
(b) calculate the current at this maximum if the input is a 15-V sinu-
soidal signal.
(c) calculate the potential drop across each of the components if the
input is a 15-V sinusoidal signal.
23. ~rive a frequency response curve for a low-pass RC filter in which
R';' 2.5 X 10' n and C •••0.015 pF. Cover a range of (Yp).!(Yp)/ of
0.01 to 0.99.
24. Derive a frequency response curve for a high-pass RC filter in which
R •••5.0 X 10' nand C •••200 pF (1 pF •••10-12 F). Cover a range of
om
(V,,).!(Yp)/ of to 0.99.
25. Derive a frequency response curve for an RL low-pass filter in which
R ••• 120 nand L ••• 8.0 H. Cover a range of (V,,).!(V,,)/ of 0.01 to 0.99.
Give a circuit design for an RC filter that would have the same
characteristics.
26. Determine the current and the voltage across !=&Chcomponent in a
series resonant circuit in which R'"' 100 n, C •••0.050 pF,
L'"' 225 mH, and (V,,)/'"' 12.0 V
(a) at the current maximum.
(b) at a frequency 10% lower than the maximum.
(c) at a frequency 10% higher than the maxim~m.
27. Calculate the capacitive and inductive reactances and the impedance
for parallel resonant circuits similar to that stfown in Figure 2-14b
assuming the values for f, R, C, and L given in Problem 21.
28. For the circuit shown in Figure 2-14b, calculate
(a) the frequency at which a current minimum would be expected.
(b) the current at a frequency that was 10% less than the minimum.
(c) the current at a frequency that was 10% more than the minimum.
A p-type or positive type semiconductor is p-type semiconductor is inherently less than
The circuit elements and circuits considered crysta~ each of these electrons is immobilized
formed when silicon is doped with a Group that of an n-type.
in Chapter 2 are linear in the sense that their by combination with an electron from
III element which contains only three valence
input and output voltages and currents are another silicon atom to form a covalent bond.
electrons. Here, positive holes are introduced
proportional to one another. Electronic cir- Thus, in principle, no free electrons' should Semiconductor Diodes
when electrons from adjoining. silicon atoms
cuits contain one or more nonlinear devices exist in crystalline silicon, and the material
jump to the vacant orbital associated with the A diode is a device that has greater conduct-
such as transistors, vacuum tubes, gas-filled should be an insulator. In fact, however,
impurity atom. Note that this process imparts ance in one direction than in another. Useful
tubes, and semiconductor diodes.l Histor. sufficient thermal agitation occurs at room
ically, electronic circuits were based upon temperature to liberate an occasional elec- a negative charge to the Group III atoms. diodes are manufactured by forming adjacent
vacuum tubes. Now, however, the vacuum Movement of the holes from silicon atom to 11- and p-type regions within a single germa-
tron, leaving it free to move through the crys-
tube has been almost totally replaced by semi- tal lattice and thus to conduct electricity. silicon atom, as described earlier, constitutes nium or silicon crystal; the interface between
conductor devices. The advantages of the This thermal excitation of an electron leaves a a current in which the majority carrier is these regions is termed a pn junction.
latter include low cost, low power consump- positively charged region, termed a hole, asso- positive. Positive holes are less mobile than ProPerties of. pn JUnctiolL Figure 3-la is
tion, small heat generation, long life, and ciated with the silicon atom. The hole, free electrons; thus, the conductivity of a a cros~ section of one type of pll junction,
small size. Most modern analytical instru- however, is mobile as wel~ and thus also
ments use only semiconductors; we shal~ contributes to the electrical conductance of
therefore, limit our discussion to these
devices.
the crystal. The mechanism of hole movement
is stepwise; a bound electron from a neigh-
,..---1 mm--.,
I I
I I
boring silicon atom jumps to the electron- I rr I
deficient region and thereby leaves a positive I +l l
IWire
hole in its wake. Thus, conduction by a semi- l Metal
I contact p.region
I
A semiconductor is a crystalline material conductor involves motion of thermal elec-
I \ I
having a conductivity between that of a con- trons in one direction and holes in the other.
ductor and an insulator. Many types of semi- The conduction of a silicon or germanium
conducting materials exist, including silicon crystal can be greatly enhanced by doping. a
and germanium. intermetallic compounds process by which a tiny, controlled amount of
(such as silicon carbide~ and a variety of an impurity is introduced into the crystal
organic compounds. Two semiconducting structure. Typically, a silicon semiconductor
materials that find wide application for elec- is doped with a Group V element such as
tronic devices are crystalline silicon and arsenic or antimony or a Group III element
germanium; we shall limit our discussions to such as indium or gallium. When an atom of
these. a Group V element replaces a silicon atom in
the lattice, one unbound electron is in-
Properties of Silicon troduced into the structure; only a small ther-
and Germanium Semiconductors mal energy is needed to free this electron for
Silicon and germanium are Group IV ele- conduction. Note that the resulting positive
ments and thus have four valence electrons Group V ion does not provide a mobile hole
available for bond formation. In a siliCQn inasmuch as there is little tendency for elec-
trons to move from a covalent silicon bond to
this nonbonding position. A semiconductor
I for further infonnation about modern electronics, see:
that has been doped so that it contains non-
A.J. Diefenderfer,Principles of Electronic Instrumentalion,
2d cd., Philadelphia: Saunders, 1979; J. J. Brophy, Basic bonding electrons is termed an n-type (nega-
ElectroniCS for Scientists, 3d cd. New York: McGraw. tive type) because electrons are the majority
Hill, 1977; H. V. Malmstadl, C. G. Enke, S. R. ca"iers of current. Positive holes still exist as (e) (d)
Crouch, and G. Horlick, Eleetronk Measurements for in the undoped crystal (associated with sili-
Screntlsrs.Menlo Park, CA: Benjamin, 1974;R. J. Smith,
con atoms~ but their number is small with FIGURE 3-1 A pn junction diode. (a) Physical appearance of
Circuits, Device3, and Systems. 3d cd. New York: Wiley. one type formed by diffusion of a p-type impurity into an n-type
1976; and B. H. Vassos and G. W. Ewing, Analog and respect to the number of electrons; thus, holes
represent minority carriers in an n-type semiconductor; (b) symbol for; (c) current under forward bias;
Digital Electronics for Scientists. New York: Wiley-
Interscience, 1972. semiconductor. (d) resistance to current under reverse bias.
we shall see, this phenomenon has practical transistors are; (1) small size; (2) heavilY
which is formed by dilTusing an excess of a
application in electronics. doped outer layers, the smaller of which is
p-type impurity such as indium into a minute
silicon chip that has been doped with an called the emitter and .the larger the collector;
n-type impurity such as antimony. A junction Transistors (3) a thin (- 0.02 mm~ lightly doped base
of this kind permits ready flow of positive layer that separates· the emitter and the
The transistor is the basic semiconductor am- collector; and (4) metallic contacts and elec-
charge from the p region through the n region
plifying device and performs the same func- trical leads to each of the three regions.
(or flow of negative charge in the reverse
tion as a vacuum amplifier tube-that is, it An alloy junction imp transistor, such as
direction); it olTers a high resistance to the
provides an output signal that is usually that shown in Figure 3-3a, is often fabricated
flow of positive charge in the other direction
significantly greater than the input. from a chip of n-type germanium. Small pel_
and is .thus a current rectifier.
Bipolar Transistors. Bipolar transistors lets of a Group III metal such as indium are
Figure 3-1b illustrates the symbol employed
consist of two back-to-back semiconductor fixed on both sides of the chip and heated
in circuit diagrams to denote the presence of
diodes. The pnp transistor consists of an until the metal melts and dissolves in the
a diode. The arrow points in the direction of
n-type region sandwiched between two p-type germanium. Upon cooling and crystallization,
low resistance to positive currents.
-I regions; the npn type has the reverse struc- two heavily doped p regions are formed on
Fi&/1re 3-lc shows the mechanism of con-
duction of electricity when the p region is FIGURE 3-2 Current voltage char- ture. Bipolar transistors are constructed in a the tw~ sides of the thin layer of n-doped
made positive with respect to the n region by acteristics of a silicon semiconductor variety of ways, two of which are shown in germanIUm.
application of a potential; this process is diode. Note that, fo~ the sake of Figure 3-3. The general features of all such A typical planar npn transistor, such as
called forward biasing. Here, the positive holes clarity, the small current under reverse
in the:p region and the excess electrons in the bias has been exaggerated.
n region (the majority carriers in both rr.-
gions)imove under the influence of the electric
field toward th~ junction, where they can
combine with and thus annihilate each other.
The negative terminal of the battery injects observed. Under reverse bias, a current on the
new electrons into the n region, which can order of microamperes flows over a consider-
then continue the conduction process; the able voltage range; in this region, conduction
positive terminal, on the other hand, extracts is by the minority carriers. Ordinarily, this
electrons from the p region thus creating new reverse current is of no consequence. As the
holes which are free to migrate toward the pn reverse potential is increased, however, a
junction. breakdown voltage is ultimately reached
When the diode is reverse-biased, as in where the reverse current increases abruptly
Figure 3-ld, the majority carriers in each to very high values. Here, holes and electrons,
region drift away from the junction to leave a formed by the rupture of covalent bonds of
depletion layer that contains few charges. the semiconductor, are accelerated by the
Only the small concentration of minority car- field to produce additional electrons and
riers present in each region drift toward the holes by collision. In addition, quantum p·layer
junction and thus carry a current. Con- mechanical tunneling of electrons through the -O.02mm
thick
sequently, conductance under reverse bias is junction layer contributes to the enhanced
typicalIy 10-6 to 10-8 that of conductance conductance. This conduction, if sufficiently (b) Id)
under forward bias. large, may result in heating and damaging of
Current-Voltage Carves for Semiconductor the diode. The voltage at which the sharp in- FIGURE 3-3 Two types of bipolar resistors. Construction
Diodes. Figure 3-2 shows the behavior of a crease in current occurs under reverse bias is details are shown in (a) for a pnp alloy junction transistor and
typical semiconductor diode under forward called the Zener breakdown voltage. By con- in (b) for an npn planar transistor. Symbols for a pnp and an
and reverse bias. With forward bias, the cur- trolling the thickness and type of the junction npn bipolar transistor are shown in (c) and (d), respectively.
rent increases nearly exponentially with layer, Zener voltages ranging from a few volts Note that aUoy junction transistors may also be fabricated as
voltage; often currents of several amperes are to several hundred volts can be realized. As npn types and planar transistors as pllp.
that shown in Figure 3-3b, is formed from a . After amplification, the dc component can
wafer of n-type silicon which has a surface then be removed by a high-pass filter (p. 23).
coating of silicon dioxide; this layer is The emitter-collector circuit is powered by
deposited by heating the wafer in an oxygen a dc source or. power supply that may consist
atmosphere. To fabricate the p-type base of a series of b~tteries or a rectifier. Typically,
region, a circle is etched in the oxide surface; the power supply will provide a potential be-
the crystal is then heated and exposed to va- tween 9 and 30 V.
porized boron to produce a lightly doped, Note that, as shown by the breadth of the
p-type region by diffusion. The surface film is arrows, the collector or output current Ie is
then reformed by oxidation, following which significantly larger than the base input cur-
a smaller circle is etched within the original rent lB' Furthermore, the magnitude of the
circle. This region is then heavily doped by collector current is directly proportional to
diffusion of phosphorus into the heated crys- the input current. That is,
tal to give an n-type emitter.
The symbols for the pnp and the npn type
of transistor are shown on the right in Figure
3-3. The arrow on the emitter lead indicates where the proportionality constant a is the
the direction of flow of positive charge. Thus, current gain, which measures the current
in the pnp type. positive charge flows from the amplification that has occurred. The collector
emitter to the base; the reverse is true for the current is also proportional to the emit.er
npntype. I current. Thus,
Electrical Charaderistics of a Bipolar
Transistor. The discussion that follows will
focus upon the behavior of a pnp-type bipolar where the proportionality constant C( is called Ie
transistor. It should be appreciated that the
the forward current transfer ratio. Since J;;
Current gain, a ,..
npn type acts analogously except for the
IE == IB + Ie (see Figure 3-4), it is readily FIGURE 3-4 Currents in a common-emitter
direction of the flow of electricity, which is
shown that cir~uit with a pnp transistor. Ordinarily, C( ~ 0.95
opposite.
to 0.995 or C( ~ 20 to 200.
When a transistor is to be used in an elec- C(
a=--
tronic device, one of its terminals is connected l-C(
to the input and the second serves as the
output; the third terminal is connected to Values for a for typical transistors range from
both and is the common terminal. Three 20 to 200. away from the junction, as shown in Figure the n layer. Thus, the holes in this pn junction
configurations are thus possible: a common- Mechanism of Amplification with a Bipolar 3-ld. have a current-carrying capacity that is per-
emitter, a common-collector, and a common- Transistor. It should be noted that the In the forward-biased pn junction il- haps one hundred times greater than that for
base. The common-emitter configuration has emitter-base interface of the transistor ,;hown lustrated in Figure 3-1c, the number of holes the electrons.
the widest application in amplification and is in Figure 3-4 constitutes a forward-biased pn in the p region is about equal to the number Turning again to Figure 3-4, it is apparent
the one we shall consider in detail. junction similar in behavior to that shown in of mobile electrons in the n region. Thus, that holes are formed at the p-type emitter
Figure 3-4 illustrates the current Figure 3-lc, while the base-collector region is except for the small difference in their mobili- junction through abstraction of electrons by
amplification that occurs when a pnp tran- a reverse-biased np junction similar to the cir- ties, conduction through the diode is shared the two dc sources, namely, the input battery
sistor is employed in the common-emitter cuit shown in Figure 3-1d. A current 18 more or less equally between the two types of and the power supply. These holes can then
mode. Here, a small dc input current IB, readily flows through the forward-biased charged bod.ies. Recall, however, that the p move into the very thin n-type base region
which is to be amplified, is introduced in the junction whett an input signal of a few tenths region of a pnp transistor is much more where some will combine with the electrons
emitter-base circuit; this current is labeled as of a volt is applied (see Figure 3-2). In heavily doped than the n region. As a from the input source; the base current IB is
the base current in the figure. As we shall contrast, passage of electricity across the consequence, the concentration of holes in the the result. The majority of the holes wil~
show later, an ac current can also be reverse-biased collector-base junction is in- p region is a hundredfold or more greater however, drift through the narrow base layer
amplified by introducing it in series with lB' hibited by the migration of majority carriers than the concentration of mobile electrons in and be attracted to the negatively charged
Power supply
collector junction, where they can combine the gate, however (as in Figure 3-5c~ the Ve ac. It is important to note that a dry cell (or
with electrons from the power supply; the col- supply of electrons in the channel is depleted other dc source generally) has, for all practi-
lector current Ic is the result. (see Figure 3-1d, p. 43); thus, the resistance of -1111 +
cal purposes, negligible resistance toward an
It is important to appreciate that the mag- the channel increases and the current ac signal; similarly, most ac sources can be
nitude of the collector current is determined decreases. assumed to olTer virtually no resistance
by the number of current-carrying holes
available in the emitter. This number,
A major advantage of the field elTect tran-
sistor arises from the fact that it is normally
, toward a dc current. Thus, the series arrange-
ment shown will produce an ac signal that is
I
however, is a fixed multiple of the number of operated in the reverse-bias mode. As a I superimposed upon the dc current. That is,
1
electrons supplied by the input base current. consequence, the resistance of the gate circuit I the base current is given by (/B + iB) where is '
I
Thus, when the base current doubles, so also is very large, and the current drawn from the Ouqx;t is due to the ac signal.
AC
does the collector current. This relationship input signal small. input f (va+ Va) Typically, an amplifier such as that shown '
leads to the current amplification exhibited
by a bipolar transistor.
~~I~
v. (~+M i
I
I
in Figure 3068 employs a power supply that
has an output potential of IS to 30 V dc'
'

AMPLIFIERS l (I. + i.1 j 'lie + IcJ I


I
normally, the current is produced b;:
Field Effect Transistors EMPLOYING TRANSISTORS .J rectification of 1l0-Vac line output. The'
Two other types of transistors, field elTect power supply then serves as the source of the .
The electrical signal from a transducer is or- ':" Ground amplified collector current. Often this current '
transistors (FET) and metal-oxide-semicon-
dinarily so small that it must be amplified, (a) is observed as the output voltage drop
ductor field effect transistors (MOSFET~ are
also encountered. An example of the former
often millions of times, before it is sufficient (vc£ + Vcd across the emitter-eollector ter-
to operate a meter or a recorder and thus be minals of the transistor.
appears in Figure 3-5. Here, the gate is a cylin-
measured. In this section, we will consider the Both the bias current and the signal cur-
drical p-type semiconductor surrounding a
details of a few single-stage, practical rent are, of course, amplified and appear as a
center core of n~oped material called thel
amplifier circuits. It should be appreciated, dc component or the collector current Ie and
channel. As with bipolar transistors, the p- and
however, that a single transistor seldom pro- an ac component ic. Note that the collector
»-semiconductor regions can be reversed. The
vides the amplification needed for most appli- current and the base current combine at the
gate serves the same function as the base in a
cations, and that a satisfactory output can common junction to form an emitter current
bipolar transistor; that is, the bias on the gate
only be achieved by cascading several (if: + If:~ which also has both ac and dc
controls the flow of electrons between the
amplifier stages. character.
source and drain leads of the transistor.
In the absence of a gate bias potentia~ Ground Connections. Electronic circuits
A Simple Amplifier Circuit normally have a low-resistance wire or foil
electrons are free to flow from the power
supply, through the channel and back to the Figure 3-6a is a circuit diagram for a simple (b)
that interconnects one terminal or each signal
power supply; conduction in the channel is amplifier capable or amplirying either an ac source and the power supply; other compo-
FIGURE 3-6 (a) A common-emitter cir- nents or the circuit may also be connected to
by electrons. When a reverse bias is applied to or a dc current by a factor of 100 or more.
cuit for amplification or a small ac signal. this common line, called the ground or the
(b) Shorthand version or (a) wherein all indi- system. The ground provides a common
Power supply cated voltages are assumed to be with respect return ror all currents to their sources. As a
to the common ground.
+ 1111- consequence, all voltages in the circuit are

-=~ ,,,.
atelead
with reference to the common ground. Figure
3-6a shows that the plus terminal or the
- ,p , + power supply is connected directly to the
n n
Source ~ ..,' - Drain - :; - ;>~'j . + With either, the input signal is placed in series ground, which is indicated by the three paral-
lead lead .. (-,. '.'
with a constant base bias VB' and is thus lei lines at the bottom orthe figure. Note that
superimposed on the steady current lB' The a~ide of the input and output signals, as well
Gate bias may be provided by a battery, as shown; as the emitter terminal of the transistor, are
more often, however, the bias is provided by similarly connected and are thus at ground
the amplifier power supply. potential.
In Figure 3-68, the signal is assumed to be Figure 3-6b is a simplified diagram of the
amplifier shown in (a). Here, the lines con- reactance to an alternating current is negli- transistors that are coupled through their
necting the power supply to the grounded gible in the frequency range for which the cir- emitters. As its name implies, this circuit
sides of the output and input have been cuit is intended. Thus, typically, the amplifies the difference between input signals
deleted, it being understood that one side of approximation 1 and 2. If the two load resistors and tran-
v., (VCIl + VcIl1 and Vc is at ground potential.
sistors are perfectly matched, the output of
Note that positive electricity flows from
is valid to 1 to 2 %. Clearly then, the emitter this device will be zero for equal input signals.
ground to the power supply and the signal
follower does not provide any voltage gain. Otherwise, a linear response to signal differ-
sources. ences is obtained.
As with the common-emitter circuit, however,
good current gains are realized. It is also im- The difference amplifier is useful for
An Emitter-Follower Amplifier several reasons. It wil~ for example, reject an
portant to note that Equation 3-4 indicates
Figure 3-7 is a diagram for an emit~er that the output signal will be in phase with unwanted part of a signal that is common to
follower. It differs from the common-emitter the input. both inputs. It is relatively unaffected by tem-
circuit in that the load resistor, RL, is locat-ed The emitter-follower circuit finds wide- perature changes and the passage of time.
between the emitter and the base. For theac spread use because it has a high input im- Finally, it is a useful device for comparing
input, the collector terminal is at ground inas- pedance and thus draws little current. High two signals by the null-point method. Because
much as the dc power supply offers essentially impedance is of importance in measuring of its stability, this circuit is widely used to
no resistance toward the ac current; thus, the electrical properties of devices whose output amplify dc signals.
circuit is a common-collector circuit. is disturbed by current passage. The input re- Amplifiers with Negative Feedback. Often
By applying Kirchhoffs voltage law to the sistance is readily found by assuming that R, it is advantageous to return or feedback a
loop that contains the input, the base, the in Figure 3-7 is so large with respect to RL fraction of the output signal from an amplifier
emitter, and the load resistor RL, we obtaip that essentially all of the input signal passes to the input terminals. The fractional signal
which rearranges to that is returned is called feedback. Figure 3-9
through the latter. Then it is seen that
V••• i.Xc, + V.1l + illRL VB
ill = iB + ic Rio = -;- = (I + a)RL (3-5) shows an amplifier with a feedback loop.
The voltage V.1l across the base-emitter ter- 'B Here, a fraction /I of the output signal is
minal is at most a few tenths of a volt and is Substituting the ac form of Equations 3-1 and where R", is the resistance to an input current sampled by means of a resistor-capacitor
generally insignificant when compared with 3-4 yields of i•.Typically, then, the input resistance of network (shown as the feedback box in the
the JR drop across RL• In addition, the capac- an emitter follower is one hundredfold or diagram); the resulting signal /Iv. is then
itance of C, is made large enough so that its greater than its output resistance, since a returned in such a way that it combines with
often ranges from 100 to 200. the input signal v,, The net input signal then
A Differential Amplifier. Figure 3-8 is a is VA, which is equal to (VI + jlv.). If the
circuit diagram for a difference or differential amplifier gain is a, we may write that

Vc t amplifier. It consists of a pair of matched I'. = aVA = a(v; + /Iv.)


I

t
R, iE: RL
i. I
----- ----.Ij
icl
a
v. = I _ pa Vi (3-6)
feedback system, in contrast t~ the amplifier
itself, can be made quite stable, inasmuch as
the resistors and capacitors it contains are
much leSs subject to temperature fluctuations
C 12V~

Thus, the net gain a' of the amplifier with


feedback is
and aging effects. For that matter, replace-
ment of one or more transistors in an
amplifier with good negative feedback often
r.~ C30V~
. Switch

a'= _a_
1- pa
(3-7)

The feedback ratio p can have either a pos-


has no discernible effect on the overall
amplification of the unit. ~L-:]
115V

Fuse
Primary

itive or negative sign. For example, the


output from the amplifier shown in Figure POWER SUPPLIES
3-9 is out of phase with the input signal. The AND REGULATORS
feedback from this signal, then, would inter- EMPLOYING TRANSISTORS
act destructively with the input, thus reducing
the signal amplitude; here p is negative, and Generally, laboratory instruments require de FIGURE 3-11 Schematicofa typical power
the circuit is said to have negative feedback. power to operate amplifiers and other reac- transformer with multiple secondary windings.
On the other hand, the output signal from the tive components. The most convenient source
amplifier in Figure 3-7 is in phase with the of electrical power, however, is llS-Vac fur-
input; here, the feedback would be positive. nished by public utility companies. As shown
Negative feedback is of great importance in in Figure 3-10,laboratory power supply units
former such as that shown sS:hematically in Filters
electronic circuits; we shall focus almost ex- increase or decrease the potential from the
Figure 3-11. The varying mal!1letic field aris-
clusively on this type of feedback. house supply, rectify the current so that it has The output of each rectifier is also shown in
ing in the primary coil in this device from the
TIle full advantage of negative feedback is a single polarity, and finally, smooth the the plots on the right of Figure 3-12. In order
passage of the IIS-Vac supply current
realized when pa ~ I in Equation 3-7. Then output to give a signal that approximates de. to minimize current fluctuations, the output is
induces currents in the secondary coils; the
Most power supplies also contain a voltage usually filtered by placing a large capacitance
potential across each is given by
regulator which maintains the output voltage in parallel with the load as shown in Figure
a' = - ~ (3-8) at a constant desired level. 3-13. The charge and discharge of the capaci-
tor has the effect of decreasing the variations
Note that under these circumstances, the
Tranefonnera where N 2 and N I are the number of turns in to a relatively small ripple. In some applica-
overall amplification depends only upon the
the secondary and primary coils, respectively. tions, an inductor in series and a capacitor in
properties of the feedback network and not Alternating current is readily increased or de-
Power supplies with multiple taps, as in parallel with the load serve as a filter; this
upon the properties of the amplifier system. The creased in voltage by means of a power trans-
Figure 3-11, are available commercially; type of filter is known as an L section. By
many voltage combinations can be had. Thus, suitable choice of capacitance and inductance,
a single transformer can serve as a power the peak-to-peak ripple can be reduced to the
supply for several components of an millivolt range or lower.
instrument.
The output voltage of a simple transformer
Voltage Regulators
fluctuates only a few percent: even when the
current drawn from the output varies widely. Often. instrument components require de volt-
I
ages that are constant and independent of
the current being drawn. Voltage regulators
Rectifiers
serve this purpose.
Figure 3-12 shows three types of rectifiers and Figure 3-14 illustrates two simple voltage
their output-signal forms. Each uses tran- regulators. Both employ a Zener diode, a pn
FIG U R E 3-10 Diagram showing the components of a power supply sistor diodes (p. 43) to block current flow in junction which has been designed to operate
and their effect on a signal. one direction while permitting it in the other. under breakdown conditions; note the special
""8 e=:
It'"
Holf __
vo~ Time
symbol for this type of diode. In Figure 3-2
(p. 44~ it is seen that at a certain reverse bias
a transistor diode undergoes an abrupt break~
down whereupon the current increases precip- .
itously. For example, under' breakdown
and resistance-the

ments.2
typical signals from the
transducers empioyed for chemical measure-·

General Characteristics
conditions, a current change of 20 to 30 mA of Operational Amplifiers

115V
ac
C
G..
V.

Full-wave
voE Time
may result from a potential change of 0.1 V
or less. Zener diodes with a variety of
specified breakdown voltages are available
commercially.
For voltage regulators, a Zener diode is
Operational amplifiers are a class of differen-
tial amplifiers that find wide application in
instrumentation. The designs of these devices
vary widely. Their performance character-
istics. on the other hand. have a number of
chosen such that it always operates under
co?,mon features which include: (1) large
breakdown conditions; that is, the input volt-

mvil
ItU
age to be regulated is greater than the break-
down voltage.
For the regulator shown in Figure 3-148,
an increase in voltage results in an increase in
gaIns (104 to 106); (2) high input impedances
(orten ~p to 1012 0 or greater); (3) low
output IDlpedances (typically I to 100); and
(4) zero output for zero input.
O~tiOll.1 AmpUfier Design. Operational
current through the diode. Because of the
~phfiers ordinarily contain several amplify-
steepness of the current-voltage curve in the
Ing stages that contain resistors. capacitors,
breakdown region (Figure 3-2), however,
and transistors. Internal negative feedback is
the voltage drop across the diode. and thus
used extensively in their design. Early opera-
the load. is virtually constant.
tional amplifiers employed vacuum tubes and
The regulator circuit shown in Figure
were consequently bulky and expensive. In
3-14b provides the most effective and widely

_., fc :: ~.,
contrast, modern commercial operational
used type of voltage regulator. It serves to
I
,\
reduce the current. and thus the power that
amplifiers are tiny; typically. they are manu-
factured on a single thin silicon chip having
I \ must be dissipated by the diode. Again. the
surface dimensions that are perhaps 0.05 in.
Time
voltage drop across the load is essentially that
on a side. Through integrated-circuit technol-
across the Zener diode.
ogy. resistors, capacitors, and transistors are
FIGURE 3-13 Filtering the current from a rectifier.
formed on the surface of the chip by photo-
lithographic techniques. After leads have bec~
soldered in place, the entire amplifier is
encased in a plastic housing that has dimen-
Regulated sions of a centimeter or less; the package. of
de
course, does not include the power supply.
output Operational amplifiers derive their name
Current-day operational amplifiers, in
from their original applications in analog
addition to being compact, are remarkably re-
computers, where they were employed to per-
liable and inexpensive. Their cost ranges
form such mathematical operations as
typically from less than a dollar to several
summing, multiplying. differentiating, and
dollars. Manufacturers offer a large number
integrating_ These operations also play an im-
Unregulated Regulated of operational amplifiers, each differing in
de de portant part in modern instrumentation; as a
input output consequence. operational amplifiers find
widespread use in instrument design. In addi- 2 For a more complete discussion of applications of
tion to their mathematic role, however, oper- operational amplifiers, see: R. Kalvoda, Opera/joMI
(bl
ational amplifiers find general application in Amplifier .• in Chemical Inslrumenra/ion. New York:
FIGURE 3-14 Voltage regulators. the precise measurement of voltage, current, Halsted Press. 1975.
such properties lIS gain, input and output the differential mode. Here, the difference be- signal input into the inverting terminal yields
impedance, operating voltages, and maximum . tween two input potentials is amplified and an output that is 180 deg out of phDse. The
power. appears as an output with respect to ground. positive input of an amplifier, on the other
Symbols fOl' Operational Amplifiel;oi.In cir- The symbol shown in Figure 3-ISd is often hand, yields an in-phase output signal or a de
cuit diagrams, the operational ampl~er is or- encountered. Unfortunately, it may lead to signal of the same polarity as the -input
dinarily depicted as a triangle having, as a confusion because it is not evident whether
minimum, an input and an output terminal the input is to the inverting or the noninvert- Circuits Employing
(Figure 3-ISd). Often, additional terminals ing terminal; thus, we shall avoid its use. Operetlonal Amplifiers
are indicated, as in (a1 (b1 aild (c) of the Inverting and Noninverting Terminals. It is Operational amplifiers are employed in cir-
figure. A symbol as complete as that in Figure important to realize that the positive and neg- cuit networks that contain various combina-
3-ISa is seldom encountered. Here, the power ative signs show the inverting aniJ noninverting tions of capacitors, resistors, and other
supply leads are shown; ordinarily the power terminals of the amplifier and do not imply electrical components. Under ideal condi- FIGURE 3-16 Inverting circuit with nega-
supply input is deleted, its presence being that they are necessarily to be :connected to tions, the output of the amplifier is tive feedback.
assumed. inputs of the corresponding signs. Thus, if the determined entirely by the nature of the
Figures 3-15a and 3-15b depict the most negative terminal of a battery is connected to network and its components and is independ-
common operational amplifier arrangement, the minus or inverting terminaI,: the output of ent tf tlte operational amplifier itself. Thus, it
in which the noninverting terminal (labeled the amplifier is positive with respect to it; i( is important to examine some of the various tremely high (see criterion (21 p. 55); thus,
+ ) is grounded. The input and output poten- on the other hand, the positive terminal of the networks that employ operational amplifiers. i, <C if and
tials are with respect to ground. Figure 3-ISc battery is connected to the minus input of the BasIc Inverting Circuit. Figure 3-16 is a
shows an operational amplifier being used in amplifier, a negative output results. An ac circuit diagram of an inverting circuit em-
ploying an operational amplifier. The
amplified output v. is 180 deg out of phase
Power supply with the input 1',. Note that part of the output V, - v, v, - v.
Invorting + T vc is fed back through the resistor Rf; because R;-=R;-
terminal I of the phase shift, the feedback is negative.

r
I Substituting the expression for the gain of the
This arrangement is perhaps the most
amplifier into Equation 3-11 and solving for
Input common of all networks associated with
the output voltage gives
operational amplifier circuits.
-:- + Output
I The gain a of the operational amplifier y -_y~_R_f_/R_,}_
_!Vc shown in Figure 3-15a is given by the • 1 + (I + Rf/R,}/a
Noninverting relationship
terminal
Note, however, that a for an operational
(I) v. amplifier is usually larger than 104• Normally,
a=---
1'+ - v_ the resistors R f and R, are chosen such that
Inp~o RJfR, is less than 100. Therefore, under usual
~ Ou~ut Since the plus terminal is grounded, however,
In~PUt conditions, (I + Rf/R,)/a <C I and the fore-
v+ = 0; furthermore, V, = 1'_ . Thus, the gain
going equation becomes
(e) (d) is given by
FIGURE 3-15 Symbols for operational ampli- Rf
fiers. In each case, the output is with respect to
1'.= -V, R,
ground. Inverting circuits with grounded non- The properties of the circuit in Figure 3-16
inverting terminals are shown in (a) and (b). More can be developed with the aid of Kirchhol1's Thus, when a high-gain amplifier is employed
detail than is-usually provided is shown in (a). current law; that is, and when Rf/R, is not too great, the circuit
The circuit in (c) is a differential amplifier in which shown in Figure 3-16 provides a gain which
i,= i,+ if depends only upon Rf/R, and is independent
neither input is grounded. A shorthand representa-
tion of an operational amplifier circuit is shown Recal~ however, that the input impedance of of fluctuations in a, the gain of the amplifier
in (d). an operational amplifier is always made ex- itself.
Light
As we have seen, however, the second term in ~source amplifiers to the measurement of each type of


the denominator is the amplifier gain. Thus, 1\ signal.
Curreat Meuurement The accurate mea-
5 A Open loop gain
Z
,
=~1 +a (3-13) I
I
I
I
surement of small currents is important to
such analytical methods as voltammetry,
4 "SlImPI.
c
Typical values for Rf and a are 10' a and coulometry, photometry, and gas chromatog-
"s,
104• Under these circumstances, the im- + + raphy.
~ 3 As we pointed out in Chapter 2 (p. 30), an
pedance between S and ground would be only
10 a. This property of the circuit, shown in important concern that arises in all physical
Figure 3-16, is of considerable importance in measurements, including that of current, is
current measurement (see p. 30~ whether the measuring process wilI, in itsel~
Frequency Response of the Basic Inverting alter significantly the signal being measured,
Circuit The gain of a typical operational thus leading to aQ error. It is inevitable that
log frequency amplifier decreases rapidly in response to any measuring prOcess will perturb a system
high-frequency input signals. This frequency under study in such a way that the quantity
FIGURE 3-17 Frequency response of a actually measured differs from its original
typical operational amplifier. (a) Without dependence arises from small capacitances FIGURE 3-19 Application of an opera-
that develop at pn junctions. The effect for a tional amplifier to the measurement of a value before the Iheasurement. All that may
negative feedback and (b) with negative
typical amplifier is shown in Figure 3-17, small photocurrent, I". be hoped is that the perturbation can be kept
feedback.
where the curve labeled open-loop gain repre- sufficiently small. For a current measurement,
sents the behavior of the amplifier in the this consideration requires that the internal
absence of the feedback resistor Rf in Figure resistance of the ;measuring device be kept
3-16. Note that both the ordinate and the input is fed into the noninverting termina~ low.
It is important to appreciate that when abscissa are log scales and that the open-loop and a feedback loop is provided which in- A low-resistanet current measuring device
Equation 3-12 applies, the voltage at point S gain for this particular amplifier decreases volves the inverting terminal. The result is an is readily obtained by deleting the resistor R,
with respect to ground is necessarily negli- rapidly at frequencies greater than about amplifier with a gain of approximately unity. in Figure 3-16 arid using the current to be
gible compared with either 1'/ or v•. That t~is 100 Hz. Because the input is not inverted, the output measured as the signal input. An arrangement
statement is correct can be seen by notmg In contrast, an operational amplifier em- potential is the same as the input. The input of this kind is shown in Figure 3-19, where a
that the only cOndition under which Equation ploying external negative feedback, as in impedance is the open-loop impedance for small direct current I" is generated by a pho-
3-12 follows from 3-11 is when v, ~ v. and 1'/. Figure 3-16, has a constant gain or bandwidth the amplifier, which can be very large totube, a transducer that converts radiant
Thus. the potential at point S must approach that extends from dc to over 10' Hz. In this (100 Ma) when a field effect transistor is energy such as IiBht into an electrical current.
ground potential; as a consequence, point S is region, the gain depends only upon Rf/R" as employed. The output impedance, on the When the cathode of the phototube is main-
said to be at virtual ground in the circuit. shown by Equation 3-12. For many purposes, other hand, is low « 1 a). As will be shown tained at a potential of about -90 V, absorp-
It is of interest to determine the effective the frequency independence of the negative later, this impedance transformation is valu- tion of radiation by its surface results in the
impedance Z, between the point Sand feedback inverting circuit is of great impor- able in the measurement of high-impedance ejection of electrons, which are then ac-
ground. This quantity is given by tance and more than offsets the loss in sources with low-impedance measuring celerated to the wire anode; a current results
amplification. devices. that is directly proportional to the power of
Z, = V,/ii the radiant beam.
The Voltage-Follower Circuit Figure 3-18
depicts a voltage follower, a circuit in which Amplification and If the conclusions reached on page 57 are
We have noted, however, that ij ~ if' and applied to this circuit, we may write
from Figure 3-16, we see that Measurement of Transducer Signals
Operational amplifiers find general applica- Ix=',If+I.~If
. . v. - v. ;
li=lf=-R- tion to the amplification and measurement of
f

Combining the two equations gives ,,$ L.. the electrical signals frolD transducers. These
signals, which are often concentration-
dependent, include current, potentia~ and re-
In addition, the point S is at virtual' ground
so that the potential Y" corresponds to the
potential drop across the resistor Rf. There-
fore, from Ohm's law,
Z= v,Rf =~ FIGURE 3-18 Voltage-fol- sistance (or conductance). This section
, v, - v. 1 - v./v, lower circuit. includes simple applications of operational Y,,= -lfRf= -I"Rf
parameter is related to the concentration of 100 Mn or more is often called an ment of conductance or resistance. In (a), the
the species responsible for the absorption. electrometer. conductance of a cell for a conductometric
I" = - V./R, = kV. Photometers are described in detail in Chap- Resistance or Conductance Mea.ements. titration is of interest. Here, an alternating-
Thus, the potential measurement V. gives the ter 7. Electrolytic cells and temperature-responsive current input signal VI of perhaps:5 to 10 V is
current, provided R, is known. By making R, Potencial Meaurements. Potential meas- devices, such as thermistors and bolometers, provided from the secondary of a filament
reasonably large, the accurate measurement urements are used extensively for the measure- are common examples of transducers whose transformer. The output signal is then
of small currents is feasible. For example, if ment of temperature and for the determination electrical resistance or conductance varies in rectified and measured as a potential. The
R, is 100 kCl, a 1-1lA current results in a of the concentration of ions in solution. In the response to a chemical signal. They are em- variable resistance R, provides a means for
potentia} of 0.1 V, a quantity that is readily former application, the transducer is a thermo- ployed for conductometric and thermometric varying the range of conductances that can be
measured with a high degree of accuracy. couple; in the latter, it is a pair of electrodes. titrations, in infrared absorption and emission recorded or read. Calibration is provided by
An important property of the circuit In Chapter 2 (p. 32~ it was demonstrated methods, and for temperature control in a switching the standard resistor R. into the cir-
shown in Figure 3-19 is its low resistance that accurate potential measurements require variety of analytical measurements. cuit in place of the conductivity cell.
toward the tra~ucer, which minimizes the that the resistance of the measuring device be The circuit shown in Figure 3-16 provides Figure 3-21b illustrates how the circuit in
measuring error. Recall from the earlier large with respect to the internal resistance of a convenient means for measurement of the Figure 3-16 can be applied to the measure-
discussion (p. 58) that S is at virtual ground, the voltage source to be measured. The need resistance or conductance of a transducer. ment of a ratio of resistances or conduct-
and the resistance of the measuring device is for a high-resistive measuring device becomes Here, a constant potential source is employed ances. Here, the absorption of radiant energy
thus low. ' particularly acute in the determination of pH for Vi and the transducer is substituted for by a sample is being compared with that for a
with a glass" electrode. For this reason, the either Ri or R, in the circuit. The amplified reference solution. The two photoconduct-
EXAMPLE basic inverting circuit shown in Figure 3-16, output potential V. is then measured with a ance transducers replace R, and RI in Figure
Assume that R, in Figure 3-19 is 200 kCl, which typically has an internal resistance of suitable meter, potentiometer, or recorder. 3-16. A battery serves as the source of power
the internal ~tance of the phototube is perhaps 10' n, is not satisfactory for voltage Thus, if the transducer is substituted for R, in and the output potential M, as seen from
5.0 x 104 Cl, ani! the amplifier gain is measurements. On the other hand, it can be Figure 3-16, the output, as can be seen from Equation 3-12, is
1.0 x 10'. Calcu~te the relative error in the combined with the voltage-follower circuit rearrangement of Equation 3-12, is
current measurement that results from the shown in Figure 3-18 to give a very high- M = k R. = k' p. (3-16)
presence of the measuring ~ircuit. impedance voltage-measuring device. An R P
Here, R••, the resistance of the measuring example of such a circuit is shown in Figure Typically, the resistance of a photoconductive
device, is given by Equation 3-13 3-20. The first stage involves a voltage cell is directly proportional to the radiant
follower, which provides an impedance of as where R" is the resistance to be measured and k
R., = Z. = R,/(I + a) power, P, of the radiation striking it. Thus,
much as 1012 n. An inverting amplifier circuit is a constant which can be calculated if Ri the meter reading M is proportional to the
- 200 x 10'/(1 + 1.0 x 10') = 2.0 n follows, which amplifies the input by R,/~ and V, are known; alternatively, k can be ratio of the power of the two beams.
or 20. An amplifier with a resistance of determined by a calibration wherein R" is Comparison of Transducer Outputs. It is
On page 31, it was demonstrated that the
replaced by a standard resistor. frequently desirable to compare a signal gen-
relative error in a current measurement is
If conductance rather than resistance is of erated by an analyte to a reference signal, as
given by
interest, the transducer conveniently replaces in Figure 3-21b. A difference amplifier, such
R R; in the circuit. Here, from Equation 3-12, it as that shown in Figure 3-22, can also be
reI error = - -_ ••-
RT + R., is found that employed for this purpose. Here, the amplifier
where RT is the resistance of the circuit in the is being employed for a temperature measure-
absence of the resistance of the measuring ..!- = G = _l = k'v (3-15) ment. Note that the two input resistors (Rj)
device R••. Thus; R" " viR, • have equal resistances; similarly, the feedback
resistor and the resistor between the non in-
i 2.0 where G" is the desired conductance. Note verting terminal and ground (both labeled Rt)
re I error = - 5.0 X 104 + 2.0 that in either type of measurement, the value are also alike.
= -4.0 x 10-' or -0.004% of k, and thus the range of the measured Applying Ohm's law to the circuit shown
values, can be readily varied by employing a in Figure 3-22 gives
The instrument shown in Figure 3-19 is a variable resistor R or R,.
j

photometer, which measures the attenuation FIGURE 3-20 A high-impedance circuit Figure 3-21 illustrates two simple applica- V1- V_
11=---
of radiation passing through a solution; this for voltage amplification. tions of operational amplifiers for the measure- Rj
The potential V+ can also be written in terms that meets these qualifications is termed a
of V2 by means of the voltage-divider equa- potentiostat.
tion. Thus, Several reference cells are available to pro-
_ V _ V2Ra: vide an accurately known voltage; one of
these, the Weston cell, was described on
+ - R1+Rt
page 34. None of these sources will, however,
Combining the last two equations gives, upon maintain its potential when a significant cur-
rearrangement, rent is required.
Figure 3-23 illustrates how a Weston cell
or some other reference cell can be employed
to provide a standard voltage source from
Thus,' it is the difference between the two sig- which relatively large currents can be drawn.
nals that is amplified. Note that in both circuits, the standard
Any extraneous potential common to the source appears in the feedback loop of an
two input terminals shown in Figure 3-22 will operational amplifier.
Rodiation be ~celed and not appear in the output. Recall (p. 58) that point S in Figure 3-23a
sou",. Thus, any slow drift in the output of the is 'at virtual ground. For this to be so, it is
A'-••
I
-~ I
transducers or any 6(kycle currents induced necessary that V. = V••••.That is, the current
I I from the laboratory power lines will be eli- through the load must be such that
minated from V.. This useful property I R,. = V..,. It is important to appreciate,
accoUnts for the widespread use of a differen- however, that this current arises from the
tial ~rcuit in the first amplifier stages of power source of the operational amplifier and
instnJments. not tire stalldard cell. Thus, the standard cell
The transducers shown in Figure 3-22 are controls v" without the passage of a
a pair of thermocouple jUllctiolls, one of which significant current through it.
is immersed in the sample and the second in a Figure 3-23b illustrates a modification of
reference solution (often an ice bath) held at the circuit in (a) which permits the output
constant temperature. A temperature-depend- voltage of the potentiostat to be fixed at a
ent contact potential develops at each of the level that is a known multiple of the output
two junctions formed from wires made of voltage of the standard cell.
copper and an alloy called constantan (other
(b)
metal pairs are also employed). The potential
FIGURE 3-21 Two simple circuits developed is roughly 5 mV per loooe temper-
conductance or resistance outputs. ature difference.

Application of
Operational Amplifiers
Figure 3-16 is at virtual ground (p. 58) can to V~ltage and Current Control
V_ - v" be applied to Figure 3-22, where it becomes
1,=--- Operational amplifiers are readily employed
Rt apparent that
to generate constant-potential or constant-
But Equation 3-10 reveals that II and I, are V_ = - V+ curreht signals.
approximately equal. Thus, COMtant-Voltage SoUrce. Several instru-
Substitution of this relationship into the
VI - V_ V_ - v" previous equation yields, upon rearrangement, mental methods require a power source
--R-i-=~ whose potential is precisely known and from
-V _ V1Rt + v"Rj which reasonable currents can be obtained FIGURE 3-22 A circuit for the amplifica-
+-
The argument demonstrating that point S in RI +Rt without alteration of this potential. A circuit tion of differences.
Figure 3-24b is an amperostat which em- evident in Chapter 6, however, it is the loga-
ploys a standard cell (V...,) to maintain a con- rithm of this ratio that is directly related to the
stant current; no significant current is, analyte concentration. Suitable modification
however, drawn from the standard. The non- of the circuit shown in Figure 3-21b permits
inverting booster amplifier permits relatively direct conversion of the outpUt signal to a
large currents to be passed through the cell. logarithmic function.
The outputs of several analytical methods,
Application of for example, chromatography, take the form
Operational Amplifiers
to Mathematical Operations
of a Gaussian-shaped peak in which an elec-
trical signal is plotted as a function of time.
Vo·-R,
. (V,-+-+-+-
V2 V, V.)
R, R2 R, R.
Integration to give the area under the peak is • -Rill, + 12+ I, + 1.1
An important property of the basic inverting
necessary in order to provide a parameter FIGURE 3-25 A summing circuit.
operational amplifier circuit shown in Figure
that is proportional to concentration. Opera-
3-16 is that substitution of various combina-
tional amplifiers can be used to perform the
tions of resistors and capacitors for Rj and R,
integration process and thus give directly a
permit mathematical operations to be per-
concentration-dependent signal.
formed on the input signal. Often, as a
Multiplication and Division by a Con-
consequence, the output is more easily related
to the concentration of the analyte. For
example, we have shown in Figure 3-21b how
stant. We have already seen that the output
of the basic inverting amplifier circuit is given
.=
y.
-R, ( -
VI
RI
+-V2 +-V3 +-V~)
R2 R3 R~
by (3-19)
(bl
the electrical output from a pair of photo-
diodes can be converted to a signal that is R, If the four input resistors and the feCdback
FIGURE 3-23 Constant p0- directly proportional to the ratio of the power v = -v·- resistor are all alike (R, = R == RI = R2 =
tential sources.
• I Rj
of two beams of radiation. As will become R3 == R~~ then the output potential is the
Here, the input signal is being multiplied by a sum of the four input potentials. That is,
constant whose magnitude can be varied by
alteration of the resistance ratio. When this V.=VI+V2+V3+V~

. Constant Current Sources. Constant cur- ratio is less than one, the operation is equiva- Another interesting case arises when
lent to division of the signal by a constant.
rent sources, called amperostats, find applica- RI = R2 = R3 == R~ = R = 4R,
Adclition.nd Subtraction. Figure 3-25 illus-
tion in several analytical instruments. These V,
devices are usually employed to maintain a h-- trates how an operational amplifier can be Here, Equation 3-19 reduces to
Rj
employed to yield an output that is the sum
constant current through an electrochemical
cell. An amperostat reacts to a change in of several input signals. It is important to Y. = _ ~(VI + V2 + V3 + \14)
note again that the point S is at virtual • 4 R R R R
input power or a change in internal resistance
ground. As a consequence, the various input c' (VI + V2 + V3 + V~)
of the cell by altering its output potential in
such a way as to maintain the current at a signals do not interfere with one another; all 4
predetermined level. flow independently to the virtual ground.
Thus, the current at S is simply the sum of the and the output gives the average value for the
Figure 3-24 shows two amperostats. The
individual currents. Because the impedance of four signals. Clearly, a weighted average can
first requires a voltage input V; whose poten-
the amplifier is large, however, be obtained by varying the values of the input
tial is constant in the presence of a current. resistors.
Recall from our earlier discussion that Subtraction becomes possible with the cir-
cuit by reversing the sign of one or more of
the input signals. •
I, = - V./R, Integration. Figure 3-26a shows a useful
Thus, the current will be constant and circuit for integration of a varying-input
independent of the resistance of the cel~ lbl Therefore, signal. Here, a capacitor CI replaces the re-
provided that V; and R remain constant.
j FIGURE 3-24 Constant current sources. V.= -R,(II + 12 + 13 + I~) (3-18) sistance R I in the basic inverting circuit.
3-28a overcomes this problem. Here the
v = --
• R,C
1
f v, dt
'
0
(3-20) introduction· of a small capaci~ C i~ the
feedback circuit and a small resisto: R in
the input circuit filters the high-frequency :olt-
Hold Thus, the output is the time integral of the
ages without significant attenuation of
lWitl:h input voltage. The definite integral is ob-
Vl~ the signal of interest. Figure 3-28b shows
tained by opening the reset switch and closing
the appearance of a typical titration curve v,
the hold switch (Figure 3-26a) at time zero.
and its derivative v•.
Then, when the hold switch is opened at time
t, the integration is stopped, thus holding v.
Generation or Logarithms and Antiloga-
rithms. A number of transducers produce an
at a constant level for measurement. Closing
electrical response which is exponentially
the reset switch then discharges the capacitor
related to concentration. Thus, it is the loga-
so that a new integration can be started.
rithm of the signal that is directly proportional
Figure 3-26b shows a chromatogram
to concentration. Figure 3-29a shows a circuit
(lower curve) and its time integral obtained
t~at will convert an input to its logarithm;
with an operational amplifier circuit. Note
FIgure 3-29b shows a circuit that will provide
that for each peak. there is a corresponding
the antilogarithm of an input. It should be
step in the intepl curve. The height of this
noted that the circuits shown in Figure 3-29
step corresponds to v. in Equation 3-20 and is
Volume, ml ---.. require that the input voltage be of a fixed
ordinarily proportional to the concentration
(bl polarity.
of the species responsible for the peale.
Differentiation. Figure 3-27 is a simple cir- FIGURE 3-28 (a) A practical circuit for
cuit for differentiation. Note that it differs differentiation. (b) Titration curve v, and its
(bl from the integration circuit only in the re- derivative v•.
spect that the positions of C and R have been
FIGURE 3-26 (a) Circuit for integration.
reversed. Here, we may write
(b) A chromatogram and its time integra~
where v. is proportional to the concentration dv, v.
of the species X. C,-== -- where the rate of change in the transducer
dt Rr
signal is low. For example, differentiation is a
useful way to treat the data from a poten-
ilv,
v = -RfC,- tiometric titration; here, the potential change
• dt of interest occl,lrs over a period of a second or
To analyze this circuit, recall that for all The circuit shown in Figure 3-27 is not, in more (f S; 1· Hz). The input signal will,
practical purposes, fact, practical for most chemical applications, however, contain extraneous 60-, 120-, and
2400Hz potentials (see Figure 3-34), which are
ij= ir induced by the ac power supply. In addition,
Because point S is at virtual ground, the signal fluctuations resulting from incomplete
; output v.' is the voltage across the capacitor. mixing of the reagent and analyte solutions
Thus, substituting Equation 2-22 (p. 16) and are often encountered. Unfortunately, the
i Ohm's law into the foregoing equality gives output of the circuit in Figure 3-27 has a
strong frequency dependence; as a
.!i = _Cdv• consequence, the output voltage from the ex-
R, dt traneous signals often becomes as great or
greater than that from the changing trans-
(bl
ducer signa~ even though the magnitude of
FIGURE 3-27 Asimplecircuitfordifferen- the former relative to the latter is small. FIGURE 3-29 Circuits for generating
tiation. The modified circuit shown in Figure (a) logarithms and (b) antilogarithms.
concentration level. At double the analyte
concentration (Figure 3-3Of~ the existence of
The ultimate limitation to the accuracy and the rectangular wave becomes obvious;
Aelative
sensitivity of an analytical method is the accurate determination of the ac amplitude enalyte
presence of extraneous and unwanted signals would, however, be difficult. At the highest conen ANllyte signal Signal + noise SIN

that are superimposed on the analyte signal concentration of analyte, the presence of the

,I ,I
being measured. These extraneous signals are repetitive analytical signal is clearly seen; the
called noise, the terminology being derived relative error associated with the measure- 0 0
"M), .• ~
from radio engineering, where the presence of ment of its amplitude would be significantly ~v VV • \\J
unwanted signals was recognizable to the ear smaller than at the lower concentrations.
0 0
as noise or static. Amplification of a noisy signal provides
(e) (bl
The effect of noise on an analyte signal is little improvement because typically both the
shown in the plots on the right-hand side of noise and the information-bearing signal are

,~""
Figure 3-30. The graphs on the left are the amplified to the same extent; in addition,
theoretical signals in the absence of noise. noise may be added by the amplification
Note that the analytical signal is a rectangu- process. This effect is shown in Figure 3-31,
. Jar wave whose current or voltage amplitude where the noise shown in Figure 3-3Ob has
is directly proportional to analyte concentra-
tion; the signals shown are for relative con-
been doubled and superimposed upon the
pure signal in Figure 3-3Oe. Here, the output
,~
o 0
centrations of 0: I : 2 : 4. The noise assumed is effectively that of Figure 3-3Od after dou-
to be associated with the measurement is bling. Again, visual detection and measurement
shown in Figure 3-3Ob, where the analyte of the analyte signal would be unsatisfactory.
concentration is zero. For the purpose of It is worth noting here that lock-in amplifiers,
illustration, this same noise has simply been
added to the three analyte signals (c, e, and
g~ Thus, the representations in graphs d, (
which are discussed in a later section, do
permit amplification and extraction of a
signal from much of the accompanying noise.
, ,In n n ,~J1I.~
L!h 'VtM- ,
~t
and h are artificial because the fluctuations are It is apparent from these illustrations that
random and would differ from measurement noise becomes increasingly important as its o 0

to measurement unless all four signals were magnitude approaches that of the analyte
observed simultaneously. Furthermore, it has signal; thus, the signal-to-noise tatio SIN is a
been assumed that no significant noise is as- much more useful parameter than the abso-
sociated with the analyte signal itself; such a lute noise level in describing the quality of the
situation is unlikely in real measurements. output from a chemical measurement.
Finally, noise is not necessarily random; it
often has both random and nonrandom com- Signal-to-Noi •• Ratio
ponents. Nevertheless, the illustration is
useful. The signal-to-noise ratio associated with a
measurement is often expressed in terms of a
Examination of the signal plus noise in
ratio of currents, voltages, or power asso-
Figure 3-3Od suggests that the presence of
noise would make detection of the analyte ciated with the signal and the noise. It may ~ 00

signal difficult, or perhaps impossible, at this also be expressed in terms of the decibel; by FIG U R E 3-30 The effect of noise on an ac signal. In each ins1ance, the
definition, noise shown in (a) has been superimposed on the signal from the first column.

!
N
(decibels)
• For a more detailed discussion of noise see: T. Coor, J.
en. Educ. 45, AS34 (1968)and A. J. Diefenderfer,
Principles of Electronic Instrumentation. Philadelphia, = 20 log signal voltage or current
Saunders. 1979.Chapter 13. noise voltage or current
.~
The random noise for an instrument can Johnson Noise. Johnson noise owes its After amplification,
often be obtained by performing 20 to 30 source to the thermal motion of electrons in
measurements on a blank. The noise then is resistors and other resistive elements of an ",ms = 0.013 x 100 = 1.3 mV
taken as the standard deviation of the set (see electrical circuit. This agitation, which Shot Noise. Shot noise arises wherever a
Appendix 1~ That is, o decreases as temperature is r~uced, imparts current involves the transfer of electrons or
a random motion to the electrons, which in other charged particles across junctions. In
1 (x. -

.rn-n
x)
turn causes inhomogeneities in electron densi- the typical electronic circuit, these junctions
(n - 1) ties throughout the resistor. Thus, momentary are found at p and n interfaces; in photocells
where x. is the observed parameter and x is small voltage differences develop within re- and vacuum tubes, the junction consists of
the mean for the n measurements. The SIN gions of the resistor, and the net voltage drop the evacuated space between the anode and
values shown in the last column in Figure across the resistor fluctuates slightly with the cathode. The currents in both result from
3-30 were based on this definition of N; that o time. Note that Johnson noise will occur even a series of quantized events. namely, the
is, they are ratios of the amplitudes of the ac in the absence of a current. transfer of individual electrons across the
signals to the standard deviation derived from The magnitude of Johnson noise is readily junction. Thus, the current results from a
the blank. As a rule of thumb, visual observa- derived from thermodynamic considerations4 series of random events, the number of which
tion of a signal becomes impossible when SIN and is given by
is subject to fluctuation from instant to
is smaller than from two to three. instant. Because the behavior is random,
. however, the magnitude of the current fluctu-
where "•••••is the root-mean-square noise volt- ation is readily susceptible to statistical treat-
Sources of Nol ••
age lying in a frequency bandwidth of ment using the equation
Noise is associated with each component of M" Hz, k is the Boltzmann constant (1.38 x
an instrument-that is with the source, the 10-23 J/deg~ T is the absolute temperature, = J2Je Af
i,•••• (3-24)
FIGURE 3-31 Effect of
transducer, the signal processor, and the read- and R is resistance in ohms of the resistive where i,ms is the root-mean-square current
amplification of a noisy
out Furthermore, the noise from each of element. fluctuation associated with the average direct
signal. Here, both the noise
these components may be of several types and It is important to note that Johnson noise. current J. e is the charge on the electron
and the signal from Figure
arise from several sources. Thus, the noise while dependent upon the frequency band- (1.6 x 10-19 q and M" is again the band-
3-3Od have been doubled.
that is actually observed is a complex compos- width. is independent of frequency itself; thus. width of frequencies being considered. Like
ite, which usually cannot be fully charac- it is sometimes termed white noise by analogy Johnson noise, shot noise has a •.white ft

terized. Certain kinds of noise are to white light, which contains all visible spectrum.
recognizable, however, and a consideration of frequencies. It is also noteworthy that John- Flicker Noise. Flicker noise is charac-
their properties is useful. example. incomplete mixing of a reagent with son noise is independent of the physical size terized as a noise whose magnitude is in-
Instrumental noise can be divided into an analyte solution during a titration often of the resistor. versely proportional to the frequency of the
four general categories. Two, thermal or John- causes initial fluctuations (or noise) in the signal being observed; it is sometimcs tcrmed
son noise and shot noise, are well understood, potential of an electrode system used to mon- EXAMPLE l/r (olle-over{) noise as a consequence. The
and quantitative statements can be made itor the analyte concentration. Uncontrol- Calculate the Johnson noise at the input causes of flicker noise are not well undcr-
about the magnitude of each. It is clear that lable variations in the temperature of an arc and the output of an amplifier having an stood; its ubiquitous presence. however, is rec-
neither Johnson nor shot noise can be totally source used to excite the atomic spectrum of input resistance of I MO and an amplification ognizable by its frequency dependence.
eliminated from an instrument. an element often cause significant fluctuations , of 100. Assume the bandwidth of the input Flicker noise becomes significant at frequen-
Two other types of noise are also recogniz- in the intensity of the line that is employed as ; signal is 104 Hz. The temperature is 25°C. cies lower than about 100 Hz. The long-term
able, environmental noise and flicker or Ilf a measure of concentration. Noise of this type Employing Equation 3-23. we find drift observed in dc amplifiers, meters, and
noise. Their sources are neither well defined
nor understood. In principle, however, they
is often more significant than instrumental i
noise.
V,ms = J4 x 1.38 x 10-23 x 298 X 106 x 104 galvanometers is considered to be a mani-
festation of flicker noise.
can be eliminated by appropriate instrument _ The sections that follow deal with in- = 1.3 x 10-5 V or 0.013 mV Environmental Noise. Environmental noise
design. strumental noise only. Noise associated with is a composite of noises arising from the sur-
In addition to instrumental noise, the re- the properties of analytes will be dealt with in roundings. Figure 3-32 suggests typical
sults of an analysis are often affected by a the chapters devoted to specific instrumental • For example, see: T. Coor, J. Chem. Edue .• 45, AS33 sources of environmental noise in a university
type of noise related to the analyte itself. For methods. and ASS3 (1968). laboratory.
Gl'OUIIlIi.. and Shielding. Noise ansmg the input signal to a narrow band of frequen-
Power line
60-
from environmenfa1ly generated electromag- cies and employing an amplifier that is tuned
180- netic radiation can often be substantially to this band. It should be appreciated,
reduced by shielding, grounding, and mini- however, that if the signal generated from the
I· Good ·1 mizing the lengths of conductors. Shielding analyte varies with time, the bandwidth must
qUtet
region consists of surrounding a circuit, or Some of be sufficiently great to carry all of the infor-
the wires in a circuit, with a conducting mate- mation provided by the signal.
rial that is attached to earth ground. Electro- Mndulation. Amplification of a low-fre-
magnetic radiation is then absorbed by the quency or dc signal is particularly"troublesome
shield rather than by the enclosed conduc- because of amplifier drift and flicker noise.
tors; noise generation in the instrument cir- Often, this Iffnoise is several times larger than
cuit is thus avoided in principle. the types of noise that predominate at higher
Shielding becomes particularly important frequencies. For this reason, low-frequency or
when the output of a high-impedance trans- dc signals from transducers are often converted
FIGURE 3-32 Some sources of environmental noises in a university ducer, such as the glass electrode, is being to a higher frequency, where Iff"noise is less
laboratory. Note the frequency dependence. [From T. Coor, J. Chern. amplified. Here, even minuscule induced cur- troublesome. This process is call~ modulation.
Educ~ 45,533 (1968). With permission.] rents give rise to relatively large voltage drops After amplification, the modulated signal can
and thus to large voltage fluctuations. be freed from amplifier Iff noise by filtering
Difference Amplifiers. Any noise generated with a high-pass filter; demodulation and
in the transducer circuit is particularly critical filtering with a low-pass filter then produces a
because it appears in an amplified form in the . dc signal suitable for driving a re8dout device.
Much environmental noise occurs because instrument readout. To attenuate this type of Figure 3-34 is a schematic diagram show-
Signat-to-Nol.e Improvement
each conductor in an instrument is poten- noise, most instruments employ a difference ing the flow of a signal thrdugh such a
tially an antenna capable of picking up elec- Many laboratory measurements require only amplifier, such as that shown on page 51, for system. Here, the original dc current is mod-
tromagnetic radiation and converting it to an minimal effort to maintain the signal-to-noise the first stage of amplification. An ac signal ulated to give a narrow-band 4OO-Hz signal
electrical signal. Numerous sources of electro- ratio at an acceptable Ieve~ because the sig- induced in the transducer circuit generally
magnetic radiation exist, including ac power nals are relatively strong and the require- appears in phase at both the inverting and
lines, ignition systems in gasoline engines, ments for precision and accuracy are low. noninverting terminals; cancellation, however,
arcing switches, brushes in electrical motors, Examples include the weight determinations occurs at the output.
lightning, and ionospheric disturbances. Note made in the course of a chemical synthesis or
that some of these sources, such as power the color comparison made in determining
Fllte,. v,~
lines. cause noises with limited-frequency the chlorine content of the water in a swim-
bandwidths. ming pool. For both examples, the signal is High, low, and resonant filters (pages 23 to
It is also noteworthy that the noise spec- large relative to the noise and the require- 27) are often employed to remove noise that
trum shown in Figure 3-32 contains a large, ments for accuracy are minimal. When the differs in frequency from the signal of interest.
continuous noise region at low frequencies.
This noise has the properties of flicker noise;
its sources are unknown. Superimposed upon
the flicker noise are noise peaks associated
with yearly and daily temperature fluctua-
tions and other periodic phenomena asso-
ciated with the use of the laboratory building.
need for sensitivity and accuracy increases,
however, the SIN ratio often becomes the lim-
iting factor of a measurement.
Two general methods exist for overcoming
signal-to-noise limitations. The first entails re-
duction of the noise by appropriate instru-
ment design. The second makes use of some
Figure 3-33 illustrates the use of a low-pass
RC filter for reducing environmental and
Johnson noise from a slowly varying dc
signal. High-pass filters are also useful. Gen-
erally, this type of filter is employed to reduce
drift and other flicker noise from an ac signal.
Resonant filters are also used to attenuate
V~t2
- Time

Finally, two quiet-frequency regions in type of signal averaging to isolate the signal noise in electronic circuits. We have pointed FIGURE 3-33 Use of a
which environmental noises are low are in- from the noise. out that the magnitude of fundamental noise low-pass filter with a large
dicated in Figure 3-32. Often, signals are con- The sections that follow deal briefly with is directly proportional to the frequency time constant to remove
verted to these frequencies to reduce noise some of the methods that are employed to bandwidth of the signal. Thus significant noise from a slowly chang-
during signal processing. overcome the effects of noise. noise reduction can be achieved by restricting ing dc voltage.
that is amplified by a factor of 105• As shown, their surroundings; that is, they suffer from Signal form.

b
however, amplification introduces Iff and serious environmental noise effects. e
power line noise; most of this noise can be In order to minimize these noise problems,
removed with the aid of a suitable filter the beams from infrared sources are generally
before demodulation. chopped by imposition of a slotted rotating I I III I I I
nnnn
Sipal Chopping; the Chopper Amplifier.
The chopper amplifier provides one means for
disk in the beam path. The rotation of this
chopper produces a radiant signal that fluctu-
t I I II I I I

I~vo --- 1.5V

accomplishing the signal flow shown in Figure ates periodically between zero and some max-
3-34. In this device, the input signal is con- imum intensity. After interaction with the Circuit
verted to a rectangular wave form by an sample, the signal is converted by the trans- A B

1
ducer to an ac electrical signal whose Vi
electronic or mechanical chopp~. Chopping V.
can be performed either on the ~source itself frequency depends upon the size of the slots
or on the electrical signal from t~ transducer. and the rate at which the disk rotates. Envi-

1~
In genera~ it is desirable to chop the signal as ronmental noise associated with infrared
close to its source as possible because it is only measurement is generally dc or low-frequency
the noise arising after chopping that is removed
by the process.
Infrared spectroscopy provides an example
of the use of a mechanical chopper for signal
. modulation. Noise is a major. concern in
detecting and measuring infran:d radiation
ac; it can be largely eliminated by use of a
high-pass filter prior to amplification of the
electrical signal. .
Another example of the use of a chopper is
shown in Figure 3-35. This device is a chopper
amplifier, which employs an ac-driven
FIGURE 3-35 A chopper amplifier. The signal forms are idealized wave
forms at the various indicated points in the circuit.
I
"='

because both source intensity ~nd detector electromagnet to operate a switch that, in its
sensitivity are low in this regio~ of the spec- closed position, shorts the input or the output
trum. As a consequence, the electrical signal signal to ground. The appearance of the
approximately rectangular wave signal with ship to the latter. Figure 3-36a shows a
from an infrared transducer is generally small signal at various stages is shown above the
an amplitude of 6 mV (B). Amplification pro- system which employs an optical chopper to
and requires large amplification. Further- circuit diagram. The transducer input is
duces an ac signal with an amplitude of 6 V provide coherent analytical and reference sig-
more, infrared transducers, which are heat assumed to. be a 6-mV dc signal (A). The
(C), which, however, is shorted to ground per- nals. The reference signal is provided by a
detectors, respond to thermal radiation from vibrating sWitch has the effect of forming an
iodically; as shown in (D), shorting also re- lamp and can be quite intense, thus freeing it
duces the amplitude of the signal to 3 V. from potential interferences. The reference
Finally, the RC filter serves to smooth the and signal beams are chopped synchronously
signal and produce a 1.5-V dc output. The by the rotating slotted wheel, thus providing
synchronous demodulation process has signals that are identical in frequency and
Demodulation to de the effect of rejecting the noise generated have a fixed phase angle to one another.
.lter filtering within the amplifier. The synchronous demodulator acts in a

\ Lock-in Amplifiers.5 Lock-in amplifiers


permit the recovery of small signals even
when SIN is unity or less. Generally, a lock-in
amplifier requires a reference signal that is
manner analogous to the double poleo(fouble
throw switch shown in Figure 3-36b. Here,
the reference signal controls the switching so
that the polarity of the analytical signal is
coherent with the signal to be amplified. That reversed periodically to provide a rectified dc
is, the reference signal must be of the same signal, as shown to the right in Figure 3-36c.
frequency as the analytical signal and, equally The ac noise is then removed by a low-pass
important, must bear a fixed phase relation- filtering system.
A lock-in amplifier is generally relatively
o
1.0 free of noise because only those signals that
Frequency, Hz Frequency, Hz Frequency, Hz
are "locked-in" to the reference signal are
• See. T. C. O'H.ver. J. Chem. Educ., 49, A131. A211 amplified; that is, those that are coherent. AU
FIGURE 3-34 Amplification of a dc signal with a chopper amplifier.
(1972~ other frequencies are rejected by the system.
PRINCIPLES OF INSTRUMENTAL ANALYSIS

with computers in order to facilitate signal appears as a small bright dot in the center of
I
Monochromotor
Synchronous averaging. Here, the measuring process is the screen.
SIm_PIe ~SignaIor demodulator triggered repetitively, and the signal is accu- Input signals are applied to two sets of
mulated in one or more memory channels. plates, one of which deflects the beam hori-
.J V.
v. After a suitable number of sweepS or measure- zontally and the other vertically. Thus, x-y
ments, an average for the signal is computed. plotting of two related signals becomes
v, The effect of averaging is shown in Figure possible. Because the screen is phosphores-
_-0'-v, l
_ 3-37; the signal· is part of the output from an
NMR spectrometer. With a single scan, only
cent, the movement of the dot appears as a
lighted continuous trace that fades after a
Refenmce
detector a few of the peaks are discernible because brief period.
they are roughly the same size as the random The most common way of operating the
noise. After averaging the data from 50 scans, cathode tube is to cause the dot to be swept
the peaks begin to emerge; averaging 200 periodically across the horizontal axis of the
scans provides an even better spectrum. tube. For this purpose, a sawtooth sweep
Double-throw switch signal (p. 12) is employed. The signal to be
opel'llted in p"-
with v"" measured is then applied across the vertical
plates. If the signal is dc, it will simply
displace the horizontal line to a position
A variety of devices are employed for convert- above or below the central horizontal axis.
ing the output from a signal processor to a In order to have a repetitive signal, such as
(bl form that can:be read and interpreted by a sine wave, displayed on the screen, it is
man. We have:already considered some of essential that each sweep begin at an identical
these in thesedtion of Chapter 2 devoted to place on the wave; that is, for example, at a
.,~ meters. We hav~ also seen how the position of
a sliding contact on a linear resistor can be
maximum, a minimum, or a zero crossing.
Synchronization is usually realized by mixing
employed to provide voltage data (p. 34) or a portion of the test signal with the sweep

'.'LEft
resistance data (p. 35). _ signal in such a way as to produce a voltage
In this section, we consider two more so- . spike for, say, each maximum or some mul-
phisticated readout devices, namely, the tiple thereof. This spike then serves to trigger
cathode-ray tube and the laboratory recorder. the sweep. Thus, the wave form can be ob-
served as a continuous image on the screen.
The Cathode-Ray Tube
(e)
The oscilloscope is a most useful and versatile Recorders'
FIGURE 3-36 Alock_inamplifiersystem:(a)s~stemfora
laboratory instrument that employs a
spectrophotometer, (b) synchronous demodulation (sche- The typical laboratory recorder is an example
cathode-ray tube as a readout device. Figure
matic), and (c) signal form. of a servosystem, a null device that compares
3-38 is a schematic drawing showing the main
two signals and then makes a mechanical
components of the tube. Here, the display is
adjustment that reduces their difference to
formed by the ;interaction of electrons in a
numerous measurements of its magnitude can zero; that is, a servosystem continuously
Signal Avereglng6 focused beam with a phosphorescent coating
be performed. If the measurements can all be seeks the null condition.
on the interior pf the large curved surface of In the laboratory recorder, shown schemat-
Signal averaging provides a powerful tool for made in exactly the same way~the signals.wi~1 the tube. Electtions are formed at a heated ically in Figure 3-39, the signal to .be
extracting a small signal fro~ larg~ ~ounts be additive, but the noise, Insofar as It IS
cathode, which is maintained at ground recorded, v,., is continuously compared With
of noise. To apply this techmque, It IS neces- random and tends to cancel, adds. as the
potential; an anode with a potential of the the output from a potentiometer powered by
sary that the signal be repeatable, so that square root of the number of measurements n.
order of kilovolts accelerates the electrons
Thus, the SIN ratio is improved by a factor of through a control grid and a second anode
~. . that serves to focus the beam on the screen. , For a good discussion of laboratory recorders, see:
Many modern instruments are Interfaced In the absence of input signals, the beam G. W. Ewing, J. Chem. £due .• 53. A361. A407 (1976).
IntensitY
control Focusing Accelerating
grid: •• ode anode

fereooe signal can be caused to drive the ser- INTERFACING OF


vomechanism to the null state from either INSTRUMENTS WITH COMPUTERS
direction.
In most laboratory recorders, the paper is Undoubtedly, one of the major developments
moved at a fixed speed. Thus, a plot of signal in analytical chemistry during the last decade
FIGURE 3-37 Effect of signal averaging. Note that the intensity as a function of time is obtained. has been the interfacing of digital computers
vertical scale is smaller as the number of scans increases. A good-quality laboratory recorder will with the various instruments listed in Table
That is, the noise grows in absolute value with increased provide a record of voltages that is accurate I-I. At least two reasons exist for connecting
number of scans; its value relative to the analytical signal to a few microvolts. a computer to an analytical instrument. The
decreases, however.

a reference signal, v".,. The latter may be one To understand the directional control of
or more mercury cells, the potential of which the motor, it is important to note that a
remains constant throughout the lifetime of reversible ac motor has two sets of coils, one
the battery; alternatively, a rectified ac signal, of which is fixed (the stator) and the other of
stabilized by a Zener diode (p. 53) serves as which rotates (the rotor). One of these, say,
the reference. Any difference in potential be- the rotor, is powered from the no-v house
tween the potentiometer output and Vx is line and thus has a continuously fluctuating
converted to a 6O-cycle ac current by a magnetic field associated with it. The output c,
mechanical chopper; the resulting signal is from the ac amplifier, on the other hand, is
then amplified sufficiently to activate a small fed to the coils of the stator. The magnetic .~_ c.
phase-sensitive electrical motor that is me- field induced here interacts with the rotor ,Qac +
chanically geared or linked (by a pulley
arrangement in Figure 3-39) to both a record-
field and causes the rotor to turn. The direc-
tion of motion depends upon the phase of the
] _
-
-
Chopper
amplifier
er pen and the sliding contact of the poten-
tiometer. The direction of rotation of the
motor is such that the potential difference be-
stator current with respect to that of the
rotor; the phase of the stator current,
however, differs by 180 deg, depending upon
,r
V.

tween the potentiometer and v.: is decreased whether Vx is greater or smaller than FIG U R E 3-39 Schematic diagram of a self-balancing recording
to zero, whereupon the motor stops. the signal fro~ v".,. Thus, the amplified dif- potentiometer.
first is that partial or complete automation of Binary Number Syoem (b) 19 10011
.measurements becomes possible. Ordinarily, +6 110
The programs, instructions, mathematical 1011000110
automation leads to more rapid data acquisi- operations, and memories of a digital com-
tion and shortens the time required for an 29 - 27 26 - - - 22 21 -
25 11001
puter are all based upon the binary system of
analysis or increaSes its precision by provid- (c) 7
numbers rather than on the more familiar deci- 111
ing time for additional replicate measure- Arithmetic with binary numbers is similar x3
mal system. The decimal system is, of course, 11
ments to be made. Automation, moreover, based on ten digits; the binary system uses to, but simpler than, decimal arithmetic. For
frequently provides better control over exper- addition, only four combinations are possible. 111
but two, 0 and 1.
imental variabl~ than a human operator can 111
Table 3-1 illustrates the relationship be-
achieve; more precise and accurate data are
tween some decimal and binary numbers. The 10101
the result.
example thaf follows illustrates methods for
A second reason for attaching computers (d) 22
binary-to-decimal conversion and the reverse. 10110
to instruments is to take advantage of their
x5 101
tremendous computational and data-handling
capabilities. These capabilities make possible 10110
EXAMPLE
the routine use of techniques that would or- (a) Convett 101011 to a decimal number. 0‫סס‬oo
dinarily be impractical because of their exces- 10110
(b) Convert 710 to a binary number.
sive computational time requirements.
(a) To understand how this conversio~ 1101110
Notable among such applications is the use of
can be made, it is useful to realize that a decJ-
Fourier transform calculations in spectros-
mal number can be readily broken down into The following example illustrates the use of
copy to extract small analytical signals from Bits. The binary digits 0 and 1 are often
integer multi~1es of powers of ~O. For exam- these operations.
noisy environments. referred to as bits. Thus, the decimal number
ple, the numljer 4753 can be wntten as
The interfacing of computers with analyti- EXAMPLE 110 in part (d) of the example is called a
cal instruments is too large and complex a 4753 = three-digit number in the decimal system; the
Perform the following calculations with
field to be treated in detail here.8 Thus, the binary arithmetic. (a) 7 + 3, (b) 19 + 6, (c) binary equivalent 1101110 is a seven-bit
4 x 103 + 7 X 102 + 5 X 101 + 3 x 10°
discussion that follows is limited to a general number.
7 x 3, (d) 22 x 2.
summary of the properties possessed by com- 4000 +700 + 50 + 3 Words. A word consists of one or more
puters and the advantages gained by interfac- By analogy, binary numbers can be expressed (a) 7 111 bits that are associated with one another. The
ing them with analytical instruments. in terms of base 2. Thus, +3 +11 number of bits per word depends upon the
For many purposes, it is convenient to size of the computer; some common sizes in-
categorize computers on the basis of size. 101011 = 1 x 2' + 0 X 24 + 1 X 23 10 1010 clude 8, 12, 16, 24, and 64. Minicomputers
Here; size refers not only to physical dimen- 32 + 0 + 8
sions but more importantly to the quantity of
+ 0 X 22 + 1 X 21 + 1 x 2°
information and instructions that can be
stored. This discussion will focus on so-called + 0 + 2 + 1 = 43
minicomputers and microcomputers rather (b) As a first step, we determine the largest TABLE 3-1 RELATIONSHIP BETWEEN SOME
than on the large, centralized, general-purpose power of 2 that is less than 710. Thus, since DECIMAL AND BINARY NUMBERS
device. 21o = 1024, ;
Decimal Binary Decimal Binary
29 = 512 and 710 - 512 = 198 Number Representation Number Representation
The process Is repeated for 198 o
• For monographs on compuler interfacing, see: Com- o 8 1000
purtrs for SpectroJcoputJ,ed. R. A. G. Carrington. New 27 = 128 and 198 - 128 = 70 1 1 9 1001
York: Wiley, 1974; J.• Finkel, Computer-Aided Exper- 2 10 10 1010
imentation. New York: Wiley-Interscience, 1974; J. S. Continuing, we find 3 11 12 1100
MatlJoD, H. B. Mark, Jr, and H. C. MacDonald, Jr, 4 100 15 1111
Chemu/$. New York: Marcel 26 = 64 and 70-64=6
Computer Fundamentabfor 5 101 16 1‫סס‬oo
Dekker, 1973; and A. J. Diefenderfer, PrincipleJ of Elec- 2
2 = 4 and 6- 4=2 6 110 32 10‫סס‬oo
tronic Instrumentation, 2d ed. Philadelphia: Saunders, 7 111 64 1000ooo
1979,Chapters 10-12. 21 = 2 and 2- 2=0
logic to perform the various functions of the Tiny single chips containing 1024 to 4096 of
often employ 16-bit words and microcom- Major Components of a Computer computer; generally, the switches control volt- these elements are employed. Binary storage
puters 8. Computer memories are described Figure 3-40 is a block diagram showing the ages that range from 0 ± 0.5 V for the off or . and retrieval depends upon which half of the
in terms of the number of words they can arrangement of the major components of a o position and 5 ± 2 V in the on or I posi- circuit is conducting at any time. An
retain. Thus, an 8K memory has space for digital computer. tion. The band between 0.5 and 3.0 V is integrated-circuit memory has the disadvan-
approximately 8000 words. The Central Proceaiag Unit. The heart of sufficiently large to prevent interference by tage of being volatile; that is, loss of electric
A single word contains a pattern of bits a computer is the central processing unit, random noise. power results in loss of the stored
that cause the computer to take a specific which contains a control unit and an arith- Memory. Storage of data and instructions information.
action. Generally, a word contains a datum for a computer take two forms. One makes Input-Output Systems. Input-output devices
metic unit. The former controls the overall
or a command and an address, both written use of magnetic tapes or disks. Access to these provide the means by which the user (or
sequence of operations by means of a
in binary notation. The latter gives the loca- program, which is stored in the memory unit is relatively slow, and they are used primarily his instrument) communicates with the com-
tion of an instruction or a datum word in the The control unit ~receives instructions from for storage of programs and data for future puter. Familiar input devices include the
memory. For example, for a computer em- use. electric typewriter, punch cards or tapes,
the input device, fetches instructions and data
ploying 12-bit words, a word might take the from the memory: unit, transmits instructions The main memory of computers is either a magnetic tapes or disks, and the electrical or
form to the arithmetic. unit for computation, and magnetic-core type or an integrated-circuit mechanical signals from analytical instru-
transmits the results of these computations to type. In both, instructions or data are stored ments. Output devices include recorders, elec-
the output and often to memory as well in binary form with a separate core or circuit tric typewriters, cathode-ray screens, and
Here, the first three bits could be the code for The program, the instructions, the math- being required for each bit. Generally, the meters. It is important to appreciate that
a command such as add, subtract, multiply, ematical operations, and the memory of a dig- individual cores or circuits are arranged in most of these devices provide or use an
or store. The next two, in this example, are ital computer are all based upon the binary square arrays of 32 x 32 or 1024 storage posi- alUllog signal while, as we have pointed out,
not used. The last seven bits give the location system made up of just two states, on and 00; tions. For a computer employing words made the computer can respond only to digital sig-
of the number upon which the command is to or 0 and 1. The \:cntral processing unit of a up of 16 bits; one of these arrays would corre- nals. Thus, an important part of the input-
be exercised. For example, the word just computer contaiJls a vast array of transis- spond to 64 words. The minimum memory output system is an analog-to-digital converter
shown might command the computer to add torized switches, ~ach of which has only two capacity required for most computational (ADC) for inputting data, which the com-
the number stored in the indicated location settings, 0 or I, corresponding to on or 01T. applications is 4096 words (a 4K memory). puter can use, and a digital-ta-analog counter
(storage number 74) to another number. These switches are employed using binary Thus, the number of individual binary cores (DAC) for converting the output from the
or circuits is enormous. computer to a usable signal.
The individual storage unit or core in a By definition, an analog signal is one that
magnetic-core memory is made of a ferrite varies continuously, whereas a digital signal is
ceramic, shaped in the form of a doughnut. one that varies in discrete increments. An
The inside and outside diameters of the ring example of an analog signal is the voltage
are 0.013 and 0.020 in~ respectively. When an output of a glass electrode, which varies con-
electrical signal passes through a wire that is tinuously from about 0 to about 1.0 V
threaded through the ring, magnetization of depending upon the pH of the solution in
the ferrite material occurs; the direction of which it is immersed. Most, but not al~ analyt-
magnetization (clockwise or counter clock- ical instruments are analog; examples of ex-
wise) depends upon the direction of the ceptions include the output from the decay of
current and remains fixed after the current a radioactive species or from a Geiger tube
ceases. Thus, the core has two stable states when it is employed to measure the intensity
and can store binary information. A sensing of an X-ray beam.

•• wire running through the ring is then Most output devices such as recorders,
employed to sense the direction of magnetiza- plotters, and meters require analog rather
tion and thus retrieve the information stored than digital signals.


in the element. Where the analog signal is the rotary
An integrated-circuit memory element gen- motion of a shaft, a digital signal is readily
erally consists of a circuit containing four produced by means of a mechanical turn
transistors, two capacitors, and two switches. counter. One method of achieving a digital
signal is to mount on the shaft a circular disk to errors. For this reason, assembly languages
used in more than one way and by more than the user is required to obtain automated in-
that is notched to give opaque and transparent have been developed in which the switch-
one user at the same time without interfer- strument control.
slots. An array of light beams and pho- setting steps are. assembled into groups which
ences. A single minicomputer is frequently Microcomputer memories tend to be small-
tocells then produces a series of digital electri- can be designated by code word. For exam-
used to control several instruments in one er than those employed in minicomputers.
cal pulses as the disk rotates. Such a device ple, the abbreviation for subtract might be
laboratory. Memory increments of 256 words are avail-
might be employed to monitor the SUB and might correspond to 101 in machine
MiniComputers can usually be pro- able; typically, eight of these are employed to
wavelength setting of a monochromator or language. Clearly, SUB is a good deal easier
grammed in one or more high-level lan- provide a capacity of 2048 words. Only one
the null point of a potentiometer. for the programmer to remember than 101.
guages. Often, however, interfacing them with or two of the memory increments is required
More commonly, analog-to-digital convert- Assembly programming, while simpler
instrument signals may require the user to for the control of many instruments. Word
ers are electronic and are based upon count- than machine programming, is still difficult
construct subroutines in assembly language. lengths for microcomputers are commonly
ing square-wave pulses from a generator. and tedious. As a consequence. a number of
The central processing unit in a mini- eight bits.
Here, the number of pulses or their frequency high-level languages, such as FORTRAN and
computer generally ranges in cost from 51000 'Microcomputers, quite adequate for the
is determined by the analytical signal. Alter- Basic, have been developed. These languages,
to 510,000. The addition of storage, input and control of a single instrument and with data-
natively, the time required for a ramp voltage which are easily learned, have been designed
output devices, and other peripherals usually acituisition speeds of 1000 to 3000 words per
to change from zero to a chosen level may be to make communication with the computer
doubles these costs at a minimum and may second, are available for 51000 or less.
measured. relatively straightforward. Here, instructions
multiply them by a factor as great as five to tIn more complex instruments, one or more
Digital-to-analog converters are electronic in Basic or FORTRAN are translated by a
ten. microcomputers may be controlled by a
devices that convert a digital number into a computer program (called a compiler) into
minicomputer. The slave microcomputer then
voltage proportional to the number. The volt- machine language which can then control the
Microcomput •••• performs one or more control functions at the
age can then be employed to drive a record- computer for computations. Unfortunately,
command of the minicomputer.
er, meter, or other output device. loss of efficiency accompanies the use of The heart of a microcomputer is the micro-
higher-level languages. processor, a device that first appeared on the
Figure 3-41 illustrates the application of market in about 1973; the applications of
Computer Programming .A~Plicetiona of Compute •••
the three languages for obtaining a sum. microprocessors have grown exponentially
Communication with a computer entails set- since that time.9 A microprocessor is a com- Computer interactions with analytical instru-
ting an enormous aggregation of switches to plete central processing unit, which has been ments are of two types, passive and active.tO
Minicomput ••.•
appropriate off or on positions. A program miniaturized and formed on a single In passive applications, the computer does
consists of a set of instructions as to how A minicomputer is sufficiently small and inex- integratcd-circuit chip having an area of a few not participate in the control or the experi-
these switches are to be set for each step in an pensive that its use can be limited to one lab- square millimeters. Combination of a micro- ment but is used only for data handling,
instruction or a computation. Each of these oratory or even to a single instrument. A processor with a memory and an input- processing, storing, file searching, or display.
instructions must be written in a form to minicomputer will have at least 4K or output device. leads to a remarkably In an active interaction, the computer is
which the computer can respond-that is, a memory and often considerably more. inexpensive microcomputer, which can be significantly involved in the control of the
binary mJJChine code. Machine coding is te- Usually, it can be used on a time-shtJred basis. conveniently employed in conjunction with a experiment. Instruments with computer con-
dious and time consuming and is often prone That is, its capabilities are such that it can be single instrument; such a computer is said to trol are said to be automated.
be dedicated. Most new analytical instruments Passive Applications. Data processing by a
are being sold with one or more programmed computer may involve relatively simple math-
microprocessors as an integral part of the ematical operations such as calculation of
equipment. With these, no programming by ~ncentrations, data averaging, least-square
analysis, statistical analysis, and integration
o 110001 100 100 001 LOA Z
to obtain peak areas; as we have seen. several
0100 001100 100010 Assembler ADA B

0100 001 100 100 011 AOAC


• For a summary of the construction and applications of
o 111 001 100 100 100 STA 0
microprocessors and minicomputers. see: R. E. Dessy. P.
Janse·Van. Vuuren, and J. A. Titus. AMI. Chem. 46 (11~
917A (1974); 46 (12~ 10SSA (1974); B. Soucek. Micro- 10 This c1assification has been IUgated by Perone in his
FIGURE 3-41 Relationships among machine, as- procnson and Microcompuler<. New York: Wiley. 1976; summary or computer applicatioDl. See: S. P. Perone
sembly, and a high-level language. (From S. Perone, and An Introduclion 10 Microcompulers. Adam Osborne and D. O. Jones. Digital Compuler III ScinIlific
and Associates. P.O. Box 2036. Berkeley. CA 94710. 1977. llUll'lUMntation. New York: McGraw·Hill, 1973. Chapter
J. Chromatog, Sci., 7, 715 (1969). With permission.)
vol O. I. and 2. 12.
of these operations can also be performed the computer exercises at least some and 1. Design a circuit whose output V. is given by
with operational amplifiers. More complex sometimes all of the control of an instrument
during an analysis. Most modem instruments
- V. = 2V, + 4V2 + 1V3 + 3V4
calculations may involve solution of several
simultaneous equations, curve fitting, and contain one or more microcomputers that 2. Design a circuit for calculating an average value for three voltage
Fourier transformations. perform such functions. Examples would in- inputs multiplied by 100.
Data storage is another important passive clude control of the slit width and wavelength 3. Design a circuit for performing the following calculation:
function of computers. For example, a power- setting ora monochromator, the temperature
ful tool for the analysis of complex mixtures of a chromatographic column, the potential y = !(2x, + 3X2)
is obtained by linking gas-liquid chromatog- applied to an electrode, the rate of addition of
raphy (GLe) with mass spectrometry (MS). a reagent, and the time at which the integra-
The former is capable of separating mixtures tion of a peak is to begin. Referring again to
on the basis of the time required for the indi- the GLCjMS instrument considered in the R" Rf2
vidual components to appear at the end of last section, a computer is often used to ini-
suitably packed columns. Mass spectrometry tiate collection of mass spectral data each R,
permits identification of each co~ponent time a compound is sensed at the end of the V,
according to the mass of the fragments chromatographic column. R.
V.
formed when the compound is bombarded Computer control can be relatively simple,
with a beam of electrons. GLCjMS equip- as in the examples just cited, or more .".
ment may produce as many as 100 spectra in complex. For example, the determination of R3
V3
less than an hour, with each spectrum being the concentration of elements by atomic emis-
made up of tens to hundreds of peaks. Con- sion involves the measurement of the heights
version of these data to an interpretable form of emission peaks that occur at wavelengths
is time consuming; thus, the data must be characteristic for each element. Here, the I
(a) write an expression giving the output voltage in terms of the three
stored in digital form for subsequent printing. computer can cause a monochromator to
input voltages and the various resistances. .
Identification of a species from its mass sweep a range of wavelengths until a peak is
(b) indicate the mathematical operation performed by the circuit
spectrum involves a search of files of spectra located, and can then seek the exact
when R1 = Rfl = 100 kn; R4 = Rf2 = SOkn; R2 = 2S kn;
for pure compounds until a match is found; wavelength at which the maximum output
R3 = 10 kn.
this process is also time consuming, but can signal is obtained. Measurements are made at
be accomplished quickly by using a this point until an average is obtained that S. Show the algebraic relationship between the voltage input and output
computer. Here, a magnetic disc or magnetic will give a suitable signal-to-noise ratio. The for the following circuit:
tape file storage is required. For example, computer then causes the instrument to
programs have been described which permit repeat this operation for each peak of interest
the search of several thousands of spectra in a in the spectrum. Finally, the computer
minute or less. Such a search will frequently processes the data and prints out the concen-
produce several possible compounds. Further tration of the elements present. 3kU
comparison of spectra by the scientist usually Because of its great speed, a computer can v,
makes identification possible. often control variables more efficiently than SkU
v,
Another important passive application uti- can a human operator. Furthermore, with 4kU
lizes the high-speed data fetching and cor- some experiments, a computer can be pro- V3
relating capabilities of the computer. Thus, grammed to alter the way in which the mea- 6kU
V.
for example, the computer can be called upon surement is being made, according to the
to display on a cathode-ray screen the spec- nature of the initial data. Here, a feedback
trum of anyone of the components after it loop is employed in which the signal output is
has exited from a gas chromatographic fed back through the computer and serves to 6. What would be the output voltage of the circuit shown in Figure
column. control and optimize the way in which addi- 3-26a after 30 s, if the input voltage is 2S mY, the resistor is 200 kQ,
Active Applicatiol& In active applications, tional measurements are made. and the capacitor is 0.26 JlF?
12. The linear slide wire AB has a length of 100 an. Where along its
length should contact C be placed in order to provide a potential of
exactly 2.50 V?

8. Show that when the four resistances are equal. the following circuit
becomes a subtracting circuit.

13. Design a circuit that will produce the following output:

V. =2 .f VJ dt +3 f V2 dt

14. Design a circuit that will produce the following output:

V. =3 f VI dt - 4(V2 + V3)
9. Derive a relationship between v..
I't. and ~ for the circuit shown in 15. The following data were obtained for repetitive weighings ofa 12S4-g
Problem 8 when Ru = Ru and RJ = R2• object with a top-loading balance. Calculate the signal-to-noise ratio
10. Derive a relationship between V. and l'/ for the following circuit: for the measurements, assuming the noise to be random.
1253 g 12S6g
1256 1.252
1.257 1.254
1.254 1.255
16. How many measurements would have to be averaged to increase SIN
in Problem 15 to l000?
17. The following data were obtained for a current measurement on a
noisy system. What is the signal-to-noise ratio. assuming that the
noise is random?
1.34 flA 1.10 flA
1.76 1.63
121 1.77
1.35 1.19

18. How many measurements would have to be averaged to obtain a


~v. signal-to-noise ratio of 10 for the mean in Problem 16; to obtain a
ratio of l00?
19. The resistance of a dc circuit is 1500 0. What is the maximum resist-
ance that a current-measuring device can havc if thc current in this
circuit is to be measured with a relativc error of less than 2 %? Thc
cxpected current range is 0 to 20 pA.
20. Devise an operational amplifier circuit employing a 0 to 10 mV mctcr
that will meet the requirements specified in Problem 20.
21. Thc resistance of a dc circuit is 200 0. What is thc maximum resist-
ance that a current-measuring device can havc if thc current in this.
circuit is to be measured with a relative error of less than 5 %? The
expected current range is 0 to 50 pA.
22. Devise an operational amplifier circuit employing a 0 to 10 mV meter:
that will meet the requiremcnts specified in Problem 21. .
the zero axis. Normally, only' the electrical maximum. This velocity, given the symbol c,
This chapter reviews several topics concerned as particles and waves are not mutually exclu- vector is employed because it is the electriclJl has been accurately determined to be
with electromagnetic radiation, including the sive. Indeed, the duality is found to apply to force that is responsible for such phenomena 2.99792 X 1010 em/s. Thus, for a vacuum,
fundamental properties of this radiation and the behavior of streams of electr~ns or. other as transmission, reflection, refraction, and
the mechanism of its interactions with matter. elementary particles as well and IS readdy ra- absorption of radiation. . . c = vl ~ 3 x 1010 emls (4-2)
The material will serve as a general introduc- tionalized by wave mechanics. Figure 4-1 is a plot ofthe electr.ical vector
In any other medium, the rate of propagation
tion to the various instrumental methods dis- of monochromatic plane-polarized radiation, is less because of interactions between the
cussed in Chapters 5 through 14} Wave Properties the simplest type. The first term indicates that electromagnetic field of the radiation and the
For many purposes, electromagnetic. ~ia- the radiation has a single. frequency; the bound electrons in the atoms or molecules
tion is conveniently treated as an OSCIllatmg second implies that it oscillates in but a single making up the medium. Since the radiant
PROPERTIES OF electrical force field in space; associated ~~ plane in space.
frequency is invariant and fixed by the ~u~ce,
ELECTROMAGNETIC RADIATION Wave Parameters. The time required for
the electrical field, and at right angles to It, IS the wavelength must decrease as radiatIon
a magnetic force field. . the passage of successive maxima through a passes from a vacuum to a me<!ium con~ai?-
Electromagnetic radiation is a type of energy fixed point in space is called the period p of
The electrical and magnetic fields ass0- ing matter (Equation 4-1). ThiS effect IS il-
that is transmitted through space at enor- the radiation. The frequency v is the number
ciated with radiation are vector quantities; at lustrated in Figure 4-2.
mous velocities. It takes numerous forms, the of oscillations of the field that occurs per
any instant, they can be represented by an It should be noted that th~ velocity of ra-
most easily recognizable being light and ra- second2 and is equal to lip. It is important to
arrow whose length is proportio~ t~ t~ diation in air differs only slightly from c (about
~iant beat. Less obvious manifestations include realize that the frequency is determined by the
magnitude of the force and whose direction IS 0.03% less); thus, Equation 4-2 is usually
X-ray, ultraviolet, microwave, and radio source and remaillS invariant regardless of the
parallel to that of the force. A ~phical rep- applicable to air as well as to a vacuum.
radiations. . media traversed by the radiation. In contrast,
resentation of a beam of radiation can be The wavenumber (1 is defined as the
Many of the properties of electromagnetIC the velocity of propagation v,, the rate at'
obtained by plotting one of these vect.or number of waves per centimeter and is yet
radiation are conveniently described by which the wave front moves through a
quantities as a function. of ~ime as the radia- another way of describing ielectromagnet~c
means of a classical wave model which em- medium, is dependent upon both the medium
tion passes a fixed potnt tn space. Alterna- radiation. When the wavelength in vacuo IS
ploys such parameters ~ wavelength, fre- and the frequency; the subscript i is employed
tively, the vector can be plotted as a function expressed in centimeters, the' wavenumber is
quency, velocity, and amplItude. In contrast to to indicate this frequency dependence.
other wave phenomena, such as sound,
of distance, with time held constant. equal to Iii..,",,"", ')
In Figure 4-1, the electrical vector .serves as Another parameter of interest is the wave- Radiant Power or Intensity. The power P
electromagnetic radiation req~ir~ no suI: length Ai' which is the linear distance between
the ordinate; a plot of the magnetIC vector of radiation is the energy of the beam that
porting medium for its transmIssIon; thus, It successive maxima or minima of a wave.J
would be identical in every regard except that -' reaches a given area per second; the intellSity
readily passes through a vacuum. Multiplication of the frequency in cycles per
the ordinate would be rotated 90 deg around '- 1 is the power per unit solid angle. These
The wave model fails to account for phe- second by the wavelength in centimet~ ~ quantities are related to the square of ~h~am-
nomena associated with the absorption or cycle gives the velocity of propagatton .m plitude A (see Figure 4-1). Although .It IS ~ot
emission of radiant energy; for these centimeters per second; that is, strictly correct to do so, power and mtenslty
processes, it is necessary to view e~ectro-
Vi = VA; (4-1) are often used synonymously.
magnetic radiation as a stream of dISCrete I
A---., •
_Wavelength. Mathematical Description of a Wave. With
particles of energy called photons. The energy I
I
I
, In a vacuum the velocity of radiation be-
I time as a variable, the wave in Figure 4-1 can
of a photon is proportional ~o the freq~e~cy comes independent of frequency and is at its be described by the equation (see Equation
of the radiation. These dual vIews of radtatlon
2-18, p. 12)
y = A sin(wt +< I/J) (4-3)
z The common unit of frequency is the hertz Hz, which is
I For a further review of these topics, see: E. J. Bair,
equal to one cycle per second. where y is the electric force, -1 is the amplitude
Introtlucrwn ro Chemical I ••.•,rumen,a,ion. New York:
or maximum value for y, t ~s time, and I/J is
McGraw-HilI, 1962, ChapletS 1-2; E, J. Bowen, The 3 The units commonly used for describing wavelength
Chemical hpecl!l of Light, 2d ed. Oxford: The Clarendon differ considerably in the various spectral regions. For
the phase angle, a term which has ~n
Press, 1946; E. J. Meehan, Treatise on Analytical Chem- example, the AngstrOm unit, A (10· I.m), is eonvenient for defined earlier (p. 12). The angular velOCIty
islry, eels. I. M. Kolthofl' and P. J. Elving, Part I, vol. ~,
FIGURE 4-1 Representation of a beam X-ray and short ultraviolet radiation; the nan~meter, ~ of the vector, w, is related to the frequency of
New York: Wiley, 1964, Chapter 53; and. F. Grum m
of monochromatic, plane-polarized radia-
(10·' m), is employed with visible ~nd ultraVlolet ~adla- the radiation v by the equation
Physical Me/hods of Chemistry, eels. A. WeJSSberger and tion; the micrometer,/lm (10·' m), IS useful for the mfra-
B. W. Rossiter. New York: Wiley-Interscience, 1912, Vol. tion. The arrows represent the electrical red region. w = 2ltv
I, Part III B, pp. 214-305. vector.
or cosine terms such as those shown in Equa- approximate the shape of a square wave. As
tion 4-5. For example, the square wave form shown by the solid line in Figure 4-4b, the
often employed in electronics can be repre- resultant more closely approaches a square
sented by an equation having the form wave when nine waves are incorporated. .
Mathematically, resolution of a complex
y = A (sin 2nvt + i sin 67tvt wave form into its sine or cosine components .t
(4-6) is tedious and time consuming; modern com-
puters, however, have made it practical to ex-
+! sin to nvt + ... + ~ sin 2n7tvt) ploit the power of the Fourier transformation
on a routine basis. The application of this
A graphical representation of the summation
Distance
technique will be considered in the discussion
process is shown in Figure 4-4. The solid line
of several types of spectroscopy.
FIGURE 4-2 Effect of change of medium on a monochromatic beam of in Figure 4-4a is the sum of three sine waves
Ditrraction. Figure _4-5 is a schematic rep-
radiation. differing in amplitude in the ratio of S : 3 : I
resentation of diffraction, which occurs
and in frequency in the ratio of 1 : 3 : S. Note
whenever waves of any type pass by a sharp
that the resultant is already beginning to
barrier or through a narrow opening. Diffrac-

Substitution of this relationship into Equa- waves. Figure 4-3b differs from 4-3a in that
tion 4-3 yields the difference in phase angle is greater; here,
the resultant amplitude is smaller than those
y = A sin(2nvt ~ tfJ) (4-4)
of the component waves. Clearly, a maximum
Superposition of Wa,ts. fhe principle of amplitude for the resultant will occur when
superposition states that when two or more the two waves are completely in phase; this
waves traverse the same space, a displacement situation prevails whenever the phase differ-
occurs which is the sum of the displacements ence between waves (tfJI - 4>2) is 0 deg, 360
caused by the individual waves. This principle deg, or an integer multiple of 360 deg. Under
applies to electromagnetic waves, where the these circumstances, maximum constructive
displacements involve an electrical force field, interference is said to occur. A maximum de-
as well as to several other types of waves, structive interference occurs when (tfJI - tfJ2) is
where atoms or molecules are displaced. For equal to 180 deg or 180 deg plus an integer ----Time---~ ----Time,----
example, when n electromagnetic waves of the multiple of 360 deg. The property of interfer-
(al (b)
same frequency but differing amplitudes and ence plays an important role in many in-
phase angles pass some point in space simul- strumental methods based on electromagnetic
taneously, the principle of superposition and radiation.
Equation 4-4 permits us to write Figure 4-3c depicts the superposition of
two waves with the same amplitude but dif-
y = Al sin(2nvt + tfJd ferent frequency. The resulting wave is no
+ A'2 sin(2nvt t 4>2) + ... (4-S) longer sinusoidal. It does, however, exhibit a
periodicity. Thus, the wave from 0 to B in the
+ A. sin(2nvt + 4>.) figure is identical to that from B to C.
. i
where y IS the resultant for~ field. An important aspect of superposition is
The solid line in Figure 4-3a shows the that the process can be reversed by a math-
application of Equation 4-S to two waves of ematical operation called a Fourier transfor-
identical frequency but somewhat different mation. Jean Fourier, an early French (el
amplitude and phase angle. Note that the re- mathematician (1768-1830~ demonstrated FIGURE 4-3 Superposition of sinusoidal waves: (a) Al < A2, (tfJI - tfJ2) = -20°,
sultant is a sine wave with the same frequency that any wave motion, regardless of complex- VI = V2; (b) Al < A2, (4)1 - 4>2) = -200°, VI = V2; (c) Al = A2, 4>1 = 4>2' VI = I.S V2' In
but different amplitude from the component ity, can be described by a sum of simple sine each instance, the solid curve is the resultant from combination of the two dashed curves.
lion is readily observed by generating waves
of constant frequency in a tank of water and
observing the wave crests before and after
Superposition of 3 .ine _ they pass through a rectangular opening or
y - Al.in 2m. t sin 81M. t sin 1000l slit. When the slit is wide relative to the wave-
length of the motion (Figure 4-Sa~ diffraction
is slight and difficult to detect. On the other
hand, when the wavelength and the slit open-
ing are of the same order of magnitude, as in
Figure 4-Sb, diffraction becomes pronounced.
Here, the slit or opening behaves as a new
source from which waves radiate in a series of
nearly 180-deg arcs. Thus, the direction of the
wave front appears to bend as a consequence
of passin.s the two edges of the slit.
- Diff~ion is a consequence of interfer-
ence. This relationship is most easily under-
stood by considering an experiment, per-
formed first by Thomas Young in 1800, by
which the wave nature of light was unambig-
uously d~monstrated. As shown in Figure 4-6,
a paralI4I beam of radiation is allowed to
pass through a na~row slit A (or in Young's
/ "\ experiment, a pinhole) whereupon it is dif- lbl
'V' \
\ Su-,tion of 9 .ine_
fracted and illuminates more or less equally FIGURE 4-5 Propagation
two closely spaced slits or pinholes Band C; of waves through a slit.
, \ - y-Ahin2lh't.tsin8""t ••..• ;!; sin 34,,,,,) the radiation emerging from these slits is then (a) x)' ~ l; (b) x)';:: l.
,""'-
, Superposition of 3 sine _
\
\
observed on the screen lying in a plane XY.1f
the radiation is monochromatic, a series of
, y - Alsin 2•.vr • t .in 8•.vr • t sin 10,M) darlc and light images perpendicular to the
plane of the page is observed. in the other light bands, are readily derived.
Figure 4-6b shows the intensities of the The angle of diffraction (J is the angle from
various bands reaching the screen. If, as in the normal formed by the dotted line extend-
this diagram, the slit widths approach the ing from a point, 0, halfway between the slits,
wavelength of radiation, the band intensities to the point of maximum intensity, D. The
decrease only gradually with increasing dis- solid lines BD and CD represent the light
tances from the central band. With wider slits, paths from the slits B an<!..f to this point.
the decrC'ase is much more pronounced. Ordinarily, the distance OE is enormous
The existence of the central band E, which compared to the distance between the slits
lies in tht shadow of the opaque material se- BC; as a consequence, the lines BD, OD, and
parating the two slits, is readily explained by CD are, for all practical purposes, paralleL
noting that the paths from B to E and C to E Line BF is perpendicular to CD and forms
lb)
are identical. Thus, constructive interference the triangle BCF, which is, to a close approxi-
of the diffracted rays from the two slits mation, similar to DOE; consequently, the
FIGURE 4-4 Superposition of sine waves to form a square wave:
occurs, and an intense band is observed. With angle CBF is equal to the angle of diffraction
(a) combination of three sine waves, and (b) combination of three, as in
(a), and nine sine waves. the aid of Figure 4-6c, the conditions for max- 9. We may then write
imum constructive interference, which result CF= BCsin (J
Substitution into Equation 4-7 gives vidual events, each of which lasts on the order
of 10-1 second. Thus, a beam of radiation
n.l. - ~ ;;;:~ (4-8) Crom this type of source is not continuous, as
DD DE . is the case in microwave or laser radiation;
).
/ Equation 4-8 permits the calculation of.~ instead, it is composed of a series of wave
I- wavelength Crom the three measurable trains that arc a few meters in length at most
quantities. Because the processes that produce a train
are random, the phase differences among the
EXAMPLE
Suppose that the screen in Figure 4-6 is trains must also be variable. A wave train
2.00 m Crom the plane of the slits and that the from slit B may arrive at a point on the
slit spacing is 0.300 mm. What is the wave- screen in phase with a wave train from C, and
length of radiation if the fourth band is constructive interference will occur; an in-
located 15.4 mm from the central band? stant later, the trains may be totally out of
DiffrllCtion Substituting into Equation 4-8 gives phase at the same point, and destructive inter-
byl ference will occur. Thus, the radiation at all
slngllsfit
' _ 0.300 )( 15.4 points on the screen is governed: by the
4
A 2)( 1000 mm random phase variations lU1,1ongthe wave
.l.- 5.78 )( 10-· mm or 578 mn trains; uniform illumination, which represents
an average for the trains, is the result.
Coherent Wiatioa. In order to produce a Sources do exist which produce- electro-
diffraction pattern such as that shown in magnetic radiation in the form of tr~ins with
Figure 4-6a, it is ncccssary that the clcctro- essentially infinite length and consfant fre-

f I Ii n n
magnetic waves traveling from slits B and C quency. Examples include radio frequency
to any given point on the screen (such as D or oscillators, microwave sources, optical lasers,
E) have sharply defined phase differences that and various mechanical sources such as a

~olLVJJJlliL
X-Diltlncl-Y
remain entirely constant with time; that is,
the radiation from slits Band C must be
coherent. The conditions for cohcrcncc are:
two-pronged vibrating tapper in a riffle tank.
When two of these are employed as sources
for the experiment shown in Figure 4-6, a reg-
(bl (el (1) the two sources of radiation must have ular diffraction pattern is observed.
FIGURE 4-6 Diffraction of monochromatic radiation by identical frequency and wavelength (or sets of Diffraction patterns can be obtained from
slits. frequencies and wavelengths); and (2) the 'random sources, such as a tungsten filament,
phase relationships between the two beams provided that an arrangement similar to that
must remain constant with time. The neces- shown in Figure 4-6a is employed. Here, the
sity for these requirements can be demon- very narrow slit A assures that the radiation
strated by illuminating the two slits in reaching Band C emanates from the same
Because BC is so very small compared to DE, for the light bands surrounding the central Figure 4-6a with individual filament lamps. small region of the source. Under these cir-
FD closely approximates BD, and the dis- band is Under these circumstances, the well-defined cumstances, the various wave trail1$ exiting
tance IT is a good measure of the difference n.l.= BC sin 8 (4-7) light and dark patterns disappear and are from slits Band C have a constant set of
in path lengths of beams BD and CD. For the replaced by a more or less uniform illumina- frequencies and phase relationships to one
two beams to be in phase at D, it is necessary where D is an integer called the order of
interference. tion of the screen. This behavior is a con- another and are thus coherent. If the slit at A
that CF correspond to the wavelength of the sequence of the incoherent character of is widened so that a larger part of the source
radiation; that is, The linear displacement DE of the dif-
fracted beam along the plane of the screen is filament sources (many other sources of is sampled, the diffraction pattern becomes
.l.z= CF z= BC sin 8 a function of the distance DE between the electromagnetic radiation are incoherent" as less pronounced because the two beams arc
screen and the plane ofthe slits, as well as the well). only partially coherent. If slit A is made
Reinforcement would also occur when the In incoherent sources, light is emitted by sufficiently wide, the incoherence may become
additional path length corresponds to 2)., 3)., spacing between the slits BC; that is,
individual atoms or molecules, and the result- great enough to produce only a constant illu-
and so forth. Thus, a more general expression DE= OD sin 8 ing beam is the summation of countless indi- mination across the screen.
ELECTROMAGNETIC RADIATION 101

Wavenumber. W••••• ngth. Frequency.


PlU1:icleProperti •• of R_iation beam front, but rather is concentrated at cer- Energy

Energy of EIedromagnetic Radiation. An
tain points or in particles of energy. Electron
volts.
a
--- A
Type
rad. Type Type
The work w required to cause emission of kCII/moi .V em-' em Hz tion IpeCtroIcopy
understanding of certain interactions between quantum transition
electrons is characteristic of the metal. The
radiation and matter requires that the radia-
alkali metals possess low work functions and 9.4 X 10' 4.1 X loa 3.3 X 10'· 3 X 10-11 1021
t
tion be treated as packets of energy called
photons or quanta. The energy of a photon
emit electrons when exposed to radiation in
the visible region. Metals to the right of the
Gemma
I'IY G"""""roy
emission
t
Nucl.lr
depends upon the freql,lency of the radiation,
and is given by
alkali metals in the periodic chart have larger
work functions and require the more ener-
getic ultraviolet radiation to exhibit the
9.4 X 10' 4.1 X 10' 3.3 X 10· 3 X 10-" 10'·
~ f
X..-oy
Ibsorption
omission
l r
Electronic
(inner sh.lI)
photoelectric effect. As we shall note in later
where h is Planck's cOnstant (6.63 x 10-27
erg see). In terms of w~velength,
chapters, the photoelectric effect has great
9.4 X 10" 4.1 X 10" 3.3 X loa 3 X 10-' 10"
1 1
practical importance in the detection of radia-
he tion using phototubes.
E=- Energy Units. The energy of a photon that
;1
is absorbed or emitted by a sample of matter 9.4 X 10' 4.1 X 100 3.3 X 10' 3 X 10-5 10'"
V VK
UV. absorption t
EllCtronic
Thus, an X-ray photon (1- 10-8 em) has ap- can be related to an energy difference between

fl'::'r
(outer sh.lI)
U\ }~ion
proximately 10,000 times the energy of a two molecular or atomic states, or to the
photon emitted by a hot tungsten wire frequency of a molecular motion of a constit- Yr· !
(1- 10-4 em~ uent of the matter. For this reason, it is often It absorption.
TIle Photoelectric ~ffed. The need for a convenient to describe radiation in energy
9.4 X 10-' 4.1 X 10-2 3.3 X 102 3 X 10-' 10"
1 1m1n Molecullr
vibration
particle model to de¢ribe the behavior of units, or alternatively in terms of frequency
electromagnetic radiation can be seen by con- (Hz) or wavenumber (cm -1), which are
1 f
Molecular
sideration of the photoelectric ~fJect. When directly proportional to energy. On the other t

- n:_I~
9.4 X 10-' 4.1 X 10-4 3.3 X 10· 3 X 10-' 10" rotation
Microwave
sufficiently energetic radiation impinges on a hand, the experimental measurement of radia- Micro-

metallic surface, electrons are emitted. The tion is most often expressed in terms of the
energy of the emitted electrons is found to be reciprocally related wavelength units such as
9.4 X 10-1 4.1 X 10'" 3.3 X 10-2 3 X 10' porlmlQnellC
related to the frequency of the incident radia- centimeters, micrometers, or nanometers. The 10·
resonence Magnetically
tion by the equation chemist must become adept at interconver-
E = hv - w (4-11)
sion of the various units employed in spec-
troscopy. Conversion factors commonly
9.4 X 10-'
Radio
J•
Nuclear magnetic
resonenc.
1 ~
--
spin states

4.1 X 10-· 3.3 X 10" 3 X 10' 10'


where w, the work function, is the work encountered for transformations are found
required to remove the electron from the FIGURE 4-7 Spectral properties, applications, and interactions of electromagnetic
inside of the front cover of this text.
radiation.
metal to a vacuum. While E is directly depend- The electron volt (eV) is the unit ordinarily
ent upon the frequency, it is found to be employed to describe the more energetic
totally independent of the intensity of the X-ray or ultraviolet radiation. The electron
beam; an increase in intensity merely causes volt is the energy acquired by an electron in sions. A logarithmic scale has been employed molecular or atomic transitions r~ponsible
an increase in the number of electrons emitted falling through a potential of one volt. Ra- in this representation; note that the portion for absorption or emission of radiation in
with energy E. diant energy can also be expressed in terms of to which the human eye is perceptive is very each region are also indicated.
Calculations indicate that no single elec- energy per mole of photons (that is, Avoga- small. Such diverse radiations as gamma rays
tron could acquire sufficient energy for ejec- dro's number of photons). For this purpose, and radio waves are also electromagnetic ra-
tion if the radiation striking the metal were units of kcal/mol or cal/mol are convenient. diations, differing from visible light only in the THE INTERACTION OF
uniformly distributed over the surface; nor matter of frequency, and hence energy. RADIATION WITH MATTER
could any electron accumulate enough energy The Electromegnetlc Spectrum Figure 4-7 shows the regions of the spec-
for its removal in a reasonable length of time. The electromagnetic spectrum covers an im- trum that are useful for analytical purposes As radiation passes from a vacuum through
Thus, it is necessary to assume that the mense range of wavelengths or energies. and the names of the spectroscopic methods the surface of a portion of matter, the electri-
energy is not uniformly distributed over the Figure 4-7 depicts qualitatively its major divi- I associated with these applications. The cal vector of the radiation interacts with the

_1_-
atoms and molecules of the medium. The Iating atoms, ions, or molecules as inter- normal dispersion region, there: is a gradual
nature of the interaction depends upon mediates .. increase in refractive index with increasing
the properties of the matter and may lead to One would expect the radiation from each frequency (or decreasing wavelength). Anoma-
transmission, absorption, or scattering of the polarized particle in a medium to be emitted lous dispersion regions are those frequency
radiation.· 'in all directions. If the particles are smal~ ranges in which a sharp change in refractive
however, it can be shown that destructive index is observed. Anomalous dispersion
Transmlulon of Rlldietlon interference prevents the propagation of always occurs at frequencies that correspond
significant amounts in any direction other. to the natural harmonic frequency associated
It is observed experimentally that the rate at with some part of the molecule, atom, or ion
than that of the original light path. On the
which radiation is propagated through a trans- . of the substance. At such a frequency, per-
other hand, if the medium contains large par-
parent substance is less than its velocity in a manent energy transfer from the radiation to
ticles (such as polymer molecules or colloidal.
vacuum; furthermore, the rate depends upon particles 1this destructive effect is incomplete: the substance occurs and absorption of the
the kinds and concentrations of atoms, ions, and a portion of the beam is scattered as a. beam is observed. Absorption is discussed in
or molecules in the medium. It follows from consequence of the interaction step. Scatter-' a later section.
these observations that the radiation must ing is considered in a later section of this' Dispersion curves are important in the FIGURE 4-9 Refraction
interact in some way with the matter. Because choice of materials for the optical compo- of light in passing from a
chapter.
a frequency change is not observed, however, Dispersion. As we have noted, the velocity' nents ofinstruments. A substance that exhibits less dense medium M1 into
the interaction CQllIIOt involve a permanent of radiation in matter is frequency-dependent; normal dispersion over the wavelength region a more dense medium M2,
energy transfer. since c in Equation 4-12 is independent of this of interest is most suitable for the manufac- where its velocity is lower.
The refractive index of a medium is a
parameter, the refractive index of a substance ture of lenses, for which a high and relatively
measure of its interaction with radiation and must also change with frequency. The varia'; constant refractive index is desirable. Chro-
is defined by
tion in refractive index of a substance with' matic aberration is minimized through the
frequency or wavelength is called its disper- i choice of such a material. In contrast, a sub- Reflection and
sion. The dispersion of a typical substance is: stance with a refractive index that is not only Scattering of Rlldletlon
shown in Figure 4-8. Clearly, the relationship' large but also highly frequency-dependent is
selected for the fabrication of prisms. The RellectioL When radiation crosses an
where IIj is the refractive index at a specified is complex; generally, however, dispersion
frequency i, Vi is the velocity of the radiation plots exhibit two types of regions. In the applicable wavelength region for the prism interface between media with differing refrac-
tive indexes, reftection occurs. The fraction
in the medium, and c is its velocity in vacuo. thus approaches the anomalous dispersion
region for the material from which it was reflected becomes larger with increasing dif-
The interaction involved in the transmis-
fabricated. ferences in refractive index; for a beam travel-
sion process can be ascribed to the alternat-
Refraction of Radiation. When radiation ing normally to the interface, the fraction
ing electrical field of the radiation, which
reflected is given by
causes the bound electrons of the particles passes from one medium to another of differ-
contained in the medium to oscillate with re- ing physical density, an abrupt change in di-
rection of the beam is observed as a
I, (n2 - nd2
spect to their heavy (and essentially fixed) - = ~-~ (4-14)
nuclei; periodic polarization of the particles consequence of differences in the velocity of 10 (n2 + nd2
thus results. Provided the radiation is not ab- . the radiation in the two media. This refraction
where 10 is the intensity of the incident beam
sorbed, the energy required for polarization is of a beam is illustrated in Figure 4-9. The
and I, is the reflected intensity; nl and "2 are
only momentarily retained (10-14 to 10-15 s) extent of refraction is given by the
the refractive indexes of the two media.
by the species and is reemitted without altera- relationship
tion as the substance returns to its original
state. Since there is no net energy change in EXAMPLE
this process, the frequency of the emitted ra- Calculate the percent loss of intensity due
diation is unchanged, but the rate of its propa- to reflection of a perpendicular beam of
gation has been slowed by the time required If the medium M1 is vacuum, VI becomes c yellow light as it passes through a glass cell
for retention and reemission to occur. Thus, and nl is unity (Equation 4-12); thus, the re- containing water. Assume that for yellow ra-
Frequency, Hz
transmission through a medium can be fractive index n2 of M2 is simply the ratio of diation the refractive index of glass is 1.5, of
viewed as a stepwise process involving oscil- FIGURE 4-8 Typical dispersion curve. the sines of the two angles. water is 1.33, and of air is 1.00.
In passing from air to glass, we find intense to be seen by the naked eye (the Tyn-
dall effect~ components of AB. A polarizing sheet, re-
!.!. = (1.50 -1.00f = 0.040 or 4.0% Scattering by molecules or aggregates of gardless of orientation, removes approxi-
10 (1.50 + l.00f molecules with dimensions significantly small- mately half of the radiation from an
er than the wavelength of the radiation is unpolarized beam, and transmits the other
In passing from glass to water, we find the
caIIcd Rayleigh scattering; its intensity is half as a plane-polarized beam. The plane of
loss to be
readily related to wavelength (an inverse polarization of the transmitted beam is
1, (1.50 - 1.33)2 0004 4. fourth-power effect~ the dimensions of the dependent upon the orien~ation of the sheet
10 = (1.50 + 1.33)2 =. or O. % scattering particles, and their polarizability. with respect to the incident beam. When two
An everyday manifestation of scattering is the polarizing sheets, oriented at 90 deg to one
The same losses would occur as the beam blueness of the sky, which results from another, are placed perpendicular to the
passes from the water and through the glass. the greater scattering of the shorter wave- beam path, essentially no ~radiation is trans-
Thus, lengths of the visible spectrum. mitted. Rotation of one results in a contin-
Scattering by larger particles, as in a col- FIGURE 4-10 (a) A few of the electrical uous increase in transmission until a
total percent loss = 2(4.0 + 0.4) = 8.8 loidal suspension, is much more difficult to vectors of a beam traveling perpendicular to maximum is reached when molecules of the
treat theoretically; the intensity varies roughly the page. (b) The resolution of a vector in two sheets have the same ltlignment.
It will become evident in later chapters as the inverse square of wavelength. plane XY into two mutually perpendicular The way in which radIation is reflected,
that losses such as those shown in this Measurements of scattered radiation can components. (c) The resultant when all vectors scattered, transmitted, and refracted by cer-
example are of considerable significance in be used to determine the size and shape of are resolved (not to scale). tain substances is also dependent upon direc-
various optical instruments. polymer molecules and colloidal particles. tion of polarization. As a consequence, a
Reftective losses at a polished glass or The phenomenon is also I1tilized in nephelom- group of analytical methods have been
quartz surface increase only slightly as the etry, an analytical method considered in developed whose selectivity. is based upon the
angle of the incident beam increases up to Chapter 10. different scale from Figure 4-lOa or 4-lOb to interaction of these substances with light of a
about 60 deg. Beyond this figure, however, Raman Scattering. The Raman effect dif- keep its size within reason. particular polarization. These methods are
the percentage of radiation that is reflected fers from ordinary scattering in that part of Removal of one of the two resultant planes considered in detail in Chapter 13.
increases rapidly and approaches 100% at 90 the scattered radiation suffers quantized of vibration in Figure 4-IOc produces a beam
deg. frequency changes. These changes are the that is plane polarized. The vibration of the
Scattering. As noted earlier, the transmis- result of vibrational energy level transitions electrical vector of a plane-polarized beam, Absorption of Radiation
sion of radiation in matter can be pictured as occurring in the molecule as a consequence of then, occupies a single plane in space. When radiation passes through a transparent
a momentary retention of the radiant energy, the polarization process. Raman spectroscopy Plane-polarized electromagnetic radiation layer of a solid, liquid, or gas, certain frequen-
which causes a brief polarization of the ions, is discussed in Chapter 9. is produced by certain radiant energy sources. cies may be selectively removed by the
atoms, or molecules present. Polarization is For example. the radio waves emanating from process of absorption. Here. electromagnetic
followed by reemission of radiation in all di- an antenna commonly have this character- energy is transferred to the atoms or
Polarization of Radiation
rections as the particles return to their original istic. Presumably, the radiation from a single molecules constituting the sample; as a result,
state. When the particles are small with re- Plane PoiarizatiolL Ordinary radiation can atom or molecule is also polarized; however, these particles are promoted from a lower-
spect to the wavelength of the radiation, de- be visualized as a bundle of electromagnetic since common light sources contain large energy state to higher-energy states, or
structive interference removes nearly all of the waves in which the amplitude of vibrations is numbers of these particles in all orientations, excited states. At room temperature, most
reemitted radiation except that which travels equally distributed among a series of planes the resultant is a beam that vibrates equally substances are in thei~ lowest energy
in the original direction of the beam; the path centered along the path of the beam. Viewed in all directions around the axis of travel. or ground state. Absorption then ordinarily
of the beam appears to be unaltered as a con- end-on, the electrical vectors would then The absorption of radiation by certain involves a transition from the ground state to
sequence of the interaction. Careful observa- appear as shown in Figure 4-lOa. The vector types of matter is dependent upon the plane higher-energy states. 1
tion, however, reveals that a very small in anyone plane, say X Y, can be resolved of polarization of the radiation. For example. Atoms, molecules, or ions have only a .
fraction of the radiation is transmitted at all into two mutually perpendicular components when properly oriented to a beam, aniso- limited number of discrete, quantized energy
angles from the original path and that the AB and CD as shown in Figure 4-10b. If the tropic crystals selectively absorb radiations levels; for absorption of radiation to occur,
intensity of this scattered radiation increases two components for each plane are combined, vibrating in one plane but not the other. the energy of the exciting photon must
with particle size. With particles of colloidal the resultant has the appearance shown in Thus, a layer of anisotropic crystals absorbs exactly match the energy difference between
dimensions, scattering becomes sufficiently Figure 4-1Oc. Note that Figure 4-1Oc has a all of the components of, say, CD in Figure the ground state and one of the excited states
4-IOc and transmits nearly completely the of the absorbing species. Since these energy
a molecule is given by
the number of possible energy levels for a
difTerences arc unique for each species, a
1. + Eribra'lanal
E - £.••••••••• + E"""IIOIIoI . (4-15) molecule Is much greater than for an atomic
study of the frequencies of absorbed radiation particle.
provides a means of characterizing the constit- where E•••et •••••l• describes the electronic
Figure 4-12 is a graphical representation of
uents of a sample of matter. For this energy of the molecule, and EribraiiOlUlI refers to the energies associated with Ii few of the elec-
purpose, a plot of absorbance as a function of the energy of the molecule resulting from var- tronic and vibrational states"'of a molecule.
wavelength or frequency is experimentally ious atomic vibrations. The third term in The heavy line labeled Eo represents the elec-
derived (absorbance, a measure of the Equation 4-15 accounts for the energy asso- tronic energy of the molecule in its ground
decrease in radiant power, is defined on ciated with the rotation of the molecule state (its state of lowest electronic energy);
p. 149). Typical plots of this kind, called around its center of gravity. For each elec- the lines labeled E1 and Ez represent the
absorption spectra, arc shown in Figure 4-11. tronic energy state of the molecule, there energies of two excited electronic states.
The general appearance of an absorption normally exist several possible vibrational Several vibrational energy levels (eo, eh .,.,
spectrum will depend upon the complexity, states, and for each of these, in turn, e.) are shown for each of these electronic
the physical state, and the environment of the numerous rotational states, As a consequence, states.
absorbing species. It is convenient to recog-
nize two types of spectra, namely, those asso-
ciated with atomic absorption and those
resulting from molecular absorption.
Atomic Absorptia The passage of poly-
chromatic ultraviolet or visible radiation
through a medium consisting of monatomic
t
I
particles, such as gaseous mercury or sodium,
results in the absorption of but a few well-
,
i
Excited
defined frequencies (see Figure 4-11d ~ The rel- £, electronic
ative simplicity of such spectra is due to the state 2

small number of possible energy states for the


particles. Excitation can occur only by an
electronic process in which one or more of
the electrons of the atom is raised to a higher-
o
220 240 260 280 300 ·3
.
~r"'"---~t -
Wovelength, nm •• L ~
CDerIYlevel. Thus, with sodium, excitation or
Cd)
the 35 electron to the 3p state requires energy
corresponding to a wavenumber of 1.697 x FIGURE 4-11 Some typical ultra- 1 .") J
eo' I E,
Excited
electronic
1()4em - I. As a result, sodium vapor exhibits violet absorption spectra. state t

a sharp absorption peak at 589.3 nm

f~:ae:'::.<--- - -
I
(yellow light). Several other narrow absorp-
tion lines, corresponding to other permitted Vi.lbl. energy
V ibrational :

~j _
-f-
electronic transitions, are also observed (see
Figure 4-11d). Regardless of the wavelength region in-
Ultraviolet and visible radiation have
sufficient energy to cause transitions of the
volved, atomic absorption spectra typically
consist of a limited number of narrow peaks. ..,_)1'
outermost or bonding electrons only. X-ray Spectra of this type are discussed in connec- •J -'~I----
frequencies, on the other hand, are several tion with X-ray absorption (Chapter 15) and
1 • 2 'I I
orders of magnitude more energetic and are atomic absorption spectroscopy (Chapter 11).
f I • " '
capable of interacting with electrons closest Molecular Absorption. Absorption by poly- • l il Eo
Gtound
electronic
to the nuclei of atoms. Absorption peaks cor- atomic molecules, particularly in the con- state
responding to electronic transitions of these densed state, is a considerably more complex fa) (bl (el
innermost electrons are thus observed in the process because the number of energy states
FIGURE 4-12 Partial energy level diagram for a fluorescent organic molecule.
X-ray region. is greatly enhanced. Here, the total energy of
As can be seen in Figure 4-12, the energy we have mentioned, many rotational energy resonance (NMR) and electron spin resonance complicated molecules possessing many
levels are associated with each vibrational (ESR) techniques, respectively; these methods closely related energy states. Continuous
difference between the ground state and an
state .. As a consequence, the spectrum for a are considered in Chapter 14. spectra also arise when the energy changes
electronically excited state is large relative to
molecule ordinarily consists of a series of Relaxation Processes. Ordinarily, the involve particles with unquantized kinetic
the energy differences be~een vibrational
closely spaced absorption bands, such as lifetime of an atom or molecule excited by energies.
levels in a given electronic s~te (typically, the
that shown for benzene vapor in Figure absorption of radiation is brief because Both continuous spectra and line spectra
two differ by a factor of 10 to 100).
4-11c. Unless a high-resolution instrument is several relaxation processes exist which are of importance in analytical chemistry. The
Figure 4-118 depicts with arrows some. of
employed, the individual bands may not be permit its return to the ground state. As former are frequently employed in methods
the transitions which result from absorption
detected, and the spectra will appear as shown in Figure 4-12b, nonradiative relaxa- based on the interaction of radiation with
of radiation. Visible radiation causes excita-
smooth curves. Finally, in the condensed tion involves the loss of energy in a series of matter, such as spectrophotometry. Line spec-
tion of an electron from Eo to any of the
state and in the presence of solvent small steps, the excitation energy being con- tra, on the other hand, are important because
vibrational levels associated with E I' The
molc:eules, the individual bands tend to verted to kinetic energy by collision with they permit the identification and determina-
frequencies absorbed would be given by
broaden to give spectra such as those shown other molecules. A minute increase in the tion of the emitting species.
in Figure 4-11a and b. Solvent effects are con- temperature of the system results.
v. = ~(E1 + e~ - Eo) (4-16)
As shown in Figure 4-12c, relaxation can
sidered in later chapters.
Pure vibrational alpsorption can be ob-
Thermel Radiation
also occur by emission of ftuorescent radia-
Similarly, the frequencies of absorbed ultra- served in the infrared region. where the tion. Still other relaxation processes are dis- When solids are heated to incandescence, the
violet radiation would be given by energy of radiation is insufficient to ca~ cussed in Chapters 10 and' 14. continuous radiation that is emitted is more
electronic transitions. Here, spectra exhibit characteristic of the temperature of the emit-
v. = ~ (E2 + e: - Eo) (4-17) narrow, closely spaced absorption peaks re- ting surface than of the material of which that
sulting from transitioBs among the various surface is composed. Radiation of this kind
Finally, the less energetic near- and mid- vibrational quantum Jeivels(see the bottom of (called black-body radiation) is produced by
infrared radiation can only bring about tran- Figure 4-118). Variatidns in rotational levels Electromagnetic radiation is often produced the innumerable atomic and molecular oscil-
sition among the vibrational levels of the may give rise to a seiies of peaks for each when excited particles (ions, atoms, or lations excited in the condensed solid by the
ground state. vibrational state. However, rotation is often molecules) return to lower-energy levels or to thermal energy. Theoretical treatment of
hindered or prevented in liquid or solid their ground states. Excitation can be brought black-body radiation leads to the following
samples; thus, the effects of these small energy about by a variety of means, including bom- conclusions: (I) the radiation exhibits a maxi-
differences are not ordinarily detected in such bardment with electrons or other elementary mum emission at a wavelength that varies in-
Although they are not shown, several rota- samples. particles, exposure to a high-potential alter- versely with the absolute temperature; (2) the
tional energy levels are associated with each Pure rotational spectra for gases can be nating current spark, heat treatment in an arc total energy emitted by a blade body (per unit
vibrational level. The energy difference be- observed in the microwave region. or a ftame, or absorption of electromagnetic of time and area) varies as the fourth power
Absorption Induced by a Magnetic Field. radiation. of temperature; and (3) the emissive power at
tween these is small relative to the energy dif-
When electrons or the nuclei of certain ele- Radiating particles that are well separated a given temperature varies inversely as the
ference between vibrational levels; transition
to excited rotational states is brought about ments are subjected to a strong magnetic from one another, as in the gaseous state, fifth power of wavelength. These relationships
by radiation in the 500- to 1QO-cm- 1 range. field additional quantized energy levels are behave as independent bodies and often pro- are reflected in the behavior of several exper-
In contrast to atomic absorption spectra, produced as a consequence of .magnetic duce radiation containing relatively few imental radiation sources shown in Figure
which consist of a series of sharp, well-defined properties of these elementary partiCles. The specific wavelengths. The resulting spectrum 4-13; the emission from these sources
lines, molecular spectra in the ultraviolet and· difference in energy' between the induced is then discontinuous and is termed a line spec- approaches that of the ideal black body. Note
visible regions are ordinarily characterized by states is small, and transitions between the trum. A continuous spectrum, on the other that the energy peaks in Figure 4-13 shift to
absorption bands that often encompass a sub- states are brought abput only by absorption hand, is one in which all wavelengths are rep- shorter wavelengths with increasing tempera-
stantial wavelength range (see Figure 4-11a, of long-wavelength (<lr low-frequency) radia- resented over an appreciable range, or one ture. It is clear that very high temperatures
b, c). Molecular absorption also involvc:selec- tion. With nucle~ radio waves ranging from in which the individual wavelengths are so are needed to cause a thermally excited
tronic transitions. As shown by Equations 4- 10 to 200 MHz are generally employed, while closely spaced that resolution is not feasible source to emit a substantial fraction of its
16 and 4-17, however, several closely spaced for electrons, microwaves with frequencies of by ordinary means. Continuous spectra result energy as ultraviolet radiation.
1000 to 25,000 MHz are absorbed. from excitation of: (1) solids or liquids, in Heated solids are used to produce infrared,
absorption lines will be associated with each
Absorption by nuclei or by electrons in which the atoms are so closely packed as to visible, and longer-wavelength ultraviolet ra-
electronic transition, owing to the existence of
numerous vibrational states. Furthermore, as magnetic fields is studied by nuclear magnetic be incapable of independent behavior; or (2) diation for analytical instruments.
Resonance fluorescence describes a process as a result of collision with other molecules.
the excited hydrogen atom, while £Ht and £H2
in which the emitted radiation is identical in Morc often than not, the energy of the ab-
represent the kinetic energies of the 'atoms.
frequency to the radiation employed for exci- sorbed radiation is greater than that of the
The sum of the latter can vary continuously
tation. The lines labeled 1 and 2 in Figure emitted radiation. Thus, for the absorption
from zero to EH•• Thus, the frequency of ~-
4-12a and 4-12c illustrate this type of fluores- labeled 3 in Figure 4-128, the absorbed
diation hv can also vary continuously" over thIS
cence. Here, the species is excited to the energy would be equal to (E2 - Eo +
range. The conseqUet1QCis a continuous s~-
trum from about 400 to 200 om that 15 a
energy states EI or E2 by radiation having an e: - eo); the energy of the fluorescent radia-
energy of (Et - Eo) or (E2 - Eo). After a tion, on the other hand, would again be given
useful source for absorption spectropho-
brief period, emission of radiation of identical by -(E2 - Eo); the emitted radiation would
tometry. energy" occurs, as depicted in Figure 4-12c. necessarily have a lower frequency or greater
Resonance fluorescence is most commonly wavelength than the radiation that excited
Emluion of X-ray Radiation
500 1000
produced by atoms in the gaseous state which the 'fluorescence. (This shift in wavelength to
Wavelength, om Radiation in the X-ray region is normally do not have vibrational energy states super- lower frequencies is sometimes called the
generated by the bomba~ment of a metal imposed on electronic energy levels. Stokes shift.) Clearly, both resonance and
target with a stream of htgh-speed electrons. Nonresonance fluorescence is brought nonresonance radiation can accompany fluo-
The electron beam causes the innermost elec- about by irradiation of molecules in solution resqence of molecules; the latter tends to pre-
trons in the atoms of the target material to be or in the gaseous state. As shown in Figure dominate,however, because of the much larger
raised to higher energy levels or to be ejected 4-128, absorption of radiation promotes the number of vibrationally excited states.
Emission of G••••
entirely, The excited atoms o~ ions th~ molecules into any of the several vibrational Phosphorescence occurs when an excited
Atoms, ions. or molecules in the gaseous state return to the ground state by vanous stepw,* levels associated with the two excited elec- molecule relaxes to a metastable excited elec-
can often be excited, by electrical discharge or electronic transitions that are accompanted tronic levels. The lifetimes of the excited troaic state which has an average lifetime of
by heat, to produce radiation in the ult~a- by the emission of photons, each having. an vibrational states are momentary, however greiter than about 10-, s. The nature of this
violet and visible regions. The process 10- energy hv. The consequeooe is the production (- 10-15 s~ and, as shown in Figure 4-12b, typtl of excited state is discussed in Chapter
volves promotion of the outerm~t electrons of an X-ray spectrum consisting of a ~ries ~f relaxation to the lowest vibrational level (eo 10. '
of a species to an excited electron~ state; ~a- lines characteristic of the target matenal. ThIS and e~) of a given electronic state takes place
diation is then emitted as the excited Species discrete spectrum is superimposed on a con-
returns to the ground state. Atomic emission tinuum of nonquantized radiation given off
spectra consist of a series of discrete Iin~ when some of the high-speed electrons are
1. Calculate the frequency in hertz and the wavenumber for
whose energies correspond to the energy dif- partially decelerated as they pass through the (a) an X-ray beam with a wavelength 4.2 A.
ferences between the various electronic states. target material. (b) an emission line for Cu at 211.0 nm.
For molecules, emission spectra are usua~ly
(c) the line at 694.3 nm produced by a ruby laser.
more complicated because there can eXISt Fluorescence and Phosphorescence
(d) the output of a CO2 gas laser at 10.6 Jim.
several vibrational and rotational states for
Fluorescence and phosphorescence are analyt- (e) an infrared absorption peak at 34.1 jJTll.
each possible electronic energy ~vel; in~t~
ically important emission processes in wh.ich (f) a microwave beam of wavelength 0.250 em.
of a single line for each electrOniC translu~n,
atoms or molecules are excited by absorption 2. Calculate the wavelength in centimeters for
numerous closely spaced lines form an emis-
of a beam of electromagnetic radiation; radiant (a) the radio frequency of an airport tower at 118.6 MHz.
sion band. .' emission then occurs as the excited species
A true continuous spectrum IS sometimes (b) a radio navigation beam !laving a frequency of 315.3 kHz.
produced from the excitation of gaseous return to the ground state. (c) an NMR signal at 42.6 MHz.
Fluorescence occurs much more rapidly (d) an EPR signal at 12,000 MHz.
molecules. For example, when hydrogen at
than phosphorescence and is generally com- 3. Calculate the energy of each bf the radiations in Problem 1 in
low pressure is subjected t~ an electric
plete after about 10-' s (or I~) from the (a) ergs/photon, (b) kcal/mol. and (c) eV. .
discharge, an excited molecule IS formed that
time of excitation; it is most easily observed
dissociates to give two hydrogen atoms and 4. Calculate the energy of each of the radiations in Problem 2 in
an ultraviolet photon. The energetics of this at a 9<k1eg angle to the excitation beam. P~-
(a) ergs/photon, (b) kcal/mol, and (c) eV.
phorescence emission takes place ove~ pen·
process is described by the equation 5. Calculate the velocity, wavelength, and frequency of the sodium D line
ods longer than 10-' s, and may lO~ecd
EH• = £HI + £H2+ hv (l = 5890 A in a vacuum) in
continue for minutes or even hours after Irra-
(a) acetone vapor (nD = l.()() 11).
Here, EH• represents the quantized energy of diation has ceased.
(b) carbon bisulfide (110 •• 1.622~
(c) a flint glass (110 •• 1.6S~
(d) fused silica (110 •• 1.44).
6. Calculate the reflective loss when a S89-nm beam of radiation
(a) passes through a piece of dense flint glass ("0" 1.63~
(b) passes through a cell having glass windows ("0" 1.46) and con-
taining an alcohol solution with a refractive index of 1.37.
COMPONENTS OF INSTRUMENTS FOR OPTICAL SPECTROSCOPY

The term spectroscopy was originally used to whether they are applied to the ultraviolet,
describe a branch of tcience based upon the visible, or infrared portion of the Spectrum.l
resolution of visible radiation into its com- Spectroscopic instruments contain five
ponent wavelengths. With the passage of components, including: (1) a stable source of
time, the meaning of the term has been ex- radiant energy; (2) a device that permits em-
panded to iitclude studies encompassing the ployment of a restricted wavelength region;
entire electromagnetic spectrum (Figure 4-7~ (3) a transparent container for holding the
The first spectroscopic instruments were sample; (4) a radiation detector or transducer
developed for use in the visible region and that cOnverts radiant energy to a usable
were thus called optical instruments. This term signal (usually electrical); and (5) a signal
has now been extended to include instru- processor and readout. Figure 5-1 shows the
ments designed for use in the ultraviolet and arrangement of these components for the four
infrared regions as well; while not strictly common types of spectroscopic measure-
correct, the terminology is nevertheless useful ments :mentioned earlier. As can be seen in
in that it emphasizes the many features that the figure, the configuration of components
are common to the instruments used for stud- (4) and (5) is the same for each type of
ies in these three important spectral regions. instruinent.
The purpose of this chapter is to describe Emission spectroscopy ditTers from the
the nature and properties of components of other three types in the respect that no exter-
instruments employed. for optical spectros- nal radiation source is required; the sample
copy. The seven chapters that follow will itself ~ the emitter. Here, the sample con-
provide details showing how the various c0m- tainer; is an arc, a spark, or a flame, which
ponents are assembled to produce optical in- both holds the sample and causes it to emit
struments, and how these instruments in turn characteristic radiation.
provide useful information. The instruments Absorption as well as fluorescence and
and techniques for spectroscopic studies in re- scattering spectroscopy require an external
sions more energetic than the ultraviolet and source of radiant energy. In the former. the
less energetic than the infrared have charac- beam from the source passes through the
teristics that dill'er substantially from optical sample after leaving the wavelength selector. Ie)
iDstrumcnts and are considered separately in In the latter two, the source iDduces
Chapters 14 and IS. the sample, held in a container, to emit char- FI~URE 6-1 Components for various types of instruments for
acteristic fluorescent or scattered radiation, ~Ptlcal spectroscopy: (a) emission spectroscopy, (b) absorption
which is measured at an angle (usually 90 pectroscopy, and (c) fluorescence and scattering Spectroscopy.
deg) with respect to the source.
Figure 5-2 summarizes the characteristics
COMPONENTS AND
of the first four components shown in Figure
CONFIGURATIONS OF
5-1. It is clear that instrument components
INSTRUMENTS FOR differ in detail, depending upon the wave-
OPTICAL SPECTROSCOPY length region within which they are to be
used. Their des~gn also depends on the pri-
~n order to be suitable for spectroscopic stud-
Spectroscopic methods are based upon the ~ use of the Instrument; that is, whether il
I For ~ more complcae discuuion of the components of I~, a s.ource must generate a beam of radia-
~ to be e~ployed for qualitative or quantita-
phenomena of emission, absorption, fluores- optical instruments, see: R. P. Bauman, Absorption Spec. tion With sufficient power for ready detection
tive ~nalyslS and whether it is to be applied to
cence, or ~cattering. While the instruments lrOICOpy. New York: Wiley. 1962, Chapten 2 and 3; E. J. and measurement. In addition, its output
Meehan, in TfflItiM 011 AlI4lytlcGl Clwmlltry, eds. I. M. atolDJC or molecular spectroscopy. Never-
for each dill'er somewhat in configuration, should be stabl~. Typically, the radiant power
Koltholl' and P. J. Elvin•. New York: Intcncicncc, 1964, thele;u, the general function and performance
their basic components are remarkably simi- Part I, yo\. 5, Chapter 55; and Pltyslc4l Mtfltob ofCIwm- o~ a source vanes exponentially with the elec-
r~~lrements of each type of component are
lar. Furthermore, the general properties of IItry. eds. A. Wciubcracr and B. W. Rouiter. New York: tncal power supplied to it. Thus, a regulated
slml~ar, .regardless of wavelength region and
their components are the same regardless of Wiley·lntcnciencc, 1912, Part IIIB, Yo\. I, Chapters 1-5. apphcatlon. power supply. is often needed to provide the
required stability. Alternatively, in some in-
-
struments. the output of the source is split An important feature of hydrogen discharge
into. a reference beam and a sample beam. lamps is the shape of the aperture between
.nm The first passes directly to a transduc:er while the two electrodes, which constricts the dis-
Spectral region VACUV UV ,I VISIBLE NEAR'IR IR • , FARIR the second interacts with the sample and is charge to a narrow path. As a consequence.
then focused on a matched transducer. (Some an intense ball of radiation about 1 to 1.5 mm
(.1 Sources
~limp I instruments employ a single transducer which in diameter is produced. Replacement of
Xenon limp
is irradiated alternately by the sample and hydrogen by deuterium results in a somewhat
'I H20rD2 1 reference beams.) The ratio of the outputs of
lamp Tungsten lamp
larger light ball.
Continuous r the two transducers serves as the analYtical Both high- and low-voltage lamps produce
Nern.t glower IZrD2 +Y20,)
parameter. The effect of fluctuations in the a continuous spectrum in the region of 160 to
Nichrome wire (Ni + Cr) ,
source output is largely canceled by. this 375 nm. Quartz windows must be employed
GlobIr (SiC!
means. in the tubes, since glass absorbs strongly in
Hollow cathode
Di1continuou. {
, Both continuous and line sources are used this wavelength region.
lamps
in optical spectroscopy. The former ;finds Tungsten Filament Lamps. The most
wide application in molecular absorption
lb) W_ength
selectors :~ prism '-.,
Fused .
.11..,. or qUlrtz prosm methods. The latter is employed in ftqores-
common source of visible and near-infrared
radiation is the tungsten filament lamp. The
GllSlprism . cence and atomic absorption spectroscopy. energy distribution of this source approxi-
Continuous

3000 Iineslmm
I I
Grltings with various number
i~~
of lineslmm
KBr pritm
50lineslmm
Figure 5-la lists the most widely used
spectroscopic sources.
mates that of a black body and is thus
temperature dependent. Figure 4-13 illus-
trates the behavior of the tungsten filament
I Interference wedges
lamp at 3OOOoK.In most absorption instru-
Interforence filters
Continuous Sources of Ultraviohtt. ments, the operating filament temperature is
OilContinuous {
GllSlabsorption
Visible, and Near-Infrared Radiat)on about 2870oK; the bulk of the energy is thus
filters .emitted in the infrared region. A tungsten
LiF
(e) M.terill. for filament lamp is useful for the wavelength
cell •• windows, qUIrtz
& lenses
Fused silica or Continuous sources provide radiation whose region between 320 and 2500 nm.
,
Corexglass! power does not change sharply among adja- In the visible region. the energy output of a
Silicate gtlSl cent wavelengths. tungsten lamp varies approximately as the
NaCI Hydrogen or Deuterium Lamps. As noted fourth power of the operating voltage. As a
K8r in Chapter 4 (p. 110~ a continuous spectrum consequence. close voltage control is required
TIBr-TII in .the ultraviolet region is conveniently for a stable radiation source. Constant volt-
produced by the electrical excitation of age transformers or electronic voltage regu-
ldl TWItducers Photornu tiplier hydrogen or deuterium at low pressure. lators are often employed for this purpose. As
Two types of hydrogen lamps are en-

{
Photon Phototube an alternative, the lamp can be operated from
detectors
, 7:otocefl countered. The high-voltage variety employs a 6-V storage battery, which provides a
Silicon diode
potentials of 2000 to 6000 V to cause a remarkably stable voltage source if it is main-
Semiconductor
discharge between aluminum electrodes; tained in good condition.
Thermocouple (volts) or Bolometer (ohms'
water cooling of the lamp is required if-high Xenon Arc Lamps. This lamp produces
Heat
detectors { I
I
I Gol.y
Pyrcetectric
pneumltic cell
cell (capocitonce)
I radiation intensities are to be produee<i. In
low-voltage lamps. an arc is formed between
intense radiation by the passage of current
through an atmosphere of xenon. The spec-
a heated, oxide-coated filament and a rhetal trum is continuous over the range between
FIGURE 5-2 Components and materials for spectroscopic instruments ..(Adapt~ ~rom
a figure by Professor A. R. Armstrong, College of William and Mary. With permission.) electrode. The heated filament provides 'elec- about 25;0 and 600 nm, with the peak inten-
trons to maintain a dc current when a voltage sity occurring at about 500 nm (see Figure
of about 40 V is applied; a regulated power 4-13, p. 110). In some instruments, the lamp is
supply (p. 53) is required for constant operated intermittently by regular discharges
intensities. from a capacitor; high intensities are obtained.
Continuous Sources age of current. A rhodium wire heater sealed
~ Nonparallel radiation •
of Infr••.••• Rlldlation in a ceramic cylinder has similar properties as
..•.._ .•..,._A'
a source.
t----i: '..>.-;""'"

'X-i- ~
The common infrared source is an inert solid
heated electrically to temperatures between
lSOO and 2OQOoK. Continuous radiation
approximating that of a black body results
(see Figure 4-13). The maximum radiant
intensity at these temperatures occurs be-
tween 1.7 and 2 #Ll11 (6000 to SOOOem -I). At
Line Sources
Sources that emit a few discrete lines find use
in atomic absorption spectroscopy, Raman
spectroscopy, refractometry, and polarimetry.
Metal Vapor Lamps. Two of the most
Mirror

Pumping
SOurce
JZ:"
Radiation ----- ~

g£~
. ".
t -.
''.i'~lv
=s --.
Active I~ing mediu'!'

. .
transmitting
mirror

common line sources are the familiar mercury FIGURE 5-3 Schematic representation of a typical
longer wavelengths, the intensity falls off co~- laser source.
and sodium vapor lamps. A vapor lamp con-
tinuously until it is about I % of the maxI-
sists of a transparent envelope containing a
mum at 151lm (667 em-I). On the short
gaseous element at low pressure. Excitation of
wavelength side, the decrease is much more
the characteristic line spectrum of the element processes with lifetimes in the range of 10-9 narrow band of radiation at any chosen
rapid , and a similar reduction in intensity is
I occurs when a potential is applied across a to 10-12 s, the detection and determination wavelength within the range of the source.
observed at about 111m (10,000 em- ).
pair of electrodes fixed in the envelope. Con- of extremely small concentrations of species Figure 5-3 is a schematic representation
TIle Nerast Glower. The Nemst glower is
duction occurs as a consequence of electrons in the atmosphere, and the induction of is0- showing the components of a typical laser
composed of rare earth oxides formed into a
and ions formed by ionization of the metaL topically selective reactions.3 In addition, laser source. The heart of the device is a lasing
cylinder having a diameter of I to 2 mm and
Ordinarily, initial heating is required to pro- sources have become important in several
a length of perhaps 20 Mm. Platinum leads medium. It may be a solid crystal, such as
duce sufficient metal vapor; once started, routine analytical methods, including Raman
are sealed to the ends of the cylinder to ruby, a semiconductor such as gallium arse-
however, the current is self-sustaining. spectroscopy (Chapter 91 emission spectros-
permit passage of current. This device has a nide. a solution of an organic dye, or a gas.
The mercury lamp produces a series of copy (Chapter 12), and Fourier transform The lasing material can be activated or
large negative temperature coefficient of elec-
lines ranging in wavelength from 254 to infrared spectroscopy (Chapter 8).
trical resistance, and it must be heated ex- pumped by radiation from an external source
734 nm. With the sodium lamp. the pair of The term laser is an acronym for light
ternally to a dull red heat before a sufficient so that a few photons of proper energy will
lines at 589.0 and 589.6 nm predominate. amplification by stimulated emission of radia- trigger the formation of a cascade of photons
current passes to maintain the desired
Hollow CathcMle Lamps. Hollow cathode tion. As a consequence of their light- of the same energy. Pumping can also be
temperature. Because the resistance decreases
lamps provide line spectra for a large number amplifying properties, lasers produce spatially carried out by an electrical current or by an
with increasing temperature, the source cir-
of elements. Their use has been confined to narrow, extremely intense beams of radiation. electrical discharge. Thus, gas lasers usually
cuit must be designed to limit the current;
atomic absorption and atomic fluorescence The process of stimulated emission produces do not have the external radiation source
othe(Wise the glower rapidly becomes so hot
spectroscopy. Discussion of this type of a beam of highly monochromatic (band- shown in Figure 5-3; instead, the power
that it is destroyed.
source is deferred to Chapter II. widths of 0.01 nm or less) and remarkably supply is connected to a pair of electrodes
The Globar Source. A Globar is a silicon
coherent (p. 99) radiation. Because of these contained in a cell filled with the gas.
carbide rod, usually about 50 mm in length
Lase,. unique properties, lasers have become impor- A laser normally functions as an oscillator,
and 5 mm in diameter. It also is electrically
heated and has the advantage of a positive tant sources for use in the ultraviolet, visible, in the sense that the radiation produced from
The first laser was constructed in 1960.2 Since
coefficient of resistance. On the other hand, and infrared regions of the spectrum. A limi- the laser mechanism is caused to pass back
that time, chemists have found numerous
water cooling of the electrical contacts is tation of early lasers was that the radiation and forth through the medium numerous
useful applications for these sources in high-
required to prevent arcing. Spectral energies from a given source was restricted to a rela- times by means of a pair of mirrors as shown
resolution spectroscopy, kinetic studies of
of the Globar and the Nernst glower are com- tively few discrete wavelengths or lines. Re- in Figure 5-3. Additional photons are gen-
cently, however, dye lasers have become erated with each passage, thus leading to
parable except in the region below 5 #Ll11
(2000 em -I), where the Globar provides a available; tuning of these sources provides a enormous amplification. The repeated pass-
significantly greater output. 1 For a more complete discussion or lasers. see: B. A. age also produces a beam that is highly par-
Inl:andeseeat Wire Source. A source of Lengye~ lAsers. New York: Wiley, 1971; S. R. Leone allel because nonparallel radiation escapes
and C. B. Moore, Chemical allll Biological A.pplications of from the sides of the medium after being
somewhat lower intensity but longer life than lAsers, eel. C. E. Moore. New York: Academic Press,
the Globar or Nernst glower is a tightly reflected a few times (see Figure 5-3~
1974; and A.nalytical lAser Spectroscopy. eel. N. Omen-
wound spiral of nichrome wire heated by pass- etlo. New York: Wiley, 1979. • For a review or some or Ihese applicalions. see: S. R. In order to obtain a usable laser beam, one
Leone, J. Chem. Educ.,53, 13 (1976~ of the mirrors is coated with a sufficiently
thin layer of reflecting material so that a frac- diagrams shown in Figure 5-4b. Note that the
tion of the beam is trilnsmitted rather than wavelength of the fluorescent radiation is
(1)
reflected (see Figure 5-3). directly related to the energy difference be-
-_ EM
. Laser action can be understood by con-
sidering the four processes depicted in Figure
tween the two electronic states, E, - Ex. It is
also important to note that the instant at
=+=~.
.,- -r· E•
:±: ~ Heat

5-4, namely, (a) pumping, (b) spontaneous which the photon is emitted, and its direction, Pumping I I
energy I I
emission (fluorescence1 (c) stimulated emis-
sion, and (d) absorption. For purposes of
will vary from excited electron to excited elec-
tron. That is, spontaneous emission is a
-+- +E.
illustration, we will focus on just .two of random process; thus, as shown in Figure 5-4,
several electronic energy levels that wil( exist the fluorescent radiation produced by one of
in the atoms, ions, or molecules making up the particles in diagram b-I differs in direc-
the laser material; as shown in the figure, the tion and phase from that produced by the
two electronic levels have energies E, and Ex. second particle (diagram b-2). Spon.taneous
Note that the higher electronic state is shown emission, therefore, yields incoherent radiation.
as having several slightly different vibrational Stimulated Emission. Stimulated emission,
energy levels, E" E~, E;, and so forth. We which is the basis of laser behavior, is
have not shown additional levels for the depicted in Figure 5-4c. Here, the' excited
lower electronic state, although such often laser particles are struck by externally
exist. produced photons having precisely the same
Pumping. Pumping, which is necessary for energies (E,- Ex) as the photons produced
laser action, is a process by which the active by'spontaneous emission. Collisions; of this
species of a laser is excited by means of an type cause the excited species to r~lax im-
electrical discharge, passage of an electrical mediately to the lower energy statei and to
current, exposure to an intense radiant emit simultaneously a photon of exactly the ---
--- ----
source, or interaction with a chemical species.
During pumping, several of the higher elec-
tronic and vibrational energy levels of the
active species will be populated. In diagram
same energy as the photon that stimulated
the process. More important, and more
remarkable, the emitted photon is precisely in
phase with the photon that triggered the event.
~~ __=cvu-
a-I of Figure 5-4, one atom or molecule is That is, the stimulated emission is totally
shown as being promoted to an energy state coherent with the incoming radiation.
E;; the second is excited to the slightly higher . Absorption. The absorption process, which
vibrational level E;. The lifetime of excited competes with stimulated emission. is
vibrational states is brief, however; after depicted in Figure 5-4d. Here, two photons
10-13 to 10-14 S, relaxation to the lowest with energies exactly equal to E, - Ex are ab-
excited vibrational level (E, in diagram a-3) sorbed to produce the metastable excited
occurs with the production of an undetectable state shown in diagram d-3; note that the
quantity of heat. Some excited electronic state shown in diagram d-3 is identical to that
states of laser materials have lifetimes con- attained by pumping diagram a-3. FIGURE 5-4 Four processes important in laser action: (a) pumping
siderably longer (often I ms or more) than Population Inversion a'" Light Amplifica- (excitation by electrical. radiant. or chemical energy). (b) spontaneous
their excited vibrational counterparts; long- tion. In order to have light amplification in a emission. (c) stimulated emission. and (d) absorption.
lived states are sometimes termed metastable laser, it is necessary that the number of~hotons
as a consequence. produced by stimulated emission exceed the
Spontaneous Emission. As was pointed QUt number lost by absorption. This condition will
in the discussion of fluorescence, a species in prevail only when the number of particles in the
an excited electronic state may lose all or part higher energy state exceeds the number in the
of its excess energy by spontaneous emission lower; that is, a population inversion from the
of radiation. This process is depicted in the normal distribution of energy states must exist.
Population inversions are brought about by system. In contrast, it is only necessary to E. E.

:rE~
pumping. Figure 5-5 contrasts the effect of pump sufficiently to make the number of par- " ____ •FISttransition
nonradiative
incoming radiation on a noninverted popula- ticles in the E, energy level exceed the number
tion with an inverted one. in Ex of a four-level system. The lifetime of a Ey

11Iree- •••• Four-Level LIser Systems. Fig- particle in the Ex state is brief, however, be-
ure 5-6 shows simplified energy diagrams for
the two common types of laser systems. In the
three-level system, the transition responsible
cause the transition to Eo is fast; thus, the
number in the Ex state will generally be negli-
gible with respect to Eo and also (with a
1E,/-- Fast transition

for laser radiation is between an excited state modest input of pumping energy) with respect
E, and the ground state Eo; in a four-level to E,. That is, the four-level laser usually
FIGURE 6-6 Energy level diagrams for two types
system, on the other hand, radiation is gen- achieves a population inversion with a smaller
of laser systems.
erated by a transition from E, to a state Ex expenditure of pumping energy.
that has a greater energy than the ground Some Examples of Useful Lasers. The first
state. Furthermore, it is necessary that transi- successful laser, and one that still finds wide-
tions between Ex and the ground state be spread use, was a three-level device in which a laser, which produces intense lines in the Tuning of dye lasers can be readily accom-
rapid. ruby crystal was the active medium. The ruby green (514.5 om) and blue (488.0 nm) regions. plished by replacing the nontransmitting
The advantage of the four-level system is was machined into a rod about 4 em in This laser is a four-level device in which mirror shown in Figure 5-3 with a monochrom-
that the population inversions necessary for length and 0.5 em in diameter. A ftash tube argon ions are formed by an electrical or ator equipped with a reftection grating or a
laser action are more readily achieved. To was coiled around the cylinder to produce radio frequency discharge. The input energy Littrow-type prism (p. 126) which will reftect
understand this, note that at room tempera- intense ftashes of light. Because the pumping is sufficient to excite the ions from their only a narrow bandwidth of radiation into
ture a large majority of the laser particles will was discontinuous, a pulsed laser beam was ground state, with a principal Quantum the laser medium; the peak wavelength can
be in the ground-state energy level Eo'in both produced. Continuous wave ruby sources are number of 3, to various 4p states. Laser activ- be varied by rotatioQ of the grating or prism.
systems. Sufficient energy must thus be now available. ity then involves transitions to the 4s state. Emission is then stirllulated for only part of
provided to convert more than 50% of the A variety of gas lasers are sold commer- Dye lasers4 have become important radia- the fluorescent spec(rum, namely. the wave-
lasing species to the E, level of a three-level cially. An important example is the argon ion tion sources in chemistry because they are length reflected from the monochromator.
tunable over a range of 20 to 50 nm; that is,
dye lasers provide bands of radiation of a
chosen wavelength with widths of a few hun- WAVELENGTH
dredths of a nanometer or less. SELECTION; MONOCHROMATORS
The active materials in dye lasers are solu-
AbIofption "', Ey tions of organic compounds capable.()fftuores- With few minor exceptions, the analytical
cing in the ultraviolet. visible. or infrared methods considered in the following seven
rv\ri I
rvv- regions. In contrast to ruby or gas lasers. chapters require dispersal of polychromatic
however, the lower energy level for laser radiation into bands that encompass a re-
E, action (Ex in Figure 5-6) is not a single energy stricted wavelength region. The most common
but a band of energies arising from the super- way of producing such bands is with a device
lal
position of a large number of small vibra- called a monochromator.'
Ey tional and rotational energy states upon the

'Ulr
base electronic energy state. Electrons in E).
V\r Light amplification
by stimulated
emission
may then undergo transitions to any of these
states, thus producing photons of slightly dif-
j
• For a more complete d~ussion or monochromators,
see: R. A. Bauman, AbJO,ption Spect,OM:OPY.New York:
Stimulated / E, ferent energies.
emission
Wiley, 1962; F. A. Jenkins and H. E. White, FUMamen-
tats of Optia. 3d ed. New York: McGraw-HiI~ 1957;
E. J. Meehan, in T,etIliM 011 A""lytiClJI CMmUt'Y. cds.
(b)
I. M. KolthofTand P. J. Elving. New York: Wiley, 1964.
FIGURE 6-6 Passage of radiation through (a) a noninverted • For rurther inrormation, see: R. B. Green, - Dye Laser Part t, vol. 5. Chapter 55: and J. F. James and R. S.
Instrumentation," J. Chem. Educ.• 54 (9) A365, (10) A407, Sternberg, The Design of Optical Spectromners. London:
population and (b) an inverted population. Chapman and Hall. 1969,
(1977~
grating, angular dispersion results. from Construedoa ~terials. Clearly, the mate-
diffraction, which occurs at the reflective rials employed for fabricating the windows
Focal
surface. In both ~esigns, the dispersed radia- lenses, and prisms of a monochromator must

~~7B
••••... A

A2/Exit
plane
tion is focused on the focal plane AB where it
appears as two "images of the entrance slit
(one for each wavelength).
transmit radiation in the frequency range of
interest •. ideally, the transmittance should
approach tOO%, although it is sometimes
/ sfit necessary to use substances with transmit-
A Types of Instruments tances as low as 20%. The refractive index of
Employing Monochromators materials used in the construction of windows
The method of detecting the radiation dis- and prisms should be low in order to reduce
persed by a monochromator varies. In a reflective losses (see Equation 4-14, p. 103).
spectroscope, detection is accomplished vis- However, for lenses which function by bend-
ually with a movable eyepiece located along ing rays of radiation, a material with a high
the focal plane. The wavelength is then refractive index is desirable in order to reduce
determined by measurement of the angle be- focal lengths. Ideal construction materials for
tween the incident and dispersed beams. In a both lenses and windows should exhibit little
spectrograph, a photographic film or plate is change in refractive index with frequency,
mounted along the focal plane; the series of thus reducing chromatic aberrations. In
darkened images of the slit along the length contrast, just the opposite property is sought
of the developed film or plate is called a for prisms, where1he extent of dispersion
spectrogram. depends upon the rate of change of refractive
The monochromators shown in Figure 5-7 index with frequency. In addition to the fore-
(bl are also utilized for dispersing elements in going optical properties, it is desirable that
spectr~ters. Such instruments have a fixed monochromator components be resistant to
FIGURE 5-7 Two types of monochromators: (a) Bunsen mechanical abrasion and attack by atmos-
prism monochromator, and (b) Czerney-Tumer grating mono- exit slit located in the focal plane. With a
continuous source, the wavelength emitted pheric components and laboratory fumes.
chromator. (In both instances, 11 > 12,) Needless to say, no single substance meets
from this slit can be varied continuously by
rotation of the dispersing element. all of these criteria and trade-offs are thus
We need to define two other types of in- required in monochromator design. The
Mooochromators for ultraviolet, visible, which projects a series of rectangular images choices here depend strongly upon the wave-
struments to complete this discussion. A
and infrared radiation are all similar in of the entrance slit upon a plane surface (the length region to be used.
spectrophotometer is a spectrometer which
mechanical construction in the sense that focal plane). In addition, most monochroma- contains a photoelectric device for determin- Table 5-1 lists the properties of the
they employ slits, lenses, mirrors, windows, tors have entrance and exit windows, which ing the power of the radiation exiting from common substances employed for fabricating
and prisms or gratings. To be sure, the mate- are designed to protect the components from the slit. A photometer also employs a photo- monochromator components. Clearly, no
rials from which these components are fabri- dust and corrosive laboratory fumes. electric detector but contains no monochro- single material is suitable for the entire wave-
cated will depend upon the wavelength region Figure 5-7 shows the optical design of two mator; instead, filters are used to provide length region. For the ultraviolet, visible, and
of intended use (see Figure 5-2b and c). typical monochromators, one employing a near-infrared regions (up to about 3000 nm~
radiation bands encompassing limited
prism for dispersal of radiation and the other a quartz prism is often employed; it should be
wavelength regions. A photometer is inca-
Components of • Monochromator a grating. A source of radiation containing noted, however, that glass provides better res-
pable of providing a continuously variable
but two wavelengths, 11 and 12, is shown for olution for the same size prism for wave-
All monochromators contain an entrance slit, band of radiation.
purposes of illustration. This radiation enters lengths between 350 and 2000 nm. As shown
a collimating lens or mirror to produce a par-
the monochromators via a narrow rectangu- Prism Monochromators in column 5 of Table 5-1, several prisms are
allel beam of radiation, a prism or grating as
lar opening or slit, is collimated, and then required fo cover the entire infrared region.
a dispersing element, 6 and a focusing element
Prisms can be used to disperse ultraviolet, Types of Prisms and Prism Monochroma-
strikes the surface of the dispersing element at
an angle. In the prism monochromator, re- visible, and infrared radiation. The material tors. Figure 5-8 shows the two most common
fraction at the two faces results in angular used for their construction will differ depend- types of prism configurations. The first is a
• Much less gencraIIy employed are interference wedges.
ing upon the wavelength region. 6O-deg design, which is ordinarily fabricated
These devices are described in the next section (p. 136). dispersal of the radiation, as shown; for the
somewhat more compact monochromator rise in the refractive index for glass below 400
designs. In addition, when quartz is Dm corresponds to the sharp increase in
employed, polarization is canceled by the absorption that prevents the use of this mate-
reversal of the radiation path. A typical rial below 350 Dm. in the region of 350 to
Littrow-type monochromator' is shown in 2000 DID, however, glass is greatly superior to
Figure 7-10. quartz for prism fabrication because of its
Angular DispersioR of ~ The angular larger change in refractive index with wave-
dispersion of a prism is the rate of change of length (dll/d).).
FIGURE 5-8 Dispersion by a prism: (a) quartz Cornu type,
the angle 8 in Figure 5-8 as a function of Focal-Plane DiSpersion of Prism Mono-
and (b) Littrow type.
wavelength; that is, d8/dl. The spectral purity chromators. The focal-plane dispersion of a
of the radiation exiting from a prism mono- monochromator refers to the variation in
chromator is dependent upon this quantity. wavelength as a function of y, the linear dis-
The angular dispersion of a prism can be tance along the line AB of the focal plane of
from a single block of material. When crystal- called a Cornu prism. Figure 5-7a shows a
resolved into two parts the instrument (see Figure 5-7). That is, the
line (but not fused) quartz is employed, Bunsen monochromator, which employs a
however, the prism is usually formed by 6O-deg prism, likewise often made of quartz. d8 d8 dn focal-plane dispersion is given by dy/d)'.
cementing two 3().deg prisms together, as Figure 5-8b depicts a Littrow prism, which d)' = dn . d)' (5-1) Figure 5-10 shows the focal-plane disper-
shown in Figure 5-8a; one is constructed is a 3().deg prism with a mirrored back. It is sion of two prism monochromators and one
from right-handed quartz and the second seen that refraction occurs twice at the same where d8/dn represents the change in 8 as a grating instrument. Note that the dispersion
interface; the performance characteristics of function of the refractive index n of the prism for the two prism monochromators is highly
from left-handed quartz. In this way, the optic-
aUyactive quartz causes no net polarization of the Littrow prism are thus similar to those of material and dn/d)' expresses the variation of nonlinear, the longer wavelengths being
the emitted radiation; this type of prism is the 6O-deg prism. The Littrow prism permits the refractive index with wavelength (that is, bunched together over a relatively small dis-
the dispersion of the substance from which tance. The two prism monochromators were
the prism is fabricated). Littrow types. cach having a height of 57 mm.
The magnitude of dO/dn is determined by Notc the much largcr dispersion exhibited by
the geometry of the prism and the angle of the monochromator with a glass prism, in the
incidence i (Figure 5-Sa~ In order to avoid region of 350 to 800 nm.
TABLE 1·1 CONSTRUCTION MATERIAL FOR SPECTROPHOTOMETER
problems of astigmatism (double images), this
OPTICS

=w
angle should be such that the path of the
Uleful HarUelsDd beam through the prism is within a few
Tl'lIDSIIlittance Refractive 1tuIe fer CIIemicaI degrees of being parallel to the base of the
Material Range, pm Index Prisms, IJI11 Resistance
prism. Under these circumstances. dO/dll
Fused silica 0.18-3.3 1.46 0.52 x 10-4 0.18-2.7 Excellent depends only upon the prism angle ClC (Figure
Quartz 0.20-3.3 1.54 0.63 x 10-4 0.20-2.7 Excellent 5-Sa) and increases rapidly with this quantity.
Flint glass 0.35-2.2 1.66 1.70 x 10-4 0.35-2 Excellent Reflection losses, however. impose an upper K 1.60
Calcium 0.12-12 1.43 0.33 x 10-4 5-9.4 Good ~
limit of about 60 deg for IX. For a prism in
fluoride .S

~fluorite)
Lit ium
fluoride
0.12-6 1.39 0.29 x 10-4 2.7-5.5 Poor
which ClC = 60 deg, it can be shown that
2
dO/dn = (I - n /4,-t (5-2)
.~
~ -----------
~OCk salt .

~ 1.50
Sodium 0.3-17 1.54 0.94 x 10-4 8-16 Poor The term dll/d)' in Equation 5-1 is related
chloride to the dispersion of the substance from which
Potassium 0.3-29 1.56 1.45 x 10-4 15-28 Poor
bromide the prism is constructed. We have noted
Cesium 0.3-70 1.79 15-55 Poor (p. 102) that the greatest dispersion for a
iodide given material is near its anomalous disper- 1.40
KRS-5 1-40 2.63 24-40 Good sion region, which in turn is close to a region o 200 400 600 800 1000 1200
(TlBr-TlI) of absorption. The dispersion of some sub-
Wavelength. nm

stances employed for construction of prisms FIGURE 5-9 Dispersion of several optical
• Radians/ l'1li at 0.589 pm is shown in Figure 5-9. Note that the rapid materials.
Figure 5-10 illustrates one ofthe important reflection from a reflection grating. A trans-
advantages of grating mopochromators- mission grating consists of a series of parallel
linear dispersion along the focal plane. and closely spaced grooves ruled on a piece of
Monochromatic
Resolving Power of Prism Monocbroma- glass or other transparent material. A grating beam It incident
tors. The resolving .power R of a prism gives suitable for use in the ultraviolet and visible ongtel .
the limit of its ability to separate adjacent region has between 2000 and 6000 lines per
images having slightly different wavelengths. millimeter. An infrared grating requires con-
Mathematically, this quantity is defined as siderably fewer lines; thus, for the far infrared
region, gratingS with 20 to 30 lines per milli-
meter may suffice. It is vital that these lines be
where dl represents the wavelength difference equally spaced. throughout the several centi-
that can just be resolved and l is the average meters in length of the typical grating. Such
wavelength of the two images. It can be gratings require elaborate apparatus for their
shown that the resolving power of a prism is production and are consequently expensive.
directly proportional to the length of the Replica gratings are less costly. They are
prism base b (Figure 5-8) and the dispersion manufactured I;>yemploying a master grating
of its construction material. That is, as a mold for' the production of numerous
plastic replicas; the products of this process, . FIGURE 6-11 Schematic diagram illustrating the mechan-
while inferior in performance to an original ism of diffraction from an echellette-type grating. (From R. P.
grating, are adequate for many applications. Bauman, Absorption Spectroscopy. New York, Wiley, 1962,
When a transmission grating is illuminated p. 65. With permission.)
from a slit, ~ch groove scatters radiation
Gl'IIting Monochromatora
and thus effedtively becomes opaque. The
for Wavelength Selection
nonruled portions then behave as a series of
Dispersion of ultraviolet, visible, and infrared closely spaced slits, each of which acts as a the radiation as shown in Figure 4-00 reflected beams 1,2, and 3 can occur. In order
radiation can be brought about by passage of new radiation source; interference among the (p. 98). The angle of diffraction, of course, for constructive interference to occur between
a beam through a transmission grating or by multitude of beams results in diffraction of depends upon the wavelength. adjacent beams, it is necessary that their path
Reflection gratings are used more exten- lengths differ by an integral multiple n of the
sively in instrument construction than their wavelength l of the incident beam.
Gl'lting
transmitting counterparts. Such gratings are In Figure 5-11, parallel beams of mono-
200 . 300 500 made by ruling a polished metal surface or by chromatic radiation 1 and 2 are shown strik-
~. om I I I evaporating a thin film of aluminum onto the ing the grating at an incident angle i to the
surface of a replica grating. As shown in grating normal. Maximum constructive inter-
figure 5-11, the incident radiation is reflected ference is shown as occurring at the reflected
Glass prism
200 350 400 450 500 600 800 from one of the faces of the groove, which angle r. It is evident that beam 2 travels a
X,om A-,j\il9£~;;; -.-, I I I I I II then acts as a new radiation source. Interfer- greater distance than beam 1 and that this
ence results in radiation of differing wave- difference is equal to (CD - AB). For con-
Quartz prism
lengths being reflected at different angles r. structive interference to occur, this difference

A,nm
200
I
250
I
300
I
350 400
I
'! I ,
500600800
II
Diffraction by a Grating. figure 5-11 is a
schematic representation of an echellette-type
must equal n).. That is,

M = (CD - AB)
A 8 grating, which is grooved or blazed such that
I it has relatively broad faces from which Note, however, that angle CAD is equal to
0 5.0 10.0 15.0 20.0 25.0 reflection occurs and narrow unused faces.
Diltance y along focal plane, em angle i and that angle BDA is identical with
This geometry provides highly efficient dif- angle r. Therefore, from simple trigonometry,
FIGURE 6-10 Dispersion for three types ofmonochromators. fraction of radiation. Each of the broad faces
The points A and B on the scale correspond to the points we may write
can be considered to be a point source of
shown in Figure 5-7. radiation; thus interference among the
where d is the spacing between the reflecting and the wavelengths for the first-, secondo, RfIOIviDg Power of a Gntiac- It can be For the grating, we employ Equation 5-7
surfaces. It is also seen that and thUd-order reflections are 543, 271, and shown 7 that the resolving power R of a grat- .t
181 nm, respectively. ing is given by the very simple expression -=aN
.u
AB= -dsin r
Similarly, when r = -10 deg (here r lies to .t For the first-order spectrum (D = I)
The minus sign, by convention, indicates that the right of the grating n0"ttal) R = .u = aN (5-7)
reflection has occurred. The angle r, then, is N 460.25 460 3'
negative when it lies on the opposite side of where n is the diffraction order and N is the = 0.10 x I =. x 10 hnes
.t = SOO[sin 48 + sine-10)] = 284.7
the grating normal from the angle i (as in n n number of lines illuminated by the radiation · I h 4.6 X 103 lines 0.1 em
Figure 5-1I~ and is positive when it is on the from the entrance slit. Thus, as with a prism gra t 109 engt = 2000 Ii x --
same side. Substitution of the last two expres- nes/mm mm
(Equation 5-4~ the resolving power depends
sions into the first gives the condition for con- upon the physical size of the dispersing •••0.23 em
structive interference. Thus, Wavelength (nm) for element.
Thus, in theory, the separation can be
III = d(sin i + sin r) (5-5) accomplished with a silica prism with a base
r n=1 n=2 n=3
EXAMPLE of 3.5 em, a glass prism with a base of 1.3 em,
Equation 5-5 suggests that several values 20 543 271 181 ~mpare the size of (I) a 6O-deg fused or 0.23 em of grating.
of.t exist for a given diffraction angle r. Thus, 10 458 229 153 silica' prism; (2) a 6O-deg glass prism; and
if a first-order line (D = I) of 800 nm is found 0 372 186 124
at r, second-order (400 nm) and third-order -10 285 142 95 (3) a grating with 2000 lines/mm, that would Gntinp versus Priims. Gratings offer
(267 nm) lines also appear at this angle. Or- be required to resolve two lithium emission several advantages over prisms as dispersing
lines at 460.20 and 460.30 nm. Average values elements. One of these is that the dispersion is
dinarily, the first-order line is the most
intense; indeed, it is possible to design grat- A concave grating can be produced by for tlje dispersion (dn/d.t) of fused silica and nearly constant with wavelength; this
ings that concentrate as much as 90% of the ruling a spherical reflecting surface. Such a glass ~n the region of interest are 1.3 x 10- 4 property makes the design of monochroma-
incident intensity in this order. The higher- diffracting element serves also to focus the and 3.6 x 10-4 nm-I, respectively. tors considerably more simple. A second
order lines can generally be removed by radiation on the exit slit and eliminates the For the two prisms, we employ Equation 5-4 advantage can be seen from the foregoing
filters. For example, glass, which absorbs ra- need for a lens. .t dn example, namely, that better dispersion can
R=-
diation below 350 nm, eliminates the higher- Dispersion fJl GntiDp. The angular dis- .u =b-d)' be expected for the same size of dispersing
element. In addition, reflection gratings pro-
order spectra associated with first-order persion of a grating can be obtained by
radiation in most of the visible region. The differentiating Equation 5-5 while holding i vide a means of dispersing radiation in the
example which follows illustrates these points. constant; thus, at any given angle of incidence far ultraviolet and far infrared regions where
.t I 460.25 I
b-- x-- ------ x-- absorption prevents the use or prisms.
EXAMPLE dr D
.u dn/d.t 460.30 - 460.20 dn/d)' Gratings have the disadvantage of produc-
An echellette grating containing 2000 -=-- (5-6) For fused silica, ing somewhat greater amounts of stray radia-
d)' dcosr
blazes per millimeter was irradiated with a tion as well as higher-order spectra. These
polychromatic beam at an incident angle 48 b = 460.25 nm x I x 10-7 em disadvantages are ordinarily not serious be-
Note that the dispersion increases as the dis-
deg to the grating normal. Calculate the 0.10 nm 1.3 x 10-4 nm-I om cause the extraneous radiation can be mini-
tance d between rulings decreases or as the
wavelengths of radiation that would appear number of lines per millimeter increases. Over mized through the use of filters and by proper
= 3.5 em
at an angle of reflection of +20, + 10,0, and short wavelength ranges, cos r does not instrument design.
-10 deg (angle r, Figure 5-11). change greatly with .t, so that the dispersion of and f<?rglass, For a number of years, high-quality grat-
To obtain d in Equation 5-5, we write a grating is nearly linear. By proper design of ings were more expensive than good prisms.
the optics of a grating monochromator, it is
b = 460.25 1 x 10-7 = 1.3 em This disadvantage has now largely disap-
Imm nm nm 10.10 x 3.6 x 10-4
d=---- x 106- =500-- possible to produce an instrument that for all peared.
2000 blazes mm blaze
practical purposes has a linear dispersion of
radiation along the focal plane of the exit slit. Monochromator Slits
When r in Figure 5-11 equals + 20 deg
Figure 5-10 shows the contrast between a The slits of a monochromator play an impor-
, SOO(. 48 . 2) 542.6 grating and a prism monochromator in this 1 R. A. Sawyer. Experimental SpeClroscoPY. 2d cd. New tant part in determining its quality. Slit jaws
A=- sm +sm 0 =--
D n regard. Jersey: Prentic:e-Hal~ 1951. p. 130. are formed by carefully machining two pieces
Monochromator A peak power. Figure 5-14 illustrates the .rela-
of metal to give sharp edges, as shown in setting ..l Movement of the .monochromator in either
Figure 5-12. Care is taken to assure that the I tionship between the effective bandwidth of direction diminishes the transmitted intensity
edges of the slit are exactly parallel to one I an instrument and. its effectiveness in resolv- of A2' but increases the intensity of one of the
I ing spectral peaks. Here, the exit slit of a
another and that they lie on the same plane. other lines by an equivalent amount. As
I monochromator is illuminated with a beam
In some monochromators, the openings of I shown bJl the solid line in the plot to the
the two slits are fixed; more commonly, the composed of just three wavelengths, AI> A2' right, no spectral resolution of the three wave-
spacing can be adjusted with a micrometer and A); each beam has the same intensity. In lengths is achieved.
mechanism. I the top figure, the effective bandwidth of the In the middle drawing of Figure 5-14, the
The entrance slit (see Figure 5-7) serves as I instrument is exactly equal to the difference in effective bandwidth of the instrument has
I wavelength between AI and AZ or AZ and A).
a radiation source; its image is focused on the I been reduced by narrowing the openings of
surface containing the exit slit. If the radia- I When the monochromator is set at AZ , radia- the exit and entrance slits to three-quarters
tion source consists of a few discrete I tion of this wavelength just fills the slit. that of their original dimensions. The solid
wavelengths, a series of rectangular images I
appears on this surface as bright lines, each I Effective bandwidth
corresponding to a given wavelength. A partic-
ular line can be brought to focus on the exit Radiant Monochromator
power sening
slit by rotating the dispersing element. If the p
entrance and exit slits are of the same size (as A, A. A"

is usually the case ~ the image of the entrance Relative


A, A. A3 slit
slit will in theory just fill the exit -slit opening width
I--Bandwidth--t
when the setting of the monochromator cor-
Monochromator
responds to the wavelength of the radiation. setting. A
Movement of the monochromator mount in
one direction or the other results in a contin- FIGURE 5-13 Illumination of an exit slit A, A2 A3
uous decrease in emitted intensity, zero being by monochromatic radiation AZ at various ;;

reached when the entrance-slit image has monochromator settings. Exit and entrance •.
'j( Effective
bandwidth
been displaced by its full width. slits are identical. g
Figure 5-13 illustrates the situation in
which monochromatic radiation of wave-
~
1p
length A2 strikes the exit slit. Here, the mono- Cl.:
chromator is set for A2 and the two slits are suits in the image being moved completely ~
0
identical in width. The image of the entrance out of the slit. The lower half of the figure Q. A, A. A3
slit just fills the exit slit. Movement of the shows a plot of the radiant power emitted as i
ia: Effectiv.

,4l
monochromator to a setting of AI or A) re- a function of monochromator setting. Note A, A. A3 bandwidth
that the bandwidth is defined as the span of

'1JJIfJl
monochromator settings (in units of wave-
length) needed to move the image of the
C entrance slit across the exit slit. If polychro-
I
I matic radiation were employed, it would also
I
represent the span of wavelengths emitted A, A2 A3
A-~.i..._':'/ +~.-"""';-""--B from the exit slit for a given monochromator Monochromator
~:~ I setting. setting A

o The effective bandwidth for a monochroma- FIGURE 5-14 The effect of the slit wiath on spectra.
The entrance slit is illuminated with AI> AZ' and A) only.
A-IIa-"'-B tor is also defined in Figure 5-13 as one-half
the wavelength range transmitted by the in- Entrance and exit slits are identical. Plots on the right show
FIGURE 5-12 Construc- strument at a given wavelength setting, or the changes in emitted power as the setting of monochromator
tion of slits. range of wavelengths encompassed at half is varied.
line in the plot on the right shows that partial if one of the elements is a gratin& higher-order zero so that the equation accompanying
resolution of the three lines results. When the wavelengths are removed by the second Figure 5-16 simplifies to
effective bandwidth is decreased to one-half element.
!lA' = 2t (5-8)
the difference in wavelengths of the three
beams, complete resolution is obtained, as where l' is the wavelength of radiation in the
shown in the bottom drawing. 0.2 WAVELENGTH dielectric and t is the thickness of the dielec-
The effective bandwidth of a monochroma- E SELECTION; FILTERS tric. The corresponding wavelength in air is
E given by
tor depends upon both the dispersion of the .<:'
prism or grating as weIl as the width of the
entrance and exit slits. Most monochroma- "
·i Absorption and interference filters are
employed for wavelength selection. The
l=l'n
tors are equipped with variable slits so that fonner are restricted to the visible region of where n is the refractive index of the dielectric
the effective bandWidth can be changed. the spectrum. Interference filters, on the other medium. Thus, the wavelengths of radiation
The use of minimal. slit widths is desirable hand, are available for ultraviolet, visible, and transmitted by the filter are
where the resolution of narrow absorption or infrared radiation.
emission bands is needed. On the other hand, 1 = Un
there is a marked ~ease in the available D

radiant power as the slits are narrowed, and Interf ••.• nce Filters The integer n is the order of interference. The
300 400 500
ac:c:urate measurement of this power becomes glass layers of the filter are often selected to
wavelength. nm
more difficult. Thus, wider slit widths may be As the name implies, interference filters rely absorb all but one of the reinforced bands;
used for quantitative analysis than for quali- FIGURE 5-15 The slit width needed to on optical interference to provide relatively transmission is thus restricted to a single
tative work, where spc;etral detail is important. maintain a constant eft'ective bandwidth of narrow bands of radiation (see Figure 5-17). order.
As noted previo~ly, the dispersion of a 1 nm from a quartz Littrow prism mono- An interference filter consists of a transparent Figure 5-17 illustrates the performance
prism is not linear (see Figure 5-10). Much chromator. dielectric (frequently calcium fluoride or mag-
characteristics of typical interference filters.
narrower slits muSt, therefore, be employed at nesium fluoride) that occupies the space be-
long wavelengths than at short in order to tween two semitransparent metalIic films
obtain radiation of a given effective band- coated on the inside surfaces of two glass
width. The slit width in millimeters required parts also causes stray radiation to reach the plates. The thickness of the dielectric layer is
to maintain an effective bandwidth of one cdt slit. Generally, the efl'ects of spurious ra- carefully controlled and determines the wave-
nanometer for a monochromator equipped diation are minimimJ by introducing bafJles in length of the transmitted radiation. When a
with. Littrow prism is shown in Figure 5-15. appropriate spots in the monochromator and perpendicular beam of collimated radiation
One of the advantages of a grating mono- by coating interior surfaces with flat black strikes this array, a fraction passes through
chromator is that a fixed slit width produces paint. In addition, the monochromator is the first metallic layer while the remainder is
radiation of nearly constant bandwidth, re- sealed with windows over the slits to prevent reflected. The portion that is passed under-
gardless of wavelength. entrance of dust and fumes. Despite these goes a similar partition upon striking the
precautions, however, some spurious radia- second metallic film. If the reflected portion
tion is still emitted; we shaIl find that its from this second interaction is of the proper FIGURE 5-16 The interference filter. At
presence can have serious effects on absorp- wavelength, it is partiaIly reflected from the point), light strikes the semitransparent film
tion measurements under certain conditions. inner side of the first layer in phase with in- at an angle 8 from the perpendicular, is
The exit beam of a monochromator is usuaIly coming light of the same wavelength. partiaIly reflected, partiaIly passed. The same
contaminated with smaIl amounts of radia- The result is that this particular wavelength is process occurs at I', 2, 2', and so forth.
Double Monochromator.
tion having wavele~hs far different from reinforced, while most other wavelengths, For reinforcement to occur at point 2, the
that of the instrument setting. There. are Many modern monochromators contain two being out of phase, suffer destructive distance traveled by the beam reflected at l'
several sources of this unwanted radiation. dispersing elements; that is, two prisms, two interference. must be some multiple of its wavelength in
Reflections of the beam from various optical gratings, or a prism and a grating. This ar- Figure 5-16 illustrates an interference filter. the medium 1'. Since the path length between
parts and the monochromator housing con- rangement markedly reduces the amount of For purposes of clarity, the incident beam is surfaces can be expressed as tlcos 8, the
tribute to it. Scattering by dust particles in stray radiation and also provides greater dis- shown as arriving at an angle 8 from the condition for reinforcement is that !lA' =
the atmosphere or on the surfaces of optical persion and spectral resolution. Furthermore, perpendicular. In ordinary use, 8 approaches 2tlcos 8 where n is a small whole number.
the wedge, one can isolate a bandwidth of ing certain portions of the spectra. The most stance employed for cell windows in the infra-
about '20 nm. common type consists of colored glass or of a red region; the other infrared transparent
Interference wedges are available for the dye suspended in gelatin and sandwiched be- materials listed in Table 5-1 (p. 126) may also
visible region (400 to 700 om~ the near-infra- tween glass plates. The former have the be used for this purpose.
red region (1000 to 2000 nm~ and for several advantage of greater thermal stability. The best cells have windows that are per-
parts of the infrared region (2.5 to 14.5 JlI11). Absorption filters have effective band- fectly normal to the direction of the beam in
widths that range from perhaps 30 to 250 om order to minimize reflection losses. The most
Absorption Filters (see Figure 5-18). Filters that provide the nar- common cell length for studies in the ultra-
rowest bandwidths also absorb a significant violet and visible regions is I em; matched,
Absorption filters, which are generally less ex-
fraction of the desired radiation, and may calibrated cells of this size are available from
pensive than interference filters, have been
lu(ve a transmittance of 0.1 or less at their several commercial sources. Other path
widely used for band selection in the visible
band peaks. Glass filters with transmittance lengths, from 0.1 em (and shorter) to 10 em,
region. These filters limit radiation by absorb-
maxima throughout the entire visible region can also be purchased. Transparent spacers
are available commercially. for shortening the path length of l-em cells to
.Cut-off filters have transmittances of nearly 0.1 em are also available.
100% over a portion of the visible spectrum, Cells for infrared studies of liquids and so-
5110 6215 6225 6940 6960
Wavelength, A but then rapidly decrease to zero transmit- lutions generally have path lengths of less than
tance over the remainder, A narrow spectral Imm.
FIGURE 5-17 Transmission character- Interle<wnc:e
band can be isolated by coupling a cut-off For reasons of economy, cylindrical cells
istics of typical interference filters, filler
Iso filter with a second filter (see Figure 5-18b).
;It is apparent from Figure 5-18a that the
are sometimes employed in the ultraviolet
and visible regions. Care must be taken to
1 pejformance characteristics of absorption duplicate the position of the cell with respect
Ordinarily, filters are characterized, as shown, S a.ndwidth
-8nm fil~rs are significantly inferior to those of to the beam; otherwise, variations in path
by the wavelength of their transmittance i interference-type filters. Not only are their length and reflection losses at the curved sur-
peaks, the percentage of incident radiation ~25
CL bandwidths greater, but the fraction of light faces can cause significant errors.
transmitted at the peak (their percent trallS-
AbsorPtion filter transmitted is also less. Nevertheless, absorp- The quality of absorbance data is critically
mitumee, p. 149~ and their bandwidths at a tion filters are totally adequate. for many dependent upon the way the matched cells
transmittance equal to one-half the maximum
applications. are used and maintained. Fingerprints, grease,
(the effective bandwidth). or other deposits on the walls alter the trans-
0
Interference filters are available through- 400 500 600
mission characteristics of a cell markedly.
out the ultraviolet and visible regions and up W-'engttI,nm
(I)
Thus, thorough cleaning before and after use
to about 141llf1in the infrared. Typically, half
Sample containers arc required for all spec- is imperative; the surface of the windows
bandwidths are about 1,5% of the wavelength 100
Absorption
troscopic studies except emission spectros- must not be touched during the handling.
at peak transmittance, although this figure is
copy. In common with the optical elements Matched cells should never be dried by heat-
reduced to 0.15% in some narrow-band •.l! filter> Orlnge
cut .••"
filter of monochromators, the cells or cuvettes that ing in an oven or over a flame-such treat-
filters; these have maximum transmittances of
!! hold the samples must be made of material ment may cause physical damage or a change
about 10%. .~ in path length. The cells should be calibrated
c: SO tha.t passes radiation in the spectral region of
S interest. Thus, as shown in Figure 5-2, quartz against each other regularly with an absorb-
Interference Wedges
•.~•.
'C or fused silica is required for work in the ing solution .
An interference wedge consists of a pair of CL ult~aviolet region (below 350 nm); both of
mirrored, partially transparent plates sepa- theSe substances are transparent in the visible
rated by a wedge-shaped layer of a dielectric 0 region and to about' 3 Ilm in the infrared
400 500 600 700
materiaL The length of the plates ranges from Wavelength, nm
region as well. Silicate glasses can be
about 50 to 200 mm. The radiation trans- lbl
employed in the region between 350 and 2000
mitted varies continuously from one end to the nm. Plastic containers have also found appli- Early instruments for the measurement of
other' as the thickness of the wedge varies. FIGURE 5-18 Comparison of various cation in the visible region. Crystalline emission and absorption of radiation required
By choosing the proper linear position along types of filters for visible radiation. sodium chloride is the most common sub- visual or photographic methods for detection.
Glass ThinI.yerof liI_
choice of exper,imental conditions minimize
These detectors have been largely supplanted ual events (absorption of one photon). the this effect.
by transducers which convert radiant energy
into an electrical signal; our discussion will
be confined to detectors of this kind. Brief
probability of which can be described by the
use of statistics. In contrast. heat transducers,
which are required for the detection of infra-
~:~~~. Barrier-type cells are widely used in
simple, portable. instruments where rug-
gedIlCSljand low cost are important. For rou-
red radiation, are nonquantized sensors. tine analyses. these instruments often provide
consideration of photographic detection will
As will be shown in Chapter 6. the distinc-
~
be found in Chapter 12. perfectly reliable analytical data.
tion between photon and heat detectors is im- FIGURE 6-19 Schematic of a typical Phototubes. A second type of photoelectric
Properties of Photoelectric Detectors. To
portant because shot noise (p. 7!) often barrier-layer cell. device is the phototube. which consists of a
be usefu~ the detector must respond to ra-
diant energy over a broad wavelength range. It limits the behavior of the former whIle John- semicylindrical cathode and a wire anode
should, in addition. be sensitive to low levels son noise may limit the latter. As a sealed inside an evacuated transparent
consequence. indeterminate errors associated envelope. The concave surface of the elec-
of radiant power. respond rapidly to the
radiation, produce an electrical signal that with the two types of detectors are different. the entire array is protected by a transparent trode supports a layer of photoemissive mate-
can be readily amplified, and have a relatively envelope. When radiation of sufficient energy rial (p. 100) that tends to emit electrons upon
low noise level (for stability). Finally, it is reaches the semiconductor. covalent bonds being irradiated. When a potential is applied
Photon Detectors are broken, with the result that conduction
essential that the signal produced be directly across the electrodes, the emitted electrons
proportional to the beam power P; that is, Several types of photon detectors are avail- electrons and holes are formed. The electrons Oow to the wire anode, generating a photo-
able. including: (1) photoooltaic cells. in then migrate toward the metallic film and the current. The currents produced are generally
G = KP. + K' (5-10) which the radiant energy generates a current holes toward the base upon which the semi- about one-tenth as great as those from a
where G is the electrical response of the detec- at the interface of a semiconductor layer and conductor is deposited. The liberated elec- photovoltaic cell fot a given radiant intensity.
tor in units of current. resistance, or emf. The a metal; (2) phototubes. in which radiation trons are free to migrate through the external In contrast, however, amplification is easily
constant K measures the sensitivity of the causes emission of electrons from a photo- circuit to interact with these holes. The result accomplished since the phototube has a high
detector in terms of electrical response per sensitive solid surface; (3) photomultiplier is an electrical current whose magnitude is electrical resistance. Figure 3-19 (p. 59) is
unit of radiant power. Many detectors exhibit tubes, which contain a photoemissive surface proportional to the number of photons strik- a schematic diagram of a typical phololube
a small constant response. known as a dark as well as several additional surfaces that emit ing the semiconductor surface. Ordinarily. the arrangement.
currenl K'. even when no radiation impinges a cascade of electrons when struck by elec- currents produced by a photovoltaic cell are The number of electrons ejected from a
on their surfaces. Instruments with detectors trons from the photosensitive area; (4) semi- large enough to be measured with a galva- photoemissive surface is directly proportional
that have a dark-eurrent response are or- conductor detectors. in which absorption of nometer or microammeter; if the resistance of to the radiant power of the beam striking that
dinarily equipped with a compensating circuit radiation produces electrons and holes, thus the external circuit is kept small, the magni- surface. As the potential applied across the
that permits application of a countersignal to leading to enhanced conductivity; and (5) sili- tude of the photocurrent is directly propor- two electrodes of the tube is increased, the
reduce K' to zero. Thus. under ordinary cir- con diodes. in which photons increase the con- tional to the power of the radiation striking fraction of the emitted electrons reaching
cumstances. we may write ductance across a reverse biased pn junction. the cell. Currents on the order of 10 to 100 the anode rapidly increases; when the satura-
Photovoltak: or Barrier-Layer Cells. The pA are typical. tion potential is achieved, essentially all of
photovoltaic cell is used primarily to detect The barrier-layer cell constitutes a rugged, the electrons are collected at the anode.
Types of Photoelectric Detectors. As in- and measure radiation in the visible region. low-eost means for measuring radiant power. The current then becomes independent of
dicated in Figure 5-2, two general types of The typical cell has a maximum sensitivity at No external source of electrical energy is potential and directly proportional to the ra-
radiation transducers are employed; one re- about SSO nm; the response falls off to per- required. On the other hand. the low internal diant power (see Figure 5-20). Phototubes are
sponds to photons, the other to heat. All haps 10"10 of the maximum at 350 and 750 resistance of the cell makes the amplification usually operated at a potential of about 90 V,
photon detectors are based upon interaction nm. Its range approximates that of the human of its output less convenient. Thus, although which is well within the saturation region.
of radiation with a reactive surface to pro- eye. the barrier-layer cell provides a readily A variety of photoemissive surfaces are
duce electrons (photoemission) or to promote The photovoltaic cell consists of a flat measured response at high levels of illumina- used in commercial phototubes. Some consist
electrons to energy states in which they can copper or iron electrode upon which is tion, it suffers from lack of sensitivity at low of one or more of the alkali metals intermixed
conduct electricity (photoconduction). Only deposited a layer of semiconducting material, levels. with alkali metal oxides and silver oxide. This
ultraviolet, visible, and near-infrared radia- such as selenium or copper(l) oxide (see Another disadvantage of the barrier-type mixture is formed on the surface of the cath-
tion have sufficient energy to cause these Figure 5-19). The outer surface of the semi- cell is that it exhibits fatigue; that is, its cur- ode. Other surfaces contain such semi-
processes to occur. Photoelectric detectors conductor is coated with a thin transparent rent output decreases gradually during con- conductors as Cs3Sb or Na2KSb mixed with
also differ from heat detectors in that their metallic film of gold, silver, or lead. which tinual illumination; proper circuit design and small amounts of the alkali metals. Some
electrical signal results from a series of individ- serves as the second or collector electrode;
electrons are accelerated toward it as a Several electrons
consequence. Upon striking the dynode, each for -" incident
electron
photoelectron causes emission of several
additional electrons; these, in turn, are ac-
celerated toward dynode 2, which is 90 V
more positive than dynode 1. Again, several
electrons are emitted for each electron strik-
ing the surface. By the time this process has
been repeated nine times, 106 to 107 electrons
have been formed for each photon; this
cascade is finally collected at the anode. The
resulting current is then electronically ampli-
FIGURE 5-20 Variation of photo- fied and measured. Anode. - 10'
tube response with the applied potential Semiconductor Detecton. The most sensi- electrons for each
photon
at three levels of iIlumination. The tive detectors for monitoring radiation in the
shaded portion indicates the saturation near-infrared region of about 0.75 to 3 JlM are
region. Note the linearity of response at photoconductors whose resistances decrease
the higher applied potentials. 900 V de
when radiation within this range is absorbed.
Crystalline semiconductors are formed
from the sulfides, selenides, and stibnides of
such metals as lead, cadmium, gallium, and
cathodes are semitransparent, with the indium. Absorption of radiation by these
photoemissive material being present on the
surface opposite that which faces the radiation. ~ Cathode
As shown in Figure 5-21, the photo- Anode \
emissive coating on the cathode determines To numbered dynodes
the spectral response of a phototube. When shown above

sensitivity to radiation below 350 nm is


required, the tube must be supplied with an
envelope or window of quartz. Phototubes
are available for various ranges within the
J~t To

Ibl
region between 180 and 1()()()nm. >
Phototubes frequently produce a small .~ 0.6 FIGURE 5-22 Photomultiplier tube: (a) cross section of the
current in the absence of radiation; this dark .~ tube; (b) electrical circuit.
~
current (see Equation 5-10) results from .~
thermally induced electron emission. ~ 0.4
Photomultiplier Tubes. For the measure- a:

ment of low radiant power, the photomulti- materials promotes some of their bound elec- the region between 0.8 and 2 JlM (12,500 to
plier tube offers advantages over the ordinary trons into an energy state in which they are 5000 em - 1). A thin layer of this compound is
phototube. Figure 5-22 is a schematic dia- free to conduct electricity. The resulting deposited on glass or quartz plates to form
gram of such a device. The cathode surface is change in conductivity can then be measured the cell. The entire assembly is then sealed in
similar in composition to that of a photo tube, with a Wheatstone bridge arrangement an evacuated container to protect the semi-
electrons being 'emitted upon exposure to o (Figure 2-24, p. 35) or, alternatively, with a conductor from reaction with the atmosphere.
200 400 600 circuit such as that shown in Figure 3-21b Silicon Diode Detecton. A silicon diode
radiation. The tube also contains additional
Wavelength, om
electrodes (nine in Figure 5-22) called (p.62). detector consists of a reverse-biased pn junc-
dynodes. Dynode 1 is maintained at a poten- FIGURE 5-21 Spectral responses of some The most widely used photoconductive tion formed on a silicon chip. As shown in
tial 90 V more positive than the cathode, and photoemissive surfaces. material is lead sulfide, which is sensitive in Figure 3-1d (p. 43), the reverse bias creates a
depletion layer which reduces the conduct- n-type surface opposite to it. Then, holes are TII'JIOt
ance of the junction to nearly zero: If radia- created which diffuse toward the other surface wray of
•••• Iconductor
tion is allowed to impinge on the n region, and discharge several of the capacitors that diodas
however, holeS and electrons are formed. The lie along the extension of the path of the inci-
former can diffuse through the depletion layer denr beam. The electron beam then recharges
to the p region where they are annihilated; an these- capacitors when it again reaches them
increase in conductance, which is propor- in the course of its sweeps (see the end view,
tional to radiant power, is the consequence. Figure S-23b). This charging current is then
A silicon diode detector is not so sensitive amplified and serves as the signal.
as a photomultiplier tube. It has, however, The width of the electron beams is such
become important because arrays of these ( - 20 JIm) that the surface of the target is
detectors can be formed on the surface of a effectively divided into SOOlines or channels.
single silicon chip. These chips are an impor- The signal from each of these channels can be
tant component of vidicon tubes.s stored separately in a computer memory.
Figure 5-23a is a schematic diagram of a Thus, if the vidicon tube is placed at the focal 5i02
vidicon tube. It is similar in construction to a plane of a monochromator, the signal from Face insulating 8"m
pl.1ll rwgIon i02
television tube in which a stream of electrons each channel corresponds to radiation of a
is employed to scan a target area with a series different wavelength. It therefore becomes . p region
of successive horizontal sweeps. The target in possible to obtain a spectrum without move-
Figure 5-23 is a silicon diode array with a ment of the monochromator. That is, scan-
diameter of 16 mm and containing over ning is accomplished electronically rather + \ Depletion
layer ~ ~ Electron
15,000 photodiodes per square millimeter. than optically, and channel numbers rather ocan beam
Figure 5-23b is a magnified cross section than monochromator settings serve as a
and end view of a part of the target array of measure of wavelength for spectra.
lb)
photodiodes. Each photodiode consists of a An instrument consisling of a monochrom-
cylindrical section of p-type silicon sur- ator, vidicon tube, and computer, such as FIGURE 5-23 A vidicon tube detector: (a) vidicon tube
rounded by an insulating layer of silicon diox- that just described, is called an optical multi- and (b) target array of silicon diodes.
ide. Each diode is thus electrically independent channel analyzer. It is of considerable impor-
of neighboring diodes; all are connected to the tance because it allows a portion or all of a
common II-type layer. It should be emphasized spectrum to be recorded essentially simultan-
that the schematic has been greatly magnified; and 307 nm. By sacrificing resolution, a much
eously. A disadvantage is that the physical positions. 11 A spectral resolution of about 1
indeed, the distances between centers of the p larger wavelength range can be covered.
size of the detector places a limit on either the nm is achieved over the range from 200 to
regions is only about 8 Jim. Thus, a compact multichannel spectrometer 800 nm.
wavelength range or the resolution of such an
When the surface of the target is swept by for field use has been designed that covers a
analyzer. For example. when a vidicon tube
the electron beam, each of the p-type cylin- 650 nm wavelength rangelO; its resolution,
was mounted at the focal plane of a typical Heat Detectors
ders becomes successively charged to the however, is only 10 nm. Instruments have
. grating monochromator, the face of the tube
potential of the beam and acts as a small also been designed to provide a two dimen- The measurement of infrared radiation is
corresponded to only about 27.0 nm for any
charged capacitor, which remains charged sional display of spectra on the surface of a difficult as a result of the low intensity of
given wavelength setting.9 The resolution,
(except for leakage) until a photon strikes the vidicon tube. In one of these, an optical ~vailable sources and the low energy of the
however, was better than I nm and allowed
system is employed which divides the ultra- mfrared photon. As a consequence of these
the simultaneous determination of lead, tin,
violet-visible spectrum into six 100 nm seg- properties, the eleclrical signal from an infra-
nickel, and iron in a brass, based upon the
ments which are displayed along the red detector is small, and its measurement re-
emission lines of these elements between 280
• For a discussion of vidicon tubes and other analogous horizontal axis of the tube at different vertical quires large amplification. It is usually the
devices, see: Y. Talmi, Anal. Chem. ~, 697A, 658A
(1975); and G. Horlick and E. G. Codding, in Conrem-
porary Topics in Analytical and Clinical Chemistry, cds.
D. M. Hercules. G. M. Hicftjc, L. R. Snyder, and M. A. • E. G. Codding and G. Horlick, Spectrosc. Lett., 7 (I~ 10 G. A. H. Walker, et al., Rev. Sci. Instrum. 45. 1349
Evenson.New York: Plenum Press. 1977,vol. I, Chapter 4. IIR. M. Hoffman and H. L. Pardue. Anal. Chem. 51,
33 (1974). (1974~ 1267 (1979).
COMPONENTS OF INSTRUMENTS FOR OPTICAL SPECTROSCOPY

detector system that limits the sensitivity and Thermocouples. In its simplest form, a Spectrophotometer
the precision of an infrared instrument.12
The convenient phototubes discussed ear-
thermocouple consists of the junctions
formed when two pieces of a given metal are
-,-
II
slit

Thermocouple
lier are generally not applicable in the infra-
red because the photons in this region lade
fused to either end of a dissimilar metal or a
semiconducting metal alloy. A potential
tt junction

the energy to cause photoemission of elec-


trons. Thus, thermal detectors and detection
develops between the two thermocouple junc-
tions which varies with the temperature differ-
~---- -;:]
based upon photoconduction are required. ence between the two junctions. For most
Neither of these is as satisfactory as the applications, one of the junctions (the refer-
photocell. ence junction) is held at a constant tempera-
Thermal detectors, whose responses ture (often in an ice bath) and the second
depend upon the heating effect of the radia- junction serves as the temperature-sensitive
tion, can be employed for detection of all but detector.
the shortest infrared wavelengths. With these The junction for infrared detection is
devices, the radiation is absorbed by a small formed from very fine wires of such metals as
black body and the resultant temperature in-
crease is measured. The radiant power level
platinum and silver or antimony and
bismuth. Alternatively, the detector junction
'=" -15V I
from a spectrophotometer beam is minute may be constructed by evaporating the metals FIGURE.5-24 Thermocouple and preamplifier. (Ad~Pted from
(10-7 to 10-9 W~ so that the heat capacity of onto a nonconducting support. lit either case. G. W. Ewmg, J. Chern. Educ., 48(9), A52l, (1971). With permission.)
the absorbing element must be as small as the junction is usually blackened (to improve
possible if a detectable temperature rise is to its heat absorbing capacity) and sealed in an
be produced. Every effort is made to mini- evacuated chamber with a window that is
mize the size and thickness of the absorbing transparent to infrared radiation. absorb the radiant heat. Bolometers are not
beam that strikes the active sur/ace of the
element and to concentrate the entire infra- The reference junction is designed to have so extensively used as other infrared detectors
phototube is changed; a change in' photocur:
red beam on its surface. Under the best of a relatively large heat capacity and is care- Pyr~r~ Detectors. Certain crystals:
rent results, which can be related to the
circumstances, temperature changes are fully shielded from the incident radiation. such as lithtum tantalate, barium titanate
power of the infrared beam.
confined to a few thousandths of a degree Because the analyte signal is chopped, only and triglycine sulfate, possess temperature~
sensitive dipole moments. When placed be- The Golay cell is more expensive than
Celsius. the difference in temperature between the two other heat detectors and no more sensitive to
The problem of measuring infrared radia- junctions is important; therefore, the refer- tween metal plates, a temperature-sensitive
near- and mid-infrared radiation; thus, it is
tion by thermal means is compounded by eoce junction does not need to be maintained capacitor ~ formed which can be employed
seldom employed for these spectral regions.
thermal effects from the surroundings. The at constant temperature. for rneasunng the power of infrared radiation.
. <?n .the other hand, the Golay detector is
absorbing element is housed in a vacuum and A well-designed thermocouple detector is Here, the transduced signal is capacitance.
slgmficantly superior for radiation with wave-
is carefully shielded from thermal radiation capable of responding to temperature differ- ~ Golay Detector. The Golay detector is
lengths greater t~an. SO Ilm (200 em - I); it
emitted by other nearby objects. In order to ences of 10-hC. This figure corresponds to a essentially a sensitive gas thermometer in
finds g~neral use m mstruments designed for
minimize the effects of extraneous heat potential difference of about 6 to 8 pV/pW. which xenon is contained in a small cylindri- the far-mfrared region.
sources, a chopped beam is always employed The thermocouple of an infrared detector is a cal chamber, which contains a blackened
in infrared measurements. Thus, the analyte low-impedance device; it is frequently con- membrane. One end of the cylinder is sealed
signal, after transduction, has the frequency of nected to a high-impedance preamplifier such with an infrared window; the other consists of
the chopper; with appropriate circuitry, this as that shown in Figure 5-24. a flexible diaphragm which is silvered on the SIGNAL
signal is readily separated from extraneous Bolometers. A bolometer is a type of re- outside. A light beam is reflected from this PROCESSORS AND READOUTS
silvered surface to the cathode of a vacuum
radiation signals, which ordinarily vary only sistance thermometer constructed of strips of
phototube. When infrared radiation enters
!
slowly with time. metals such as platinum or nickel, or from a
the cell, the blackened membrane is warmed The. signal. processor is ordinarily an elec-
semiconductor; the latter devices are some- tromc deVice that amplifies the electrical
times called thermistors. These materials ex- which in turn heats the xenon by conduction:
signal from the detector; in addition, it may
12 For further information on infrared deteaors, see:
hibit a relatively large change in resistance The resulting increase in pressure causes dis-
alter the signal from dc to ac (or the reverse~
G. W. Ewing, J. CMm. Educ ••• (9~ AS21 (1971);and H. as a function of temperature. The responsive tortion of the silvered diaphragm. As a
change the ph~ of the signal, and filter it to
Levinstein, Anal. CMm. 41 (14~ 81A (1969~ element is kept small and blackened to consequence, the fraction of the reflected light
remove unwanted components. Furthermore,
the signal processor may be called upon to decay of radioactive species (these techniques order for the first-order diffraction line for 1 - 500 nm to be observed
perform such mathematical operations on the are considered in detail in Chapter 15~ Only at a reftection angle of -40 deg when the angle of incidence is 60 deg?
signal as differentiation, integration, or con- recently, however, has photon counting been
version to a logarithm. 6. Consider an infrared grating with 72.0 lines per millimeter and
applied to ultraviolet and visible radiation. 13
Several types of the readout devices 5.00 mm of illuminated area. Calculate the first-order resolution (l/Al)
Here, the output of a photomultiplier tube is
described in Chapters 2 and 3 are found in of this- grating. How far apart (in an -I) must two lines centered at
employed. In the previous section, it was in-
modern instruments. Some of these include dicated that a single photon striking the cath- 1000 an-I be if they are to be resolved?
the d'Arsonval meter, digital meters, the ode of a photomultiplier ultimately leads to a 7. Calculate the angular dispersion (d8/d1) at 500 nm for a Cornu type
scales of potentiometers, recorders, and cascade of 106 to 10' electrons or a pulse of prism having a prism angle IX of 60 deg em -1 assuming the material of
cathode-ray tubes. current which can be amplified and counted. construction is
Generally, the equipment for photon (a) a dense flint glass.
Photon Counting counting includes a pulse-height discrimina- (b) fused quartz.
Ordinarily, the output from the p]1otoelectric tor (see Figure 15-13~ which rejects pulses 8. Use the curves in Figure 5-9 to estimate dn/dl and n at 500 nm for
detectors described in the previous section unless they exceed some predetermined mini- prisms constructed from the materials in Problem 7.
is processed and displayed by analog mum voltage. Such a device is useful because 9. A common infrared source consists of an inert solid that behaves as a
techni~ues. That is, the aver. current, dar1c current and instrument noise are often black-body radiator when electrically heated. The energy density of
potential, or conductance associated with the significantly smaller than the signal pulse and black-body radiation is given by the equation
detector is amplified and recorded or fed into are thus not counted; an improved signal-ta-
a suitable meter. As was indicated in Chapter noise ratio results. enelJY density at wavelength 1- 8;h (e<6'.IlT -1)-1
3 (p. 83~ analog signals such as these vary Photon counting has a number of advan-
Continuously; in' spectroscopy, th~ are gen- tages over analog signal processing, including where c is the speed of light, h is Planck'. constant, Ie is Boltzmann's
erally proportional to the aver.ge radiant improved signal-ta-noise ratio, sensitivity to constant, and T is temperature in OK. The total amount of radiant
power of the incident beam. I low radiation levels, improved precision for a energy per second per square centimeter of surface of a black body is
In some instances, it is possible 'and advan- given measurement time, and lowered sensitiv- given by E= (5.70 X 10-5 ergan-Z s-loK-4)'f4, where Tis tempera-
tageous to employ direct digital' techniques ity to voltage and temperature changes. The ture. The wavelength of maximum emission energy is given by
wherein electrical pulses produced by individ- required equipment is, however, more com- l..x - 0.289 an °K/T OK.A graph of black-body radiation intensity is
ual photons are counted. Here, radiant
power is proportional to the number of pulses
rather than to an average current or
plex and expensive; the technique has not yet
been widely applied for routine measurements
in the ultraviolet and visible regions.
(a) Calculate l...
shown in Figure 4-13.
for black body radiators at l500 K and at 1000oK.
0

(b) Calculate the ratio of energy density at 1500 an-I compared to


potential that at 1-. for black bodies at lSOO"Kand at l000"K.
Counting techniques have been used for (c) The center of a typical infrared spectrum is at a wavenumber near
IS For a review or photon countill& see: H. V. Malm-
many years for measuring the power of X-ray stadt, M. L. Franklin, and G. Horlick. Allal. CMm., •••• 1500 em-I. Why is it advantageous in infrared spectroscopy for
beams and of radiation produced by the (8~ 63A (1972~ the source to be heated to 1500 K instead of looo K?
0 0

I. State the differences between and the relative advantages of Littrow


and Bunsen prism monochromators. _
2. W~at ar~ the relative advantages and disadvantages of gratings and
pnsms for dispersion of light?
3. Why m4st the slit width of a prism monochromator be varied to
provide constant resolution. whereas a nearly constant slit width may
be used with a grating monochromator?
4. Why do quantitative and qualitative analyses often require different
monochromator slit widths?
5. For a grating, how many lines per centimeter would be required in
This chapter provides introductory material
which is applicable to ultraviolet-visible,
infrared, and atomic absorption spectroscopy. 1
Detailed discussion of these methods appears
in later chapters.
Po ,--I -I'------~
Absorbing
solution of
P
T =!...

A -log
Po
Po
Ii"

concentration c
TERMS EMPLOYED IN FIGURE 6-' Attenuation of a beam of
ABSORPTION SPECTROSCOPY radiation by an absorbing solution.

Table 6-1 lists the common terms and sym-


bols employed in absorption spectroscopy. In
recent years, a consi~erable effort has been of incident radiation transmitted by the solu-
made by the American Society for Testing tion. That is,
Materials to develop a standard nomencla-
ture; the terms and symbols listed in the first T = PIPo (6-1)
two columns of Table 6-1 are based on Transmittance is often expressed as a
ASTM recommendations. Column 3 contains percentage.
alternative symbols that will be found in the
older literature. A standard nomenclature
seems most worthwhile in order to avoid am-
biguities; the reader, is, therefore, urged to
learn and use the ri:commended terms and The absorbance of a solution is defined by the
equation
symbols.
Po
A = - Ioglo Toc log-
P
Transmittance
Note that, in contrast to transmittance, the
Figure 6-1 depicts a beam of parallel radia- absorbance of a solution increases as the
tion before and after it has passed through a attenuation of the beam becomes larger.
layer of solution with a thickness of b em and
a concentration of c of an absorbing species.
As a consequence of interactions between the Absorptivity and Molar Absorptivity
photons and absorbing particles, the power of
the beam is attenuated from Po to P. The As will be shown presently, absorbance is
transmittance T of the solution is the fraction directly proportional to the path length
through the solution and the concentration of
the absorbing species. That is,

1 For more detailed trelitment or absorption spectros-


A = abc (6-3)
copy. see: R. P. Bauman, Absorption Sp"ctroscopy. New where a is a proportionality constant called
York: Wiley, 1962; J. Ri Edisbury, Practical Hints on
Absorption Sp"ctrometryi New York: Plenum Press, the absorptivity. The magnitude of a will
1968;F. Grum, in Physical Methods o/Chemistry,eds. A. clearly depend upon the units used for b and
Weissberger and B. W. Roositer. New York: Wiley- c. When the concentration is expressed in
Interscience, 1972, Vol. I, Part III B, Chapter 3; G. F. moles per liter and the cell length is in cen-
Lothian, Absorption Sp"ctrophotometry, 3d ed. London: timeters, the absorptivity is called the molar
Adam Hilger Ltd., 1%9; and J. E. Crooks, The Sp"ctrum
absorptivity and given the special symbol e.
in ChemLstry. London: Academic Press, 1978.
Thus, when b is in centimeters and c is in by large molecules or inhomogeneities. Reflec- radiation after it has passed through a cell The meter is thus made direct reading in per-
moles per liter, tion losses can be appreciable; for example, containing the solvent or the solution of the cent transmittance. Obviously, an absorbance
about 4% of a beam is reflected upon vertical anaIyte. scale can also be scribed on the meter face.
passage of visible radiation across an air-to-
glass or glass-to-air interface (see example, M••••• rement of
Experimental p. 103). Trllll8lltlttance and Absorbance QUANTITATIVE ASPECTS OF
M••••• rement of P and Po In order to compensate for these effects, ABSORPTION MEASUREMENTS
Transmittance (or absorbance) measurements
the power of the beam transmitted through
The relationship given by Equation 6-1 or 6-2 are made with instruments having· the com-
the absorbing solution is generally compared The remainder of this chapter is devoted to
is not directly applicable to chemical analysis. ponents arranged as shown in Figure S-Ib
with that which passes through an identical an examination of Equation 6-4 (A = ebc~
Neither P nor Po, as defined, can be conven- (p. 115). The electrical output G of the detec-
cell containing the solvent for the sample. An with particular attention to causes of devia-
iently measured in the laboratory because the tor of such an instrument is given by (see
experimental absorbance that closely approx- tions from this relationship. In addition, con-
solution to be studied must be held in some p. 138).
imates the true absorbance of the solution sideration is given to the effects that
sort of container. Interaction between the ra-
diation and the walls is inevitable leading to a
can then be obtained; that is, G = KP + K' (6-5) uncertainties in the measurement of P and Po
have on absorbance (and thus concentration~
loss by reflection at each interface; moreover, p •••_. Po where K' is the dark current that often exists
significant absorption may occur within the A = Iog-- =Iog- when no radiation strikes the transducer and
p•••
8liea P Bee". Law
walls themselves. Finally, the beam may suffer K is a proportionality constant.
a diminution in power during its passage The terms Po and P, when used hence- In many instruments, the readout device Equations 6-3 and 6-4 are statements of
through the solution as a result of scattering forth, refer to the power of a beam or will consist of a linear scale calibrated in units Beer's law. These relationships can be ration-
of 0 to 100% T. A transmittance measure- alized as follows.2 Consider the block of
ment with such an instrument requires three absorbing matter (solid, liquid, or gas) shown
steps. First, with the light beam blocked from in Figure 6-2. A beam of parallel monochro-
the transducer by means of a shutter, an elec- matic radiation with power Po strikes the
TABLE &-1 IMPORTANT TERMS AND SYMBOLS EMPLOYED IN
trical adjustment is made until the pointer of block perpendicular to a surface; after pass-
ABSORPTION MEASUREMENT
the readout device is at exactly zero; this step ing through a length b of the material, which
Term ••• Symbol" Definition Akerutin Name and Symbol is called the dark current or 0% T adjustment.
A 100% T adjustment is then carried out with
Radiant power, P, Po Energy ofradiation (in ergs) Radiation intensity, I, 10 the shutter open and the solvent in the light
impinging on a l-an2 area
of a detector per second path. This adjustment may involve increasing
or decreasing the power output or the source
P electrically; alternatively, the power of the
Absorbance, A log~ Optical density, D; extinction, E
P beam may be varied with an adjustable
P diaphragm or by appropriate positioning of a
Transmittance, T Transmission, T comb or optical wedge. After this step, we
Po
may write Equation 6-5 in the form
Path lent!th of I,d
radiation, in em b Go = 100 = KPo + 0.00
A In the final step of the measurement, the FIGURE 6-2 Attenuation of radiation
AbsorptivitY,b a Extinction ~fficient, k with initial power Po by a solution con-
be solvent cell is replaced by the sample. The
meter reading is then given by taining c moles per liter of absorbing solute
A with a path length of b ern. P < Po·
Molar absorptivity,' £ Molar extinction coefficient
be G = KP +0.00
Dividing this equation by the previous one
• TerminololY recommended by the American Chemical Society. Reprinted with permission from AMI. CItmt., 24, 1349 yields 2 The discussion that follows is based on a paper by
(1952); •••• 2298 (1976~ Copyrisht by the American Chemiall Society. F. C. Strong. AMI. Ch"",. 24. 338 (t952). For a rigorous
• e may be expressed in g/liter or other specified concentration units; b may be expressed in em or in other units or length. P derivation of tbe law, see: D. J. Swinehart. J. Chem.
, e is expressed in units or mol/liter. G = - x 100 = T x 100
Po Edue.• 39. 333 (1972~
contains n absorbing particles (atoms, ions, or Upon converting to base lO logarithms and among the various species, the total absorb- absorber alters the molar absorptivity of the
molecules ~ its power is decreased tp P as a inverting the fraction to change the sign, we ance for a multicomponent system is given latter by electrostatic interactions; the effect is
result of absorption. Consider now a cross obtain by lessened by dilution.
section of the block having an area S and an Po an While the effect of molecular interactions
infinitesimal thickness dx. Within this section log Ii = 2.303S (6-8) A._I = Al+ A2 + ... + A. . (6-11) is ordinarily not significant at concentrations
there are dn absorbing particles; associated = £Ibel + £2be2 + ... + £.bc. below O.OlM, some exceptions are en-
with each particle, we can imagine a surface where n is the total number of particles within countered among certain large organic ions
at which photon capture will occur. That is, if the block shown in Figure 6-2. The cross- where the subscripts refer to absorbing com- or molecules. For example, the molar absorp-
a photon reaches one of these areas by sectional area S can be expressed in terms of ponents I, 2, ... , n. tivity at 436 nm for the cation of methylene
chance, absorption will follow immediately. the volume of the block V and its length b. blue is reported to increase by 88% as the dye
The total projected area of these capture sur- Thus, Limitation. to the concentration is increased from 10-5 to
faces within the section is designated as dS; Applicability of B•••.•• Law lO-2M; even below 10-6M, strict adherence
V
the ratio of the capture area to the total area, S=-cm2 to Beer's law is not observed.
b The linear relationship between absorbance
then, is dS/S. On a statistical average, this and path length at a fixed concentration of Deviations from Beer's law also arise be-
ratio represents the probability for the cap- Substitution of this quantity into Equation absorbing substances is a generalization for cause £ is dependent upon the refractive index
ture of photons within the section. 6-8 yields of the solution.3 Thus, if concentration
which no exceptions are known. On the other
The power of the beam entering the Po' anb hand, deviations from the direct proportional- changes cause significant alterations in the re-
section, P", is proportional to the number of log Ii = 2.303V (6-9) ity between the measured absorbance and fractive index n of a solution, departures from
photons per square centimeter per second, concentration when b is constant are fre- Beer's law are observed. A correction for this
and dP" represents the quantity removed per Note that n/V has the units of concentration quently encountered. Some of these devia- effect can be made by substitution of the
second within the section; the fraction ab- (that is, number of ~icles per cubic centi- tions are fundamental and represent reallimi- quantity en/(n2 + 2)2 for £ in Equation 6-10.
sorbed is then -dP"/P,,, and this ratio also meter); we can readily convert n/V to moles tations of the law. Others occur as a In general, this correction is never very large
equals the average probability for capture. per liter. Thus, i consequence of the manner in which the and is rarely significant at concentrations less
The term is given a minus sign to indicate absorbance measurements are made or as a than O.OlM.
that P undergoes a decrease. Thus, n particles .
c = --------- result of chemical changes associated with Chemical Deviatiol& Apparent deviations
6.02 )( lO23 particles/mol from Beer's law are frequently encountered as
concentration changes; the latter two are
dP" dS
-p;=S 1000 cm3/liter lOOOn . sometimes known, respectively, as instrumen- a consequence of association, dissociation, or
x Vem3 = 6.02 )( lO23V mol/llter tal deviations and chemical deviations. reaction of the absorbing species with the
Real Limitations to Beer's La". Beer's law solvent. A classic example of a chemical de-
Recall, now, that dS is the sum of the cap-
Combining this relationship with Equation is successful in describing the absorption be- viation occurs in unbuffered potassium dichro-
ture areas for particles within the section; it
6-9 yields havior of dilute solution only; in this sense, it mate solutions, in which the following
must therefore be proportional to the number
of particles, or is a limiting law. At high concentrations equilibria exist:
I Po 6.02 x lO23abe
(usually > O.OlM~ the average distance be- + H20
og Ii = 2.303 x 1000 Cr20~-
tween the species responsible for absorption
Finally, the constants in this equation can be is diminished to the point where each affects ~2HCr04 ~2H+ + 2CrO~-
where tin is the number of particles and a is a
proportionality constant, which can be called collected into a single term £ to give the charge distribution of its neighbors. This At most wavelengths, the molar absorptivities
the capture cross section. Combining Equa- interaction, in turn, can alter their ability to of the dichromate ion and the two chromate
Po .
tions 6-6 and 6-7 and summing over the inter- log Ii =;ebc = A (6-10) absorb a given wavelength of radiation. Be- species are quite different. Thus, the total
val between zero and n, we obtain cause the extent of interaction depends upon absorbance of any solution depends upon the
concentration, the occurrence of this phe- concentration ratio between the dimeric and
_ (dP" = (adn nomenon causes deviations from the linear re-
lationship between absorbance and concen-
the monomeric forms. This ratio, however,
·"0 P" ·0 S Application of changes markedly with dilution, and causes a
B•••.•• Law to Mixtu •••• tration. A similar effect is sometimes
which, upon integration, giv~ encountered in solutions containing low
Beer's law also applies to a solution contain- absorber concentrations and high concentra-
P an ing more than one kind of absorbing
-In-=- tions of other species, particularly elec- • G. Kortum and M. Seiler. Allgew. C"""'" 51, 687
Po S substance. Provided there is no interaction trolytes. The close proximity of ions to the (1939~
We are now able to calculate the absorbance ually, we may write for radiation A'
pronounced deviation from linearity between into the expression for K. makes it apparent
the absorbance and the total concentration of that, in the presence of strong base, essentially at the two wavelengths. Thus, , I P'o 'be
A=og-=e
chromium(VI). Nevertheless, the absorbance all of the indicator is dissociated to In ~ ; there- A430 = e1nb[In-] + t:Hlnb[HIn] P'
due to the dichromate ion remains directly fore,
= 2.06 x 104 x 1.00 x 1.12 X 10-5
proportional to its molar concentration; the
[In-] ~ (1.02 x 10-4)/2 = 5.10 x 10-5
same is true for the chromate ions. This fact is .+ 6.3 x 10 2 X 1.00 x 0.88 X 10-5 P'o = 10"'"
easily demonstrated by measuring the absorb- Similarly, in the strong acid [HIn] ~ 5.10 x P'
ance of chromium(VI) solutions in strongly 10-5. The molar absorptivities for the two = 0.236
acidic solution, where dichromate is the prin- species can thus be determined. At 430 nm, A570 .: 9.6 x 102 x 1.12 X 10-5
cipal species, and in strongly alkaline solu-
tion, where chromate predominat:s. Thus, ,A 1.051 + 7.12 X 103 x 0.88 X 10-5
deviations in the absorbance of this system £in= be = 1.00 x 5.10 X 10-5
=0.073
from Beer's law are more apparent than real = 2.06 X 104
because they result from shifts in chemical The accompanying data were obtained simi-
equilibria. These deviations can, in fact, be , _ 0.032 _ 2 larly.
readily predicted from the equilibrium con- tHin- 5.10 X 10-5 - 6.3 x 10 [lIt-] A430 A570
FH1n [HIn]
stants for the reactions and the molar absorp-
and at 570 nm, 2.00 X 10-5 0.88 X 10-5 1.12 X 10-5 0.236 0.073
tivities of the dichromate and chromate ions. 10-5 0.175
4.00 x 10-5 2.22 X 10-5 1.78 X 0.381
M 0.049 ,,2 8.00 x 10-5 5.27 X 10-5 2.73 X 10-5 0.596 0.401
EXAMPLE £k,= 5.10 X 10-5 = 9.6 x hr
12.00 x 10-5 10-5 10-5 0.771 0.640
Two solutions of the acid-base indicator
HIn were prepared by diluting 1.02 x 1O-4F M 0.363
_ _ 3
16.00 x 10-5
8.52 X
11.9 X 10-5
3.48 X
4.11 * 10-5 0.922 0.887

£Hln- 5.10 X 10-5 - 7.12 x 10


solutions of HIn (K. = 1.42 x 10-5) with
equal volumes of: (a) O.2F NaOH; and (b) Figure 6-3 is a plot of the data derived in
Let us calculate the concentration of HIn the foregoing example and illustrates the kind
O.2F HCI. When measured in a 1.0Q.an cel~
and In- in an unbuffered 2.00 x 1O-5F of departures from Beer's law that arises
the following absorbance' data were obtained
solution of HIn. From the equation for the when the absorbing system is capable of
at 430 and 570 nm:
dissociation reaction, we see that undergoing dissociation or association. Note
SohIrioB Dilutell With that the direction of curvature is opposite for
the two wavelengths.
0.2FNaOH 1.051 0.049 Furthermore, Instrumental DeYiations with Polychromatic ~ 0.600
O.2F HCI 0.032 0.363 Radiation. Strict adherence to Beer's law ?!
[In-] + [HIn] = 2.00 x 10-5 1:
is observed only when the radiation employed j
Derive absorbance data at the two wavelengths Substitution of these relationships into the is truly monochromatic; this observation is <l: 0.400
for unbuffered solutions having indicator con- expression for K. gives yet another manifestation of the limiting
centrations in the range of 2 x 10-5 to [In-]2 _ -5 character of the law. Unfortunately, the use of
16 X 10-5 F. Plot the data. 2.00 X 10-5 _ [In-] - 1.42 x 10 radiation that is restricted to a single wave-
The absorbance of the various solutions length is seldom practical; devices that isolate
depends upon the equilibrium Rearrangement yields the quadratic expression portions of the output from a continuous
[In -]2 + 1.42 x 10- 5[In-] source produce a more or less symmetric
HIN~H+ + In- band of wavelengths around the desired one
i
O.QOO
0.00 4.00 8.00 12.00
- 2.84 X 10-10 = 0 (see Figures 5-14 and 5-17, for example). Indicator conCentration. F X 105
which gives The following derivation shows the effect
K = 1.42 X 10-5 = [H+][In-] [In-] = 1.12 x 10-5 of polychromatic radiation on Beer's law. FIGURE 6-3 Chemical deviations from
• [HI~ Beer's law for unbuffered solutions of the
Consider a beam comprising of just two
[HIn] = 2.00 x 10-5 - 1.12 X 10-5 wavelengths A' and AM. Assuming that Beer's indicator HIn. For data, see example on this
Substitution of the hydrogen ion concen-
law applies strictly for each of these individ- page.
tration of the strong acid or base solutions = 0.88 X 10-5
When an absorbance measurement is made It is an txperimental fact that deviations tration for various levels of p. as compared to !!.. x 100
with radiation composed of both wave- from Beer's law resulting from the use of a Po'
lengths, the power of the beam emerging from Po·
polychromatic beam are not appreciable, Note that the instrumental deviations il-
the solution is given by (P' + P") and that of provided the radiation used does not en- lustrated in Figures 6-4 and 6-5 result in
the beam from the solvent by (PO + PO). compass a spectral region in which the absorb- absorbanees that are smaller than theoretical.
Therefore, the measured absorbance is er exhibits large changes in absorbance as a It can be shown that instrumental deviations
function of wavelength. This observation is always lead to negative absorbance errors.4
A _ I (P'o + PO)
101 - og (P' + P") illustrated in Figure 6-4.
Instrumental Deviations in the Presence of "~
which· can be rewritten as
.e 1.0
Stray Radiation. We have noted earlier
(Chapter 5) that the radiation exiting from a
B
«
The Effect of
monochromator is ordinarily contaminated Instrumental Noise on the Precision
with small amounts of scattered or stray ra- of Spectrophotometric Analy••• S
diation which reaches the exit slit owing to
AM = 10g(P o + PO) reflections from various internal surfaces. The accuracy and precision of spectropho-
tometric analyses are often limited by the un-
- 10g(P'olO-"" + POIO-'-") Stray radiation often differs greatly in wave-
certainties or noise associated with the
length from that of the principal radiation 2.5 5.0 7.5
Now, when t = t', this equation simplifies to and, in addition, may' not have passed instrument. A general discussion of in- Concentrotion. M X 103
through the sample. strumental noise and signal-to-noise optimi-
AM = e'bc FIGURE &-5 Apparent deviation from
When measurements are made in the zation is found in Chapter 3; the reader may
and ~r's law is followed. If the two molar find it helpful to review this material before Beer's law brought about by various amounts
presence of stray radiation, the observed
absofP;livities differ, however, the relationship absorbance is given by undertaking a detailed study of the effect of of stray radiation.
betw~n AM and concentration will no longer instrumental noise on the precision of spec-
be linear; moreover, greater departures from Po + p. trophotometric measurements. In addition, a
linearity can be expected with increasing dif- A' = Iog--- review of the contents of Appendix 1 on the
P+P. uncertainty in concentration can be derived
ferences between e' and t'. This derivation use of standard deviation as a measure of
can be expanded to include additional wave- where p. is the power of the stray radiation. precision may also prove worthwhile. by writing Beer's law in the form
lengths; the relationships remain the ~e. Figure 6-5 shows a plot of A' versus concen- As was pointed out earlier, a spectro- 1 -0.434
photometric measurement entails three steps: C = -£b log T= --e;;-In T
a 0% T adjustment, a 100% T adjustment,
and a measurement of % T with the sample The partial derivative of this equation, hold-
in the radiation path. The noise associated ing band c constant, is given by
with each of these steps combines to give a
net uncertainty for the final value obtained oc = -0.434 iJT
for T. The relationship between the noise en- rbT
countered in the measurement of T and the where oc is the variation in c that results from
the noise or uncertainty oT in T. Dividing by
the first equation yields

~ = 0.434 x oT (6-12)
• I. M. Kolthoff and P. J. Elving. eds, Treatise on Anal.l·t- c 10gT T
ical Chemistry. New York: Interscience Publishe ••• 1964.
Part I. vol. S, pp. 2767-2773. Here, oT/T is the net relative uncertainty or
FIGURE &-4 The effect of polychromatic radiation upon • See: L. D. Rothman, S. R. Crouch, and J. D. Ingle, Jr. noise in T that arises from the three measure-
the Beer's law relationship, Band A shows little deviation Anal. Chem. 47. 1226 (197S); J. D. Ingle, Jr. and S. R. ment steps and oc/c is the resulting relative
since e does not change greatly throughout the band. Band B Crouch, Anal. Chem., 44, 137S (1972); J. O. Erickson and
uncertainty in concentration.
T. Surles, American Laboratory, g (6~ 41 (1976); OplilflJJm
shows marked deviations since e undergoes significant changes The best and most useful measure of the
Parameters for Spectrophotometry. Varian Instrument
in this region.
Division, Palo Alto, CA. noise oT is the standard deviation ST, which
for a series of replicate measurements is given creases or decreases the average number of lation for 20 or 30 independent transmittance
by (see Appendix 1Y' 6.0
electrons moving in a given direction in the measurements oTan absorbing solution.

ST-
_ aT - f'i]••N-l
-...;-
(7; - T)2 (6-13)
5.0
resistor element. Johnson noise does not
depend on the magnitude of the current;
Column 3 of Table 6-3 shows the relative
error in concentration that would be en-
~ indeed, it occurs even in the' absence of a count~red with a spectrophotometer for
In this expression, 7; represents the individual x current. which 'the performance was limited by John-
values of T obtained for a set of N replicate .:~4.0 Johnson noise exists in all transducers and son or 0% T noise. Here. an absolute uncer-
measurements, and t is the average or mean 1/1
electronic comPonents of photometers and tainty ST in transmittance of ±0.OO3, or
for the set. From Equation 6-12, the relative g'•. 3.0 spectrophotometers. With one important ±OJ% T, was assumed. These data are
standard deviation in terms of concentration •... exception. however, its magnitude is so small plotted in Figure 6-6, curve A, where it is seen
(s.Ie) is then i 2.0 relative to other sources of noise that for all that the relative standard deviation of c reaches
II:
practical purposes it can be neglected. The a minimum at an absorbance of about 0.4. (By
~ = 0.434 x ~ (6-14) exception is for the thermal detectors for setting the derivative of Equation 6-14 equal
c 10gT T 1.0
infrared radiation. With such detectors, John- to zero, it can be shown that this minimum
Experimentally. ST can be evaluated by son noise may predominate and limit the occurs at a transmittance of 0.368 or an
C
malting, say. 20 replicate measurements precision of spectral measurements. This situ- absorbance of 0.434.)
1.0 2,0 3.0
(N •• 20) of the transmittance of a solution in Absorbonee ation is in distinct contrast to the ultraviolet The standard deviation in transmittance
exactly the same way and substituting the and visible regions. where Johnson noise in measurements for most commercial spectro-
data into Equation 6-1l FIGURE 6-6 Error curves (or various in- phototubes. photomultiplier tubes. and photometers will range from ± 0.000 1 to
It is clear from an examination of Equa- strumental uncertainties. Data from Table 6-3, amplifiers is inconsequential. ±O.OI (T = 0.01 to 1%); the lower figure is
tion 6-14 that the uncertainty in a photomet- Curve A: 0 % T and Johnson noise. Curve Elects or 0"10T and Johnson Noise. As we found in high-quality ultraviolet-visible in-
ric concentration measurement varies in a B: Signal shot noise. Curve C: Flicker noise. have noted. the magnitude of Johnson noise struments, whereas the upper figure might be
complex way with the magnitude of the trans- and 0% T noise is independent of the trans- expected in a low-cost infrared instrument.
mittance, The situation is even more com- mittance. Thus. ST in Equation 6-14 is The figure of ±0.OO3 in Table 6-3 is perhaps
plicated than is suggested by Equation 6-14. method thus requires an understanding of the independent of T and can be assumed to be typical for medium quality O/'o T or Johnson
however. because the uncertainty ST is, under likely type (or types) that predominate in a constant in determining the effect of ST on s•. noise-limited instruments; concentration er-
many circumstances, also dependent upon T. given set of circumstances, For a given instrument, a good estimate of ST rors of about 1 to 2% relative are thus in-
In a detailed theoretical and experimental 0"10 T Uacertaiaties. As shown in Table can be derived by a standard deviation calcu- herent when measured absorbanees lie in the
study, Rothman. Crouch and Ingle have 6-2, 0% T noise includes three types of noise
described several sources of instrumental un- associated with transducers and amplifiers; as
certainties and shown their net effect on the the name implies all are associated with the
precision of absorbance or transmittance uncertainties in making the 0% T adjustment.
measurements. These uncertainties fall into Sources would include light leaks around the
three categories: those for which the magni- shutter that screens the radiation from TABLE 6-2 SOURCES OF UNCERTAINTY IN A SPECTROPHOTOMETRIC
tude of ST is~portional to T. (2) propor- the detector, vibrational effects associated MEASUREMENT
tional to .jT2 + T. and (3) independent of T. with movement of the shutter. and tempera-
Table 6-2 summarizes information about ture fluctuations in the transducer and
ST Independent of T sT=k'JT!+T ST = k"T

these sources of uncertainty. Clearly, each amplifier systems. In modem spectropho- 0% TNoise Detector shot noise Source nicker noise
type of source has a different effect upon the tometers, 0% T noise is seldom important (photon detectors only)
magnitude of the concentration error. An except at the wavelength extremes of the Dark current shot noise
Dark current excess noise
error analysis for a given spectrophotometric instrument. where K' in Equation 6-S may
Amplifier excess noise
approach KP and KPo. Note that 0% T
noise is independent of the magnitude of T. Detector Johnson noise
JoblBOll Noise. As noted in Chapter 3
(thermal detectors only)
• For a discussion of the standard deviation as a measu",
of experimental uncertainties, see: D. A. Skoog and D. M. (p, 71). Johnson noise results from thermally Limited readout resolution
Wesl, FundamentaLs of Analytical Chemistry, 3d ed. New induced motion of electrons in resistive cir- Amplifier shot noise
York: Holl, Rinehart and Winston, 1976, Chapter 4. cuit elements. This motion momentarily in-
range of 0.15 to 1.0. The higher error ass0- thus, its effect on the relative error in concen- around a mean value. Here, the magnitude of Ordinarily, Po will not change greatly during
ciated with more strongly absorbing samples tration will be similar to 0% T and Johnson the current fluctuationS is proportional to the a series of measurements of T. When this situ-
can usually be avoided by suitable dilution noise and will be described by Equation 6-14. square root of current (see Equation 3-24, ation prevails, we may write
prior to measurement. Relative errors greater Substitution of 0.005 into this equation p. 71~
than 2"/0 are inevitable, however, when absorb- reveals that such an instrument will yield con- In contrast to the types of ·noise we &ave
ST= Tk'J~+ 1 = k'~ (6-15)
ances are less than 0.1. centration data that are uncertain to the thus far considered, shot noise increasei in
It is important to emphasize that the con- extent of 1.5 to 3 % relative, provided percent magnitude with increases in the size of the
clusions just described are based on the transmittances are limited to a range of 10 to quantity being measured. As a consequence, where k' = k..jP;. Substitution of Equation
assumption that the limiting uncertainty in an 80%. ST in Equation 6-14 varies as a function of T. 6-15 into 6-14 yields
analysis is a constant indeterminate error ST
that is independent of transmittance; this type
Signal Shot Noise. As was pointed out ear-
lier (p. 71~ shot noise must be expected
The effect of shot noise on ST and S. is readily
derived, as shown in the following example. ~=
clog
0.434 x k'~
T T
= 0.434k' J.!. +
log T ~ T
1
of uncertainty is most commonly encountered whenever the passage of electricity involves
in instruments employing thermal detectors. transfer of charge across a junction, such as EXAMPLE : (6-16)
Limited Readout Resolution. The precision the movement of electrons from the cathode Derive an expression relating ST and s. to
of some commercial spectrophotometers is shot noise. The data in column 4 of Table 6-3 were
to the anode of a photomultiplier tube or
limited by the resolution of their readout de- across the junction of a barrier-type pho- Shot noise is associated with each of the obtained with the aid of Equati~n 6-16.
vices. For example, an instrument with a meter tocell Here, the current results from a series of three steps required for measuring transmit- Figure 6-6, curve B is a plot of such data. Note
readable to ±O.s% full scale will have an un- discrete events (emission of electrons from a tance or absorbance. Generally, however, the the much larger range of absorbances that can
certainty (ST) of ±0.005T due to this source. cathode); the number of these events per noise associated with the 0% T adjustment is be encompassed without serious loss of ac-
This uncertainty is also independent of T; unit time is distributed in a random way small relative to the measurement of P and curacy when shot noise, rather than Johnson or
Po because the magnitude of the noise in- 0% T noise, limits the precision. This increased
creases as the square root of the signal range represents a major advantage 4f photon
(p. 71). Thus, the uncertainty in T is gen- type detectors over thermal types.· As with
erally determined largely by noise associated Johnson or O'i~T noise-limited instruments,
TABLE &-3 RELATIVE CONCENTRATION ERROR AS A FUNCTION OF with the measurement of P and Po. Since T is shot-limited instruments do not give very re-
TRANSMITTANCE AND ABSORBANCE FOR VARIOUS SOURCES OF the quotient of these two numbers, its relative liable concentration data at transmittances
UNCERTAINTY standard deviation can be obtained from the greater than 95% (or A < 0.02).
relative standard deviation of P and Po Signal shot noise sometimes limits the
Relative Error, ~ x 100 (Appendix I). Thus, accuracy of transmittance and absorbance
c measurements with ultraviolet and visible
For Measurements Limited by: . photometers and spectrophotome1ers. It is
Absorbance, usually of little importance with infrared
Transmittance, 0% T or Signal Flicker
T A Johnson Noise" Shot Noise" Noise' But shot noise is proportional to the square instruments.
root of the measured quantity (p. 71). Thus, Flicker Noise. Flicker noise in photom-
0.95 0.022 ± 6.2 ±8.4 ±5.8 eters and spectrophotometers is largely as-
0.90 0.046 ± 3.2 ±4.1 ±2.8 s,.=kjP sociated with the radiation source. As was
0.80 0.097 ± 1.7 ±2.0 ±1.3
0.60 0.222 ± 0.98 ±0.96 ±0.s9 where k is a proportionality constant. Also, pointed out in Chapter 3, the magnitude of
0.40 0.398 ± 0.82 ±0.61 ±0.33 flicker noise cannot be predicted frotn theory.
0.20 0.699 ± 0.93 ±0.46 ±0.18
s,.o=k~ It has been shown. however.7 that uncertain-
0.10 1.00 ±1.3 ±0.43 ±0.13 ties from the source can be described by the
Substitution of these quantities into the first
0.032 1.50 ± 2.7 ±0.50 ±0.09 equation and rearrangement gives relationship ,
0.010 2.00 ± 6.5 ±0.65 ±0.06
0.0032
0.0010
2.50
3.00
±16.3
±43.4
±0.92
±1.4
±0.05
±0.04 (~)2
T
('!' +~)
= k2
P Po
2
= k (Po
Po P
+ I) _
• From Equation 6-14 employing IT - ±O.OO3.
• From Equation 6-16 employing k' - ±O.OO3. 7 See: L. D. Rothman. S. R. Crouch. and J. D. Ingle, Jr.
, From Equation 6-18 employing k- - ±O.OO3. Anal. Ch""" 47. 1226 (197S~
and Equation 6-14 becomes ent sections of the cell are exposed to the resulted from this source. Indeed, it seems of random error is probably cell-positioning
source; corresponding variations in measured probable that imprecision in cell positioning uncertainty. It is evident· that the fullest
~ = _ 0.434 k N (6-18)
C 10gT transmittances must then be expected. For is often the most important limitation to the potential of such instruments can only be
high-quality cells, such imperfections will be accuracy of spectrophotometric measure- realizea by filling and rinsing the cells in
As shown by the data in column 5 of Table minimal. If cells become scratched or dirty, ments. One method of reducing the effect of place; ·even with this precaution, cell uncer-
6-3 and by Figure 6-6, curve C, the concentra- however, the dependence of transmittance cell positioning is to leave cells in place tainty often appears to limit the performance
tion error due to flicker noise decreases rapidly upon cell position may become significant. during calibration and analysis; new stand- of these instruments.
and approaches zero at high absorbances. Rothman, Crouch, and Ingle have shown ards and samples are introduced after wash- The performance of inexpensive spectro-
The overall uncertainty in a spectro- that uncertainties due to cell positioning var- ing and rinsing the cell in place with a photometers tends to be limited, at least in
photometric measurement at low absorbance iations behave in the same way as source syringe. Care must be taken to avoid touch- part, by their readout devices; it has been
frequently results from a combination of flicker noise; that is, ST is proportional to ing or jarring the cells during this process. shown that significant improvements can be
flicker noise and the other types mentioned transmittance and the concentration uncer- Stmmary. It is apparent that the random achieved by replacing these components with
earlier; occasionally, all three may contribute. tainty is given by Equation 6-18. Figure 6-7A errors associated with a spectrophotometric more sensitive ones.8
CeO Positioning Uncertainty. All cells have is a plot of some of their data. Even though analysis depend upon several variables, in-
minor imperfections. As a consequence, painstaking care was followed in cleaning cluding instrument design, the wavelength
reflection and scattering losses vary as differ- and positioning the cells, significant errors region, the source intensity and stability, the Effect of Slit Width
sensitivity of the transducer, the slit widths, on Abaorptlon M ••••• nments
and the concentration of the analyte. The sit-
The ability of a spectrophotometer to resolve
uation is made even more complicated with a pair of adjacent absorption peaks depends
5.0
recording instruments, where several of these
upon the slit width employed (p. 133). Narrow
/8 variables may change as the spectrum is
slits are required to obtain the maximum
I recorded.
I detail from a complex spectrum.9 Figure 6-8
Infrared instruments are most commonly
4.0 I illustrates the loss of spectral detail that ac-
limited by the Johnson noise associated with
I companies the use of wider slits. In this
the detector and by the cell position uncer-
I example, the transmittance spectrum of a
~ I
tainty. The latter is due to the fact that
didymium glass was obtained at slit settings
x 3.0 narrow cells must often be used and the width
~':lu I that provided bandwidths of 0.5, 9, and 20

-
~.
, /
I
I
of the windows, and thus the relative sample
thickness undoubtedly fluctuates considerably
from one part of the cell to another. Further-
om. The progressive loss of spectral detail is
clear. For qualitative studies, such losses
often loom important.
..
.~ 2.0
/ more, infrared windows are readily scratched,
Figure 6-9 illustrates another elTect of slit
11
a:
'---_/ abraded, and attacked by the atmosphere;
width on spectra. Here, the spectrum of a
transmission is thus highly variable across
praseodymium chloride solution was ob-
their surfaces. Source flicker may also con- tained at slit settings of 1.0, 0.5, and 0.1 mm.
tribute to the uncertainty of infrared measure-
Note that the peak absorbance values. in-
ments, but is probably less important than
crease significantly (by as much as 70% in
Johnson noise and cell uncertainty. Figure
one instance) as the slit width decreases. At
6-7, curve B is an error plot assuming the
slit settings less than about 0.14 mm, absorb-
same cell positioning uncertainty as in Figure
ances were found to become independent of
6-7, curve A and a Johnson noise uncertainty
FIGURE 6-7 Curve A: Uncertainty due to un- equal to that in Figure 6-6, curve A. For most
certainties in cell positioning. (Data adapted. Re- infrared measurements, the errors would prob-
printed with permission from L. D. Rothman, S. R. • For example, see: J. D. Inlle. Jr~ AII/IL Clam. Actll, •••
ably be larger than those indicated by the plot. 131 (1977~
Crouch, and J. D. Ingle, Jr .• AruJL Chern., 47, 1231 Considerably greater variation in behavior • For • discussion ol the ell"ectsof slit width on .pec:t•.••
(1975~ Copyright by the American Chemical is encountered among ultraviolet and visible see: OptimumSp«trophotometer P-en. Applictllion
Society.) Curve B: Combined effect of cell position- photometers and spectrophotometers. For the Report AR •.•.•1, Monrovia, CA: Cary Instruments; and
ing uncertainty and Johnson noise. best quality instruments, the limiting source F. C. Strons III, AnDl. CMm.,48, 2155 (1976~
100 slit width. Careful inspection of Figure 6-8
90 reveals the same type of effect. In both sets of
spectra, the areas under the individual peaks
80
are the same, but wide slit widths result in
i 70 broader, lower peaks.
';60
~
~ 60

i40
It is evident from both of these illustra-
tions that quantitative measurement of
narrow absorption bands demands the use of
1
<
04
.
0.3
narrow slit widths or, alternatively, very re-
.f3O
producible slit-width settings.
Bandwidth· 0.5 nm
20 Unfortunately, a decrease in slit width is
accompanied by a second-order power reduc-
10
tion in the e~itted radiant energy; at very
100
narrow settings, spectral detail may be lost
90 owing to an increase in the signal-to-noise
4600 6000 4200 4600 6000 4200 4600
80 ratio. The situation becomes particularly ser·
WIY8Iength. It. WIYeI8ngth. It. WIY8Iength. It.
ious in spectnU regions where the output of
I 70 the source or the sensitivity of the detector is FIGURE 8-9. Effect of slit width and bandwidth on peak heights. Here, the sample was
solution of praseodymium chloride. (From Optimum Spectrophotometer Parameters, Applica-
a
"'60 low. Under such circumstances, noise in
either of these components or their associated tion Report AR 14-2, Monrovia, California. Cary Instruments. With permission.)
160
electronic cinoUits may result in partial or
i40
total loss of spectral fine structure.
.•.~ 30 In general, lit is good practice to narrow
20

10
slits no more than is necessary for resolution
of the spectrum at hand. With a variable slit
1.00
,
spectrophotometer, proper slit adjustment \ 8
100
can be determined by obtaining spectra at \
90 progressively narrower slits until peak heights- 0.75 \
80
become constant. \

~ 70
Effect of Scattered ..
.Ii 60
Radiation at Wavelength .e~ 0.50
!i 60

40
Extremes of a Spectrophotometer ~
<
We have already noted that scattered radia-
.•.~ 30 Bandwidth· 20 nm tion may cause instrumental deviations from
0.25
20 Beer's law. When measurements are at-
10
tempted at the wavelength extremes of an
instrument, the effects of stray radiation may
0
300 360 400 460 500 550 600 650 be even more serious, and on occasion lead to
Wavelength, om the appearanee! of false absorption peaks. For 380
example, consider a spectrophotpmeter for Wavelength, nm
FIGURE 6-8 Effect of band width on the visible region equipped with glass optics, FIGURE 6-10 Spectra of cerium (IV)
spectral detail. The sample was a didymium a tungsten source, and a photocell detector. obtained with a spectrophotometer
glass. [Spectra provided through the courtesy At wavelengths below about 380 nm, the win- having glass optics (A) and quartz optics
of Amsco Instrument Company (formerly dows, cells, and prism begin to absorb radia- (B). The false peak in A arises from
Turner Associates) Carpinteria, California.] tion, thus reducing the radiant energy transmission of stray radiation of longer
wavelengths.
reaching the transducer. The output of the the wavelength extremes of a visible-region (a) the absorbance of a 3.77 x 10-4M solution of the complex at 580
source falls off rapidly in this region as well; spectrophotometer is shown in Figure 6-10. nm when measured in a 0.75O-an ceIL
so also does the sensitivity of the photo- The spectrum of a solution of cerium(IV) ob- (b) the transmittance of a 2.85 x 1O-4M solution employing the SlPIle
electric device. Thus, the radiant power for tained with an ultraviolet-visible spectro- cell as in (a).
the 100% T adjustment may be as low as 1 to photometer, sensitive in the region of 200 to (c) the absorbance (0.75O-cm cell) of a solution that has half the trans-
2% of that in the region between SOOand 650 750 DID, is shown by curve B. Curve A is a mittance of that described in (a~
nm. spectrum of the same solution obtained with 5. A solution containing the thiourea complex of bismuth(III) has a
The scattered radiation, however, is often a simple visible spectrophotometer. The ap- molar absorptivity of 9.3 x 103 liter em-I mol-I at 470 nm.
made up of wavelengths to which the instru- parent maximum shown in curve A arises (a) What will be the absorbance of a 5.0 x lO-sM solution of the
ment is highly sensitive. Thus, its effects can from the instrument responding to stray complex when measured in a 0.5OO-cmcell?
be enormously magnified. Indeed, in some in- wavelengths longer than 400 nm, which (as (b) What will be the percent transmittance of the solution described in
stances the output signal: produced by the can be seen from the spectra) are not ab- (a)?
stray radiation may exceed that from the sorbed by the cerium(IV) ions. (c) What concentration of bismuth could be determined by means of
monochromator beam; uDder these circum- This same effect is sometimes observed this complex if 1.00-cm cells are to be used and the absorbance is
stances, the measured absorbance is as much with ultraviolet-visible instruments when to be kept between 0.100 and 1.500?
that for the stray radiatioq as for the radia- attempts are made to measure absorbances at 6. In ethanol, acetone has a molar absorptivity of 2.75 x 103 liter em - I
tion to which the instrument is set. wavelengths lower than about 200 nm. mol-1 at 366 nm. What range of acetone concentrations could· be
An example of a false peak appearing at determined if the percent transmittances of the solutions are to be
limited to a range of 10 to 90% and a 1.().cm cell is to be used?
7. In neutral aqueous solution, it is found that log t for phenol at 211 nm
1. Use the data provided to evaluate the missing quantities. Wherever
is 3.79. What range of phenol concentration could be determined spec-
n~ry, assume that the molecular weight of the absorbing species is
trophotometrically if absorbance values in a 2.00-cm cell are to be
2~0.
limited to a range of 0.100 to 1.50?
8. The equilibrium constant for the reaction
A %T t b, em c,M C

2CrO~- + 2W ~ Cr10~- + H10


(a) 0.2% 2.50 4.41 x 10- 4 mg,lliter
(b) 19.6 -1.50 6.91 x 10-3 ppm 14
has a value of 4.2 x 10 . The molar absorptivities for the two princi-
(e) 0.877 0.500 g,IIiter pal species in a solution of K1Cr10, are
(d) 84.2 7.85 x 102 3.00 X 10-4 II1II100 ml
(e) US x 104 1.25 3.33 ppm
(f) 6.42 x 103 1.10 7.84 X 10-5 mg,lliter
(g) 43.2 0.100 4.11 ppm
(h) 1.025 2.86 x 103 1.00 g,IIiter 1.84 X 103 10.7 X 101
(i) 87.2 0.980 mg,lliter 4.81 X 103 7.28 X 101
(j) 6.74 9.16 x 102 1.86 X 10-4 mg,lliter 1.88 X 103 1.89 X 102

2. A,solution that was 1.54 x 1O-4M in X had a transmittance of 0.0874 Four solutions were prepared by dissolving the following number of
when measured in a 2.00-cm cell. What concentration of X would be moles of K1Cr107 in water and diluting to 1.00 liter with a pH 5.40
required for the transmittance to be increased by a factor of 3 when a buffer: 4.00 x 10-4,3.00 X 10-4,2.00 X 10-4, and 1.00 x 10-4• Derive
1.@O-cmcell was used? theoretical absorbance values (1.00-cm cells) for each and plot the data
3. A 'compound had a molar absorptivity of 6.74 x 103 liter em-I mol-I. for (a) 345 nm, (b) 370 nm, (c) 400 nm.
What concentration of the compound would be required to produce a 9. A species Y has a molar absorptivity of 2400. Derive absorbance data
solution having a transmittance of 7.77% in a 2.5O-an cell? (1.QO-cm cells) for solutions of Y that are 10 x 10-4, 6.00 X 10'4,
4. At 580 nm, the wavelength of its maximum absorption, the complex 3.00 X 10-4, and 1.00 x 1O-4M in Y, assuming that the radiation
FeSCN2+ has a molar absorptivity of 7.00 x 103 liter em-I mol-I. employed was contaminated with the following percent of nonad-
Calculate sorbed radiation: (a) 0.000; (b) 0.300; (c) 2.00; (d) 6.00. Plot the data.
Absorption measurements employing ultra- more, the amount of thermal energy evolved
violet or visible radiation find widespread is usually not detectable. Thus, absorption
application to quantitative and qualitative measurements have the advantage of creating
analysis. Some of these applications are con- minimal disturbance of the system under
sidered in this chapter} study (except when photochemical de-
composition occurs ~
The absorption of ultraviolet or visible
radiation generally results from excitation of
bonding electrons; as a consequence, the
The absorPtion of ultraviolet or visible radia- wavelengths of absorption peaks can be cor-
tion by some species M can be considered to related with the types of bonds that exist in
be a two-siep process, the first of which in- the species under study. Absorption spectros-
volves excitation as shown by the equation copy is, therefore, valuable for identifying
functional groups in a molecule; it also pro-
M + hv --+ M· vides a somewhat selective method of quanti-
where M· represents the atomic or molecular tative analysis for compounds containing
particle in the electronically excited state re- absorbing bonds.
sulting from absorption of the photon h,•.The For purposes of this discussion, it is useful
lifetime of the excited state is brief (10-1 to to recognize three types of electronic transi-
10-9 s~ its existence being terminated by any tions and to categorize absorbing species on
of several relaxation processes (p. 1(9). The this basis. The three include transitions,
most common type of relaxation involves involving: (1) n, CT, and n electrons, (2) d and!
conversion! of the excitation energy to heat; electrons, and (3) charge-transfer electrons.
that is,
(M· ...•M + heat Absorbing Species
Containing n, CT, and n Electrons
Relaxation may also occur by decomposition
of M· to form new species; such a proceSs is Absorbing species of this type include organic
called a photochemical reaction. Alternatively, molecules and ions as well as a number of
relaxation may involve fluorescent or phos- inorganic anions, Our discussion will deal
phorescent reemission of radiation. It is im- largely with the former. although brief men-
portant to note that the lifetime of M* is tion will be made of absorption by certain
usually so very short that its concentration at inorganic systems as well.
any instant is ordinarily negligible. Further- All organic compounds are capable of
absorbing electromagnetic radiation because
all contain valence electrons that can be
excited to higher energy levels. The excitation
I Some useful references on absorl'lion methods include: energies associated with electrons forming
R. P. Baum.tn. Ahsorption Spectroscop)". New York: most single bonds are sufficiently high that
Wiley. 1962: J. R. Edisbury. Practical Hints on Ahsorp- absorption by them is restricted to the so-
tion Spectr0rlletr}". Nc"W York: Plenum Press. I96X: F.
Grum. Physisal Methods of Chemistry. eds. A. Wcissber-
called vacuum ultraviolet region (A. < 185 nm),
ger and B. W. Rossiter. New York:. Wiley-Interscience, where components of the atmosphere also
1972, Volume I, Part 111 B. Chapter 3; H. H. Jaffe and absorb strongly. The experimental difficulties
M. Orchin, TMory and Applictltions of Vltrariole! associated with the vacuum ultraviolet are
Spectroscopy. New York: Wiley, t962; G. F. Lothian, formidable; as a result, most spectrophoto-
Absorplion SpectropMt"""'try, 3d cd. London: Adam
metric investigations of organic compounds
Hilger Ltd, 1969; and J. E. Crooks, The Sptrtrum in
Ch •••••istry. London: Academic Press. 1978.
have involved the wavelength region greater
than 185 om. Absorption oflonger-wavelength
ultraviolet and visible radiation is restricted to
molecule occupy th~ former in the ground
state.
In addition to a and Jr electrons, many
organic compounds contain nonbonding elec-
~.
·c xx
• ••
00
00
0
.-a
x_.
a limited number of functional groups (called The molecular orbitals associated with
trons. These unshared electrons are designated 1-. O-n
chromophores) that contain valena: electrons . single bonds in organic molecules are desig- by the symbol n. An example showing the three
with relatively low excitation energies. . nated as sigma (0-) orbitals, and the' cor- FIGURE 7-2 Types of
types of electrons in a simple organic molecule
The electronic spectra of organic molecules responding electrons are a electrons: As molecular orbitals in form-
is shown in Figure 7-2.
containing chromophores are usually complex, sbown in Figure 7-1a, the distribution of aldehyde.
As shown in Figure 7-3, the energies for
because the superposition of vibrational transi- charge density of a 'sigma orbital is rota- the various types of molecular orbitals differ
tions on the electronic transitions leads to an tionally symmetric around the axis. of the significantly. Quite .generally, the energy level
intricate combination of overlapping lines; the bond. Here, the average negative charge den- of a non bonding electron lies between those
result is a broad band of continuous absorp- sity arising from the motion of tbe two elec- of the bonding and the antibonding Jr and a and can be brought about by radiation in the
tion. The complex nature of the spectra makes trons around the two positive nuclei is orbitals. region of between 150 and 250 nm, with most
detailed theoretical analysis difficult or im- indicated by the degree of shading. Electronic transitions among certain of the absorption peaks appearing below 200 nm.
possible. Nevertheless, qualitative or semi- The double bond in an organic molecule energy levels can be brought about by the Table 7-1 shows absorption data for some
quantitative statements cona:ming the types rontains two types of molecular orbitals: a absorption of radiation. As shown in Figure typical n -+ cr* transitions. It will be seen that
of electronic transitions responsible for a given sigma (a) orbital corresponding to one pair of 7-3, there are four: of these: a ....•a*, n -+ a*, the energy requirements for such transitions
absorption spectrum can be deduced from tbe bonding electrons; and a pi (Jr) molecular n ....•Jr*, and It -+ Jr•. depend primarily upon the kind of ~tomic
molecular orbital considerations. orbital associated with the other pair. Pi orbi- a ....•cr* Transitions. Here, an electron in a bond and to a lesser extent upon the structure
Types of AlIIorbing Electroas.1 The elec- tals are formed by the parallel overlap of
bonding a orbital of a molecule is excited to of the molecule. The molar absorptivities (e)
trons that contribute to absorption by an atomic p orbitals. Their charge distribution is associated with this type of absorption are
the corresponding antibonding orbital by the
organic molecule are: (1) tbose that partici- characterized by a nodal plane (a region of
absorption of radia~ion. The molecule is then intermediate in magnitude and usually range
pate directly in bond formation between low-charge density) along the axis of the
described as being Iin the a,a· excited state. between 100 and 3000 liter em -I mol- I.
atoms and are thus associated witb more than bond and a maximum density in regions
Relative to other; possible transitions, the Absorption maxima for formation of the
one atom; (2) nonbonding or unshared outer above and below the plane (see Figure 7-1b).
energy required to induce a (T'....•(T. transition n,(T· state tend to shift to shorter wavelengths
electrons that are largely localized about such Also shown in Figure 7-1c and 7-1d are the in the presence of polar solvents such as
is large (see Figure 7-3), corresponding to
atoms as oxygen, the halogens, sulfur, and charge-density distributions for antibonding
radiant frequencies in the vacuum ultraviolet water or ethanol. The number of organic
nitrogen. sigma and pi orbitals; these orbitals are _functional groups with n ....•cr* peaks in the
region. Methane, for example, which contains
Covalent bonding occurs because the elec- designated by cr* and Jr•.
only single C-H bonds and can thus readily accessible ultraviolet region is rela-
trons forming the bond move in the field
undergo only a -+ a· transitions, exhibits an tively small.
about two atomic centers in such a manner as
absorption maximum at 125 nm. Ethane has

,/-a,.....
to minimize the repulsive coulombic forces
between these centers. The nonlocalized fields ~;,.",.........
,~.. an absorption peak at 135 nm, which must
also arise from the same type of transition,
between atoms that are occupied by bonding
electrons are called molecular orbitals and can ( but here, electrons of the C-C bond appear
to be involved. Because the strength of the
be considered to result from the overlap of r
C-C bond is less than that of the C-H
atomic orbitals. When two atomic orbitals
combine, either a low-energy bonding molecu-
bond, less energy is required for excitation;
thus, the absorptioJl peak occurs at a longer '0 .~ .~'0

lar orbital or a high-energy antibonding


molecular orbital results. The electrons of a
~

c:
. .
;>
.o wavelength.
Absorption maxima due to a"'" a* transi-
tions are never dbserved in the ordinary,
f
0
1t
~~ f
~

\J accessible ultraviolet region; for this reason,


they will not be considered further. .
n -+ a· Transitions. Saturated compounds
2 For funher details, see: C. N. R. Rao, Ultra-Viol~t and containing atoms with unshared electron
Visib/~ Spectroscopy, 3d ed. London: Butlerwonhs, 1975; pairs (nonbonding electrons) are capable of
and R. P. Bauman, Absorption Spectroscopy. New York: FIGURE 7-1 Electron distribution in FIGURE 7-3 Electronic molecular
n ....•a· transitions. In general, these transi-
Wiley, 1962. Chapters 6 and 8. sigma and pi molecular orbitals. energy levels.
tions require less energy than the a ....•a* type
/I •...• n· and n ....•

Transitions. Most alcohols, in which hydrogen-bond formation groups, since the positions of maxima are also absorptions of multichromophores in a single
applications of absorption spectroscopy to between the solvent protons and the non- affected by solvent and structural details. organic molecule are approximately additive,
organic compounds are based upon transi~ bonded electron pair is extensive. Here, the Furthermore, the peaks are ordinarily broad provided the chromophores are separated
tions for /I or n electrons to the n· excited energy of the n orbital is lowered by an because of vibrational effects; the precise from one another by more than one single
state, because the energies required for these amount approximately equal to the energy of determination of the position of a-maximum bond. Conjugation of chromophores, how-
processes bring the absorption peaks into an the hydrogen bond. When an n ....•7t. transi- is thus difficult. ever, has a profound effect·on spectral proper-
experimentally convenient spectral region tion occurs, however, the remaining single n Effect of Conjugation of Chromopbores. In ties. For example, it is seen in Table 7-3
(200 to 700 nm). Both transitions require the electron cannot maintain the hydrogen bond; the molecular-orbital treatment, n electrons that 1,3-butadiene, CH -eHCH -=CH2, has
presence of an unsaturated functional group thus, the energy ofthe n,n· excited state is not are considered to be further delocalized by a strong absorption band that is displaced
to provide the n orbitals. Strictly speaking, it affected by this type of solvent interaction. A the conjugation process; the orbitals thus in- to a longer wavelength by 20 nm as compared
is to these unsaturated absorbing centers that blue shift, also roughly corresponding to the volve four (or more) atomic centers. The with the corresponding peak for an uncon-
the term chromophore applies. energy of the hydrogen bond, is therefore effect of this delocalization is to lower the jugated diene. When three double bonds are
The molar absorptivities for peaks asso- observed. energy level of the n· orbital and give it less conjugated. the bathochromic effect is even
ciated with excitation to the n,n· state are A second solvent effect that undoubtedly antibonding character. Absorption maxima larger.
generally low and ordinarily range from 10 to influences both 7t •...• ~ and n ....•n· transitions are shifted to longer wavelengths as a Conjugation between the doubly bonded
100 liter em - 1 mol-I; values for n ....•n· tran- leads to a bathochromic shift with increased consequence. oxygen of aldehydes, ketones, and carboxylic
sitions, on the other hand, normally fall in the solvent polarity. This effect is small (usually . As seen from the data in Table 7-3, the acids and an olefinic double bond gives rise
range between 1000 and 10,000. Another less than 5 nm~ and as a result is completely
characteristic difference between the two overshadowed in n ....•n· transitions by the
types of absorption is the effect exerted by the hypsochromic effect just discussed. Here,
solvent on the wavelength of the peaks. Peaks attractive polarization forces between the sol- TABLE 7-2 ABSORPTION CHARACTERISTICS OF SOME COMMON
associated with n ....•7t. transitions are gen- vent and the absorber tend to lower the CHROMOPHORES
erally shifted to shorter wavelengths (a hyp- energy levels of both the unexcited and
sochromic or blue shift) with increasing the excited states. The effect on the excited Type of
polarity of the solvent. Usually, but not state is greater, however, and the energy dif- Chromophore Example Solvent ).••••(nm) e..... transition
always, the reverse trend (a bathochromic or ferences thus become smaller with increased Alkene C6HI3CH-eH2 n-Heptane 177 13,000 n ....•n·
red shift) is observed for n ....•n· transitions. solvent polarity; small bathochromic shifts CSH11CI!iEC-CH3 n ....•n·
Alkyne n-Heptane 178 10,000
The hypsochromic effect apparently arises result. 196 2000
from the increased solvation of the unbonded Orpnic CIIromophores. Table 7-2 lists 225 160
electron pair, which lowers the energy of the n common organic chromophores and the 0
orbital. The most dramatic effects of this kind approximate location of their absorption I
Qubonyl CH3CCH3 n-Hexane 186 1000 n ....•u·
(blue shifts of 30 nm or more) are seen with maxima. These data can serve only as rough n ....•n-
280 16
polar hydrolytic solvents, such as water or guides for the identification of functional
°
CH3CH
I
n-Hexane 180 large n ....•u·
293 12 n ....•n·
TABLE 7-1 SOME EXAMPLES OF ABSORPTION DUE TO
n ....•
u· TRANSITIONS- Carboxyl
°I
CH3COH Ethanol 204 41 /I •..•• n·

Compound
H2O
J.",•• (nm)
167
em••

1480
Compound J.",..( nm)
229
Em..

140 Amido
°•
CH3CNH2 Water 214 60 n -+ 7[*

CH30H 184 ~H3~2Sb Azo CH3N=NCH3 Ethanol 339 5 n ....•n·


150 CH320 184 2520
CH3Q 173 200 H3NH2 215 600 Nitro CH3N02 Isooctane 280 -22 n ....•n·
CH3I 258 365 (CH3hN 227 900 Nitroso C4H9NO Ethyl ether 300 100
665 20 n ....•n·
• Samples in vapor state. Nitrate C2HsON02 Dioxane 270 12 n ....•n·
• In ethanol solvent.
to similar behavior (see Table 7-3). Analo- Absorption by Aromatic Systems. The tional group that does not itself absorb in the carbonate (217 om), nitrite (360 and 280 om~
gous effects are also observed when two car- ultraviolet spectra of aromatic hydrocarbons ultraviolet region, but has the effect of shifting azido (230 nm), and trithiocarbonate (500 nm)
bonyl or carboxylate groups are conjugated are characterized by three sets of bands that chromophore peaks to longer wavelengths as ions. .
with one another. For «-p unsaturated originate from n ...•7t* transitions. For exam- well as increasing their intensities. It is seen in
aldehydes and ketones,· the weak absorption ple, benzene has a strong absorption peak at Table 7-4 that -oH and -NH2 have an Absorption
peak due to n ...• 1!* transitions is shifted to 184 om (e.- - 60,(00); a weaker band, called auxochromic effect on the benzene chromo- Involving d and f Electrons
longer wavelengths by 40 nm or more. In the E2 band, at 204 om (e.- = 79(0); and a phore, particularly with respect to the B Most transition-metal ions absorb in the
addition, a strong absorption peak corre- still weaker peak, termed the B band, at band. Auxochromic substituents have at least ultraviolet or visible region of the spectrum.
sponding to a II:.•.•1'. transition appears. This 256 nm (t",.. = 2(0). The long-wavelength one pair of n electrons capable of interacting For the lanthanide and actinide series, the
latter peak occurs only in the vacuum ultra- band of benzene, and many other aromatics, with the II: electrons of the ring. This interac- absorption process results from electronic
violet if the carbonyl group is not conjugated. contains a series Qf sharp peaks (see Figure tion apparently has the effect of stabilizing transitions of 4f and Sf electrons; for elements
The wavelengths of absorption peaks for 4-11b, p. 106) due to the superposition of the n· state, thereby lowering its energy; a of the first and second transition-metal series,
conjugated systems are sensitive to the types vibrational transitions upon the basic elec- bathochromic shift results. Note that the the 3d and 4d electrons are responsible for
of groups attached to the doubly bonded tronic transitions. Polar solvents tend to elim- auxochromic effect is more pronounced for absorption.
atoms. Various empirical rules have been inate this fine structure, as do certain types the phenolate anion than for phenol itself, Absorption by Lanthanide and Actinide
developed for predicting the effect of such of substitution. ; probably because the anion has an extra pair Ions. The ions of most lanthanide and acti-
substitutions upon absorption maxima and All three of the characteristic bands for of unshared electrons to contribute to the nide elements absorb in the ultraviolet and vis-
have proved useful for structural determina- benzene are strongly affected by ring substitu- interaction. With aniline, on the other hand, ible regions. In distinct contrast to the
tions.3 tion; the effects on the two longer-wavelength the nonbonding electrons are lost by forma- behavior of most inorganic and organic
bands are of particular interest because they tion of the anilinium cation, and the auxo- absorbers, their spectra consist of narrow,
can be readily stlK\ied with ordinary spectro- chromic effect disappears as a consequence. well-defined, and characteristic absorption
I For a summary or these rules, see: R. M. SilveRtein,
G. C. Bassler, and T. C. Morrill, Sp«tromerrlc Idmtljica- photometric equipment. Table 7-4 illustrates Absorption by Inorganic Anions. A number peaks, which are little affected by the type of
tiottofOr(lllllic COffIpowllds, 3d ed. New York: Wiley, 1974, the effects of some k:ommon ring substituents. of inorganic anions exhibit ultraviolet absorp- ligand associated with·the metal ion. Portions
pp. 241-255. By definition, an auxochrome is a fune- tion peaks that are a consequence of " ...•11:. of a typical spectrum are shown in Figure 7-4.
transitions. Examples include nitrate (313 nm), The transitions responsible for absorption

TABLE 7-3 EFFECT OF MULTICHROMOPHORES ON ABSORPTION

Compounll TABLE 7-4 ABSORPTION CHARACTERISTICS OF AROMATIC


Type )-Jnm) tmu
COMPOUNDS
CH3CH2CH2CH=CH2 Olefin 184 - 10,000
CH2 -eHCH2CH2CH==CH2 Diolefin (unconjugated) E2 Band B Band
185 - 20,000
H2C-eHCH-eH2 Diolefin (conjugated) 217 21,000 Compound ;""'.(nm) f"maX ;""'.(nm) & ••••
H2C-eHCH-eHCH==CH2 Triolefin (conjugated) 250
Benzene C6H6 204 7900 256 200
0
I Toluene C6HsCH3 207 7000 261 300
CH3CH2CH2CH2CCH3 Ketone ~82 27 m-Xylene C6H4(CH3h 263 300
Chlorobenzene C6HsCI 210 7600 265 240

CH2 =CHCH2CH2CCH3
°I Unsaturated ketone !78
;
30
Phenol
Phenolate ion
C6HsOH
C6HsO-
211
235
6200
9400
270
287
1450
2600
(unconjugated ) Aniline C6HsNH2 230 8600 280 1430
0
I Anilinium ion C6HsNHj 203 7500 254 160
CH2-eHCCH3 Unsaturated ketone 324 24 Thiophenol C6HsSH 236 10,000 269 700
(conjugated) Naphthalene C10He 286 9300 312 289
219 3600 Styrene C6HsCH==CH2 244 12,000 282 450
~y elements of the lanthanide series appear to
mvolve the various energy levels of 4f elec-
trons, while it is the Sf electrons of the acti-
nide teries that interact with radiation. These
inner orbitals are largely screened from exter- .
nal influences by electrons occupying orbitals ".
with higher principal quantum numbers. As a
consequence, the bands are narrow and rela-
tively unaffected by the nature of the solvent
or the species bonded by the outer electrons".
Absorption by Elements of tile First lUIlI
Second TraJlSition-Metal Series. The ions and
complexes of the 18 elements in the first two
transition series tend to absorb visible radia-
tion in one if not all of their oxidation states.
In contrast to the lanthanide and actinide ele- WlM!length. nm
ments, however, the absorption bands are
FIGURE 7-5 Absorption spectra of ~me d.2_.'- dil
?ften broad (Figu~ 7-S) and are strongly
transition-metal ions. FIGURE 7-6 Electron density distribution in various d orbitals.
mfluenced by chemlClll environmental factors.
An example of the environmental effect is
found in the pale blue color of the aquo
copper(II) ion and the much darker blue of
the copper complex with ammonia. i Both theories are based upon the premise except for their spatial orientation. Note that
Metals of the transition series are charac- that the energies of d orbitals of the these orbitals occupy spaces between the three
terized by having five partially occupied d transition-metal ions in solution are not iden- axes; consequently, they have minimum elec-
orbitals (3d in the first series and 4d in the tical, and that absorption involves the transi- tron densities along the axes and maximum
second~ each capable of accommodating a tion of electrons from a d orbital of lower densities on the diagonals between axes. In
pair of electrons. The electrons in these orbit- energy to one of higher energy. In the absence contrast, the electron densities of the d x' -"
als do not aenerally participate in bond of an external electric or magnetic field (as in and the d.. orbitals are directed along the
formation; nevertheless, it is clear that the the dilute gaseous state~ the energies of the axes .
.s.. 0.04
sJ'CC!raicharacteristics of transition metals in-
~olve electronic transitions among the var-
five d orbitals are identical, and absorption of
radiation is not required for an electron to
Let us now consider a transition-metal ion
that is coordinated to six molecules of water
IOUSenergy levels of these d orbitals. move from one orbital to another. On the (or some other ligand). These ligand
Two theories have been advanced to ra- other hand, complex formation in solution molecules or ions can be imagined as being
tionalize the colors of transition-metal ions occurs between the metal ion and water or symmetrically distributed around the central
and the profound influence of chemical envi- some other ligand. Splitting of the d-orbital atom, one ligand being located at each end of
ronment ~ these colors. The crystal-field' energies then results, owing to the differential the three axes shown in Figure 7-6; the result-
t~eory, which we shall discuss briefly, is. the forces of electrostatic repulsion between the ing octahedral structure is the most common
~lm~ler of the two and is adequate for a qual- electron pair of the donor and the electrons in orientation for transition-metal complexes.
Wavelength. nm Itative understanding. The more complex the various d orbitals of the central metal ion. The negative ends of the water dipoles are
FIGURE 7-4 The absorption spec- ~olecular-orbital treatment, however, pro- In order to understand this effect, we must pointed toward the metal ion, and the electri-
trum of a praseodymium chloride solu- vides a better quantitative treatment of; the first consider the spatial distribution of elec- cal fields from these dipoles tend to exert a
tion; a = absorptivity in liter em -, g - '. phenomenon.4 repulsive effect on all of the d orbitals, thus
trons in the various d orbitals.
(Reprinted with permission from T. The electron-density distribution of the increasing their energy; the orbitals are then
Moeller and J. C. Brantley, Anal. Chern., five d orbitals around the nucleus is shown in said to have become destabilized. The maxi-
• For a DOnmathemalicaldiscussion of these theories,
22,433 (1950). Copyright by the Ameri- Figure 7-6. Three of the orbitals, termed d"" mum charge density of the dx• orbital lies
see: L. E. Orgel, An Introduction 10 Traruition Metal
can Chemical Society.) Chemistry, Ligand-FiLld Theory. New York: Wiley, 1960.
along the bonding axis. The negative field of a
dx.' and dr-' are similar in every regard
---------------'1,--;--------
countered: the tetrahedral, in which the four ~ansier rq,tion· and are therefore called,
groups are symmetrically distributed around charge-transfer complexes. Common examples
the metal ion; and the square planar, in which of such complexes include the thiocyanate
dz2. d.2_r' dK2_~2
the four ligands and the metal ion lie in a and phenolic complexes of iron(I1I), the
, ,,
II
d.y, dllz. dY1
single plane. Unique d-orbital. splitting pat- o-phenanthroline complex of iron(II), the
, I
I,
d•• terns for each configuration can be deduced
by arguments similar to those used for the
iodide complex of molecular iodine, and the
ferro-ferricyanide complex responsible for
~ I 6- I II
!! I octahedral structure. the color of Prussian blue.

/~
dlfl,dY1
w Iii
I
I 11/ The magnitude of 11 (Figure 7-7) depends In order for a complex to exhibit a charge-
d,2
/II, upon a number of factors, including the transfer spectrum, it is necessary for one of its
I, I 'd z2••d 2_r'
11/1 valence state of the metal ion and the position components to have electron-donor charac-
I / d.y, d." dyz II
II ,II
,I of the parent element in the periodic table. An teristics and for the other component to have
1/ /I tI
dley, dill, dY1 ( -J --.1 important variable attributable to the ligand electron-acceptor properties. Absorption of
dz2, d.2_y2 is the ligand .field strength, which is a measure radiation then involves transfer of an electron
No Octahedral
of the extent to which a complexing group from the donor to an orbital that is largely
T etrlhedral Square planar
ligand ligand ligand ligand will split the energies of the d electrons; that associated with the acceptor. As a conse-
field foeld foeld field is, a complexing agent with a high ligand field quence, the excited state is the product of an
FIGURE 7-7 Effect of ligand field on d-orbital energies. strength will cause 11 to be large. internal oxidation-reduction process. This be-
It is possible to arrange the common lig- havior differs from that of an organic chromo-
ands in the order of increasing ligand field phore, where the electron in the excited state
bonding ligand therefore has a greater effect strengths: 1- < Dr- < CI- < F- < OH- < is in the molecular orbital formed by two or
is destabilized to the same extent as the d••
on this orbital than upon the dq, du, and d,. C20~- - H20 < SCN- < NH3 < ethylenedi- more atoms.
orbital. Thus, the energy-level diagram for the
orbitals, whose charge densities do not coin- amine < o-phenanthroline < NOi < CN-. A well-known example of charge-transfer
octahedral configuration (Figure 7-7) shows
cide with the bonding axes. These latter orbit- With only minor exceptions, this order absorption is observed in the iron(I1I)jthio-
that the energies of all of the d orbitals rise in
als will be destabilized equally, inasmuch as applies to all transition-metal ions and per- cyanate complex. Absorption of a photon
the presence of a ligand field but. in addition,
they differ from one another only in the mits qualitative predictions about the relative results in the transfer of an electron from
that the d orbitals are split into levels differ-
matter of orientation. The effect of the electri- positions of absorption peaks for the various the thiocyanate ion to an orbital associated
ing in energy by 11. Also shown are energy
cal field on the dJ<2-,. orbital is less obvious, complexes of a given transition-metal ion. with the iron(I1I) ion. The product is thus
diagrams for complexes involving four co-
but quantum c:alcuIations have shown that it Since 11 increases with increasing field an excited species involving predominantly
oldinated bonds. Two configurations are en-
strength, the wavelength of the absorption (iron(II) and the neutral thiocyanate radical
maxima decreases. This effect is demonstrated SCN. As with other types of electronic excita-
by the data in Table 7-5, tion, the electron, under ordinary circum-
stances, returns to its original state after a brief
TABLE 7-& EFFECT OF L1GANDS ON ABSORPTION MAXIMA Charge-Transfer Absorption period. Occasionally, however, dissociation of
ASSOCIATED WITH d-d TRANSITIONS the excited complex may occur, producing
For analytical purposes, species that exhibit
photochemical oxidation-reduction products.
Worn) for the indicated ligands charge-transfer absorption' are of particular
As the tendency for electron transfer in-
importance because molar absorptivities are
creases, less radiant energy is required for the
Central
Increasing lipnd field strength very large (e.....
> 10,(00). Thus, these com-
charge-transfer process, and the resulting
ion 6H20 6NH3 6CN- plexes provide a highly sensitive means for
complexes absorb at longer wavelengths. For
detecting and determining absorbing species.
Cr(lII) 573 462 example, thiocyanate ion is a better electron
456 380 Many inorganic complexes exhibit charge-
538 435 donor (reducing agent) than is chloride ion;
CoiIt) 1345 980
428 294
thus, the absorption of the iron(I1I)jthio-
Co~ ·909
Ni I 1279 925 863 cyanate complex occurs in the visible region,
Cu I) 794 663 610 • For a brief discussion or this type or absorption, see:
whereas the absorption maximum for the
C. N. R. Rao. Ullra-Violel allli Vi.sible SpeClrOSMPY.
Chmlical Appliealions. 3d eel. London: Butterworths, corresponding yellow chloride complex is in
1975.Chapter tI. the ultraviolet region. Presumably, the iodide
Variable
complex of iron(I1I) would absoJ;.b at still first is a single-beam, direct-reading instru- diephragm
longer wavelengths. Such an absorption peak ment consisting of a tungsten-filament lamp, to set , 00'l' T
has not been observed, however, because the r"
a lens to provide a parallel beam of light, a
electron-transfer process is complete, giving • Single- __ --, r-- II
filter, and a photovoltaic cell. The current

*
beam "'~ --1 1---
iron(II) and iodine as products. produced is indicated with a microammeter, photometer - ---I J..--
I I
In most charge-transfer complexes involv- the face of which is ordinarily scribed with a LJ
ing a metal ion, the metal serves as the elec- linear scale from 0 to 100. In some instru-
tron acceptor. An exception is the o-phenan-
throline complex of iron(II) or copper(I),
where the ligand is the acceptor and the metal
ments, adjustment to obtain a full-scale re-
sponse with the solvent in the light path
involves changing the voltage applied to the
------
t
ion is the donor. Other examples of this type lamp. In others, the aperture size of a dia-
are known. phragm located in the light path is altered. -------
~::. -------
",""

Organic compounds form many interesting


charge-transfer complexes. An example is
Since the signal from the photovoltaic cell is
linear with respect to the radiation it receives,
~~ -:::::: :.~ Mirror

quinhydrOJle (a 1 ; 1 complex of quinone and the scale reading with the sample in the light -filII'" Solvent
hydroquinone), which exhibits strong absorp- path will be the percent transmittance (that is, ~ cell

tion in the visible region. Other examples in- the percent of full scale). Clearly, a logarith- I I I
I I I
clude iodine .complexes with amines, mic scale could be substituted to give Reference 21 I I
aromatics, and sulfides, among others. . the absorbance of the solution directly. photovoltaic
cell
Also shown in Figure 7-8 is a schematic
representation of a double-beam, null-type
photometer. Here, the light beam is split by a
mirror. One part passes through the sample,
and thence to a photovoltaic cell; the other
A host of instruments are available commer-
part passes through the solvent to a similar FIGURE 7-8 Schematic diagram for a single-beam and a double-beam
cially for absorbance measurements in the
detector. The currents from the two photo- photometer.
ultraviolet and visible spectral regions. The
voltaic cells are passed through variable re-
simplest of these are colorimeters, in which
sistances; one of these is calibrated as a
the eye serves as the transducer. The most
transmittance scale in linear units from 0 to double-beam principle because this design plement of the solution being analyzed. If
complex are double-beam; recording instru-
100. A sensitive galvanometer, which serves as largely compensates for fluctuations in the several filters possessing the same general hue
ments that cover the entire spectral range
a null indicator, is connected across the two source intensity due to voltage variations. are available, the one that causes the sample
from about 185 to 3000 nm. A few typical
resistances. When the potential drop across Filter Selection for Photometric Analysis. to exhibit the greatest absorbance (or least
instruments are considered in this section.
AB is equal to that across CD, no electricity Photometers are generally supplied with transmittance) should be used.
passes through the galvanometer; under all several filters, each of which transmits a Ultraviolet Photometers. The most impor-
other circumstances, a current is indicated. At different portion of the spectrum. Selection of tant application of ultraviolet photometry is
The photometer provides a simple, relatively the outset, the solvent is placed in both cells, the proper filter for a given application is im- as a detector in high-performance liquid chro-
inexpensive tool for performing absorption and contact A is set at 100; contact C is then portant inasmuch as the sensitivity of the matography. In this application, a mercury-
analyses. Convenience, ease of maintenance, adjusted until no current is indicated. Intro- measurement is directly dependent upon this vapor lamp serves. as a source and the
and ruggedness are properties of a filter duction of the sample into the one cell results choice. The color of the light absorbed is the emission line at 254 nm is isolated by filters.
photometer that may not be found in the in a decrease in radiant power that reaches complement of the color of the solution itself. This type of detector is discussed in more
more sophisticated spectrophotometer. More- the working phototube and a corresponding For example, a liquid appears red because it detail in Chapter 25.
over, where high spectral purity is not impor- decrease in the potential drop across CD; this transmits the red portion of the spectrum but
tant to a method (and often it is not), analyses lack of balance is compensated for by moving absorbs the green. It is the intensity of green S~rophotomet.r.
can be performed as accurately with a photom- A to a lower value. At balance, the percent radiation that varies with concentration; a A large number of spectrophotometers are
eter as with more complex instrumentation. transmittance is read directly from the scale. green filter should thus be employed. In available from commercial sources. A few of
Visible Photometers. Figure 7-8 presents Commercial photometers usually cost a general, then, the most suitable filter for a these are considered in the paragraphs that
schematic diagrams for two photometers. The few hundred dollars. The majority employ the photometric analysis will be the color com- follow.
IIIItnneIICs for the Vin"ble Range. Several The Bausch and Lomb Spectronic 20 spec- can be moved in or out of the beam in order As might be expected, performance
spcctropootometers are available that are trophotometer shown in Figure 7-9a employs to set the meter to 100% T. specifications vary considerably among in-
designed to operate within the general region a reference phototube, which serves to com- The range of the Spectronic 20 is from 340 struments and are related, at least to some
.of 380 to 800 nm; within these limits, the pensate for fluctuations in the output of the to 625 nm; an accessory phototube extends degree, to instrument price. Typically, band-
range varies somewhat from instrument to tungsten-filament light source; this design this range to 950 nm. Other specifications for widths vary from 2 to 8 nm; wavelength
iDstrument. eliminates the need for a stabilized lamp the instrument include a bandwidth of 20 nm accuracies of ±0.5 to ±2 nm are reported.
Spectrophotometers designed for the vis- power supply. The amplified difference signal and a wavelength accuracy of ± 2.5 nm. The optical designs for the various grating
ible region are frequently simple, single-beam, . from the two phototubes powers a meter with a The Turner instrument, shown schemat- instruments do not differ greatly from those
grating instruments that are inexpensive (less . 5 1/2-in. scale calibrated in transmittance and ically in Figure 7-9b, makes use of a of the two instruments shown in Figure 7-9.
than SIOOO~ rugged, and readily portable. At absorbance. tungsten-filament bulb as a source, a plane One manufacturer, however, employs a con-
least one is battery operated and light enough: The Spectronic 20 is equipped with an reflection grating in an Ebert mounting for cave rather than a plane grating; a simpler
to be hand-held. The most common applica- . occluder, which is a vane that automatically dispersion, and a phototube detector that is and more compact design results.
tion of these instruments is to quantitative: falls between the beam and the detector sensitive in the range between 210 and Figure 7-10 is a schematic diagram of the
analysis, although several produce surprisingly whenever the cuvette is removed from its 710 nm. The readout device is a meter cal- Beckman DU-2 spectrophotometer, a high-
good absorption spectra as well. . holder; the 0% T adjustment can then be ibrated both in transmittance and absorb- quality, single-beam instrument for the ultra-
Figure 7-9 depicts two typical visible: made. The light control device shown in ance; instruments with 4- or 7-in. scales are violet-visible region. This instrument, which
spectrophotometers. Figure 7-9a consists of a V-shaped slot that offered. The transmittance is first set to zero employs a Littrow prism rather than a grat-
. by adjustment of the amplifier output while a ing, originally appeared on the market in
shutter screens the lamp from the detector. 1941, at which time it was the first ultraviolet-
With the solvent in the light path, the 100% visible spectrophotometer to become com-
T adjustment is then accomplished by vary- mercially available to chemists. The optics
i_Objective lens
ing the output of the stabilized lamp power were of remarkably high quality; many of the
--~-..-- supply. Finally, the transmittance or absorb- early instruments are still in use today,

,
ance is read with the sample in the beam. The although most have been refitted with more
_-=:::::::::::: - instrument specifications include a bandwidth modern electronic systems. The optical design
of 9 nm, a wavelength accuracy of ±2 nm, of the instrument is still being employed by
and a photometric accuracy of 0.5% A. one instrument manufacturer.6
Light control Several accessories are offered with the The DU-2 spectrophotometer is equipped
Turner instrument. One, which includes a with quartz optics, which permits its use in
deuterium lamp, extends the range of the in- both the ultraviolet and visible regions of the
strument to 210 nm; another provides an spectrum. Interchangeable radiation sources
additional phototube that permits measure- are provided, including a deuterium or
ments to 1000 nm. hydrogen discharge tube for the lower wave-
Single-Beam Instruments for the Ukra- lengths and a tungsten-filament lamp for the
violet-Visible Region. Several instrument visible region. A pair of mirrors reflects radia-
manufacturers offer single-beam instruments tion through an adjustable slit into the
which can be used for both ultraviolet and monochromator compartment. After tra-
visible measurements. The lower wavelength versing the length of the instrument, the ra-
extremes for these instruments vary from 190 diation is reflected into a Littrow prism; by
to 210 nm and the upper from 800 to adjusting the position of the prism, light of
1000 nm. All are equipped with interchange- the desired wavelength can be focused on the
able tungsten and hydrogen or deuterium exit slit. The optics are so arranged that the
lamps. Most employ photomultiplier tubes as entrance and exit beams are displaced from
{bl
detectors and gratings for dispersion. Some one another on the vertical axis; thus, the exit
FIGURE 7-9 Two examples of simple spectrophotometers. (a) The Spectronic 20; diagram are equipped with digital readout devices;
courtesy of Bausch and Lomb, Rochester, New York. (b) The Turner 350; diagram courtesy others employ large meters. The prices for
of Amsco Instrument Company (formerly Turner Associates) Carpinteria, California. these instruments range from $2000 to $7000.
discussed earlier are operated at fixed slit The performance characteristics of the two
widths because of the linear characteristics of instruments differ significantly. The stray light
their dispersing elements ~ at 220 nm, the wavelength accuracy, and the
The manual DU-2 is particularly well wavelength repeatability for the Hitachi in-
suited for quantitative analytical measure- strument are < 0.07%. ±0.4 nm, and
ments that require absorbance data at a ±0.2 nm, respectively; the corresponding
limited number of wavelengths. figures for the Cary instrument are < 0.002%.
Double-Beam Instruments for the Ultra- ±0.2 nm, and ±O.l nm. The photometric lin-
violet-Visible Region. NumeroUs double-beam earity of the Hitachi instrument is 0.005 to
spectrophotometers for the ultraviolet-visible 0.01 A compared with 0.0016 to 0.003 A for
region of the spectrum are now available. the Cary. In addition, the Cary instrument
GCnerally, these instruments are more expen- has eight recorder ranges (compared with
sive than their single-beam coQnterparts, with three for the Hitachi) and a number of au-
the nonrecording variety ranging in cost from tomated features such as an automatic slit
about $4000 to SI0,000. The most sophisticated control to provide either constant resolution
instrument may cost as much. as S30,ooo or or constant noise, automatic 0% T set, and
more when all accessories a~ included. The digital display of concentrations.
latter will generally provide spectra in the For many applications, simpler and less
wavelength range of 185 to 3000 nm. expensive instruments, such as that shown in
Figure 7-11 shows construction details of Figure 7-11&, provide just as satisfactory re-
two commercially available, ultiaviolet-visible, sults as do the more expensive ones.
double-beam instruments. The:upper diagram Comparison of Single- •••• Double-Beam
is for a relatively inexpendive (-$5000~ Instruments. Because the measurement of P
FIGURE 7-10 Schematic diagram of the Beckman DU-2@lSpectrophotom-
manual instrument (the Hitachi Model and Po is made simultaneously or nearly si-
eter. (By permission, Beckman Instruments, Inc., Fullerton, California.)
100-60~ which employs a conCave grating for multaneously, a double-beam instrument
dispersing and focusing the radiation on the compensates for all but the most short-term
entrance slit. A motor-driven sector mirror electrical fluctuations, as well as other time-
divides the beam; after passage through the dependent irregularities in the source, the
beam passes above the entrance mirror as it is directly proportional to absorbance; the sample or solvent, the beams are recombined detector, and the amplifier. Therefore, the
leaves the monochromator. readout is digital Gilford also offers an acces- to give an ac sipal (unless their powers are electrical components of a double-beam pho-
The ceIl compartment accommodates as sory that oscillates the cuvette carriage at a identical) which is transduced, amplified, and tometer need not be of such high quality as
many as four rectangular l-cm cells, anyone rate of five cycles per second, alternately plac- converted to a ratio of P and Po. Either a those for a single-beam instrument. Offsetting
of which can be positioned in the path -of the ing the sample and reference cells in the light meter or a digital readout is available; a this advantage, however, is the greater
beam by movement of a carriage arrange- path. With this addition, the readout of the recorder can also be purchased as an accessory. number and complexity of components asso-
ment. Compartments are also available that instrument is automatic in absorbance. Addi- Figure 7-11b shows details of a high- ciated with double-beam instrumentation.
will hold both cylindrical cells and cells up to tion of wavelength and slit-drive mechanisms performance, recording spectrophotometer Single-beam instruments are particularly
10 em in length. and a recorder, makes possible the direct re- (Varian Cary Model 219). Its cost is signifi- well adapted to the quantitative analysis that
The original DU-2 employed interchang- cording of spectra. cantly greater than the sim~ler instrument involves an absorbance measurement at a
able detectors-one for the range of 190 to The modified DU-2 design achieves pho- (-$12,000). A unique feature of the Cary single wavelength. Here, the simplicity of the
625 nm and the other for 625 to 800 nm. Gil- tometric accuracies as great as ± 0.5 % A by instrument is the double-pass 9ptics by means instrument and the concomitant ease of main-
ford Instrument now offers a photometer employing high-quality electronic compo- of which radiation is reflected and diffracted tenance offer real advantages. The greater
modernization kit for the DU-2 in which the nents that are opCpited well below their rated by the grating twice before .being split for speed and convenience of measurement, on
phototubes are replaced by a single photo- capacities. Narrow effective bandwidths (less passage through the sample and reference the other hand, make the double-beam instru-
multiplier with the same range as the original than 0.5 nm) can be obtained throughout the cells. The effect is the same as with a double ment particularly useful for qualitative
two. An amplifier circuit is then employed to spectrum region by suitable variation of the monochromator-that is, enhanced disper- analyses, where absorbance measurements
convert the transducer signal to a voltage that slit adjustment (note that grating instruments sion and reduction in scattered radiation. must be made at many wavelengths.
APPLICATION OF
ABSORPTION MEASUREMENT
TO QUALITATIVE ANALYSIS

Ultraviolet and visible spectrophotometry


have somewhat limited application for quali-
tative analysis because the number of absorp-
g
tion maxima and minima are relatively few.
Thus, unambiguous identification is frequently
i
1i
1.2
1.6
impossible. «

Qualitative Techniques
Solvents. In choosing a solvent, considera-
tion must be given not only to its transpar-
----e~, but also to its possible effects upon the 280 300
absorbing system. Quite generally, polar sol- WIIIeleng1h,nm
vents such as water, alcohols, esters, and ke- FIGURE 7-12 Effect, of solvent on the
tones tend to obliterate spectral fine structure absorption spectrum of acetaldehyde.
arising from vibrational effects; spectra that
approach those of the gas phase (see Figure'
7-12) are more likely to be observed in non-
polar solvents such as hydrocarbons. In addi-
tion, the positions of absorption maxima are found at low transmittance or high absorb-
also influenced by the nature of the solvent. ance (0.8 to 1.6~
Clearly, the same solvent must be used when The log absorbance plots shown in Figure
comparing absorption spectra for identifica- 7-13c are advantageous for the comparison of
tion purposes. spectra. If Beer's law is written in the form
Table 7-6 lists some common solvents and
log A. = log e + log be
the approximate wavelength below which
they cannot use because of absorption. it is seen that of the terms on the right-hand
These minima depend strongly upon the side of this equation, only e varies with wave- .
purity of the solvent. 7 length. Thus, if the concentrations or cell
Methods for Plotting SpectrL Figure 7-13 lengths for the spectra being matched are not
shows absorption spectra for solutions of per- identical, log A. is displaced by exactly the
manganate plotted with three different ordi- same absolute amount at each wavelength.
nates. Note that the greatest differences among This effect is shown in Figure 7-13c. When
transmittance plots occur when the transmit- the plots of log A. are compared with those of
tance ranges between 20 to 60%. In contrast, A (Figure 7-13b1 the advantage of the former
greater differentiation of absorbance plots is for qualitative analysis becomes evident. It
should also be noted that plots of log e are
also readily compared for the same reason.
, Most major suppliers or reapt chemicals in the
(bl United States olTerspectrochemical grades or solvents; Detection of Functional Groupl
these meet or exceed the requirements set rorth in Rea-
FIG U R E 7-11 Schematic diagrams for two ultraviolet-visible spectrophotometers. (a) Hitachi gent Chemicals, American Chemical Society S~cifications,
Even though they may not provide the un-
Model 100-60 (courtesy of Hitachi Scientific Instruments, Sunnyvale, California); (b) Varian Sth ed. Washington, D.C.: American Chemical Society, ambiguous identification of an organic
Cary Model 219 (courtesy of Varian Instrument Division, Palo Alto, California). 1974. compound, absorption measurements in the
visible and the ultraviolet regions are never- m~tion. In addition, many nonabsorbing Scope
theless useful for detecting the presenc:e of SpeCIes can be analyzed after conversion to
The applications of quantitative absorption
absorbing species by suitable chenrlca1
..
t!
certain functional groups that act as
treatment .
methods are not only numerous, but also
.E chromophores.8 For example, a weak absorp- touch upon every field in whic~ quantitative
tion band in the region of 280 to 290 nm, 2. High sensitiVity. Molar absorptivities in
~ chemical information is require4. The reader
which is displaced toward shorter wave- the range of 10,000 to 40,000 are common,
•..~ lengths with increased solvent polarity, particularly among the charge-transfer com-
can obtain a notion of the scope of spectro-
photometry by consulting a series of review
40 strongly indicates the presence of the car- plexes of inorganic species. Thus, analyses for
articles published periodically in Analytical
bonyl group. A weak absorption band at concentrations in the range of 10-4 to 10-'M
Chemistry9 and from monographs on the
about 260 nm with indications of vibrational are ordinary; the range can often be extended
subjeet.IO
fine structure constitutes evidence for the to 10- 6 or even 10- 7M with suitable
20 Applications to Absorbing Species. Tables
existence of an aromatic ring. Confirmation procedural modifications.
7-2, 7·3, and 7-4 list many common organic
of the presence of an aromatic amine or a 3. Moderate to high selectivity. It may be
chromophoric groups. Spectrophotometric
phenolic structure may be obtained by com- possible to locate a wavelength region in
0 analysis for any organic compound contain-
paring the effects of pH on the spectra of so- which the only absorbing component in a
ing ?ne or more of these groups is potentially
lutions containing the sample with those sample is the substance being determined.
20.00 feasible; many examples of this type of
shown in Table 7-4 for phenol and aniline. Furthermore, where overlapping absorption
analysis are found ~ the literature.

r
1.6 bands do occur, corrections based on addi-
A number of inorganic species also absorb
tional measurements at other wavelengths are
sometimes possible. As a consequence, the
1.2 QUANTITATIVE ANALYSIS BY separation step can be omitted.
j
« 0.6
ABSORPTION MEASUREMENTS

AbsorPtion spectroscopy is one of the most


4. Good accuracy. For the typical spectro-
photometric or photometric procedure, the
• See, for example, D. F. Boltz and M. G. Mellon, Arud.
0-.. •• 216R (1976); and J. A. Howel~ and L. G.
Harp, AIIIJl. CheIn~50, 243R (1978~
relative error in concentration measurements 10 See, for example: E. B. Sandel~ Colorimetric Determi-
useful tools available to the chemist for quan- lies in the range of 1 to 3 %. Such errors can ""'ion of Traces of M eta/ •• 4th ed. New York: Interscicnce,
titative analysis. Important characteristics of often be reduced to a few tenths of a percent 1978; CoIor"rk Determi""'ion of N_Is, 2d ed.
0.4 spectrophotometric and photometric methods with special precautions. cds. D. F. Boltz and J. A. Howell. New York: Wiley,
include: 1978;Z. Marczenko, Spectrophotorrwtrk Determination of
S. Ease and convenience. Spectrophotomet-
1. Wide applicability. As we have pointed £/ements. New York: Halsted Press, 1975;and M. Piscz
ric or photometric measurements are easily and J. BartOl, CoIorinwIric I11III Fluorometric AlIIJlysls of
out, numerous inorpnic and orpnic species and rapidly performed with modern instru- Orpnic COfIIPOIItWI I11III Drugs. New York: Maroel
absorb in the ultraviolet and visible ranges ments. Delcker. 1974.
and are thus susceptible to quantitative deter-

• See: R. M. Silventein, G. C. Busler, and T. C. Morril~


Sp«trometric ldentijiCfltiOtl of Organic CompountIs, 3d ed.
New York: Wiley, 1974,Chapter S.
TABLE 7-6 SOLVENTS FOR THE ULTRAVIOLET AND THE
VISIBLE REGIONS

Approximate Approximate
transparency transparency
Solvent minimum (nm) Solvent minimum (nm)
FIGURE '-13 Three methods for plotting
absorption spectra. The numbers beneath the Water 180 Carbon tetrachloride 260
Ethanol 220 Diethyl ether 210
curves indicate the ppm KMn04 in the solu- Hexane 200 Acetone 330
tion; 2.O-cm cells were employed. (Plots from Cyc10hexane 200 Dioxane 320
M. G. Mellon, Analytical Absorption Spec- Benzene 280 Cellosolve 320
troscopy.NewYork:Wiley,19SO,pp.l04-106.)
and are thps susceptible to direct determina-
tion; we have already mentioned the various
ance per unit of concentration is greatest at
this point; the maximum sensitivity is thus
proper cell cleaning and drying techniques.
Erickson and Surlesl1 recommend the follow- /
.•..
-1--....
I ••••
,
l
I
/ I •••. I
transition metals. In addition, a number of realized. In addition, the absorption curve is ing cleaning sequence for the outside win- / I --"" M+N
other species also show characteristic absorp-
tion. Examples include nitrite, nitrate, and
often flat in this region; under these circum-
stances, good adherence to Beer's law can be
dows of cells. Prior to measurement, the cell
surfaces are cleaned with a lens paper soaked
/ I',,
chromate ions; osmium and rut"henium tetrox- expected (p. 156). Finally, the measurements in spectrograde methanol. The paper is held N

ides; molecular iodine; and ozone. are less sensitive to uncertainties arising from with a hemostat; after wiping, the methanol is
Applications to Nonabsorbing Species. Nu- failure to reproduce precjsely the wavelength allowed to evaporate, leaving the cell surfaces
merous reagents react with nonabsorbing setting of the instrument. free of contaminants. The authors showed
species to yield products that absorb strongly The absorption spectrum, if available, aids that this method was far superior to the usual
in the ultraviolet or visible regions. The suc- in choosing the most suitable filter for a procedure of wiping the cell surfaces with a
cessful application of such reagents to quan- photometric analysis; if this information is dry lens paper, which apparently leaves lint
titative analysis usually requires that the color- lacking, the alternative method for selection, and films on the surface.
forming reaction be forced to near completion. given on page 181, may be used. Analysis of Mixtures of Absorbing Substan-
It should be noted that these reagents are Variables that InfluenCe Absorbance. Com- ces. The total absorbance of a solution at a
FIGURE 7-14 Absorption spec-
frequently employed as well for the determina- mon variables that influence the absorption given wavelength is equal to the sum of the
tion of an absorbing species, such as a transi- absorbances of the individual components trum of a two-component mixture.
spectrum of a substance Include the nature of
tion-metal ion; the molar absorptivity of the the solvent, the pH of the solution, the present (p. 153~ This relationship makes pos-
product will frequently be orders of magnitude temperature. high electrolyte concentrations, sible the analysis of the individual compo-
greater than that of the uncombined species. and the presence of interfering substances. nents of a mixture, even if their spectra
A host of complexing agents have been The effects of these variables must be known; overlap. Consider, for example, the spectra of
employed for the determination of inorganic conditions for the analrsis must be chosen M and N, shown in Figure 7-14. Obviously, in the mixture, CMand CN,can be readily cal-
species. Typical inorganic reagents include such that the absorbance I.villnot be materially no wavelength exists at which the absorbance culated. These relationships are valid only if
thiocyanate ion for iron, cobalt, and molyb- inftuenced by small, uncotttrolled variations in of this mixture is due simply to one or the Beer's law is followed. The greatest accuracy
denum; the anion of hydrogen peroxide for their magnitudes. components; thus, an analysis ror either M or in an analysis of this sort is attained by
titanium, vanadium, and chromium; and Determination of the Relationship Between N is impossible by a single measurement. choosing wavelengths at which the differences
iodide ion for bismuth, palladium, and telluri- Absorbance and Concentration. After deciding However, the absorbances of the mixture at in molar absorptivities are large.
um. Of even more importance are organic upon the conditions for the analysis, it is the two wavelengths ).' and ).H may be ex- Mixtures containing more than two absorb-
chelating agents which form stable, colored necessary to prepare a calibration curve from pressed as follows: ing species can be analyzed, in principle at
complexes with cations. Examples include a series of standard solutions. These stand- least, if a further absorbance measurement is
o-phenanthroline for the determination 'of iron, ards should approximate the overall compo- A' = ti.cbCM+ ENbcN (at ).') made for each added component. The uncer-
dimethylglyoxime for nickel, diethyldithio- sition of the actual samples and should cover tainties in the resulting data become greater,
A" = tMbCM+ eNbcN (at ;.")
carbamate for copper, and diphenyldithi6- a reasonable concentration range of the however, as the number of measurements
carbazone for lead. analyte. Seldom, if ever, is it safe to assume The four molar absorptivities ti.c, EN,tM, and increases.
adherence to Beer's law and use only a single eN can be evaluated from individual standard
standard to determine the molar absorptivity. solutions of M and of N, or better, from the EXAMPLE
Procedural Details
The results of an analysis should never be slopes of their Beer's law plots. The absorb- A 0.246-g sample of steel was analyzed for
The first steps in a photometric or spectro- based on a literature value for the molar ances of the mixture, A' and A", are exper- Cr and Mn by solution of the sample in acid
photometric analysis involve the establish- absorptivity. ' imentally determinable, as is b, the cell and dilution to exactly 250 ml. A 50.0-ml ali-
ment of working conditions and the prepara- Cleaning and Handling of Cells. It is thickness. Thus, from these two equations, the quot was treated with potassium persulfate.
tion of a calibration curve relating concentra- apparent that accurat~ spectrophotometric concentration of the individual components which, in the presence of a catalytic amount
tion to absorbance. analysis requires the use of good-quality, of Ag+, converted the Cr and Mn to Cr20~-
SelectiOll of Wavelength. Spectrophotomet- matched cells. These should be regularly cal- and MnO;. The resulting solution was then
ric absorbance measurements are ordinarily ibrated against one another to detect differ- diluted to 100 ml. In a 1.00-cm cell, the absorb-
made at a wavelength corresponding to an ences that can arise from scratches, etching, 11 J. O. Erickson and T. Surles, American Laboratory, • ance of the solution at 440 and S45 nm was
absorption peak because the change in absorb- and wear. Equally important is the use of (6~ SO(1916~ found to be 0.932 and 0.778, respectively. Cal-
eulate the percent Mn and Cr employing the of a recording spectrophotometer can be dif- ordinary double-beam mode instrument. In
following previously determined data: ferentiated electronically (p. 66) or mechan- the dual mode, the reference cell is not
ically and recorded as a function of required and the sample cell is alternately
l,nm 6, MnO; wavelength. Alternatively, use is made of a exposed to radiation from each of the grat-
dual-wavelength spectrophotometer such as ings. In the double-beam mode, only mono-
440 95 that shown in Figure 7-16. Note that this in- chromator 1 is used and the beam is directed
545 2350
strument is designed to operate either in a first through the sample and then through the
dual-wavelength mode, as shown, or in an reference.
For the absorbance at 440 nm, we may
write
0.932 = 369 x 1.00 x MCr
+ 95 x 1.00 x MMn

0.778 = 11 x 1.00 x MCr


+ 2350 x 1.00 x M ••••

Mcr = (0.932 - 95MMn)f369

Substituting into the second equation gives

0.778 x 369 = 11 x 0.932 - 11 X 95MMn


+ 2350 X 369MMn
FIGURE 7-15 Comparison of
4
MM. = 3.20 X 10- a derivative curve (a) with a
standard transmittance curve (b). AlterMte
Substitution of this into the first equation cuvette
position
yields
FIGURE 7-16 Schematic diagram of a dual-wavelength spectrophotometer. Radia-
Mcr = 2.44 X 10-3
Derivative Spectrophotometry!2 tion from the source is focused by lamp mirror MI on the slit of the duochromator.
This mirror and the aperture define two independent beams used in the dual wavelength
Employing 54.9 as the atomic weight of
In derivative spectrophotometry, the first or mode. After collimation by mirror M2 and diffraction by gratings MONO I and
manganese gives
second derivative of absorbance or transmit- MONO 2, the beams are focused by mirror M3 on the exit slit. After passing through
3.20 X 10-4 x 100(50.0;250) tance with respect to wavelength is employed the exit slit and the zoom lens, the beams alternately pass through, or are deflected
x 0.0549 x 100 as the ordinate in plotting spectral data. As by the rotating chopper mirror M4. During one half of the cycle, light from MONO I
shown by Figure 7-15; such plots often reveal is reflected by mirror M5 toward a light trap, and light from MONO 2 is reflected by
0246
spectral detail that is less apparent in an mirror M6, located behind the chopper mirror, toward the sample cuvette. During the
ordinary absorption spectral plot. second half of the cycle, light from MONO I is reflected by the chopper mirror
Derivative spectra are obtained in two toward the sample cuvette while light from MONO 2 is trapped. For ordinary double-
Similarly, ways. The out~ut signal from the transducer beam spectra, shutter I is turned so that it deflects beam I away from the entrance
2.44 x 10- 3 x 100(50.0/250) aperture, shutter 2 is also moved so that the second half of the radiation passes through
x 0.0520 x 100 the solvent contained in the alternate cuvette. Twin detectors are then employed to
12 For additional information, see: T. C. O'Haver and compare the two beams. (Courtesy of American Instruments Company, Division of
0246 G. L. Green, Ammcan Laboratory, 7 (3t 15 (1975); and Travenol Laboratories, Silver Springs, Md.)
T. J. Porro, Anal. ChmI, 44 (4t 93A (1972~
photometer or spectrophotometer rather than Pen:ent T,ordinary
A spectrum such as that shown in Figure
7-15 is obtained with a dual-wavelength in- the shutter and the solvent. The three types of o 20 40 60 80 100
strument by offsetting one monochromator
from the other by one or two nanometers and
dilTerential methods are compared with the
ordinarY method in Table 7-7. ell Shutter
... oJ
I";;r-L:
san,Pl!
-- I I I
Reference 1
I i I

recording the resulting derivative spectrum. Higb-A1JIorbIlnce Method. In. the high- t' I I I" I I
---
I r ~_,J
Dual-wavelength spectrometers are also absorbance procedure, the zero adjustment is o 20 40 60 80 100
Pen:ent T, expanded lCIle
useful for quantitative analysis of one species carried out in the usual way. with the shutter
in the presence of a second species, when the imposed between the source and the detector. Pen:ent T, ordinary
spectra of the two are alike. Here, the instru- x The 100"10 transmittance adjustment, however.
.:-1'" o 20 40 60 60 100
ment is operated in the nonscanning mode is made with a standard solution of the . j I I I I I ,__1-,> I
with one monochromator set at a wavelength
~. analyte, which is less concentrated than the _-- "" I
_Reference 2 §lmple I Solr:"t
at which the two species have identical molar .~ ±2.0 sample, in the light path. Finally. the stan- Lc:.,..••.. r -, ! I kl, I . r
absorptivities. The other monochromator is '! dard is replaced by the sample, and a relative o 20 40 60 80 100
then set to a wavelength where the analyte ti transmittance is read directly. As shown in Percent T,expanded scole
absorbance exhibits a large change as a func- Figure 7-18a, the elTect of this modification is
~ Percent T,ordinary
tion of concentration. The resulting signal is to expand a small portion of the trans-
then proportional to the concentration of the
o 20 40 60 80 100
mittance scale to a full 100%; thus. as shown I I , I 'PO K'I .••I I I
analyte under most circumstances. by the upper scale of Figure 7-18a, the trans- .,,'" \ •...
~;rence 2 SImple Referenc~ 1
Dual-wavelength spectrophotometers are
also useful for monitoring the concentration
mittance of the reference standard employed
is 10% when compared with pure solvent. while ••••
O.
I I
20
I I
40
I"
60
I
80
,=" 100
of two species in kinetic studies, and for deriv- the sample exhibits a 4 % transmittance Percent T.expanded scole
ing absorption spectra of a species in a turbid against the same solvent. That is, the sample
medium. transmittance is four-tenths that of the stan- FIGURE 7-18 Scale expansion by variohs
dard when compared with a common solvent. methods: (a) high-absorbance method; (b) low-
Differentiaf Absorption Methodst3 Figure 7-17 Relative errors in spectro- When the standard is substituted as the refer- absorbance method; and (c) method of ulti·
photometric analysis. Curve 0: ordinary ence, however. its transmittance becomes mate precision.
In the previous chapter (p. 160~ it was
method; reference solution has transmittance
pointed out that the accuracy of many
of 100%. Curve A: high-absorbance method;
photometers and spectrophotometers is
limited by the sensitivities of their readout reference solution has transmittance of 10%.
Curve B: low-absorbance method; reference
devices. With such instruments, the un-
solution has transmittance of 90% Curve C: TABLE 7-7 COMPARISON OF METHODS FOR ADSORPTION
certainty ST in the measurement of T is
method of ultimate precision; reference solu- MEASUREMENTS
constant, and the resulting relative concen-
tration errors are given by Equation 6-14. tions have transmittances of 45 % and 55 %,
respectively. ST is 0.5% for each curve. See Impose.t in Beam for Indicator Setting of:
Curve 0 in Figure 7-17 gives a plot of the
Figure 7·18 for the type of scale expansion Method Designation- 0% T 100% T
relative concentration errors to be expected
corresponding to curves A, B, and C. Solvent
from a readout-limited uncertainty of ST = Ordinary 0 Shutter
±0.005. This uncertainty would be typical High absorbance A ora Shutter Standard solution
for an instrument equipped with a small less concentrated
D'Arsonval-type meter. Clearly, serious ana- than sample
lytical errors are encountered when the analyte Low absorbance Bor b Standard solution Solvent
10"10(A = 1.0) or greater than 70"10 (A = 0.15). more concentrated
concentration is such that % T is smaller than
DilTerential methods provide a means of than sample
expanding the scale of a readout-limited Ultimate precision Core StaQdard solution Standard solution
instrument and thus reducing this type of more concentrated less concentrated
than sample than sample
IS For a complete analysis or these methods, see: C. N. error significantly. These methods employ
Reilley and C. M. Crawford, AIItIl. CIrem., 17. 716 (19SS); solutions of the analyte for adjusting the zero
C. F. Hiskey, Anal. Chern. 11, 1440 (1949~ and/or the 100% transmittance setting of the • See Figures 7-17 and 7-18.
100% (lower scale of Figure 7-18a). The expansion is again achieved; here, two refer- amplification of the photoelectric current. tions with differing slopes, one occurring at
transmittance of the sample remains four- ence solutions, one having a smaller and one The method of ultimate precision requires in- the outset of the titration and the other lo-
tenths that of the reference standard; here, having a greater transmittance than the struments with both of these qualities. cated well beyond the equivalence-point
then, it is 40%. sample, are employed to adjust the 100% and The capability of a spectrophotometer to region; the end point is taken as the inter-
If the instrumental uncertainty in the the 0% scale readings. A nonlinear relation- be set to full scale with an absorbing solution section of extrapolated linear portions. Figure
measurement of absorbance is not affected by ship between relative absorbance and con- in the radiation path will depend both upon 7-19 shows some typical titration curves.
the modification, use of the standard as a centration is again observed. the quality of its monochromator and the Titration of a nonabsorbing species with a
reference brings the transmittance into .the Precision Gain by DitJerential Methods. It stability of its electronic circuit. Furthermore, colored titrant that is decolorized by the re-
middle of the scale, where instrumental uncer- is possible to derive relationships analogous this capacity will be wavelength-dependent, action produces a horizontal line in the initial
tainties have a minimal effect on the relative to Equation 6-14 for the three differential since the intensity ofthe source and the sensi- stages, followed by a rapid rise in absorbance
concentration error. methods. IS Figure 7-17 compares the cor- tivity of the detector change as the wave- beyond the equivalence point (Figure 7-19a).
It can be shown that a linear relationsbip responding error curves with that of the length is varied. In regions where the intensity The formation of a colored product ftom
exists between concentration and the relative ordinary method. From curve A, it is appar- and sensitivity are low, an increase in slit colorless reactants, on the other hand, ini-
absorbance measured by the high-absorbance ent that a significant gain in precision results width may be necessary to realize a full-scale tially produces a linear rise in the absorbance,
technique.14 The method is particularly useful when the high-absorbance method is applied setting; under these circumstances, scattered followed by a region in which the absorb-
for the analysis of samples with absorblU)ces to samples having high absorbances or low radiation may lead to errors unless the qual- ance becomes independent of reagent volume
greater than unity. . transmittances. The plot was based upon the ity of the monochromator is high. Alter- (Figure 7-19b). Depending upon the absorp-
Low-Allsorbuce Method. In the low-ab- reference standard employed in Figure 7-18a; natively, a very high current amplification tion characteristics of the reactants and the
sorbance procedure, a standard solution some- that is, for a reference having a normal may be necessary; again, unless the electronic products, the other curve forms shown in
what ",ore concentrated than the sample transmittance of 10%. The error curve stability is good, significant photometric error Figure 7-19 are also possible.
is employed in lieu of the shutter for zeroing approaches that for the ordinary procedure may result. In order to obtain a satisfactory photomet-
the indicator scale; the full-scale adjustment (see curve 0, Figure 7-17) as the percent ric end point, it is necessary that the ab-
is made in the usual way, with the solvetit in transmittance of the reference increases and sorbing system(s) obey Beer's law; otherwise,
the light path. The sample transmittand: is approaches 100. Error curves for the low- the titration curve will lack the linear por-
then obtained by replacing the solvent with absorbance method (curve B) and the method tions needed for end-point extrapolation.
the sample. The effect of this procedure is of ultimate precision (curve C) are based on Photometric or spectrophotometric measure- Further, it is necessary to correct the absorb-
shown in Figure 7-18b; note that, again, a the corresponding data in the caption accom- ments can be employed to advantage in lo- ance for volume changes. The observed
small segment of the scale is expanded to panying Figure 7-18. - cating the equivalence point of a titration.16
100% transmittance, and that the transmit- IIIStnMnentaI RequiremenlB for Precision The end point in a direct photometric titra-
tance or the sample now lies near the center Methods. For the low-absorbance method, it tion is the result of a change in the con-
of the expanded scale. is clearly necessary that the spectrophotom- centration of a reactant or a product, or both;

\:
The low-absorbance method is particularly eter possess a dark-current compensating clearly, at least one of these species must (aJ Es• ED =0 Ibl t:c >0 Ie) (,>0
applicable to samples having absorbances of circuit capable of offsetting larger currents absorb radiation at the wavelength selected. 0

:J r
(,. £t·

less than 0.1. Here a nonlinear relationship than are normally produced when no radia- In the indirect method, the absorbance of an
exists between relative absorbance and con- tion strikes the photoelectric detector. For the indicator is observed as a function of titrant
centration. As a consequence, a number of high-absorbance method, on the other hand, volume.
standard solutions must be employed to
establish accurately the calibration curve" for
the analysis.
The Method of Ultimate Precision. ;rhe
two techniques just described can be ciJm-
bined to give the method of ultimate preci-
.
the instrument must have a sufficient reserve
capacity to permit setting the indicator at
100% transmittance when an absorbing solu-
tion is placed in the light path. Here, the full-
scale reading may be realized either by
increasing the radiation intensity (most often
Titration Curve.
A photometric titration curve is a plot of cor-
rected absorbance as a function of the volume
of titrant. If conditions are chosen properly,
the curve will consist of two straight-line por-
~
'y & If)

(, = 0

sion. As shown in Figure 7-18c, a scale by widening the slits) or by increasing the
FIGURE 7-19 Typical photometric titra-
" For further information concerning this technique,
tion curves. Molar absorptivities of the sub-
14 See: C. N. Reilley and C. M. Crawford, AMI. Chem. IS See: C. N. Reilley and C. M. Crawford, AMI. Chem. see: J. B. Hcadridge, Photometric Titratioru. New York: stance titrated, the product, and the titrant
27, 716 (1955~ 27, 716 (1955). Pergamon Press, 1961. are given by c., lp, l •• respectively.
absorbing species may not interfere, since end points. Acid-base indicators have been cat Iaboratories.19 As an outgrowth, automat-
values are multiplied by (V + v)/V, where V is
only a change in absorbance is being mea- employed for photometric neutralization ic instruments are now beginning to find
the original volume of the solution and v is
the volume of added titrant. sured. titrations. The photometric end point has also application in such diverse fields as analyses
The photometric end point possesses the been used to great advantage in titrations for the control of industrial processes and the
advantage over many other commonly used with EDT A and other complexing agents. routine analyses of air, water, soil, and agri-
Instrumentation end points in that the experimental data are Figure 7·20 illustratC5 the application of this cultural products.
Photometric titrations are ordinarily per- taken well away from the equivalence-point end point to the successive titration of Most clinical analyses are based upon pho-
formed with a spectrophotometer or a pho- region. Thus, the titration reactions need not bismuth(I1I)and copper(II). At 745 nm, neither tometric or spectrophotometric measurements
tometer that has been modified to permit have such favorable equilibrium constants as cation nor the reagent absorbs, nor does the employing 'ultraviolet or visible radiation.
insertion of the titration vessel in the light those required for a titration that depends more stable bismuth complex, which is formed Thus, an important component of most auto-
path. I , After the zero adjustment of the meter upon observations near the equivalence point in the first part of the titration; the copper matic instruments is a photometer or spectro-
scale has been made, radiation is allowed to (for example, potentiometric or indicator end complex, however, does absorb. Thus, the photometer: As will be pointed out later,
pass through the solution of the analyte, and points). For the same reason, more dilute s0- solution exhibits no absorbance until essen- however, au.tomatic systems may also employ
the instrument is adjusted by varying the lutions may be titrated. tially all of the bismuth has been titrated. atomic speCtroscopy and electroanalytical
source intensity or the detector sensitivity The photometric end point has been With the first formation of the copper methods to measure analyte concentrations.
until a convenient absorbance reading is ob- applied to all types of reactions.18 Most of complex, an increase in absorbance occurs. Two exaqtples of automatic systems are dis-
tained. Ordinarily, no attempt is made to the reagents used in oxidation-reduction titra- The increase continues until the copper equiv- cussed briefly in the paragraphs that follow.
measure the true absorbance, since relative tions have characteristic absorption spectra alence point is reached. Further reagent addi-
values are perfectly adequate for the purpose and thus produce photometrically detectable tions cause no further absorbance change.
Clearly, two well-defined end points result. Continuous-Flow Analyzer
of end-point detection. Data for the titration
are then collected without alteration of the 0.15
The photometric end point has also been In a continiJous analyzer, successive samples
instrument setting. adapted to precipitation titrations; here, the pass through the same system of tubes and
The power of the radiation source and the 0.12 suspended solid product has the effect of di- chambers. thus, samples must be isolated
response of the detector must be reasonably •. minishing the radiant power by scattering; from one another to avoid cross contamina-
constant during the period required for a ..
u
c 0.09 tit rations are carried to a condition of con- tion, and means must be provided for rinsing
photometric titration. Cylindrical containers
are ordinarily used, and care must be taken to
1« 0.06
stant turbidity. the system between samples. In the Technicon
AutoAnalyzer®, the earliest of the commer-
avoid any movement of the vessel that might cial instruments, movement of sample, rea-
0.03
alter the length of the radiation path. gents, and diluent through plastic tubing is
AUTOMATIC PHOTOMETRIC
Both filter photometers and spectrophotom- accomplished by means of a peristaltic pump.
0 AND SPECTROPHOTOMETRIC
eters have been employed for photometric Successive samples are isolated from one
4
ANALYSIS
titrations. The latter are preferred, however, 0 2 3 5 another by introducing bubbles of air into the
because their narrower bandwidths enhance mlO.l FEOlA One of the major instrumental developments tubing.
the probability of adherence to Beer's law. during the last two decades has been the Figure 7-21 is a schematic diagram of a
FIGURE 7-20 Photometric titra- appearance, from commercial sources, of auto- single-channel autoanalyzer used for the
tion curve of 100 ml of a solution matic analysis systems, which provide ana- analysis of one of the constituents of blood.
Application of Photometric Titretions that was 2.0 x 10-3 Fin Bi3+ and lytical data with a minimum of operator Here, samples are removed automatically and
Cu2+. Wavelength: 745 nm. (Re- intervention. The need for these systems has
Photometric titrations often provide more successively-from containers held in the rotat-
printed with permission from A. L. been greatest in clinical laboratories, where ing table sampler and are mixed with diluent
accurate results than a direct photometric Underwood, Anal. Chern., 26, 1322
analysis because the data from several mea- perhaps thirty or more analyses are routinely and air bubbles. The latter promote mixing
(1954). Copyright by the American used in large number for diagnostic and and serve t~ separate the sample from earlier
surements are pooled in determining the end Chemical Society.)
point. Furthermore, the presence of other screening purposes. The number of such
analyses required by modem medicine is
enormous; the need to keep their costs at a ,. For a description of some commereially available in-
•• See, for ewnple, the review: A. L. Underwod,
reasonable level is obvious. These two con- struments. see: R. H. Laessig. Anal. CIrnn. 43 (8), 18A
'7 Titration ftub and cells for use in a Spectronic 20 AdIlQ"",, in A/llJlyrieal Clrnnutry aM Inst~atiDn, ed.
siderations have led to a focus of effort (1971); and J. K. Foreman and P. B. Stockwell, AIIt-'c
spectropbotometer are available from tbe Kontes Manu- C. N. Reilley. New York: Intencience, t964, vol. 3. pp.
31-104. toward using automatic instruments for c1ini- Chemical A/llJI}"si•. New York: Wiley, 1975, Chapter 4.
facturin8 Corp, Vineland, NJ.
1. Sample photocell
2.Ref.~ephotoc~1
Figure 7-23, which is a cross-sectional view of teenth. When the rotor reaches a rotation rate
3. Flowcell the circular, plastic rotor of a centrifuge. The of about 3.50 rpm, reagent and samples in the
4. Light IOUrce
rotor has 17 dual compartments arranged ra- 17 compartments are mixed simultaneously
dially around the axis of rotation. Samples and carried: into individual cells located at
and reagents are pipetted automatically into the outer edge of the rotor; these cells are
16 of the compartments as shown; solvent equipped with horizontal quam windows.
and reagent are measured into the seven- Mixing is hastened by drawing air through

Inor. Uric Alk.


Ca~ Phos. Glu. BUN Acid Chal. T.P. Alb. T. Bill. Phos. LDH SGOT
Diffusable
constituonU PISS
through mombrone
to be picked up by
.............
Air bubbles
0
rng%

15
mg%P
10
mg%
500
rng%
100
mg%
12
mg%
500
gm%
10
gm%
8
mg% mU.lml

10 350
mU.lml

BOO
mU.lml

250
0
flowing reagent removed before
0
stream. colored solution 0
14
onten floweell. 0

where color
intensity is 0
monitored. 0 13
0
0
COLORIMETER 0
0
0 12

131' ...:
0
0
00 0
11

SImples recorded
as • series of peaks.

RECORDER

FIGURE 7-21 A single-c:hannel Technicon AutoAnalyzer System@. (Courtesy of Technicon


Instruments Corporation, Tarrytown, New York.) 50

40
and later ones. The diluted sample then Centrlfugel Fen Analyzer
passes into the dialyzer, which contains mem-
branes through which the small analyte ions A type of batch analyzer that is capable of
or molecules are free to diffuse into the rea- analyzing as many as 16 samples simultan- 3 30

gent stream. The residual large-protein eously for a single constituent is based upon 4

molecules of the blood remain in the diluent the use of a centrifuge to mix the sample with
stream and pass from the system to waste. the reagent and to transfer the mixture to a 3 2 100 2
The remainder of the system shown in Figure cell for photometric or spectrophotometric 2
7-21 is self-explanatory. analysis.20 The system is such that conversion 2
A system for routine blood analysis in- from one type of reagent to another is usually 50 10 50

cludes a module for diluting and partitioning easy.


the sample into several aliquots, each of The principle of the instrument is seen in
which passes through a separate channel simi- 0 0 0 0 0 0 0 o 0 0 0
lar to the one shown in Figure 7-21. The final Ca" lnor. Glu. BUN Uric Chol. T.P. T. Bill. Alk. LDH SGOT
Phos. Acid Phos.
output from a typical system is shown in 2.For a brief description of a typical system, see: R. L
Figure 7-22. The shaded areas show the range Coleman, w. D. Shults, M. T. Kelley, and J. A. Dean, FIGURE 7-22 Readout from a multichannel Technicon AutoAna-
of concentrations which are considered to be Amer. lAb, 3 (7~ 26 (1971); and C. D. Scali and C. A. Iyzer System®. (Courtesy of Technicon Instruments Corporation, Tarry-
normal for the population. Bunis, Anal. Chmt, 45 (3~ 327A (1973~ town, New York.)
the analysis. The centrifugal fast analyzer per-
mits the simultaneous determination of the
rates of 16 reactions under exactly the same
conditions; thus, 16 simultaneous enzyme
an~yses are feasible.

Photoacoustic spectroscopy is a recently


developed method for determining the ult~a-
violet and visible absorption spectra of soh~s
Light
from
or semisolids.21 Acquisition of spectra for thIS
monochromator type of sample by ordinary methods is ~sually
difficult at best because of light scattenng.
FIGURE 7-23 Rotor for a centrifugal fast
analyzer. (From R. L Coleman, W. D. Shults, The PhotoaeoUatic Effect
M. T. Kelley, and J. A. Dean, Amer. ~ .• 3(7~ Photoacoustic spectroscopy is bas,ed u~n an
26 (1971). With permission. Copyn~t ~971 absorption effect which was first 10vestlgated
by International Scientific Communications, in the 18805 by Alexander Graham Bell ~nd
Inc.) others. This effect is observed when a gas 10a
FIGURE 7-24 Block diagram ofa single-beam photoacoustic spectrometer
closed cell is irradiated with a chopped beam
with digital data processing. (Reprinted with permission from A. Rosencwaig,
of radiation of a wavelength that is absorbed
Anal. Chern., 47, 592A (1975). Copyright by the American Chemical Society.)
the mixtures. Radiation from an interference- by the gas. The absorbed radiation is ulti-
filter photometer or a spectrophotometer mately converted partially or complet~ly to
passes through the cells and falls u~n a kinetic energy of the gas molecules, ow1Ogto
photomultiplier tube. For each rotation, a non radiative relaxation of the excited absorb-
series of electrical pulses are produced, 16 for ing species. The kinetic energy of the .gas
the samples and 1 for the blank. ~n sitive microphone. The solid is then irradiated the components of a single-beam photo-
molecules gives rise to pressure fluctuatiOns
each of these pulses is a signal corresponcbng with a chopped beam from a monochroma- acoustic spectrometer. Figure 7-25 shows a
within the cell, which, because of the chop-
to the dark current. tor. The photoacoustic effect is observed typical application of the instrument. Here,
ping rate: are of acoustical freq~~ncies. and
The successive signals from the instr~ment can thus be detected by a senSitIve mIcro- provided the radiation is absorbed by the solid; photoacoustic spectra of smears of whole
are collected in the memory of a dedteated phone. The photoacoustic effect has been the power of the resulting sound is directly blood, blood cells freed of plasma, and he-
minicomputer for subsequent manipu~ti~n. related to the extent of absorption. Radiation moglobin extracted from the cells are shown.
used since the turn of the century for gas
Signal averaging can be employed to optimize reflected or scattered by the sample has no Conventional spectroscopy, even on dilute
analysis, and has recently taken on new im-
the signal-to-noise ratio. ., effect on the microphone and thus does not solutions of whole blood, does not yield satis-
portance for this purpose with the advent of
One of the most important apphcatlons of interfere. This latter property is perhaps the factory results because of the strong light-
tunable infrared lasers as sources.
the centrifugal fast analyzer is for the deter- most important characteristic of the method. scattering properties of the protein and lipid
In photoacoustic' studies of solids,. !he ,
mination of enzymes. Ordinarily, enzyme The source of the photoacoustic effect in molecules present. Photoacoustic spectros-
sample is placed in a closed cell contammg
analyses are based upon the catalytic effect of
the analyte upon a reaction that involves f~r-
air or some other nonabsorbing gas and a sen- ! solids appears to be similar to that in gases.
That is, nonradiative relaxation of the absorb-
copy clearly permits spectroscopic studies of
blood without the necessity of a preliminary
mation or consumption of an absorb1Og ing solid causes a periodic heat flow from the separation of large molecules.
species. Here, a calibration curve relating the .i For a review on applications, see: A. R~ig,
solid to the surrounding gas; the resulting Other applications which suggest them-
rate of appearance or disappearance of the AlIIIL CMnL, 47 (6~ S92A (1975); and OploaaJllSrIC S~- pressure fluctuations in the gas are then selves include the study of minerals, semi-
absorbing species as a function of ~e troscopy /Inti Detection, cd. Y. Pao, New York: AcademIC detected by the microphone. conductors, and the spots on thin-layer or
Press, 1977. Chapter 8. Figure 7-24 is a block diagram showing paper chromatograms.
concentration serves as the basIS for
(b) A solution containing both X and Y was found to have an absorb-
ance of 0.431 at 320 nm and 0.686 at 460 nm when measured in
0.5OO-cmcells. Calculate the molar concentration of each species
in the ~olution assuming both species obey Beer's law.
2. A. J. Muk~edkar and N. V. Deshpande [Anal. Chern., 35, 47 (1963)]
report on a simultaneous determination for cobalt and nickel based
upon a~~~tion by thei! respective 8-quinolinol complexes. Molar
absorptlvltles correspondmg to absorption maxima are: .

Wavelength
365 nm 700 nm
3529 428.9
3228 0.00
:
Red
blood ~culat~ the concentration of nickel and cobalt in each ~f the follow-
cells mg solutiOns on the basis of the accompanying data:

Absorbance, 1.0lk:m Cell


365 nm 700 nm
0.816 0.074
0.516 0.033

FIGURE 7-25 Photoacoustic


spectra of smears of blood and 3. Solutio.ns of P and Q individually obey Beer's law over a large con.
blood components. (Reprinted with centratlon range. Spectral data for these species, measured in I.oo-cm
permission from A. Rosencwaig, ceUs, are tabulated below.
Anal. Chem., 47, 592A (1975). Copy-
800 right by the American Chemical Absorbance, A
Society.)
l,nm 8.55 x 10"M P 2.37 x 10-4M Q
400 0.Q78 0.550
420 0.087 0.592
1. The following data were obtained for solutions of substance X and Y 440 0.096 0.599
(1.20-cm cells): 460 0.102 0.590
480 0.106
500
520
0.110
0.113
8:~~
0.433
C,M A at 320 nm A at 460 nm 540 0.116 0.343
560 0.126 0.255
4.60 x 10-3 0.965 0.106 580 0.170
1.213 0.170
5.11 X 10-3 0.094 600 0.264 0.100
620 0.326 0.055
640 0.359 0.030
660 0.373 0.030
680 0.370 0.035
700 0.346 0.063
(a) Plot an absorption spectrum (1.00-cm cells) for a solution that is Estimate the wavelength at which absorption by the indicator be-
7.76 x 1O-5M with respect to P and 3.14 x 1O-4M with respect comes independent of the pH (that is, the isosbestic point).
toQ. 7. Calculate the absorbance (1.00-cm cells) at 4SO nm of a solution in
(b) Calculate the absorbance (1.00-cm cells) at 440 nm of a solution which the total formal concentration of the indicator described in .
that is 200 x 1O-5M in P and 400 x 10-4M in Q. Problem 6 is 7.50 X 10-5 and the pH is (a) 4.670, (b) 5.217, and (c),
4. Use the data in the previous problem to calculate the quantities 5.866.
sought (1.5O-cm cells are used for all measurements). 8. Wha~ will be the_:bsor~ce at 595 nm (1.00-cm cells) of a solution
that IS 1.25 x 10 F WIth respect to the indicator in Problem 6 and
whose pH is (a) 5.451, (b) 6.213, and (c) 5.158?
Concentration, mol/liter
9. Several buffer solutions were made 1.25 x 1O-4F with respect to the
440 nm 620 nm
indicator in Problem 6; absorbance data (1.00-cm cells) at 4SO nm
0.862 0.450 and 595 nm were as follows:
0.313 0.799
0.276 0.347
0.897 0.350 Absorbance
0.495 0.816
0.202 0.297 Solution 4SO nm 595 nm
A 0.333 0.307
5. Absorptivity data for the cobalt and nickel complexes with 2,3- B 0.497 0.193
quinoxalinedithiol at the respective absorption peaks are as follows: C 0.666 0.121
D 0.210 0.375
Wavelength
Calculate the pH of each solution.
510 nm 656 nm
10. Construct an absorption spectrum for an 8.00 x 1O-5F solution of
36,400 1,240 the indicator in Problem 6 when measurements are made with
5,520 17,500 1.00-cm cells and

A 0.376-g soil sample was dissolved and subsequently diluted to ( ) [Hln] = 3


SO.Oml. A 25.Q-ml aliquot was treated to eliminate interferences; after a [In-]
addition of 2,3-quinoxalinedithiol, the volume was adjusted to
50.0 ml. This solution had an absorbance of 0.467 at 510 nm and (b) [Hln] =
[In-] I
0.347 at 656 nm in a l.OQ-cmcell. Calculate the respective percentages
of cobalt and nickel in the soil.
6. The indicator Hln has an acid-dissociation constant of 5.20 x 10-6 at
() [Hln] =!
c [In-] 3
ordinary temperatures. The accompanying absorbance data are for
7.50 x 10- 5 F solutions of the indicator measured in 1.00-an cells in II. The chelate ZnQ~ - exhibits a maximum absorption at 480 nm. When
strongly acidic and strongly alkaline media .. the chelating agent is present in at least a fivefold excess, the absorb-
, ance depends only upon the formal concentration of Zn(lI) and
Absorbance obeys Beer's law over a large range. Neither Zn2+ nor Q2- absorbs at
Absorbance
480 nm.
;',hm pH = 1.00 pH = 13.00 )., nm pH = 1.00 pH = 13.00 A solution that is 2.30 x 10-4F in Zn2+ and 8.60 x 1O-3F in Q
0.119 0.324 has an absorbance of 0.690 when measured in a 1.00-cm cell at
420 0535 O.OSO 550
570 0.068 0.352 480 nm. Under the same conditions, a solution that is 2.30 x 1O-4F
445 0.657 0.068
4SO 0.658 0.076 585 0.044 0.360 in Zn2+ and 5.00 x 1O-4F in Q2- has an absorbance ofO.S40. Calcu-
455 0.656 0.085 595 0.032 0.361 late the numerical value of K for the process.
470 0.614 0.116 610 0.019 0.355
0.353 650 0.014 0.284 Zn2+ +2Q2-~ ZnQ~-
510 0.223
12. What colored glass filter should be employed for a photometric
analysis based upon
(a) the blue color of the ammonia complex of Cu(II)?
(b) the red SCN- complex of Fe3+?
(c) the yellow complex formed when H202 is added to a solution
containing Ti(IV)?
4
13. Derive titration curves for 50.0 ml of 2.00 x 10- F A with 0.001ooF

M if the reaction is
M+A~MA
Assume that the path length of the cell is 5.5 em and that the equilib-
rium is forced far to the right after the following additions of rea-
gent: 2.00.4.00.6.00. 14.00. 16.00. and 18.00 ml. Assume the following
molar absorptivities for the three species:

11M SA I1MA

0.0 18.6 6530


(a)
762 0.0 6530
(b)
8.2 6530 0.0
(c) 0.0 36.5
(d) 6680
6680 0.0 1430
(e)
the abscissa in this chart is linear in units 2.0 X 1013 Hz). A scale in terms of em -1 is
The infrared region of the spectrum encom- trum of an organic compound represents one of reciprocal centimeters; other instruments, often referred to as a frequency scale; the stu-
passes radiation with wavenumbers ranging of its unique physical properties; with the ex- particularly older ones, employ a scale that dent should bear in mind that this termi-
from about 12,800 to 10 em-lor wavelengths ception of optical isomers, no two com- is linear in wavelength (an easy way to nology is not strictly correct, the wavenumber
from 0.78 to 1000 JlIIl.1 From the standpoint pounds, in theory, absorb in exactly the same remember how to convert from one scale to being only proportional to frequency.
of both application and instrumentation, it is w~.
In addition to its application as a qualt-
. the other is the relationship em
-I x JlIIl=
Finally, it should be noted that the hori-
convenient to subdivide the spectrum into 10,000). For comparison, a wavelength scale zontal scale of Figure 8-1 changes at 2000
near, middle, and far-infrared radiation; the tative-analytical tool, infrared spectrophoto- has been added on the upper axis of the em - I, with the units at higher wavenumbers
limits of each are shown in Table 8-1. The metry is finding increasing use for quantitative original chart paper shown in Figure 8-1. being represented by half the linear distance
large majority of analytical app~ications ~re analysis as well. Here, its great advantage Several brands of chart paper contain both of those at lower. This discontinuity is intro-
confined t9 a portion of the middle regton lies in its selectivity, which often makes wavelength and wavenumber scales; ob- duced for convenience, since much useful qual-
extending from 4000 to 670 em -lor 2.5 to 15 possible the quantitative estimation of an viously, only one can be linear. itative infrared data appear at wavenumbers
pm. analyte in a complex mixture with little or
The preference for a linear wavenumber smaller than 2000 em - I.
Infrared:spectroscopy finds widespread ap- no prior separation steps. The most impor- scale is based upon the direct proportionality Dipole Changes During Vibrations and Ro-
plication to qualitative and quantitative anal- tant analyses of this type have been of atmos- between this quantity and both energy and tations. Most electronic transitions require
yses.) Its most important. use has been for pheric pollutants from industrial processes. frequency; the frequency of the absorbed ra- energies in the ultraviolet or visible regions;
the identification of organIC compounds, be- diation is, in turn, the molecular vibrational absorption of infrared radiation is thus
cause their spectra are generally complex and frequency actually responsible for the absorp- confined largely to molecular species for
provide numerous maxima a~ minima that THEORY OF tion process. Frequency, however, is seldom ir which small energy differences exist between
can be employed for companson purposes INFRARED ABSORPTION ever employed as the abscissa, probably be- various vibrational and rotational states.
(see Figu~e 8-1). Indeed, the infrared spec- cause of the inconvenient size of the unit (for In order to absorb infrared radiation, a
Introduction example, the frequency scale of the plot in molecule must undergo a net change in dipole
Figure 8-1 would extend from 12 x 1014 to moment as a consequence of its vibrational or
A typical infrared spectrum, obtained with .a
I Until receiltly. the unit' of wavelength that is 10-' m
double-beam recording spectrophotometer, IS
was called the micron, I'; it is now more properly tenned shown in Figure 8-1. In contrast to most
lhe micrometer. I'm. ultraviolet and visible spectra, a bewildering
J For detailed discussion of infrared spectroscopy. see: array or maxima and minima are observed.

J~'.
N. B. Collhup. L. N. Daly. and S. E. Wiberley. IlIfrotluc- 2.5 3.0 4.0 50
Spectral Plots. The plot shown in Figure 100 I
1/011 10 Infra Rat •••• R-.. Sp«trtnCopy. 2d ed. New
York: Academic PreIs, 1974; K. Nakanishi, l'lfrared 8-1 is a reproduction of the record~r output
of a widely used commercial infrared spectro- 80
Absorption Sp«trMCop)'. San Francisco: Holden-Da~. I~
1977: N. L. Alpert. W. E. Keiser. and H. A. SzymanskI. photometer. As is ordinarily the case, the \ 1,/\
Theor)" and Practice of Infrawl Sp«tros<:opy. 2d ed. New ordinate is linear in transmittance; it should
80
V ,filii
York: Plenum Press. 1970; R. T. Conley. Infrared S""c-
Iroscopy. 2d ed. Boston: Allyn and Bacon. 1972; and
Infrared and Raman S""clros<:op)'.cds. E. G. Brame and
be noted, however, that the chart paper of
some manufacturers also contains a nonlinear
ii
"";60
i
70
uj+tIii I
-- ----I' ...
J. G. Grasselli. New York: Marcel Dekker, 1977. absorbance scale ror reference. Note also that .; 50
i
,-- .~

~ i
~40 ! I !
I-
30 lJ-1-
I
INFRARED SPECTRAL REGIONS 20
!
10
Wa~eleogth ().) Wavel\lllllber (q) Frequency (v)
Region ~nge, JlIIl Range, em - I Ringe, Hz 0 IIII
4000 3600
I,
3200 2800
1111 2400 2llOO
Near 0.78 to 2.5 12800 to 4000 3.8 X 1014 to 1.2 X 1014
Middle l,S to so' :1000 to 200 1.2 X lO'4 to 6.0 X 1012
Far 50 to 1000 200 to 10 6.0 X 1012 to 3.0 X 1011 FIGURE 8-1 Infrared absorption spectrum of a thin polystyrene film recorded with a modem
Most used 2.5 to 15 4000 to 670 1.2 x 1014 to 2.0 X 1013 IR spectrophotometer. Note that the abscissa scale changes at 2000 em -1.
rotational motion. Only under these circum-
stances can the alternating electrical field of
the radiation interact with the molecule and
cause changes in the amplitude of one of its
motions. For example, the charge distribution
around a molecule such as hydrogen chloride
of a gas usually consists of a series of closely
spaced lines, because there are several rota-
tional energy states for each vibrational state.
On the other hand, rotation is highly re-
stricted in liquids and solids; in such samples,
discrete vibratioool-rotationallines disappear,
yy Symmetric
(al Stretching vibrations
Asymmetric
proportional to the displacement
law). That is,

F = -ky
(Hooke's

(8-1)
where F is the restoring force and k is the
is not symmetric, the chlorine having a higher leaving only somewhat broadened vibrational force constant, which depends upon the stiff-
electron density than the hydrogen. Thus, peaks. Our concern is primarily with the ness of the spring. The negative sign indicates
hydrogen chloride has a significant dipole
moment and is said to be polar. The dipole
moment is determined by the magnitude of
the charge difference and the distance be-
tween the two centers of charge. As a
spectra of solutions, liquids, and solids, in
which rotational effects are minimal.
Types of Molecular Vibrations. The rela-
tive positions of atoms in a molecule are not
exactly fixed but instead fluctuate con-
X In-plane scissoring
that F is a restoring force.
Potential Energy of a Harmonic Oscilla-
tor,. The potential energy of the mass and
sprmg can be considered to be zero when the
mass is in its rest or equilibrium position As

'y' 'y-
hydrogen chloride molecule vibrates longi- tinuously as a consequence of a multitude of the spring is compressed or stretched, h~w-
tudinally, a regular fluctuation in dipole different types of vibrations. For a simple dia- ever, the potential energy of this system in-
moment occurs, and a field is established tomic or triatomic molecule, it is easy to creases by an amount equal to the work
which can interact with the electrical field as- define the number and nature of such vi- required t? displace the mass. I( for example,
sociated with radiation. If the frequency of brations and relate these to energies of the mass IS m~ved from some position y to
absorption. An analysis of this kind becomes Out..,fopl.,. wagging Out..,f.plane twisting
the radiation matches a natural vibrational (y + dy~ the work and hence the change in
frequency of the molecule, there occurs a net difficult if not impossible for polyatomic Ibl Bending~ibrations potential energy E is equal to the force F
transfer of energy that results in a change in molecules, not only because of the large FIGURE 8-2 Types of molecular vibra- times the distance dy. Thus,
the amplitude of the molecular vibration: number of vibrating centers, but also because tions. Note: + indicates motion from plane
absorption of the radiation is the conse- interactions among several centers occur and toward reader; - indicates motion from
quence. Similarly, the rotation of asymmetric must be taken into account. plane ~way from reader.
molecules around their centers of mass results Vibrationsfall into the basic categories of
in a periodic dipole fluctuation; again, inter- stretching and bending. A stretching vibration
action with radiation is possible. involves a continuous change in the inter-
No net change in dipole moment occurs atomic distance along the axis of the bond
up next. Finally, the effects of vibrational
during the vibration or rotation of homo- between two atoms. Bending vibrations are Integrating between the equilibrium position
interactions in molecular systems will be
nuclear species such as Oz , N z, or Oz; con- characterized by a change in the angle be- discussed. (y = 0) and y gives
sequently, such compounds cannot absorb in tween two bonds and are of four types: scis-
the infrared. soring, rocking, wagging, and twisting. The ,Ii ,
Rotational Transitions. The energy required
to cause a change in rotational level is minute
various types of vibrations are shown
schematically in Figure 8-2.
Mechanical Model of
Stretching Vibrations
J
o
dE = k 1
0
ydy

and corresponds to radiation of 100 pm and All of the vibration types shown in Figure E =!kl (8-3)
~e ~haracteristics of an atomic stretching
greater ( < 100 em - I). Because rotational levels 8-2 may be possible in a molecule containing
vlbr~tlon can be approximated by a me-
are quantized, absorption by gases in this far- more than two atoms. In addition, interaction The .pote~tial:energy curve for a simple har-
chamcal model consisting of two masses con-
infrared region is characterized by discrete, or coupling of vibrations can occur if the vi- ?tome oscl~latlon, derived from Equation 8-3,
nected by a spring. A disturbance of one of
well-defined lines. In liquids or solids, intra- brations involve bonds to a single central IS plotted In Figure 8-3a. It is seen that the
molecular collisions and interactions cause these .mass~ alo.ng the axis of the spring re-
atom. The result of coupling is a change in potential energy is a maximum when the
sults In a Vibration called a simple harmonic
broadening of the lines into a continuum. the characteristics of the vibrations involved. motion. spring is stretched or compressed to its
Vibrational-Rotational Transitions. Vibra- In the treatment that follows, we shall first ma~imum amplitude A, and decreases para-
tional energy levels are also quantized, and Let us first. consider the vibration of a
consider isola~ vibrations with a simple bohcally to zero at the equilibrium position.
~ attached to a spring that is hung from
the energy differences between quantum mechanical model called the harmonic oscil- Vibrational Frequency. The motion of the
states correspond to the readily accessible re- an ~o.vable object (see Figure 8-3a). If the
lator. Modifications to the theory of the har- mass as a function of time t can be deduced
",lass ISd~?laced a distance y from its equilib-
gions of the infrared from about 13,000 to 675 monic oscillator, which are needed to as follows. Newton's law states that
em-I (0.75 to 15 pm). The infrared spectrum describe a molecular system, will be taken rium ~slhon by a~plication of a force along
the aXIs of the spring, the restoring force is
where m is the mass and a is its acceleration. Thus, the vibrational frequency for such a
But acceleration is the second derivative of system is' given by
distance with respect to time:
.
0= df
d1y
v.=-
1
27£
A
- =-
11
1
27£
k(ml + m2)
--~
ml ml
(8-9)

Substituting these expressions into 8-1 gives Molecular Vibrations. The approximation
is ordinarily made that the behavior of a
d1y
m dt2 = -ky (8-4) molecular vibration is analogous to the me.·
chanical model just described. Thus, the
One solution (but not the only one) to Equa- frequency of the molecular vibration is
tion 8-4 can be shown by substitution to be calculated from Equation 8-9 by substituting:
the masses of the two atoms for ml and m2;"
y = A sin( ~t) (8-5)
the quantity k becomes the force constant for:
the chemical bond, which is a measure of its:
stiffness (but not necessarily its strength).
where A is the amplitude of the vibration, a
-A
constant that is equal to the maximum value
_Ditplacementy-
ofy. Quantum Treatment of Vibrations
Equation. 8-5 is analogous to the sinusoi-
dal function described by Equation 2-16, page Rannonk Oscillaton. The equations of
12. That is, ordinary mechanics, such as we have used"
y = A sin 27£Vt thus far, do not completely describe the be- ~
havior of particles with atomic dimensions. I
\1
\
Comparison of the two equations reveals that For example, the quantized nature of molec-,

Jfn
\ ular vibrational energies (and of course"
\
\ t = 2xvt (8-6) other atomic and molecular energies as well)
\
\ does not appear in these equations. It is
\
\ This equation can be rearranged to give possible, however, to employ the concept of
\
the simple harmonic oscillator for the develop-
Jfn
1
\
\ v. =~ (8-7) ment of the wave equations of quantum
1
1 mechan~. Solutions of these equations for
1
1 where v.. is the natural frequency of the potential energies are found when
1
\ mechanical oscillator. While it is dependent
\
\
\
upon the force constant of the spring and the £= (v + !)!!- ~
2 27£\/ P
(8-10)
\ mass of the attached body, the natural fre-
quency is independent of the energy imparted . where v, the ribrational quantum number, can
to the system; changes in energy merely result take only positive integer values (including
in a change in the amplitude A of the zero). Thus, in contrast to ordinary me-.
(b) vibration. chanics where vibrators can have any positive'
FIGURE 8-3 Potential energy diagrams. Curve I, The equations just developed can be potential energy, quantum mechanics requires .
harmonic oscillator. Curve 2, anharmonic oscillator. readily modified to describe the behavior of a that only certain discrete energies be assumed 1
system consisting of two masses ml and ml by a vibrator. '
connected by a spring. Here, it is only neces- It is of interest to note that the term
sary to substitute the reduced mass 11 for the (.jiC[iJ)/27£ appears in both the mechanical
single mass m where and the quantum equations; by substituting
Equation 8-9 into 8-10, we find
£ = (v + 1)111'.. (8-11)
em/s, and ml and mi are the masses of atoms Selection Rules. As gjveJl by Equations accounts for the success of the approximate
where v•• is the vibrational frequency of the
1and2ing. 8-11 and 8-12,' the energy for a transition methods described.
mechanical model.
Equation 8-14 and infrared measurements from energy level 1 to 2 or from level 2 to 3 Anharmonicity leads to deviations of two
We now assume that transitions in vibra-
have been employed to evaluate the force should be identical to that for the 0 - 1 tran- kinds. At higher quantum numbers, AE be-
tional energy levels can be brought about by
constants for various types of chemical bonds. sition. Furthermore, quantum theory indi- com~ smalle~ (see Figure 8-3b). and the
radiation, provided the energy of the radi-
Generally, k has been found to lie in the cates that the only transitions that can take selection rule IS not rigorously followed' as a
ation exactly matches the difference in energy
range between.3 x 105 and 8 x 105 dynes/em place are those in which the vibrational quan- result, transitions of Av = ±2 or ±3 ar~ ob-
levels AE between the vibrational quantum
for most single bonds, with 5 x 105 serving as tum number changes by unity; that is, the served. Such transitions are ,responsible for
states (and provided also that the vibration
a reasonable average value. Double and triple so-called selection rule states· that Av = ± 1. the appearance of overtone lines at frequencies
causes a fluctuation in dipole). This difference
bonds are found by this same means to have Since the vibrational levels are equally approximately twice or three times that of the
is identical between any pair of adjacent levels,
force constants of about two and three times spaced, only a single absorption peak should fundamental line; the intensity of overtone
since v in Equations 8-10 and 8-11 can
this value, respectively. With these average be observed for a given transition. absorption is frequently low,: and the peaks
assume only whole numbers; that is,
experimental values, Equation 8-14 can be Anharmonic Oscillator. Thus far, we have may not be observed. ,
used to estimate the wavenumber of the considered the classical and quantum me- Vibrational spectra are further compli-
AE=hv =~ ~ (8-12)
•• 21t ~ ~ fundamental absorption peak (the absorption chanical treatments of the harmonic oscil- cated by the fact that two 'different vibra-
peak due to'the transition from the ground lator. The potential energy of such a vibrator tions in a molecule can ilJteract to give
At room temperature, the majority of mole- changes periodically as the distance between
cules are in the ground state (v = 0); thus, state to the first excited state) for a variety of absorption peaks with frequencies that are
bond types. The following example demon- masses fluctuates (Figure 8-3a). From qualita- approximately the sums or differences of their
from Equation 8-11,
strates such a calculation. tive considerations, however, it is apparent fundamental frequencies. Again, the intensi-
Eo = thv•• that this description of a molecular vibration ties of combination and difference peaks are
is imperfect. For example, as the two atoms' generally low. '
Promotion to the first excited state (v = 1) EXAMPLE
Calculate the approximate wavenumber approach one another, coulombic repulsion
with energy
and wavelength of the fundamental ab- between the two nuclei produces a force that
EI = fhv •• acts in the same direction as the restoring Vibrational Modes
sorption peak due to the stretching vibration
force of ~he bond; thus, the potential energy Il is ordinarily possible to deduce the number
requires radiation of energy of a carbonyl group C-o. can be expected to rise more rapidly than the and kinds of vibrations in simple diatomic
(fhv•• - thv ••) = hv•• The force constant for a double bond has harmonic approximation predicts. At the and triatomic molecules and whether these
an approximate value of 1 x 106 dynes/em. other extreme of oscillation, a decrease in vibrations will lead to absorption. Complex
The frequency of radiation v that will bring The masses of the carbon and oxygen atoms the restoring foree, and thus the potential molecules may contain several types of atoms
about this change is identical to the classical energy, OCCUrsas the interatomic distance as well as bonds; for these, the multitude of
are approximately 12/6.0 x IOU and 16/6.0 x
vibration frequency of the bond v••. That is, IOU, or 2.0 x 10-23 and 2.7 x 10-13 g per approaches that at which dissociation of possible vibrations gives rise to infrared
atom. Substituting into Equation 8-14, we atoms takes place. spectra that are difficult, if not impossible. to
h In theory, the wave equations of quantum
Eradia.lon= hv = AE = hv•• = 2 1t~ obtain analyze.
mechanics permit the derivation of more The number of possible vibrations in a
1 nearly correct potential-energy curves for polyatomic molecule can be calculated as fol-
(1= 2 x 3.14 x 3 x 1010 molecular vibrations. Unfortunately, how- lows. Three coordinates are needed to locate
v= v =
••
-.!.-
21t ~ ~
fE (8-13)
1 x 10 (2.0 + 2.7) x 10-
ever, the mathematical complexity of these a point in space; to fix N points requires a set
equations precludes their quantitative appli- of three coordinates for each for a total of 3N.
2.0 x 2.7 X 10-46 cation to all but the very simplest of systems. Each coordinate corresponds to one degree of
If we wish to express the radiation in It is qualitatively apparent, however, that the
= 1.6 X 103 em - I freedom for one of the atoms in a polyatomic
wavenumbers, curves must take the anharmonic form shown molecule; for this reason, a i molecule con-
!if in Figure 8-3b. These curves depart from har- taining N atoms is said to have 3N degrees of
(1=1... ~=_1 k(ml + m2) (8-14) A = 1.6 X 103 = 6.3 JlI1l monic behavior by varying desrees, de- freedom.
21tC ~ ~ 21tC mlm2
pending upon the nature of the bond and the In defining the motion of a molecule, we
where (1 is the wavenumber of an absorption The carbonyl stretching band is found ex- atoms involved. Note, however, that the har- need to consider: (1) the motion of the entire
peak in em - I, k is the force constant for the perimentally to be in the region of 5.3 to 6.7 monic and anharmonic curves are nearly molecule through space (that is, the trans-
bond in dynes/em, c is the velocity of light in JlI1lor 1500 to 1900 em -I. alike at low potential energies. This fact lational motion of its center of gravity); (2)
the rotational motion of the entire molecule Vibrlltlonal Coupling of the coupled vibrations is symmetric and Since the central atom is not in line with the
around its center of gravity; and (3) the the other is asymmetric. other two, a symmetric stretching vibr~tion
The energy of a vibration, and thus t.he
motion of each of its atoms relative to the wavelength of its absorption peak, may be ro- will produce a ch~ge in dipole and will thus
other atoms (in other words, its individual fluenced (or coupled) by other vibrators in be responsible for infrared absorption. For
vibrations). ~finition of translational motion the molecule. A number of factors that example, stretching peaks at 3650 and 3760
requires three coordinates and uses up three influence the extent of such coupling can be The symmetric vibration causes no change in em-I (2.74 and 2.66 pm) are observed for the
degrees of freedom. Another three degrees of identified. dipole, since the two oxygen atoms simulta- symmetric and asymmetric vibrations of the
freedom are needed to describe the rotation 1. Strong coupling between stretching neously move away from or toward the central water molecule. Only one component to
of. the molecule as a whole. The remaining vibrations occurs only when there is an atom carbon atom. Thus, the symmetric vibration the scissoring vibration exists for this non-
(3N - 6) degrees of freedom involve inter- common to the two vibrations. is infrared-inactive. One oxygen approaches linear molecule, since motion in the plane of
atomic motion, and hence represent the 2. Interaction between bending vibrations the carbon atom as the other moves away the molecule constitutes a rotational degree
number of possible vibrations within the requires a common bond between the vibrat- during asymmetric vibration. As a con- of freedom. For water, the bending vibration
molecule. A linear molecule is a special case ing groups. sequence, a net change in charge distribution causes absorption at 1595 em-I (6.27 pm).
since, by definition, all of the atoms lie on a 3. Coupling between a stretching and a occurs periodically; absorption at 2330 em-I The ditTerence in behavior of linear and
single, straight line. Rotation about the bond bending vibration can occur if the stretching results. nonlinear triatomic molecules (two and three
aXis is not possible, and two degrees of bond forms one side of the angle that varies The remaining two vibrational modes of absorption peaks, respectively) illustrates how
freedom suffia: to describe rotational motion. in the bending vibrati9n. carbon dioxide involve scissoring, as shown infrared-absorption spectroscopy can some-
Thus, the number of vibrations for a linear 4. Interaction is greatest when the coupled below. times be used to deduce molecular shapes.
molecule is given by (3N - 5). Each of the groups have individual energies that are ap- Coupling of vibrations is a common phe-
(3N - 6) or (3N - 5) vibrations is called a proximately equal. nomenon; as a result, the position of an
nOrmal mode. 5. Little or no interaction is observed be- absorption peak corresponding to a given
I For each normal mode of vibration, there tween groups separated by two or more o~ organic functional group cannot be specified
e~ists a potential-energy relationship such as bonds. exactly. For example, the C-O stretching
that shown by the solid line in Figure 8-3b. 6. Coupling requires that the vibrations be The two bending vibrations are the resolved frequency in methanol is 1034 em - I (9.67
The same selection rules discussed earlier of the same symmetry species.3 components (at 90 deg to one another) ofthe pm); in ethanol it is 1053 em-I (9.50 pm);
apply for each of these. In addition, to the As an example of coupling etTects, let us bending motion in all possible planes around and in methylethylcarbinol it is 1105 em-I
extent that a vibration appro.ximates har- consider the infrared spectrum of carbon the bond axis. The two vibrations are identi- (9.05 pm). These variations result from a cou-
monic behavior, the ditTerenceS between the dioxide. If no coupling occurred between the cal in energy and thus produce but one peak pling of the C-O stretching with adjacent
energy levels of a given vibration are two C-O bonds, an absorption peak would at 667 an - I. (Quantum energy differences C-C or C-H vibrations.
the same; that is, a single absorption peak be expected at the same wavenumber as the that are identical, as these are, are said to be While interaction effects may lead to un-
should appear for each vibration in which peak for the C-O stretching vibration in an degenerate. ) certainties in the identification of functional
there is a change in dipole. aliphatic ketone (about 1700 em - I, or 6 pm; It is of interest to compare the spectrum groups contained in a compound, it is this
In fact, however, the number of normal see example, p. 216). Experimentally, carbon of carbon dioxide with that of a nonlinear, very etTect that provides the unique features
modes does not necessarily correspond ex- dioxide exhibits two absorption peaks, the triatomic molecule such as water, sulfur di- of an infrared-absorption spectrum that are
actly to the number of observed absorption one at 2330 em -I (4.3 pm) and the other at oxide, or nitric oxide. These molecules have so important for the positive identification of
peaks. The number of peaks is frequently less 667 em -I (IS pm). (3 x 3 - 6), or 3, vibrational modes which a specific compound.
l!ecause: (I) the symmetry of the molecules is Carbon dioxide is a linear molecule and take the following forms:
such that no change in dipole results from a thus has four normil1 modes (3 x 3 - 5). Two
particular vibration; (2) the energies of two or stretching vibrations are possible; further-
~ore vibrations are identical or nearly identi- more, interaction between the two can occur INFRARED
Cal; (3) the absorption intensity is so low that since the bonds involved are associated with a ~ INSTRUMENT COMPONENTS
as to be undetectable by ordinary means; or common carbon atom. As may be seen, one
(4) the vibrational energy is in a wavelength Infrared spectrophotometers have the same
region beyond the range of the instrument. As basic components as do instruments used for
we have pointed out, additional peaks arise • For a discussion of symmctry operations and symmetry the study of absorption in the ultraviolet and
from overtones as well as from combination species, see: R. P. Bauman, Absorption Spectroscopy. visible regions of the spectrum (see Figure
New York: Wiley, 1962, Chapter 10. 5-lb, p. IS).
or ditTerence frequencies.
PRINCIPLES OF INSTRUMENTAL ANALYSIS

red region (15 to 40 pm or 670 t~ 250 em-I~ synchronized with the recorder pen so that its source output. This feature permits the detec-
Wevelength Selection position gives a measure of the relative power
while lithium fluoride is useful m the near- tor to discriminate between the signal from
Infrared wavelength selection can be accom- infrared region (1 to 5 pm or 10,000 to 2000 of the two beams and thus the transmittance
the source and signals from extraneous radia-
plished by means of interference filters, -I). Many spcctrophotometers are de- of the sample.
tion, such as infrared emission from various
prisms, or gratings. ~ed for convenient prism interchange. Many instruments also employ a beam bodies surrounding the detector. Low chop-
Interference Filters. Interference ~lters ~nfortunatelY, all common infrared-tra~- attenuator in conjunction with the sample ping rates are demanded by the slow response
(p. 135) for quantitative infrared analYS15 of mitting materials except quartz are. easl1y beam for the purpose ofmaJcing the 100% T times of most infrared detectors.
specific compounds are availab~e commer-, adjustment.
scratched and are water-soluble. Desiccants Infrared instruments are generally of the
cially.· For example, a filte~ With a trans- or heat are required to protect them from the null type with the beam being attenuated by a
mission maximum of 9.0 pm 15offered for the
condensation of moisture. Infrered Source., comb or an absorbing wedge.
determination of acetaldehyde; one with ~ Re8ection Gratings. Reflection gratings Semple Contlliners, end Detectors Several dozen infrared instruments, rang-
maximum at 13.4 pm is suggested for O-dl-
offer a number of advantages as dispersing Details regarding the nature and properties of ing in cost from $3000 to more than $30,000,
chlorobenzene; and a filter exhibiting a trans-
elements for the infrared region and are re- infrared sources and detectors are found in are available from various instrument manu-
mission peak at 4.5 Jim is employed for the
placing prisms as a consequence. Inherentl.y Chapter 5. Discussion of infrared cells is facturers. In general, the optical design of
determination of nitrous oxide. In general,
better resolution is possible beca~e ther~ 15 deferred to the section in this chapter that these instruments does not differ greatly from
these filters have effective bandwidths of about
less loss of radiant energy than m a prlSm describes sample handling. the ultraviolet-visible spectrophotometers dis-
1.5% of their peak wavelength.
system; thus, narrower slit widths can be cussed in the previous chapter except that
Fitter Weqes. Wilks Scientific Co~ra-
realized. Other advantages include more the sample and reference compartment is
tion also offers an instrument that provides
nearly linear dispersion and r:esis~nce to always located between the source and the
narrow and continuously variable bands of
attack by water. l.n infrared gratmg 15usually monochromator in infrared instruments. This
infrared radiation by means of three filter
constructed from glass or plastic that has A large number of infrared instruments for arrangement is desirable because the mono-
wedges (p. 136) formed as segments of a
been coated with aluminum. qualitative applications are available com- chromator then removes most of the stray
circle. The ranges of the three segments are
The disadvantage of a grating lies in the mercially. In addition, photometers and radiation generated by the sample, the refer-
2.5 to 4.5, 4.5 to 8, and 8 to 14.5 Jim. The
greater amounts of scattered radiation and spectrophotometers designed specifically for ence, and the cells. In ultraviolet-visible spec-
wedges are mounted on a wheel which can be
the appearance of radiation of other spectral quantitative analysis have recently become troscopy, stray radiation from these sources is
rotated to provide the desired wavelength. at
orders. In order to minimize these eff~ts, available from instrument manufacturers. seldom serious. Location of the cell compart-
the slit. The effective bandwidth of the device
gratings are blazed to cona:ntrate the radia- ment between the monochromator and the
is about 1.5% of the wavelength that is trans-
tion into a single order. Fl1te~ (a~d occ::a- Infrered detector of such instruments, however, has
mitted to the slit. sionaIly prisms) are used in conjunctIon WIth Speetrophotometers the advantage of protecting the sample from
PriImL Several materials have been used
the grating to minimize these problems. for Qulllftetive A••••yal. the full power and wavelength range of the
for prism construction. Quartz is employed
ultraviolet source; thus, photochemical de-
for the near-infrared region (0.8 to 3 Jlm~ Because of the complexity of infrared spectra, composition is less likely.
even though its dispersion characteristics for Seem Attenuetion a recording instrument is required for qualita- Typical Instrument DesiglL Figure 8-4
this region are far from ideal. It absorbs tive work. As a consequence, all commercially
Infrared instruments are generally of a null shows schematically the arrangement of com-
strongly beyond about 4 JlID p500 em - I). type. in which the power of the reference available instruments of this type are of ponents in a typical infrared spectrophotom-
Crystalline sodium chlonde 15 .the. m~t beam is reduced, or attenuated, to match that double-beam design. An added reason for em- eter. Note that three types of systems link
common prism material; its dispersIon 15high of the beam passing through the sample. ploying two beams is that this design is less the components: (1) a radiation linkage in-
in the region between 5 and 15 Jim (2000 and Attenuation is accomplished by imposing a demanding in terms of performance of source dicated by dashed lines; (2) a mechanical link-
670 em - I) and is adequate to 2.5 pm (~ and detector than the single-beam arrange-
device that removes a variable fraction of the age shown by thick dark lines; and (3) an
em -I). Beyond 20 Jim (500 em - 1 ), sodIum ment (see p. 185). This property is important
reference beam. The attenuator commonly electrical linkage shown by narrow solid lines.
chloride absorbs strongly and cannot be used. because of the low energy of infrared radia-
takes the form of a fine-toothed comb: the Radiation from the source is split into two
Crystalline potassium bromide and cc:sium teeth of which are tapered so that a bnear tion, the low stability of sources and de- beams, half passing into the sample-cell com-
bromide prisms are suitable for the far-mfra- relationship exists between the later~l move- tectorS, and the need for large signal partment and the other half into the reference
ment of the comb and the decrease m power amplification. area. The reference beam then passes through
of the beam. Movement of the comb occurs All commercial infrared spectrophotom- an attenuator and on to a chopper. The chop-
when a difference in power of the two beams eters incorporate a low-frequency chopper per consists of a motor-driven disk that alter-
• Wilks Scienlific Corporation. Soulh Norwalk, CT.
06856. is sensed by the detector. This movement is (5 to 13 cycles per minute) to modulate the nately reflects the reference beam or transmits
by a thermocouple. The ac output of the sample. The resulting peak profile is thetl
latter, after amplification, drives the reference compared with profiles of pure compounds
attcnuator to a null position; the pen drive is stored in the comput~r memory to provide
coupled mechanically to this movement. positive or tentative identification of the
Instrument Automation. As· with ultra- analyte. The memory bank of one of these in-
violet and visible instruments, newer infrared struments contains 2000 compounds; the other
instruments usually contain microprocessor holds 4800. The systems cost about $25,000.
systems for partial automation of the measure-
ment process. In the less-expensive instru-
ments, automation includes slit programs, Instruments for Quantitative Work
scan-time adjustments, calibration modes for For two reasons double-beam recording
checking the frequency scale, and gain adjust- spectrophotometers are, in principle, less sat-
ment for survey scans. More sophisticated in- isfactory for infrared quantitative work than
struments offer push-button control of the single-beam instruments. First, the double-
normal scanning parameters, a visual display beam instrument, with its more complex elec-
ofsclected nonstandard parameters, automat- tronic and switching devices, is inherently
ic means for superimposing spectra, and more noisy. Second, the zero-transmittance
computer storage of spectra for subsequent calibration and low-transmittance measure-
signal averaging. ments with most double-beam instruments
Computer Search Systems. Two companies are subject to an error arising from their opti-
FIGURE 8-4 Schematic diagram of a double-beam spectro-
have recently offered computer search sys- cal null design. The zero-percent transmit-
photometer. Heavy dark line indicates mechanical linkage; light
tems. Both are ftoppy.<fisk data storage sys- tance setting for such an instrument is
line indicates electrical linkage; dashed line indicates radiation
tems that attach to existing spectrophotom- obtained by blocking the sample beam from
path.
eters. Spectral data are processed to identify the detector. The reference-beam attenuator is
the characteristic absorption peaks for the then driven in such a way as to decrease the

the sample beam into the monochromator. the attenuator and the pen drive of the record-
After dispersion by a prism or grating, the er and causes both to move until a null is
alternating beams fallon a detector and are achieved. A second synchronous motor drives
r-----I
converted to an electrical signal. The signal is the chart and varies the wavelength simultan- I
amplified and passed to the synchronous eously. There is frequently a mechanical link- I
I
rectifier, a device that is mechanically or elec- age between the wavelength and slit drives so
trically coupled to the chopper to cause the that the radiant power reaching the detector
rectifier switch and the beam leaving the is kept approximately constant by variations
chopper to change simultaneously. If the two in the slit width.
beams are identical in power, the signal from
the rectifier is an unftuctuating direct current.
Figure 8-5 shows the optics of a simple,
inexpensive commercial spectrophotometer
.--,
Sample

-_.:.I
If, on'the other hand, the two beams differ in manufactured by Beckman Instruments, Inc.
powet;, a ftuctuating or ac current is Radiation from the hot nichrome-wire source
prod~, the polarity of which is determined is split and passes through the sample com-
by which beam is the more intense. The cur- partment. Both beams have attenuators,
rent from the rectifier is filtered and further one for the 100% T adjustment and the other
amplified to drive a synchronous motor in for reducing the reference beam to null for the
one direction or the other, depending upon %T measurement. The beams are recombined
the polarity of the input current. The syn- by means of the segmented chopper, dis- FIGURE 8-5 Schematic diagram of the Beckman AccuLab® Infrared Spectro-
chronous motor is mechanically linked to both persed by the monochromator, and detected photometer. (Courtesy, Beckman Instruments, Inc., Fullerton, Ca. With permission.)
power of the reference beam to zc:ro also. matic diagram of a portable (weight = 18 heated equally by infrared radiation from the
Under these circumstances, essentially no Ib~ infrared filter photometer designed for
two sources. If the sample contains carbon

a-_ u*w
energy reaches the detector and the exact null quantitative analysis of various organic sub- monoxide, howev~r, the right-hand beam is
position cannot be located with precision. I~ stances in the atmosphere. The source is a attenuated somewhat and the corresponding
practice, the intensity of the sampl~ beam IS nichrome, wire-wound, ceramic rod; the I------~
detector chamber becomes cooler with re-
slowly diminished by the gradual mtroduc- transducer is a pyroelectric detector. A vari-
spect to its reference counterpart; the con-
tion of a shutter across its path; in this way, ety of interference filters, which transmit in Beam • sequence is a movement of the diaphragm to

--
the range between about 3 and 14 I.I1II (3000 attenuator
the tendency for the pen drive to overshoot the right and a change in capacitance of the
the zero is avoided, and a more accurate cali- to 750 em - 1~ are available; each is designed Sample
capacitor. This change in capacitance is
bration is achieved. Figure 8-14 illustrates the for the analysis of a specific compound. The
sensed by the amplifier system, the output of
tendency of a typical recording instrument to filters are readily interchangeable. Sample which drives a servomotor that moves the
overshoot when measuring zero or low trans- The gaseous sample is introduced into the cell

E-
beam attenuator into the reference beam until
mittances; again, the cause is the optical null cell by means of a battery-operated pump.
the two compartments are again at the same
deSign. Fortunately, the inevitable small error The path length of the cell as shown is O.S m;

-~~,.,~
temperature. The instrument thus operates as
a series of reflecting mirrors (not shown in

~=~~.-.
arising from the uncertainty in the zero in low Capacitor a null device. The chopper serves to provide a
transmittance measurements is not serious for Figure 8-6) permit increases in cell length to
with CO dynamic, ac-type signal which is needed be-
qu8litative work, provided the transmittances 20 m in increments of I.S m. This feature
cause the electrical system responds more re-
being measured are greater than-S to 10%. greatly enhances the concentration range of
producibly to an ac signal than to a slow dc
Because of these limitations inherent in the instrument. drift.
typical infrared-recording spectrophotometers, The photometer is reported to be sensitive with CO To 8tI1plifier diaphragm
The instrument is highly selective because
a :number of simple, rugged (and usually to a few tenths parts per million of such sub-
FIGURE 8-7 A nondispersive infrared heating of the detector gas occurs only from
si~gIe-beam) instruments have been designed stances as acrylonitrile, chlorinated hydro-
photometer for monitoring carbon mon- that narrow portion of the spectrum that is
for quantitative work. Some are simple filter carbons, carbon monoxide, phosgene, and
oxide. absorbed by the carbon monoxide in the
or' nondispersive photometers; others are hydrogen cyanide.
sample. Clearly, the device can be adapted to
spectrophotometers that employ filter wedges Nondispersive Photometers. Nondispersive
the analysis of any infrared-absorbing gas.
as the dispersing element. photometers, which employ no wavelength-
Automated Instruments for Quantitative
Fiker Photometers. Figure 8-6 is asche- restricting device, are widely employed to monitor gas streams for a single component. Analysis. Figure 8-8 is a schematic diagram
Figure 8-7 shows a typical nondispersive in- of a computer-eontrolled instrument designed
strument designed to determine carbon mon- specifically for quantitative infrared analysis.
oxide in a gaseous mixture. The reference cell The dispersing element, which consists of
is a sealed container filled with a nonabsorb- three filter wedges (p. 136) mounted in the
Inlet/outlet ing gas; as shown in the figure, the sample form of a segmented circle, is shown in Figure
IPO~ flows through a second cell that is of similar 8-8b. The motor drive and potentiometric

81
length. The chopper blade is so arranged that control permit rapid computer-controlled
the beams from identical sources are chopped wavelength selection in the region between
B.tte~ simultaneously at the rate of about five times 4000 and 690 em -1 (or 2.5 to 14.5 Jim) with
per second. Selectivity is obtained by filling an accuracy of 0.4 em - I. The source and
both compartments of the detector cell with detector are similar to those described in the
the gas being analyzed, here, carbon monox- earlier section on filter photometers; note
ide. The two chambers of the detector are that a beam chopper is used here. The sample
separated by a thin, flexible, metal diaphragm area can be readily adapted to solid, liquid, or
that serves as one plate of a capacitor; the gaseous samples. The instrument can be pro-
second plate is contained in the detector com- grammed to determine the absorbance of a
FIGURE 8-6 A portable, infrared photometer designed for gas partment on the left. multicomponent sample at several wave-
analysis. (Courtesy of Foxboro Analytical, Wilks Infrared Center, In the absence of carbon monoxide in the lengths and then compute the concentration
South Norwalk, cr. With permission.) sample cell, the two detector chambers are of each component.
purpose, a variety of cells are available with to allow variation in path length. Fixed path-
Chopper / Kevl!oard
path lengths that range from a few centi- length cells can be filled or emptied with a
ROOS- meters to several meters. The longer path hypodermic syringe.
•••• lengths are obtained in compact cells by pro- Sodium chloride windows are most com-
viding reftecting internal surfaces, so that the monly employed; even with care however
beam malces numerous passes through the their surfaces eventually become fogged dU~
sample before exiting from the cell. to absorption of moisture. Polishing with a
buffing powder returns them to their original
Solutions condition.
Solvents. Figure 8-9 lists the more com- The thickness of a narrow infrared cell can
mon solvents employed for infrared stud- be determined by measuring the transmit-
ies of organic compounds. It is apparent tance of the empty cell with air as the refer-
that no single solvent is transparent through- ence. Radiation reftected off the two walls of
out the entire middle-infrared region (Table the cell interacts with the transmitted radia-
UMotor & bearing pivot tion to produce an interference pattern such
8-1).
\I •
Swing-mounted
drive motor
Water and the alcohols are seldom
employed, not only because they absorb
as that shown in Figure 8-10. The thickness b
of the cell in centimeters is obtained from the
strongly, but also because they attack alkali- relationShip
metal halides. the most common materials
used for cell windows. For these reasons also, b= N
2(0"1 - 0"2)
care must be taken to dry the solvents shown
in Figure 8-9 before use. where N is the number of peaks between two
(b) Cells. Because of the tendency for solvents wavenumbers 0"1 and 0"2 •
FIGURE 8-8 An infrared instrument for quantitative analysis. to absorb, infrared cells are ordinarily much It should be noted that interference fringes
(a) Schematic of the instrument; (b) circular variable filter wheel. narrower (0.1 to I mm) than those employed are ordinarily not seen when a cell is filled
(Courtesy of Foxboro Analytical, Wilks Infrared Center, South in the ultraviolet and visible regions. Light with liquid because the refractive index of
Norwalk, cr.
With permission.) paths in the infrar~ range normally require
sample concentratit1ns from 0.1 to 10%. The
most liquids approaches that of the window
material; refteetion is thus reduced (Equation
cells are frequently demountable, with spacers 4-14, p. 103~ On the other hand, interference

SAMPLE centration or the cell length. Unfortunately,


HANDLING TECHNIQUES' this approach is not generally applicable for
infrared spectroscopy because no good sol-
As we have seen, ultraviolet and visible vents exist that are transparent throughout 5000 2500 1667 1250 1000 B33 714 625
Carbon disulfide
spectra are most conveniently obtained from the region. As a consequence, techniques Carbon tetrachloride
dilute solutions of the analyte. Absorbance must often be employed for liquid and solid
T elrachloroethylene
measurements in the optimum range are ob- samples that make -the accurate determina-
Chloroform
tained by suitably adjusting either the con- tion of molar absorptivities difficult if not
Dimethvlformamide _
impossible. Some of these techniques are dis-
Dioxane
cussed in the paragraphs that follow.
Cyclohexane
s For a more complete discussion, see: N. B. Collhup,
L. H. Daly, and S. E. Wiberley, 11ItrodllCfion to Infra Red Benzene
tUtd Raman Spectroscopy, 2d cd. New Yorlt: Academic Gas Samples
Press, 1975, p. 85; R. P. Bauman, AMorption Spectros-
copy. New Yorlt: Wiley, 1962, p. 184; Treatise on AlIOIyt- The spectrum of a low-boiling liquid or gas Wavelength, jlm

leal Chemistry, cds. I. M. Kolthoff and P. J. Elving. New can be obtained by permitting the sample to FIGURE 8-9 Infrared solvents. Black lines indicate
York: Interscienc:e, 1965, Part 1. vol. 6. p. 3582. expand into an evacuated cell. For this useful regions.
a two-phase mixture called a mull. An essen-
tial condition for the acquisition of satisfac-
tory spectra is that the particle size of the
suspended solid be smaller than the wave-
length of the infrared beam; if this condition
is not realized, a significant portion of the
-~"- From

Sample.

(e'
.

Solid with high


refractive index

radiation is lost to scattering.


b= __ N_
2(0, - 02) Two techniques are employed. In one, 2 to
12 5 mg of the finely ground sample (particle size
= 2(3200 _ 2000) = 0.0050 em
< 2 JlUl) is further ground in the presence of
one or two drops of a heavy hydrocarbon oil
(Nujol). If hydrocarbon bands are likely to
FIGURE 8-10 Determination of path length interfere, Fluorolube, a halogenated polymer,
b from interference fringe produced by an can be used. In either case, the resulting mull
is then examined as a film between fiat salt
empty cell.
plates.
In the second technique, a milligram or
less of the finely ground sample is intimately
mixed with about 100 mg of dried potassium
can be observed between 2800 and 2000 em - 1
bromide powder. Mixing can be carried out (b)
in Fiaure 8-1 (p. 211). Here, the sample is a
with a mortar and pestle; a small ball mill is
sheet of polyethylene, which has a refractive FIGURE 8-11 Internal reflectance apparatus. (a) Sample mounted on
more satisfactory, however. The mixture is
index Considerably different from air; s~g- reflection plate; (b) 'internal reflection adapter. (Courtesy of Foxboro
then pressed in a special die at 10,000 to
nificant reflection occurs at the two interfaces Analytical, Wilks Infrared Center, South Norwalk, CT. With permission.)
15,000 pounds per square inch to yield a trans-
of the sheet.
parent disk. Best results are obtained if the
disk is formed in a vacuum to eliminate
Pure Liquids occluded air. The disk is then held in the in-
When the amount of sample is small or when strument beam for spectroscopic examina- occurs.7 The depth of penetration, which incident angle, th~ radiation undergoes mul-
• suitable solvent is unavailable, it is common tion. The resulting spectra frequently exhibit varies from a fraction of a wavelength up to tiple internal reflections before passing from
practice to obtain spectra on the pure (neat) bands at 2.9 and 6.1 JlUl (3400 and 1600 several wavelensths, depends upon the wave- the crystal to the detector. Absorption and
liquid. Here, only a very thin film has a em - I) due to absorbed moisture. length, the index of refraction of the two attenuation takes place at each of these
sufficiently short path length to produce satis- materials, and the angle of the beam with re- reflections.
factory spectra. Commonly, a drop of the spect to the interface. If the less dense Figure 8-lIb is an optical diagram of a
neat liquid is squeezed between two rock-salt Internal-Reflection SpectroscoPY' medium absorbs the radiation, attenuation of commercially available adapter that will fit
plates to give a layer that has a thickness of the beam also occurs. into the cell area of most infrared spectrom-
When a beam of radiation passes from a Figure 8-11 shows an apparatus that takes
O.ot mm or less. The two plates, held together eters and will permit internal-reflectance
more dense to a less dense medium, reflec- advantage of the internal-reflectance phe-
by capillarity, are then mounted in the beam measurements. Note that an incident angle of
tion occurs; beyond a certain critical angle, nomenon for infrared-absorption measure-
path. Clearly, such a technique does not give 30, 45. or 60 deg can be chosen. CeUs for
reflection becomes complete. It has been
particularly reproducible transmittance data, ments. As will be seen from the upper figure, liquid samples are also available.
shown both theoretically and experimentally the sample (here, a solid) is placed on oppo-
but the resulting spectra are usually satisfac- Internal-reflectance spectra are similitr to,
that during the reflection process the beam site sides of a transparent crystalline material
tory for qualitative investigations. but not identical with, ordinary absotption
acts as if, in fact, it penetrates a small distance of high refractive index; a mixed crystal of spectra. In genera~ while the same peaks are
into the less dense medium before reflection thallium bromide/thallium iodide is frequent- observed, their relative intensities differ. The
Solids ly employed. By proper adjustment of the absorbances, while dependent upon the angle
Solids that cannot be dissolved in an infrared- of incidence, are independent of sample
transparent solvent can be suspended in a • For a review of this technique, see: P. A. Wilks. Jr. thickness, since the radiation penetrates only
suitable nonabsorbing liquid medium to form Amer. Lab .• 4 (11~ 42 (1972). a few micrometers into the sample.
DIO I'" t 1_ 1CIOO IlID •• ••

Internal-reflectance spectroscopy has been


applied to many substances such as polymers,
rubbers, and other solids. It is of interest to
note that the resulting spectra are free from
Correlation charts permit intelligent guesses
to be made as to what functional groups
are likely to be present or. absent in a
molecule. Ordinarily, it is impossible to iden- _.J~:';l~~+~j=
~~!~'~~.~~~~.~~~.
...'
the interference fringes mentioned in the tify unambiguously either the sources of an of
previous section. the peaks in a given spectrum or the exact
identity of the molecule. Instead, correlation

-~~.··==:,,~
..~:-~~F~;;~~~····~
QUALITATIVE charts serve as a starting point in the
APPLICATIONS identification process.
OF INFRARED ABSORPTION
Important Spectral
We have noted that the approximate fre-
Regions in the Infrared
quency at which an organic functional group, tOfftZlDCAIlIOXnlSAl,.TI, NlTT1I1 ••••. nc:.t - C;:Ol-j":T:"".--,-- .• - _..,..... .~ ..• - - - .•••••. -"....... •••••• "'---1
such as C-O, c-e, CHJ, and cae, The chemist interested in identifying an
will absorb infrared radiation can be cal- organic compound by the infrared technique
culated from the masses of the atoms and the usually examines lrertain regions of the spec-
force constant of the bond between them trum in a systematic way in order to obtain
(Equation 8-14). These frequencies, called clues as to the presence or absence of certain
(JI'OIlP frequencies, are seldom totally invariant group frequencies. Some of the important re-
because of interactions with other vibrations gions are considered briefly.
associated with one or both of the atoms Hydrogen Stretching Region 3700 to 2700
comprising the group. On the other hand, em-I (2.7 to 3.7 JII1l~ The appearance of
'---.-
i ••••.. ~~

such interaction effects are frequently small; strong absorption peaks in this region usually
as a result, a range of frequencies can be as- results from a stretching vibration between h:r:J< .. ,.; •..... ~.1 .. , It i~--'!- ..
-- • ...,...
signed within which it is highly probable that hydrogen and some other atom. The motion I"Y~l~tr~ ~;. ~ . Tl flNlT lNE$". : -o-~.t,-;;::-~.. {_-[-1--!--:~::~: -~ 2:'.""t:t.
the absorption peak for a given functional . is largely that of the hydrogen atom since it is i ; ~ .. ~ ,~ " .• _.. ~.. ""'I"S ;~::.~~ ,_~-.._~ __•. t i . ~ , -I
I I ••.. L., I ,~T •••••• :.c:-"-C-:-··_·--'·--r~--r~j~~ I
group will be found. so much lighter than the species with which it I
Group frequencies often make it possible bonds; as a consequence, the absorption is .-e.. idEY~Di(:~-'::·-'t-~-.- ~'--'I ! I I Ii; :
--I,-J.
I I '

--·~·---;"--,···_·_-
"', 'I'
ItSOCV~TEI. 1-'-C»8lIolOUI:.t I 1 ' 0 ,

to establish the probable presence or absence not greatly affected by the rest of the I
I
'.
I

:g.;:';/!a:'. .-

I
ooor--~$~e-OIILACfAllSJ
__ I.~ ATfC_o

I I
,-
HOllYflltHG

~rll\·."c.";'._

c~'I:
,- -
~

~I·ra-
• ' I
I '

m
I ~;------:_':'T."''':.. '-1-" ,,\c~o
of a Jiven functional group in a molecule. molecule. Furthermore, the hydrogen stretch- I I I ,I : '. I ' I I

ing frequency is much higher than that for I


~.~~~f=:~=
--'--, " ,I I ,I
I j I 2.
I I . "i':';'~'
,C~''''.I
! I I I: .
Correlation Charts other chemical bonds, with the result that
I I I ~~':-.:
I' I
::
- I
I; _.-
I I
•.• --
;
•. , •.. q •••• ,. ~J""PrJ:_-;r-~..:gIcc.~ I - j

interaction of this vibration with others is


c-..1 .••••••I •. ---r-- l.-

Over the years, a mass of empirical informa-


usually small.
tion has been accumulated concerning the
Absorption peaks in the region of 3700 to
frequency range within which various func-
3100 cm-1 (2.7 to 3.2 11m)are ordinarily due
tional groups can be expected to absorb.
to various O-H and N-H stretching vibra-
Correlation charts provide a concise means
tions, with the former tending to appear at
for summarizing this information in a form
higher wavenumbers~ The O-H bands are
that is useful for identification purposes. A
often broader than N-H bands and appear
number of correlation charts have been
developed,S one of which is shown in Figure
8-12. Summary Charts of Principol Group Frequencies. Ottawa:
National Research Council of Canada, 1959; K. Nakan-
ishi and P. H. Solomon, Infrared Absorption Spectros- ,
....,,
...•. "j,
'N. B. Colthup, J. Opt. Soc. Am., <to, 397--400(1950); copy, 2d ed. San Francisco: Holden-Day, 1977, pp. !

A. D. Cross, IntroduClion to Practical Infra-Red Spectros-


copy, 2d ed. Washington: Butterworths, 1964,pp. 56--{j2;
R. N. Jones, Infrared Spectra of Organic Compounds:
10-56; and R. M. Silverstein, G. C. Bassler, and T. C.
MorrilL Spectrometric Identification of Organic Com-
pounds. 3d ed. New York: Wiley. 1974.pp. 135-152.
.. t
U
, ,
U .., .. .. I I
.• I

" "
t
.. .. ..
I I I
only in dilute, nonpolar solvents. Hydrogen frequently impossible to determine the ty~ of
in this region is seldom possible because of
bonding tends to broaden the peaks and carbonyl that is present solely on the basl~ of resonance, and elemental analysis, usually
their complexity; on the other hand, it is this
move them toward lower wavenumbers. absorption in this region; how~er, examma- makes possible the positive identification of a
complexity that leads to uniqueness and the species.
Aliphatic C-H vibrations fall in the tion of additional spectral regions may pro-
consequent usefulness of the region for
region between 3000 and 2850 em -I (3.3 to vide the supporting evidence needed for The examples that follow illustrate how.
identification purposes.
3.5 pm). Most aliphatic compounds have a clear-cut identification. For example, esters infrared spectra are employed in the iden- ..
A few important group frequencies are to
sufficient number of C-H bonds to make have a strong C-O-R stretching peak at tification of pure compounds. In every case,
be found in the fingerprint region. These in-
this a prominent peak. Any structural varia- about 1200 em - I (8.3 pm ~ while aldehydes confirmatory tests would be desirable; a
clude the C-Q-C stretching vibration in
tion that affects the C-H bond strength will have a distinctive hydrogen stretching peak comparison of the experimental spectrum
ethers and esters at about 1200 em - I (8.3 11m)
cause a shift in the maximum. For example, just above 2700 em-I (3.7 pm~ as noted with that of the pure compound would suffice
and the C-Cl stretching vibration at 700 to for this purpose.
the band for CI-C-H lies just above 3000 previously.
800 em-I (14.3 to 12.5 JIm). A number of
em - I « 3.3 pm), as do the bands for olefinic Absorption peaks arising from C-e and
inorganic groups such as sulfate, phosphate,
and aromatic hydrogen. The acetylenic C-H C-N stretching vibrations are located in the
nitrate, and carbonate also absorb at wave-
bond is strong and occurs at about 3300 1690 to 1600 em -I (5.9 to 6.2 pm) range. Val- EXAMPLE
numbers below 1200 em-I.
em -I (3.0 11m). The hydrogen on the carbonyl uable information concerning the structure of The spectrum in Figure 8-13 was obtained
Umitations to the Use of Correlation
group of an aldehyde usually produces a dis- olefins can be obtained from the exact posi- for a pure colorless liquid in a O.OI-mm cell;
CbllI1s. The unambiguous establishment of
tinct peak in the region of 2745 to 2710 an-I tion of such a peak. the liquid boiled at 190°C. Suggest a structure
the identity or the structure of a compound for the sample.
(3.64 to 3.69 pm). Substitution of deuterium The region between 1650 and 1450 em-I
is seldom possible from correlation charts
for hydrogen causes a shift to lower wave- (6.1 to 6.9 pm) provides important informa- The four absorption peaks in the region of
alone. Uncertainties frequently arise from
numbers by the factor of approximately tion about aromatic rings. Aromatic com- 1450 to 1600 em-I are characteristic of an
I/J'i, as would be predicted from Equation overlapping group frequencies, spectral varia-
pounds with a low degree of substitution aromatic system, and one quickly learns to
8-14; this effect has been employed to identify tions as a function of the physical state of the
exhibit four peaks near 1600, 1580, 1500, and recognize this grouping. The peak at 3100
C-H stretching peaks. 1460 em-I (6.25, 6.33, 6.67, and 6.85 pm). sample (that is, whether it is a solution, a
em -I corresponds to a hydrogen-stretching
. The Triple-Bond Region Between 2700 and mul~ in a pelleted form, and so forth~ and
Variations of the spectra in this region with vibration., which for an aromatic system is
instrumental limitations.
1850 an -I (3.7 to 5.4 J&I11~ A limited number the number and arrangement of substituent usually above 3000 em -I. For aliphatic
of groups absorb in this spectral region; their groups are usually consistent but independent In employing group frequencies, it is essen-
hydrogens, this type of vibration causes
presence is thus readily apparent. Triple-bond tial that the entire spectrum, rather than a
of the type of substituent; considerable struc- absorption at or below 3000 em - I (see Figure
stretching results in a peak at 2250 to 2225 small isolated portion, be considered and in-
tural information can thus be gleaned from 8-12). Thus, it would appear that the com-
terrelated. Interpretation based On one part of
an-I (4.44 to 4.49 pm) for -C-N, at 2180 careful study of aromatic absorption in the pound is predominately, if not exclusively,
infrared region. the Spectrum should be confirmed or rejected aromatic.
to 2120 an -I (459 to 4.72 pm) for
by study of other regions.
-N+-e-, and at 2260 to 2190 em-I (4.42 The "Fingerprint" Region Between 1500 We then note the sharp peak at 2250 em-I
to 4.57 JIm) for -CSC-. Also present in and 700 em -I (6.7 to 14 J&I11~ Small differ- To summarize, then, correlation charts
and recall that only a few groups absorb in
this region are peaks for S-H at 2600 to serve only as a guide for further and more
ences in the structure and constitution of a this region. From Figure 8-12, we see that
2550 em -I (3.85 to 3.92 11m), P-H at 2440 to careful study. Several excellent monographs
molecule result in significant changes in the these include -CSC-, -C==CH, -CsN,
2350 em -I (4.10 to 4.26 11m), and Si-H at describe the absorption characteristics of
distribution of absorption peaks in this region and Si-H. The -CSCH group can be elim-
2260 to 2090 em-I (4.42 to 4.78 pm). functional groups in detail.9 A study of these
of the spectrum. As a consequence, a close inated, however, since it has a hydrogen
The DoabIe-Bond Region Between 1950 a'" characteristics, as well as the other physical
match between two spectra in the fingerprint stretching frequency of about 3250 em - I.
1550 an-I (5.1 to 6.5 J&I11~ The carbonyl properties of the sample, may permit unambig-
region (as well as others) constitutes strong The pair of strong peaks at 680 and 760
stretching vibration is characterized by ab- uous identification. Infrared spectroscopy,
evidence for the identity of the compounds em - I suggests that the aromatic ring may be
sorption throughout this region. Ketones, when used in conjunction with other methods
yielding the spectra. Most single bonds give singly substituted, although the pattern of
aldehydes, acids, amides, and carbonates ~I! such as mass spectroscopy, nuclear magnetic
rise to absorption bands at these frequencies; peaks in the regions of 1000 to 1300 em - I is
have absorption peaks around 1700 em because their energies are about the same, confusing and may lead to some doubt about
(5.9 pm). Esters, acid chlorides, and acid strong interaction occurs between neighbor- this conclusion.
anhydrides tend to absorb at slightly higher ing bonds. The absorption bands are thus • N. B. Collhup, L. N. Daly, and S. E. Wilberley, In/,o- Two likely structures that fit the infrared
wavenumbers; that is, 1770 to 1725 em-I composites of these various interactions and dlletion10 Infra RftI and Raman S""t:I,o.t:opy, 2d cd. New data would appear to be C6HsC=N and
(5.65 to 5.80 11m). Conjugation tends to lower depend upon the overall skeletal structure of York: Academic Press, 1975; and H. A. Szymansk~ C6HsCSC-C6Hs' The latter, however, is a
the absorption peak by about 20 em - I. It is the molecule. Exact interpretation of spectra Theory and P,at:lice of I~r,arftl S""t:I,OSt:opy. New York: solid at room temperature; in contrast, ben-
Plenum Press. 1964.
zonitrile has a boiling point (191°C), similar
co
to that of the sample. Thus, we tentatively for absorption in this region is usually the
•..
8
conclude that the compound under investiga- result of an N-H or an O-H stretching
tion is C6H,C=N.
8 vibration; the possibility that the substance is.
CIl
an aliphatic alcohol must thus be considered;
EXAMPLE
The empirical formula, however, is inconsis~
§ A colorless liquid was found to have the tent with a structure that contains both an
empirical formula C6H120 and a boiling
aldehyde and a hydroxyl group. Examination
§ point of 130°C. Its infrared spectrum (neat in
of the lower-frequency region fails to yield
a O.025-mm cell) is given in Figure 8-14. Sug- further information. This dilemma can be
§1 gest a structure.
resolved by attributing absorption at 3450
The presence of an aliphatic structure is em-I to the strong O-H stretching band of
, 8 suggested in Figure 8-14 by both the strong
~ water, which is often present as a contami-
band at 3000 em -I and the empirical for-
nant. Thus, we conclude that the sample is
8 mula. The intense band at 1720 em - I and
probably an aliphatic aldehyde. We are un-
::? the single oxygen in the formula strongly
able to determine the extent of chain branch-
indicate the likelihood of an aldehyde or a
8 ing from the spectrum; to be sure, the peak
:! ketone. The band at about 2800 em - I could
at 730 em - I suggests the presence of four or
be attributed to the shift in stretching vibra- more methylene groups in a row. The boiling
=~ ~ tion associated with hydrogen that is bonded
directly to a carbonyl carbon atom; if this
point of n-hexanal is 131°C; we are thus
inclined to conclude that the sample is this
~ interpretation is correct, the substance is an
compound. Confirmatory tests would be
aldehyde. The peak at 3450 cm - I is puzzling, needed, however.
co
8

4000 3000
·" ••• f, I
100

II>

80

• .,
, .
I 60
I
-.... .- , _I
I
.; --_ .. --'t-
!!
::;
"
I 40
I
"' V~
M-· " I-

1/111
III"

, N
0 0 ~
"!
,
0 1 11 12 13 14

FIGURE 8-14 Infrared spectrum of an unknown. (From "Catalog of Selected


Ultraviolet Spectral Data," Thermodynamics Research Center Data Project,
Thermodynamics Research Center, Texas A&M University, College Station, Texas.
Loose-leaf data sheets, extant 1975. With permission.)
236

3000 2500 2000 1500 1000 900 We now turn to the low-frequency end of the DevIation. from ••••••• Law
1~ I spectrum, where correlations of peaks with
It' aromatic substitution patterns can often be With infrared radiation, instrumental devia-
found. The two strong peaks at 680 and 825 tions from Beer's law are more common than
em - 1 and the weaker one at 865 em -I with ultraviolet and visible wavelengths be-
appear to fit best with the pattern for a sym- cause infrared absorption bands are relatively
metrically trisubstituted benzene. Thus, 3,5- narrow. Furthermore, the low intensity of
i 1\
T 5 di-methylpheJ1ol seems the most logical sources and low sensitivities of detectors in
choice. Confirmatory data would be needed, this region require the use of relatively wide
I however, to· establish the identity unam- monochromator slit widths; thus, the band-
biguously. widths employed are frequently of the same
V order of magnitude as the widths of absorp-

,.
, . 1
tion peaks. We have pointed out (see Figure
i 1 I.
I I Collections of Spectra 6-4) that this combination of circumstances
0 usually leads to a nonlinear relationship be-
3 4 5 6 9 10 11 12 13 14 15 As may be ~n from the foregoing examples,
tween absorbance and concentration. Cali-
correlation charts seldom suffice for the posi-
W•••• ength. /1m f Selected bration curves, determined empirically, are
. 8) (From "Catalog 0 tive identification of an organic compound
FIGURE 8-15 Spectru~ of a samp~ml~ ~~~~~ht~ent:n Data Project, Thermodynamics therefore often required for quantitative
from its infrared spectrum. There are avail- work.
Ultraviolet Spectral Data, Thermu~yn't C liege Station, Texas. Loose-leaf data sheets,. able, however, several catalogs of infrared
Research Center, Texas A&M mversl y, 0
extant 1975. With permission.) spectra that ~ist in qualitative identification
by providing romparison spectra for a large Absorbance Mea.urement
number of purb compounds. These collections
have become so extensivi: as to require edge- Matched absorption cells for solvent and so-
EXAMPLE 1100 to 1400 em - 1 is compatible with the lution are ordinarily employed in the ultra-
punched cards; IBM cards, or magnetic tapes
1be spectrum shown in Figure 8-15. w~ postulated O-H group, but we are unable to for efficient retrieval. Data presentation takes violet and visible regions, and the measured
obtained from a 0.5% solution of ~ hquid decide from the spectral data whether the several forms including replica of spectra on absorbance is then found from the relation
sample in a 0.5-mm cell. For the region of 2 sample is a phenol or an alcohol. notebook-size paper, cards, microfiches and
to 8 pm, the solvent was co.,
and for the From our observations thus far, and from microfilms, as well as magnetic tapes or disks. A-I= og--
P••I.on,
higher wavelengths, CSz was employed. The the empirical formula, likely structures for the P 101•• 1•••
In some instances, digitalized peak positions
empirical formula of the compound was unknown appear to be: . are employed in lieu of replicas of spectra. A The use of the solvent in a matched cell as a
found to be CsHIOO. Suggest a probable
structure.
OH CHzOH list of sources of these collections is found in reference absorber has the advantage of
a paper by Gevantman.1O largely canceling out the effects of radiation
The empirical formula indicates that t~e losses due to reflection at the various inter-
compound is probably aromatic, and .th~s CH'CH,-6 CH,-6 faces, scattering and absorption by the
supposition is borne out by the characterlS~~ solvent, and absorption by the container win-
pattern of peaks in the 1450 to 1600 em OH dows (p. ISO). This technique is seldom prac-
range. The C-H stretching bands, however, I tical for measurements in the infrared region
Quantitative infrared-absorption methods dif-
are somewhat lower in wavenumber t~an CHJCHZOH C-1CHJ
H-6. fer somewhat' from those discussed in the
because of the difficulty in obtaining cells
those for a purely aromatic system,. WhlC.h whose transmission characteristics are identi-
suggests the existence of some ahpha~l~ previous chapter because of the greater com-
6~ I ~ cal. Most infrared cells have very short path
groups as well. The broad band at 3300 em plexity of the $pectra, the narrowness of the
lengths that are difficult to duplicate exactly.
appears important, and suggests an O-H or absorption bands, and the instrumental limi-
tations of infrared instruments. In addition, the cell windows are readily at-
an N-H stretching vibration. The formula OH tacked by contaminants in the atmosphere
permits us to conclude. that the molecu~e and the solvent; thus, their transmission char-
must contain a phenohc or an al~hohc
O-H group. The spectrum in the region of
CH,OCH' acteristics change continually with use. For
these reasons. a reference absorber is often
dispensed with entirely in infrared work, and
238

100 reference. If P, remains constant during the


two measurements, then the transmittance of
the sample with respect to the solvent ~ be
i obtained by division of the two equations.
~
Co
,; To Po That is,
A· 109- • 109-
T = To/To = P/Po
i
.~
50 T. P

c
An alternative way of obtaining Po and T
!! is the base-line method, in which the solvent
I-

transmittance is assumed to be constant or at 012 13 14 12 13 14


0 least to change linearly between the shoulders A.I'fTl A,l'm
of the absorption peale This technique is
FIGURE 8-16 Base-line method for deter- demonstrated in Figure 8-16.
mination of absorbance.
Applications of
QuantiUtive Infrared Spectroscopy
With the exception of homonuclear mole-
the intensity of the radiation WSing through cules, all organic and inorganic molecular
the sample is simply compared with that of species absorb in the infrared region; thus,
the unobstructed beam; alternatively, a salt infrared spectrophotometry offers the poten·
plate may be placed in the reference beam. In tial for determining an unusually large o
12 13 14 15 12 13 14 15
either case, the resulting transmittance is or· number of substances. Moreover, the unique- A.l'm A I'm
dinarily less than lOO'i" •••even in regions of the ness of an infrared spectrum leads to a
spectrum where no absorption by the sample degree of specificity that is matched. or FIGURE 8-17 Spectra of C.Hto isomers in cyc1ohexane.
occurs; this effect is readily seen by examining exceeded by relatively few other analytical (From R. P. Bauman, Absorption Spectroscopy. New York:
the several spectra that have appeared earlier methods. This specificity has found particular John Wiley & Sons, Jnc .• 1962, p. 406. With permission.)
in this chapter. application to analysis of mixtures of closely
For quantitative work, it is necessary to related organic compounds. Two examples
correct for the scattering and absorption by that typify these applications follow.
the solvent and the cell. Two methods are A8aIysiI of • Mixa.e of AreIIIatic Hy4ro- Then four simultaneous equations can be Aulysis of Air Contaminants. The recent
employed. In the so-called cell in-cell out carboas. A typical application of quantitative written which permit the calculation of the proliferation of government" regulations with
procedure. spectra of the solvent and sample infrared spectroscopy involves the resolution concentration of each species from four respect to atmospheric contaminants has
are obtained successively with respect to the of C.H 10 isomers in a mixture which includes absorbance measurements (see p. 191). Such demanded the development of sensitive,
unobstructed reference beam. The same cell is o-xylene, m-xylene, p-xylene, and ethylben- calculations are most easily performed with a rapid, and highly specific methods for a vari-
used for both measurements. The transmit- zene. The infrared absorption spectra of the computer. ety of chemical compounds. Infrared absorp-
tance of each solution versus the reference individual components in the 12 to IS JlIIl When the relationship between absorbance tion procedures appear to meet this need
beam is then determined at an absorption range is shown in Figure 8-17; cyclohexane is and concentration is nonlinear (as frequently better than any other single analytical tool.
maximum of the analyte. These transmit- the solvent. Useful absorption peaks for occurs in the infrared region), the algebraic Table 8-2 demonstrates the potential of
tances can be written as determination of the individual compounds manipulations associated with an analysis infrared spectroscopy for the analysis of mix-
occur at 13.47, 13.01, 12.58, and 14.36 Jim, of several components having overlapping tures of gases. The standard sample of air
respectively. Unfortunately, however, the absorption peaks are considerably more containing five species in known concentra-
absorbance of a mixture at anyone of these complex." tion was analyzed with the computerized in-
To= PIP, wavelengths is not entirely determined by the strument shown in Figure 8·8; a 2O-m gas cell
concentration of just one component, because was employed. The data were printed out
where P, is the power of the reference beam of overlapping absorption bands. Thus, molar II The treatment or inrrared data for various types or within a minute or two after sample injection.
and To and To are the transmittances of the absorptivities for each of the four compounds mixtures is discussed in: R. P. Bauman, Absorption Sf'«- Table 8-3 shows potential applications of
solvent and sample, respectively, against this must be determined at the four wavelengths. Iroscopy. New York: Wiley, 1962,pp. 403-419. infrared filter photometers (such as that
shown in Figure 8-6) for the quantitative absorption characteristics suitable for deter- Disadvantages and LimlUti_
transfo?J1 spectroscopy, which wer~ reported
determination of various chemicals in the mination by means of infrared filter pho- to Quantitative Infrared Methods
approxunately a decade later, were to the
atmosphere for the purpose of assuring com- tometers or spectrophotometers. Obviously, energy-starv~ far-infrared region; by the
Several disadvantages attend the application
pliance with OSHA regulations. among all of these absorbing compounds, late 19605, Instruments for chemical studies
of infrared methods to quantitative analysis.
Of the more than 400 chemicals for which peak overlaps are to be expected; yet the
Among these are the frequent nonadherence in bot~ the. far- (10 to 400 em - t) and
maximum tolerable limits have been set by method should provide a moderately high the orchnary-mfrared regions were available
to Beer's law and the complexity of spectra;
the Occupational Safety and Health Admin- degree of selectivity. commercially. I 2
the latter enhances the probability of the
istration, more than half appear to have
overlap of absorption peaks. In addition, the Before considering details of Fourier trans-
narrowness of peaks and the effects of stray form spectroscopy, it is worthwhile to de-
radiation make absorbance measurements scribe the general principle of the procedure
critically dependent upon the slit width and and its inherent advantage in recovering
the wavelength setting. Finally, the narrow weak signals from noisy backgrounds.
cells required for many analyses are inconven-
TABLE 8-2 AN EXAMPLE OF INFRARED ient to use and may lead to significant
ANALYSIS OF AIR CONTAMINANTS- analytical uncertainties. The Inherent Adventege
The analytical errors associated with a of Fourier Transfonn Spectroscopy
ConcD, Found, Relative quantitative infrared analysis often cannot be
Contaminants ppm ppm EnOr,% reduced to the level associated with ultra- For purposes of this discussion, it is conven-
violet and visible methods, even with con- ient to think of an experimentally derived
Carbon monoxide SO 49.1 spectrum as being made up of m individ-
Methylethylketone 100 98.3 siderable care and effort.
Methyl alcohol 100 99.0 ual transmittance measurements at equally
Ethylene oxide SO 49.9 spaced frequency or wavelength intervals
Chloroform 100 99.5 called resolution elements. The quality of the
INFRARED FOURIER spectrum-that is, the amount of spectral
TRANSFORM SPECTROSCOPY detail-increases as the number of resolution
elements becomes larger or as the frequen-
The sensitivity of several spectroscopic meth- cy intervals between measurements become
ods can be enhanced significantly by appli- smaller. t3 Thus, in order to increase spectral
cation of the Fourier transform mentioned quali!y, m must be made larger; clearly, in-
in Chapter 4 (p. 94~ Use of the trans- creasmg the number of resolution elements
TABLE 8-3 SOME EXAMPLES OF INFRARED VAPOR
form requires a dedicated high-speed must also increase the time required for
ANALYSIS FOR OSHA COMPLIANCE-
computer and often, spectroscopic equipment obtaining a spectrum.
AUo".bIe Exposure, Minimum Detectable that differs considerably from the conven- Consider, for example, the derivation of an
Compound ppmb ;., JIm Concentration, ppm< tional dispersive type. The Fourier transform infrared spectra from SOO to SOOO em - I. If
technique has been applied advantageously resolution elements of 3 em - t were chosen, m
Carbon disulfide 20 4.54 0.5
Chloroprene 25 11.4 4 !o infrared, nuclear magnetic resonance,
Diborane 0.1 3.9 0.05 Ion cyclotron resonance, atomic absorption,
Ethtenediamine 10 13.0 0.4 mass, and microwave spectroscopy. In this IZ For a more complete discussion of Fourier transform
Hy roren cyanide 10 3.04 0.4 section, we will discuss its use for infrared infrared spectroscopy. consult the following references:
Methy mercaptan 10 3.38 0.4 studies. P. R. Griffiths. Chemicallnfrarrd Fourier Transform Spec-
Nitrobenzene 1 11.8 02 Iroscopy. New York: Wiley. 1915; Fourier Transform
Pyridine 5 14.2 02 Fourier transform infrared spectroscopy
Infrored Speclroscopy: Applicarions 10 Chemical SyslelfU,
Sulfur dioxide 5 8.6 0.5 was first developed by astronomers in the eds. J. R. Ferraro and L. J. Basile. New York: Academic
Vinyl chloride 1 10.9 0.3 early 19505 in order to study the spectra of Press. Vol. I. t918 and Vol. II, t919.
distant stars; only by the Fourier technique I J With a recording speetropho!ometer, of course, individ-
could the very weak signals from these ual point-by-point measurements are not made; never-
• (Courtesy of FoxborofWilks, Inc, South Norwalk, CT.) theless. the idea of a resolution element is useful, and the
• 1911 OSHA exposure limits for 8-hour weighed average. sources be isolated from environmental noise.
ideas generated from it apply to recording instruments as
, For 2O.2S-m cell. The first chemical applications of Fourier well.
would be 1500; if one-half second were amount of time as the measurement of a
required for recording the transinittance of single resolution element in conventional spec-
each resolution element, 750 s or 12.5 min troscopy (one-half second in our example~
would be nCeded to obtain the spectrum. Re- then in the 12.5 min required to record a
ducing the width of the resolution element to conventional spectrum, 1500 (12.5 x 60/0.5)
1.5 an - 1 would be expected to provide replicate Fourier transform measurements
significantly greater spectral detail; it would could be averaged. The resulting improve- Frequency

iU
also double the number of resolution ele- ment in signal-to-noise ratio would be JlSOO
ments as well as the time required for their or about 39. This advantage of Fourier trans-
measurement. form spectroscopy was first suggested by P.
The quality of an infrared spectrum de- Fellgett in 1958 and has become known as
Frequency
pends not only upon the number of resolu- the F ellgett advantage.
tion elements, but also upon the signal-to- Spectroscopic methods, such as the Fou-
noise ratio for the instrument. Unfortunately, rier transform technique, in which all resolu-
an increase in the number of resolution ele- tion elements are measured simultaneously
ments is usually accompanied by a decrease are often called multiplex methods. Measure-
in the signal-to-noise ratio, because narrower ment of ultraviolet-visible spectra with a fixed
slits are required. Thus, the signal reaching monochromator and a vidicon tube (p. 142)
the transducer is weaker; the noise, however, is another example of a multiplex method.
ordinarily remains undiminished.
As was pointed out in Chapter 3 (p. 76~
Time Domain Spectroscopy
an effective method of extracting a weak
signal from a noisy environment is by signal Conventional spectroscopy can be termed
averaging. This technique is effective because frequellCY domain spectroscopy, in that raC
in summing replicate measurements, the sig- diant power data are recorded as a function of (bl
nals contained therein are additive; the radiant frequency (or the inversely related
wavelength). Time domain spectroscopy, in
FIGURE 8-18 Illustration of time domain (a) and (b) and frequency domain (c), (d~ and
random noise, on the other hand, increases (e) plots.
with the square root of the number of meas- contrast, is concerned with changes in radiant
urements only. The improvement in signal-to- power with time. Figure 8-18 illustrates the
noise ratio brought about ~ signal averaging difference.
is thus proportional to .j n, where n is the The cueves in (c) and (d) are conventional the time domain spectrum of the source con-
number of replications. spectra of two monochromatic sources with It is important to appreciate that a time
taining the two frequencies. As is shown by
The application of signal averaging to con- frequencies VI and V: Hz. The curve in (e) is domain spectrum contains the same informa-
the horizontal arrow, the plot exhibits a pe_
ventional infrared spectroscopy is, unfor- the spectrum of a source containing both tion as does a spectrum in the frequency
riodicity or beat as the two waves go in and
tunately, costly in terms of time. Thus, in the frequencies. In each case, some measure of the domain, and in fact, one can be converted to
out of phase.
example just considered, 12.5 min were radiant power, P(l') is plotted with respect to the other by mathematical manipulations.
Examination of Figure 8-19 reveals that Thus, Figure 8-18b was derived from Figure
required to obtain a spectrum of 1500 resolu- the frequency in hertz. The symbol in paren- the time domain spectrum for a source con-
tion elements. Improvement of the signal-to- theses is added to. emphasize the frequency 8-18e by means of the equation
taining several wavelengths is considerably
noise ratio by a factor of two, would require dependence of the power; time domain power more complex than those shown in Figure pet) = k(cos 2nv1 t + cos 2nv: t)
4 x 1500 measurements, or SOmin. will be indicated by pet). 8-18. Because a large number of wavelengths
Fourier transform spectroscopy differs from The curve in Figure 8-18a shows the time where k is a constant and t is the time. The
are involVed, a full cycle is never realized in difference in frequency between the two lines
conventional spectroscopy in that all of the domain spectra for each of the mono- the. time shown. To be sure, a pattern of beats
resolution elements for a spectrum are meas- cllromatic sources. The two have been plotted was approximately 10% of v:.
can be observed as certain wavelengths pass The interconversion of time and frequency
ured simultaneously. Significant gains in together in order to make the small frequency in and out of phase. In general, the signal domain spectra becomes exceedingly complex
signal-to-noise ratio accrue as a consequence. difference between them more obvious. Here, power decreases with time as a consequence and mathematically tedious when more than
For example, if we assume that a single Fou- the instantaneous power pet) is plotted as a of the various closely spaced wavelengths be- a few lines are involved; the operation is only
rier transform measurement requires the same function of time. The curve in Figure 8-18b is coming more and more out of phase. practical with a high-speed computer.
Method. of obtain time domain spectra requires, there~ The Michelaon Interferometer
Obtllining Time Domein Spectre and the other of which is movable. The beams
fore, a method of converting (or modulDting) a The device used for modulating infrared radi-
Time domain spectra, such as those shown in high-frequency signal to one of measurable then meet again at the beam splitter, with half
ation is an interferometer similar in design to
Figures 8-18 and 8-19, cannot be acquired ex- frequency without distort~g the tim~ rela- of each beam being directed toward the
one first described by Michelson in 1891. The
perimentally with radiation of the frequency tionships carried in the Signal; that IS, the sample and detector and the other two halves
Michelson interferometer is a device that
range of absorption spectroscopy (1015 Hz frequencies in the modulated ~ignal m?S~ be being directed back toward the source. Only
splits a beam of radiation into two beams of
for ultraviolet to 107 Hz for nuclear mag- directly proportional to those In the onginal. the two halves passing through the sample to
nearly equal power and then recombines
netic resonance) because transducers with suf- Different signal-modulation procedures are the detector are employed for analytical pur-
them in such a way that intensity variations poses, although the other halves contain the
ficiently rapid response times are not avail- employed for the various wavelength regions of the combined beam can be measured as
able. Thus, a typical transducer responds of the spectrum. The Michelson interferom- same information about the sourCe.
a function of differences in the lengths of
only to the average power of a high-frequency eter is used extensively for measurements in Horizontal motion of the movable mirror
the paths of the two halves. Figure 8-20
signal and not to its periodic variation. To the infrared region. will cause the power of the radiation reaching
is a schematic diagram of such a device as
the detector to t€ictuate in a; predictable
it is used for infrared Fourier transform
manner. When the two mirrors are equidis-
spectroscopy.
tant from the splitter (position 0 in Figure
As shown in the figure, a beam of radiation 8-20~ the two parts of the recombined beam
from a source is collimated and impinges on a will be totally in phase and the power will be
beam splitter, which transmits approximately
a maximum. For a monochromatic source,
half of the radiation and reflects the other
motion of the movable mirror in either direc-
half. The resulting twin beams are then
tion by a distance equal to exactly one-
reflected from mirrors, one of which is fixed quarter wavelength (position B or C in the

t tI
I
~
I
F I

.1111111111
ABOCD
[1 n n f1 n.J:l.ll.+}
11/1111111
~
111111'111
q..f.J...- __
t
_
,----,.
I
I So~rce(
.r
~'r: 1"'11111111 l!IIT~M
11111I1111 1111111111 I
U u
I I
UU
I II
U i::tlj-U-tJu - -r ~
I J 1 '" I
--.:>. "
-u, -~A 0 +4>. +U. ~I ~I~Sam-pl-e-+'
0'
lStance.cm
I
1oe~tor I
/ r I \ t
~ :/ \ I
o~ tf}l ~
-2A -lA 0 +lA +2A
6.cm
FIGURE 8-19 Time domain spectrum of a source made up of FIGURE 8-20 Schematic diagram of a Michelson inter-
several wavelengths. ferometer illuminated by a monochromatic source.
INFRARED ABSORPTION SPECTROSCOPY 247

figure) will change the path length of the cor- where v is the frequency of the radiation and

~I -
Inlerferograms
Spectra

-Li
responding reflected beam by one-half wave- c is the velocity of light (3 x lO' 0 em/s).
length (one-quarter wavelength for each When VAlis constant, it is evident that the
direction). Under this circumstance, destruc- interferogram frequency f is directly propor-
tive interference will reduce the radiant power tional to the optical frequency v. Furthermore, is:
of the recombined beams to zero. Further the proportionality constant will generally be
motion to A or D will bring the two halves a very small number. For example, if the 0 • +
back in phase so that constructive interfer- o,cm-1
mirror is driven at a rate of 1.5 em/s, 6.em
ence can again occur. (al
The difference in path lengths for the two 2VM= 2 x 1.5 em/s = 10-10
beams, 2(M - F) in the figure is termed the c 3 X 1010 em/s _ A
retardation f>. A plot of the output power
from the detector versus f> is called an inter- :0
f= 1O-lov is: ti ·2
ferogram; for monochromatic radiation, the is: "
interferogram takes the form of a cosine curve
such as that shown in the lower part of As shown by the following example, the
Figure 8-20 (cosine rather than sine because frequency of infrared radiatic)n is readily mod-
the power is always a maximum when f> is
zero and the two paths are identical).
ulated into the audio range by a Michelson
interferometer.
6.cm
0

lb)
• +
t m
-'

In the typical infrared scanning interfer-


ometer, the mirror is motor-driven at a con:: EXAMPLE
stant velocity of VAlem/so Since peaks occur Calculate the frequency range of a mod-
at higher multiples of the wavelength of the ulated signal from a Mij::helson interfer-
incident radiation, the frequency f of the inter- ometer with a mirror veloc:;ity of 0.50 em/s,
ferogram is given by for infrared radiation of 2.5 and i6 Jim
(1.2 x 10.4 to 1.9 X 1013 Hz).
= 2vAlem/s Employing Equation 8-15, we find
-------0 ---- .+
f ). em (8-15) 6.cm

2 x 0.5 em/s leI


Here, ). is the wavelength of the incident
II = 2.5 JiM x 10-4 em/JiM = 4000 Hz FIGURE 8-21 Comparison of interferograms and optical spectra.
beam (in em) and the factor 2 is required be-
cause the reflection process doubles the effect
of the retardation. 2 x 0.5
In some instances, it is convenient to
f2 = 16 X 10-4 = 625 Hz
Fourier
express Equation 8-15 in terms of wave- interferogram signal. The parenthetical sym-
Transformation of Interferograms
number of the incident radiation u rather bols emphasize that one power (u) is in the
than in wavelength. Thus, Certain types of infrared transducers are The .cosine wave of the interferogram shown frequency domain and the other (f» in
capable of following fluctuations in signal 10 Flgur~ 8-2.la (and also in Figure 8-20) can the time domain. In practice, the foregoing
f= 2VMU (8-16) power that fall into the _ audio-frequency be descnbed 10 theory by the equationl4 equation is modified to take into account
The relationship between the optical fre- range. Thus, it becomes possible to record a the fact that the interferometer ordinarily will
modulated time domain spectrum which P(f» = tP(u) cas 2nft (8-18) not split the source exactly in half and that
quellCy of the radiation and the frequency
of the interferogram is readily obtained by reflects exactly the appear~nce of the very where P(u) is the radiant power of the infra- the detector response and the amplifier be-
substitution of ;. = c/v into Equation 8-15. high-frequency time domain spectra of an red beam incident upon the interferometer havior are frequency-dependent. Thus, it is
Thus, infrared source. Figure 8-21 shows three and P(f» is the amplitude or power of the useful to introduce a new variable B(u) which
examples of such time domain interferograms depends upon P(u) but also takes these other
on the left and their frequency domain coun- factors into account. Therefore, we rewrite the
terparts on the right. equation in the form
" This equa.ion can be derived in lhe same way as
Equa.ion 2-16 (p. (2).
Substitution of Equation 8-16 into Equation output must be sampled periodically and EXAMPLE
8-19 leads to stored in digital form. Equation 8-22, how- What length of mirror drive will provide a
ever, demands that the sampling intervals d6 resolution of 0.1 em -I? Several Fourier transform spectrometers are
P(6) ••• B(O") cos 271:2v"O"t (8-20) be infinitely small; that is, d6 --+ O. . Substituting into Equation 8-25 gives available commercially. Their prices lie in the
But the mirror velocity can be expressed in From a practical standpoint, only a finite- S35,000 to SI20,OOO range (including the
terms of retardation or sized sampling interval can be summed over a ., 0.1 •••~ dedicated computer for performing the Fou-
finite retardation range (a few centimeters). rier transformation).
6 These limitations have the effect of limiting
v,,'" 2t 6 •••10 em Drive I,\fechanism. A requirement for satis-
the resolution of a Fourier transform instru- factory interferograms (and thus satisfactory
Substitution of this relationship into Equa- ment and restricting its frequency range. The mirror motion required is one-half the spectra) is that the speed of the moving
tion 8-20 gives Resolution of • Fourier Transform Spec- retardation, or 5 em. mirror be relatively constant and its position
trometer. The resolution of a Fourier trans- exactly known at any instant. The planarity
P(6) ••• B(O") cos 271:60" form spectrometer can be described in terms In addition to the length of mirror sweep,
resolution is also influenced by the planarity of the mirror must also remain constant
which expresses the magnitude of the interfer- of the difference in wavenumber between two during its' entire sweep of to em or more.
lines that can be just separated by the instru- of the mirrors and the coplanarity of the
ogram signal as a function of the retardation beam splitter. Resolution is also lost if the In the far-infrared region, where the wave-
factor and the wavenumber of the optical ment. That is, lengths .re in the micrometer range, dis-
radiation entering the interferometer con-
input signal. verges or diverges significantly. placement of the mirror by a fraction of a
The interferograms shown in part (b) of Sampling the InterfefOlraDL In order to wavelength, and accurate measurement of its
Figure 8-21 would be described by position, can be accomplished by means of
where 0"1 and 0"2 are wavenumbers for a pair compute the Fourier transform of an inter-
a motor-driven micrometer screw. A more
P(6) •••Ba(a) cos 271:60"1 + B2(0") cos 271:60"2 of barely resolvable infrared lines. ferogram, it is necessary' to digitalize and
store the radiant power at equal intervals of precise and sophisticated mechanism is re-
It is possible to show that in order to re-
(8-21) quired for,the mid- and near-(!iTrared regions,
solve two lines, it is necessary to scan the time retardation. That is, the interferogram must
be broken into resolution elements and the however. Here, the mirror mount is generally
For a continuous source, such as in part domain spectrum long enough so that one
power of each of these measured. digitalized. floated oil air cushions held within c1ose-
(c) of Figure 8-21, the interferogram can be complete cycle or beat for the two lines is
and stored in the memory of a computer. In- fitting stainless steel sleeves (see Figure 8-22).
represented as a sum of an infinite number of completed; only then will all of the in-
formation theory permits the calculation of The mount is driven by an electromagnetic
cosine terms. That is, formation contained in the spectra have been
the minimum number of resolution elements coil similar to the voice coil in a loudspeaker;
• + 00 recorded. For example, resolution of the two
a slowly increasing current in the coil drives
I X that will permit reconstruction of a spec-
P(6) •••
..-- •. B(O") cos 271:60"•• d6 (8-22) lines 0"1 and 0"2 in Figure 8-21b would require
recording the interferogram from the maxi- trum from an interferogram without a loss of
detail. The number is gi.ven by
the mirror at constant velocity. After reaching
its terminus, the mirror is returned rapidly to
mum A at zero retardation to the maximum
the starting point for the next sweep by a
B where the two waves are again in phase.
• +00 rapid reversal of the current. The length of
P(O")'" I B(6) cos 271:0"6•. dO" (8-23) The maximum at B occurs, however, when
°.-- 00

Infrared Fourier transform spectroscopy


60"2 is larger than 60"1 by I in Equation 8-21.
That is, when
travel varies from 2 to about 18 em; the scan
rates range from 0.05 em/s to 4 em/so
Two additional features of the mirror
consists of recording P(6) as a function of 6 system are necessary for successful operation
(Equation 8-22) and then mathematically EXAMPLE in the infrared regions. The first of these is
transforming this relation to one that gives Calculate the number of resolution ele- a means 'of sampling the interferogram at
P(O") as a function of 0" (the spectrum) as ments that would be needed to reconstruct an precisely 'spaced retardation intervals. The
shown by Equation 8-23. infrared spectrum from 4000 to 625 cm - I if a second is, a method for determining exactly
Equations 8-22 and 8-23 cannot be em- resolution of 1.5 cm - I is desired. the zero-retardation point to permit signal
ployed as written because they assume that averaging. If this point is not known pre-
the beam contains radiation from zero to Substitution into Equation 8-24 reveals that X ••• 2'4000 - 625) •••4500 cisely, the signals from repetitive sweeps
infinite wavenumbers and a mirror drive of the resolution is given by 1.5 would not be fully in phase; averaging would
infinite length. Furthermore, Fourier trans- tend to degrade rather than improve the
formations with a computer require that the Thus, the power for 4500 points would need signal.
detector output be digitalized; that is, the to be measured, digitalized, and stored. The problem of precise signal sampling
of the radiation is reflected. A widely used shorter. than those of the other infrared '
material for the far-infrared region· is a thin detectors.
film of Mylar sandwiched between two plates DoabIe-Beam Design. Instruments for the
of a low refractive-index solid. Thin films of far-infrared region are often of single-beam
germanium or silicon deposited on cesium design. Most instrumen~ for the higher-
iodide or bromide, sodium chloride, or potas- frequency range are double-beam; after exit-
IA sium bromide are satisfactory for the mid-
source ing from the interferometer, the beam is
5, infrared region. A film of iron(III) oxide is alternated between a sample compartment
deposited on calcium fluoride for work in the and a reference compartment by means of a
near-infrared. moving mirror. The beams are then recom-
Sources .ad Detectors. The sources for bined and pass on to the detector.
Fourier transform infrared instruments are Performance Characteristics of Commer-
similar to those discussed earlier in this chap- cial Instruments. Several instrument makers
ter. Generally, pyroelectric detectors must be olTer Fourier transform spectrometers which
employed because their response times are cover various frequency ranges and have var-

FIGURE 8-22 Interferometers in an infrared Fourier transform spectrometer. Subscripts I


define the infrared interferometer; subscripts 2 and 3 refer to the laser and white-light interferometers,
respectively. (Courtesy of Nicolet Instrument Corp., Madison, Wisconsin.)

~ ~ Zero
and signal averaging is accomplished in mod- each successive zero crossing. The laser-fringe --..,J~_re_ta_rda_tion _
em instruments by using three interferom- reference system gives a highly reproducible
- eters rather than one, with a single mirror and regularly spaced sampling interval. In
mount holding the three movable mirrors. most instruments, the laser signal is also
Figure 8-22 is a schematic diagram showing employed to control the speed of the mirror-
the arrangement employed by one manu- drive system at a constant level.
facturer. The components and radiation paths The third interferometer system, some-
for each of the three interferometer systems times called the "white-light" system, em-
are indicated by the subscripts I, 2, and 3, ploys a tungsten source S3 and transducer 13
respectively. System I is the infrared system sensitive to visible radiation. Its mirror
that ultimately provides an interferogram system is fixed to give a zero retardation that
similar to that shown as curve A in Figure is displaced to the left from that for the analyt-
8-23. System 2 is a so-called laser{ringe refer- ical signal (see interferogram B, Figure 8-23).
ence system, which provides sampling-interval Because the source is polychromatic, its
information. It consists of a helium neon power at zero retardation is much larger than FIGURE 8-23 Time domain signals for the three
laser S2, an interferometric system including any signal before and after that point. Thus, interferometers contained in Fourier transform infrared
mirrors MM2 and M2, a beam splitter this maximum can be employed to trigger the instruments. Curve A: Infrared signal. Curve B: White
B2, and a transducer T2• The output from start of data sampling for each sweep at a light signal. Curve C: Laser-fringe reference signal. •
this system is sinusoidal, as shown in C of highly reproducible point. Curve D: Square-wave electrical signal formed from the
Figure 8-23. This signal is converted elec- Beam Splitters. Beam splitters are con- laser signal. (From P. G. Griffiths, Chemical Infrared
tronically to the square-wave form shown structed of transparent materials with re- Fourier Transform Spectroscopy. New York: John Wiley
in D; sampling begins or terminates at fractive indices such that approximately 50% & Sons, Inc., 1975, p. 102. With permission.)
ious resolutions. Some instruments are much larger energy throughput (approxi- upon both the metal atom and the organic distinguishable with a dispersive-type spec-
designed exclusively for the· far-infrared mately 50%) than do dispersive optics, which portion of the species. trometer. A particularly important application
region (- 10 to SOO em -I ); the best reso- are limited by the necessity of narrow slit Far-infrared studies of inorganic solids has been to studies of aqueous solutions.
lution of these instruments is about 0.1 em - I. widths. The higher energy reaching the detec- have also provided useful information about Identification of Species After Gas Chro-
tor with Fourier transform instruments often lattice energies of crystals and transition ener- matographic Separatioas. Perhaps the most
Their cost ranges from '$40,000 to $60,000, in-
cluding the computer. '. results in an increase in sensitivity. Finally, it gies of semiconducting materials. important application of the infrared Fourier
Several instruments are available for the should be noted that the interferometer is free Molecules composed only of light atoms technique in the mid-infrared region .has
mid-infrared region; most have interchange- from the problem of stray radiation because absorb in the far-infrared if they have skeletal been to the on-line identification of species
able sources and beam splitters that permit each IR frequency' is, in effect, chopped at a bending modes that involve more than two eluted from a gas chromatographic column.
coverage of several ranges of wavenumbers. different frequency. atoms other than hydrogen. Important An analysis of this sort must be completed
When interchangeable beam splitters, detec- The Fourier transform method offers little examples are substituted benzene derivatives, in a few seconds; thus, rapid data acquisition,
tors, and sources are used, one instrument is re- or no advantage o'ver a good-quality grating which generally show several absorption such as that provided by a scanning inter-
ported to produce spectra from the visible to spectrophotometer. for routine qualitative ap- peaks. The spectra are frequently quite ferometer, is essential. Further discussion of
the far-infrared (16,000 em-I to 10 em-lor plications in the' region between 650 and specific and useful for identifying a particular this application will be found in the chapter
0.6 to 1000 JlI11~ 4000 an-I (15 and 2.5 JlI11~ Furthermore, it compound; to be sure, characteristic group devoted to gas chromatography.
Resolutions for commercial instruments suffers by compa~on in terms of high initial frequencies also exist in the far-infrared Emission Spectroscopy. Upon being heated,
vary from about 2 em -I to a low of 0.06 cost and substantial maintenance problems. region. molecules that absorb infrared radiation are
an -I for the more expensive instruments. The latter arise because the quality of a Fou- Pure rotational absotption by gases is ob- also capable of emitting characteristic infrared
Several minutes are required to obtain a com- rier transform spectrum degrades much more served in the far-infrared region, provided the wavelengths. The principal deterrent to the
plete spectrum with the highest ~Iution. rapidly with instrument maladjustment than molecules have permanent dipole moments. analytical application of this phenomenon
On the other hand, a complete spCctrum can does a grating s~rum. Examples include H20, 03, Hel, and AsH3 • has been the poor signal-to-noise ratio char-
be displayed on a television screen in less The paragraph$. that folIow outline briefly Absorption by water is troublesome; elimi- acteristic of the infrared emission signal,
than one second if a resolution of several some of the important applications of Fourier nation of its interference requires evacuation particularly when the sample is at a tempera-
em - I is tolerable. transform spectrosCopy. or at least purging of the spectrometer. ture only slightly higher than its surroundings.
Far-infrared SPectrostOpy. The earliest Mid-lnfrared SpectrostOPy. Application of With the interferometric method, interesting
chemical applications of the Fourier trans- Fourier transform spectroscopy in the 650 to and useful applications are now appearing.
form technique involved absorption studies in 4000 em - I range has been largely confined to An early example of the applicatiol)--(>f
Applications of the range between 400 and 10 em -I (25 and special problems where some sort of energy infrared emission spectroscopy is found (in a
FT Infrared Spectroacopyll 1000 JlI11). The energy advantage of the inter- limitation exists. For example, it has proved paperl6 which describes the use of a Fourier
As noted early in this discussion, most of ferometric system over a dispersive one gen- useful for studying micro samples where transform spectrometer for the identification
the applications of Fourier transform spec- eralIy results in a significant improvement in absorption is confined to a very limited area; of microgram quantities of pesticides. Sam-
troscopy have been in areas where low ra- spectral quality. In addition, the use of grat- spectra for particles as small as 100 Ilm can ples were prepared by solution in a suit-
diant energy is a serious problem. Here, the ing spectrometers in this wavelength region is be acquired. able solvent and evaporation on a NaO or
Fellgett advantage of the interferometric complicated by the overlapping of several The method has also been employed for KBr plate. The plate was then heated elec-
system offers significant improvements over orders of diffracted radiation. studying transient species that would other- trically near the spectrometer entrance. Pest-
the signal-ta-noise ratios found with prism or The far-infrared region is particularly use- wise require rapid scanning. The gain here icides such as DOT, malathion, and diel-
grating instruments. Another advantage is ful for inorganic studies because absorp- results from the fact that the entire spectrum drin were identified in amounts as low as I to
that the optics of the interferometer provide a tion due to stretch)ng and bending vibrations is observed simultaneously. 10 Ilg.
of bonds between metal atoms and both inor- Single-beam Fourier transform spectros- Equally interesting has been the use or the
ganic and organic, ligands generalIy occur at copy provides a useful method for the study interferometric technique for the remote
frequencies lower than 600 em-I (> 17 JlI11). of dilute solutions. Here, the interferograms detection of components emitted from in-
For example, heavy-metal iodides generalIy for the solvent and sample are obtained indi-
•• For an excellenl summary of applications of FT infra- absorb in the region below 100 em -I, while vidualIy and stored in the computer. The
red spectroscopy, see: J. R. Ferraro and L. J. Basile, the bromides and chlorides have bands at spectrum of the pure solute is then obtained
FolllUr Transform l'lfrared Spectroscopy: Applications to
Chemical Systcnu. New York: Academic Press, vol. I.
higher frequencies. Absorption frequencies for by subtraction. For dilute solutions, the dif- " I. Coleman and M. J. D. Low, Speetrochim. Acta, 22,
1978 and vol. II, 1979. metal-organic bonds are ordinarily dependent ference will often be so small as to be in- 1293 (1966~
dustrial stacks. In one of these applications,l' steps, a sufficient number of linear simultan- (a) aromatic C-H, q = 3030 em-I
an interferometer was mounted on an 8-inch eous equations can be written to permit the (b) alkyne C-H, q = 3300 em-I
reflecting telescope. With the telescope fo- calculation of the power p. C9rresponding to (c) aldehydic C- H, q = 2750 em - 1
cused on the plume from an industrial plant, each of the resolution elements. That is, a (d) alkane C-H, q = 2900 em-I
COz and 50z were readily detected at a spectrum can be mathematically derived by 3. Carbon disulfide is a linear molecule.
distance of several hundred feet. the computer. The array of equations used (a) Draw schematic diagrams for the different modes of vibration of
takes the form CSz·

HADAMARD (PT)1 = al,lPl + al,Z Pz + + al •• P. (b) Indicate which of the normal vibrational modes are infrared
active.
TRANSFORM SPECTROSCOPY (PT)z = aZ.IPI + az. z Pz + + az .• p. (c) Assuming a standard double bond force constant (about 10 x 105
dynes/em~ calculat~ the expected absorption for the stretching
Another multiplexing technique that provides vibrations. Compare your answer with the standard correlation
the Fellgett advantage is Hadamard trans- (PT). = a.,IPI + a.,zPz + ... + a. .• p•
charts. Explain the large difference.
form spectroscopy.lS This method employs a
The quantity (PT). corresponds to the power 4. Shown below are some of the normal modes of vibration for ethylene,
modified-grating infrared spectrometer in
at step n, while a•.• is a coefficient that has a HzC-eHz· Indicate whether the vibration will be active or inactive
which the dispersed radiation is focused on a
value of 0 for an opaque slot and 1 for a in the infrared spectruni.
mask located on the focal plane of the mono-
transparent slot. When n = In, sufficient data
chromator. The mask is somewhat longer (a) C-H stretch
are available to solve the set of equations for
than twice the length of the focal plane and is
PI through p •. "" r/ $ $
covered with a continuous random arrange- H, /H
As in the interferometric experiment, the H, ~
ment of opaque and transparent slots, each of
which has the dimensions of the entrance slit.
spectrum is viewed continuously. Thus, the
signal-to-noise ratio is in theory improved by
H/
C==C
'H H
/
c==c ,H
The mask in its various positions has the
effect of dividing the dispersed radiation into
the factor of jrn.
The Hadamard transform
? '" $ $

method does not, however, offer the advan-


m resolution elements, each of which has a
tage of greater energy throughout.
bandwidth determined by the width of the
To date, the Hadamard method has not
slots (or the entrance slit). The radiation that
passes through the mask is reflected back to
the grating, where it is recombined and then
been widely used, despite the fact that the
equipment is simpler and should be less ex-
""
,
H
C==C
/
H
?

passes out of the entrance slit; the total power


pensive and easier to maintain. H/ 'Ii
of the exiting beam is measured with a radi- ? ""
ation detector.
The measurement process consists of step- (f) C-H bend
ping the mask across the focal plane one slot 17 M. J. D. Low and F. K. Clam:y. E.n•. Sci. Technol. I.
width at a time and measuring the total 13 (1961~ (H, /H\
power at each step. The power data are dig- IIFor a discussion or Ihis method. see: J. A. Dec:lcer. \H/C==C'H)
italized and stored in a computer. After m App.Opl. 10,510 (1911) and A•••.•'. Lab. 4 (n
29 (1912~

5. The force conslanl for the bond in HF is about 9 x 105 dynes/em.


1. For CHCI3, the C-H stretch occurs at 3030 em -I and the c-a (a) Calculate the vibrational absorption peak for HF.
stretch occurs at 758 em-I. (b) Repeat the calculati~n for DF.
(a) Calculate the position of the C-D stretch for CDCh. 6. A compound isolated from the kernels of almonds was determined to
(b) Calculate the expected frequency of the C-Br stretch of bromo- have the molecular formula C,H60. From the infrared spectrum
form, CHBr 3' given in Figure 8-24, determine the structure of this unknown.
2. Most organic compounds contain C-H bonds, whose IR absorptions 7. Acetone reacts with hydrocyanic acid to give a moderately stable
appear at approximately 3000 em - 1. Calculate the force constants for adduct whose infrared spectrum is shown in Figure 8-25. Interpret the
the following types of C-H bonds: major peaks in this spectrum, and deduce the structure of the adduct.
8. From the infrared spectrum given in Figure 8-26, determine the struc-
ture of an unknown having the formula C,H9N. By examining the
fingerprint region, determine which isomer corresponds to this
spectrum.
9. For the infrared spectrum shown in Figure 8-27, assign probable
700 650 625
functional groups to all the peaks. What class of molecules would give , 1 1 ,I 1
100 100 100
this spectrum?
90 90 90
10. From the infrared spectrum given in Figure 8-28, determine what 80 """"- 80 80
functional groups are present in this molecule. What gives rise to the
peak at 1250 em -I? This compound is often used as a solvent for glue
70
80
, If
T
70-
80-
r II III.
!iW I, "' f\
70

and has a strong odor of bananas. Given the molecular formula, 60 50- r 60
50
C,HI40Z' suggest the structure of this unknown. (
40 40- - 40
11. Deduce the structure of the compound CJH40 whose infrared spec- 30 30- 30
trum is given in Figure 8-29. Explain the double peak at 3300 em - I. 20 20- 20
12. Deduce the functional groups present in the unknown whose IR spec- 10 10 , 10
trum is shown in Figure 8-30.
o Neat 0 o
2.5 3 8 10 11 12 13 14 16 16 .
13. The compound, whose spectrum is shown in Figure 8-31, was used as
an antiseptic for early surgical work and is still found in some popular WMeng1h,,..,.

face creams. From the infrared spectrum determine its structure. The FIGURE 8-24 See Problem 6. (From C. J. Poucl)ert, The Aldrich Library of Infrared Spectra,
molecular weight of this unknown has been determined to be 94 ± 1. 2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc.,
14. The compound, whose infrared spectrum is shown in Figure 8-32, can Milwaukee, Wisconsin.) .
be found in small quantities in black pepper. It has a characteristic
fishy odor. What functional groups are indicated by the infrared spec-
trum? Given its molecular weight of 85, suggest possible structures for
this compound.
15. From the infrared spectrum given in Figure 8-33, determine what
functional groups are present in this molecule. Explain the peaks at
4llOO
1700 em - I. The structural features of this unknown are all repre- , I
sented in its infrared spectrum. Determine the structure of this
compound.
16. An empty cell showed 14 interference peaks in the wavelength range
100
90
80 f l
100

80
- 1 ..• 1/ i'\ r' - .•....
100
90
80

of 6.0 to 12.0 /lm. Calculate the path length of the cell.


70
60
- f( 70
60 \
70
60
17. An empty cell exhibited 8.5 interference peaks in the region of 1000 to 50 50
\ 50
1250 cm -I. What was the path length of the cell? 40 40 40
18. What length of mirror drive in a Fourier transform spectrometer 30 30 30
would be required to provide a resolution ,?f (a) 0.05 em-I and (b) 20 20 20
,
om em-I? 10 10 10
o Neat 0 o
2.5 3 9 10 11 12 13 14 115 16

FIGURE 8-25 See Problem 7. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
2d ed., 1975 and published with permission of the Aldrich Chemical Company, Inc., Milwaukee,
Wisconsin.)
100
1 . . ."
700
.1
650 625
,I
4000
I
700
,I
650 626
I

,- ,
100 100 100 100
~ 100
90
80
.fl
II
80
-
V
If' ~ ...•..
90
80
90
80 80
"l ,..., '" - - 90
80

- =~
70 70- 70 70 70
60 60 - - I 80 80 ~ J 70
60
50 r--'- 501- -
\ I' -' 50
.•... 60
50 50
\ 50
40 40 40 40 40 40
30
20
30 •... 30 30 30 30
20 20 20 20 20
10 10 10 10 10 10
o Melt 0 o o NNt 0 o
2.5 3 5 6 7 8 II 10 11 12 13 14 15 16 2.6 3 6 6 7 8 II 10 11 12 13 14 16 16
WIVelength, jUl1
WIVelength, jUl1

FIGURE 8-26 See Problem 8. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra, FIGURE 8-28 See Problem 10. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
2d "4 •• 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee, 2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee,
WiscOnsin.) Wisconsin.)

1 . . .1 ,I 4000 700 650 626


100
90 , ,rr 100 100
100
I
100
I
100

-
'V 90
80 1'1' I' 90 90
80 80 J 1/ V. I~r
70 701- r-r-
I"-. 80 80 LL l-
80
\
70
-) 70 II \ I \ 1\ '"
80 80- -- 70 70
50 50
'\, 60
50
60
50
W 60
50
'f 1/
\
60
40
30
20
40
30
20
40
30
40
30
40
30
, 50
40
30
20 \ I'\.
10
o
2.5 3
10
Neat 0
10
o
20
10
o
20
10
Neat 0
- 20
10
o
6 9 10 11 12 13 14 15 16
2.5 3 7 8 9 10 11 12 13 14 15 16
Wovelength. jUl1
W••• length, jUl1
FIGURE 8-27 See Problem 9. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee,
FIGURE 8-29 See Problem II. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
Wisconsin.) 2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc .• Milwaukee,
Wisconsin.)
-,- . -,- ,
700 650 625
I , 700
, ,
850625
100 100 100 100
100 100
90
80
••••••••••
\
:/'
90
80
ft.tl
- / V-I" IV \./
/'

,...
90
80
70
90
80
70
-\. r ;T
90
80
70
f\ JI IV
.....•
90
60

'''' ..
70 70 70
_I I ~
60
50
40
\
\ I
60
50
40
" I
1/ c/'
60
50
40
60
50
40
I
60
50
40
" II
V '\,
II
60
50
40
30 30 30 30 30 30
20 20 20 20 20 20
10 10 10 10
10 NujOI'0
o mull 0 o o Nut 0 o
2.6 3 8 10 11 12 13 14 15 18 2.6 3 7 8 8 10 11 12 13 14 15 16
WIVMngth. I'm wwelength. pm
FIGURE 8-30 See Problem 12. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra, FIGURE 8-32 See Problem 14. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee, 2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee,
Wisconsin.) Wisconsin.)

-.-
700
.1 ,
850 625

',' ., 650 825

-
100 100 100 I .

-..,
90 100 100 100
10 ~
80
70
60
50
,J
TI
I
801-'-
701- l-
60
501- l-
~

V
/
r ""'" II
V
~
II
1-1.

1\
80
70
60
50
'1
90
80
70
60
I
V
90

70
60 -I-
iii

~ I
,./ ""
I'"
90
80
70
60
401- l- 40 50 50- .\J 50
40 1
\ 301- l-
I 30 40 40
1\ 40
30
2O~ I-
II II 20 30 30 ~ 30
20 I\J 1/
10 10 10
o
20
10
20
10
-~ "-
20
o Melt 0 10
2.5 16 o Melt 0 o
2.5 15 16

FIGURE 8-31 See Problem 13. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee, FIGURE 8-33 See Problem IS. (From C. J. Pouchert, The Aldrich Library of Infrared Spectra,
Wisconsin.) 2d ed., 1975, and published with permission of the Aldrich Chemical Company, Inc., Milwaukee,
Wisconsin.)
When radiation passes through a transparent ~n important advantage of Raman spectra
medium, a fraction of the radiant power is over infrared lies in the fact that water does
scattered at all angles from the direction of not cause interference; indeed, Raman spectra
the beam by the molecules or aggregates of can be obtained from aqueous solutions. In
moiecules present (p. 104). The 'fraction of the. addition, glass or quartz cells can be em-
beam that is scattered is undetectable to the ployed, thus avoiding the inl.:Onvenience of
eye. when the particles are of molecular working with sodium chloride or other atmo-
dimensions; this type of scattering is termed spherically unstable windows.
Rayleigh scattering. If the medium contains
aggregates of particles with dimensions ap-
proximating that of the wavelength of the THEORY OF
radiation, scattering becomes evident as the RAMAN SPECTROSCOPY
Tyndall effect or as a turbidity of the solution.
In 1928, the Indian physicist C. V. Raman Raman spectra are obtained by irradiating a
discovered that the wavelength of a small sample with a powerful source of yisible
fraction of the radiation scattered by certain monochromatic radiation. A mercury arc was
molecules differs from that of the incident employed in early investigations; by now, this
beam and furthermore that the shifts in wave- source has been superseded by high-intensity
length depend upon the chemical structure of gas or solid lasers. During irradiation, the
the molecules responsible for the scattering. spectrum of the scattered radiation is: mea-
He was awarded the 1931 Nobel prize in phys- sured at some angle (usually 90 deg) with a
ics for this discovery and his systematic ex- suitable visible-region spectrophotometi:r. At
ploration of it. the very most, the intensities of Raman lines
The theory of Raman scattering, which by are 0.01 ~, of the source; as a consequence,
now is well understood, shows that the phe- their detection and measurement creates some
nomenon results from the same type of experimental problems.
quantized vibrational changes that are asso-
ciated with infrared adsorption. Thus, the dif- Exdu~onofR~nS~~
ference in wavelength between the incident
and scattered radiation corresponds to wave- Figure 9-1b shows a photograph of the
Raman spectrum for carbon tetrachloride
lengths in the mid-infrared region. Indeed, the
excited by an intense mercury arc. For refer-
Raman scattering spectrum and infrared ab-
ence, the spectrum of the source is shown in
sorption spectrum for a given species often
part (a) of the figure. Note that the two mer-
resemble one another quite closely. There are,
cury lines appear unchanged in wavelength
however, enough differences between the kinds
on the lower spectrum as a consequence of
of groups that are infrared and Raman active
normal Rayleigh scattering. In addition,
to make the techniques complementary rather
however, two sets of five lines appear cin the
than competitive. For some problems, the
infrared method is the superior tool; for
i
others, the Raman procedure offers advan- lice or Raman spectroscopy, see: T. R. Gilson alid P. J.
tages. Because of this close relationship, we Hendra, Laser Raman Spectroscopy. New York: Wiley·
have chosen to discuss Raman spectroscopy at_ IntcncicllCiC,1970; Raman Spectroscopy, Theory OM
this point.! Prtu:tice, 2 vols~ eel. H. A. Szymanski. New York:
Plenum Press, 1967 and 1970; J. R. Duria and W. C.
Harris, PhysiCGI Methods of Chemistry. New York:
Wilcy-Inlerscicnce, 1972, Vol. I, Part. III B, Chapter 2;
and Infrared aM Raman Spectroscopy, cds. E. G. Brame
and J. G. Grassclli. New York: Dekker, 1977.
particular source line and (J is the wave- first vibrational level. After a retention time
number for the Raman peak. Thus, the spec- of perhaps 10-15 to 10-14 s, the species ordi-
trum for carbon tetrachloride in Figure 9-1 narily return to their ground state, as shown
A· 404.7 nm A = 435.9 nm reveals Raman peaks at 218, 324, 459, 762, by the dashed arrows, with the release in all
o• 24710 em-I 0= 22940 em-' and 790 em -1. It should be rioted that anti- directions, of radiation with exactly th~ same
Stokes shifts may be more convenient to energy as the source. The small fraction of this
measure in samples that also fluoresce. radiation that is transmitted at an angle to the
co"' •.• Trusitions Responsible for RaIJWl Peaks. beam is Rayleigh scattering. Under some cir-
"''''co
"'CD",
The discrete energy shifts that characterize
"''''''' cumstances, however, the molecule may return
"''''''' Raman spectra indicate the presence of quan-
II : I I
from the distorted or virtual state to the first
'I
II I ,
I I I
I
tized energy transitions. These shifts can be
rationalized by assuming that the electrical
excited vibrational level of the ground state
II I I I (broad arrow in Figure 9-2). The frequency
'1 I I I field of the radiation interacts with the elec- emitted in this transition will be less by a
trons of the sample and causes periodic polar- quantized amount corresponding to the dif-
ization and depolarization. As a consequence, ference in energies between the ground state
Iin..
Stokes Anti·Stok
•• lines the energy of the radiation is momentarily and the first vibrational leveL dE. That is,
(b)
retained in a so-called virtUQI state as a dis-
torted polarized species. The energetics of this hv = hv,-dE
FIGURE 9-' (a) Intense lines of a mercury arc spectrum and (b) Raman process is shown by the two vertical arrows
spectrum for CCl4 • on the left in Figure 9-2; note that the inter-
actions, in contrast to absorption, do /lOt dv= dE/h
involve transition to a higher electronic energy where v and v, are the frequencies of the
level. Note also that the second arrow rep- Raman peak and the source peak, respec-
longer wavelength side of the two intense fluorescence experiment; thus, Raman shifts resents the interaction between photons and a tively, and h is Planck's constant.
mercury lines at 404.7 and 435.8 nm. If it is in this direction are sometimes called Stokes molecule that has been thermally excited to the As shown to the right in Figure 9-2, anti-
assumed that each set of Raman lines is shifts (Chapter 10). We shall see, however, Stokes scattering results from interaction of
excited by the nearer of the two intense mer- that the two phenomena arise from funda- Vibr.tional 3------- _ the radiation with a molecule that is initially
cury lines, it is found that the pattern of mentally different processes; for this reason, levels.2------- _ in its first excited vibrational level. The result-
displlcement in wavenumbers (em - 1) is iden- the application of the same nomenclature to first excited
eIectronlc 1------- _ ing emission is increased in frequency by the
tical for the two sets. Thus, the 404.7-nm line both fluorescent and Raman spectra is unfor- SlIte 0----- _ amount dE if the distorted species returns to
with a wavenumber of 24,710 em-I exhibits tunate. the ground vibrational state. Since the frac-
Raman lines at 24,492, 24,386, and 24,251. As Shifts toward higher energies are termed R'V1eigh tion of molecules in the first excited state is
Raman
shown by the lower scale of Figure 9-1b, these anti-Stokes; quite generally, anti-Stokes lines scattering scattering smaller than in the ground state at room
E=hv E= hI' ilE
lines correspond to Raman shifts (du) of 218, have appreciably weaker intensities than the 1
temperature, anti-Stokes radiation is weaker
I
324, and 459 em - 1. Note that lines having corresponding Stokes lines. I than Stokes.
I
identical shifts appear to the right of the Raman spectra are characteristic of the
I '"
435.8 nm (22,946 em - 1) line. Additional lines molecules responsib,le for the scattering and I
I
""
I
I ~ Il '"
.•• -1
appear at shifts of 762 and 790 em -,I. are thus useful for analytical purposes. Most I '" S +
I II <I) •
...:. ..c: Vibrational Modes
Another set of less intense lines appear to commonly, the spectra are plotted in terms of
II "'. C , Associated with the Raman Effect
the left of both of the mercury Iin~ (the set the wavenumber shift du with respect to the I Il <{ '"

associated with the 404.7-nm lint: is not source line. Since Stokes peaks are the more
Vibration.13
--/---l-I __":_
I S The transitions responsible for the Raman
shown for lack of space). These lines are intense, dt1 (the shift in units of em - 1) is effect are of the same type as those described
levels, 2 I J- in Chapter 8 for infrared absorption; that is,
displaced by amounts identical to the usually evaluated on the basis of the shift to ground
electronic 1 I J__.
displacements to the right. lower wavenumbers; that is, SlIte 0
transitions between the ground state of a
Superficially, the appearance of Raman molecule and its first vibrational states are
spectral lines at lower energies (longer wave- involved in both phenomena. Therefore, for a
lengths) is analogous to what is observed in a given molecule the energy shifts observed in a
between the two nuclei. Thus, absorption of
radiation of the vibration frequency cannot
~ 8 80 a occur. On the other hand, the polarizability
.~ 6
60 ! of the bond between the two atoms of such a
.5 i molecule varies periodically in phase with the
Ii •• 40 ~ stretching vibrations, reaching a maximum at
~
.0:: 2 20 !!: ~is8. tI [
0 -------------
Infrared
the greatest separation and a minimum at the
, inactive
closest approach. A Raman shift corresponding
in frequency to that of the vibrational mode
thus results.
~H It is of interest to compare the infrared and

...~ ~[
the Raman activities of coupled vibrational
modes such as those described earlier (p. 218)
for carbon dioxide. In the symmetric mode,
.~ 8
shown in Figure 9-48, it is seen that no
!6
.5 change in dipole occurs as the two oxygen
! .. atoms move away from or toward the cen~ral
0:: 2 carbon atom; thus, this mode is infrared-in-
active. The polarizability, however, fluctuates
in phase with the vibration since distortion of
foOO 3500 3000 2500 2000 1800 1800 1400 1200 1000 800 600 400 200
bonds becomes easier as they lengthen and
a or <1a(cm-')
more difficult as they shorten; Raman activity
FIGURE 9-3 Comparison of Raman and infrared spectra. (Courtesy Perkin- is associated with this mode.
Elmer Corp., Norwalk, Connecticut.) As shown in Figure 9-4b, the dipole
moment of carbon dioxide fluctuates in phase
with the asymmetric vibrational mode; thus,
an infrared absorption peak arises from this
(bl Diode. The polarizability also changes during
Raman experiment should be identical to the molecule have a change in dipole or charge
FIGURE 9-4 Raman and infrared activity the vibration, reaching a maximum at the
energies of its infrared absorption bands, pro- distribution associated with it. Only then can
of two vibrational modes of carbon dioxide: equilibrium position and a minimum at either
vided that the vibrational modes involved are radiation of the same frequency interact with
(a) symmetric, (b) asymmetric. vibrational extreme. Note, however, that the
active toward both infrared absorption and the molecule and promote it to an excited
fluctuations in polarizability are not in phase
Raman scattering. Figure 9-3 illustrates the vibrational state. In contrast, scattering in-
with the vibrational changes but Occur at
similarity of the two types of spectra; it is seen volves a momentary elastic distortion of the
twice the frequency. As a consequence, Raman
that several peaks with identical u and Au electrons distributed around a bond in a
scattering is not associated with these asym-
values exist for the two compounds. It is also molecule, followed by reemission of the radia- creasing bond strength, and decreasing bond metric vibrations.
noteworthy, however, that the relative size of tion in all directions as the bond returns to length. The Raman shift in scattered radia- As shown earlier (p. 219), a dipole change,
the corresponding peaks is frequently quite its normal state. 111' its distorted form, the tion, then, requires that there be a change and thus infrared absorption, is also asso-
dilferent; moreover, certain peaks that occur molecule is temporarily polarized; that is, it in polarizability-rather than a change in ciated with the two degenerate scissoring
in one spectrum are absent in the other. develops, momentarily, an induced dipole dipole-associated wit#l the vibrational mode vibrational modes of carbon dioxide. Here
The differences between a Raman and an which disappears upon relaxation and re- of the molecule; as a cOnsequence, the Raman again, the change in polarizability does not
infrared spectrum are not surprising when it emission. The effectiveness of a bond toward activity of a given mode may differ markedly occur at the vibrational frequency but at twice
is considered that the basic mechanisms, scattering thus depends directly upon the ease from its infrared activity. For example, a homo- the frequency; these modes do not, therefore,
although dependent upon the same vibra- with which the electrons of the bond can be nuclear molecule such as nitrogen, chlorine, produce Raman scattering.
tional modes, arise from processes that are distorted from their normal positions (that is, or hydrogen has no dipole moment either in Often, as in the foregoing examples, parts
mechanistically different. Infrared absorption the polarizability of the bonds); polarizability ~ts equilibrium position or when a stretching of Raman and infrared spectra are comple-
requires that a vibrational mode of the decreases with increasing electron density, in- vibration causes a change in the distance mentary, each being associated with a dilferent
set of vibrational modes within a molecule. radiation is produced at 632.8 nm; several
Other vibrational modes may be both Raman- other lower-intensity nonlasing lines accom-
and infrared-active. Here, the two spectra pany the principal line and must be rem.oved
resemble one another with peaks involving the by suitable narrow band filters. J\lte!Datlvely, Excitation
the effect of these lines can be eliminated by collector
same energies. The relative intensities of corre- mirror
sponding peaks may differ, however, because taking advantage of the fact t~t nonlasing
the probability for the transition may be lines diverge much more rapidly than the
different for the two mechanisms. lasing line; thus, by making the distance be-
tween the source and the entrance slit large,
the intensities of the former can be made to
Intensity of Raman Peaks approach zero.
Argon-ion lasers, with lines at 488.0 and
The intensity or power of a Raman peak 514.5 nm, are also employed, particularly
depends in a complex way upon the polar- when higher sensitivity is required. Because
izability of the molecule, the intensity of the the intensity of Raman scattering varies as the
source, and the concentration of the active
group, as well as other factors. In the a~ce
fourth power of the frequency of the exci!ing
source, the argon line at 488 nm provides
I Monoc:hromator
entrance
of absorption, the power of Raman emISSion Raman lines that are nearly three times as
increases with the fourth power of the fre- intense as those excited by the helium/neon II
quency of the source; however, advantage can source, given the same input power. .
--l..l-. Interference
L..++-J filter
seldom be taken of this relationship because A variety of other laser sources are avatl-
of the likelihood that ultraviolet irradiation II
will cause photodecomposition.
able; undoubtedly new and improved ~ources
will appear in the future. The need eXISts for
H
Raman intensities are usually directly pro-
portional to the concentration of the active
several sources, inasmuch as one must be
chosen which is not absorbed by the sample
DL~ef
species. In this regard, Raman spectroscopy (b)
or solvent. Furthermore, care must be taken to
more closely resembles fluorescence than ab- avoid wavelengths which will cause the sample FIGURE 9-5 Two sample excitation systems.
sorption, where the concentration-intensity to fluoresce or to photodecompose.
relationship is logarithmic.

Sample-Illumination Systems details ofthe reflection ofthe Raman radiation incorporate photomultipliers to record peak
off the walls; in fact, the holder is a I-mm 0.<1. intensities are now available. Their designs
Sample handling for Raman studies tends to glass capillary that is about 5 em long.
Instrumentation for modern Raman spectros- are not significantly different from the record-
be simpler than for infrared because the mea- For unusually weak signals such as might
copy consists of three components, namely, ing spectrophotometers discussed in Chapter
sured wavenumber differences are between occur with a dilute gas sample, the cell is
an intense source, a sample illumination 7. Most employ double monochromators to
two visible frequencies. Thus, glass can be often placed between the mirrors of the laser
system, and a suitable spectrophotometer.2 minimize spurious radiation reaching the
employed for windows, lenses, and other opti- source; enhanced excitation power results. detector. In addition, a split-beam design is
cal components. In' addition, the source is
Sources employed to compensate for the effects of
readily focused on a small area and the
fluctuations in the intensity of the source.
The most widely used Raman source is prob- emitted radiation can be efficiently focused on Raman Spectrophotometera
ably the helium/neon laser, which operates a slit. Very small samples can be examined as
All early Raman studies were performed with Comparison of
continuously at a power of SO mW. Laser a consequence. In fact, a common sample Raman and Infrared
holder for liquid samples is. an ordinary glass prism spectrographs, with the spectra being
recorded on photographic films or plates. Sample-Handling Techniques
melting-point capillary.
Figure 9-5 shows two of many configura- Intensities were determined by measuring the In terms of ease of sample preparation and
2 For a description of commercially available instru-
tions for the handling of liquids. The size of blackness of the lines. handling, Raman spectroscopy possesses the
ments up to 1969. see: B. J. Bulldn, J. CIrnn. Educ, ~
A781. A859 (1969~ , the tube in (b) has been enlarged to show Several Raman spectrophotometers that great advantage of permitting use of glass
cells instead of the more fragile and atmo- ment, however, must be of higher quality be-
!_. oA
I

--..-.,..
cause spectral artifacts arising from imper- • _ p'henYI
spherically less stable crystalline haiides re- NH
quired for work in the infrared. fections have more serious consequences. .·H

Substances of limited solubility can often Detectors. Raman instruments use photo- C-C·H
multiplier detectors which offer distinct signal- CIl•
be finely ground and tapPed into an open- CH. .s·s- ~
ended cavity for Raman exariiination; the more
troublesome mull technique is required for in-
to-noise advantages over the thermal detectors
employed in infrared instruments (see p. 143).
_S·H


.C=N -- c·s

•_~-o
frared studies. Polymers can often be examined ResolutiolL The resolution of the best !IIIC';;C- ---',c-c
for Raman activity directly with no preliminary infrared and Raman spectrometers is about • mono. and 1.2.
sample treatment; comparable infrared analy- the same (- 0.2 em-I). -.ICOOH mono •• l,3,.an~ ~~
N-N·Rr""I N-N'f "mono.
ses may require that the polymer be com- Cost. There are no low-cost Raman spectro-
J
pressed, molded, or cast into a thin film prior photometers equivalent to bench-top infrared
spectrometers. The p~ices of Raman instru-
all. S~
••••
R·.
I .1,4.
_ b••
I
nd l.~-
to examination.
Another important difference is that water ments have decreased significantly recently; ,
I

it is to be hoped that instruments for routine liNO,. _NO.


is a weak scatterer of Raman radiation but a
-
strong absorber of infrared. Thus, aqueous
solutions of samples can be employed for
Raman studies. This advantage is particularly
applications will appc!ar shortly at a cost of
SSOOOto SI0,000. ---•.Br~,'''--
CI.altanos

_'.~Ik~-
I I
pronounced for the study of biological systems,
inorganic substances, and water pollution APPLICATIONS OF
~o.cm-l
problems. RAMAN SPECTROSCOPY
FIGURE 9-6 Correlation chart showing group frequencies for various organic functional
Liquid samples containing colloidal or sus-
Raman spectroscopy ~as been applied to the groups. (Cary Instruments, Monrovia, California, with permission.)
pended particles ordinarily scatter sufficient
amounts of the laser beam to make observa- qualitative and quantitative analysis of or-
tion of the Raman effect difficult or impos- ganic, inorganic, and biological systems.
sible. Such samples must be treated to remove Raman Spectra
Raman Spectra of of coordination compounds. For example,
solids before a Raman spectrum can be ob-
Inorganic Sped •• 4 numerous halogen and halogenoid complexes of Organic Specie.
tained. produce Raman spectra and thus are suscep-
The Raman technique is often superior to the Raman spectra are similar to infrared spectra
tible to investigation by this means. Metal-
Comperlson of Raman infrared for investigating inorganic systems in that they have regions that are useful for
oxygen bonds are also active. Spectra for such
and Infrared Instrumentation' because aqueous solutions can be employed. functional group detection and fingerprint re-
species as VO:-, AI(OH)i, Si(OH)~-, and
gions that permit the identification of specific
H. J. Sloane has compared Raman and infra- In addition, the vibrational energies of metal- Sn(OH~- have been obtained; Raman stud-
compounds. Figure 9-6 shows a correlation
red spectroscopy on the basis of instrumenta- ligand bonds are generally in the range of 100 ies have permitted conclusions regarding the
chart that can be employed for functional
tion, sample handling, and applicability.3 to 700 em - " a region of the infrared that is probable nature of such species. For example,
group recognition. DolIish' has published a
Optics. Raman spectrophotometers are experimentally difficult to study. These vibra- in perchloric acid solutions, vanadium(IV)
comprehensive treatment of functional group
somewhat less complicated from the stand- tions are frequently Raman-active, however, appears to be present as V02 + (aq) rather
and peaks with !:J.(1 .values in this range are frequencies. Catalogs of Raman spectra for
point that glass or quartz optics can be used than as V(OHW (aq); studies of boric acid
organic compounds are also available.6
throughout. Furthermore, Raman spectra in readily observed. Rafuan studies are poten- solutions show that the anion formed by acid Raman spectra yield more information
both the mid- and the far-infrared regions tially useful sources of information concern- dissociation is the tetrahedral B(OH)i rather about certain types of organic compounds
(4000 to 25 em-I) can be examined with a ing the composition, !structure, and stability than H2BOi. Dissociation constants for
single optical system; in contrast, infrared strong acids such as H2S04, HN03, H2Se04
studies require several gratings to cover the and H,I06 have been obtained by Raman
• F. R. Dollish, W. G. Faleley, and F. F. Bentley, Cllarac-
same range. The grating for a Raman instru- • For a review of inorganic applications, see: K. Naka- measurement. teristic Raman Frequencies of Organic Compounds. New
moto. Infrared lUll! Raman Spectra of 11IOrgtl1lreand Co- It seems probable that the future will see York: Wiley-Inlerscienc:e. 1971.
ordination Compounds. 3d ed. New York: Wiley, 1978; D. even wider use of Raman spectroscopy for 6 Samuel P. Sadtlcr and Sons, Inc, 2100 Arch Slreel,
J. Gardiner, AIUII. Chem, !G, I3IR (1978); and W. E. L Philadelphia. PA.
theoretical studies of inorganic systems.
Grossman, Anal. Chem, 48, 261R (1976); 46, 34SR (1974~
than do their infrared counterparts. For ex- to attack by moisture, and small amounts of past the laser beam or to use a pulsed beam.
ample, the double-bond stretching vibration water in a sample do not interfere. Despite Coherent ~okes Raman Spectroscopy
for olefins results in weak and sometimes un- these advantages, Raman spectroscopy has not (CARS).10 This technique has been employed Vr = v" - V~

detected infrared absorption. On the other yet: been exploited widely for quantitative to overcome some of the drawbacks of con- Under these circumstances, a beam or radia-
hand, the Raman band, (which, like the infra- analysis. . ventional Raman spectroscopy, namely, its tion is generated having a frequency v that is
red band, ocx:urs at about 1600 em-I) is in- An example of the potentialities of the low efficiency, its limitation to the visible and given by •
tense, and its position is sensitive to the nature method for the analysis of mixtures is near-ultraviolet regions, and its susceptibility
of substituents as well as to their geometry. provided in a paper by Nicholson, 8 in which to interference from fluorescence. V. = 2v" - v~
Thus, Raman studies are likely to yield useful a procedure for the determination of the con- One way of carrying out a CARS experi- This radiation differs from normal Raman
information about the olefinic functional group stituents in an eight-component mixture is ment is to employ a laser to pump a tunable radiation in that it is coherent rather than
that may not be revealed by infrared spectra. described. The components included benzene, dye laser. The two lasers: are arranged in such scattered in all directions. Furthermore, it is
This statement applies to cycloparaffin deriva- isopropyl benzene, three diisopropyl ben- a way that part of the beam from the pump- emitted at an angle such that it can be readily
tives as well; these compounds have a charac- zenes, two triisopropyl derivatives, and 1,2,4,5- ing laser of frequency: v" is employed for separated from the excitation beam without a
teristic Raman peak in the region of 700 to tetraisopropyl benzene. The power of the pumping and part for sample excitation. The monochromator. Finally, because it occurs in
1200 em - I. This peak has been attributed to characteristic peaks, compared with the power dye laser is also employed for sample excita- the anti-Stokes region (a similar beam occurs
a "breathing" vibration in which the nuclei of a reference peak (CCI4~ was assumed to tion, and its frequency ,~ is varied until the at v. == 2v" + v~in the Stokes region~ interfer-
move in and out symmetrically with respect to vary linearly with volume percent of each com- difference between the two excitation frequen- ence from Stokes fluorescence by the sample
the center ofthe ring. The position ofthe peak ponent. Analysis of synthetic mixtures of all cies is equal to the resonance frequency Vr for is avoided. .
decreases continuously from 1190 em -1 for the components by the procedure produced The efficiency of coherent anti-Stokes emis-
cyclopropane to 700 em-I for cyclooctane; results that agreed with the preparatory data
sion is high, with as much as 1 % of the excita-
Raman spectroscopy thus appears to be an to about 1% (absolute). tion radiation being converted to the new
excellent diagnostic tool for the estimation of frequency. Furthermore, because the beam is
ring size in paraffins. The infrared peak asso-
i
Application of Other 10 For brief reviews. see: R. F.; Begley, A. B. Harvey, R. coherent, high detection efficiencies are real-
ciated with this vibration is weak or non- L. Beyer, and B. S. Hudson, AlIJ<'r. Lab, 6 (I"~ II (1974); ized. As a consequence of these two proper-
Types of Raman Spectroscopy
existent. and A. B. Harvey, A ••••/. Cltem, 50 (9~ 90SA (1978). ties, sensitivity is enhanced.
With the development of tunable lasers, two
Biological Applications new Raman spectroscopic methods were de-
of Raman Spectroscopy veloped in the early 19705. Neither has as yet
become widely employed. Thus. only brief
Raman spectroscopy has been applied advan- 1. For two energy levels separated by the energy 11£, the population of
mention will be made of the techniques,
tageously for the study of biological systems.' the higher state compared with that of the lower state at thermal
Resonance Raman Spectrascopy.9 Reso-
The advantages of this technique include the equilibrium is given by the Boltzmann distribution:
nance Raman spectroscopy is based upon the
small sample requirement, the minimal sensi-
observation that a laser source with a frequency population of higher state _ _ M:/lT
tivity toward interference by water, the spec-
near or coincident with an electronic absorp- population of lower state - e
tral detail, and the conformational and en-
tion peak of a substance causes a marked en-
vironmental sensitivity. where k is Boltzmann's constant and T is the temperature in OK. The
hancement ofthe Raman scattering associated
with the chromophore. Significant increases in intensity of a Raman line is proportional to the population of
Quantitative Applications ·molecules from which the transition originates. If all other factors are
sensitivity and selectivity result. Because local
Raman spectra tend to be less cluttered with heating is likely to bring about sample decom- 'equa~ calculate the ratio of anti-Stokes intensity to that of Stokes
peaks than infrared spectra. As a consequence, position when this technique is employed, it is .intensity for the CCl4 lines at 218 and 459 em - I at 3OOoK.
peak overlap in mixtures is less likely and often necessary to rotate or flow the sample 2. ITwo reasonable structures for the Cu(CN)i- ion are planar or bent.
quantitative measurements are simpler. In ad-
dition, Raman instrumentation is not subject N
HI

• D. E. Nicholson. Anal. Cltem, 32, 1634 (1960~


~
Cu
1 For a review of biological applications. see: B. P. 9 For a brief review of this topic, see: D. P. Sirommen c:" 'c
Gaber, Amer. Lab, 9 (3~ IS (1977~ and K. Nakamoto, J. Cltem. Edue, 54, 474 (1977). ~ ~
Each of these species is expected to have three types of C-N stretch-
ing modes, two of which have the same energy. Vibrations with the·
same energy are said to be degenerate and can give only a single IR or
Raman band. The three C-N stretching modes are therefore expected
to give, at most, two bands in the C-N stretching region (near
2100 em -I) of the IR or Raman spectrum. The symmetric and asym-
metric stretching modes of the planar species are as follows:

symmetric one of two degenerate


stretching asymmetric stretching motions FIGURE 9-7 IR and Raman spectra of a compound with the
formula C2H402S. (Reproduced with permission of the Sadtler
Research Laboratories, Philadelphia, Pennsylvania.)
The asymmetric stretching mode is allowed by both the IR and Raman
selection rules for both the planar and bent structures. Using reasoning
analogous to that for Figure 9-4, decide whether the symmetric stretch-
ing is allowed in either the IR or Raman spectrum, or in both. Use
this reasoning for both the planar and bent structures.
The IR spectrum ofCu(CN)J- in aqueous solution exhibits one band
in the C-N stretching region at 2094 em -I. The Raman spectrum
. ..
A<aptlhr,
_1It.Ey
Cdl:
COU:
•• t
0.015_

exhibits bands at 2094 and 2108 em-I. Is Cu(CN)J- planar or beDt?


3. Suggest structures for the compounds whose lR and Raman spectra
are shown in Figures 9-7 through 9-14.

FIGURE 9-8 IR and Raman spectra of a compound with the


formula C3H402• (Reproduced with permission of the Sadtler
Research Laboratories, Philadelphia, Pennsylvania.)
FIGURE 9-9 IR and Raman spectra of a compound with the FIGURE 9-11 IR and Raman spectra of a compound with
formula C7H,N04• (Reproduced with permission; of the Sadder the formula CaHaOz . (Reproduced with permission of the"Sadder
Research Laboratories, Philadelphia, Pennsylvania.~ Research Laboratories, Philadelphia, Pennsylvania.)
i

I-
' '-

4"d
_.~
C-
-

FIGURE 9-10 IR and Raman spectra of a compound with FIGURE 9-12 IR and Raman spectra of a compound with
the formula C4Hz03• (Reproduced with permission of the Sadder the formula C3H7NO. (Reproduced with permission of the Sadtler
Research Laboratories, Philadelphia, Pennsylvania.) Research Laboratories, Philadelphia, Pennsylvania.)
FIGURE 9-13 IR and Raman spectra of a compound with
the formula C,H,NO. (Reproduced with permission ohhe Sadtler
Research Laboratories, Philadelphia, Pennsylvania.)

FIGURE 9-14 IR and Raman spectra of a compound with


the formula C3H3N. (Reproduced with permission of the Sadtler
Research Laboratories, Philadelphia, Pennsylvania.)
Many chemical systems are photoluminescent; orbital have opposing spins (the spins are singlet transition. As a consequence, the aver-
that is, they can be excited by electromagnetic then said to be paired). As a consequence of age lifetime of an excited triplet state may
radiation and, as a consequence, reemit radia- Fluorescent behavior occurs in simple as well spin pairing, most molecules have no net elec- be as long as a second or more, as compared
tion either of the same or longer wavelengths. as in complex gaseous, liquid, and solid chemi- tron spin and are, therefore, diamagnetic.2 A with an average lifetime of about 10- 8 S for an
The two most common manifestations of pho- cal systems, The simplest kind of fluorescence molecular electronic state in wliich all elec- excited singlet state. Furthermore, radiation-
toluminescence are fluorescence and plwspho- is that exhibited by dilute atomic vapors. For tron spins are paired is called a singlet state, induced excitation of a ground-state molecule
rescence which, as we shall see, are produced example, the 3s electrons of vaporized sodium and no splitting of the energy level occurs to an excited triplet state does not occur
by somewhat different mechanisms. The two atoms can be excited to the 3p state by absorp- when the molecule is exposed to a magnetic readily, and absorption peaks due to this
phenomena can be distinguished experimental- tion of radiation of 5896 and 5890 A. After field (we are here neglecting the effects of process are several orders of magnitude less
ly by observing the lifetime of the excited state. approximately 10-8 s, the electrons re9irn to nuclear spin). The ground state for a free intense than the analogous singlet-singlet tran-
With fluorescence, the luminescent process the ground state, and in so doing emit radia- radical, on the other hand, is a doublet state; sition. We shall see, however, that an excited
Ceases almost immediately « 10-6 s) after tion of the same two wavelengths in all direc- here, the odd electron can assume two orien- triplet state can be populated from an excited
irradiation is discontinued; phosphorescence tions. This type of fluorescence, in which the tations in a magnetic field and thus gives rise singlet state of certain molecules; the conse-
ilsually endures for an easily detectable length absorbed radiation is reemitted without altera- to a splitting of the energy level. quence of this process is phosphorescent
of time. Fluorescence is more important than tion, is known as resonance radiation or res- When one of the electrons of a molecule is behavior.
phosphorescence from an analytical stand- onance fluorescence. excited to a higher energy level, a singlet or a Figure 10-1 is a partial energy-level diagram
Point and will be emphasized in the discussion Polyatomic molecules or ions also exhibit triplet stale can result. In the excited singlet for a typical photoluminescent molecule. The
that follows. resonance radiation; in addition, charader- state, the spin of the promoted electron is still lowest heavy horizontal line represents the
Measurement of fluorescent intensity per- istic radiation oflonger wavelengths is emitted. paired with the ground-state electron; in the ground-state energy of the molecule, which is
mits the quantitative determination of many in- This phenomenon is called the Stokes shift, triplet state, however, the spins of the two normally a singlet state and is labeled So.
organic and organic species in trace amounts; electrons have become unpaired and are thus At room temperature, this state represents the
inany useful fluorometric methods exist, par- parallel. These states can be represented as energies of essentially all of the molecules in
Excited Stat ••
ticularly for biological systems. follows: a solution.
, One of the most attractive features offluoro- To understand the characteristics of fluo- The upper heavy lines are energy levels for
inetry is its inherent sensitivity. The lower rescence and phosphorescence phenomena, -L --.L the ground vibrational states of three excited
limits for the method frequently are less than recall that (p. 170) a bond between two atoms i.-.L --.L --.L electronic states. The two lines on the left
those for an absorption method by a factor of consists of one or more molecular orbitals Ground Excited Excited represent the first (S.) and second (S2) elec-
ten or better and are in the range between a formed from the overlap of the atomic orbit- singlet state singlet state triplet state
tronic singlet states. The one on the right (7iJ
few thousandths to perhaps one-tenth part per als of the electron pair that forms the bond. represents the energy of the first electronic
The nomenclature of singlet, doublet, and
million. In addition, lC1ec:tivityis at least as Combination of two atomic orbitals gives rise triplet state. As is normally the case, the
good and may be better than that of other triplet comes from spectroscopic multiplicity
to both a bonding and an antibonding molec- energy of the first excited triplet state is lower
methods. Fluorometry, however, is less widely considerations which need not concern us
ular orbital; the former has the lower energy than the energy of the corresponding singlet
applicable than absorption methods because here.
and is thus occupied by the electrons in the state.
of the relatively limited number of chemical The properties of a molecule in the excited
ground state. Superimposed upon the elec- Numerous vibrational energy levels are as-
systems that can be made to fluoresce.t triplet state differ significantly from those of
tronic energy level of each molecular orbital is sociated with each of the four electronic
the corresponding singlet state. For example,
a series of closely spaced vibrational energy states, as suggested by the lighter horizontal
a molecule is paramagnetic in the former and
levels. As a consequence, each electronic ab- lines.
1 For further discussion or fluorescence and phospho- diamagnetic in the latter. More important,
sorption band contains a series of closely As shown in Figure 10-1, excitation of this
rcsccncc theory and applications, sce: G. G. Guilbault, however, is the fact that a singlet-triplet tran-
PrtlClical Fluoresunce: Theory, Methods, and Technique. spaced vibrational peaks corresponding to molecule can be brought about by absorption
sition (or the reverse) that also involves a
:New York: Marcel Dekker, 1913; S. G. Schulman, Flu- transitions from the ground state to the several of two bands of radiation, one centered about
change in electronic state is a significantly less
lwesulICe and Phosphorescence Spectroscopy: Physio- vibrational levels of an excited electronic state. the wavelength At (So ....•St) and the second
~hmlical Prlncip/u tmd PrtlClice. :New York: Perpmon Most molecules contain an even number of probable event than the corresponding singlet-
Press, 1977; Motion FlUorescence Spectroscopy, cd. E. L around the shorter wavelength A2 (So ....•S2)'
electrons; in the ground state, these electrons Note that the excitation process can result in
Wehry. :New York: Plenum Press, 1976; and J. D.
Winefordner, S. G. Schulman, and T. C. O'Haver, Lumi- exist as pairs in the various atomic or molecu- conversion of the molecule to any of the
2 Diamagnetic species tend to move out from magnetic
nescence Spectrometry in Analytical Chemistry. :NewYork: lar orbitals. The Pauli exclusion principle several excited vibrational states. Note also
fic1ds,whereas paramagnetic species tend to move into a
Wiley-Interscicncc,1972. demands that the two electrons in a given that direct excitation to the triplet state does
field.
perhaps not surprising that a simple inverse version is employed to describe intermolecular
relationship exists between the lifetime of an processes by which a molecule passes to a
excited state and the molar absorptivity of the lower energy electronic state without emission
absorption peak corresponding to the excita- of radiation. These processes are neither well
tion process. This relationship (and experi- defined nor well understood, but it is apparent
ment as well) shows that typical lifetimes of that they are often highly efficient, since rela-
excited states that are deactivated by emission tively few compounds exhibit fluorescence.
are 10-7 to 10-9 s, when the molar absorp- Internal conversion appears to be partic-
tivities range between 103 and 105; for weakly ularly efficient when two electronic energy
absorbing systems, where the probability of levels are sufficiently close for the existence of
the transition process is low, the lifetimes may an overlap in vibrational levels. This situation
be as long as 10-6 to 10-5 s. Any deactiva- is depicted for the two excited singlet states
tion process that occurs in a shorter time thus in Figure 10-1. At the overlaps shown, the
reduces the fluorescent intensity. potential energies of the two excited states
Vibrational Relaxatioa. As shown in are identical; this equality apparently permits
Figure 10-1, a molecule may be promoted to an efficient transition. Internal conversion
any of several vibrational levels during the through overlapping vibrational levels is
electronic excitation process. In solution, usually more probable than the loss of energy
however, the excess vibrational energy is im- by fluorescence from a higher excited state.
So
mediately lost as a consequence of collisions Thus, referring again to Figure 10-1, excita-
between the molecules of the excited species tion by radiation of 12 frequently produces
Ground
state and those of the solvent; the result is an fluorescence of wavelength 13to the exclusion
energy transfer and a minuscule increase in of a band that would result from a transition
temperature of the solvent. This relaxation between S2 and So' Here, the excited mole-
process is so efficient that the average lifetime cule proceeds from the higher electronic state
of a vibrationally excited molecule is 10- u s to the lowest vibrational state of the lower
or less, a periQd significantly shorter than the electronic excited state via a series of vibra-
not occur to any significant extent because tionless path has a more favorable rate average lifetime of an electronically excited tional relaxations, an internal conversion, and
this process involves a change in multiplicity, constant, fluorescence is either absent or less state. As a consequence, fluorescence from then further relaxations. Under these circum-
an event which, as we have mentioned, has a intense. solution, when it occurs, always involves a stances, the fluorescence would be of 13only,
low probability of occurrence. The fluorescence phenomenon is limited to transition from the lowest vibrational level of regardless of whether radiation of wave-
a relatively small number of systems incor- an excited state. Several closely spaced peaks length Al or 12 was responsible for the excita-
porating structural and environmental fea- are produced, however, since the electron can tion.
o..ctivnion Processes
tures that cause the rate of radiationless return to anyone of the vibrational levels of The mechanisms of the internal conversion
An excited molecule can return to its ground relaxation or deactivation processes to be the ground state (Figure 10-1~ whereupon it processes S I -+ So and TI -+ So shown in
state by a combination of several mechanistic slowed to a point where the emission reaction will rapidly fall to the lowest ground state by Figure 10-1 are not well understood. The
steps. As shown by the straight vertical can compete kinetically. Information concern- further vibrational relaxation. vibrational levels of the ground state may
arrows in Figure 10-1, two of these steps, flu- ing the emission process is sufficiently com- A consequence of the efficiency of vibra- overlap those of the first excited electronic
orescence and phosphorescence, involve the plete to permit a quantitative accounting of tional relaxation is that the fluorescent band state; under such circumstances, deactivation
release of a photon of radiation. The other its rate. Understanding of other deactivation for a given electronic transition is displaced will occur rapidly by the mechanism just
deactivation steps, indicated by wavy arrows, routes, however, is rudimentary at best; for toward lower frequencies or longer wave- described. This situation prevails with aliphat-
are radiationless processes. The favored route these, only qualitative statements or specula- lengths from the absorption band; overlap ic compounds, for example, and accounts for
to the ground state is the one that minimizes tions can be put forth. Nevertheless, the inter- occurs only for the resonance peak involving the fact that these species seldom fluoresce;
the lifetime of the excited state. Thus, if deac- pretation of fluorescence requires considera- transitions between the lowest vibrational that is, deactivation by energy transfer through
tivation by fluorescence is rapid with respect tion of these other processes. level of the ground state and the correspond- overlapping vibrational levels occurs so rapidly
to the radiationless processes, such emission Emission Rate. Since fluorescent emission ing level of the excited state. that fluorescence is prevented.
is observed. On the other hand, if a radia- is the reverse of the excitation process, it is Internal Conversion. The term internal con- Internal conversion may also result in the
phenomenon of predissociation. Here, the elec- sition shown in Figure 10-1 is an example; molecules (the q~ntum yield for phospho- magnitude. As a consequence, fluorescence
tron moves from a higher electronic state to here, the lowest singlet vibrational state over· rescence can be defined in an analogous way). due to u· ....• u transitions is seldom observed;
an upper vibrational level of a lower electronic laps one of the upper triplet vibrational levels For a highly fluorescent molecule such as flu- instead, such emission is confined to the less
state in which the vibrational energy is great and a change in spin state is thus more prob- orescein, the quantum efficiency undc;r some energetic It· ....•It and It· ....•n processes (see
enough to cause rupture of a bond. In a large able. conditions approaches unity. Chemical ~pecies Figure 7-3, p. 171 for the relative energies as-
molecule, there is an appreciable probability Intersystem crossings are most common in that do not fluoresce appreciably have an
effi- sociated with these transitions).
for the existence of bonds with strengths less molecules that contain heavy atoms, such as ciency that approaches zero. As we have noted, an electronically excited
than the electronic excitation energy of the iodine or bromine~ Apparently interaction be- From a consideration of Figure 10-1 and molecule ordinarily returns to its lowest
chromophores. Rupture of these bonds can tween the spin and orbital motions becomes our discussion of deactivation prOcesses, it is excited state by a series of rapid vibrational
occur as a consequence of absorption by the large in the presence of such atoms, and a apparent that the fluorescent quantum yield relaxations and internal conversions that pro-
chromophore followed by internal conversion change in spin is thus more favored. The 4> for a compound must be determined by the duce no emission of radiation. Thus, any ftu-
of the electronic energy to vibrational energy presence of paramagnetic species such as relative rates for the processes by which the orescence observed most commonly arises
associated with the weak bond. molecular oxygen in solution also enhances lowest excited singlet state is deactivated- from a transition from the first excited state
A predissociation should be differentiated intersystem crossing and a consequent decrease namely, ftuorescence, intersystem crossing, ex- to the ground state. For the majQrity of ftuo-
from a dissociation, in which the absorbed in fluorescence. ternal and internal conversion, predis- rescent compounds then, radiation is prod uced
radiation excites the electron of a chromo- PbosphoresceDCe. Deactivation may also sociation, and dissociation. We may express by deactivation of either the n,lt~ or the It,lt*
phore directly to a sufficiently high vibra- involve phosphorescence. After intersystem these relationships by the equation excited state, depending upon which of these
tional level to cause rupture of the chromo- crossing to an excited triplet state, further is the less energetic.
photic bond; no internal conversion is in- deactivation can occur either by internal Quantum EllideDcy ud TrusitiOll Type. It
4> = kc (10-1)
volved. Dissociation processes also compete or external conversion or by phosphorescence. kc + k; + k.c + kic + kpd + kd ·is observed empirically that fluorescent be-
with the fluorescence process. A triplet-singlet transition is much less prob- havior is more commonly found in~compounds
Extemal Com'ersion. Deactivation of an able than a singlet-singlet conversion, and the where the k terms are the respective rate con- in which the lowest energy exciteji state is or
excited electronic state may involve interac- average lifetime of the excited triplet state stants for the several processes enumerated a 1t,1t· type than in those with a IQwest energy
tion and energy transfer between the excited with respect to emission ranges from 10-4 to above. n,lt· state; that is, the quantum ,efficiency IS
molecule and the solvent or other solutes. several seconds. Thus, emission from such a Equation 10-1 permits a qualitative inter- greater for It· ....•It transitions.
These processes are called external conver- transition may persist for some time after pretation of many of the structural and envi- The greater quantum efficiency associated
sions. Evidence for external conversion in- irradiation has been discontinued. ronmental factors that inftuence fluorescent with the 1t,1t· state can be rationalized in two
cludes the marked effect upon fluorescent External and internal conversions compete intensity. Clearly, those variables that lead to ways. First, the molar absorptivity of a It ....•It·
intensity exerted by the solvent; furthermore, so successfully with phosphorescence that this high values for the ftuorescence rate constant transition is ordinarily 100- to lOOO-fokl
those conditions that tend to reduce the kind of emission is ordinarily observed only kc and low values for the other k terms en- greater than for n ....•It· process, and this quan-
number of collisions between particles (low at very low temperatures or in viscous media. hance ftuorescence. The magnitude of kc, the tity represents a measure of transition prob-
temperature and high viscosity) generally lead predissociation rate constant kpd , and the dis- ability in either direction; thus, the inherent
to enhanced fluorescence. The details of exter- sociation rate constant kd are mainly depend- lifetime associated with a It ....•It· transition is
nal conversion processes are not well under- Variabl •• that Affect ent upon chemical structure; the remaining shorter (10-7 to 10-9 s compared with 10-1
stood. Fluorescence and Phosphoreacence constants are strongly inftuenced by environ- to 10-7 s for an n,lt· state) and kr in Equation
Radiationless transitions to the ground ment and to a somewhat lesser extent by 10-1 is larger.
state from the lowest excited singlet and trip- Both molecular structure and chemical envi- structure. It is also believed that the rate constant for
let states (Figure 10-1) probably involve ex- ronment are inftuential in determining whether Tra.ilion Types in Fluorescence. It is im- intersystem crossing kj is smaller for n,n·
ternal conversions, as well as internal a substance will or will not fluoresce (or portant to note that fluorescence seldom re- excited states because the energy difference
conversions. phosphoresce ~ as well as the intensity of the sults from absorption of ultraviolet radiation between the singlet-triplet stat~ is larger;
Intersystem Crossing. Intersystem crossing emission that might occur. The effects of of wavelengths lower than 250 nm, because that is, more energy is required td unpair the
is a process in which the spin of an excited some of these variables are considered briefly such radiation is sufficiently energetic to electrons of the It,n* excited state. As a con-
electron is reversed and a change in multi- in this section. cause deactivation of the excited states by sequence, overlap of triplet vibrational levels
plicity of the molecule results. As with inter- Quantum Yield. The quantum yield, or predissociation or dissociation. For example, with those of the singlet state is less, and the
nal conversion, the probability of this transi- quantum efficiency, for a fluorescent process is 200-nm radiation corresponds to about 140 probability of an intersystem crossing is
tion is enhanced if the vibrational levels of simply the ratio of the number of molecules kcaljmol; most molecules have at least some smaller.
the two states overlap. The singlet-triplet tran- that fluoresce to the total number of excited bonds that can be ruptured by energies of this In summary, then, fluorescence is more
commonly associated with 7t,n· states than ph ene, and pyrrole, do not exhibit fluorescent The influence of halogen substitution is
with n,n* statc5 because the fonner possess behavior; on the other hand, fused-ring struc- striking; the decrease in fluorescence with in-
shorter average lifetimes (k, is larger) and be- tures ordinarily do. With nitrogen hetero- creasing atomic number of the halogen is
cause the deactivation processes that compete cyclics, the lowest-energy electronic transition thought to be due in part to the heavy atom
with fluorescence are less likely to occur. is believed to involve an n....• n· system which effect (p. 2841 which increases the probability
F'Iuoresceucle •••• Structure. The most rapidly converts to the triplet state and pre- for intersystem crossing to the triplet state.
intense and most useful fluorescent behavior vents fluorescence. Fusion of benzene rings to a Predissociation is thought to play an impor-
is found in compounds cont;lining aromatic heterocyclic nucleus, however, results in an tant role in iodobenzene and in nitro deriva-
functional groups with low-energy n ....•n* increase in the molar absorptivity of the tives as well; these compounds have easily Lack of rigidity in a molecule probably
transition levels. Compounds containing ali- absorption peak. The lifetime of an excited ruptured bonds that can absorb the excitation causes an enhanced internal conversion rate
phatic and alicyclic carbonyl structures or state is shorter in such structures; fluorescence energy following internal conversion. (kic in Equation 10-1) and a consequent in-
highly conjugated double-bond structures is thus observed for compounds such as Substitution of a carboxylic acid or car- crease in the likelihood for radiationless de-
may also exhibit fluorescence,. but the number quinoline, isoquinoline, and indole. bonyl group on an aromatic ring generally activation. One part of a nonrigid molecule
of these is small compared with the number Substitution on the benzene ring causes inhibits fluorescence. In these compounds, the can undergo low-frequency vibrations with
in the aromatic systems. shifts in the wavelength of absorption maxima energy of the n,n· syStem is less than in the respect to its other parts; such motions un-
Most unsubstituted aromatic hydrocarbons and corresponding changes in the fluorescence n,n· system; as we have pointed out earlier, doubtedly account for some energy loss.
fluoresce in solution, the quantum efficiency peaks. In addition, substitution frequently the fluorescent yield from the fonner type of Temperatwe •••• Solvent Elects. The quan-
usually increasing with the number of rings affects the fluorescent efficiency; some of these system is ordinarily low. tum efficiency of fluorescence by most mole-
and their degree of condensation. The simplest effects are illustrated by the data for benzene Effect of Structural Rigidity. It is found cules decreases with increasing temperature
heterocyclics, such as pyridine, furan, thio- derivatives in Table 10-1. experimentally that fluorescence is particularly because the increased frequency of collisions
favored in molecules that possess rigid struc- at elevated temperatures improves the prob-
tures. For example, the quantum efficiencies ability for deactivation by external conversion.
for fluorene and biphenyl are nearly 1.0 and A decrease in solvent viscosity also increases
TABLE 10-1 EFFECT OF SUBSTITUTION ON THE 02, respectively, under similar conditions of the likelihood of external conversion and leads
FLUORESCENCE OF BENZENEob to the same result.
measurement.
The polarity of the solvent may also have
Wavelength of Relative Intensity
Compou'"
Benzene
Formula

C6H6
FIuoresceace,- DID

270-310
270-320
of Fluorescence

10
O=:J) Hz
an important influence. In Chapter 7, we
pointed out that the energy for n ....•n· tran-
sitions is often increased in polar solvents,
while that for an ....•n· transition suffers the
Toluene C6HsCH3 17
Propylbenzene C6H,C3H7 270-320 17 Fluorene opposite effect. Such shifts may occasionally
Fluorobenzene C6HsF 270-320 10 be great enough to lower the energy of the
Chlorobenzene C6HsCI 275-345 7 The difference in behavior appears to be n -+ n* process below that of the n ....•n· tran-
Bromobenzene C6H,Br 290-380 5 largely a result of the increased rigidity fur- sition; enhanced fluorescence results.
lodobenzene C6H,I 0 nished by the bridging methylene group in The fluorescence of a molecule is decreased
Phenol C6H,OH 285-365 18 fluorene. Many similar examples can be cited. by solvents containing heavy atoms or other
Phenolate ion C6H,O- 31{}-4()() 10 In addition, enhanced emission frequently re- solutes with such atoms in their structure;
Anisole ;C6H,OCH3 285-345 20' sults when fluorescing dyes are adsorbed on a carbon tetrabromide and ethyl iodide are
Aniline C6H,NHz 310-405 20 solid surface; here again. the added rigidity examples. The effect is similar to that which
Anilinium ion iC6H,NHj 0 provided by the solid surface may account for occurs when heavy atoms are substituted into
Benzoic acid ,C6HsCOOH 310-390 3 the observed effect. fluorescent compounds; orbital spin interac-
Benzonitrile C6HsCN 280-360 20 The influence of rigidity has also been tions resuh in an increase in the rate of triplet
Nitrobenzene C6HsNOz 0 invoked to account for the increase in fluo- formation and a corresponding decrease in
rescence of certain organic chelating agents fluorescence. Compounds containing heavy
when they are complexed with a metal ion . atoms are frequently incorporated into sol-
• In ethanol solution.
For example, the fluorescent intensity of 8- vents when enhanced phosphorescence is
• Taken from W. West, Chemical Applicatimu o/Spectroscopy (Techniques ojOrgonic Chemistry, Vol.
IX). New York: Intencience Publishers, Inc, 19S6, p. 730. With permission. hydroxyquinoline is much less than that of desired.
El'ect of pH 011 Fluorescence. The fluores- F is proportional to the radiant power of the Two other factors responsible for further Exciwion
filter or
cence of an aromatic compound with acidic excitation beaID that is absorbed by the negative departures from linearity at high monochromotor
or basic ring substituents is usually pH- system. That is, concentration are self-quenching and self-
dependent. Both the wavelength and the absorption. The former is the result of col-
emission intensity are likely to be different for F = K'(Po - P) (10-2) lisions between excited molecules. Radiation-
the ionized and nonionized forms of the com- where Po is the power of the bean1 incident less transfer of energy occurs, perhaps in a
pound. The data for phenol and aniline upon the solution and P is its power after fashion analogous to the transfer to solvent
shown in Table 10-1 illustrate this effect. The traversing a length b of the medium. The con- molecules that occurs in an external con-
changes in emission exhibited by these com- stant K' depends upon the quantum efficiency version. Self-quenching can be expected to in-
pounds are analogous to the absorption of the fluorescence process. In order to relate crease with concentration.
changes that occur \\lith acid-base indicators; F to the concentration c of the fluorescing Selkbsorptic:m occurs when the wavelength
indeed, fluorescent indicators have been pro- particle, we write Beer's law in the form of emission overlaps an absorption peak; fluo-
posed for acid-base titrations in highly colored rescence is the~decreased as the emitted beam
solutions. For example, fluorescence of the !- = 10-<6< (10-3) traverses the solution.
phenolic form of 1-naphthol-4-sulfonic acid is Po The effects 'of these phenomena are such
not detectable by the eye since it occurs in the where Il is the molar absorptivity of the fluo- that a plot relating fluorescent power to con-
ultraviolet region. When the compound is con- rescent molecules and ebc is the absorbance A. centration often exhibits a maximum.
verted to the phenolate ion by the addition of
base, the emission peak shifts to visible wave-
By substitution of Equation 10-3 into Equa- FIGURE 10-2 Components of a fluorom-
tion 10-2, we obtain eter or a spectrofluorometer.
lengths where it can readily be detected. It is INSTRUMENTS FOR
of interest that this change occurs at a different F = K'Po(1 - 10-<6<) (10-4)
FLUORESCENCE ANALYSIS
pH than would be predicted from the acid dis-
The exponential term in Equation 10-4 can be
sociation constant for the phenol. The expla- The various ebmponents of fluorescence in-
expanded to
nation of this discrepancy is that the acid dis- struments are similar to those found in ultra-
sociation constant for the excited molecule F = K'Po (10-5) violet-visible photometers or spectropho- The reference beam passes through an
differs from that for the same species in its tometers. Figure 10-2 shows a typical con- attenuator that reduces its power to approxi-
ground state. Changes in acid or base dissocia- x [2.3ebc _ (-2:cj2 figuration for these components in a jluo: mately that of the fluorescent radiation (the
tion constants with excitation are common and rometer or a spectrojluorometer. Nearly all flu- power reduction is usually by a factor of 100
are occasionally as large as four or five orders (-23ebc)3 -' .. J orescence instruments employ double-beam or more). The outputs from the reference and
of magnitude. optics as shown in order to compensate for sample phototubes are then fed into a dif-
3!
It is clear from these observations that fluctuations in the power of the source. The ference amplifier and the output into a meter
analytical procedures based on fluorescence Provided &be = A < 0.05, all of the subse- sample beam first passes through an excita- or recorder. Many fluorescent instruments are
frequently require close control of pH. quent terms in the brackets become small tion filter or monochromator which serves to of the null type, this state being achieved by
Effect of Dissolved OxygelL The presence with respect to the first; under these condi- transmit that part of the beam which will optical or electrical attenuators.
of dissolved oxygen often reduces the emis- tions, we may write excite fluorescence but exclude wavelengths The sophistication, performance character-
sion intensity of a fluorescent solution. This which are subsequently produced by the ir- istics. and costs of fluorometers and spectra-
F = 2.3K'&bcPo (10-6)
effect may be the result of a photochemically radiated sample. Fluorescent radiation is fluorometers differ as widely as do the corre-
induced oxidation of the fluorescent species. or at constant Po, . emitted by the sample in all directions but is sponding instruments for absorption measure-
More commonly, however, the quenching most conveniently observed at right angles to ments. Fluorometers are analogous to absorp-
takes place as a consequence of the para- the excitation; beam; at other angles, in- tion photometers in that filters are employed
magnetic properties of molecular oxygen that Thus, a plot of the fluorescent power of a creased scatteJing from the solution and the to restrict the wavelengths of the excitation
can be expected to promote intersystem cross- solution versus concentration of the emitting cell walls can be the cause of large errors in and emission beams. Spectrofluorometers are
ing and conversion of excited molecules to species should be linear at low concefltrations the measurement of fluorescent intensity. The of two types. The first employs a suitable
the triplet state. Other paramagnetic species c. When c becomes great enough so that the emitted radiation reaches a photoelectric filter to limit the excitation radiation and a
also tend to quench fluorescence. absorbance is larger than about 0.05, linearity detector after passing through a second filter grating or prism monochromator to isolate a
Effect of Concentration on Fluorescent is lost and F lies below an extrapolation of or monochromator that isolates a fluorescent peak of the fluorescent emission spectrum.
Inteusity. The power of fluorescent radiation the straight-line plot. peak for measurement. Several commercial spectrophotometers can
be purchased with adapters that permit their SoureeL In most applications, a more in- or silica are employed for fluorescence mea- ence beam and the fluorescent beam to strike
use as spectrofluorometers. tense source is needed than the" tungsten or . surements. Care must be taken in the design of the detector surface alternately, thus producing
True spectrofluorometers are specialized hydrogen lamp employed for the measurement the cell compartment to reduce the amount an ac signal whenever the power of one differs
instruments equipPed with two monochroma- of absorption. A mercury or xenon arc lamp of scattered radiation reaching the detector. from the other; the phase of the ac signal,
tors. One of these restricts the excitation radia- is commonly employed. Baftles are often introduced into 'the compart- however, will depend upon which is the
tion to a narrow band; the other permits the The xenon arc lamp produces intense ra- ment for this purpose. stronger. A phase-sensitive device is employed
isolation of a particular fluorescent wave- diation by the passage of current through an to translate this difference and its sign into a
length. Such instruments permit measurement atmosphere of xenon. The spectrum is con- deflection of a meter needle (not shown). The
Instrument Designs power ofthe reference beam can then be varied
of fluorescence, excitation, and absorption tinuous over the range between about 250
spectra. An excitation spectrum is obtained and 600 om, with the peak intensity occurring Fluorometers. Figure 10-3 is a schematic by rotation ofthe light earn, which mechanical-
with the emission monochromator fixed at a at about 470 om (see Figure 4-13, p. 110). In diagram of a double-beam fluorometer that ly increases or decreases the fraction of the
wavelength of strong fluorescence; a plot of some instruments, regularly spaced flashes are employs a mercury lamp and a single photo- reference beam that reaches the detector. The
fluorescence output as a function of the excita- obtained by discharging a capacitor through multiplier tube as a detector. Part; of the ra- earn is equipped with a linear dial, each incre-
tion wavelength is then obtained. A fluores- the lamp; high intensities result. In addition, diation from the lamp passes through a filter ment of which corresponds to an equal fraction
cence spectrum, on the other hand, involves a the output of the phototubes is then ac, which to the sample. The fluorescent radiation then of light.
fixed excitation wavelength while the fluores- can be readily amplified and processed. passes through a second filter to the detector. Accurate adjustment of any null device re-
cent wavelengths are scanned. With suitable Mercury arc lamps produce an intense line A reference beam is reftected off the mirrored quires that the condition of null be approach-
corrections for variations in source output as a spectrum. High-pressure lamps ("" 8 atmo- surface of the light earn to a lucite light pipe able from both directions. For a totally non-
function of wavelength, an absolute excitation spheres) give lines at 366,405,436, 546, 577, which directs it to the photomultiplier tube. fluorescent sample, however, no light would
spectrum for a substance is obtained which is 691, and n3 om. Low-pressure lamps, equip- The rotating light interrupter caUSC!ithis refer- reach the detector in one of the phases and
similar in appearance to the absorption spec- ped with silica windows, additionally provide
trum. intense radiation at 254 om. Inasmuch as
The selectivity provided by spectrofluorom- fluorescent behavior can be induced in most
eters is of prime importance to investigations fluorescing compounds by a variety of wave-
concerned with the electronic and structural lengths, at least one of the mercury lines
characteristics of molecules, and is of value in ordinarily proves suitable.
both qualitative and quantitative analytical A recent development has been the use of
work as well. For quantitative purposes, how- various types of lasers as excitation sources
ever, the information provided by simpler for fluorometry. Of particular interest is a
iDstruments is entirely satisfactory. Indeed, tunable dye laser employing a pulsed nitrogen
relatively inexpensive fluorometers that have laser as thC primary source. Radiation in the
been designed specifically to meet the measure- region between 360 and 650 nm is produced.
ment problems peculiar to fluorescent analysis Such a device eliminates the need for an
are frequently as specific and selective as excitation monochromator.
modified spectrophotometers. Filters and MonocbromatOl'S. Both inter-
The discussion that follows is largely fo- ference and absorption filters have been
cussed on the simpler instruments for fluo- employed in fluorometers. Most spectroflu-
rescence analysis. orometers are equipped with grating mono-
chromators.
Components of DetectOl'S. The typical fluorescent signal is
of low intensity; large amplification factors
Fluorometers and
Spectrofluorometers are thus required for its measurement.
Photomultiplier tubes have come into wide-
The components of fluorometers and spectro- spread use as detectors in sensitive flu- Filter
(passes onlv UV)
fluorometers differ only in detail from those orescence instruments.
of photometers and spectrophotometers; we Cells and Cell Compartments. Both cylin- FIGURE 10-3 Optical design of the Turner Model 110 ~Iuoromet~. (C~urtesy
need to consider only these differences. drical and rectangular cells fabricated of glass of Amsco Instrument Company (formerly G. K. Turner ASSOCIlltes
~ Carpmtena, CA.)
vary with wavelength and differ from instru- ties that are two to four orders of magnitude
ment to instrument. better than the corresponding spectrophoto-
A number of methods have been developed .metric procedures.
for obtaining a corrected spectrum, which is
the true fluorescent spectrum freed from in- Inorganic Analysis
strumental effects; many of the newer and
Inorganic fluorometric methods are of two
more .sophisticated commercial instruments
types. Direct methods involve the formation
provide a means for obtaining corrected
of a fluorescent chelate and the measurement
spectra directly.3
/ of its emission. A second group is based upon
Fluorescence the diminution of fluorescence resulting from
detector
the quenching action of the substance being
Detector
determined. The latter technique has been
~~ for
Ibsorbanee APPLICATIONS
most widely used for anion analysis.
OF FLUOROMETRY
CatiGDs that Form Fluorescent OJelates.
Two factors greatly limit the number of
Fluorometric methods are inherently appli- transition-metal ions that form fluorescent
FIGURE 10-4 A spectrofluorometer. Note that the instrument can cable to lower concentration ranges than are chelates. First, many of these ions are para-
also be employed for absorbance measurements. spectrophotometric determinations and are magnetic; this property increases the rate of
thus among the most sensitive analytical tech- intersystem crossing to the triplet state. Deacti-
niques available to the scientist. The basic vation by fluorescence is thus unlikely, al-
difference in the sensitivity between the two though phosphorescent behavior may be ob-
methods arises from the fact that the con- served. A second reason is that transition-metal
the null point could then be approached from simpler single-beam instruments. The cost of
centration-related parameter for fluorometry complexes are characterized by many closely
one direction only. To avoid the resultant such fluorometers ranges from a few hundred
F can be measured independently of the power spaced energy levels, which enhance the likeli-
error, a third beam of constant intensity (the dollars to perhaps $2000.
Spedr'OIIuorom«ers. Several companies of- of the source Po. In contrast, a spectro- hood of deactivation by internal conversion.
forward light path) is directed to the detector
photometric measurement requires evaluation Nontransition-metal ions are less susceptible
ill phase with the fluorescent beam so that, fer spectrofluorometers. An example of a
of both Po and P, because the concentration- to the foregoing deactivation processes; it is
under all conditions, some radiation strikes the typical simple instrument is shown in Figure
dependent parameter A is dependent upon the for these elements that the principal applica-
photomultiplier. Correction for the effect of 104. Two grating monochromators are em-
ratio between these two quantities. The sensi- tions of fluorometry are to be found. It is
the third beam on the measured fluorescence ployed. Radiation from the xenon lamp is
tivity of a fluorometric method can be im- noteworthy that nontransition-metal cations
is accomplished by setting the fluorescence dial dispersed by the first and serves to excite the
proved by increasing Po or by further amplify- are generally colorless and tend to form
to zero with a solvent blank or a nonfluorescing sample. The resulting fluorescent radiation,
ing the fluorescent signal. In spectrophoto- chelates which are also without color. Thus,
dummy cuvette in the cell compartment; the after dispersion by the second monochro-
metry, an increase in Po results in a propor- fluorometry often complements spectrophoto-
intensity of the reference beam can then be mator, is detected by a photocell. Readout is
tionate change in P and therefore fails to affect metry.
varied by means of the blank shutter until an by means of a meter or recorder.
A; thus, no improvement in sensitivity results. FJuorometric Reagents. 4 The most success-
optical null is indicated; this operation must Note that the instrument can be employed
Similarly, amplification of the detector signal ful fluorometric reagents for cation analyses
be carried out frequently during a set of for absorbance measurements by use of the
has the same effect on both P and Po and have aromatic structures with two or more
analyses. first monochromator" only.
results in no net gain with respect to A. Thus, donor functional groups that permit chelate
The single-detector aspect of this double- An instrument such as that shown in
fluorometric methods generally have sensitivi-
beam instrument imparts high reproducibil- Figure 10-4 provides perfectly satisfactory
ity, even with long-term changes in detector spectra for quantitative analysis. The spectra
sensitivity and output of the source. Thus, obtained will, however, not necessarily com- • For a more detailed discussion of ftuorometric: re-
only occasional checks of calibration curves pare well with spectra from other instruments agents, see: T. S. West, in Trace Characterization, Chemi-
with a standard are required. because the output depends not only upon • For a summary of correction methods, see: N. Wother- a:rl anti PhysiclJl, eds. W. W. Meinke and B. E. Scribner.
spoon. G. K. Oster. and G. Oster. in PhysicDI Methods of Washington, D.C.: National Bureau or Standards Mono-
The instniment just described is represent- the fluorescence but also upon the character- C"","istry. ods. A. Weissberaer and B. R. Rossiter. New graph 100. 1967. pp. 237-266; and C. E. White. in Trace
ative of the dozen or J1\ore fluorometers istics of the lamp, detector, and monochro- York: Wiley-Interscience, 1972, vol. I. Part III B. pp. AlIlI1ysis. ods. J. H. Yoe and H. J. Koch, New York:
available commercially. Some of these are mators. All of these instrument characteristics 460-462 and pp. 473-478. Wiley 1957, Chapter 7.
formation with the metal ion. The structures Organic Specie. late matter.9 The instrumentation required tered radiation, usually at a right angle to the
of four common reagents follow: for nephelometry is similar to the fluorom- incident beam. Nephelometry is generally more
The number of applications of ftuorometric
eters we have just considered, whereas a sensitive than turbidimetry for the same rea-
analysis to organic problems is impressive.
filter photometer can be used for turbidimet- sons that fluorometry is more sensitive than
Weissler and White have summarized the
ric measurements. It is convement to con- photometry.
most important of these in several tables.8
sider these two methods briefty at this point. The choice between a nephelometric and a
~ Under a heading of Organic and General Bio-
~ When light passes through a transparent turbidimetric measurement depends upon the
chemical Substances are found over one
medium in which solid particles are dispersed, fraction of light scattered. When scattering is
8-Hydroxyquinoline hundred entries that include such diverse sub-
part of the radiation is scattered in all direc- extensive, owing to the presence of many par-
(Reagent for AI, Be, and . stances as adenine, anthranilic acid, aromatic
tions, giving a turbid appearance to the ticles, a turbidimetric measurement is the
other metal ions) polycyclic hydrocarbons, cysteine, guanidine,
mixture. The diminution of power of a col- more satisfactory. If scattering is minimal and
indole, naphthols, certain nerve gases, proteins,

HO-ON~:h
limated beam as a result of scattering by par- the diminution in power of the incident beam
salicylic acid, skatole, tryptophan, uric acid,
ticles is the basis of turbidimetric methods. is small, nephelometric measurements provide
and warfarin. Some fifty medicir$l agents that
Nephelometric methods, on the other hand, more satisfactory results.
can be determined ftuorometrically are listed.
are based upon the measurement of the scat-
"=\~Na Included among these are adrenaline, a1kyl-
morphine, chloroquin, digitalis principles, ly- Theory of
Nepheiometry and Turbidimetry
Alizarin garnet R sergic acid diethylamide (LSD), penicillin,
(Reagent for AI, F-) phenobarbital, procaine, and reserpine. Meth- It is important to appreciate that the scatter-
• For a more complete discussion, see: F. P. Hochpang,
ods for the analysis of ten steroids and an in Treatise 011 APItlIytlcal Chemistry, eds. 1. M. Koltholl' ing associated with nephelometry and turbi-
~or(") . equal number of enzymes and cl;Ienzymes are
also found in these tables. Somo of the plant
and P. J. Elving. New York: Interscience, 1964, Part I,
vol. S, Chapter 63.
dimetry (in contrast to Raman spectroscopy)

~O~ products listed include chlorophyll, ergot


o alkaloids, rauwolfia serpentina alkaloids, fia-
vonoids, and rotenone. Some eighteen listings
Flavanol
for vitamins and vitamin products are also TABLE 10-2 SELECTED FLUOROMETRIC METHODS FOR
(Reagent for Zr and Sn)
included; among these are ascorbic acid, folic INORGANIC SPECIES-

o-
acid, nicotinamide, pyridoxal, riboflavin, thi·
o OH
amin, vitamin A, and vitamin Bu . W.,eleegth, nm
H I
C-~~ Without question, the most important ap- Seasitivity
Ion Reagent Absorption F1uorescence /-II1m1 Interference
~ HV plications of ftuorometry are in the analyses
of food products, pharmaceuticals, clinical AI3+ Alizarin garnet R 470 500 0.007 Be, Co, Cr, Cu,
Benzoin samples, and natural products. The sensitivity
(Reagent for B, Zn, Ge, and Si) F-, NO" Ni,
and selectivity of the method make it a particu- PO~-, Th, Zr
Selected ftuorometric reagents and their larly valuable tool in these fields. F- AI complex of 470 500 0.001 Be, Co, Cr, Cu,
applications are presented in Table 10-2. For Alizarin garnet R Fe, Ni, PO~-,
a more complete summary see Meites,5 St. (quenching) Th,Zr
John, 6 or the review articles in Analytical NEPHELOMETRY B40~- Benzoin 370 450 0.04 Be,Sb
Chemistry.' AND TURBIDIMETRY Cd 2+ 2-(o-Hydroxyphenyl)- 365 Blue 2 NH3
benzoxazole
Nephelometry and turbidimet~ are closely Li+ 8-Hydroxyquinoline 370 580 0.2 Mg
• L Mcites, Handbook of APItlIyticlJl Chemistry. New related analytical methods which are based Sn4+ Flavanol 400 470 0.1 F-, po~-,Zr
York: McGraw-Hill, 1963, pp. 6-178 to 6-181. _ upon the scattering of radiation by particu- Zn2+ Benzoin Green 10 B, Be, Sb,
• P. A. 51. John, in Trace APItlIysls,ed. J. D. Winefordner. Colored ions
New York: Wiley, 1976, pp. 263-271.
• A. Weiasler and C. E. White, in Handbook of APItlIytical
, C. M. O'DonneD and T. N. SoDe, APItlI.CIoem.,50, 189R Chemistry, ed. L Meites. New York: McGraw-Hill, 1963, • L Mcites, Handbook of Anolytical Chemistry. New York: McGraw-HiD Book Company, (DC., 1963, pp. 6-178 to 6-181.
(1978); 48, 17SR (1976~ pp. 6-182 to 6-196. With permission.
involves no net loss in radiant power; only Effect or PIU1ide Size on ScatteriJl&. The reflection. of unwanted radiation to the de- centration of a variety of ions can be deter-
the direction of propagation is affected. The fraction of radiation scattered at any angle tector. mined by using suitable precipitating reagents.
intensity of radiation appearing at any angle depends upon the size and shape of the par- Figure 10-5 shows a simple visual turbi- Conditions must be chosen so that the solid
depends upon the number of particles, their ticles responsible for the scattering; the effect dimeter. The viewing tube !s adjusted in the phase forms as a stable colloidal suspension.
size and shape, the relative refractive indexes is large. Since most analytical applications of suspension until the special, 8-shaped lamp Surface-active agents (such as gelatin) are
ofthe particles and the medium, and the wave- scattering involve the generation of a col- filament just disappears. The length of solu- frequently added to prevent coagulation of the
length of the radiation. The relationship among loidally dispersed phase in a solution, those tion is then related to concentration by cali- colloid. As noted earlier, reliable analytical
these variables is complex. A theoretical treat- variables that influence particle size during bration. This very elementary device yields data are obtained only when care is taken to
ment is feasible but seldom applied to specific precipitation also affect both turbidimetric remarkably accurate data -for the analysis of control scrupulously all of the variables that
analytical problems because of its complexity. and nephelometric measurements. Thus, such sulfate in low concentrations. Here, a BaS04 affect particle size.
In fact, most nephelometric and turbidimetric factors as concentration of reagents, rate and suspension is formed by the addition of Table 10-3 lists sonie of the species that
procedures are highly empirical. order of mixing, length of standing, tempera- BaCl2• have been determined. by turbidimetric or
Effect of Concentration on Scattering. The ture, pH, and ionic strength are important ex- nephelometric methods; Perhaps most widely
attenuation of a parallel beam of radiation by perimental variables. Care must be exercised employed is the method for sulfate ion.
Applications of
scattering in a dilute suspension is given by to reproduce all conditions likely to affect Nephelometric methods permit the determi-
Scattering Methoda
the relationship particle size during calibration and analysis. nation of concentrationS as low as a few parts
Effect of Wavelength OR ScatteriDg. It has Turbidimetric or nephelometric methods are per million with a precision of 1 to 5 %.
been shown experimentally that the turbidity widely used in the analysis of water, for the Turbidimetric methods are reported to give
coefficient varies with wavelength as given by determination of clarity, and for the control the same degree of reproducibility with more
where Po and P are the power of the beam the equation of treatment processes. In addition, the-con- concentrated solutions. -
before and after passing through the length b
of the turbid medium. The quantity t is called
the turbidity coefficient, or the turbidity; its
where s is a constant for a given system. The
value is often found to be linearly related to
quantity t is dependent on particle size and
the concentration c of the scattering particles.
has a value of 4 when the scattering particles
As a consequence, a relationship analogous to
Beer's law results; that is, are significantly smaller than the wavelength TABLE 10-3 SOME TURBIDIMETRIC AND NEPHELOMETRIC METHODS·
of radiation (Rayleigh scattering); for par-
ticles with dimensions similar to the wave- Element MetW· Suspensioas Reagent Interferences
Po
Ioglo P = kbc (10-9) length (the usual situation in a turbidimetric
Ag T,N AgCl NaCl
analysis) t is found to be about 2.
As T As KH2P02 Se, Te
where k = 2.3 tic. For purposes of analysis, ordinary white
Au T Au SnCl2 Ag, Hg, Pd, Pt,
Equation 10-9 is employed in turbidimetric light is employed. If the solution is colored, it
Ru, Se, Te
is necessary to select a portion of the spec-
analysis in exactly the same way as Beer's law Ca T CaC204 H2C204 Mg, Na, SO~ - (in high
is used in photometric analysis. The relation- trum in which absorption by the medium is
minimized. concentration)
ship between loglo PolP and c is established CI- T,N AgCl AgNO) Br-, 1-
with standard samples, the solvent being used K T K2NaCo(N02)6 Na)Co(N02)6 SO~-
as a reference to determine Po. The resulting Na T,N Nazn(U02h(OAc)9 Zn(OAch and Li
Instruments
calibration curve is then used to determine U02(OAch
the concentration of samples. Nephelometric and turbidimetric measure- SO~- T,N BaS04 BaCl2 Pb
For nephelometric measurements, the power ments can be readily made with the various Se T Se SnCl2 Te
of the beam scattered at right angles to the fluorometers and photometers discussed ear- Te T Te NaH2P02 Se, As
incident beam is normally plotted against con- lier. Rectangular cells are ordinarily employed.
centration; a linear relationship is frequently With the exception of those areas through • Dam talcen rrom L. Meites, Handbook of A1IQlyticalChemistry. New York: McGraw-Hili Book Company, Inc, 1963, p.
obtained. The procedure here is entirely analo- which the radiation is transmitted, the cell ~17S. Wilh permission.
gous to a fluorometric method. walls are given a dull black coating to eliminate • T - turbidimetric; N - nephelometric.
Turbidimetric measurements have also been
employed for determining the end point in 5. In which solvent would the fluorescence of naphthalene be expected
precipitation titrations. The apparatus can be to be greatest: 1-chloropropane, 1-bromopropane, or 1-iodopropane?
very simple, consisting of a light source and a Why?
photocell located on opposite sides of the titra- 6. WouJd the fluorese:ence of aniline (C6HsNH2) be greater at pH 3 or
tion vessel. The photocurrent is then plotted as at pH 10? Explain.
a function of volume of reagent. Ideally, the 7. (a) Assume that the rate constants for various electronic processes of
turbidity increases linearly with volume of a certain molecule are: k(fluorescence) = 2 x 108 5-1, k(internal
reagent until the end point is reached, where- conversion, SI ...•S0)= 5 x 1075-1, k(external conversion,
upon it remains essentially constant. 1 0 SI ...•So) = 5 x 107 5-1, k(dissociation) = 3 x 107 5-1, k(predis-
sociation) = 1 x 10S 5-1• Referring to Figure 10·1, calculate the
quantum yield for fluorescence.
'0 For an analysis of the factors affecting turbidity titra- (b) Assume that in addition to the processes in part (a~ the following
tion curves, see: F. J. Meehan and G. Chiu, AMI. Chem, rate constants also apply: k(intersystem crossing, SI ...• Td = 2 x
36, 536 (1964).
1085-1, k(internal conversion, T1 •.•• So) = 0.15-1, k(external
conversion, T1 •.•• So) •••0.2 5-1, k(phosphorese:enc:e) = 0.7 5-1•
Calculate the quantum yield for fluorescence and the quantum
yield for phosphoresCence. What fraction of molecules which
reach the T1 state phosphoresc:es?
8. The reduced form of nicotinamide adenine dinucleotide (NADH) is
an important and highly fluorescent coenzyme. It has an absorption
FIGURE 10-5 A simple turbidim· maximum at 340 nm and an emission maximum at 465 nm. Standard
eter. solutions of NADH gave the following fluorese:ence intensities:

Concn Relative Concn Relative


NADH,IJM Intensity NADH,IJM Intensity
I. Define the following terms: (a) fluorese:ence, (b) phosphoresc:ence, (c) 0.100 13.0 0.500 59.7
resonanc:e fluorescence, (d) singlet state, (e) triplet state, (f) vibrational 0.200 24.6 0.600 71.2
relaxation, (g) internal conversion, (h) external conversion, (i) inter- 0.300 37.9 0.700 83.5
system crossing, (j) predissociation, (Ie)dissociation, and (I) quantum 0.400 49.0 0.800 95.1
yield.
2. Explain the difference between a fluorescence emission spectrum and
a fluorese:ence excitation spectrum. Which most closely resembles an Construct a calibration curve and estimate the concentration of
absorption spectrum? NADH in an unknown with a relative fluorescence intensity of 42.3
units.
3. Why is nephelometry more sensitive than turbidimetry?
9. In the absence of self-absorption, the fluorescence intensity of a
,4. Which compound below is expected to have- a greater fluorese:ence
sample is proportional to concentration only at low concentrations.
quantum yield? Explain.
Use Equation 10-4 to calculate the relative fluorescence intensities of
2.5 x IO-s, 2.5 X 10-4, and 1.0 x 1O-3M solutions of a compound

0'C(X)" with £ = 4.0 X 103M -t em -t if b = 020 em. Suppose that a fluo-


rometer is calibrated with a 2.5 x 10- sM solution of the compound
and the response is assumed to be linear. What would be the percent-
,}yC02H age of error when a 2.5 x 1O-4M solution is analyzed?

lJ 10. When a fluorescing sample absorbs appreciably at the exciting and


emitting wavelengths (Au and A_, respectively) the observed emission
Fluorescein intensity does not follow Beer's law. Consider the cell below in which
the observed fluorescenc:c originates from the central region with
approximate dimensioDS 12 x ;2: .

Incoming radiation is diminished by sample absorbance of A..,. over


'1'
the distanc:e Outgoing radiation is diminished by absorbance of .l.-
over the distance 13 • 1be fluorescence inteDSity depends upon absorb-
ance of A..,. over the distance 12• It can be shown that the net ob-
served emission intensity, F, is given by:l1
F == K(lO-···ltC(l - lO-""IzC)lO-'''''')
where eu is the molar absorptivity at ).u, e_ is the molar absorptivity
at 1_, C is the sample concentration, and K is a constant which
incorporates various instrumental parameters and the fluorescence
quantum yield. Construct a graph showing the relative fluorescence
intensity of 2,3-butanedione (CH3coCocH3) at the following con-
centratioDS: 0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, O.os and O.l2M. Use
the parameters ).0 == 422 om, en == 19.2M- I an -I, ).••••:;= 464 nm,
e••••== 2.3M-I an-I, II == 0.400 em, '2 == 0.200 an, and '3 = 0.500 an.
Atomic spectroscopy is based upon the ab- number of discrete lines at wavelengths that
sorption, emission, or fluorescence of electro- are characteristic of the element. To the extent
magnetic radiation by atomic particles. Two that molecules or complex ions are absent,
regions of the spectrum yield atomic spectral band spectra are not observed because vibra-
data-the ultraviolet-visible and the X-ray. tional and rotational quantum states cannot
This and the following chapter will be con- exist; thus, the number of possible transitions
cerned with the former; a discussion of X-ray is relatively small.
spectroscopy will be found in Chapter IS. Table 11-1 outlines the various analytical
In order to obtain ultraviolet and visible methods based upon atomic spectroscopy.
atomic spectra, it is necessary to atomize the These procedures frequently offer the advan-
sample. In this process, the constituent mole- tages of high specificity, wide applicability,
cules are decomposed and converted to gas- excellent sensitivity, speed, and convenience;
eous elementary particles. The emission, ab- they are among the most selective of all

/fr_--~"
sorption, and fluorescent spectrum of an atom- analytical methods. Perhaps 70 elements can
ized element consists of a relatively limited be determined. The sensitivities are typically

TABLE ".,

Emission
CLASSIFICATION

Type of Spectroscopy
Arc
OF ATOMIC SPECTRAL

Method of Atomization
Sample heated in an
METHODS

Radiation Source

Sample
Souru

VI
./)~
K'
~ SOO \ Fluorescent

Sample
radiation

FIGUR~ 11 ~1. Three types of flame spectroscopy:


(a) atomic emiSSion, (b) atomic absorption, and (c) atomic
electric arc fluorescence.
Spark Sample excited in a Sample
high voltage spark
Argon plasma Sample heated in an Sample
argon plasma in the parts-per-million to parts-per-billion THEORY OF
Atomic or flame Sample solution Sample range. An atomic spectral analysis can fre- FLAME SPECTROSCOPY
emission aspirated into a flame quently be completed in a few minutes.
X-Ray emission or None required; sample Sample
This chapter is concerned principally with When an aqueous solution of inorganic salts
electron probe X-ray bombarded with
electrons ultraviolet and visible atomic emission, ab- is aspirated into the hot flame of a burner, a
sorption, and fluorescence spectroscopy that substantial fraction of the metallic constitu-
Absorption Atomic absorption Sample solution Hollow cathode employ flames or other relatively low-temper- ents are reduced to the elemental state; to a
(flamek aspirated into a flame tube
Atomic a rption Sample solution Hollow cathode ature sources.! Arc, spark, and argon plasma much lesser extent, monatomic ions are also
(nonflame) evaporated and ignited tube atomic spectroscopy are considered in Chapter formed. Thus, within the flame is produced
on a hot surface 12. a gaseous solution or plasma containing
X-Ray absorption None required X-Ray tube
a significant concentration of elementary par-
Fluorescence Atomic fluorescence Sample solution SamPleJtexcited I For a more detailed discussion or these topics. see: ticles. Flame spectroscopy employs the emis-
(flame) aspirated into a flame by r iation G. F. Kirkbrigh! and M. Sargent. Atomic Absorplion and sion, absorption, or fluorescence by these
from a pulsed Fluorescence S[H!ctroscopy. New York: Academic Press,
particles as a basis for analyses.
lamp) 1974; Flame Emission and Atomic Absorption S[H!C-
Atomic fluorescence Sample solution Same as above trOfftetry, eels. J. A. Dean and T. C. Rains. New York:
(nonflame) evaporated and ignited Marcel Dekker, vol. 1: Theory, 1969;vol. 2: COffIPOMnlS Types of .Flame Spectroscopy
on a hot surface and Tec/udqws, 1971; vol. 3: E'-nu and Matrices.
X-Ray fluorescence None required Sample (excited 1974; W. G. Schrenk, Allillytical Atomic Spectroscopy. Figure 11-1 illustrates the three types of flame
by X-ray New York: Plenum Press, 1975; and E. L. Grove, spectroscopy. The first, called atomic emission
radiation) Applied Atomic Absorplion S[H!ctroscopy. New York: or flame emission, is based upon the measure-
Plenum Press, 1978. ment of the radiant power of a characteristic
line emitted by the analyte when aspirated Note that the energy scale is linear in units of magnesium, with two outer electrons, excited moments lie in the same direction). The effect
into a hot flame..A monochromator or filter electron volts (eV1 with the 3s orbital being singlet and triplet states with different ener- of the orbital magnetic moment on the mag-
is employed to isolate the line from the radia- assigned a value of zero. The scale extends to gies exist (Figure 11-3). In the singlet state, netic field of the combined spins produces a
tion produced by the flame and the other about 51 eV, the energy necessary to remove the spins of the two electrons arc opposed (or splitting of the p level into a triplet. This
constituents of the sample. the single 3s electron from the influence of the paired); in the triplet state the spins are paral- behavior is characteristic of all of the alkaline-
Figure 11-1b is a schematic diagram show- central atom, thus producing a sodium ion. lel (p. 281). The p, d, and f orbitals of the earth atoms, singly charged aluminum and
ing the components of an atomic absorption The energies of several atomic orbitals are triplet state are split into three levels that beryllium ions, and so forth.
or flame absorption apparatus. Here, the flame indicated on the diagram by horizontal lines. differ slightly in energy. These phenomena As the number of electrons outside the
serves the same purpose as the cell or cuvette Note that the p orbitals are split into two can be also rationalized by taking into account closed shell increases, the energy level dia-
in an ordinary photometer or spectrophotom- levels that differ but slightly in energy. This the interaction between the fields associated grams become more and more complex. Thus,
eter; that is, the flame can be considered to difference is rationalized by assuming that an with the spins of the two outer electrons and with three outer electrons, a splitting of energy
be a dilute gaseous solution of the atomized electron spins about its own axis and that the the net field arising from the orbital motions levels into two and four states occurs; with
sample held in place by the aspirator-burner. direction of this motion may either be the of all the electrons. In the singlet state, the four outer electrons, singlet, triplet, and quintet
Radiation from a suitable source is passed same as or opposed to its orbital motion. two spins are paired, and their respective states exist.
through the atomized sample and into the slit Both the spin and the orbital motions create magnetic effects cancel; thus, no energy split- Although correlation of atomic spectra with
of a photometer or spectrophotometer. In ma,gnetic fields owing to the rotation of ting is observed. In the triplet state, however, energy level diagrams for elements such as
order to discriminate between radiation from the charge carried by the electron. The two the two spins are unpaired (that is, their spin sodium and magnesium is relatively straight-
the source and that which is emitted by the fields interact in an attractive sense if these
flame, it is common practice to chop the two motions are in the opposite direction; a
beam from the source before it reaches the repulsive force is generated when the motions
ar~ parallel. As a consequence, the energy of

.)
flame. The detector'circuitry is then designed 's,;; 'P,
;;
2p3
;;
'0H3• 's,;; 2p, 2P3 '0,;; 'D.;;
to reject the dc output from flame emission the electron whose spin opposes its orbital ;; ;;

and measure the ac absorption signal from mdtion is slightly smaller than one in which
5.0 10.0 4p 4p
the source and sample. the motions are alike. Similar differences exist
Atomic fluorescence or flame fluorescence is in the d and f orbitals, but their magnitudes
illustrated in Figure 11-1c. Here again, the are ordinarily so slight as to be undetectable; 4.
flame serves to form atomic particles and thus, only a single energy level is indicated for
4.0 8.0
hold them in the light path for excitation by a d orbitals in Figure 11-2a.
The splitting of higher energy p, d. and f
~
suitable source. The resulting fluorescence is
then directed to a photometer or spectro- orbitals into two states is characteristic of all '"
photometer. species containing a single external electron. :l
g 3.0 6.0
Thus, the energy level diagram for the singly c:

Flame Spectra
charged magnesium ion, shown in Figure
11-2b, has much the same general appearance
i
0;

The emission, absorption, and fluorescence as that for the uncharged sodium atom. So ~ 2.0 4.0
spectra of gaseous, atomic particles consist of also do the diagrams for the dipositive alu- "
c:
w
well-defined narrow lines arising from elec- minum ion and the remainder of the alkali-
tronic transitions of the outermost electrons. metal atoms. It is important to note, however,
For metals, the energies of many of these thltt the energy difference between the 3p and Mg+
1.0 2.0
transitions appear in the ultraviolet and vis- 3s states is approximately twice as great for
ible regions. th~ magnesium ion as for the sodium atom
Energy Level Diagrams. The energy level beCause of the large,: nuclear charge of the
diagrart1§ for the outer electrons of an element former.
provide a convenient basis for describing the A comparison of Figure 11-2b with Figure
processes upon which the various types of 11-3 shows that the energy levels, and thus ~ ~
atomic spectroscopy are based. The diagram the spectrum, of an ion is significantly differ- FIGURE 11-2 Energy level diagrams for (a) atomic sodium and (b) mag-
for sodium shown in Figure 11-2a is typical. ent from that of its parent atom. For atomic nesium(I) ion.
Singlet

-;,i-
''"-~::O~~:~ simultaneous measurement of the two lines is
frequently used.
Atomic Absorption Spectra. In a flame
5p 4d 4'~ plasma, sodium atoms are capable of absorb-
6.0 ...
",fi ing radiation of wavelengths characteristic of
3d electronic transitions from the 3s state to
II
higher excited states. For example, sharp
~ 47~; - r- 6.' p'l1"
absorption peaks at 5890, 5896, 330~ and
..i 4.0
3303 A are observed experimentally; referring
again to Figure 11-2a, it is apparent that each
, 2 0 ...•.•..••.•....•..
6. 6p P'12
---L:'--_ Radiationless
deactivation

f
w
adjacent pair of these peaks corresponds to
transitions from the 3s level to the 3p aftd the
FIGURE 11-4 Energyleveldi-
agram for thallium showing the
4p levels, respectively. It should be mentioned source of two fluorescence lines.
2.0
that absorption due to the 3p to 5s transition
is so weak as to go undetected because the
number of sodium atoms in the 3p stat, is so
very small in a flame. Thus, typicaliy, an in two steps, including a fluorescent step pro"
atomic absorption spectrum from a flame ducing a line at 5350 A; radiationless deactiva-
consists predominately of resonance lines, tion to the ground state quickly follows. Res-
FIGURE 11-3 Energy level diagram for atomic magnesium. which are the result of transitions from the onance fluorescence at 3776 A is also ob-
ground state to upper levels. . served.
fo~ard and amenable to theoretical interpre- Atomic Fluorescence Spectra. Atom~ in a lAne Widths. Atomic emission and absorp-
tation, the same cannot be said for the heavier tra of the transition metals are considerably flame can be made to fluoresce by irradiation tion peaks are much narrower than the bands
elements, and particularly the transition met- more ~plex that the Spectra of species with with an intense source containing \wave- . resulting from emission or absorption by
als. These species have larger numbers of low atomic numbers. lengths that are absorbed by the element. As molecules. The natural width of atomic lines
closely spaced energy levels; as a consequence, Atomic Emission Spectra. At room tem- in molecular fluorescence, the fluorescent can be shown to be about 10-4 A. Two
the number of absorption or emission lines perature, essentially all of the atoms of a sample spectra is most conveniently observed at an elTects, however, combine to broaden the lines
c:an be enortnous. For example, Harvey2 has of ~tter are in the ground state. For example, angle of 90 deg to the light path. The ob- to a range from o.oi
to 0.05 A.
listed the Dumber of Jines observed in the arc: the s~e outer electron of metallic sodium served radiation is frequently the result of Doppler broadening arises from the rapid
aDd spark spectra of neutral and singly ionized occupteS the 3.t orbital under these circum- resonance fluorescence. For example, when motion of the atomic particles in the flame
atoms for.a variety of elements. For the alkali s~. Excitation of this electron to higher magnesium atoms are exposed to an ultra- plasma. Atoms that are moving toward the
metals, thIS number ranges from 30 for lithium orbitals can be brought about by the heat of violet source, radiation of 2852 A is absorbed monochromator emit lower wavelengths ow-
to 645 .for cesium; for the alkaline earths, ~ flame or an electric are or spark. The life- as electrons are promoted from the 3s to the ing to the well-known Doppler shift; the effect
magnesl~ has 173, calcium 662, and barium ttme. of the excited atom is brief, however, 3p level (see Figure 11-3); the resonance flu- is reversed for atoms moving away from the
472. Typical of the transition series, on the and. Its return to the ground state is aCCOm- orescence emitted at this same wavelength monochromator.
~her hand, are chromium, iron, and cerium ~D1ed by the emission of a quantum of radia- may then be used for analysis. In contrast, Doppler broadening is also observed for
With 22?7, 4757, and 5755 lines, respectively. tIOn. The vertical lines in Figure ll-la indicate when sodium atoms absorb radiation of wave- absorption lines. Those atoms traveling toward
F~er Imes are excited in flames because of some of the common electronic transitions length 3302 A, electrons are promoted (0 the the source absorb radiation of slightly shorter
their lower temperatures; still, the flame spec- that foUow excitation of sodium atoms. the 4p state (see Figure ll-la). A radiationless wavelength than that absorbed by particles
~a~elength of the resulting radiation is' also transition to the two 3p states takes iplace traveling perpendicular to the incident beam.
ilxhcated. The two lines at 5890 and 5896 A more rapidly than resonance fluorescenee. As The reverse is true of atoms moving away
are t~~ most intense under usual excitation a consequence, the observed fluorescence oc- from the source.
2 C. E. ~l1'Vey, sP«troc'-koJ Proudura. GIeDdaIe, COnditIOns and are, therefore, the ones most curs at 5890 and 5896 A. Figure 11-4 illustrates Pressure broadening also occurs; here colli-
CA: Applied Research Laboratorica, 1950, Chapter 4. CO~o~y employ~ for analytical purposes. yet a third type of mechanism for atomic sions among atoms cause small changes in
A slit width that IS wide eDough to pertnit fluorescence. Here, some of the thallium atoms the ground-state energy levels and a conse-
excited in a flame return to the ground state quent broadening of peaks.
the reactants are present in stoichiometric served for an element. For example, when a
proportions. sample containing magnesium is aspirated into
The temperatures provided by the burning a low-temperature flame (2000 to 2500·K~ the
of natural or manufactured gas in air are so element is converted almost completely to the
low that only the alkali and alkaline-earth atomic state; the spectrum, therefore, consists
metals, with very low excitation energies, pro- of the lines shown in Figure 11-3. On the
duce useful spectra. Acetylene/air mixtures other hand, more than 50 % of the element
give a somewhat higher temperature. Oxygen may be present as Mg+ in a cyanogen flame;
or nitrous oxide must be employed as the oxi- under these circumstances, the spectrum will
dant in order to excite the spectra of many contain wavelengths indicated in Figure il-2b
metals; with the common fuels, temperatures as well as those in Figure 11-3.
of 2500 to 3100°C are obtained. The hottest Effect of Temperature on Emission, Ab-
practical flame results from the combustion of sorption, and Fluorescence. The flame tem-
cyanogen in oxygen: perature also determines the fraction of
species that exists in excited states, and thus
CzNz + Oz ..• 200 + Nz influences emission intensities. The Boltz-
6~1 ~
6636 Temperatwe Profiles. Figure 11-6 is a mann equation permits the calculation of this
WMength,A· temperature profile of a typical flame for fraction. If NJ and No are the number of
FIGURE 11-5 Molecular flame and flame absorption spectra for atomic spectroscopy. The maximum tempera- atoms in an excited state and the ground state
CaOH and Ba. (Adapted from L. Capacho-Delgado and S. Sprague, ture is located somewhat above the inner respectively, their ratio is given by
A.tomic A.bsorption Newsletter, 4, 363 (1965). Courtesy of Perkin-Elmer cone. Clearly, it is important-particularly for
Corporation, Norwalk, Connecticut.) emission methods-to focus the same part of NJ = PJ exp( _ EJ) (H-I)
the flame on the entrance slit for all cali- No Po kT
brations and analytical measurements.
Effect of Temperature on the Types of where k is the Boltzmann constant (1.38 x
Spectra. The temperature of the flame plasma 10-16 erg/deg~ T is the temperature in degrees
Molecular Spectra in Flames. Hydrogen or both emission and absorption spectroscopy, determines the nature of the emission, ab- Kelvin, and EJ is the energy difference in ergs
hydrocarbon fuels produce absorption and the presence of such bands represents a sorption, or fluorescence spectrum that is ob- between the excited state and the ground state.
emission bands over certain wavelength ranges, potential source of interference, wllli:h must
owing to the presence of such species as OH be dealt with by proper choice of wavelength,
and CN radicals and Cz molecules. Some by background correction, or by a change in
aIkaline-earth and rare-earth metals also form combustion conditions.
volatile oxides or hydroxides that absorb and TABLE 11-2 MAXIMUM FLAME TEMPERATURE
emit over broad spectral ranges. One example FOR VARIOUS FUELS AND OXIDANTS·
is shown in Figure 11-5 where the emission
and absorption spectra of CaOH are shown. Oxidant Measured Temperatures, °C
The dotted line in the figure shows the wave- The greatest source of uncertainty in flame
spectroscopic methods arises from variations Natural gas Air 1700-1900
length of the barium resonance line. The Natural gas Oxygen 2740
potential interference of calcium in the atomic in the behavior of tne flame. Thus, it is impor-
Hydrogen Air 2000-2050
absorption determination of barium can be tant to understand the important character- Hydrogen Oxygen 2550-2700
avoided by employing a hotter flame which istics of flames and the variables that affect Acetylene Air 2125-2400
these characteristics. Acetylene Oxygen 3060-3135
dCC9mposes the CaOH molecule, thus causing
Acetylene Nitrous oxide 2600-2800
the molecular absorption band shown in the Oxygen 4500
Flame Tempenature Cyanogen
figure to disappear.
A band spectrum is useful for the deter- Table 11-2 lists the common fuels and oxi-
• Data from R. N. Kniselcy in Flome Emission and Aromic Absorption Spec/rascopy,
mination of perhaps one-third of the elements dants employed in flame spectroscopy and eels. J. A. Dean and T. C. Rains, London: Marcel Dekker, 1969, voL I, p. 191. With
that are amenable to emission analysis. For the maximum observed temperatures when permission.
considered, only about 0.Ql7 % of the sodium
The quantities PJ and Po are statistical factors
5.0
atoms were thermaIly excited at the tempera-
that are determined by the number of states
ture of a hydrogen/oxygen flame. An emission
having equal energy at each quantum level.
method is based upon this small fraction of
As shown by the following example, the frac-
the analyte. In contrast, absorption and fluo-
tion of excited atoms in a typical gas flame 4.0
rescence measurements use the 99.8 % of the
(T = 2SOO"K)is very small.
E
analyte present as unexcited sodium atoms to
u
EXAMPLE ,; produce the analytical signals. Note also that
Calculate the ratio of sodium atoms in the ~ 3.0 while a looK temperature change causes a
"g
3p excited states to the number in the ground 4 % increase in sodium ions, the correspond- FIGURE 11-7 Flame absorption profiles
.8~
state at 2500 and 251OoK. .••. 2.0
ing change in percent of sodium atoms is
inconsequential.
for three elements. (Reprinted from J. W.
Robinson, Atomic Absorption Spectroscopy,
In order tQ calculate EJ in Equation H-l,
we employ an average wavelength of 5893 A ~ Temperature fluctuations actually do exert 2d ed. New York: p. 69. Courtesy of Marcel
~ an indirect influence on flame absorption and Dekker, 1975, p. 56.)
for the two sodium emission lines involving
the 3p ...• 3s transitions. To obtain the energy 1.0 1830 fluorescence measurements in several ways.
1800
in ergs, we employ the conversion factors An increase in temperature usually increases
1700 Flame Profil ••
found inside the front cover. the efficiency of the atomization process and
hence the total number of atoms in the flame. Important regions of the flame, from bottom
1 In addition, line broadening and a con- to top, include the base, the inner cone, the
Wavenumber = 5893 A x 10-8 em/A
sequent decrease in peak height occurs be- reaction zone, and the outer mantle (see Figure
. = 1.697 X 104 em-I cause the atomic particles travel at greater 11-7). The sample enters the base of the flame
FIGURE 11-8 Temperature profiles (in 0c) rates which enhance the Doppler effect. In- in the form of minute droplets. Within this
EJ = 1.697 X 104 em -I x 1.99 for a natural gas-air flame. (From B. Lewis creased concentrations of gaseous atoms at region, water evaporates from a substantial
i and G. van Elbe, J. Chern. Phys .• II, 75 (1943).
. x 1O~16erg/em-I higher temperatures also cause pressure fraction of the droplets; some of the sample
With permission.) broadening of the absorption lines. Because thus enters the inner cone as solid particles.
= 3.38 x 10-12 erg of these indirect effects, a reasonable control Here, vaporization and decomposition to the
There are two quantum states in the 35 level of the flame temperature is also required for atomic state occurs; it is here, also, that
and six in the 3p. Thus, quantitative absorption and fluorescence mea- the excitation and absorption processes com-
surements. mence. Upon entering the reaction zone, the
Replacing 2500 with 2510 in the foregoing
PJ =~ = 3 equations yields .
The larac: ratio of unexcited to excited atoms are converted to oxides; these then pass
Po 2 atoms in flames leads to another interesting into the outer mantle and are subsequently
Substituting into Equation 11-1 yields comparison of the three atomic flame methods. ejected from the flame. This sequence is not'
NJ = 1.74 X 10-4
No Because atomic absorption and fluorescence necessarily suffered by every drop that is
methods are based upon a much larger pop- aspirated into the flame; indeed, depending
N ( 3.38 X 10-12 )
~ = 3 exp -1.38 x 10-16 x 2500 The foregoing example demonstrates that ulation of particles, both procedures might upon the droplet size and the flow rate, a
a temperature fluctuation of only lOOKresults be expected to be more sensitive than the significant portion of the sample may pass
in a 4 % increase in the number of excited emission procedure. This apparent advant- essentially unaltered through the flame.
sodium atoms. A corresponding increase in age is offset in the absorption method, how- A flame profile provides useful information
N· emitted power by the two lines would result. ever, by the fact that an absorbance measure- about the processes that go on in different
loglo 3':0 Thus, an analytical method based on the ment involves evaluation of a difference parts of a flame; it is a plot that reveals re-
i measurement of emission requires close con- (log Po -log P); when the two numbers are gions of the flame that have similar values for
3.38 x 10-12 trol of flame temperature. nearly alike, larger errors must be expected in a parameter of interest. Some of these pa-
2.303 x 1.38 x 10-16 x 2500 = -4.254 Absorption and fluorescence methods are the difference. As a consequence, atomic emis- rameters include temperature (Figure H-6~
theoretically less dependent upon tempera- sion and atomic absorption procedures tend chemical composition, absorbance, and ra-
ture because both measurements are based to be complementary in sensitivity, the one diant or fluorescent intensity.
N being advantageous for one group of elements Emission Pro6les. Figure 11-8 is a three-
upon initially unexcited atoms rather than
N~ = 1.67 X 10-4 and the other for a different group. dimensional profile showing the emission in-
thermally excited ones. In the example just
responsible for the emission dissociate at the and fluorescence measurements. Nonftame
. higher temperatures of the latter region. atomizers have been confined to the latter
The more sophisticated instruments for two procedures.
flame emission spectroscopy are equipped
with monochromators that sample the radia- Flame Atomizer.
tion from a relatiVl:ly small part of the flame; for Atomic Spectroscopy
adjustment of the position of the flame with
The most common atomization device for
respect to the entrance slit is thus critical.
· atomic spectroscopy consists of a nebulizer
Filter photometers, on the other hand, scan a
and a bumer.3 The nebulizer produces a fine
much larger portion of the flame; here, con-
trol of flame position is less important. · spray or aerosol from the liquid sample,
Flame Absorbance Profiles. Figure 11-7 : which is then fed into the flame. Both total
shows typical absorbance profiles for three
consumption (turbulent flow) and premixed
elements. Magnesium exhibits a maximum in
: (laminar flow) burners are encountered.
· Turbulent Flow Burner. Figure 11-9 is a
absorbance at about the middle of the flame FIGURE 11-9 A turbulent flow bum-
FIGURE 11-8 Flame profile for calcium · schematic diagram of a commercially avail- er. (Courtesy of Beckman Instrument,
because of two opposing effects. The initial
line in a cyanogen-oxygen flame for difJ'erent , able turbulent flow burner. Here, the nebu- Ioe .• Fullerton, California.)
increase in absorbance as the distance from
sample flow rates. (Reprinted with permission lizer and burner are combined into a single
the base becomes larger results from an in-
from K. Fuwa, R. E. Thiers, B. L. Vallee, and unit. The sample is drawn up the capillary
creased number of atomic magnesium par-
M. R. Baker, A.nal. Chem .• 31, 2039 (1959). and nebulized by venturi action caused by the
ticles produced with longer exposure to the
Copyright by the American Chemical Society.) flow of gases around the capillary tip. Typical
heat of the flame. In the" reaction zone, of the mixing chamber where it is drained to
however, appreciable oxidation of the mag- sample flow rates are 1 to 3 ml/min. a waste container. The aeroso~ oxidant, and
Turbulent flow burners offer the advantage fuel are then burned in a slotted burner that
nesium starts to occur. This process leads to
an eventual decrease in absorbance because of letting.a relatively large and representative provides a flame which is usually 5 or 10 em
the oxide particles formed are nonabsorbing sample reach the flame. Furthermore, no pos- in length.
tensity for a calcium line produced in a cyano- at the wavelength used. To obtain maximum sibility of flashback and explosion exists. The Laminar flow burners provide a relatively
gen ftame. Note that the emission maximum is analytical sensitivity, then, the flame must be disadvantages of such burners include a rela- quiet flame and a significantly longer path
found just above the inner cone. adjusted with respect to the beam until a tively short flame path length and problems length. These properties tend to enhance sen-
Fiaure 11-8 also demonstrates that the maximum absorbance reading is obtained. with clogging of the tip. In addition, turbulent sitivityand reproducibility. Furthermore, clog-
inteasity of emission is critically dependent The behavior of silver, which is not readily flow burners are noisy both from the elec- ging is seldom a problem. Disadvantages in-
upon the rate at which the sample is in- oxidized, is quite different; here, a continuous tronic and auditory standpoint. Although clude a lower rate of sample introduction
troduced into the flame. Initially, the line increase in atoms is observed from the base to sometimes used for emission and fluorescence (which may offset the longer path length
intensity rises rapidly with increasing flow the periphery of the flame. In contrast, chro- analyses, turbulent flow burners find little use advantage) and the possibility of selective
rate as a consequence of the increasing number mium, which forms very stable oxides, shows in present-day absorption work. evaporation of mixed solvents in the mixing
of calcium particles. A rather sharp maximum a continuous decrease in absorbance begin- Laminar Flow Burner. Figure 11-10 is a chamber, which may lead to analytical uncer-
in intensity occurs, beyond which the water, ning at the base; this observation suggests diagram of a typical commercial laminar flow tainties. Furthermore, the mixing chamber
which is aspirated with the sample, lowers the that oxide formation predominates from the burner. The sample is nebulized by the flow contains a potentially explosive mixture
flame temperature. start. Clearly, a different portion of the flame of oxidant past a capillary tip. The resulting which can be ignited by a flashback. Note
Where molecular band spectra form the should be used for the analysis of each of aerosol is then mixed with fuel and flows past that the burner in Figure 11-10 is equipped
basis for emission analysis, the peaks often these elements. a series of baffies that remove all but the with pressure relief vents for this reason.
appear lower in the flame than line maxima. finest droplets. As a result of the baffies, the Fuel and Oxidant Regulators. An impor-
For example, calcium produces a useful band majority of the SIpllple collects in the bottom tant variable that requires close control in
ATOMIZERS FOR
in the region of 540 to S60 DID, probably due ATOMIC SPECTROSCOPY flame spectroscopy is the flow rate of both
to the presence of CaOH in the flame (see oxidant and fuel. It is desirable to be able to
Figure 11-5). The maximum intensity of this Two types of atomizers are encountered in vary each over a considerable range so that
3For a detailed discussion of ftame atomizeR, see: R. D.
emission band occurs in the inner cone and atomic spectroscopy. Flame atomizers are ideal atomization conditions can be found ex-
Dresser, R. A. Mooney, E. M. Heithmar, and F. W.
decreases rapidly in the mantle as the molecules employed for atomic emission, absorption, Pankey, J. Chern. Edue, 51, A403 (197S~ perimentally. Fuel and oxidant are ordinarily
combined in approximately stoichiometric the smaller end down. A light-weight conical the drain (laminar burner) or is not C9mpletely The relative precision of nonflame methods
amounts. For the analysis of metals that form or spherical float is lifted by the gas flow; its atomized (turbulent burner). Second, the resi- is generally in the range of 5 to 10 % compared
stable oxides, however, a reducing flame, vertical position is determined by the flow rate. dence time of !ndividual atoms in the optical with the 1 to 2 % that can be expected for
which contains an excess of fue~ may prove path in the flame is brief ( -10 - 4 s). flame atomization.
more desirable. Flow rates are ordinarily con- Since 1970, nonflame atomizers have ap-
Nonflame Atomize,.
trolled by means of double-diaphragm pres- peared on the market. 4 These devices gen-
sure regulators followed by needle valves in In terms of reproducible behavior, flame atom- erally provide enhanced sensitivity because ATOMIC
the instrument housing. ization appears to be superior to all other the entire sample is atomized in a short ABSORPTION
A requirement for reproducible analytical methods that have been thus far developed. In period and the average residence time of the SPECTROSCOPYlI
conditions is that both the fuel and oxidant terms of sampling efficiency (and thus sensi- atoms in the optical path is a second or more.
systems be equipped with some type of flow- tivity~ however, other atomization methods In a nonflame atomizer, a few microliters Atomic absorption spectroscopy finds con-
meter. The most widely used of these is the are markedly better. Two reasons for the lower of sample are evaporated and ashed at low siderably wider use than either of the other
rotameter, which consists of a tapered trans- sampling efficiency of the flame can be cited. temperatures on an electrically heated surface two atomic spectroscopic methods because it
parent tube that is mounted vertically with First, a large portion of the sample flows down of carbon, tantalum, or other conducting is best suited to routine analyses in the hands
J material. The conductor can be a hollow of relatively unskilled operators.
tube, a strip or rod, a boat, or a trough. After
ashing, the current is increased to 100 A or Sources for
more, which causes the temperature to soar Atomic Absorption Methods
to perhaps 2000 to 3OOO"C;atomization of
Analytical methods based on atomic absorp-
the sample occurs in a period of a few sec-
tion are potentially highly specific because
onds. The absorption or fluorescence of the
atomic absorption lines are remarkably narrow
atomized particles can then be measured in
and because electronic transition energies are
the region immediately above the heated
unique for each element. On the other hand,
conductor. At a wavelength where absorption
the limited line widths create a measurement
or fluorescence takes place, the signal is ob-
problem not encountered in molecular ab-
served to rise to a maximum after a few sec-
sorption. Recall that Beer's law applies only
onds and then decay to zero, corresponding
for monochromatic radiation; a linear relation-
to the atomization and subsequent escape of
ship between absorbance and concentration,
the volatilized sample; analyses are based
however, can be expected only if the bandwidth
upon peak height or area. Atomization is or-
of the source is narrow with respect to the
dinarily performed in an inert gas atmosphere
width of the absorption peak (p. 155). No
to prevent oxidation of the conductor.
ordinary monochromator is capable of yielding
Nonflame atomizers olTer the advantage of
a band' of radiation as narrow as the peak
unusually high sensitivity for small volumes
width of an atomic absorption line (0.002 to
of sample. Typically, sample volumes between
0.005 nm). Thus, when a continuous source is
0.5 and 10 pi are employed; under these cir-
employed with a monochromator for atomic
cumstances, absolute limits lie in the range of
absorption, only a minute fraction of the radi-
10-10 to 10-13 g of analyte. These sensitivi-
ation is of a wavelength that is absorbed,
ties are up to 1000 times greater than those
obtained with flames.

• Reference books on atomic absorption spectroscopy


includl: G. F. Kirkbright and M. Sargent, Atomic
4 See: R. D. Dresser, R. A. Mooney, E. M. Heithmar, and Absorption aIIII Fluorescence Spectroscopy. New York:
F. W. Plankey, J. Chem. £duc, 52, A451, As03 (1975); R. Academic Press, 1974; J. W. Robinson, Atomic Absor~
E. Sturgeon, Arral. Chem, 49, 1255A (1977); and C. w. lion S~rrroscopy, 2d ed. New York: Marcel Dekker,
FIGURE 11-10 A laminar flow burner. (Courtesy of Perkin-Elmer Corporation, Fuller, Electrothermal Atomization for Atomic Absorption 1975; and W. Slavin. Alomic Absorption S~crroseopy, 2d
Norwalk, Connecticut.) Spectroseopy. London: The Chemical Society. 1978. ed. New York: Interscience, 1978.
and the relative change in intensity of the and produce II.I1 atomic cloud; this process GlISeOlB l>isc:hlrge Lamps. Gas discharge
emergent band is small in comparison to the is called sputtering. A portion of the sputtered lamps produce a line spectrum as a con-
change suffered by radiation that actually cor- metal atoms are in excited states and thus sequence of the passage of an electrical cur-
responds to the absorption peale. Under these Emiaion emit their characteristic radiation in the usual rent through a vapor of metal atoms; the
conditions, Beer's law is not followed; in ad- spectrum
of way. Eventually, the metal atoms'!iift'use back familiar sodium and mercury vapor lamps are
dition, the sensitivity of the method is lessened source to the cathode surface or the glass walls of examples. Sources of this kind are partic-
significantly. the tube and are redeposited. ularly useful for producing spectra of the
This problem has been overcome by em- The cylindrical configuration of the cath- alkali metals.
ploying a source of radiation that emits a line ode tends to concentrate tlie radiation in a Source ModuiatiolL In the typical atomic
of the same wavelength as the one to be used limited region of the tube; this design also absorption instrument, it is necessary to elim-
for the absorption analysis. For example, if enhances the probability that redeposition inate interferences caused by;emission of radi-
the S89.6-nm line of sodium is chosen for the will occur at the cathode rather than on the ation by the flame. Most of the emitted radi-
absorption analysis of that element, a sodium glass walls. ation can be removed by locating the mono-
vapor lamp is useful as a source. Gaseous The efficiency of the hollow cathode lamp chromator between the flame and the detector;
sodium atoms are excited by electrical dis- Absorption depends upon its geometry and the operating nevertheless, this arrangement does not remove
charge in such a lamp; the excited atoms then spectrum potential. High potentials, and thus high cur- the flame radiation correspoBding to the wave-
emit characteristic radiation as they return to of
samp'e rents, lead to greater intensities. This advan- length selected for the analysis. The flame
lower energy levels. An emitted line will thus tage is 01Tset somewhat by an increase in will contain such radiation, since excitation
have the same wavelength as the resonance Doppler broadening of the emission lines. and radiant emission by some atoms of the
absorption line. With a properly designed Furthermore, the greater currents result in an analyte can occur. This difficulty is overcome
source (one that operates at a lower tempera- ,
I
increase in the number of unexcited atoms in by modulating the output ofihe source so that
ture than the flame to minimize Doppler I
the cloud; the unexcited atoms, in turn, are its intensity fluctuates at a atnstant frequency.
I I
broadening1 the emission lines will have band- I I capable of absorbing the radiation emitted by The detector then receives two types of signal,
widths significantly narrower than the absorp- m • monochromator
bandwidth the excited ones. This se!{-absorption leads to an alternating one from the source' and a
tion bandwidths. It is only necessary, then, for II
lowered intensities, particularly at the center continuous one from the flame. These signals
the monochromator to have the capability of i
I
Emission spectrum
efter passage
of the emission band. are converted to the corresponding types of
isolating a suitable emission line for the ab- I through sample A variety of hollow cathode tubes are electrical response. A simple high-pass, RC
sorption measurement (see Figure 11-11). The I end monochrolMtor available commercially. The cathodes of some filter (p. 23) is then employed to remove the
radiation employed in the analysis is thus consist of a mixture of several metals; such unmodulated dc signal and pass the ac signal
sufficiently limited in bandwidth to permit Wovelength
lamps permit the analysis of more than a for amplification.
measurements at the absorption peak. Greater FIGURE 11-11 Absorption of a resonance single element. . A simple and entirely satisfactory way of
sensitivity and better adherence to Beer's law line by atoms. modulating radiation from the source is to
result. interpose a circular disk in the beam between
A separate lamp source is needed for each the source and the flame. Alternate quadrants
element (or sometimes, group of elements). of this disk are removed to permit passage of
To avoid this inconvenience, attempts have sealed in a glass tube that is filled with neon light. Rotation of the disk at constant speed
been made to employ a continuous source or argon at a pressure of 1 to 5 torr (see provides a beam that is. chopped to the
with a very high resolution monochromator Figure 11-12). The cathode is constructed of desired frequency. As all; alternative, the
or, alternatively, to produce a line source by the metal whose spectrum is desired or serves power supply for the source can be designed
introducing a compound of the element to be to support a layer of that metal. for intermittent or ac operafion.
determined into a high-temperature flame. Ionization of the gas occurs when a poten-
Neither of these alternatives is as satisfactory tial is applied across the electrodes, and a cur-
as individual lamps. rent of about 5 to 10 mA is generated as the Instruments for Atomic
Hollow Cathode Lamps. The most common ions migrate to the electrodes. If the potential Absorption Spectroscopy
Gila NeorAr
source for atomic absorption measurements is sufficiently large, the gaseous cations ac- shield at 1·5 torr
Instruments for atomic absorption work are
is the hollow cathode lamp, which consists of a quire enough kinetic energy to dislodge some FIGURE 11-12 Schematic cross section of offered by numerous manufacturers. The
tungsten anode and a cylindrical cathode of the metal atoms from the cathode surface a hollow cathode lamp. range of sophistication and cost (upward
PRINCIPLES OF INSTR~NTAL ANALYSIS

Czemey- in the several reference works listed in foot- into the flame.) Both the sensitivity and the
Turner
monochromator note 5 (p. 315). detection limiU vary widely with such variables
SelBitil'ity •••• Detection Limits. Two as flame temperature, spectral bandwi.dth,
terms are employed in characterizing atomic detector sensitivity, and type of signal pro-
absorption methods. The sensitivity is defined cessing. Small differences among quoted values
as the concentration of an element in /lg!ml of the two parameters are not significant; for
~--!;:---~ (or ppm) which produces a transmittance example, whereas an order of magnitude differ-
I I signal of 0.99 or a corresponding absorbance ence is certainly meaningful, a factor of 2 or 3
I I
t t signal of 0.0044. Modern atomic instruments is probably not.
I Flame, ~ have adequate precision to discriminate be- Detection limits ranging from about 3 x
tween absorbance signals that differ by less 10-4 ppm to 20 ppm are observed for the
than 0.0044. For this reason, the term detec- various metallic elements when flame atomi-
tion limit has been introduced, which is defined zation is employed in atomic absorption. Non-
as the concentration of the element that pro- flame atomization often enhances this limit
duces an analytical signal equal to twice the by a factor of 10 to 1000.
FIG U RE 11-13 A typical double-beam atomic absorption spectro- standard deviation of the background signal. Columns three and four of Table 11-3 list
photometer. (For flame atomimtion, the standard deviation detection limits for several common elements
of the background signal is obtained byobserv- by flame and nonflame atomimtion proce-
ing the signal variation when a blank is sprayed dures.
from a few thousand dollars) is substantial; as and passed into a Czerney-Turner grating
always, the potential user ~ust select the most monochromator; a photomultiplier tube serves
appropriate design for the intended use. as the transducer. The output from the latter
S •..••••• Spec:t~ometers. A typical is fed to a lock-in amplifier which is synchron- TABLE 11-3 DETECTION LIMITS FOR THE ANALYSIS OF
single-beam instrument for multielement anal- ized with the chopper drive. The ratio between SELECTED ELEMENTS BY ATOMIC ABSORPTION AND
yses consists of several hollow cathode sources, the reference and sample signal is then ampli- FLAME EMISSION SPECTROMETRY
a chopper, an atomizer, and a simple grating fied and fed to the readout which may be a
spectrophotometer with a photomultiplier Detection Limit, Ilgfml
mete~,digital device, or recorder. Alternatil'ely,
transducer. It is used in the same way as a the amplified signal from the reference beam Wavelength, Noa8ame Flame Flame
singlc-beam instrument for molecular absorp-
tion work. Thus, the dark current is nulled
may be attenuated to match the sample signal
by means of a potentiometer; the transmittance
Element
••• Absorptioa" AlJsorptionb EmisskMf

Aluminum 396.2 0.03 0.005 (NzO)


with a shutter in front of the transducer. The or absorbance is then read from the position 309.3 0.1 (N O~
100% T adjustment is then made while a of the slide wire contact. Calcium 422.7 0.0003 0.002 (air 0.005 gir)
blank is aspirated into the flame (or ignited in It should be noted that the reference beam Cadmium 326.1 0.0001 2 (Nz )
• nonflame atomizer). Finally, the transmit- in atomic absorption instruments does not 228.8 0.005 (air)
tance is obtained with the sample replacing pass through the flame and thus does not cor- Chromium 425.4 0.005 0.005 (NzO)
357.9 0.005 (air)
the blank. rect for loss of radiant power due to absorp- 0.05 (NzO)
Iron 372.0 0.003
Single-beam atomic absorption instruments tion or scattering by the flame itself. 248.3 0.005 lairl
have the same advantageS and disadvantages Lithium 670.8 0.005 0.005 air 0.‫סס‬OO3(N~O)
as their molecular absorption counterparts, Magnesium 285.2 0.00006 0.‫סס‬OO3(air) 0.005 (NzO
which were discussed on page 185. Potassium 766.5 0.0009 0.005lair~ 0.0005 lairl
Applications of Atomic Sodium 589.0 0.0001 0.002 air 0.0005 air
Double-..... SpectroPbotometers. Figure
Absorption Spectroscopy
11-13 is a schematic diagram of a typica1
double-beam instrument. The beam from the Atomic absorption spectroscopy is a sensitive • Data from J. W. Robinson and P. J. Slevin. Ammcan Laboratory, 4 (8~ 14 (1972~ With permissioD,
hollow cathode source is split by a mirrored International Sc:ientific Communications. Inc. Fairfield, cr C 1972; waveJensth not reported. Copy-
means for the determination of more than 60
riPt 1972 by International Scientific Communications. Inc.
chopper, one half passing through the flame elements. Details concerning the methods for • Taken from data compiled by E. E. Pickett and S. R. KoirtyohanD, Anal. Chem. 41 (14~ 28A (1969)
and the other half around it. The two beams sample preparation and quantitative deter- and reprinted with permission. Copyright by the American Chemical Society. Data are for an acetylene
are then recombined by a half-silvered mirror mination of individual elements are available flame with the oxidant shown in parentheses.
Accurecy A much more troublesome problem is en- ployed.6 A description of some of the most broad band absorption or scattering by the
Under usual conditions, the relative err~r as- countered when the source of absorption or commonly used techniques follows. flame components. A background correction
sociated with a flame absorption analysis is of scattering originates in the sample matrix; The Two-Liae Correctioa Method. The two- is thus achieved.
the order of 1 to 2%. With special precau- here, the power of the transmitted beam, P, is line correction procedure requires the presence The Zeeman Elect Correction MetbocL One
tions, this figure can be lowered to a few reduced by the nonanalyte components of the of a reference line from the source; this line commercial atomic absorption instrument
tenths of 1 %. sample matrix, but the incident beam power, should lie as close as possible to the anaIyte achieves a background correction by taking
Po is not; a positive error in absorbance and line but must not be absorbed by the analyte. advantage of the Zeeman effect to split the
Spectral Interferences thus concentration results. An example of a If these conditions are met, it is assumed that anaIyte line into two components, one of which
potential matrix interference due to absorp- any decrease in power of the reference line is displaced from the other by a small wave-
Interferences of two types are encountered in tion occurs in the determination of barium in from that observed during calibration arises length increment (about O.ol nm). The two
atomic absorption methods. Spectral inter- alkaline-earth mixtures. :As shown by the from absorption or scattering by the matrix components are also polarized at 90 deg to
ferences arise when the absorption of an dotted line in Figure 11-5, the wavelength of products of the sample; this decrease is then one another and can be monitored alterna-
interfering species either overlaps or lies so the barium line used for; atomic absorption used to correct the power of the analyte line. tively by inserting a rotating polarizer into
close to the analyte absorption that resolu- analysis appears in the center of an absorp- The reference line may be from an impur- the beam path. Here, the displaced compo-
tion by the monochromator becomes impos- tion band for CaOH; clearly, interference of ity in the hollow cathode lamp, a neon or nent is sufficiently separated from the absorp-
sible. Chemical interferences result from various calcium in a barium arlaIysis is to be ex- argon line from the gas contained in the tion peak to provide a means of correction
chemical processes occurring during atomiza- pected. In this particular situation, the elfeet lamp, or a nonresonant emission line of the for background absorption or scattering.
tion that alter the absorption characteristics is readily eliminated by substituting nitrous element being determined. . Zeeman splitting can be accomplished by
of the analyte. A brief discussion of spectral oxide for air as the oxidant for the acetylene; Unfortunately, a suitable reference line is exposing the atomizer device or the light
interferences follows; sources of chemical inter- the higher temperature decomposes the often not available. source to a strong magnetic field.7
ference are considered in the next section .. CaOH and eliminates the: absorption band. The ContiDuous-Soan Correction Metholl.
Because the emission lines of hollow cath- Spectral interference ctpe to scattering by A second method for background correction Chemical Interferences
ode sources are so very narrow, interference products of atomization, is most often en- is available with suitably equipped double-
due to overlap of atomic spectral lines is rare. countered when concentrated solutions con- Chemical interferences are more common than
beam instruments. Here, a hydrogen or deute-
For such an interference to occur, the separa- taining elements such as Ti, Zr, and W- spectral ones. Their effects can frequently be
rium lamp is employed as a source of continu-
tion between the two lines would have to be which form refractory oxides-are aspirated minimized by suitable choice of conditions.
ous radiation throughout the ultraviolet re-
less than perhaps 0.1 A. For example, a vana- into the flame. Metal oxide particles with Both theoretical and experimental evidence
gion. The configuration of the chopper is then
dium line at 3082.11 A interferes in an anal- diameters greater than the wavelength of light rearranged so that the reference beam shown
suggest that many of the processes occurring
ysis b8sed upon the aluminum absorption line appear to be formed; significant scattering of in the mantle of a flame are in approximate
in Figure 11-13 is dispensed with, andinstcad,
at 3082.lS A. The interference is readily the incident beam results. equilibrium. As a consequence, it becomes
radiation from the continuous source and the
avoided, however, by employing the aluminum Fortunately, spectral interferences by matrix possible to regard the burned gases ofthe flame
hollow cathode lamp are passed alternatively
line at 3092.7 A instead. products are not widely encountered and often as a solvent medium to which thermodynamic
, through the flame. The power of the beams
Spectral interferences also result from the can be avoided by variations in the analytical calculations can be applied. The processes of
from the two sources is then compared rather
presence of combustion products that exhibit parameters, such as temperature and fuel-to- principal interest include formation of com-
than the power of a sample and reference
broad band absorption or particulate products oxidant ratio. Alternatively, if the source of pounds onow volatility, dissociation reactions,
beam from the hollow cathode source. The
that scatter radiation. Both diminish the power interference is known, an excess of the inter- and ionizations.
slit width is kept sufficiently wide so that the
of the transmitted beam and lead to positive fering substance can be ad.ded to both sample Formatioa of Compounds of Low Volatility.
fraction of the continuous source that is ab-
analytical errors. Where the source of these and standards; provided the excess is large with Perhaps the most common type of inter-
sorbed by the atoms ofthe sample is negligible.
products is the fuel and oxidant mixture alone, respect to the concentration from the sample ference is by anions which form compounds
Thus, the attenuation of its power during
corrections are readily obtained from ab- matrix, the contribution i of the latter will passage through the flame reflects only the
sorbance measurements while a blank is aspi- become insignificant. The1added substance is
rated into the flame. Note that this correction sometimes ca1Ied a radiation buffer. , For a detailed discussion of the application of the
must be employed with a double-beam as well The foregoing technique is not effective Zeeman elfect to atomic absorption, see: T. Hadeilhi and
as a single-beam instrument because the refer- R. D. McLaupwn, SCWlta, 17'" <lO4(1974); T. Hadeilbi
ence beam of the former does not pass
with complex samples because the source of
the interference may not be known; a back- • For a critical discussion of the various methods for and R. D. McLaulh\in, A••••I.
Koizuma and K. Yasuda, A••••I.
C"-.C"-.
48, 1009 (1976); H.
48, 1178 (t976);
bacltsround correction, see: A. T. Zander, Amer. Lab. 8
through the flame (see Figure 11-13). ground correction must, therefore, be em- (11~ 11 (1976). and S. D. Brown, A••••I.C"-.'" (14~ 1269A(1977~
of low volatility with the analyte and thus Dillociatioa fAIaiIiIIria. In the hot, gaseous Chlorine atoms formed from the added HQ The equilibrium constant K for this reac-
reduce the rate at which it is atomized. Low environment of a flame, numerous dissoci- decrease the atomic sodium concentration tion may take the form
results are the consequence. An example is ation and association reactions lead to C9n- and thereby lower the line intensity.
the decrease in calcium absorbance that is ob- version of the metallic constituents to the Another example of this type of interfer-
served with increasing concentrations of sul- elemental state. It seems probable that: at ence involves the' enhancement of the absorp-
fate or phosphate. For a fixed calcium con- least some of these reactions are reversible tion by vanadium when aluminum or titanium
where the bracketed terms are activities, x is
centration, the absorbance is found to fan off and can be treated by the laws of thermody- is present. The interference is significantly more
the fraction of M that is ionized, and p is the
nearly linearly with increasing sulfate or phos- namics. Thus, in theory, it should be possible pronounced in fuel-rich flames than in lean.
partial pressure of the metal in the gaseous
phate concentration until the anion to calcium to formulate equilibria such as These effects are readily explained by assum-
solvent before ionization.
ratio is about 0.5; the absorbance then levels ing that the three metals interact with such
The effect of temperature on K is given by
off at about 30 to 50% of its original value species as 0 and OR, which are always present
theSaha eqUation; namely,
and becomes independent of anion concentra- in flames. If the oxygen-bearing species are
tion. given the general formula Ox, a series of equi- " -5041E; .5.
Examples of cation interference have also librium reactions can be postulated. Thus, 10gK= T +zlogT
been recognized. Thus, aluminum is found to
cause low results in the determination of VOx~V+Ox
- 6.49 + log g••••g.- (11-4)
magnesium, apparently as ·a result of the for- AlOx~Al+Ox g... .
mation of a heat-stable aluminum/magnesium In practice, not enough is known about the
compound (perhaps an oxide). TiOx~Ti+Ox where E; is the ionization potential for the
nature of the chemical reactions in a flame to
Interferences due to formation of species of metal in electron volts, T is the absolute tem-
permit a quantitative treatment such as that
low volatility can often be eliminated or mod- In fuel-rich combustion mixtures, the concen- perature of the medium, and 9 is a statistical
for an aqueous solution. Instead, reliance must
erated by use of higher-temperature flames. tration of Ox is sufficiently small that its con- weight for e!lch of the three species indicated
be placed on empirical observations.
Alternatively, releasing agents, which are ca- centration is lowered significantly when alu- by the subsi:ripts. For the alkali metals, the
Dissociation reactions involving metal ox-
tions that react preferentially with the inter- minum or titanium is present in the sample. final term has a value of zero while for
ides and hydroxides clearly play an important
ference and prevent its interaction with the This decrease causes the first equilibrium to the alkaline "earths it is equal to 0.6.
part in determining the nature of the emission
analyte, can be employed. For example, ad- shift to the right with an accompanying in- Table 11-4 shows the calculated degree of
or absorption spectra for an element. For
dition of an excess of strontium or lanthanum crease in metal concentration as well as ab- ionization for several common metals under
example, the alltaline-earth oxides are rela-
ion minimizes the interference of phosphate sorbance. In lean mixtures, on the other hand, conditions that approximate those used' in
tively stable, with dissociation energies in
in the determination of calcium. The same the concentration of Ox is apparently high flame emission spectroscopy. The tempera-
excess of S eV. Molecular bands arising from
two species have also been employed as releas- relative to the total concentration of metal tures correspond roughly to conditions that
the presence of metal oxides or hydroxides
ing agents for the determination of magnesium atoms. Thus, addition of aluminum or titanium exist in air/fuel, oxygen/acetylene, and oxy-
in the flame thus constitute a prominent
in the presence of aluminum. In both instances, scarcely changes the concentration of Ox. gen/cyanogen flames, respectively.
feature of their spectra. Except at very high
the strontium or lanthanum replaces the ana- Therefore, the position of the first equilibrium It is important to appreciate that treatment
temperatures, these bands are more intense
lyte in the compound formed with the interfer- is not disturbed significantly. of the ionization process as an equilibrium-
than the lines for the atoms or ions. In contrast,
ing species. Ionization in Flames. Ionization of atoms with free electrons as one of the products-im-
the oxides and hydroxides of the alkali metals
Protective agents prevent interferences by and molecules is small in combustion mix- mediately implies that the degree of ionization
are much more readily dissociated so that line
forming stable but volatile species with the tures that involve air as the oxidant, and gen- of a metal will be strongly influenced by the
intensities for these elements are high, even at
analyte. Three common reagents for this pur- erally can be neglected. At the higher tempera- presence of other ionizable metals in the flame.
relatively low temperatures.
pose are EDT A, 8-hydroxyquinoline, and tures of flames employing oxygen or nitrous Thus, if the medium contains not only species
It seems probable that dissociation equilib-
APDC (the ammonium salt of I-pyrrolidine- oxide, however, ionization becomes important, M but species B as well, and if B ionizes
ria involving anions other than oxygen may
carbodithioic acid). The presence of EDTA and a significant concentration of free electrons a~rding t4 the equation
also influence flame emission. For example,
has been shown to eliminate the interference exists as a consequence of the equilibrium
the line intensity for sodium is markedly de- B~B+ +e- .
of aluminum, silicon, phosphate, and sulfate creased by the presence of HQ. A likely ex- M·~M+ + e- (11-2)
in the determination of calcium. Similarly, 8- planation is the mass-action effect on the then the degree of ionization of M will be
hydroxyquinoline suppresses the interference equilibrium where M represents a neutral atom or decreased by the mass-action effect of the
of aluminum in the determination of calcium molecule and M+ is its ion. We will focus electrons formed from B. Determination of
and magnesium. upon equilibria in which M is a metal atom. the degree of ionization under these condi-
ATOMIC SPECTROSCOPY 325

O.B
tions requires a calculation involving the dis- can frequently be eliminated by addition of K 25000 "g/ml
K to 000 "g1ml
sociation constant for B and the mass-balance an ionization suppressor, which provides a
expression relatively high concentration of electrons to
0.6 K rooo"g1ml
the flame; suppression of ionization of the
analyte results. The effect of a suppressor is
demonstrated by the calibration curves for (N,O·acetylenel
The presence of atom-ion equilibria in strontium shown in Figure 11-14. Note the 0.4

flames has a number of important conse- marked steepening of these curves as stron-
quences in flame spectroscopy. For example, tium ionization is repressed by the increasing K o "g1ml
the intensity of atomic emission or absorption concentration of potassium ions and elec- 0.2 K tooo "g1ml (Air·acetylenel
lines for the alkali metals, particularly potas- trons. Note also the enhanced sensitivity that
sium, rubidium, and cesium, is affected by tem- results from the use of nitrous oxide instead
perature in a complex way. Increased tempera- of air as the oxidant;. the higher temperature
tures cause an increase in the population of achieved with nitrous oxide undoubtedly en-
excited atoms, according to the Boltzmann hances the rate of decomposition and volatili-
relationship (p. 3(9); counteracting this effect, zation of the strontium compounds in the FIGURE 11-14 Effect of potassium concentration on
however, is a decrease in concentration of plasma. the calibration curve for strontium. (Reprinted with per-
atoms as a result of ionization. Thus. under mission from J. A. Bowman and J. B. Willis, Anal. Chern.•
some circumstances a decrease in emission or 39, 1210 (1967). Copyright by the American Chemical
Analytical Technique.
absorption may be observed in hotter flames. Society.)
It is for this reason that lower excitation Both calibratiOn curves and the standard ad-
temperatures are usually specified for the anal- dition method are suitable for atomic absorp-
ysis of alkali metals. tion spectroscopy.
o The effects of shifts in ionization equilibria Calibration Curves. While in theory, ab- sorbance should be proportional to concentra- following relationships apply:
tion, deviations from linearity often occur. A, = kC,
Thus, empirical calibration curves must be pre-
pared. In addition, there are sufficient un- AT = k(C. + Co)
controllable variables in the production of an where C, is the analyte concentration in the
atomic vapor to warrant measuring the ab- diluted sample and C. is the contribution of
TABLE 11-4 DEGREE OF IONIZATION OF METALS AT FLAME sorbance of at least one standard solution each the added standard to the concentration; A,
TEMPERATURES- time an analysis is perfonned. Any deviation and AT are the two measured absorbances.
of the standard from the original calibration Combination of the two equations yields
Fraction Ionized at the Indicated Pressure and Temperature curve can then be employed to correct the
Ionization p = 10-4 atm p= 10-6 atm analytical results. C, = C. (AT ~ A,) (11-5)
Potential, Standard Addition Method. The standard
Element eV 2000 K
0
3SOOoK 5000 K
0
2000 K0
3500 K
0
5000 K0
addition method is widely used in atomic If several additions are made. AT can be
0.11 absorption spectroscopy. Here, two or more plotted against C•. The resulting straight line
Cs 3.893 0.01 0.86 >0.99 ;> 0.99 >0.99
Rb 4.176 0.004 0.74 >0.99 0.04 >0.99 >0.99 aliquots of the sample are transferred to vol- can be extrapolated to AT = O. Substituting
K 4.339 0.003 0.66 0.99 0.03 0.99 >0.99 umetric flasks. One is diluted to volume, and this value into Equation 11-5 reveals that at
Na 5.138 0.0003 0.26 0.98 0.003 0.90 >0.99 the absorbance of the solution is obtained. A the intercept. C, = -C •.
Li 5.390 0.0001 0.18 0.95 0.001 0.82 >0.99 known amount of analyte is added to the The standard addition method has the
Ba 5.210 6 x 10-4 0.41 0.99 6 x 10-3 0.95 >0.99
5.692 1 x 10-4 0.21 1 x 10-3 second, and its absorbance is measured after advantage that it tends to compensate for
Sr 0.97 0.87 >0.99
Ca 6.111 3 x 10-5 0.11 0.94 3 x 10-4 0.67 0.99 dilution to the same volume. Data for other variations caused by physical and chemical •
Mg 7.644 4 x 10-7 0.01 0.83 4 x 10-6 0.09 0.75 standard additions may also be obtained. If a interferences in the sample solution.
linear relationship between absorbance and Use of Organic Solvents. It has been found
concentration exists (and this should be that the heights of flame absorption and emis-
• Data calculated from Equation 11-4; from B. L. Vallee and R. R. ThielS. in Treatise on Analytical Chemistry.
cds. I. M. Kolthoff and P. J. Elving. New York: Interscience, 1965, Part I, vol. 6. p. 3500. With pennission. established by several standard additions 1the sion peaks are generally enhanced in the
presence of lower-molecular-weight alcohols, sidered to be one of the important tools for into a separate photometric system consisting sidered one example, the simple photometer
ketones, and esters, both in the presence and analysis.' of an interference filter (which transmits an for the simultaneous determination of sodium
in the absence of water. The effect is largely emission line of one of the elements while and potassium. Some of these efforts have
attributable to an increased nebulizer effi- Instrumentation absorbing those of the other two 1a photo- resulted in. the development of computer-
ciency; the lower surface tension of such solu- tube, and an amplifier~ The outputs can then controlled m?nochromators that permit rapid
Instruments for flame emission work are simi-
tions results in finer drop sizes and a con- be measured separately if desired. Ordinarily, sequential measurement of radiant power at
lar in construction to the flame absorption
sequent increase in the amount of sample that however, lithium serves as an internal stand- several wavelengths corresponding to peaks
instruments except that the flame now acts as
reaches the flame. ard for the analysis. For this purpose, a fixed for various elements. With such instruments,
the radiation source; the hollow cathode lamp
Leaner fuel-oxidant ratios must be em- amount of lithium is introduced into each two to· three seconds are required to move
and chopper are, therefore, unnecessary. Many
ployed with organic solvents in order to offset standard and sample. The ratios of outputs of from one peak to the next. The detector
modern instruments are adaptable to either
the effect of the added organic material. Un- the sodium and lithium transducer and the photocurrent is then measured for one or two
emission or absorption analysis.
fortunately, however, the leaner mixture results potassium and lithium transducer then serve seconds. Thus, it is possible to determine the
Much ofthe early work in atomic emission
in lower flame temperatures and a consequent as analytical parameters. This system pro- concentration of as many as ten elements per
analyses was accomplished with turbulent
increase in the possibility for chemical inter- vides improved accuracy because the intensi- minute. Instruments of this type have been
flow burners. Laminar flow burners, however,
ferences. ties of the three lines are affected in the same employed with all three types of flame
are becoming more and more widely used.
An important analytical application of way by most analytical variables, such as flame methods, although the emission method has
Speetrophotometers. For nonroutine analy-
organic solvents to flame spectroscopy is en- temperature, fuel flow rates, and background the distinct advantage of not requiring several
sis, a recording, ultraviolet-visible spectropho-
countered when it is necesSary to concentrate radiation. Clearly, lithium must be absent in sources.
tometer with a resolution of perhaps 0.5 A is
the analyte by solvent extraction with a
desirable. The recording feature provides a the sample. Simultaneous, multielement, flame emission
water-immiscible organic compound. Often in Automated Flame Photometers. Fully au- analyses have been made possible by the use
simple means for making background cor-
these applications a chelating agent, such as tomated photometers now exist for the deter- of the optical multichannel analyzer that was
rections (see Figure 11-(7).
dithizone or 8-hydroxyquinoline, converts the mination of sodium and potassium in clinical described earlier (p. 142). As an example,tO a
Photometers. Simple filter photometers
analyte ion to a species that is soluble in samples. In one of these, the samples are silicon diode vidicon tube was mounted on
often suffice for routine analyses of the alkali
organic solvents. Ordinarily, the organic ex- withdrawn sequentially from a sample turn- the optical plane originally occupied by the
and alkaline-earth metals. A low-temperature
tract is nebulized directly into the flame for table, dialyzed to remove protein and particu- slit of an ordinary grating monochromator.
flame is employed to eliminate excitation of
atomic absorption, emission, or fluorescence lates, diluted with the lithium internal stand- The diameter of the tube surface was such
most other metals. As a consequence, the
measurement. ard, and aspirated into a flame. Sample and that a 2O-nm band of radiation was contin-
spectra are simple, and glass or interference
reagent transport is accomplished with a roller- uously monitored; by adjustment of the tube
filters can be used to isolate the desired emis-
sion line. type pump. Air bubbles serve to separate along the focal plane of the monochromator,
ATOMIC samples. Results are printed out on a paper various 2O-nm bands of the spectrum could
Several instrument manufacturers supply
EMISSION SPECTROSCOPY tape. Calibration is performed automatically be observed. The resl>lution within the 20-nm
flame photomllters designed specifically for
after every nine samples. The instrument costs band was such that lines 0.14 nm apart could
Atomic emission spectroscopy (also called the analysis of sodium, potassium, and lith-
about $9000. be resolved.
flame emission spectroscopy or flame pho- ium in blood serum and other biological
Figure 11-15 shows a spectrum obtained
tometry) has found widespread application to samples. In these instruments, the radiation Instruments for simultaneously for eight elements that have
elemental analysis. Its most important uses from the flame is split into three beams of Simultaneous Multielement Analyses emission peaks in the wavelength range of
have been in the analysis of sodium, potas- approximately equal power. Each then passes
During the past decade, considerable effort 388.6 to 408.6 nm. A nitrous oxide/acetylene
sium, lithium, and calcium, particularly in has been made toward the development of flame was employed for excitation. Slightly
biological fluids and tissues. For reasons of instruments for rapid sequential or simultane- more than half a minute was required to
• Fo~ a. mOre complete discussion of the theory and
convenience, speed, and relative freedom from ous flame determination of several elements accumulate data for the eight analyses. A rel-
applications of flame emission spectroscopy, see: Flame
interferences, flame emission spectroscopy has Emission tmd Atomic Absorption Spectroscopy, vol. I: in a single sample.!! We have already con- ative precision better than 5 % was obtained.
become the method of choice for these other- TMory; voL 2: CompoMnt& tmd Teclrniqua; vol. 3: EIe-
wisedifficult-to~etermine elements. The meth- - tmd Matrices, eels. J. A. Dean and T. C. Rains.
od has also been applied, with varying degrees New York: Marcel Dekker, 1974; J. A. Dean, Flame
Photometry. New York: McGraw-Hill, 1960; and B. L.
of success, to the determination of perhaps
Vallee and R. E. Thien, in Treatiu on AnalytictU Chem-
half the elements in the periodic table. Thus, istry, eels. 1. M. Kolthofr and P. J. EJvins. New York' lO K. W. Busch. N. G. Howell, and G. H. Morrison,
• For a review of this topic, see: K. W. Busch and G. H.
flame emission spectroscopy must be con- Interscience, 1965, Part I, vol. 6, Chapter 65. . Anal. Chnn. 46, 575 (1974~
Morrison. Anal. Chern. 45 (8~ 712A (1973~
the analyte peak. For nonrccording instru- in the center are surrounded by a cooler
ments, a measurement on either side of the region which contains a higher concentration
peak suffices. The average of the two measure- of unexcited atoms; self-absorption of the res-
ments is then subtracted from.the total peak onance wavelengths by the atoms in the cooler
< height. layer will occur. Doppler broadening of the
Ie Chemic:aI Interferences. Chemical interfer- emission line is greater than the corresponding
; ences in flame emission studies are essentially
the same as those encountered in flame ·ab-
broadening of the resonarice absorption line,
however, because the particles are moving
:"<
« sorption methods. They are dealt with by ju- more rapidly in the hotter-emission zone. Thus,
S~ dicious choice of flame temperature and the self-absorption tends to alter the center of a
......
§~
-c: "
use of protective agents, releasing agents, and
ionization suppressors.
line more than its edges. In the extreme, the
center may become less intense than the edges,
:; :;
Self-Absorption. The center of a flame is or it may even disappear; the result is division
hotter than the exterior; thus, atoms that emit of the emission maximum into what appears

-'"
~;S
",,,,
MM
zz

FIGURE 11-15 Multielement flame emission spectrum


from 388.6 to 408.6 nm. [K. W. Busch, N. G. Howell, and
G.H.Morrison,Anal.Chem.,46,578 (1974). With permission
of the American Chemical Society.]

Interferences entirely upon the monochromator for selectiv-


ity; the probability of spectral interference
The interferences encountered in flame emis- due to line overlap is consequently greater.
sion spectroscopy arise from the same sources Figure 11-16 shows an emission spectrum for
as those in atomic absorption methods (see three transition elements, iron, nickel, and
p. 320); the severity of any given interference chromium. Note that several unresolved peaks
will often differ for the two procedures, exist and that care would have to be taken
however. to aVQid spectral inteiference in the analysis
Spectral Line Interference. Interference be- for anyone of these elements.
tween two overlapping atomic absorption Ba", Interference; Background Correction.
peaks occurs only in the occasional situation Emission lines are often superimposed on
where the peaks are within about 0.1 A of one bands emitted by oxides and other molecular
another. That is, the high degree of spectral species from the sample, the fuel, or the oxi- Wavelength, nm
specificity is more the result of the narrow dant. An example appears in Figure 11-17. As FIGURE 11-16 Partial oxyhydrogen flame-emission spectrum for a
line properties of the source than the high shown in the figure, a background correction sample containing 600 ppm Fe, 600 ppm Ni, and 200 ppm Cr. (Taken
resolution of the monochromator. Atomic for band emission is readily made by scan- from R. Herrmann and C. T. J. Alkemade, Chemical Analysis by Flame
emission spectroscopy, in contrast, depends ning for a few angstrom units on either side of Photometry, 2d ed. New York: Interscience, 1963, p. 527. With permission.)
1. Instruments. A major advantage of the
emission procedure is that the flame serves as
the source. In contrast, absorption methods
require an individual lamp for each element
(or sometimes, for a limited group of ele-
ments). On the other hand, the quality of the
monochromator for an absorption instrument
does not have to be so great. to achieve the
same degree of selectivity because of the
narrow lines emitted by the. hollow cathode
lamp.
2. Operator Skill. Emission methods gen-
erally require a higher degree of operator skill
because of the critical nature of such adjust-
ments as wavelength, flame zone sampled,
and fuel-to-oxidant ratio.
I
3. Background Co"ection. Correction for
2851 2853 band spectra arising from sample constituents
FIGURE 11-18 Curve A illustrates is more easily, and often more exactly, carried
out for emission methods.
the self reversal that occurs with high
concentration of Mg (2000 Jlg/ml).
4. Precision and Accuracy. In the hands of
Curve B shows the norma) spectrum a skilled operator, uncertainties are about the
same for the two procedures (±O.5 to ) ~:)
of 100 pg/ml of Mg.
relative). With less skilled personnel, atomic
absorption methods have an advantage.
FIGURE 11-17 Flame emission spectrum for a natura) brine 5. Interferences. The two methods suffer
showing the method used for correcting for background radiation. from similar chemical interferences. Atomic
(Taken from R. Herrmann and C. T. J. Alkemade, Chemical Analysis (p. 324). Both calibration curves and the stand-
by Flome Photometry, 2d ed. New York: Interscience, 1963, p. 484. ard addition method are employed. In addi-
With permission.) tion, internal standards may be used to com-
pensate for flame variables. 100

Comparison of
Atomic Emission and 75
Atomic Absorption Methods ~
to be two peaks by self-reversal. Figure 11-18 linear relationship between intensity and con- For purposes of comparison, the main advan-
'i
shows an example of severe self-absorption and centration is observed (Figure 11-19). At low tages and disadvantages of the two widely ..
.E
>
50
self-reversal. concentrations, curvature is due to the in- ";
used flame methods are listed in the para- a;
Self-absorption often becomes troublesome creased degree of ionization in the flame. Self- a:
graphs that follow. II The comparisons apply 25
when the analyte is present in high concentra- absorption, on the other hand, causes negative to versatile spectrophotometers that are readily
tion. Under these circumstances, a nonreso- departures from a straight line at higher adapted to the determination of numerous
nance line, which cannot undergo self-absorp- concentrations. elements .
.tion, may be preferable for an analysis.
Self-absorption and ionization sometimes Analytical Techniques
result in S-shaped emission calibration curves The analytical techniques for flame emission II For an excellent comparison or the two methods, see:
FIGURE 11-19 Effects of ionization and
with three distinct segments. At intermediate spectroscopy are similar to those described E. E. Piclcett and S. R. Koirtyohann, Anal. Chern. 41 self-absorption on a calibration curve for po-
concentrations of potassium, for example, a earlier for atomic absorption spectroscopy (14~ 28A (1969). tassium.
absorption procedures are less subject to spec- when nonftame atomization is employed. On as an atomic line width is so low as to restrict tube in a microwave field generated by an
the other hand, for many elements. the 'the sensitivity of the method. antenna or a resonant cavity.
tral line interferences, although such inter-
ferences are usually easily recognized and method is less sensitive and appears to have a Conventional hollow cathode lamps oper- Tunable dye lasers are beginning to find
avoided in emission methods. Spectral band smaller useful concentration range. Further- ated continuously do not provide enough use as sources for atomic fluorescence studies.
more, for comparable pCrCormance, fluores- .radiant power for fluorescence analysis. How-
interferences were considered under back- Interferences
ground correction. cence instruments appear to be more complex ever, very intense bursts of radiant energy can
be achieved by pulsing the lamps with large Interferences encountered in atomic fluo-
6. Detection Limits. The data in Table and potentially more expensive to purchase
ll-S provide a comparison of detection limits and maintain,u currents for brief periods. Provided the pulse rescence spectroscopy appear to be of the
and emphasizes the complementary nature of width and period are chosen to give a low same type and of about the same magnitude
the two procedures. average current. destruction of the lamp is as those found in atomic absorption spectros-
Inatrumentation avoided. The detector must be gated to observe; copy.t4
An instrument for atomic fluorescence mea- the fluorescence only during the pulse.
surements contains a modulated source, an Gaseous discharge lamps with sufficient: Applications
ATOMIC
FLUORESCENCE SPECTROSCOPY atomizer (flame or nonflame), a monochro- radiant power for fluorescence measurement: Atomic fluorescence methods have been ap-
mator or an interference-filter system, a de- are available for several of the more volatile plied to the analysis of metals in such materials
Since 1966, considerable research has been tector, and a signal-processing system. With elements such as the alkalis, mercury, cad-: as lubricating oils, seawater biological mate-
carried out on the principles and applications the exception of the source, most of these mium, zinc, thallium, aDd gallium. rial, graphite, aDd agricultural samples.15
components are similar to those discussed in Electrodeless discharge lamps provide high-
of atomic fluorescence spectroscopy.12 It is
evident from such work that this procedure is earlier parts of this chapter. . intensity line sources for the various elements.
somewhat more sensitive for perhaps five to A continuous source would clearly be de- Typically, these lamps consist of a sealed " See: J. D. Winefordner and R. C. Elser, Anal. Chem.
ten elements than either of the two atomic sirable for atomic fluorescence measurements. quartz tube containing an inert gas at perhaps; 43 (4~ 2SA (1971).
methods we have just considered, particularly Unfortunately, however, the output power of 1torr and the element or a salt of the analytei " For a summary of applications, see: J. D. Wineford-
a continuous source over a region as narrow element. Excitation is obtained by placing the: nero J. Chem. Educ .• 55. 12 (1978).

12 For further information on atomic ftuorescence spec-


troseopy, see: C. Veillon, in Trace Analysu. ed. J. D.
Winefotdner. New York: Wiley. 1976, Chapter VI; and USee: W. B. Barnett and H. L. Kahn, Anal. Chem •••••• -
J. D. Winefordner. J. Chem. Edue. 55, 72 (1978~ 935 (1972~ 1. Describe the basic differences between atomic emission and atomic
fluorescence spectroscopy.
2. Define the following terms: (a) atomization. «b) pressure broadening,
(c) Doppler broadening, (d) turbulent flow nebulizer, (e) laminar flow
TABLE 11-5 COMPARISON OF DETECTION nebulizer, (f) hollow cathode lamp. (g) sputtering. (h) self-absorption.
LIMITS FOR VARIOUS ELEMENTS BY (i) sensitivity. (j) detection limit. (k) spectral interference, (I) chemical
FLAME ABSORPTION AND FLAME interference, (m) radiation buffer. (n) releasing agent. (0) protective
EMISSION METHODS· agent, and (p) ionization suppressor.
Flame Absorption Flame Emission 3. Why is the CaOH spectrum in Figure 11-S so much broader than
Se•• itivity About
More Sensitive the Same More Se.-itive the Ba resonance line?
4. Why is atomic emission more sensitive to flame instability than
AI. Da, Ca. Eu, Cr, Cu, Dy. Er. Ag, As, Au, B. atomic absorption or fluorescence?
Ga, Ho. In. K. Gd. Ge, Mn, Mo, Be, Bi, Cd, Co,
La. Li. Lu. Na, Nb, Pd. Rh. Se. Fe, Hg, Ir, Mg, S. Describe the effectk which are responsible for the three different
Nd. Pr, Rb. Re. Ta, Ti, V, Y, Ni, Pb, Pt, Sb. absorbance profiles'in Figure 11-7.
Ru, Sm, Sr, Th, Zr Se, Si, So, Te,
TI. Tm, W. Yb 6. Why is a nonflame atomizer more sensitive than a flame atomizer?
Zn
7. Why is source modulation employed in atomic absorption spectros-
copy?
• Adapted from E. E. Pickett aDd S. R. Koirtyobann, Anal. Chem. 41 (141
42A (1969)and reprinted with permission. Copyright by the American Chemical 8. Describe how a deuterium lamp can be employed to provide a back-
Society. ground correction for an atomic absorption spectrum.
9. Why do organic solvents enhance atomic absorption and emission
peaks?
10. Explain why curve A in Figure 11-18 is split into a doublet.
11. The cobalt in an aqueous sample was determined by pipetting 10.0 m1
of the unknown into each of four SO.o-ml volumetric flasks. Various
volumes of a standard containing 6.23 ppm Co were added to the
flasks following which the solutions were diluted to volume. Calculate
the parts per million of cobalt in the sample from the data that follow.

Sample Unknown, ml Standard, ml Absorbance

blank 0.0 0.0 0.042


A 10.0 0.0 0.201
B 10.0 10.0 0.292
C 10.0 20.0 0.378
0 10.0 30.0 0.467
E 10.0 40.0 0.5S4

12. What is the equilibrium ratio of excited n atoms to those in the


ground state when radiation of 3776 A is absorbed and the tempera-
ture is (see Figure 11-4)
(a) 22000K?
(b) 2800 K?
0

13. For the flame shown in Figure 11-6, calculate the percent decrease in
sodium emission intensity (5893 A) when the flame is observed at
4.0 em rather than 3.0 em above the orifice?
14. Calculate the equilibrium constant for the reaction Ca ~Ca + + e- at
2800oK.
IS. Calculate the equilibrium constant for the reaction Ca + + Rb ~
Ca + Rb + at 2800oK.
16. The Doppler effect is one of the sources of line broadening in atomic
absorption spectroscopy. Atoms moving toward the light source see a
higher frequency than do atoms moving away from the source. The
difference in wavelength, lU, experienced by an atom moving at speed
v (compared to one at rest) is !M/A. = v/e, where e is the speed of light.
Estimate the line width (in A) of the sodium 0 line (5893 A) when
the absorbing atoms are at a temperature of 2()()()OK. The average
speed of an atom is given by v = J8kT/mn, where k is Boltzmann's
constant, T is temperature, and m is the mass.
The term emission spectroscopy usually refers under a single set of excitation conditions; as a advantages of the classical emission methods. certain wavelength regions. This type of emis-
to a type of atomic spectroscopy that employs consequence. spectra for dozens of elements An exception is encountered in the analysis of sion is due to molecular species in the vapor
more energetic excitation sources than the can be recorded simultaneously. This property powdered samples such as minerals or ores; state. which produce bands resulting from the
flames and ovens described in Chapter 11. is of particular importance for the multi- no practical method has yet been developed superposition of vibrational energy levels
Until recently, two types of sources were element analysis of very small samples. In con- for direct plasma excitation of such materials, upon electronic levels. The cyanogen band,·.
employed in emission methods. the electric trast. optimum flame excitation conditions and solution in a suitable solvent is required. caused by the presence of CN radicals, is
arc and the electric spark.l During the last vary widely from element to element; high Plasma excitation also appears to have all of always observed when carbon electrodes are
decade. however. argon plasma sources have temperatures are needed for excitation of some the advantages of flame excitation with few if employed in an atmosphere containing
been developed that combine many of the elements and low temperatures for others. any of the disadvantages. Thus, it is to be nitrogen. Samples containing a high silicOn
best features of flame sources with the attri- Reducing conditions work best in some analy- expected that the applications of plasma spec- content may yield an additional molecular
butes of the classical arc and spark. As a ses and oxidizing conditions in others. Finally, troscopy will continue to expand in the near band spectrum due to SiO. Another common
consequence. there has been a recent resur- the region of the flame that gives the optimum future. source of bands is the OH radical. If these
gence of interest in emission methods, and line intensity varies from element to element. bands obscure line spectra of interest, precau-
several new instruments employing argon This sensitivity to operating variables limits tions must be taken to eliminate them.
plasma sources have appeared on the market the applicability of flame methods for simul- Emission spectroscopy is based upon the
since 1977. taneous multielement analyses. line spectra produced by excited atoms. The
Arc and spark emission methods have Several disadvantages associated with arc The central' problem in analytical emission source and nature of these spectra were dis-
several advantages that account for their and spark spectroscopy account for its spectroscopy is the large influence of the cussed in the previous chapter (p. 306).
widespread use since the 19305 or earlier. displacement by flame methods for many source upon both the pattern and the inten- Arc, spark. and plasma sources are gen-
Among these is their wide applicability (70 or analyses. One disadvantage arises from the sity of lines produced by a species. It is appar- erally richer in lines than the sources
more metal or metalloid elements ~ their high fact that except for the most expensive instru- ent that the source serves two functions. First, described in Chapter 11 because of their
degree of specificity. and their sensitivity. which ments, arc and spark spectra must be it must provide sufficient energy to vaporize greater energies. Arc sources ordinarily pro-
often permits analyses in the parts-per-million recorded photographically. The time needed the sample; in this process. it is essential that vide lower energies than do spark sources; as
. or parts-per-billion range. Arc and spark emis- to develop and read a plate or film is substan- the distribution of the elements in the vapor a result. lines of neutral atoms tend to pre-
sion methods offer certain advantages over tially greater than that involved in the direct be reproducibly related to their concentration dominate in an arc spectrum. On the other
flame methods. which account for their con- acquisition of spectral data with a photo- or distribution in the sample. The second hand. spectra excited by a spark typically
tinued use even after the remarkable growth electric detector. To be sure, direct-reading function is to supply sufficient energy to cause contain lines associated with excited ions.
ofthe latter beginning in the mid-1960s. One photoelectric arc and spark instruments, which electronic excitation of the elementary par- Similarly, many of the lines found in plasma
of these advantages is the minimal sample permit the determination of two dozen or ticles in the gas, again in a reproducible spectra also arise from ions rather than
preparation required for emission methods. be- more elements in a matter of minutes. are manner. . atoms.
cause excitation can usually be carried out available. but their costs are prohibitive An examination of the emission produced
directly on liquids. powders. metals, and glass- (- $150.(00) for many laboratories. In addi- by an electrical arc or spark or by an argon
es. Second. the higher energies employed tend tion, the quantitative applications of arc and plasma reveals three types of superimposed
to reduce interelement interference. Third, spark source emission spectroscopy are spectra. First, there is continuous background
good spectra for most elements can be obtained limited, primarily because of difficulties en- radiation. For arc and spark sources. this ra- In arc and spark sources, sample excitation
countered in reproducing radiation intensi- diation is emitted by the heated electrodes and occurs in the gap between a pair of electrodes.
ties. Only with the greatest care can relative perhaps also by hot particulate matter de- Passage of electricity through the electrodes
errors be reduced to I to 2%; uncertainties tached from the electrode surface. The fre- and the gap provides the necessary energy to
on the order of 10 to 20% or greater are more quency distribution of this radiation depends atomize the sample and excite the resulting
I For a more complete discussion of arc and spark emis-
common. Where traces are being determined, upon the temperature and approximates that atoms to higher electronic states.
sion spectroscopy. see: B. F. Scribner and M. Margoshes,
in Treatise on Analytical Chemistry. cds. I. M. Kolthoff
errors of such magnitude are often tolerable of a black body (see p. 109). As will be
and P. J. Elving. New York: Interscieooe, 1965, Pan I, and not significantly greater than those asso- pointed out later. the continuum found with Sample Introduction
vol. 6. Chapter 64; J. Mika and T. Torok, Analyt/eQI ciated with other methods. In the determina- plasma sources apparently arises from a re- Samples for arc and spark source spectros-
Emission Spectroscopy. New York: Crane, Russak &: Co, tion of a major constituent, however. arc and combination of thennally produced electrons copy may be solid or liquid; they must be
1974; M. Slavin, Emission Specrroc#remieDl A/lQlys&. New
spark emission spectroscopy suffers by com- and ions. distributed more or less regularly upon the
York: Wiley-Intencience, 1971; and E. L. Grove, Analyt-
ical Emission Spectroscopy. New York: Man:el Dekker,
parison with flame methods. Band spectra, made up of a series of closely surface of at least one of the electrodes that
Part I. 1971. Part II. 1972. Plasma methods appear to offer all of the spaced lines, are also frequently observed in serves as the source.
at the analytical gap are subject to. continual
Are Sources end Are Spectre interferences, such as those found in flame change during operation. Accordingly, elec-
Metal Samples. If the sample is a metal
The usual arc source for a spectrochemical spectroscopy, are much less common. For
or an alloy, one or both electrodes can be trode spacing is arranged so that the voltage
analysis is formed with a pair of graphite or these reasons, an arc source is often preferred
formed from the sample by milling, by turn- breakdown is determined by the more repro-
metal electrodes spaced about 1 to 20 mm for qualitative analysis as well as for .the
ing, or by casting the molten metal in a mold. ducible control gap. The employment of air-
apart. The arc is initially ignited by a low cur- quantitative determination of trace COnstItu-
Ideally, the electrode will be shaped as a cooling tends to maintain constant operating
cylindrical rod that is on~ighth inch to one- rent spark that causes momentary formation ents. conditions for the control gap; a higher degree
quarter inch in diameter and tapered at one of ions for current conduction in the gap. of reproducibility is thus impart~ to the
end. For some samples, it is more convenient Once the arc is struck, thermal ionization Spark Source. analytical results than would otherwISe be at-
to employ a polished, flat surface of a large maintains the current. Alternatively, the arc Fonnation of a Spark. Figure 12-1 shows tained.
piece ofthe metalas one electrode and a graph- can be started by bringing the electrodes a typical electrical circuit for driving an The average current with a high-voltage
ite or metal rod as the other. Regardless of together to provide the heat for ionization; alternating current across a spark gap. The spark is usually significantly less than that of
the ultimate shaPe of the sample, care must they are then separated to the desired transformer converts line power to 15,00? to the typical arc, being on the order of a few
be taken to avoid contamination of the distance. 40,000 V, which then charges the capaCItor. tenths of an ampere. On the other hand,
surface while it: is being formed into an In the typical arc, currents lie in the range When the potential becomes large enough to during the initial phase of the discharge the
electrode. of 1 to 30 A. A dc source usually has an open break down the two air gaps shown on the instantaneous current may exceed 1000 A; here,
Electrodes for Noaconducting Samples. For circuit voltage of about 200 V; ac source volt- right, a series of oscillating discharges fo~lo~. electricity is carried by a narrow streamer that
nonmetaIlic mateiiaIs, the sample is supported ages range from 2200 to 4400 V. The voltage then drops until the capaCItor IS involves but a minuscule fraction of the total
on an electrode whose emission spectrum will Electricity is carried in an arc' by the incapable of forcing further curren .•across the space in the spark gap. The temperature with-
not interfere with the analysis. Carbon is an motion of the electrons and ions fonned by gaps, following which the cycle IS repeated. in this streamer is estimated to be as great as
ideal electrode material for many applications. thermal ionization; the high temperature that The frequency and current of the discharge 4O,ooooK. Thus, while the average electrode
It Can be obtained in a highly pure fonn, is a develops is a result of the resistance to this are determined by the magnitude of the cap~c- temperature of a spark source is much lower
good conductor, t\as good heat resistance, and motion by the atoms in the arc gap. Thus, the itance, inductance, and resistance of the cIr- than that of an arc, the energy in the s~all
is readily shaped. jManufacturers offer carbon arc temperature depends upon the composi- cuit in Figure 12-1; these parameters also volume of the streamer may be several times
electrodes in many sizes, s~apes, and fonus. tion of the plasma, which in turn depends affect the relative intensities of the neutral greater. As a consequence, ionic spectra are
Frequently, one of the electrodes is shaped as upon the rate of formation of atomic particles atom and ion lines as well as the background more pronounced in a high voltage spark than
a cylinder with a small crater drilled into one from the sample and the electrodes. Little is in an arc.
intensity.
end; the sample is then packed into this cavity. known of the mechanisms by which a sample Most spark sources employ a pair of spark
The other electrode is commonly a tapered is dissociated into atoms and then volatilized gaps arranged in series (Figure. 12-1). The
carbon rod with a slightly rounded tip. This in an arc. It can be shown experimentally, Laser Excitation
analytical gap is formed by a pair of carbon
configuration appears to produce the most however, that the rates at' which various or metal electrodes that contain the sample. A pulsed ruby laser provides sufficient energy
stable and reproducible arc or spark. species are volatilized differ widely. The The control gap may be formed from smooth- to a small area on the surface of a sam~le. to
Silver or copper rods are also employed to spectra of some species appear early and then ly rounded and carefully spa~ elect~~es cause atomization and excitation of emIssIon
hold samples when these elements are not of disappear; those for other species reach their over which is blown a stream of aIr. Conchtlons lines. This source is far from i~cal, ~owever,
analytical interest. The surfaces of these elec- maximum intensities at a later time. Thus, the because of high backgroun~ IOte~s.lty, self-
trodes can be cleaned and reshaped after each composition of the plasma, and therefore absorption, and weak line IOt.ens.llIes.One

{8
Inductor
analysis. the temperature, may undergo variation with technique that has shown promIse ISthe ~aser
Excitation of the Constituems of Solutions. time. Typically, the plasma temperature is microprobe, which is available comme~ClaIlY.
eontrOI
Several techniques ,are employed to excite the 4000 to 5OOOoK.
components of solutions or liquid samples.
One common method is to evaporate a
The precision obtainable with an arc is
generally poorer than that with a spark and
II Capacitor 1
spark gap
Analytical
spark gap
Here, the laser is employed to vapo~ze the
sample into a gap between two graphlt~ el.ec-
t odes which serves as a spark eXCItatIon
measured quantityiof the solution in a small much poorer than that with a plasma or r ce' The device is applicable even to non-
sour . . rface
cup into the surfade of a graphite or il metal flame. On the other hand, an arc source is Resistor conducting materials and prOVIdes a su .
electrode. Alternatively, a porous graphite more sensitive to traces of an element in a analysis of a spot no greater than 50 1JD1 m
electrode may be saturated by immersion in sample than is a spark source. In addition, FIGURE 12-1 Power supply for a high-
diameter.
the solution; it is then dried before use. because of its high temperature, chemical voltage spark source.
an aerosol, a thermally generated vapor, or a
fine powder.
By definition, a plasma is a gaseous mixture The most widely used apparatus for sample
in which a significant fraction of the atomic injection is similar in construction to the
or molecular species present are in the form nebulizer employed for flame methods. Figure
of ions. The plasma employed fol' emission 12-3 shows a typical arrangement. Here, the
analyses is generally made up of a mixture of sample is nebulized by the flow of argon, and
argon ions and atoms. When a sample is in- the resulting finely divided droplets are carried
jected into this medium, atomization occurs into the plasma. Aerosols have also been pro-
as a consequence of the high temperature, duced from liquids and solids by means of an
which may be as great as 10,OOO°K. Argon ultrasonic nebulizer. Still another method of
ions, once formed in a plasma, are capable of sample introduction involves deposition of the
absorbing sufficient power from an external sample on a tantalum strip and vaporization
source to maintain a sufficiently high temper- by passage of a large current. The vapors are
ature to continue the ionization process and then swept into the plasma by the argon stream.
thus sustain the plasma indefinitely. Three No practical method has yet been devised for
power sources have been employed in argon the introduction of powder samples such as
plasma spectroscopy. One is a de electrical those encountered in minerals and ores.
source capable of maintaining a current of Plasma Appearance and SpectrL The typi-
several amperes between electrodes immersed cal plasma has a very intense, brilliant white,
in a stream of argon. The second and third nontransparent core topped by a flamelike
are powerful radio frequency and microwave tail. The core, which extends a few millimeters

i ~
tangential above the tube, is made up of a continuum
frequency generators through which the argon ~gon plasma
is flowed. Of the three, a radio-frequency 11 - support flow . upon which is superimposed the atomic spec-
source appears to offer the greatest advantage Sample trum for argon. The source of the continuum
in terms of sensitivity and freedom from inter- _0101 or vapor apparently arises from recombination of argon
In_ FIGURE 12-3 A typical nebulizerfor sam-
ference. Thus, the discussion will focus on ple injection into a plasma source. [From V. A. and other ions with electrons. In the region 10
these, which are often called inductively coupled FIGURE 12-2 Atypical inductively to 30 mm above the core, the continuum fades,
Fassel, Science, 202, 186 (1978). With per-
p/4sma sources (sometimes abbreviated ICP). coupled plasma source. (From V. A. and the plasma is optically transparent. Spec-
mission. Copyright 1978 by the American As-
FasseJ, Science, 2O:Z, 185 (1978). With tral observations are generally made at a height
sociation for the Advancement of Science.]
permission. Copyright 1978 by the of 15 to 20 mm above the induction coil. Here,
The Inductively American Association for the Advance- the background radiation is remarkably free of
Coupled Plasma Source" ment of Science.) argon lines and is well suited for analysis.
Many of the most sensitive analyte lines in this
Figure 12-2 is a schematic drawing of an in-
The temperature of the plasma formed in region of the plasma are from ions such as
ductively coupled plasma source. It consists
tion coil that is powered by a radio frequency this way is high enough to require thermal Ca(I), Ca(I1), Cd(I), Cr(I1), and Mn(I1).
of three concentric quartz tubes. through
generator, which produces 2 kW of energy at isolation from the outer quartz cylinder. This Analyte Atomization and Ionization. Figure
which streams of argon flow at a total rate
about 27 MHz. Ionization pf the flowing isolation is achieved by flowing argon tangen- 12-4 shows temperatures at various parts of
between 11 and 17 liters/min. The diameter of
argon is initiated by a spark from a Tesla coil. tially around the walls of the tube as in- the plasma. By the time the sample atoms
the largest tube is about 2.5 em. Surrounding
The resulting ions, and their associated elec- dicated by the arrows in Figure 12-2; the flow have reached the observation point, they will
the top of this tube is a water-cooled induc-
trons, then interact with the ffuctuating mag- rate of this stream is 10 to 15 liters/min. The have had a residence time of about 2 ms at
netic field (labeled H in Figure 12-2) pro- tangential flow cools the inside walls of the temperatures ranging from 6000 to 8OOOoK.
duced by the induction coil. This interaction center tube and centers the plasma radially. These times and temperatures are roughly
causes the ions and electrons within the coil Sample Injection. The sample is carried twice as great as those found in the hotter
• For a more complete discussion, see: V. A. Fassel, into the hot plasma at the head of the tubes combustion flames (acetylene/nitrous oxide)
Scienu,m. 183 (1978); A. F. Ward, Amer. Lob. 10 (11~ to flow in the closed annular paths depicted
in the figure; ohmic heating is the consequence by argon flowing at about 1 liter/min through employed in flame methods. As a consequence,
79 (1978); and R. M. Ajhar, P. D. Dalager. and A. L
Davison. Amer. Lob. 8 (3~ 71 (1976~ of their resistance to this movement. the central quartz tube. The sample may be atomization is more complete, and there are
Temperature. oK INSTRUMENTS FOR
(+10%)
EMISSION SPECTROSCOPY
6000 .' ~~.i:.- _,.,~J!,_'C>,¥;~r r "

Instruments for emission spectroscopy are of -.


:••••.
..
;"--'~"
~..~.•~.....
·•.•.••
;•iW
f'""
' ' ' ii' •.....'.
I
....••·..-.. :.:..•••.- ....••...
r,

two types, simultaneous multielement spec- tz:JIJ.J.· ~Y~I~~~'~',.....


6500
trometers and sequential spectrophotometers.
The former are capable of detecting and
~l':F"';:' ~fr·"' I' .
6800 measuring emission lines for a large number
"f'l( , '111 III r'~
6 , 'II I'l'l III I f~
"
8000
of elements (sometimes as many as SO or 60)

=
(estimate)
10,000 simultaneously. Sequential instruments, on the
5
.4 I , II ~
other hand, measure line intensities for var- H.~ ,
I II,.
ious elements on a one-by-one basis. Here, a 2 '~
wavelength change is required after the signal I '1'. 'P1'!' 1111'!'II'fP"
for each element has been monitored. Clearly,
the total excitation time required for a multi-
element analysis will be significantly greater
with the sequential instrument; as a conse- FIGURE 12-5 Typical spectra obtained with a 3.4-meter grating spectrograph. Numbers on
FIGURE 12-4 Temperature in a typical
quence, such instruments are considerably horizontal axis are. wavelengths in A. Spectra: (1) iron standard; (2)-(5) casein samples;
more costly in terms of elapsed time and (6)-(8) Cd-Ge arsenide samples; (9)-(11) pure Cd, Ge, and As, respectively; (12) pure graphite
inductively coupled plasma source. (From V.
sample consumption. electrode.
A. Fassel, Science, 102, 187 (1978). With
Arc and spark sources demand the use of
permission. Copyright 1978 by the American
a simultaneous multielement instrument be-
Association for the Advancement of Science.)
cause of their instability. In order to obtain
reproducible line intensities with these sources,
it is necessary to integrate and average a signal elements in a few minutes. They are, however, by photographic means, the exit slit of a
for at least 20 s and often for a minute or expensive (-SISO,OOO) and not very versatile; monochromator (see Figure 5-7) is replaced
more. When several elements must be deter- they cannot ordinarily be switched from the by a photographic plate (or film) located
mined, the time required and the sample con· determination of one set of elements to another. along the focal plane of the instrument.3 After
fewer interference problems due to dissociation sumed by a sequential procedure is usually un· Their sum can only be justified for a situation exposure and development of the emulsion,
reactions (p. 322~ Surprisingly, ionization aa:eptable. demanding a large number of routine analyses the various spectral lines of the source appear
interference effects (p, 323) are small or non- Simultaneous multielement instruments are for a certain set of elements. as a series of black images of the entrance slit
existent, perhaps because the electron con· of three types. One is a spectrograph, which Plasma sources are sufficiently stable to distributed along the length of the plate (see
centration from ionization ofthe argon is large employs a photographic film or plate as a permit completion of line-intensity measure- Figure 12-5). The location of the lines pro-
compared with that resulting from ionization detector and integrator. The second is a photo- ments in one or two seconds. As a conse- vides qualitative information concerning the
of sample components. electric spectrometer that employs an array of quence, sequential instruments have become a sample; the darkness of the images can be
Several other advantages are associated photomultiplier tubes (one for each element) reasonable alternative to simultaneous instru- related to line intensities and hence to con-
with the plasma source. First, atomization located at appropriate positions on the focal ments. While not as rapid as the photoelectric centrations.
occurs in a chemically inert environment, plane of a spectrometer; a suitable integrating multielement spectrometer, sequential instru-
;

which should also enhance the lifetime of the circuit is required for each detector. The third ments often compete favorably in time con- Types of Monochromators
analyte. In addition, and in contrast to arc, type employs a vidicon tube as a detector sumption with spectrographs and provide the
convenience of photoelectric detection. In ad- The dispersing element in a spectrograph may
spark, and flame, the temperature cross sec- (see p. 143). Spectrographs are relatively in· be either a prism or. a grating. High-
tion of the plasma is relatively uniform; as a expensive. On the other hand, processing of dition, the cost of sequential spectrophotom-
consequence, self-absorption and self-reversal films or plates is somewhat time consuming eters is comRCtitive with spectrographs.
effects (p. 329) are not encountered. Conse- and generally inconvenient when compared
• The focal points of a prism spectrometer lie on a very
quently, linear calibration curves over several with photoelectric detection. Photoelectric
Spectrographs slightly curved surface. To compensate for this curvature,
orders of magnitude of concentration are often multielement spectrometers are capable of pro- spectrographic plate holders are designed to force a slight
observed. viding for the determination of 20 or 30 To permit simultaneous intensity recording bend in the photographic plate.
,.,/
.".-- -------- -' . ....,
Reflecting
Treatment with a reducing agent results in the
formation of a large number of silver atoms
/// prism '" for each absorbed photon. This process is an
'\ example of chemical amplification; the number 1.5
Photographic of black silver particles, and thus the darkness ~2
plate or
film of the exposed area, is a function of the Cl

exposure E, which is defined by i


.~
..,•.
Short
wavelangths 1.0
I 1ii
I .!!
Q.
0
where 11 is the intensity of radiation and t is
the exposure time.' In obtaining spectra for 0.5

quantitative purposes, t remains constant for


both sample and standards. Under these cir-
cumstances, exposure is directly proportional
to the line intensity and thus to the concen- 1.0
FIGURE 12-6 The Eagle mounting for a grating spectro- log,. exposure
tration of the emitting species.
graph.
The blackness of a line on a photographic FIGURE 12-8 Typical plate calibration
plate is expressed in terms of its optical den- curves at two wavelengths.
sity D, which is defined as
disPc:rsion prism instruments employ a Lit- are generally provided with a masking device 10
called a Hartmann diaphragm, by means of D=log
trow:mount monochromator; a Cornu-type is T only relative exposures can be calculated from
ofte~ used in lower dispersion spectrographs. which only a fraction of the vertical length of
Prism spectrographs suffer from the disad- the slit is iIluminated (Figure 12-7). Adjust- where 10 is the intensity (power) of a beam of Equation 12-1.
vantage of nonlinear dispersion, which compli- ment of the diaphragm permits the successive radiation after it has passed through an Figure 12-8 shows typical calibration
cates the identification oflines and also crowds recording of several spectra on a single plate unexposed portion of the emulsion, and 1 is curves for an emulsion obtained at two wave-
the longer wavelength lines together. or film (see Figure 12-5). its intensity after being attenuated by the line. lengths. With the aid of such curves, the
Reflection gratings are widely used in the Photopapbie Detection. A photographic Note the similarity between optical density measured optical densities of various analyti-
manufacture of spectrographs. One of the emulsion can serve as a detector, an amplifier, and absorbance; indeed, the former can be cal lines can be related to their rekltive inten-
common arrangements is shown in Figure and an integrating device for evaluating the measured with an instrument that is similar sities. These relative intensities are the concen-
12-6.4 intensity-time integral for a spectral line. When to the photometer shown in Figure 7-8. tration-dependent parameter employed in
The spectral region of greatest importance a photon of radiation is absorbed by a silver In order to convert the optical density of a quantitative spectrographic analysis. Several
for emission analysis lies between 2500 and halide Particle of the emulsion, the radiant line to its original intensity (or exposure), it is calibration curves are needed if a spectrum
4000 A, although both longer and shorter energy is stored in the form of a latent image. necessary to obtain empirically a plate cali- covers a large wavelength range because of the
wavelengths are occasionally useful. Thus, the bration curve, which is a plot of the optical dependence ofthe slope on wavelength region
optical components through which radiation density as a function of the logarithm of rela- (Figure 12-8).
is transmitted must be of quartz or fused tive exposure. Such a plot is obtained experi-
silica.. and the detector must be sensitive to mentally by exposing portions of a plate to
• ; Entrarw:.e slit Simultaneous
ultraviolet radiation. radiation of constant intensity for various
Multielement Spectrometers

~.
The Hartmann Diaphragm. Spectrographs lengths of time; generally, the absolute inten-
n sity of the source will not be known, so that

! II
.11
For routine emission analyses, photoelectric
detection of integrated line intensities can be
employed. Instruments for this purpose are
HeR also, the spectral lines are focused on a curved
equipped with as many as 48 photomultiplier
4

surface. With smaller instruments, the curvatuR may be


< ) s Here, we follow tbe convention or emission spectros- tubes, which are located behind fixed slits
copists and employ the term intensity I, rather tban the
10 areat u to preclude the use of ••••• photographic FIGURE 12-7 The Hart- power P. The dift'erenceis small (p. 93~ and for practical
along the focal curve of the spectrometer (see
plates; 8exible 111mmust tbus be employed. mann diaphragm. purposes, the terms can be considered synonomous. Figure 12-9). The position of each slit must
Sequential Spectrometer. Qualitative Analy.i.
It is practical to employ automated sequential Photographic emission spectroscopy permits
spectrometers for emission analysis with an the detection of some seventy elements by
inductively coupled plasma source because brief excitation of a few milligrams of sample.
the relative stability of this source permits Furthermore, from a subjective judgment of
intensity measurements to be completed in a line blackening, the chemist with a little ex-
few seconds. Instruments of this kind are sim- perience can rapidly estimate the concentra-
ilar to the automated flame emission instru- tion of an element with an accuracy of perhaps
ment described earlier (p. 327). one order of magnitude. In this application,
photographic emission spectroscopy is unsur-
passed.
APPLICATIONS OF The sensitivity of spectrographic methods
EMISSION SPECTROSCOPY depends upon the nature and amount of sam-
ple, the type of excitation, and the instrument
Spectroscopic instruments that rely upon pho- employed. Table 12-1 lists some representative
tographic detection are particularly useful for data for lower limits of detection that can be
qualitative or semiquantitative analyses. Pho- achieved with dc arc excitation. .
toelectric instruments find wide use for the Excitation Technique. For qualitative work,
routine quantitative analysis of a limited a dc arc is most often employed. The current
number of elements. and exposure times are adjusted so that com-

--1-8 lodicllto' TABLE 12-1 DETECTION LIMITS WITH A


crilis
DC ARC·
FIG UR E 12-9 Optical system of a direct-reading grating
spectrometer. (Courtesy of Baird Corporation, Bedford, Mas- Lower Umit of Detection"
sachusetts. ) ra"eIength,
Element Percent Micrograms

Ag 3280.68 0.0001 0.01


As 3288.12 0.002 0.2
B 2497.73 0.0004 0.04
Ca 3933.66 0.0001 0.01
correspond exactly to the position of a line of which are chosen at the time of instrument Cd 2288.01 0.001 0.1
interest. manufacture. For rapid routine analysis, Cu 3247.54 0.‫סס‬OO8 0.008
K 3446.72 0.3 30
The output of each photomultiplier is fed however, such spectrometers are often ideal. Mg 2852.12 0.00004 0.004
into a suitable capacitor-resistor circuit which For example, in the production of alloys, Na 5895.92 0.0001 0.01
integrates the output over a period of 10 to quantitative analyseS for 5 to 20 or more ele- P 2535.65 0.002 02
40 s. The capacitor voltage at the end of the ments can' be accomplished within five min- Pb 4057.82 0.0003 0.03
exposure is a function of its total charge utes of receipt of a sample; close control over Si 2516.12 0.0002 0.02
Sr 3464.45 0.02 2
which, in turn, is the product of the detector the composition of the final product thus be- 0.001 0.1
Ti 3372.80
current multiplied by time; thus, this voltage comes possible. Zn 3345.02 0.003 0.3
provides a measure of the time integral of the In l£ddition to speed, photoelectric spec-
line intensity. trometers offer the advantage of better analyt-
• Data taken from N. W. H. Addink,Speclrochim. Acro, II, 168 (1957). With
Photoelectric multielement instruments are ical precision. Under ideal conditions repro-
permission.
highly automated, complex, expensive, and ducibilities of the order of 1 % relative of the • Values are based upon complete volatilization of a IQ-mg sample in a de
permit measurement of only certain elements, amount present have been demonstrated. arc.
plete volatilization of the sample occurs; quantitative emission analysis is the very large 5. The lines of the standard and the emission method involves the preparation· or
currents of 5 to 30 A for 20 to 100 s are number of variables that aifect the blackness analyte should be similar in intensity and acquisition of a set of standard samples from
common. Generally, 2 to SO mg of sample are of the image of a spectral line on a photo- should be in the same spectral region so as to which calibration curves are prepared. For
packed into the cavity of a graphite electrode graphic plate or the intensity of a line reaching provide adequate compensation for emulsion the ultimate in accUracy, the standards must
or evaporated upon its surface; a second (or the photoelectric detector. Most variables that variables (or differences in detector response closely approximate the samples both in
counter) electrode, also of graphite, then com- are associated with the excitation and the with photoelectric spectrometers). chemical composition and physical form;
pletes the arc. The sample electrode is ordi- photographic processes are difficult or im- It is seldom possible to find an internal their preparation often requires a large expend-
narily made the anode because of its higher possible to control completely. In order to standard that will meet aU of these criteria, iture of time and effort-an expenditure that
temperature. On the other hand, an enhance- compensate for their effects, an internal stan- and compromises must be made, particularly can be justified economically only if a large
ment of spectral intensities is sometimes ob- dard is generally employed. where the same internal standard is used for number of analyses is anticipated.
served in the vicinity of a sample electrode An internal standard is an element incor- the determination of several elements. In some instances, standards can be syn-
that is operated as a cathode, as a result of porated in a fixed concentration into each If the samples to be analyzed are in solu- thesized from pure chemicals; solution samples
retarded diffusion by positive ions in the arc sample and each calibration standard. The rel- tion, considerable leeway is available in the are most readily prepared by this method.
plasm~. Advantage is sometimes taken of this ative intensity of a line of the internal stan- choice of internal standards, since a fixed Standards for an alloy analysis might be pre-
phenomenon by using radiation adjacent to dard is then determined for each sample or amount can be introduced volumetrically. pared by melting together weighed amounts of
the cathode as the source. standard at the same time the relative intensi- Here, an element must be chosen whose con- the pure elements.
~ification of Elements. For a qualita- ties of the elements of interest are measured. centration in the sample is small relative to Another common method involves chemi-
tive analysis, a standard iron spectrum is A ratio of relative intensity of the analyte to the amount to be added as the internal stan- cal analysis of a series of typical samples en-
usually recorded as a reference on each plate that of standard is then employed for the con- dard. compassing the expected concentration range
or film. Identification of unknown lines is centration determination. Experimentally, a The introduction of a measured amount of of the elements of interest. A set of standards
greatly facilitated with a projector that per- direct proportionality often exists between an internal standard is seldom possible with is then chosen on the basis of these results.
mits the simultaneous comparison of this this ratio and concentration. Occasionally, metallic samples. Instead, the major element The National Bureau of Standards has
plate iwith a master iron spectrum upon however, the relationship is nonlinear; the re- in the sample is chosen, and the assumption is available a large number of carefully analyzed
which' the wavelengths of the more intense sulting curve is then employed for concentra- made that its concentration is essentially metals, alloys, and mineral materials; occa-
iron lines have been identified. Also indicated tion determinations. invariant. For example, in the quantitative sionally, suitable standards can be found
on the master spectrum are the positions of Criteria in the Choice of an Internal Stam- analysis of the minor constituents in a brass, among these. Standard samples are also avail-
several of the most sensitive lines for each of ani. The ideal internal standard has the fol- either zinc or copper might be employed as able from commercial sources.6
the elements. The analytical plate' and the lowing properties. the internal standard. The Method or Standard AdditioM. When
comparison plate are then aIigned so that the 1. Its concentration in samples and stan- For powdered samples, the internal stand- the number of samples is too small to justify
two iron spectra coincide exactly. Identifi- dards is always the same. ard is sometimes introduced as a solid. Weigh- extensive preliminary work, the method of
cation of three or four of the prominent lines 2. Its chemical and physical properties are ed quantities of the finely ground sample and standard additions described on page 325
of a particular element constitutes nearly as similar as possible to those of the element the internal standard are thoroughly mixed may be employed.
certain evidence for its presence in the sample. being determined; only under these circum- prior to excitation. Semiquantitative Methods. Numerous semi-
stances will the internal standard provide The foregoing criteria provide theoretical quantitative spectrographic methods have
Quantitative Applications
adequate compensation for the variables as- guidelines for the selection of an internal been described which provide concentration
Quantitative emission analyses demand precise sociated with volatilization. standard; nevertheless, experimental verifica- data reliable to within 30 to 200% of the true
control of the many variables involved in 3. It should have an emission line that has tion of the effectiveness of a particular element amount of an element present in the sample.7
sample preparation and excitation (and also in about the same excitation energy as one for and the line chosen is necessary. These ex-
film processing with spectrographs). In ad- the element being determined, so that the two periments involve determining the effects of • For a listing of sources. see: R. E. Michaelis. Report on
dition.. quantitative measurements require a set lines are similarly affected by temperature variation in excitation times, source tempera- Available Standard Samples and Related Materials for
of cantully prepared standards for calibration; fluctuations in the source. tures, and development proced ures on the Spectrochemical Analysis, Am. Soc. Testing Materials
these standards should approximate as closely 4. The ionization energies of the internal Spec. Tech. PubI. No. j8-E. Phil8delphia: American Soci-
relative intensities of the lines of the internal
ety for Testinl aDd Materials. 1963.
as possible the composition and physical standard and the element of interest should standard and the analytes. 7 For references to a more complete description of IOIIIe
properties of the samples to be analyzed. be similar to assure that both have the same Standard Samples. In addition to the ICIlIiquantitative procedun:s, lee: C. E. HArvey, Spectra-
Internal Standanls •••• nata Treatment. As distribution ratio of atoms to ions in the choice of an internal standard, a most critical CMniclll Procedures. Glendale, CA: Applied Research
we have pointed out, the central problem of source. phase in the development of a quantitative Laboratories. 1950. Chapter 7.
x Deionized water These methods are useful where the prepara- ments, along with fewer chemical interference In gener~ the detection limits with the in-
Ii. 5000 I'g/ml NI ductively coupled plasma source appear com-
tion of good standards is not economical problems. For example, Figure 12-10 shows
D TIP Wiler (500 ppm hlrdnessf Be
o 200 I'g/ml Ca, 200 I'g/ml Mg Several such procedures are based upon the calibration curves for several elements. Note parable to or better than other atomic
total vaporization of a measured quantity (1 that some of the data were obtained from so- spectral procedures. Table 12-2 compares the
to 10 mg) of sample in the arc. The concen- lutions prepared by dissolving Ii compound of sensitivity of several of these methods. Note
tration estimate may then be based on a the element in pure water; others contained that more elements can be detected in the
knowledge of the minimum amount of an ele- high concentrations of different species, indi- ten-parts-per-billion (or less) range with plas-
ment required to cause the appearance of each cating a freedom from interelement interfer- ma excitation than with any other of the
of a series of lines. In other methods, optical ences. Note also that the curves cover a methods .
.2 concentration range of nearly three orders of
densities of elemental lines are measured and
compared with the line of an added matrix magnitude.

f~4/
material, or with the :background. Concentra-
tion calculations are then based on the assump- 1. State the differences between emission spectroscopy described in this
tion that the line intensity is independent of chapter and atomic emission spectroscopy described in Chapter 11.

W
the state in which the element occurs. The 2. Explain how an inductively coupled plasma source produces high
effects of other elemc!nts on the line intensity intensity emission spectra for elements.
can be minimized by mixing a large amount 3.· What is an internal standard and how is it used?
of a suitable matrix material (a spectroscopic 4. Explain the statement that ••a photographic emulsion can serve as a
buffer) with the sample. detector, an amplifier, and an integrating device for evaluation of the
0.1 0.3 1.0 3.0 It is sometimes possible to estimate the
Concentrotian,/'g/ml
intensity-time integral for a spectral line."
concentration of an :e1ement that occurs in
5. Element 1 exhibited an emission line at wavelength At in Figure 12-8.
small amount by cor+paring the blackness of
A standard mixture containing this element at a concentration of 20.0
FIGURE 12-10 Calibration curves several of its Iineswi(h a number of lines of a
ppb exhibited optical density of 1.10 at AI' An alloy containing this
with an inductively coupled plasma major constituent. Matching denSities are then
element produced an optical density of 0.68 at this same wavelength.
source. Here, an yttrium line at 242.2 nm used to establish the concentration of the minor
Use Figure 12-8 to estimate the concentration (in ppb) of the element
served as an internal standard. Notice constituent.
in the alloy.
the lack of interelement interference. Quantitative Applications with the Induc-
(From V. A. Fasse~ Science, 202, 187 tively Coupled Plasma Source. As we have 6. Element 2, which emitted at Az (Figure 12-8) was used as an internal
(1978). With permission. Copyright noted earlier, the inductively coupled plasma standard for element 1 (with emission at A-1 ~ A standard containing
1978 by the American Association for source appears to offer all the advantages of 11.8 ppb of element 1 and 94.7 ppb of element 2 yielded optical densi-
the Advancement of Science.' the flame methods for quantitative measure- ties of 1.00 at Al and 0.92 at Az . Element 2 was mixed at a concentra-
tion of 85.0 ppb with an unknown containing element 1. This sample
produced optical densities of 0.75 at Al and 0.95 at Az. Assuming that
the emission from each element has been affected by the matrix
TABLE 12-2 COMPARISON OF DETECTION LIMITS FOR SEVERAL
equally, calculate the concentration of element 1 in the unknown.
ATOMIC SPECTRAL METHODS·
Number of Elements Detected at Concentratiom; of
Method <Ippb 1-10 ppb 11-100 ppb 101-soO ppb > 500 ppb
Inductively coupled
plasma emission 9 32 14 6 0
Atomic emission 4 12 19 6 19
Atomic fluorescence 4 14 16 4 6
Atomic absorption -I. 14 25 3 14

• Data from V. A. Fassel and R. N. Kniseley. Anal. C"-.


48 (13~ IIIIA (1974~ Reprinted with permission.
Copyright by the American Chemical Society. Detection limits correspond to a signal that is twice as great as the
deviation for the background noise.
This chapter contains brief discussions of ing refractive indexes and 01 and O2 are the
some optical methods which, while impor- angles of incidence and refraction, respec-
tant, find less general use by the average prac- tively (see Figure 4-9). When MI is a vacuum,
"I is unity because VI becomes equal to c in
ticing chemist.
Equation 13-1. Thus,
c sin 0
"2 = "va< = -V2 = -.--
sm O
1

2
(13-3)

When radiation passes through a transparent where n..< is the absolute refractive index of
medium, interaction occurs between the elec- M2• Thus, "va< can be obtained by measuring
tric field of the radiation and the bound elec- the two angles 01 and 2, °
It is much more convenient to measure the
trons of the medium; as a consequence, the
rate of propagation of:the beam is less than in refractive index with respect to some medium
a vacuum (see p. 102). The refractive index of a other than vacuum, and air is commonly em-
substance "I at a wavelength i is given by the ployed as a standard for this purpose, Most
relationship compilations of" for liquids and solids in the
literature are with reference to air at laboratory
temperatures and pressures, Fortunately, the
change in the refractive index of air with
where v, is the velocity of propagation of ra- respect to temperature and pressure is small
diation in the medium' and c is the velocity in enough so that a correction from ambient
i
vacuum (a constant under all conditions). laboratory conditions to standard conditions
is needed for only the most precise work. A
The refractive index Qf most liquids is in the
range between 1.3 and 1.8; it is' 1.3 to 2.5 or refractive index "0 measured with respect to
higher for solids. I air with radiation from the D line of sodium
can be converted to "va< with the equation

The Measurement
of Refl'llctive Index This conversion is seldom required.
It is usually necessary to measure a refrac-
The refractive index ofa substance is ordinarily
tive index to an accuracy of at least 2 x 10- 4.
determined by measuring the change in direc-
Accuracies on the order of 6 to 7 x 10- S may
tion (refraction) of collimated radiation as it
be required for the routine analysis of solu-
passes from one medium to another. As was tions. For the detection of impurities, a differ-
shown on page 103, ence in refractive index between the sample
"2 VI sin 01 and a pure standard is measured; here, the
"I = V2; = sin O2 capability of detecting a difference on the
order of I x 10-6 or better is needed.
where v I is the velocity of propagation in the
less dense medium M,and V2 is the velocity
in medium M 2; "I and "2 are the correspond- Variables That Affect
Refractive Index Measurements
Temperature, wavelength, and pressure are the
1 For a more complete discussion or rerractometry, see: most common experimentally controllable
S. Z. Lewin and N. Bauer, in Treatise on Analytical variables that affect a refractive index measure-
Chemistry, eds. I. M. Kolthoff and P. J. Elving. New
York: Interscience, 1965, Part I, vol. 6, Chapter 70. ment.
Temperature. Temperature influences the only for precise work with gases and for the terms of ~he sodium D line. The compensator
refractive index of a medium primarily be- most exacting work with liquids and solids is consists of one or two Amici prisms, as shown
cause of the accompanying change in density. the variation in atmospheric pressure im- in Figure 13-2. The properties of this complex
For many liquids, the temperature coefficient portant. prism are such that the dispersed radiation is
lies in the range of -4 to -6 x 10-4 deg-1• converged to give a beam of white light that
Water is an important exception. with a Instruments for travels in the path ofthe yellow sodium D line.
coefficient of about -1 x 10-4; aqueous so- MNauring Refractive Index FIGURE 13-2 Amici prism for compensa- Ablie Refractometer. The Abbe instrument
lutions behave similarly. Solids have tempera- tion of dispersion by sample. Note that the is undoubtedly the most convenient and widely
Two types of instruments for measuring re- used refractometer; Figure 13-3 shows a sche-
ture coefficients that are roughly an order of yellow radiation (sodium D line) suffers no net
fractive index are available from commercial maticdiagram ofits optical system. The sample
magnitude smaller than that of the typical deviation from passage through the prism.
sources. Refractometers are based upon mea- is contained as a thin layer (-0.1 mm)
liquid.
surement of the so-called critical angle or upon between two prisms. The upper prism is firmly
It is apparent from the foregoing that the
the determination of the displacement of an mounted on a bearing that permits its rotation
temperature must be controlled closely for
image. Interferometers utilize the interference
accurate refractive index measurements. For by means of the side arm shown in dotted
phenomenon to obtain differential refractive lines. The lower prism is hinged to the upper
the average liquid, temperature fluctuations Figure 13-1a illustrates the critical angle that
indexes with very high precision. We shall
should be less than ±O.2°C if fourth place is formed when the critical ray approaches to allow separation for cleaning and for intro-
consider only refractometers.
accuracy is required, and ±0.02°C for mea- the surface of the medium M 2 at 90 deg to the
Critical ADgIe Refractometen. The .most
surements to the fifth place. normal and is then refracted at some point 0
widely used instruments for the measurement
W.,elength of Wiatioa. As noted on page on the surface. Note that if the medium could
of refractive index are of the critical angle
102, the refractive index of a transparent me- be viewed end-()fi, as in Figure 13-1b, the crit-
type. The critical angle is defined as the angle
dium gradually decreases with increasing wave- ical ray would appear as the boundary be-
of refraction in a medium when the angle of
length; this effect is referred to as normal dis- tween a dark and a light field. It should be
the incident radiation is 90 deg (the grazing
persion. In the vicinity of absorption bands, noted, however, that the illustration is unre- ~ .•.•'.. Field of
angle); that is, when 61 in Equation 13-2 is 90
rapid changes in refractive index occur; here, alistic in that the rays are shown as entering ~View
deg, 62 becomes the critical angle 6•. Thus,
the dispersion is referred to as anomalous (see the medium at but one point 0; in fact, they Eyepiece
Figure 4-8, p. 102). n2 sin 90 would be expected to enter at all points along
Dispersion phenomena make it essential n1 = sin 6. = sin 6. (13-5) the surface and thus create an entire family of
that the wavelength employed be specified in critical rays with the same angle 6•. A con-
quoting a refractive index. The D line from a densing or focusing lens is needed to produce
sodium vapor lamp (l •••589 nm) is most com- a single dark-light boundary such as shown in
monly used as a source in refractometry, and
""""
Figure 13-tb.
the corresponding refractive index is desig- It is important to realize that the critical " //
/
/
nated as no (often the temperature in °C is also Adjustable ,,/
angle depends upon wavelength. Thus, if poly- Amici ~/ //
indicated by a superscript; for example nf,0). chromatic radiation is employed, no single prisms // /
Other lines commonly employed for refractive sharp boundary such as that in Figure 13-1b '//
" ,,/' /
"" / ""NoncritK:al
index measurements include the C and F lines is observed. Instead, a diffuse chromatic region Critical /
from a hydrogen source (l = 656 nm and between the light and dark areas develops; the rays rays
486 nm, respectively) and the G line of mercury precise establishment of the critical angle is Sample
(l = 436 nm). thus impossible. This difficulty is often over-
Pressure. The refractive index of a sub- o 0
come in refractometers by the use of mono-
(a) (bl
stance increases with pressure because of the chromatic radiation. As a convenient alterna-
accompanying rise in density. The effect is FIGURE 13-1 (a) Illustration of the criti- tive, many critical angle refractometers are
most pronounced in gases, where the change cal angle, 6., and the critical ray AOC; equipped with a compensator that permits the
in n amounts to about 3 x 10-4 per atmos- (b) end-on view showing sharp boundary use of radiation from a tungsten source, but
phere; the figure is less by a factor of 10 for between the dark and light fields formed at compensates for the resulting dispersion in
liquids, and it is yet smaller for solids. Thus, the critical angle. such a way as to give a refractive index in
duction of the sample. The lower prism face that shown in Figure 5-7a (p. 124). Liquid tometers, can also be employed as a reference. Transmission .nd
is rough-ground; when light is reflected into samples are contained in a prism-shaped con- The difference between the refractive index of RefrllCtion of RBdi.tion in
the prism, this surface effectively becomes the tainer mounted at the center of a large circu- the standard and the instrument scale reading Optically Anisotropic Medi.
sowce for an infinite number of rays that pass lar metal table; solid samples are cut into the is applied as an arithmetic correction to sub-
through the sample at all angles. The radiation shape of a prism and are similarly mounted. sequent determinations. With· the Abbe re- Optically isotropic substances transmit radia-
is refracted at the interface of the sample and A light source, a slit, and a collimator are fractometer, the objective of the telescope can tion at equal velocities in all directions re-
the smooth-ground face of the upper prism, employed to direct a parallel beam of radia- be adjusted mechanicatly so that the instru- gardless of the polarization of the radiation.
whereupon it passes into the fixed telescope. tion onto one surface of the prism. The re- ment indicates the proper refractive index for Examples of isotropic material include homo-
Two Amici prisms, which can be rotated with fracted image of the slit is then viewed with a the standard. geneous gases and liquids, solids that crys-
respect to one another, serve to collect the telescope mounted on the circle. Since the slit tallize in the cubic form, and noncrystalline
divergent criti~ angle rays of different colors image can be made very sharp, the accuracy solids such as glasses and many polymers.
into a single white beam, which corresponds in of the determination depends only upon the Noncubic crystals, on the other hand, are ani-
path to that ofthe sodium D ray. The eyepiece accuracy of the angular measurements and sotropic and may transmit polarized radiation
of the telescope is equipped with crosshairs; in the control of temperature; uncertainties of Optical activity is a measure of the ability at different velocities depending upon the an-
making a measurement, the prism angle is 1 x 10-6 or smaller can be attained. An in- of certain substances to rotate plane-polar- gular relationship between the plane of polar-
changed until the light-dark interface just strument of this type is often used as a detec- ized light. This phenomenon, first reported for ization and a given axis of the crystal.
coincides with' the crosshairs. The position of tor in liquid chromatography (see Figure 25-6). quartz in 1811, has been studied intensely since Transmission of Polarized Radiation Through
the prism is then established from the fixed that time. By the mid-nineteenth century, A_ropic CrystaL Recall (p. 102) that radia-
scale (which is normally graduated in units of many of the laws relating to optical activity tion is slowed as it passes through a medium
no). Thermostating is accomplished by circu- had been formulated; these in turn played a containing atoms, ions, or molecules because
1&tion of water through jackets surrounding direct part in the development of the ideas of its electrical vector interacts momentarily with
the prisms. .
Applications of Refractometry
organic stereochemistry and structure later in the electrons ofthese particles, causing tempo-
. .The Abbe r'cfraetometer owes its popular- In common with density, melting point, and the same century. Some of the early concepts rary polarization. Reemission of the radiation
ity to its convenience, its wide range (/10 = 1.3 boiling point, the refractive index is one of the of optical activity have stood the test of time occurs after 10-14 to 10-13 s as the polarized
to 1.7), and to the minimal sample required. classical physical constants that can be used and remain essentia1ly unaltered today. It is of particles return to their original state. A beam
The accuracy of the instrument is about to describe a chemical species. While it is a interest, however, that despite this long history, of radiation traversing an isotropic medium
±0.OOO2; its precision is half this figure. The nonspecific property, few substances have the interactions of radiation with matter that encounters a symmetrical distribution of par-
most serious error in the Abbe instrument is- identical refractive indexes at a given temper- cause the rotation of polarized light are less ticles around its path of travel. Thus, the slow-
CIWlIed·by. the fact that the nearly grazing rays ature and wavelength. Thus, this constant is clearly understood than the processes respon- ing of a beam of polarized radiation is the same
~ .~t. oIfby the arrangement of the two useful for confirming the identity of a com- sible for absorption, emission, or refraction. regardless of the angle of the plane of polar-
pnsms; the boundary is thus less sharp than pound and measuring its purity. In addition, The term polarimetry, as it is used by most ization around the direction of travel. In con-
is desirable. refractive index determinations are useful for chemists, can be defined as the study of the trast, the distribution of atomic or molecular
A precision Abbe refractometer, which di- the quantitative analysis of binary mixtures. rotation of polarized light by transparent sub- particles around most paths through an aniso-
minishes the uncertainties of the ordinary in- Finatly, combined with other measurements, stances. The direction and the extent of rota- tropic crystal is not symmetrical. Thus, radia-
stJ:ument by a factor of about three, is also the refractive index provides structural and tion (the optical rotatory power) is useful for tion vibrating in one plane along this path
available; the improvement in accuracy is ob- molecular weight information about a sub- both qualitative and quantitative analysis and encounters a different particle environment
tained by replacing the compensator with a stance. for the elucidation of chemical structure as from that in another; a difference in rate of
monochromatic source and by using larger For all applications, periodic calibration of well.2 transmission results.
and more precise prism mounts. The former refractometers is necessary. Standards for this All anisotropic crystals have at least one
provides a much sharper critical boundary, purpose include purified liquids such as water axis, called the optic axis, around which there
and the latter Permits a more accurate deter- (n5° = 1.3330), toluene (nl,0 = 1.4969), and exists a symmetrical distribution of the par-
mination of the' prism position. methylcyclohexane (nl,0 = 1.4231). The latter 2 For a more complete discussion of the various aspects ticles making up the crystal. Polarized radia-
Image DispIaeement Refractometer. The two compounds can be obtained from the of polarimetry, see: Physical Methods of Chemistry. cds. tion travels along the optic axis at a constant
most straightforward method for determining A. Weissberger and B. W. Rossiter. New York: Wiley. rate regardless of its angle of polarization
National Bureau of Standards as certified
1972, vol. I. Part III Co Chapters t and 2; and w. A. with respect to that axis.
refractive index involves measurement of the samples with five-decimal indexes at 20, 25, Struck and E. C. Olson, in Treatise Oil AnalytiCllI Chem-
angles of incidence and refraction by means and 30°C and for each of seven wavelengths. Figure 13-4&depicts the electrical vectors
istry. eels. I. M. Kolthoff and P. J. Elving. New York:
of a spectrometer arrangement somewhat like A glass test piece, supplied with most refrac- Intersciencc. 1965.Part I. vol. 6. Chapter 71. at maximum amplitude for two beams of
radiation that are in phase and polarized at Figure 13-4b contrasts the behavior of the much, and the two beams become out of the extraordinary ray is clearly propagated at
90 deg to one another. The arrows are the two beams when they strike the crystal at an phase not only within the crystal but also a greater rate than the ordinary ray; in quartz,
vectors for the beam which is vibrating in the angle that is 90 deg to the optic axis. Here, after emerging from it. the reverse is the case. Because anisotropic
plane of the paper. The dots represent the the beam that is vibrating in the plane of the The polarized beam that. travels at the crystals have two characteristic refractive in-
vector that is fluctuating in a plane perpen- paper encounters'!U1 atomic or molecular en- same velocity along the optic axis and per- dexes, they can be made to refract the ordinary
dicular to the page. Figure 13-4a shows that vironment similar' to that depicted in Figure pendicular to it is called the ordinary beam; and the extraordinary rays at different angles;
the wavelengths of both beams are decreased 13-4a; its wavelength and velocity behavior the faster moving beam is the extraordinary anisotropic crystals are thus double-refracting.
equally as they travel along the optic axis of are thus similar to the two beams moving beam. (In some instances, the relative veloci- This property provides a convenient means for
the crystal; their velocities remain the same as along the optic axis. The beam vibrating in ties of the two types of beams are reversed.) separating an unpolarized beam into two
a consequence (see Figure 4-2, p. 94). Thus, the the plane perpendicular to the page, however, The velocity of the extraordinary beam varies beams that are plane-polarized at 90 deg to
two beams remain in phase both within the encounters a less dense environment; thus, its continuously as the angle of its travel with one another.
crystal and as they emerge from it. velocity and wavelength are not decreased as respect to the optic axis is varied from 0 to 90 The Nicol Prism. Figure 13-5 depicts a
deg, and reaches a maximum (or minimum) Nicol prism, a device that exploits the double-
at the latter angle. refracting properties of crystalline calcite
Transmission of Unpolarizecl Radiation (Caco3) to produce plane-polarized radiation.
1'hrolIgb Anisotropic Crystals. When a beam The end faces of a natural crystal are trimmed
of unpolarized monochromatic radiation passes slightly to give an angle of 68 deg, as shown;
through an anisotropic crystal at an angle to the crystal is then cut across its short diagonal.
the optic axis, separation into an ordinary and A layer of Canada balsam, a transparent sub-
extraordinary beam occurs. To understand stance with a refractive index intermediate
this behavior, recall that the various electrical between the two refractive indexes of calcite, is
vectors of unpolarized monochromatic radia- placed between the two crystal halves. This
tion can be resolved into two mutually per- layer is totally reflecting for the ordinary ray
pendi~ular vectors, as shown in Figure 4-10 with its greater refractive index but, as shown
(p. 105); thus, a beam of ordinary radiation in the figure, transmits the extraordinary ray
can be considered to be made up of two almost unchanged.
J.
OptIC plane-polarized beams of equal amplitude Pairs of Nicol prisms are employed in
axis whose planes are oriented at 90 deg to one measurements involving the rotation of the

I another. Therefore, Figure 13-4 applies to an


unpolarized beam as well as to two plane-
polarized beams; here, it is only necessary
plane of polarized light. One prism serves to
produce a polarized beam that is passed into
the medium under study. A second analyzer
to specify that the amplitudes of the latter prism then determines the extent of rotation
• • Vector component beams are identical. caused by the medium. If the two Nicol
~ perpendicular
__ ~""-_ to page •
Double Refraction by Anisotropic Crystal. prisms are identically oriented with respect to
From the foregoing discussion, it is evident

Vector component
in plane
7. that the velocity, and thus the index of refrac-
tion ofthe extraordinary ray, in an anisotrop-
of page ic crystal is dependent upon direction, being
identical with that for the ordinary ray along TABLE 13-1 REFRACTIVE
the optic axis and changing continuously to a INDEX DATA FOR SELECTED
maximum or a minimum along a perpendicu- ANISOTROPIC CRYSTALS
lar axis. Refractive indexes for the extra-
FIGURE 13-4 Schematic representation of passage bf a mono- ordinary ray are normally reported in terms Crystal
chromatic beam through two axes of an anisotropic crystal of calcite. of this perpendicular axis. Table 13-1 com- Calcite 1.6583 1.4864
Arrows represent the electric vector component in the plane of the page. pares the refractive index for the ordinary ray 1.544 1.553
Quartz
Dots represent the vector component in a plane perpendicular to the n. with that for the extraordinary ray n. in Ice 1.306 1.307
page. some common anisotropic crystals. In calcite,
Interference Effects along MB and MA. These components are
with Polarized Rediation also indicated in Figure 13-8. Upon ei1tering
the crystal, the component lying in the MA
To account for many of the experimental direction travels at the rate of an ordinary ray
observations regarding the interactions of p0- since this direction lies along the optic axis of
larized radiation, it is necessary to assume the crystal; the orientation of component MB,
that interference between polarized beams on the other hand, corresponds to that for an
can occur, provided the beams are coherent extraordinary ray, and its rate of propagation
Ordinary (p. 99). The effect of interference can then be is thus different. As a result of the velocity
Crystalline beam
calcite (CoC03) visualized by vector addition of the electro- difference, the two components are no longer
FIGURE 13-5 A Nicol prism for resolving an unpolarizcd magnetic components ofthe individual beams. coherent and thus cannot interfere. That is,
beam into two beams plane-polarized at 90 deg to one another. Figure 13-7a illustrates the interference be- within the crystal, the beam can be considered
tween two plane-polarized beams of equal to consist of two components having different
amplitude that are in phase but oriented 90 velocities; because of their incoherence; inter-
the beam, and if the medium has no effect, the of the analyzer. Rotation of the polarizer by deg to one another. Addition of the electrical ference cannot occur.
extraordinary ray, comprising nearly 50% of 90 deg results in a beam from the analyzer vectors of the two: beams is shown schemat- When the two rays leave the crystal, their
the original intensity, is emitted from the that has no vertical component. Thus, no ra- ically in the end-on view to the right. Note velocities again become equal in the isotropic
analyzer (see Figure 13-68). If the polarizer is diation is observed at the face of the analyzer. that the numbered points along the axis MN air medium; thus, they can again interfere
rotated (Figure 13-6b~ only the vertical com- If a medium affecting the rotation of light is have been projected onto the plane ABCD since they are once more coherent. The nature
ponent m'n' of the beam mn emitted from the interposed between the two Nicol prisms, the which is perpendicular to MN, thus giving a ofthe resultant wil~ however, depend upon the
polarizer is transmitted through the analyzer, relative orientations needed to achieve a maxi- two-dimensional representation of the result- phase relationship between the two that exists
the horizontal component being now reflected mum or a minimum in transmission are chang- ant and its compqnents. It will be seen that at the instant they emerge from the crystal
from the Canada balsam layer. Here, less than ed by an amount corresponding to the rotatory the resultant is a: vector that oscillates in a surface. This phase relationship is determined
50% of the incident beam appears at the face power of the medium. plane oriented at 45 deg to the planes of the by the relative velocities of the two rays in the
two component beams; interference thus pro- medium as well as by the length of traverse.
duces a single, plane-polarized beam. If, for example, the path in the crystal is such
Figure 13-7b demonstrates interference _ that the two rays are completely in phase
when the phase relationship between the two upon exiting, then constructive interference
plane-polarized beams differs by one-half similar to that shown in Figure 13-7a occurs.
wavelength (180 deg). Here, the waves are That is, the resultant beam will be polarized at
again in phase, and the resultant is a plane- the same angle as the entering beam. If, on
polarized beam which, however, is perpendi- the other hand, the crystal thickness is such
cular with respect to the resultant of Figure that the phase relationship between the two
13-7a. rays at the face is shifted exactly one-half
Circularly and Elliptically Polarized Radia- wavelength, interference such as shown in
tion. It is of interest now to examine the be- Figure 13-7b results. Here, the plane of the exit

<1>0<8>6
havior of a plane~polarized beam of mono- beam is oriented 90 deg with respect to the
chromatic radiatioo as it passes through an entering beam.
anisotropic crystal. In Figure 13-Sa, the path In Figure 13-8a, we have shown the emerg-
of the incident bea'll is normal to the optic axis ing waves as one-quarter wavelength out of
of the crystal, with; the plane of the polarized phase and have indicated their relationship to
0' 0' 0' 0' radiation oriented at 45 deg to that axis (the one another, assuming that no interference
angle of the plane of polarization is given by takes place. In fact, however, interaction does
1_) Ibl (el (d)
the arrow MP). As shown by Figure 13-7, occur as the two rays enter the air medium,
FIGURE 13-6 End view of a polarizer and analyzer Nicol. mn = electrical however, the plane-polarized beam can be and the path of the electrical vector for the
vector of beam transmitted by polarizer. m'n' = vertical component of beam considered to consist of two coherent compo- resulting wave can be obtained by adding the
transmitted by polarizer and analyzer. nents lying in perpendicular planes oriented two vectors. The resulting vector quantity is
seen to travel in a helical pathway around the the electrical vector lies in a single plane. A Anglo of
plano-polarizod
direction of travel (Figure 13-Sb). If the helical beam of this type is cal1ed circularly beam
vector sum is plotted in tw<Hlimensional polarized light. Note that if the two waves had (45 dog to optic axis)
form (Figure 13-8c1 a circle is obtained. This been out of phase one-quarter wavelength in
condition is in distinct contrast to the orig- the other sense, the direction of travel of the
inal, linearly polarized radiation in which vector would have been clockwise rather than

6
8
8
I 9
I

$
I 2 8
6,8
7
---
_
I 3 7 End-onView
A C~
o
6
5
o
Two dimension'- representation
of interference
VICtor
o
FIG U R E 13-8 Circular polarization of light by an anisotropic crystal.

Resultant
I
I
I counterclockwise. employed experimentally to produce circularly
6 I
7 8 I Thus far, we have considered the nature of polarized radiation. Such crystals are called
AI C the exit beam from the anisotropic crystal quarter-wave plates and find use in circular
2 3
I 4
I
I
when the phase difference created was 0, i, t, dichroism studies.
I ---- 6,8 or some multiple of these fractions. If the Relationship Between Plane Polarized. and
----- 07 light path in the crystal is such as to produce Circularly Polarized RadiatiolL In the preced-
phase differences other than these, the path ing section, we have seen that the behavior
traced by the resultant electrical vector is an of plane-polarized radiation upon passage
ellipse, and the radiation is called elliptically through an anisotropic crystal can be ratio-
polarized light. Figure 13-9 summarizes the nalized by considering the beam to be the re-
states that result when the components of a sultant of two plane-polarized rays that are in
plane-polarized beam are emitted from an phase and oriented at 90 deg to one another.
anisotropic crystal with various phase differ- It is of equal importance to understand that
ences plane-polarized radiation can also be treated
Anisotropic crystals of suitable length are as the interference product of two coherent
circular rays of equal amplitude that rotate in plane-polarized beam are different in the
opposite directions. Figure 13-10 shows how presence of such a species; that is, the refrac-
the vectors for the d and I circular compo- tive index of the substance with respect to d
nents are added to produce the equivalent vec- radiation (n4) is different from that for I (n,).
tors of a plane-polarized beam. From the Thus, optically active substances are anisotrop-
middle figure, it can be seen that each of the ic with respect to circularly polarized light
two rotating vectors describes a helical path and show circular double refraction. Note that
around the axis of travel of the beam. the two circular components are no longer
The rationalization of many phenomena to coherent in the anisotropic medium and
be considered in this chapter is based upon cannot interfere until they again reach an iso-
the idea that plane-polarized radiation con- tropic medium.
sists of a d and an I circular component. Here, The rotation of a beam of light by an op-
d (dextrorotatory) refers to the clock~ise rota- tically active medium is shown schematically
tion as the beam approaches the observer; in Figure 13-11. Initially in Figure 13-11a, the
I (levorotatory) is the counterclockwise compo- beam is polarized in a vertical plane with
nent. the circular d and I components rotating at
equal velocities. Upon entering the anisotropic
medium, the rate of propagation of the d
component is slowed more than that of the
Circu"r Double Refraction
I because "" > n,. Thus, at some location
The rotation of plane-polarized light by an (Figure 13-11b) in the medium, the d vector 7

optically active species can be exp~ined if will lag behind the I; if at this point the two FIGURE 13-10 Equivalency of plane-polarized beam to two (d,/)
one assumes that the rates of propagation of rays could interfere, the resultant would still circularly polarized beams.
the d and the I circular components of a be a plane, but one that was rotated from the

Plane-
polarized
radiation

FIGURE 13-9 Effect of an anisotropiccrys-


tal on plane-polarized radiation. Each diagram
corresponds to a different path length in the FIGURE 13-11
crystal. n, < n4'
vertical. At a further point (Figure B-llc), the It is apparent from this calculation that a It is important to appreciate that the inter- clockwise, or I rotation as the observer faces
d component would be still further retarded relatively small difference in refractive index actions which lead to rotation of polarized the beam, is given the negative sign. Clockwise
and a greater rotation would result. Figure has a large effect in terms of optical rotatory light are not peculiar to molecules with asym- (d) rotation is positive.
13-11 has been drawn to show that, upon . power. Note that if the refractive index for the metric centers but instead are characteristic of The term molecular rotation [M] is also en-
emerging from the medium, retardation has . sodium 0 line is about 1.5 (a typical value~ most molecules. However, rotation is not ob- countered; it is defined as
been such that the resultant plane is now hori- then a 2.2 ppm difference between ~ aitd n, served in noncrystalline samples that lack
zontal. As shown in Figure 13-11c, the d and is responsible for the lOQ-deg rotation shown asymmetric character because the molecules
I components are again propagated at identical in the example. are randomly oriented; rotation caused by a where M is the molecular weight.
rates in the isotropic air medium, and inter- molecule in one orientation is canceled by the It is frequently necessary to measure the
ference of the coherent rays occurs. Passage Optically Active Compounds equal and opposite rotation of another ori- optical rotatory power of a substance as a
through the anisotropic medium has had the ented as its mirror image. For the same reason, solute. Unfortunately, the specific rotation of
Optical activity is associated with two types
effect, however, of shifting the phase relation- samples containing d and I isomers in equal a compound is nearly always found to vary
of species: (1) crystalline compounds, which
ship such that the observed plane of polar- concentrations el\hibit to net rotation because with the nature of the solvent. Because of
lose their activity when the crystal is con-
ization is now 90 deg with respect to the their individual effects cancel. It is only when solubility considerations, no single standard
original. verted to a liquid, a gas, or a solution (quartz
one form of a chiral molecule is present in solvent can be designated. In addition, the
is the classic example of this manifestation of
Quantitative Relationships. It is readily excess that a net :rotation can be observed. specific rotation in a given solvent may not
optical activity); (2) compounds in which the
shown that the rotation (CXA) in degrees caused be entirely independent of concentration, al-
optical activity is inherent in the molecule
by an optically active substance is given by though variations in dilute solutions are usual-
itself and is observed regardless of the physi-
the relationship Variables that ly small. Because of these effects, it is common
cal state of the compound. It is with this
180 I second type that we are concerned. Affect Optica~Rotation practice to designate both the kind of solvent
CXA = -;.- (n, - n,,) (13-6) The structural requirements for an op- and the solute concentration in reporting a
The rotation of plane-polarized radiation by specific rotation.
tically active molecule are well understood optically active l:ompounds can range from
where I is the path length of the medium in and treated in detail in organic chemistry The variation in specific rotation with tem-
several hundred' to a few· hundredths of a perature is approximately linear, but the tem-
centimeters and ;. is the wavelength of the textbooks. It is sufficient to note that the op- degree. Experimental variables that influence perature coefficient differs widely from sub-
plane-polarized radiation (in vacuum~ also in tically active forms of a molecule, called enan- the observed rotation include the wavelength stance to substance. For example, the specific
centimeters. The quantity (n, - n,,) is termed tiomers, are mirror images which cannot be of radiation, the optical path length, the tem- rotation of tartaric acid solutions may vary as
the circu/Qr birefringence. The following calcu- superimposed, regardless of the orientation of perature, the density of the substance if un- much as 10% per degree; on the other hand,
lation demonstrates the magnitude of the cir- one with respect to the other. Enantiomers diluted, and its concentration if in solution. For the variation for sucrose is less than 0.1 %
cular birefringence required to bring about a thus bear the same relationship to one another solutions, the rotation may also vary with the per degree.4
typical rotation. as the left to the right hand. The two isomers kind of solvent. The effect of wavelength on rotation is
then rotate polarized light equally but in op- The specific rotation or the specific rotatory
posite directions; if one is in excess in a considered in a later section on optical rota-
EXAMPLE power [cx]I is widely employed to describe the tory dispersion.
A solution contained in a 1O-cm cell is mixture or is isolated from the other, rotation rotatory characteristics of a liquid. It is defined
found to rotate the plane-polarized radiation is observed. The most common form of optical as Mechanism of Optical Rotation
of the sodium 0 line by 100 deg. What is the activity results from the presence of four dif-
difference in refractive index of the medium ferent substituents on a tetrahedral carbon CX While the structural requirements for optical
,
[ex]I =-1c
for the I and d circularly polarized compo- atom in an organic compound; this feature activity can be precisely defined, the mecha-
;

nents? leads to two nonsuperimposable chiral mole- nism by which a beam of circularly polarized
where cx is the observed rotation in degrees, I
cules. Other types of asymmetric centers can radiation interacts with matter and is thus
is the path lengtl) in decimeters, and c is the
also occur, both in organic and inorganic
grams of solute [in 100 cc of solution. The
structures.3
wavelength ;. and temperature T are usually
Substituting into Equation 13-6 yields • It is important to remember that most organic solvents
specified with a subscript and a superscript,
have temperature c:oeftic:ients of expansion of about 0.1 %
180 deg x 10 em
100 deg = 5.89 x 10-5 em (n, - n,,)
as shown. Most specific rotations are measured
at 20·C with sodium 0 line and are thus
r.
per dell C. Therefore. accurate measurements of [II for
• For a discussion of opti<:aJ activity in inorganic systems, solutions require close control of temperature both
see: J. C. Bailar and D. H. Busch, TM Chemistry of 1M reported as [ex]~t For a pure liquid, c is re- during the measurement and during 1M solution prepora-
(n, - n,,) = 3.3 x 10-6 Coordination Compounds. New York: Reinhold. 1956. placed by its density. By convention, counter- lion.
retarded is much less obvious. That is, we can employed. Historically, the sodium D line was HaIf-Sbldow Derices. Figure 13-12 shows or ~fractive index. Optical activity is charac-
predict with certainty that a compound such obtained by introducing a sodium salt into a a typical half-shadow device consisting of a teristic of many naturally occurririg substances
as 2-iodobutane is chiral because it possesses gas flame. Suitable filters then removed other small Nicol prism (called a Ljppich prism) such as amino acids, steroids, alIcaloids, and
an asymmetric center. On the other hand, it is lines and background radiation. Sodium vapor that intercepts about half the beam emerging carbohydrates; polarin1etry represents a valu-
difficult to account for the specific rotations lamps with a filter to remove all but the D line from the polarizer. The position of the Lip- able tool for identifying such compounds.
([a]f.° = ±32 deg) for these isomers, com- are now employed. Mercury vapor lamps are pich prism is adjusted to alter the plane of StructIiraI Determination. In this applica-
pared with those of2-butanol ([a]f.° = ±13.5 also useful, the line at 546 nm being isolated polarization by a few degrees; thus, in the tion, the change in optical rotation resulting
deg). by a suitable filter system. absence of sample and with the analyzer Nicol from a chemical transformation is measured.
Several theories concerning the mechanism Polarizer and Analyzer. Nicol prisms are prism at 90 deg with respect to the polarizer, Empirical correlations obtained from the study
of optical rotation have been developed. Gen- most commonly employed to produce plane- a split, light-dark field is observed. The light of known structures are then employed to
erally, these are couched in terms of quantum polarized light and to determine the angle portion, of course, corresponds to that half of deduce information about the unknown com-
mechanics and are of sufficient complexity to through which the light has been rotated by the beam that has been rotated by the Lippich pound. Details of steroid structures, in parti-
evade al1but the most mathematically oriented the sample. In principle, the measurement prism and the dark part of the field corres- cular, have been acquired from polarimetric
chemist. Furthermore, while in principle it could be made by first adjusting the two ponds to the unobstructed beam. The intensity measurements; similar information has been
may be possible to predict optical rotatory prisms to a crossed position that yields a of the two halves is then balanced by rotation obtained for carbohydrates, amino acids, and
power for certain types of compounds with minimum in light intensity in the absence of of the analyzer. The analyzer scale is adjusted other organic compounds.
the aid of these theories, such calculations are sample. With a sample in place, rotation of to read zero at this point. With the sample Qantitative Analysis. Polarimetric meas-
not practical in terms of effort; nor has it been the beam would cause an increase in light in place, the analyzer is rotated until the same urements are readily adapted to the quantita-
demonstrated that the values obtained from intensity which could then be offset by rota- balance is obtained. The rotation of the sample tive analysis of optically active compounds.
such calculations can be sufficiently precise to tion ofthe analyzer prism. The angular change can then be read directly from the circular Empirical calibration curves are used to
be useful. required to minimize the intensity would cor- analyzer scale. relate optical rotation to concentration; these
respond to the rotatory power of the sample. Other end point devices, which operate plots may be linear, parabolic, or hyperbolic.
Polarimete,. Unfortunately, however, the position of mini- upon the same principle as the Lippich prism, The most extensive use of optical rotation
The basic components of a polarimeter in- mum intensity cannot be determined accu- permit the determination of optical rotatory for quantitative analysis is in the sugar in-
clude a monochromatic light source, a polar- rately with the eye (nor with a photoelectric power with a precision of 0.005 to om deg dustry. For example, if the only optically
izing prism, a sample tube, an analyzer prism detector, for that matter) because the rate of under ideal conditions. Photoelectric detec- active constituent is sucrose, its concentration
with a circular scale, and a detector (see change in intensity per degree of rotation is at tors yield a precision of about 0.001 deg. can be determined from a simple polarimetric
Figure 13-12). The eye serves as the detector a minimum in this region. Therefore, polari- Sample Tubes. The sample for polarimetry analysis of an aqueous solution of the sample.
for most polarimeters although photoelectric meters are equipped with half-shadow devices is contained in a cylindrical tube, usually 5, The concentration is directly proportional to
polarimeters are becoming more common. which permit the determination to be made 10, or 20 em in length. The ends are plane- the measured rotation. If other o{)tically active
Sources. Because optical rotation varies by matching two halves of a field at a radiation parallel glass disks that are either fused to the materials are present, a more complex pro-
with wavelength, monochromatic radiation is intensity greater than the minimum. tube walls or held in place with screw-cap cedure is required; here, the change in rotation
holders. For precise measurements, the tubes resulting from the hydrolysis of the sucrose is
are surrounded by a jacket for temperature determined. The basis for this analysis is shown
control. Tubes can be calibrated for length by by the equation
measuring the rotation of a liquid of known ·.cid
rotatory power; nicotine/alcohol or sucrose/ C12H210ll + H20 --0 C6H1206 + C6HI206
water mixtures are often used for this purpose.
Sucrose Glucose Fructose

Applications of Polarimetry
Qualitative Analysis. The optical rotation This reaction is termed an inversion because
of a pure compound under a specified set of of the change in sign of the rotation that occurs.
conditions provides a basic physical constant The concentration of sucrose is directly pro-
that is useful for identification purposes in the portional to the difference in rotation before
same way as its melting point, boiling point, and after inversion.
OPTICAL MISCELLANEOUS OPTICAL METHODS

ROTATORY DISPERSION
AND CIRCULAR DICHROISM Optical Rotatory Di.peralon Curv••
I component of
.
An optical rotatory dISpersIon . curve fconsists
as a plane-polarized
Optical rotatory dispersion and circular di- beam
Of a plot of specific or molecular rota Ion
chroism are two closely related physical meth-
ods that are based upon the interaction of function of wavelengt.h T wO'.types of curva-nnal
be discerned The first IS the no
circulatory polarized radiation with an optical- ture can . . [] ban es only
dispersion range wherem lX c g th Amplitude
ly active species. $ The former method measures gradually with wavelength. !he second, ofthed
e :omponent
the wavelength dependence of the molecular
rotation of a compound. As we have already region of anoma Ious dispersIOn, occurs near
. . Iated
indicated, optical rotation at any wavelength an absorption peak. If one peak ;:; dis-
depends upon the difference in refractive index from others, th~ anomalous part ranee of the
rsion curve will have the appca .
of a Substance toward d and I circularly POlar-
pe labeled (n - n,,) in Figure 13-13. That IS,
ized radiation-that is, upon the circular bire- curve I id ban e to some I More strongly
fringence (n, - ~); this quantity is fOund to the ~otatio(n Und~~~~ue~ alte~ direction : absorbed d component
vary in a Characteristic way as a function of maximum or m . ) d then finally I
wavelength. In contrast, circular dichroism to a minimum (or maXImum. ' an I dis. I Plane of
: FIGURE 13-13 The Cotton ef- reverts to values correspondmg to norma Ioriginal beam
depends upon thefact that the molar absorpti_ . fect. (Adapted from W. Heller and I
.vityofan optically active compound is different H. G. Curme in Physical Methods FIGURE 13-15 Elliptical!y ~lari~
for the two types of circularly polarized radi- ofChem~try, cds. A. Weiss berger radiation after leaving a medIUm In whIch
ation. Here, the wavelength dependence of and B. W. ROSSiter. New Yorlc: e,,>e,andn,>~,
(e, - e,,) is studied, where £, and e" are the ;WileY-Interscience, 1972, Vol. I,
respective molar absorptivities. ;Part III C, p. 66. With permission.)
I
The inequality of molar absorptivities was 0.024
first reported by A. Cotton in 1895, and the perslon.
. As I'nd;~~ted
~ in Figure 13-13, a change
whole complex relationship between absorp-
o~ in the sign of the rotation may accompany
.~:g
_ .9-
tivity and refractive index differences is now 0.012
termed the Cotton effect. The application of ei the:; c~~::~ies have multiple abs~rption
the Cotton effect to chemical investigations At the same time, Cotton observed that dra- ~~
::Oil as is usually the case, overlapping. re-
was delayed until convenient instruments for matic changes OCCurredin the optical rotation je; 0 ~:::'Of anomalous dispersion lead to OPt;cal
~E
its study became aVailable. (n, - n,,1 as well as in the difference in the d' rsion curves that are comp ex,
absorptivities (e, - e,,) of the beams in the rotatory ~pe. 13 14 Note that the
as shown In FIgure -. _
region of an absorption maximum. These -0.012
ultraviolet absorption spectrum for the com
General Principtes 4
effects are shown in igure 13-13. It is imPOrtant pound is also included for reference.
to appreciate that the (n, - n,,) curve for a
Cotton emplOYedsolutions of POtassium chro- .•. 3
Substance is similar in shape to curves index Circular Dichroi8nt Curves
mium tartrate, which absorb in the visible as a function of the wavelength of unpolarized l 2
region, to demonstrate that right circularly light (see Figure 4-8, p. 102); here, too, marlced "
In circular dIchroISm, one I of'zed
the beam
circularis
polarized radiation was not only refracted but nents of a plane-po an
changes in the refractive index (anomalous compo gI bsorbed than the other. The
also absorbed to a different extent than the dispersion) OCCurin the region of absorption. more stron y a . . t n-
left cirCUlarly polarized beam; that is, e" oF £,. effect of this differential abs?r~tlon IS 0 COil'
I ized
ar radlahon to an e Ip-
vert the Pla?e-pobeam F' re 13-15 illustrates
tically polarized . Igu ual
ireular components of uneq
• For a more c:onipJetediscussion of optical rotatory how . two c hich result from the diftieren tial
dispenion and ciJcular dichroism, see: K. P. Wong, J. amplitude, w . are combined to
Chem. £due. 52, A9, AS3, A573 (1975); C. DjcrassL absorption by a mecbum,. lI'ptical
give a resultant that travels m ~\:.' beam
th The I component of the OrIS'
pa.
is shown as retarded more than the d compo-
nent because n, > ""; on the other hand, the Iyzer and is detected with a photomultiplier must be provided. We have noted (p. 361) depends upon the refractive index of the
amplitude of the d component is less than that tube. The signal from the detector is amplified; that circularly polarized radiation can be ob- medium, the angle of incidence of the reflected
of the I component because we have assumed it is then employed to adjust the analyzer tained by passing plane-polarized radiation beam, and the number of reflections. Quarter-
that its molar absorptivity is greater; that is, position to compensate for rotation caused by through an anisotropic crystal which has a . wave retardation, and hence circular polar-
e,,>/lt. the sample and to position a recorder pen as thickness such that the extraordinary and ization, can be attained through proper aiijust-
The angle of rotation IX is taken as the well. As with a visual polarimeter, the half- ordinary rays are one-quarter wavelength out ment of these variables.
angle between the major axis of the emergent shade method is the most efficient way of of phase. Rotation of the optic axis of the The unit shown in Figure 13-16 is fitted
elliptical beam and the plane of polarization determining the null position of the analyzer. quarter-wave plate by 90 deg yields either d into the sample compartment of a double-
of the incident beam. The ellipticity is given In one instrument, the polarizer is mechanical- or I circularly polarized radiation. For the beam spectrophotometer, with a similar unit
by the angle 6; the tangent of 6 is clearly ly rocked through a small angle at low fre- measurement of circular dichroism with a in the reference beam. The units are so ad-
equal to the ratio of the minor axis of the quency. The amplifier system of the detector single-beam spectrophotometer, a polarizer justed that d radiation is emitted from one
elliptical path to the major (that is,OB/OA). responds to the resulting ac signal and adjusts followed by a quarter-wave plate is inserted and I radiation from the other. The two beams
It can be shown that the ellipticity is ap- the analyzer until the signal is symmetric in the cell compartment of the instrument, then pass through identical cells containing
proximated by the relation around the null point. Another spectropolari- provision being made to allow rotation of the the sample and their relative powers are com-
meter employs an ordinary double-beam spec- plate by ± 45 deg. The sample cell is then pared photometrically.
6 ••• i{k, - k~) (13-9) trometer with two sets of polarizer-analyzer placed between the plate and the detector;
where k, and k~arc the absorption coefficients6 prisms. The two analyzers arc offset from one with the plate set to produce d circular radia-
of the circularly polarized I and d radiation another by a few degrees. and both beams are tion, the instrument is set on 100% transmit- Applications of
and 6 is expressed in radians. The quantity passed through the sample. The ratio of the tance, or zero absorbance. The plate is then Optical Rotatory
(k, - k,,) is termed the circular dichroism. The power of the two beams is compared electron- rotated 90 deg; the new absorbance reading Dispersion and Circular Dichroism
molecular ellipticity [6] can be shown to be ically and gives a measure of the optical then corresponds to log (P~/P,) for the sample.
equal to rotation of the sample. In order to cover a very wide spectral range, Optical rotatory dispersion and circular di-
Circular Dichroism. A conventional spec- several plates of different thicknesses must be chroism studies often provide spectral details
[6] •••3305(61- 6~) (13-10) trometer can be adapted to measure molecu- employed. for optically active compounds that are absent
where [6] has the units of degree-cm2 per lar ellipticity. From Beer's law, Equation Other methods exist for producing cir- in their ultraviolet spectra. Thus, in the lower
decimole and the e's arc the molar absorp- 13-10 can be written in the form cularly polarized radiation. One of these, a plot in Figure 13-14, the absorption spectrum
tivities of the respective circular components. Fresnel rhomb, is incorporated as an adapter is seen to consist of a group of overlapping
Circular dichroism curves consist of a plot "]
[u 3305(1
•••--og-- P,o to g- P~o) for a double-beam spectrophotometer; see peaks that would be difficult to interpret. On
of [6] versus wavelength. Note that [6] can be be P, P~ Figure 13-16. When a polarized beam under- the other hand, the molecular rotation and
either negative or positive depending on the where Pro and P, represent the power of the goes internal reflection in this device, one of ellipticity curves for the optically active groups
relative magnitudes of 6, and e". The dashed circularly polarized I beam before and after the perpendicular components is retarded are much more clearly defined and lend them-
line in Figure 13-14 is a typical curve. it has passed through a solution of length b with respect to the other. The retardation selves to detailed analysis.
and containing a molar concentration c of
Instn-mentation the sample. The terms P~ and P~o have equiva-
Optical Rotatory DispersiolL A number of lent meanings for the d radiation. If now
recording spectropolarimeters are now manu- P~o'" Pro, then
factured that directly provide optical rota- [6] •••~ log P~ (13-11) Beam from
tory dispersion curves in the ultraviolet and be P, spectro-
visible regions. In these instruments, radiation photometer
,
Thus, the molecular ellipticity can be ob- l,
from a conventional monochromator is passed
tained directly by comparing the power of the
thr~ugh a polarizer, the sample, and an ana-
two transmitted beams, provided the intensi-
:
I

'------ -
I

ties of the incident I and d circularly polarized


beams arc identical
• The absorption coefficient k is related to the more
common molar absorptivity £ by the equation In order to employ Equation 13-11 with an
k ~ 2.303u, where c is the cooc:cntration in moles per ordinary spectrophotometer, a device for pro- FIGURE 13-16 Spectrophotometer adapter for pro-
liter. ducing d and I circularly polarized radiation duction of circularly polarized radiation.
Optical rotatory dispersion curves have ketone groups, conformational analysis of sub- (c) Using nj,0 = 1.4664 and M = 153.82, calculate the density of
been mainly applied to structural detennina-- stituents exerting a vicinal action on an optical- CC4. Compare your calculated value to the observed value of
tions in two major areas: (1) amino acids, ly active chromophore, the degree of coiling 1.5942 g/ml.
polypeptides, and proteins; and (2) complex of protein helices, and the type of substitution 4. What is the difference between unpolarized, plane polarized, circularly
natural products such as steroids, terpenes, in amino acids. potarized, and elliptically polarized light?
and antibiotics. Most of the structural conclu- The applications of circular dichroism are
5. Distinguish between ordinary and extraordinary beams of radiation.
sions from this work are empirical, being less developed than optical rotatory disper-
based upon spectral observations- of known sion; it appears, however, that the technique 6. Define the following terms: (a) optic axis, (b) Nicol prism, (c) quar-
structures. The curves can provide informa- will also provide much useful 'structural infor- ter wave plate, (d) double refraction, (e) circular double refraction, (f)
tion concerning the configuration of angular mation regarding organic and biological sys- circular birefringence, (g) ellipticity, (h) enantiomers, (i) dextrorota-
substituents at ring junctures, the location of tems as well as metal-ligand complexes. tory, and (j) levorotatory.
7. What is the difference between optical rotatory dispersion and circu-
lar dichroism?
8. Describe two methods of producing circularly polarized radiation.
1. In binary mixtures of liquids that approach ideal behavior, the refrac- 9. How can it be determined whether a sample is rotating the plane of
tive index varies linearly with volume percent. Mesitylene and pen- polarization of light by + 20 deg or -160 deg?
tane exhibit this behavior. For mesitylene, rt° = 1.4998, the molecular 10. If the circular birefringence of a solution at 360 nm is 1.0 x 10-5,
weight is 12020, and the density is 0.8642 g/ml (all at 20°C). For what will be the angle of rotation, «360' for a 5-cm path length? If the
pentane, rt° = 1.3579, the molecular weight is 72.15 and the density is circular birefringence is 1.0 x 10-5 at 720 nm, what will be «720 for a
0.6262 g/ml. Find the molarity of pentane in a mixture of pentane and 5-an path length?
mesitylene having a refractive index of 1.4522. : 11. For sucrose, C12HzzOu, [«]60 = +66.5 deg.
2. Binary liquid mixtures that do not form ideal solutions exhibit refrac- (a) Calculate the molecular rotation of sucrose.
tive indexes which are complicated functions of the composition, One (b) Calculate the angle of rotation expected for a solution containing
of the simplest functions applies to an acetone/carbon tetrachloride sucrose at a concentration of 5.0 g/Iiter in a 1O-an cell.
solution whose refractive index is given by rt° = DC + b, where c is (c) A solution originally containing 5.0 g of sucrose per liter exhibits
the molar concentration of acetone and a and b are constants. For «= + 18.6 deg in a 1O-cm cell after a period of heating with acid
acetone, nj,0 = 1.3588, the molecular weight is 58.08 and the density is (reaction on page 369). Calculate the fraction of sucrose which
0.7908 g/ml (all at WOe). For carbon tetrachloride, rt° = 1.4664, the has been hydrolyzed.
molecular weight is 153.82 and the density is 1.5942 g/ml. Calculate
the refractive in4ex of 100M acetone in CCl4•
3. The molar refraction, R, of a compound is defined as

R=(::~~)~
where n is the index of refraction, M is the molecular weight, and d is
the density. To a good approximation, the molar refraction of a com-
pound is the sum of refractions contributed by ea!=h atom in the
compound. Tables of atomic contributions to molar refraction can be
found in such reference books as the Handbook of Chemistry and
Physics. For example, carbon contributes a refr~ion of 2.591,
hydrogen contributes 1.028, oxygen (doubly bonded) contributes
2.122, and chlorine contributes 5.844.
(a) Calculate the molar refraction of acetone from the data in Prob-
lem 2 and compare it to the value calculated from atomic
contributions.
(b) Calculate the molar refraction of CCl4 from atomic contributions.
A strong magnetic field causes the energies of sequence, exposure to a magnetic field would
certain nuclei to be split into two or more lead to splitting of their energy levels. During
quantized levels, owing to the magnetic prop- the next decade, experimental verification of
erties of these particles. Transitions among these postulates was obtained. It was not until
the resulting magnetically 'induced energy 1946, however, that Bloch at Stanford and
levels can be brought about by the absorption Purcell at Harvard, working independently,
of electromagnetic radiation of suitable fre- were able to demonstrate that nuclei absorbed
quency,just as electronic transitions are caused electromagnetic radiation in a strong magnetic
by the absorption of ultraviolet or visible field as a consequence of the energy level
radiation. The energy differences between mag- splitting induced by the magnetic force. The
netic quantum levels for atomic nuclei co~- two physicists shared the 1952 Nobel prize for
spond to radiation energies in the frequency their work. In the first five years following the
rangeofO.l to l00MHzl (wavelengths between discovery of nuclear magnetic resonance,
3000 and 3 m), which is in the radio-frequency chemists became aware that molecular en-
portion of the electromagnetic spectrum (see vironment influences the absorption by a
Figure 4-7, p. 101). nucleus in a magnetic field and that this effect
The study of absorption of radio-frequency can be correlated witli molecular structure.
radiation by nuclei is called nuclear magnetic Since then, the growth of NMR spectroscopy
resonance2 (often abbreviated NMR or nmr); has been explosive, and the technique has had
it has pr.oved to be one of the most powerful profound effect on the development of organic,
tools available for determining the structUre inorganic, and biochemistry. It is doubtful that
of both organic and inorganic species. i there has ever been as short a delay between
This chapter is mainly concerned with ~he an initial discovery and its widespread applica-
theory, instrumentation, and applications: of tion and acceptance.
NMR spectroscopy; a somewhat analogous In common with optical spectroscopy, both
method called electron spin resonance is classical and quantum mechanics are useful to
briefly discussed at the end. provide an explanatiQn of the nuclear mag-
netic resonance phenomenon. The two treat-
THEORY OF NUCLEAR ments yield identical relationships. Quantum
MAGNETIC RESONANCE mechanics, however, is more useful for relating
absorption frequencies to energy states of
As early as 1924, Pauli suggested that certain molecules, while classical mechanics is more
atomic nuclei might have the properties of helpful in providing a physical picture of the
spin and magnetic moment and that as a con- absorption process and how it is measured.

I MHz ~ 10' Hz _ 10' cycles per second.


2 The following references are recommended for at'di-
tiona! study: E. D. Becker. High Resolution NMR. New
To account for some properties of nuclei, it is
York: Academic Press. 1969; A. Ault and G. O. Dudek.
An Introduction to Proton Magnetic Resonance Spectros- necessary to assume that they rotate about an
copy. San Francisco: Holden-Day. 1976; L. M. Jackman axis and thus have the property of spin.
and S. Stembal1, Nuclear Magnetic Resonance Spectros- Furthermore, it is necessary to postulate that
• copy. 2d ed. New York: Perpmon Press, 1969; D. E. the angular momentum associated with the
Leyden and R. H. Cox, Analytical Applicalions of NMR.
spin of the particle is an integral or a half-
New York: Wiley. 1977; and D. W. Mathieson, Nuclear
Magnetil: Resonance for Organic Chemists. New York: integral multiple of h/21C, where h is Planck's
Academic Press. 1967. constant. The maximum spin component for
a particular nucleus is its spin quantum number the potential energy of a magnet in a field is No Applied
1; it is found that a nucleus will then have given by th~ relationship field field
(21 + 1) discrete states. The component of Hb;:0 ~--! Ho

angular momentum for these states in any


chosen direction will have values of 1, 1 - 1, ",
1- 2, ... , -1. In the absence of an external where p'z is the component of magnetic moment /-! -- "
field, the various states have identical energies. in the direction of an external field having a m - +!.-! Ho
The spin number for the proton is t; strength of Ho.
".17"
thus two spin states exist, corresponding to
1 = +t and 1 = -l Heavier nucle~ being
assemblages of various elementary particles,
have spin numbers that range from zero (no
net spin component) to at least !. As shown
The quantum character of nuclei limits the
number of possible energy levels to a few.
Thus, for a particle with a spin number of 1
and a magnetic quantum number of m, the
energy of a quantum level is given by
Ho
r m-+!

in Table 14-1, the spin number of a nucleus is


related to the relative number of protons and E= -7 fJHo (14-3)
,.L•. -,
neutrons it contains.
Mapetie Properties of EIemeatary Parti- ~~"
cles. Since a nucleus -bears a charge, its spin where H 0 is the strength of the external field
Ho
gives rise to a magnetic field that is analogous in gauss (0) and fJ is a constant called the
to the field produced when electricity flows
through a coil of wire. The resulting magnetic
nuclear magneton (5.051 x 10-24 erg 0-1); p.
is the magnetic moment of the particle ex-
pressed in units of nuclear magnetons. The
/-1
m-I.O.-I
",_J-=o
I ."
dipole p. is oriented along the axis of spin and
has a value that is characteristic for each type value of p. for the proton is 2.7927 nuclear
of nucleus. magnetons. Ho
The interrelation between particle spin and Turning now to the proton, for which
1= t, we see from Equation 14-1 that this ".17"
magnetic moment leads to a set of observable
magnetic quantum states m given by particle has magnetic quantum numbers of
+t and -to The energies of these states in a
magnetic field (Equation 14-3) take the fol-
lowing values:
r m-+I

FIGURE 14-1 Orientations of magnetic moments and


Energy Levels in a Magnetic Field. When energy levels for nuclei in a magnetic field H 0 •
__ t(p.fJHo)
brought into the influence of an external mag- E
netic field, a particle possessing a magnetic
- t
moment tends to become oriented such that = -p.fJHo
its magneticdipole-and hence its spin axis-is
parallel to the field. The behavior of the particle E = _ -i(p.fJHo) TABLE 14-1 SPIN QUANTUM NUMBER FOR
is somewhat like that of a small bar magnet t VARIOUS NUCLEI
when introduced into such a field, the potential
= +p.fJHo Number of Number of Spin Quantum
energy of either depending upon the orienta- Protons Neutrons Number I Examples
tion of the dipole with respect to the field. These two quantum energies correspond to
The energy of the bar magnet can assume an Even Even 0 12e, 160. 32S
the two possible orientations of the spin axis
infinite number of values depending upon its Odd Even t IH, 19F, 31p
with respect to the magnetic field; as shown
alignment; in contrast, however, the energy in Figure 14-1, for the lower energy state ! 118,798r
of the nucleus is limited to (21 + 1) discrete Even Odd t 13e
(m = t) the vector of the magnetic moment is
values (that is, the alignment is limited to aligned with the field, and for the higher energy ! 127
1
2H, 14N
21 + 1 positions). Whether quantized or not, state (m = -t) the alignment is reversed. The Odd Odd 1
Applied
energy difference between the two levels is semblage of protons will contain an identical field
and the angular momentum of a rotating
given by number of nuclei with m = +t and m = -t. Ho particle; that is,
When placed in a field, however, the nuclei +I

1 "
p.P
tiE = 2p.PHo
Also shown in Figure 14-1 are the orienta-
tend to orient themselves so that the lower
energy state (m = +t) predominates. Beca~
Prec•••ion.t,.,......
orbit (
---~---
1- \
I '" y = l(h/21t) (14-7)
tions and· energy levels for a nucleus such as thermal energies at room temperatures are \ T --- The magnetogyric ratio has a characteristic
'.... I _/
14N, which has a spin number of 1. Here,
three energy levels (m = I, 0, and - 1) are
several orders of magnitude greater than these
magnetic energy differences, thermal agitation
---r-- Magnetic
dipole value for each type of nucleus.
II" Equation 14-6 can be converted to a
found, and the difference in energy between tends to offset the magnetic effects, and only a \ frequency of precession vo (the Larmor fre-
I
each is p.PHo. In genera~ the energy differ- minute excess « 10 ppm) of nuclei persists in I quency) by division by 21t. Thus,
ences are given by the lower-energy _s~~. The success of nuclear I
magnetic resonance, however, depends upon 18 (l)o yHo
Ho vo = 21t = 21t (14-8)
tiE = p.Py (14-4) this:slight excess. If the number of protons in
the 'two states were identical, the probability Equations 14-8 and 14-7 can also be
As with other types of quantum states, ex- for absorption of radiation would be the same combined to give
citation to a higher nuclear magnetic quan- as the probability for reemission by particles
p.P .
I tum level can be brought about by absorption
of a photon with energy hv that just equals
passing from the higher energy state to the
lower; under these circumstances, the net ab-
hvo = 1Ho (14-9)
---- tiE. Thus, Equation 14-4 can be written as sorption would be nil.
A comparison of Equations 14-9 and 14-5
. Ho FIGURE 14-2 Precession of a ro- suggests that the precessional frequency of the
hv = p.Py (14-5) particle derived from classical mechanics is
tating particle in a magnetic field.
Cla"ical Description of NMR identical to the quantum mechanical frequency
of radiant energy required to bring about the
To understand, the absorption process, and in
EXAMPLE transition of a rotating particle from one spin
particular the measurement of absorption, a
Many NMR instruments employ a magnet state to another; that is, Vo = v. Substituting
more classical picture of the behavior of a
that provides a field strength of 14,092 G. At this equality in Equation 14-8 gives a useful
charged particle in a magnetic field is helpful. gyroscope when it is displaeed from the verti-
what frequency would the proton nucleus Precession of Particles in a Field. Let us relationship between the frequency of absorbed
cal by application of a force. radiation and the strength of the magnetic
absorb in such a field'1 first consider the behavior of a nonrotating From classical mechanics, it is known that
Substituting into Equation 14-5 we find magnetic body, such as a compass needle, in field:
the angular veloc;ity of precession is directly
an external magnetic field. If momentarily proportional to the applied force and in- yHo
2.7927 x 5.051 x 10-24 x 14,092 v=-
displaced from alignment with the field, the versely proportional to the angular momen- 21t
v= 6.6256 X 10-27 x t needle will swing in a plane about its pivot, as tum of the spinning body to which the force is
a consequence of the force exerted by the field Absorption Process. The potential energy
= 60.0 X 106 Hz applied. The force on a spinning nucleus in a
on its two ends; in the absence of friction, the magnetic field is the product of the field
E of the precessing particle shown in Figure
= 60.0 MHz ends of the needle will fluctuate back and strength H 0 and the magnetic moment p.P of
14-2 is given by
forth indefinitely about the axis of the field. A the particle, or p.pHo; as noted earlier, the E = -p..Ho = -p.Ho cos (J
The foregoing example reveals that radio- quit~ different motion occurs, however, if the angular momentum is given by l(h/21t).
frequency radiation of 60.0 MHz will bring magnet is spinning rapidly around its north- Therefore, the precessional velocity is Thus, when radio-frequency energy is ab-
about a change in alignment of the magnetic soutp axis. Because of the gyroscopic effect, sorbed by a nucleus, its angle of precession
[, moment of the proton from a direction that the force applied by the field to the axis of 21tp.P must change. Hence, we imagine that absorp-
.-'- parallels the field to a direction that opposes rotation causes movement not in the plane of (l)o = IiI . Ho = yHo (14-6) tion involves a flipping of the magnetic
it. the force but perpendicular to this plane; the moment that is oriented in the field direction
Distn1lution of Particles Between Magnetic axis of the rotating particle, therefore, moves where y is a constant called the mDgnetogyric to a state in which the moment is in the
QuanhBD States. In the absence of a magnetic in a circular path (or precesses) around the ratio (or, less appropriately, the gyromagnetic opposite direction. The process is pictured in
field, the energies of the magnetic quantum magnetic field. This motion, illustrated in ratio). The magnetogyric ratio expresses the Figure 14-3. In order for the dipole to flip,
states are identical. Consequently, a large as- Figure 14-2, is similar to the motion of a relationship between the magnetic moment there must be present a magnetic force at
NMR signal; to some extent, these rates are the precessional frequency of the magnetic
. subject to experimental control because they nuclei of interest. These vibrationally and
I depend upon the physical state of the sample. rotationally developed components interact
,_/Ill,
I
~ It is clear that relaxation processes are with and convert nuclei from a higher to a
Precessing Ho
Circularly
panicle
needed for the observation of Ii steady NMR lower spin state; the absorbed energy then
polarized Precessing
radiation panicle absorption signal. Recall that such a signal simply increases the amplitude of the thermal
depends upon the few parts-per-million excess vibrations or rotations. This change corre-
of lower energy nuclei that exist in a strong sponds to a minuscule temperature rise for the
magnetic field. Since absorption depletes this sample.
FIGURE 14-3 Model for the absorption of radiation by a
excess, the signal would rapidly fall to zero if Spin-lattice relaxation is a first-order
precessing particle.
additional low energy particles: were not process that can be characterized by a time
being produced at a sufficient rate by some T., which is a measure of the average lifetime
radiationless energy-transfer processes. These of the nuclei in the higher energy state. In
processes must also be responsible for the addition to depending upon the magnetogyric
right angles to the fixed field and one with a As shown in Figure 13-10 (p. 365), plane-
creation of the slight excess of lower energy ratio of the absorbing nuclei, T1 is strongly
circular component that can move in phase polarized radiation can be considered to be
particles when the sample is first introduced affected by the mobility of the lattice. In crys-
with the precessing dipole. Circularly p0- composed of two circularly polarized beams
into the magnetic field. talline solids and viscous liquids, where mo-
larized radiation (p.361 )ofasuitable frequency rotating in opposite directions, in phase, and
_ To produce a readily detectable absorption bilities are low, T. is large. As the mobility
has these necessary properties; that is, its in a plane perpendicular to the plane of linear
signal, then, the relaxation process should be increases (at higher temperatures, for exam-
magnetic vector has a circular component, as polarization. Thus, by irradiating nuclear par-
as rapid as possible; that is, the lifetime of ple), the vibrational and rotational frequen-
represented by the dotted line in Figure 14-3. ticles with a beam polarized at 90 deg to the
the excited state should be small.: A second cies increase, thus enhancing the probability
If the rotational frequency of the magnetic direction of the fixed magnetic field, circularly
factor-the inverse relationship bcitween the for existence of a magnetic fluctuation of the
vector of the radiation is the same as the polarized radiation is introduced in the proper
lifetime of an excited state and the width of its proper magnitude for a relaxation transition;
precessing frequency, absorption and flipping plane for absorption. Only that component of
absorption line-negates the advantage of T1 becomes shorter as a consequence. At very
can occur. The process is reversible, and the the beam that rotates in the precessional direc-
very short lifetimes. Thus, when relaxation high mobilities, on the other hand, the fluctu-
excited particle can thus return to the ground tion is absorbed; the other half of the beam,
rates are high, or the lifetimes low, line broad- ation frequencies are further increased and
state by reemission of the radiation. being out of phase, passes through the sample
ening is observed, which prevents high-resolu- spread over such a broad range that the prob-
unchanged. The process is depicted in Figure
tion measurements. These two opposing factors ability of a suitable frequency for a spin-
14-4.
,...... cause the optimum half-life for an excited lattice transition again decreases. The result is
/ \ species to range from perhaps 0.1 to 1 s. a minimum in the relationship between T1

Ii
I
\
I
I

_//
f
I
\
I
Beam
after
absorption
Relaxation Processes
We must now consider mechanisms by which
Two types of nuclear relaxation are recog-
nized. The first is called longitudinal or spin-
lattice relaxation; the second is termed trans-
and lattice mobility.
The spin-lattice relaxation time is greatly
shortened in the presence of an element with
a nucleus in an upper energy or excited spin verse or spin-spin relaxation. an unpaired electron which, because of its
"7 state can return to its lower energy state. One
obvious path would involve emission of ra-
Spin-Lattice RelaxatiolL The absorbing nu-
clei in an NMR experiment are part of the
spin, creates strong fluctuating magnetic fields.
A similar effect is caused by nuclei that have
diation of a frequency corresponding to the larger assemblage of atoms that constitutes spin numbers greater than one-half. These
+ energy difference between the states. Radia- the sample. The entire assemblage: is termed particles are characterized by a nonsym-
I
\ tion theory, however, predicts a low proba- the lattice, regardless of whether the sample is metrical charge distribution; their rotation

y
/J' bility of occurrence for this process; it is
necessary, therefore, to postulate radiationless
pathways by which energy is lost by nuclei
higher spin states. The various mechanisms
for radiationless energy transfers are termed
in
a solid, a liquid, or a gas. In the ~atter two
states particularly, the various nucl~i compris-
ing the lattice are in violent vibrational and
rotational motion, which creates a complex
field about each magnetic nucleus. The result-
also produces a strong fluctuating field that
provides yet another pathway for an excited
nucleus to give up its energy to the lattice.
The marked shortening of T. causes line
broadening in the presence of such species. An
FIGURE 14-4 Absorption of one circular nuclear relaxation processes. ing lattice field thus contains an infinite number example is found in the NMR spectrum for
component of a beam that is polarized in The rates at which relaxation processes of magnetic components, at least some of which the proton attached to a nitrogen atom
the xy plane. occur affect the nature and quality of an must correspond in frequency and phase with (for 14N, 1= 1).
SpiD-Spin Relaxation •••• Line BroMeoiDg. consisted of determining the decrease in the
Several other efl'ects tend to diminish relax- power (the attenuation) of radiation caused
ation times and thereby broaden NMR lines. by the absorbing sample. While this same
These effects are normally lumped together technique has been applied in NMR spectros-
and described by a spin-spin relaxation time copy, it suffers from the disadvantage that
T2• Values for 72 are generally so small for the excess of absorbing particles is so small
crysta1line solids or viscous liquids (as low as that the resulting attenuation is difficult to
10-4 s) as to prohibit the use of samples of measure accurately. Consequently, nearly all
these kinds for high-resolution spectra. NMR spectrometers employ an arrangement
When two neighboring nuclei of the same by which the magnitude of a positive ab-
kind have identical precession rates, but are sorption-related signal is determined.
in different magnetic quantum states, the Figure 14-5b is a schematic representation
magnetic fields of each can interact to cause of the major components of an NMR spec-
an interchange of states. That is, a nucleus in trometer; these are described in greater detail
the lower spin state can be excited while the in a subsequent section. The radiation source
excited nucleus relaxes to the lower energy is a coil that is part of a radio-frequency oscil-
state. Qearly, no net change in the relative lator circuit (p. 24 to 27). Electromagnetic
spin state population results, but the average radiation from such a coil is plane-polarized
lifetime of a particular excited nucleus is (in the xz plane in the figure). The detector is SigMl
component
shortened. Line broadening is the result. a second coil located at right angles to the in y direction
Two other causes of line broadening UllIbsorbed
source (on the y axis in the figure) and is part
component y
should be noted. Both arise if H 0 in Equation of a radio-receiver circuit. The magnetic field
14-10 differs slightly from nucleus to nucleus; used in NMR experiments has a direction (d)

under these circumstances, a band of frequen- along the z axis, perpendicular with respect to
cies, rather than a single frequency, is ab- both the source and the detector.
sorbed. One cause for such a variation in the As shown in Figure 14-5c, the plane-
static field is the presence in the sample of polarized radiation can be resolved into two
other magnetic nuclei whose spins create local circularly polarized vectors that rotate in
fields that may act to enhance or diminish opposite directions to one another (in the xy
the external field acting on the nucleus of plane). Vector addition of these components,
of the nuclei in the magnetic field. The extent data is to employ a constant-frequency radio
interest. In a mobile lattice, these local fields regardless of their angular position, indicates
tend to cancel because the nuclei causing of coupling, and thus the signal strength, oscillator and sweep the magnetic field Ho
that there is no net component along the y
depends upon the number of absorbing nuclei. continuously. Since for a given nucleus, fre-
them are in rapid and random motion. In a axis. Thus, no signal is received by a detector
AbsorptionSpectra. Two methods are avail- quency and field strength are directly propor-
solid or a viscous liquid, however, the local located on that axis.
able for producing NMR spectra. The first is tional (Equation 14-10), Ho can serve equally
fields may persist long enough to produce a Figure 14-5b and 14-Sd shows the efl'ect of a
analogous to the method for obtaining optical well as the abscissa for an absorption spectrum.
range of field strengths and thus a range of sample positioned at the origin of the three
absorption frequencies. Variations in the static spectra; it involves measuring an absorption Figure 14-6 shows such a spectrum.
axes. If the source has a frequency that is ab-
field also result from small inhomogeneities signal as the electromagnetic frequency is Early commercial NMR instruments em-
sorbed by a particular type of nucleus in the
in the field source itself. This efl'ect can be varied. No dispersing elements such as prisms ployed the field sweep method of producing
sample, the power of one of the two circular
largely offset by rapidly spinning the entire or gratings exist for radio frequencies, however. spectra because the electronic equipment re-
components of the beam is diminished.
sample in the magnetic field. Thus, a variable-frequency oscillator with a quired to produce a linear variation in mag-
Vector addition of the resulting components
linear sweep is required to produce radio netic field is considerably simpler and less ex-
indicates that the radiation now has a fluctu-
frequencies that can be varied continuously pensive than a linear sweep oscillator. The
ating component in the y direction, which
Measurement of Absorption over a range of 1 kHz for hydrogen and as latter, however, has the advantage of being
causes the detector to respond. Thus, the
Absorption Signal The analytical measure- great as 10 kHz for such nuclei as 13C and more efficient in producing spin-decoupled
sample serves to couple the generator to
ment in all of the types of absorption spectros- 19F. spectra (see p. 402). As a consequence, several
t~e receiver, provided the frequency of radia-
copy that we have discussed thus far has An alternative way of collecting NMR more recent instruments make use of frequency
tion corresponds to the precessing frequency
sweep; some are capable of both sweep modes.
Table 14-2 provides spectral data for several
common nuclei employed for NMR studies.

EXPERIMENTAL METHODS
OF NMR SPECTROSCOPY

Nuclear magnetic resonance instrumentS are


either high-resolution or wide-line instruments.
Only the former can resolve the fine struc-
tures associated with absorption peaks; the
nature of this fine structure is determined by
the chemical environment of the nucleus. FIGURE 14-6 A low-resolution NMR spec-
High resolution requires the use of magnetic trum of water in g1ass container. Frequency =
5 MHz. (Courtesy of Varian Associates, Palo
fields greater than 7000 G. Wide-line instru-
Alto, California.)
ments are useful for quantitative elemental FIGURE 14-7. Schematic diagram bfan NMR spectrometer.
analysis and for studying the physical envi- (Courtesy of Varian Associates, Palo Alto, California.)
ronment of a nucleus. Figure 14-6 is an
example of a low-resolution spectrum ob- much simpler and less expensive than their
tained with a wide-line instrument. Wide-line high-resolution counterparts.
spectrometers, which employ magnets with From an instrumental standpoint, the
field strengths of a few thousand gauss, are equipment required for high-resolution NMR spectroscopy is the most elaborate needed for cre~e with increases in field strength; in
any absorption method. Thus, most chemists addition, however, the field must be highly
are forced to accept only a general under- horltogeneousllnd reproducible. These require-
standing of the operating principles of the ments make the magnet by far the most ex-
NMR spectrometer, and leave its design and pensive component of an NMR spectrometer.
TABLE 14-2 SPECTRAL AND MAGNETIC maintenance to those skilled in electronics. Spectrometric magnets are of three types:
PROPERTIES OF SOME COMMON NUCLEI On the other hand, the sample-handling permanent magnets, conventional electro-
techniques, the interpretation of spectra, and magnets, and superconducting solenoids. Per-
A1IIorptioB Isotopic
NlIdeas Spin Allanilance, % an appreciation of the effects of variables are manent magnets with field strengths of 7046
~ ~ no more complex than for other types of or 14,092 G are used in several commercial
IH ! 60.0 99.98 1.000 absorption spectroscopy; these areas are of instruments; corresponding oscillator frequen-
7Li i 23.3 92.57 0294 greatest interest to the chemist. cies for proton studies are 30 and 60 MHz.
13C 1.11
14N
! 15.1 0.00018 Permanent magnets are highly temperature-
1 4.3 99.63 0.001 Inatrumentation3 sensitive and require extensive thermostating
170
19F
i 8.1 0.04 0.‫סס‬OO1
Figure 14-7 is a schematic diagram showing
and shielding as a consequence.
! 56.5 100 0.833
the important components. of an NMR spec-
Electromagnets are relatively insensitive to
23Na i 15.9 100 0.093 temperature fluctuations but require cooling
25Mg trometer. A brief description of each follows.
i 3.8 10.05 0.027
The Magnet. The sensitivity and resolu-
systems to remove the heat generated by the
27AI i 15.6 100 0206 lar,e currents that are required. Elaborate
29Si 11.9 tion of an NMR spectrometer are critically
31p
! 4.70 0.00037
dependent upon the strength and quality of
power supplies are also necessary to provide

33g
! 24.3 100 0.066
its magneJ. Both sensitivity and resolution in-
the required stability: Commercial electro-

I09Ag
i 4.6 7.67 0.002 magnets generate fields of 14,092, 21,140, and
! 2.8 48.65 0.0001 23,490 G, corresponding to proton absorp-
tion frequencies of 60, 90, and 100 MHz.
• MHz in a 14092.0 mapelic field. • For a concise description of NMR spectrometers, see: Superconducting magnets are used in the
• Sensitivity relative to an equal,number of protons at a constant field. D. G. Howery, J. Chern. Edue, •••• A327. A389 (1971~ highest resolution and most expensive instru-
ments. Here, fields as great as 110,390G are system gives control of the order of parts in mit estimation of the relative number of ab-
~.12.'8-4·90.,.488!l1,.5
attained, corresponding to a proton frequency 1010. CH21r81 4.90 sorbing nuclei in each chemical environm~t
of 470 MHz. The Field SweepGeaerator. A pair of coils (Figure 14-8).This informationis vital to the
The performance specificationsfor a spec- located parallel to the magnet faces (see Integrol CUM deduction of chemical structure..In addition,
trometer magnet are stringent. The field pro- Figure 14-7)permits alteration of the applied 4.90 V peaIcareas are usefulforquantitativeanalytical
duced must be homogeneous to a few parts field over a small range. By varying a direct worlc.
per billion within the sample area and must current through these coils, the effectivefield Peale heights are not particularly satisfac-
be stable to a similar degree for short periods can be changed by a few hundred milligauss tory measures of concentration because a
of time. Unfortunately, the inherent stability without loss of field homogeneity. number of variables that are difficult to con-
of most magnets is considerably lower than Ordinarily, the field strength is changed trol can alter the widths and thus the heights
this; variations as large as one part in 10' are automatically and linearly with time, and this of the pealcs.Peale areas, in contrast, are not
often observed over a period of 1 hr. In change is synchronized with the linear drive affectedby these variables.
order to offset the effectof field fluctuations, of a chart recorder. For a 6O-MHz proton The Sample Holder •••• SampleProbe. The
a frequency lock system is employed in all instrument, the sweep range is 1000 Hz (235 usual NMR sample cell consists of a 5-mm
FIGURE 14-8 Absorption and integral
commercial NMR instruments. Here, a refer- milligauss)or some integral fraction thereof. 0.0. glass tube containing about 0.4 ml of
curve for a dilute ethylbenzene solution (ali-
enee nucleus is continuously irradiated and For nuclei such as I'F and nC, sweeps as liquid. Microtubes for smaller sample vol-
phatic region).(Courtesy of Varian Associates,
monitored at a frequencycorrespondingto its great as 10 kHz are required. umes are also available.
Palo Alto, California.)
resonancemaximum at the rated fieldstrength As was noted earlier, several more recent The sample probe is a device for holding
of the magnet. Changes in the intensity of the instruments employ a frequencysweep system the sample tube in a fixedspot in the field.As
reference absorption signal, resulting from in lieu of or in addition to magnetic field will be seen from Figure 14-9,the probe con:
drifts in the magnetic field,control a feedbaclc sweep. tains not only the sample holder but the
circuit, the output of which is fed into coils in The Radio-Frequency Source. The signal withelectronicintegrators to provide informa- sweep source and detector coils as well. to
the magnetic gap in such a way as to correct from a radio-frequency oscillator (transmit- tion regarding areas under absorption peaks assure reproducible positioningof the sample
for the drift. ter) is fed into a pair of coils mounted at 90 (seeFigure 3-26,p. 66 for a simple integration with respect to those components; it may also
Recall that the ratio between the field deg to the path of the field.A fixed oscillator circuit). Usually, the integral data appear as contain the referencecell and liquid used as
strengths and resonance frequenciesis a con- of exactly 60, 90, or 100 MHz is ordinarily step functions superimposed on the NMR an external lock. The detector and receiver
stant regardless of the nucleus involved (Equa- employed; for high-resolution worlc, the fre- spectrum (seeFigure 14-8).Generally,the area coils are oriented at right angles to one
tion 14-10).Thus, the drift correction for the quencymust beconstant to about 1 part in 10'. data are reproducibleto a fewpercent relative. another to minimize the transfer of power in
referencesignal is applicable to the signals for The power output of this source is less than Pealeareas are important becausethey per- the absence of sample.Even with this arrange-
all nuclei in the sample area. 1 W and should be constant to perhaps 1 %
Two types of locIcsystemsare encountered. over a period of several minutes. The output
The external lock employs a separate sample is plane polarized.
container for the referencesubstance; this con- The Signal Detector •••• Recorder System.
tainer must be located as close as possible to The radio-frequency signal produced by the
the analyte container. In an internal lock resonatingnucleiisdetected by mellnsof a coil
system,the referencesubstance is dissolved in that surrounds the sample and is at right
the solution containing the sample. Here, the angles to the source coil. The electrical signal
reference compound is usually tetramethyl- generated in the coils 'is small and must be
silane (TMS~ which also servesas an internal amplifiedby a factor of 10' or more before it
standard (see p. 392). can be recorded.
The external locIc is simpler to operate, The abscissa drive of an NMR recorder is
and control is not lost during sample changes. synchronizedwith the spectrum sweep; often,
It does not, however,provide as good control in fact, the recorder controls the sweep rate.
over field changes encountered by the sample The responseof the recorder must be rapid; it
because of its displacement from the sample. is also desirable that the sweep rate be
Thus, the external system provides control of variable. FIGURE 14-9 NMR probe. (Courtesy of Varian Associates,Palo
the order of parts in 10', while the internal All modern NMR recorders are equipped Alto, California.)

~ -~_~ __ l__-
I
ment, some leakage of power betweeQ the absorption peaks is examined in detail with resents a profitable approach of the study of be increased by 100/60 (see Figure 14-11). In
source and receiver does occur; the so-called an instrument that permits the determination nuclei that possess magnetic properties. The contrast, the distance between the fine-struc-
paddies, shown in Figure 14-9, reduce this of absorption over very much smaller incre- fact that the proton appears in most organic ture peaks within a group (lower spectrum)
leakage to a tolerable leveL ments of the abscissa (Ho or v~ the single compounds, and in many inorganic ones as would not be altered by this frequency change.
The sample probe is also provided with an peak is ',usually found to be composed of well, has resulted in a concentration of effort
air-driven turbine for rotating the sample several Peaks (for example, see the proton upon this particular nucleus; our discussion
tube at several hundred rpm along its longitu- spectra in Figure 14-10); moreover, the posi- follows this trend. Measurement of the Chemical
dinal axis. This rotation serves to average out tion and intensity of the component peaks Shift and Spin-Spin Splitting
the effects of inhomogeneities in the field; depend critically upon the chemical environ-
Types of Source of the Chemical Sbift. The chemical
sharper lines and better resolution are ob- ment of the nucleus responsible for the
absorption. It is this dependence of the fine-
Environmental Effects shift arises from a circulation of the electrons
tained as a consequence.
structure spectrum upon the environment of a The spectra for ethyl alcohol, shown in Figure surrounding the nucleus under the influence
Cost of NMR Instruments. The foregoing
nucleus that makes NMR spectroscopy such 14-10, illustrate two types of environmental of the applied magnetic field. This phenome-
brief discussion suggests that NMR instru-
a valuable tool. effects. The curve in Figure 14-10., obtained non is discussed later in greater detail; for the
ments are complex. As might be expected,
The effect of chemical environment on with a lower resolution instrument, shows three present, it suffices to say that this movement of
their cost is high, ranging from about 56000
nuclear magnetic absorption spectra (Figure electrons creates a small magnetic field that
to $200,000. The typical60-Hz high-resolution proton peaks with areas in the ratio 1 : 2 : 3
14-10) is general; as a consequence, NMR rep- ordinarily opposes the applied field. As a con-
instrument employed for routine work in many Qeft to right). On the basis of this ratio, it
appears logical to attribute the peaks to the sequence, the nucleus is exposed to an effective
laboratories is priced from· approximately
field that is somewhat smaller (but in some
530,000 to $40,000. hydroxyl, the methylene, and the methyl pro-
instances, larger) than the external field. The
tons, respectively. Other evidence confirms
magnitude of the field developed internally is
Sample Handling this conclusion; for example, if the hydrogen
directly proportional to the applied external
atom of the hydroxyl group is replaced by
For high-resolution work, samples must be in field, so that we may write
deuterium, the first peak disappears from this
a nonviscous liquid state. Most commonly,
part of the spectrum. Thus, small differences
solutions of the sample (2 to 15%) are
occur in the absorption frequency of the
employed. The sample can also be examined = Happ,(1 - (1)
proton; such differences depend upon the
neat if it has suitable physical properties.
group to which t!te hydrogen atom is bonded. where Happl is the applied field and Ho is the
The best solvents for proton NMR spec-
This effect is called the chemical shift. resultant field which determines the resonance
troscopy contain no protons; from this stand-
The higher resolution spectrum of ethanol, behavior of the nucleus. The quantity (1 is the
point, carbon tetrachloride is ideal. The low
shown in Figure 14-1Ob, reveals that two of shielding parameter, which is determined by
solubility of many compounds in carbon tetra-
the three proton peaks are split into addi- the electron density around the nucleus and
chloride limits its value, however, and a variety
tional peaks. This secondary environmental which depends upon the structure of the com-
of deuterated solvents are used instead. Deu-
effect, which is superimposed upon the chemi- pound containing the nucleus.
terated chloroform (CDCI3) and deuterated
cal shift, has a different cause; it is termed The shielding parameter for protons in a
benzene are commonly encountered.
spin-spin splitting. methyl group is larger than (1 for methylene
Both the chemical shift and spin-spin split- protons, and the parameter is even smaller for
ting are important in structural analysis. the proton in an -OH group. It is, of course,
ENVIRONMENTAL EFFECTS Experimentally, the two types of NMR zero for an isolated hydrogen nucleus. Thus,
ON PROTON NMR SPECTRA peaks are readily distinguished, for it is found in order to bring any of the protons in
that the peak separations (in units of v or Ho) ethanol into resonance at a given oscillator
Figure 14-6 is an example of a low-resolution resulting from a chemical shift are directly frequency v, it is necessary to employ a field
NMR spectrum. Each kind of nucleus is char- propOrtional to the field strength or to the Happl that is greater than Ho (Equation 14-
acterized by a single absorption peak, the lo- oscillator frequency. Thus, if the spectrum in 11~ the resonance value for the isolated
cation of which appears to be independent of FIGURE 14-10 NMR spectra of ethanol Figure 14-10. were to be obtained at 100 proton. Since (1 differs for protons in various
the chemical state of the atom. It: however, at a frequency of 60 MHz. (a) Resolution MHz rather than at 60 MHz, the horizontal functional groups, the required applied field
the spectral region around one of the nuclear _1/106; (b) resolution _1/107• distance between any pair of the peaks would differs from group to group. This effect is
shown in the spectrum of Figure 14-108, considered later, the shielding parameter for But Hapl• is very nearly the same as Ho, so large. As shown, the zero value for the ~ scale
where the hydroxyl proton appears at the TMS is larger than for most other protons. that the ratio HoIH •• is very nearly unity. cOrresponds to the TMS peak and the value
Thus, the compound provides, at a high applied 'f!tus,
lowest applied field, the methylene protons of ~ increases from right to left. The t scale, of
next, and finally the methyl protons. Note field, a single sharp peak that is isolated from course, changes in the opposite way. Refer-
most of the peaks of interest in a spectrum.
~ :0: H,e( - Hapl• 106
that all of these peaks occur at an applied - Href x ring again to Figure 14-11, note that the var-
field greater than the theoretical one for the In addition, TMS is inert, readily soluble in ious peaks appear at the same values of ~ and
isolated hydrogen nucleus, which would lie most organic liquids, and. easily removed from The quantity ~ is dimensionless and expresses t in spite of the fact that the two spectra were
far to the left in Figure 14-10a. Note also that samples by distillation (bp = 27°C). Unfortu- the relative shift in parts per million; for a obtained with instruments having different
if the applied field is held constant at a level nately, TMS is not water soluble; in aqueous given peak, ~ will be the same regardless of fixed fields. .
necessary to excite the methyl proton, an in- media, the sodium salt of 2,2-dimethyl-2-sila- whether a 60-, 90-, or l00-MHz instrument is For the purpose of reporting spin-spin
crease in frequency would be needed to bring pentane-S-sulfonate, employed. Most proton peaks lie in the ~ splitting, it is desirable to utilize scalar units
the methylene protons into resonance. range of 1 to 12. of milligauss or hertz. The latter scale has
Source of Spin-Spin Splitting. The splitting Another chemical shift parameter t is now come into more general use and is the
of chemical shift peaks can be explained by defined as one shown in Figure 14-11. The position of
assuming that the effective field around one may be used in its stead. The methyl protons the reference TMS peak is arbitrarily taken as
t = 10 - ~ (14-16)
nucleus is further enhanced or reduced by of this compound produce a peak analogous zero, with frequencies increasing from :right to
local fields generated by the hydrogen nuclei to that of TMS; the methylene protons give Generally, NMR plots have linear scales in left. The effect of this choice is to make the
bonded" to an adjacent atom. Thus, the fine a series of small peaks that are readily identified ~ (and sometimes t1 and the data are plotted frequency of a given peak identical with the
structure of the methylene peak shown in and can thus be ignored. . with the field increasing from left to right (see increase in oscillator frequency that would be
Figure 14-10b can be attributed to the effect The field strength H ref required to produce Figure 14-11). Thus, if TMS is employed as needed to bring that proton into resonance if
of the local fields associated with the adjacent the TMS resonance line at frequency v is given the reference, its peak will appear on the far the field were maintained constant at tbe level
methyl protons. Conversely, the three methyl by Equation 14-11 right-hand side of the plot, since (T for TMS is required to produce the TMS peak. '
peaks are caused by the adjacent methylene
protons. These effects are independent of the
applied field and are superimposed on the
effects of the chemical shift. -, l-J-7Hz
Abscissa Scales for NMR Spectra. The I
I
determination of the absolute field strength I
I
with the accuracy required for high resolution
is difficult or impossible; on the other hand, it
is entirely feasible to determine within a few Similarly, for a given absorption peak of the
milligauss the change in field strength caused sample, we may write O_IHz)
by the subsidiary sweep coils. Thus, it is ex- o 6
10 T
pedient to report the position of resonance HaPI• - Ho
absorption peaks relative to the resonance (Tapl. = H
splc --ll-J=7Hz
peak for a standard substance that can be
measured at essentially the same time. In this where H.PI• is the field necessary to produce
way, the effect of fluctuations in the fixed the peak. We then' define a chemical shift
TMS
magnetic field is minimized. The use of an parameter ~ as
internal standard is also advantageous in that
chemical shifts can be reported in terms that
are independent of the oscillator frequency.
~':'",,', I I l:zbo' I I
'''~'~ "I I I ~O _(Hz)
4 3 2 1 o 6
A variety of internal standards have been and substitute Equations 14-12 and 14-13 6 7 "8 9 10 T

proposed, but the compound now most gen- into this expression to obtain Lowfield Ho---------- High
field

erally accepted is tetramethylsilane (TMS1 Lowshielding High


shielding
High
f'equency-------- ----------Lowfrequency
(CH3)4Si. All of the protons in this com-
pound are identica~ and for reasons to be
It can be seen in Figure 14-11 that the \ , compounds have been found to differ ap-'
spin-spin splitting in frequency units (J) is the; --,',11/
'\ I I .••.
_
..... preciably, depending upon the orientation of
same for the 6O-MHz and the l00-MHz in-
/
/ "\' \ \\ II /
f I /Circulating \ \
the ring with respect to the applied field. This
anisotropy is readily understood from the
struments. Note, however, that the chemical
I \\ / I electrons ,
shift in frequency units is enhanced with the I \ I I \ model shown in Figure 14-13. Here, the plane
higher frequency instrument. I , of the ring is perpendicular to the magnetic
, I
I I field; in this position, the field can induce a
The Ci1emical Shift I I flow of the It electrons around the ring (a ring
\ I I ' 1\ \ I current). The consequence is similar to that of
As noted earlier, chemical shifts arise from \ ;'1/11\ /
1 a current in a wire loop; namely, a secondary
the secondary magnetic fields produced by \
the circulation of electrons in the molecule. ,',_ / / /II1I 1\'\"\ " _ //I field Is produced that acts in opposition to
These electronic currents (local diamagnetic the applied field. This secondary field, how-
/ I ali ,
o \----. Secondary ever, exerts a magnetic effect on the protons
currents~) are induced by the fixed magnetic , field
field and result in secondary fields that may attached to the ring; as shown in Figure 14-13,
either decrease or enhance the field to which this etTectis in the direction of the field. Thus,
a given: proton responds. The effects are FIGURE 14-13 Deshielding of the aromatic protons require a lower external
complex, and we consider only the major aromatic protons brought about by field to bring them into resonance. This effect
aspects of the phenomenon here. More com- ring current. is either absent or self~nceling in other
plete treatments can be found in the several orientations of the ring.
Applied field
reference- works listed under footnote 2, page A somewhat analogous model can be envi-
377. : FIGURE 14-12 Diamagnetic sioned for the ethylenic or carbonyl double
shielding of a nucleus. Effect of Magnetic Anisotropy. It is appar- bonds. Here, one can imagine circulation of
Unde( the influence of the magnetic field,
ent from an examination of the spectra of the It electrons in a plane along the axis of the
electroM bonding ~he proton tend to precess
compounds containing double or triple bonds bond when the molecule is oriented to the
around the nucleus in a plane perpendicular
that local diamagnetic effects do not suffice to field as shown in Figure 14-14a. Again, the
to the magnetic field (see Figure 14-12). A
explain the position of certain proton peaks. secondary field produced acts upon the proton
consequence of this motion is the develop-
rounding it. Thus, in the absence of the other Consider, for example, the irregular change to reinforce the applied field. Thus, deshielding
ment of a secondary field, which opposes the
influences, shielding would be expected to in lJ values for protons in the following hy-
primary field; the behavior here is analogous
decrease with increasing electronegativity of drocarbons, arranged in order of increasing
to the passage of electrons through a wire
adjlWC11tgroups. This etTect is illustrated by acidity (or increased electronegativity of the
loop. The nucleus then experiences a resultant
the lJ values for the protons in the methyl
field, which is smaller (the nucleus is said to
be shielded from the full effect of the primary
halides, CH3X, which lie in the order I (2.l6~
groups to which the protons are bonded):
CH3 -CH3 (15 = 0.9~ CH2 -eH2 (lJ = S.8~ ~lfOHO
Br (2.68~ CI (3.0S~ and F (426). Here, iodine and HCSCH (lJ = 2.9). Furthermore, the
field); as a consequence, the external field
(the least electronegative) is the least effective aldehydic proton RCHO (15 - 10) and the .."'\--_c.,.
must be increased to cause nuclear resonance.
-
protons on benzene (15 - 7.3) appear consider-

$
of the halogens in withdrawing electrons from ....••.•
The frequency of the precession, and thus the ..

magnitude of the secondary field, is a direct


the protons; thus, the electrons of iodine ably farther downfield than might be expected ,- -,
provide the largest shielding effect. Similarly, on the basis of the electro negativity of the
function of the external field.

'=~1~
..
electron density around the methyl protons of groups to which they are attached.
The shielding experienced by a given nucleus
methanol is greater than around the proton The effects of multiple bonds upon thechem-
is directly related to the electron density sur-
i associated with oxygen because oxygen is more ical shift can be explained by taking into
electronegative than carbon. Thus, the methyl account the anisotropic magnetic properties
• The intensity or mapetization induced in a diomDgllelk peaks are upfield from the hydroxyl peak. The of these compounds. For example, the mag-
substance is smaller than that produced in a vacuum
with the same field. Diamagnetism is the result or motion
position of the proton peaks in TMS is also
explained by this model, since silicon is
netic susceptibilities 5 of crysta1line aromatic 1 f Ho 1 1
induced in bonding electrons by the applied field; this (oJ (bl
motion (a diomDgllelicCIl"~nt) creates a secondary field
relatively electropositive. Finally, acidic pro-
that opposes the applied field. Paramag~tlsm (and the tons have very low electron densities, and the FIGURE 14-14 Deshielding of ethylene
resulting paramagnetic CIlrr~nts) operates in just the peak for the proton in RS03H or RCOOH and shielding of acetylene brought about by
opposite sense. lies far doWnfield (15 > 10). electronic currents.
Correlation of Chemical Shift with Struc-
shifts the peak to larger values of ~. With. an
ture. The chemical shift is employed ·for the
aldehyde, this effect combines with the de-
identification of functional groups and as an
shielding brought abj.>utby the electronegative
aid in determining structural arrangements of
nature of the carbonyl group; a very large value
groups. These applications are based upon
of ~ results.
empirical Correlations between structure and
In an acetylenic bond, the symmetrical dis-
shift. A number of correlation charts6 and
tribution of 11: electrons about the bond axis
tables 7 have been published. Two of these are
permits electron circulation around the bond
shown in Figure 14-15 and Table 14-3. It
(in contrast, such circulation is prohibited by
the nodal plane in the electron distribution of
a double bond). From Figure 14-14b, it can
be seen that in this orientation the protons • N. F. Chamberlain, Anal. CIIem. 31, 56 (1959).
are shielded. This effect is apparently large 7 R. M. Silverstein, G. C. Bassler, and T. C. Morril~
enough to offset the deshielding resulting Spectrometric Identifialtion oJ Organic Compounds, 3d ed.
from the acidity of the protons and from the New York: Wiley, 1974,Chapter 4; and L. M. Jackman
electronic currents at perpendicular orienta- and S. Sternhal1, Nuc/eQr Magnetic Resonance Spectros-
copy, 2d ed. New yort: Perpmon Prea, 1969.
tions of the bond.

TABLE 14-3 APPROXIMATE CHEMICAL SHI"S


FOR CERTAIN METHYL, METHYLENE, AND
METHINE PROTONS

a,ppm
Structure M=CH3 M=CHz M=CH

Aliphatic p substituents 2.0


M-e-CI 1.5 1.8
1.8 1.8 1.9
M-C-Br 2.5
M-e-NOz 1.6 2.1
1.2 1.5 1.8
M-e-OH (or OR)
1.3 1.6 1.8
M-e-OC(41R 1.7
1.1
M-e-Ct-o~ 1.1 1.6 2.0
M-C-C-oR
M-e-C-o R 1.1 1.7 1.9
M-e-t/J 1.1 1.6 1.8

Aliphatic lX substituents
M-Cl 3.0 3.5 ~tO
4.1
'Norm.Ily. absorptions for the functtonal groups i~
Approtlliml'le limits for this are indicilted bv ~lion
Me
wi" be found with' the
'HIues in parenlheMsl~ .:v-:::~' ~~IY.
.

h;;':; ~v~::dilute
.
a functlONI grouP will abJoJb outside this range.

M-Br 2.7 3.4 'The absorptton positions of these groups cDnl;eOtration-dependen1 and we shifted 10 more solutions.

M-NOz 4.3 4.4 4.6


FIGURE 14-15 f
Abso rP.lon. 't'
posllOns 0f protons In
.. vanous structural environments. (Table
M-OH~orOR) 3.2 3.4 3.6
3.8 4.0 4.6 t~en from J. R. D!er, AppltcatlOns 0/ Absorption Spectroscopy o/Organic Compounds. Englewood
M-O- 5.0
M-OC(=O)R 3.6 4.1 Cliffs, N.J.: PrentIce-Hall, Inc., © 1965, p. 85. With permission.)
M-e-e 1.6 1.9
M-eac 1.7 2.2" 2.8

Ml-<>i
M-e=OR
M-C-o
M-e=o R
2.2
2.1
2.4
2.2
2.4
2.4
2.7
2.2
2.6
3.4
2.5
M-t/J 2.2 2.6 2.8
398

tion of each nucleus with a small arrow, the Here, we have eight possible spin combina- given by (n + 1). Thus, the multiplicity for the
should be noted that the exact values for lJ tions; however, among these are two groups methylene band in ethanol is determined by
may depend upon the nature of the solvent as four states are
containing three combinations that have equiv- the number of protons in the adjacent methyl
well as upon the concentration of the solute. alent magnetic effects. The methylene peak is groups and is equal to (3 + I). .
These effects are particularly pronounced for thus split into four peaks having areas in the For nuclei having spins (I) other than 1/2,
protons involved in hydrogen bonding; an ratio 1: 3: 3: 1. the multiplicity is given by (2nl + 1).
example is the hydrogen atom in the alco- The interpretation of spin-spin splitting pat- 4. If the protons on atom B are affected by
holic functional group. tern is relatively simple and straightforward protons on atoms A and C that are non-
Spin-Spin Splitting for first-order spectra. First-order spectra are equivalent, the multiplicity of B is equal to
Possible spin orientations those in which the chemical shift between (nA + l)(nc + 1~ where nA and nc are the
As may be seen in Figure 14-10, the absorp- of methylene protons interacting groups of nuclei is large with number of equivalent protons on A and C,
tion bands for the methyl and methylene pro- respect to their coupling constant J. Rigorous respectively.
tons in ethanol consist of several narrow In one combination, the spins of the two first-order behavior requires that Av/J be 5. The approximate relative areas of a
peaks that can be sep~ated only wi!h a. high- methylene protons are paired and ali~ed greater than 20; frequently, however, analysis multiplet are symmetric around the midpoint
resolution instrument. Careful exammatlon of against the field, while in a second, the palr~ of spectra by first-order techniques can be ac,. 0' of the band and are proportional to the coeffi-
these peaks shows that the spacing ~o~ the
three components of ~e methyl band ISiden-
spins are reversed; there are also two combI- complished down to value of Av/J of some- cients of the tenns in the expansion (x + 1 r.
nations in which the spins are opposed to one what less than 10. The ethanol spectrum The application of this rule is demonstrated in .
tical to that for the four peaks of the m~hy!- another. The magnetic effect that is trans- shown in Figure 14-10 is an example of a pure Table 14-4 and in the examples that follow.
ene band' this spacing in frequency urnts IS mitted to the methyl protons on the adjacent first-order spectrum with J for the methyl and 6. The coupling constant is independent of
called the' coupling constant for the interaction carbon atoms is determined by the spin com- methylene peaks being 7 Hz and the separation the applied field; thus, multiplets are readily
and is given the s~bol ~. Moreove~, the binations that exist in the methylene group at between the centers of the two multiplets being distinguished from closely spaced chemical
areas of the peaks in a multIplet approximate any instant. If the spins are paired and op- about 140 Hz. shift peaks.
an integral ratio to oJ)e another. ~us, for the posed to the external field, the effective applied Interpretation of second-order NMR spec-
methyl triplet, the ra\io of area.~ IS 1: ~ : ~; field on the methyl protons is slightly lessened; EXAMPLE
tra is difficult and complex and will no.t be
for the quartet of methylene - peaks, It IS thus, a somewhat higher field is needed ~o For each of the following compounds, cal-
dealt with in this text. It is noteworthy, how-
1:3:3:1. bring them into resonance, and upfield shIft culate the number of multiplets for each band
ever, that because Av increases with increases
OrigiD. It seems plausible to attn"bute. these results. Spins paired and aligned with the field and their relative areas.
in the magnetic field while J does not, spectra
observations to the effect that the spms of (a) ClCH2CH2CH2Cl. The multiplicity of
result in a downfield shift. Neither of the obtained with an instrument having a high
one set of nuclei exert upon the resonance the band associated with the four equivalent
combinations of opposed spin has an effect magnetic field are much more readily inter-
behavior of another. That is to say, a small protons on the two ends of the molecule
on the resonance of the methyl protons. Thus, preted than those produced by a spectrometer
interaction or coupling exists between the two would be determined by the number of pro-
splitting into three ~s ~ults. The ~rea with a weaker magnet.
groups of protons. This explanation ?~Up- tons on the central carbon; thus, the multi-
under the middle peak IS twICe that of eIther Rules Governing the Interpretation of First-
poses that such coupling takes place vIa mt.er- plicity is (2 + 1) = 3 and the areas would be
of the other two, since two spin combinations Order Spectra. The following rules govern
actions between the nuclei and the bondmg 1 : 2 : 1.The multiplicity of two central methyl-
are involved. the appearance offirst-order spin-spin spectra.
electrons rather than through free space. For ene protons would be determined by the four
Let us now consider the effect of the three 1. Equivalent nuclei do not interact with
our purpose, however, the details of this mech- equivalent protons at the ends and would thus
methyl protons upon the methylene peak. Pos- one another to give multiple absorption
anism are not important. be (4 + 1) = 5. Expansion of (x + 1)4 gives the
sible spin combinations for the methyl protons peaks. The three protons in the methyl group
Let us first consider the effect of the methyl- following coefficients (Table 14-4), which are
are in ethanol give rise to splitting of the adjacent proportional to the areas of the peaks
ene protons in ethal)ol on the resonance of methylene protons only and not to splitting
the methyl protons. 'we must first remember +- +- 1:4:6:4:1.
+- -+ among themselves. (b) CH3CHBrCH3• The band for the six
that the ratio of protons in the two possible
-+ -+ 2. Coupling constants decrease with separa- methyl protons will be made up of (1 + 1) = 2
spin states is very n~rl~ one: even in a strong
-+
tion of groups, and coupling is seldom ob- peaks having relative areas of 1 : 1; the proton
magnetic field. We can Imagme, then, that the +- +- -+
+- -+ served at distances greater than three bond on the central carbon atom has a multiplicity
two methylene protons in a molecule can Ho-+ +- -+
lengths.
have four possible combinations of spin states
+- +- -+ -+
of (6 + 1) = 7. These peaks will have areas
3. The multiplicity of a band is determined
and that in an entire sample each of these -+ -+ (Table 14-4) in the ratio of
+- -+ by the number n of magnetically equivalent
combinations will be approximately equally 1 : 6 : 15 : 20 : 15 : 6: 1
+- +- protons on the neighboring atoms and is
represented. If we represent the spin orienta-
(c) CH3CH2OCH3• The methyl protons
on the right are separated from the other pro-
tons by more than three bonds so that only a
single peak will be observed for them. The
CH3CHzCHzI
(a) (b) (c)
I
1 32) (b)
Protons
evident from the difference between J1•b) and
J (be)' Thus, invoking rule 4, the number of
peaks will be (3 + 1)(2 + 1) = 12. In cases
such as this, deriyation of a splitting pattern,
alcohol. Note the triplet OH peaks and the
eight methylene peaks in this spectrum. If, now,
a trace of acid or base is added to the pure
sample, the spectrum reverts to the form shown
1.-7.311/~ ~,
protons of the central methylene group will as shown in Figure 14-16, is helpful. Here, the in Figure 14-10.
I Hz / I I '" , (b) Protons split
have a multiplicity of (3 + 1) = 4 and a ratio II / ,by 3 protons on (a) effect of the (a) proton is first shown and The exchange of OH protons among alco-
of 1 : 3 : 3 : 1. The methyl protons on the left 1// " Jlabl· 7.3 Hz leads to four peaks of relative areas 1 : 3 : 3 : 1 hol molecules is known to be catalyzed by
V ,
have a multiplicity of (2 + 1) = 3 and an area spaced at 7.3 Hz. Each of these is then split both acids and bases, as well as by the impuri-
1(4) (12) (12) 1(4)
ratio of 1 : 2 : 1. into three new peaks spaced at 6.8 Hz, having ties that commonly occur in alcohol. It is thus
1', /i\ 1"\ It, Split by 2 relative areas of 1 : 2 : 1. (The same final pat- plaUSible to associate the decoupling observed
" I, I I \ I I, I I , protons on (c)
The foregoing examples are relatively I ' ,I I 'I I ,I I \ Jtbol· 6.8 Hz
tern is produced if the original band is first in the presence-of these catalysts to an ex-
simple because all of the protons influencing 1'(IXIVI\ split into a triplet.) At very high resolution, change process. If exchange is rapid, each OH
I I I, I 1\ II 1\ I \
the multiplicity of any single peak are mag- the spectrum for l-iodopropane exhibits a

"I II'I \ \1II 'II


II 'II: \\
I II 1/'1\ \ group will have several protons associated
netically equivalent. A more complex splitting series of peaks that approximates the series with it during any brief period; within this
pattern results when a set of protons is af- (1) (2)(3) (1)(6)(3) (3)(6)(1) (3)(2) (1)
shown at the bottom of Figure 14-16. At interval, all ofthe OH protons will experience
fected by two or more nonequivalent protons. 10-6.8""; lower resolution (so low that the instrument
I Hz I
the effects of the three spin arrangements of
As an example, consider the spectrum of -----Frequency p. Hz----- does not detect the difference between J(.b) the methylene protons. Thus, the magnetic
l·iodopropane, CH3CH2CH2I. If we label the and J (be), only six peaks are observed with effects on the alcoholic proton are averaged
FIGURE 14-16 Splitting pattern for meth-
three carbon atoms (a1 (b1 and (c) from left relative areas of and a single sharp peak is observed. Spin
ylene (b) protons in CH3CH2CH2I. Figures
to right, the chemical shift bands are found at decoupling always occurs when the exchange
in parentheses are relative areas under peaks.
c5(a) = 1.02, b(b) = 1.86, and b(e) = 3.17. The
frequency is greatet than the separation (in
band at b(a) = 1.02 will be split by the two frequency units) between the interacting com-
methylene protons on (b) into (2 + 1) = 3 Sec:ond-<>nlerSpectra. Coupling constants ponents.
peaks having relative areas of 1 : 2 : 1. A simi· are usually smaller than 20, whereas chemical Chemical exchange can affect not only spin-
lar splitting of the band b(e} = 3.17 will also J (be) = 6.8. The band for the methylene pro- shifts may be as high as 1000. Therefore, the
be observed. The experimental coupling con- tons (b) is affected by two groups of protons splitting behavior described by the rules in
stants for the two shifts are J(ab) = 7.3 and which are not magnetically equivalent, as is the previous section is common. However,
when AvlJ becomes less than perhaps 7, these
rules no longer apply. Generally, as Av ap-
proaches J, the peaks on the inner side of two
multiplets tend to be enhanced at the expense
of the peaks on the outer side, and the sym-
'1i
metry of each multiplet is thus destroyed, as 6.
.;;;
TABLE 14-4 RELATIVE INTENSITIES OF FIRST-ORDER MULTIPLETS noted earlier. Analysis of a spectrum under c
a
(/= !> these circumstances is difficult. .~
NlBDber of Equivalent
meet of Chemical Exchange on SpectrL .s« -OH
Multi~liCity, Turning again to the spectrum of ethanol
Protons, n (n + 1 Relative Peak Areas (Figure 14-101 it is interesting to consider
0 1 why the OH proton appears as a singlet
1 2 rather than a triplet. The methylene protons
2 3 2 1 and the OH proton are separated by only
3 4 3 3 three bonds; coupling should occur to in-
4 5 1 4 6 4
6 1 5 10 10 5 crease the multiplicity of both OH and the FIGURE 14-17 Spectrum of highly puri-
5
6 7 1 6 15 20 15 6 methylene peaks. Actually, as shown in Figure fied ethanol showing additional splitting of
7 8 1 7 21 35 35 21 7 1 14-17, the expected multiplicity is observed by OR and CH2 peaks (compare with Figure
employing a highly purified sample of the 14-10).
spin spectra but also chemical shift spectra. pany spin decoupling. Spectrum B shows the . sweep
absorption associated with the four protons ,po.: '" 'I'
Purified alcohol/water mixtures have two well-
on the pyridine ring of nicotine. Spectrum C
20_

,o_
'PO .", III
.", '10
defined and easily separated OH proton peaks.
was obtained by sweeping the same portion
110 .", 31l '10
Upon addition of an acid or base as a catalyst, .
however, the two peaks coalesce to form a of the spectrum while simultaneouslY'irradiat-
s_ 10 '"
10
'''' '" ""·10
2_ 'I- ''''
single sharp line. Here, the catalyst enhances ing the sample with a strong, second, radio- b ,
the rate of proton exchange between the alco- frequency signal having a frequency corre- b , ,
hol and the water and thus averages the shield- sponding to the absorption peaks of protons
ing effect. A single sharp line is obtained when (d) and (c) (about 8.6 ppm). The consequence _. (ll)
I
the exchange rate is significantly greater than is a decoupling of the interaction between
these two protons and protons (a) and (b).
n 3

the separation frequency of the individual lines C r


of alcohol and water. On the other hand, Here, the complex absorption spectra for (a)
and (b) collapse to two doublet peaks which
r I
if the exchange frequency is about the same as n

this frequency difference, shielding is only arise from coupling between these protons.
partially averaged and a broad line results. Similarly, the spectra for (d) and (c) could be
simplified by decoupling with a beam having
y .
,. ~III.
The correlation of line breadth with exchange
rates has provided a direct means for investi-
gating the kinetics of such processes and rep-
a frequency corresponding to the peaks for
protons (a) or (b).
The interaction between dissimilar nuclei
B
~
I"'u,
R :
\U
- -
/
resents an important application of the NMR / cl.
experiment. can also be decoupled. An important example A '\. jJ kJ / lWL
is the decoupling of 14N nuclei and protons,
which can be accomplished by irradiation of
"- I"

""
Aida in the the sample with a beam having a frequency of 61 10 9 ! 8 7 6 5 4 3 2 1 0
Analysis of Complex Spectra about 4.3 MHz when the magnetic field is
14,092 G. Absorption of this radiation by the FIG U R E 14-1~ Effect of spin decoupling on the NMR spectrum of nicotine dissolved in CDCI3 •
Several methods are available for the simplifi-
14N nuclei results in a complete decoupling of <:urve A: the entire spectrum. Curve B: expanded spectrum for the four protollS on the pyridine
cation of spectra that are too complex for
their interaction with protons and often pro- n~g. Curve C: spectrum for protons (a) and (b) when decoupled from (d) and (c) by irradiation
ready analysis. The more common of these are
duces spectral simplifications. with a second beam that has a frequency corresponding to about 8.6 ppm. (Courtesy Varian
discussed briefly in this section.
Instrument Division, Palo Alto, California 94303.)
IIIaftIeI Ia the Mqnetic FieW. As men- The theory of spin-spin decoupling is com-
tioned eartier, coupling constants are unaf- plex and beyond the scope of this book.
fected by increases in the magnetic field, where- Chemical Shift Reagents.' Important aids in
as chemical shifts are enhanced. Thus, the use the interpretation of proton NMR spectra are praseodymium(lII) [usually abbreviated as plex and molecules containing oxygen, nitro-
of an instrument with a stronger magnet may shift reagents, which have the effect of dis- Pr(DPMh]· gen, or other atoms that contain free electron
convert an uninterpretable second-order spec- persing the absorption peaks for certain types pairs.
tra to one that is susceptible to first-order of compounds over a much larger frequency The DPM complexes of europium and
analysis. range. This dispersion will frequently separate praseodymium are generally employed in non-
Isotopic SubstitutiolL Substitution of deu- otherwise overlapping peaks and permit easier polar solvents such as CCI4, CDCI3, and C6D6
terium for one or more of the protons in interpretation. to avoid solvent competition with the analyte
a molecule simplifies the spectrum by re- Shift reagents are generally complexes of for electron receiver sites on the metal ion.
moval of the absorption peaks corresponding europium or praseodymium. A typical ex- Figure 14-19 illustrates the dramatic effect
to the substituted protons. Further simplifi- ample is the dipivalomethanato complex of that Pr(DPMh has on the complex spectrum
cation may also be observed because the cou- for styrene oxide. Here, the hydrogens closest
pling between a deuterium and a proton is to the oxygen binding site are shifted to higher
significantly less strong than between two pro- The praseodymium ion in this neutral com- fields and actually above that for the TMS
IFor a brief lummary or shift reagents, see: SIIIfl Re-
toIlS. agnulnNMR,Pnkl ••..EImeNMR Quarrerly,Number 7, plex is capable of increasing its coordination reference. Note also that the ortho hydrogen
Spin DeeonpIing. Figure 14-18 illustrates August 1977, The Perkin-Elmer Corporation, Norwalk, by interaction with lone electron pairs. There- in the ring is shifted to a greater extent than
the spectral simplification that may accom- CT06352 fore, reactiollS can take place between the com- the meta or para hydrogens, again as a con-
: ... •• H. ~ H. 20
!" 10 •••
•••
1
Pr(d mb
T S
Pr( ~o /0, /H
......"c
-
'Ii / "'-
u u
,r
6.2V
Ifk 1\ ~1-
8

II ./
./'

:;=::::::::- ~
~ ::-::
-
A A
6·4
't. /

II / ~ ~ St'j eneo ide

V V
-- ~ ~~
20 ,nm ns;
A CI.se lution
Oh 106
8 .25n lesP (DPM3 edd
U4,<
.JJ.~ lAA J T S

A I
o 2 3 4 5 6 8 9 10 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 T
ModelR128;60MHz
FIGURE 14-20 N~R spectrum and pea~ integral curve for the organic
FIGURE 14-19 Effect of Pr(DPMh on the NMR spectrum of styrene oxide. Spectrum A:
comp?und CSH1002 In CC14. (From R. M. Silverstein, G. C. Bassler, and T. C.
in the absence of the reagent. Spectrum B: in the presence of the reagent. (Spectra courtesy of
Morrill, Spectrometric Identification of Organic Compounds, 3d ed. New York: John
Perkin-Elmer, Norwalk, Connecticut.)
Wiley & Sons, Inc., 1974, p. 296. With permission.)

sequence of its being closer to the metal ion cal molecules. In addition, however, the meth-
than the other two. od often proves useful for quantitative deter- We obtain relative areas from left to right (3 x 4) at about b = 1.7.Only six are observed,
Similar effects are observed with Eu(DPMh mination of absorbing species. of about 6.1, 4.2, 4.2, and 6.2 from inte- presumably because the resolution of the in-
except that the shifts are to lower fields. Again, gral plot. These figures suggest a distribution strument is insufficient.
the protons closest to the metal ion are affected of the 10 protons of 3, 2, 2, and 3. The
most. Identification of Compounds single peak at b •••3.6 must be due to an iso- EXAMPLE
The primary source of the chemical shift lated methyl group; upon inspection of Fig- The spectra shown in Figure 1~21 are for
An NMR spectrum, like an infrared spec- colorless, isomeric, liquids containing only car-
arises from the secondary magnetic field gen- ure 14-15 and Table 14-3, the functional group
trum, seldom suffices by itself for the identifi- bon and hydrogen. Identify the two com-
erated by the large magnetic moments of CH30q-o)- is suggested. The empirical
cation of an organic compound. However, in pounds.
the paramagnetic praseodymium or europium formula and the 2: 2 : 3 distribution of the
conjunction with other observations such as The single peak at about 0 = 7.2 in the
ions. If the geometry of the complex between remaining protons indicate the presence of
elemental analysis, as well as ultraviolet, in- upper figure suggests an aromatic structure;
the analyte and the shift reagent is known, an n-propyl group as well. The structure
frared,andmassspectra, NMR is an important the relative area of this peak corresponds to 5
reasonably good estimates can be made as to CH3oc(-o)CH2CH2CH3 is consistent with
tool for the characterization of a pure com- protons; from this we conclude that we may
the extent of shift for various protons in the all of these observations. In addition, the po-
pound. The simple' examples that follow give have a monosubstituted derivative of ben-
analyte. sitions and the splitting patterns of the three
some idea of the kinds of information that zene. The seven peaks for the single proton
remaining peaks are entirely compatible with
can be extracted from NMR studies. appearing at 0 = 2.9 and the six-proton dou-
this hypothesis. The triplet at b = 0.9 is typi-
cal of a methyl group adjacent to a methyl- blet at 0 = 1.2 can only be explained by the
EXAMPLE ene. From Table 14-3, the two protons of the structure
Unquestionably, the most important applica- The NMR spectrum shown in Figure 14-20 methylene adjacent to the carboxylate peak CH3
tions of proton NMR spectroscopy have been is for an organic compound having the em- should yield the observed triplet peak at about I
to the identification and structural elucida- b = 2.2. The other methylene group would be -C-CH3
pirical formula CSH1002• Identify the com-
tion of organic, metal-organic, and biQChemi- pound. expected to produce a pattern of 12 peaks ~
weight of72 and containing carbon, hydrogen, APPLICATION OF PROTON NMR
and oxygen only. Identify the compound. TO QUANTITATIVE ANALYSIS
The triplet peak at ~ = 9.8 appears (Figure
14-15) to be that of an aliphatic aldehyde, Quantita1iv8 Analysis
RCHO. If this is the case, R h~ a molecular
A unique ~pect of NMR spectra is the direct
weight of 43, which corresponds to a C3H,
proportionality between peak areas and the
fragment The triplet nature of the peak at
~= 9.8 requires· that there be a methylene number of nuclei responsible for the peak. As
group adjacent to the carbonyl. Thus, the a consequence, a quantitative determination
of a specific compound does not require pure .
compound would appear to be
samples for calibration. Thus, if an identifi-
able peak for one of the constituents of a
sample does not overlap the peaks of the
other constituents, the area of this peak can
The triplet peak at ~ = 0.97 appears to be be employed to establish the concentration of
that of the terminal methyl. The protons on the species directly, provided only that the
the adjacent methylene would be expected to signal area per proton is known. This latter
show a complicated splitting pattern of 12 parameter can be obtained conveniently from

.
o~
~~H'
40 H' 30 H' 2Cl H' '0 Hz H' peaks (4 x 3); the grouping of peaks around a known concentration of an internal stand-
~ ~= 1.7 is compatible with this prediction . ard. For example, if the solvent present in a
.•. -Relat Ite peok a, as- 9
Finally, the peak for the protons on the methyl-
ene group adjacent to the carbonyl should
known amount were benzene, cyclohexane, or
water, the areas of the single proton peak for
appear as a sextet downfield from the other these compounds could be used to give the
methylene proton peaks. The group at ~ = 2.4 desired information; of course, the peak of
is consistent with this conclusion . the internal standard should not overlap with

.-

~
i
..Iii

.-j-----H
i \
FIGURE 14-21 NMR spectra for two organic isomers in CDCI3 solution.
(Courtesy of Varian Associates, Palo Alto, California.)
--t--J
j
!
I:
.._---+-1
I i

Thus, we conclude that this compound is that the compound is C6H3(CH3h . The rela-
cumene. tive peak areas confirm this diagnosis. We

0-,V
cannot, however, decide which of the three
trimethylbenzene derivatives we have from the
9-CH3
NMRdata.
~ CH3
The isomeric compound has an aromatic EXAMPLE
peak at ~ = 6.8; its relative area suggests a The spectrum shown in Figure 14-22 is for FIGURE 14-22 NMR spectrum of a pure organic compound containing C,
trisubstituted benzene, which CdIl only mean an organic compound having a molecular H, and 0 only. (Courtesy of Varian Associates, Palo Alto, California.)
any of the sample peaks. Organic silicon mination of water in food products, pulp and fluoriDe. Fluorine, with an atomic number good proton spectra for materials available in
derivatives are uniquely attractive for calibra- paper. and agricultural materials. The water in of 19, has a spin quantum number of i and a microgram quantities has been difficult, time
tion purposes, owing to the high upfield loca- these substances is sufficiently mobile to give a magnetic moment of 2.6285 nuclear mag- consuming, and sometimes impossible. This
tion of their proton peaks. narrow peak suitable for quantitative measure- netons. Thus, the resonance frequency of fluo- lack of sensitivity also seriously inhibited the
The widespread use of NMR spectroscopy ment,u rine in similar fields is only slightly lower than growthofcarbon-13 NMRspectroscopy. Here,
for quantitative work has been inhibited. by Elemeata1 Analysis. NMR spectroscopy can the proton (56.5 MHz, as compared with the low natural abundance of this isotope and
the cost of the instruments. In addition, the be employed to determine the total concentra- 60.0 MHz at 14,092 gauss). Therefore, with its relatively small magnetogyric ratio provides
probability that resonance peaks will overlap tion of a given kind of magnetic nucleus in a relatively minor changes, a proton NMR spec- an NMR signal that is less by a factor of
becomes greater as the complexity of the sample. For example, Jungnickel and Forbest2 trometer can be adapted to the study of fluorine about 6000 than that for the proton (see
sample increases. Often, too, analyses that are have investigated the integrated NMR inten- resonance. Table 14-2).Thus, producing a useful spectrum
possible by the NMR method can be as con- sities of the proton peaks for numerous orpnic It is found experimentally that fluorine for carbon in unenriched samples required
veniently accomplished by other techniques. compounds and have concluded that accurate absorption is also sensitive to the environ- repeated scans and signal averaging for periods
One of the main problems in quantitative quantitative determinations of total hydrogen ment; the resulting chemical shifts, however, of 24 hr or more.
NMR methods is the result of the saturation in organic mixtures is possible. Paulsen :and extend over a range of about 300 ppm com- The commercial development of pulsed,
effect. As we have pointed out, the NMR ab- Cooket3 have shown that the resonance of pared with a maximum of 20 ppm for the pro- Fourier transform NMR spectrometers by sev-
sorption signal depends upon a very minute fluorine-19 can be used for the quanti~tive ton. In addition the solvent plays a much more eral companies since 1970 has resulted in
excess of nuclei in the lower magnetic energy analysis of that element in an organic rom- important role in determining fluorine peak dramatic increases in the sensitivity of NMR
state and that the absorption process tends to pound-an analysis that is difficult to carry positions than with the proton. measurements; as a result, the routine appli-
depopulate this excess. Whether or not de- out by classical methods. For quantitative Empirical correlations of the fluorine shift cation of this technique to naturally occurring
population has a significant effect on the ab- work, a low-resolution or wide-line spectro- with structure are relatively sparse when com- carbon-I 3 and to protons in microgram quan-
sorption intensity depends upon the relaxation meter can be employed. pared with information concerning proton tities of chemical or biological materials has
time for the species, the power of the source, behavior. It seems probable, however, that the become widespread. The basis of the increased
and the rate at which the spectrum is scanned. future will see further developments in this sensitivity is the same as that discussed in the
Errors arising from saturation can usually be Study of Isotopes field, particularly for structural investigation section on Fourier transform infrared spectros-
avoided through control of these variables. Other Than the Proton of organic fluorine compounds. copy (p. 241). In both, all of the resolution
Aaalysis of Multicompoaent Mixtures. Phosphorus. Phosphorus-31, with spin num- elements of a spectrum are observed in a very
Methods for the analysis of many multicom- Table 14-2 lists several nuclei which, in addi- ber i, also exhibits sharp NMR peaks with brief period by measurements that yield a
ponent mixtures have been reported. For ex- tion to the proton, have magnetic moments chemical shifts extending over a range of 700 time-domain rather than a frequency-domain
ample, Hollis9 has described a method for the and can thus be studied by the magnetic res- ppm. The resonance frequency of 31p at spectrum. A time-domain spectrum can be
determination of aspirin, phenacetin, and caf- onance technique. More than one hundred 14,092 gauss is 24.3 MHz. Several investiga- obtained in a few seconds or less; thus, it
feine in commercial analgesic preparations. other isotopes also possess magnetic tions correlating the chemical shift of the phos- becomes practical to replicate spectra hun-
The procedure requires about 20 min, and the moments14; the resonance behavior of only a phorus nucleus with structure have been re- dreds of thousands of times and average the
relative errors are in the range of 1 to 3 %. few of these has been investigated to date. ported. measurements to give a vastly improved signal-
ChamberlainlO describes a procedure for the Other NucleL Other nuclei that offer con- to-noise ratio. The frequency-domain spectrum
rapid analysis of benzene, heptane, ethylene siderable potential for NMR studies include can then be obtained by a Fourier transform
glycol, and water in mixtures. A wide range of 13C, t70, 2H, ItB, l09Ag, and 29Si; an in- employing a digital computer.
these mixtures was analyzed with a precision lIT. M. Shaw and R. H. Elsken, J. Chern. Phys, 18, 1113 creasing amount of work is being reported for
of 0.5%. (1950); J. Appl. Physia,Z6, 313 (1955); and T. M. Shaw,
each of these. Carbon-13 NMR spectroscopy
An important quantitative application of R. H. Elsken, and C. H. Kunsman, J. Assoc. OfJic.' A.lJr.
Chemists, 36, 1070 (1953~
will be considered in the next section, which
the NMR technique has been to the deter- deals with the application of Fourier trans- IS For a more complete discussion of Fourier transform
12 J. L. Jungnickel and J. W. Forbes, A.nal. Chem.,~, 938
form methods to NMR spectroscopy. NMR spectroscopy, see: E. D. Becker and T. C. Farrar,
(1963~ ' Scinlce, 178, 361 (1972); D. Shaw, Fourier TraMform
up. J. Paulsen aDd W. D. Cooke, Anal. Chern, 36, 1721 NMR SpectrOSCOpy. New York: Elsevier, 1976; R. J.
• D. P. Hollis, AnaL C"-., 35, 1682 (1963~ (1964~ FOURIER TRANSFORM NMR1S Abraham and P. Loftus, Prolan and Carbo •••.13 NMR
Spectroscopy. Philadelphia: Heyden and Sons, 1978; aDd
.0 N. F. ChamberWn, in 1mJtge 011 Analyt/cQl CMm- .4See, for example: J. A. Pople, W. G. Schneider, aDd Conventional NMR spectroscopy is not very T. C. Farrar and E. D. Becker, Pulse and Fouri", TrOllS-
glry, cds. I. M. KoItholF aDd P. J. Elvin.. New York: H. J. Bernstein, High-raolutlon Nuc1«Jr MII(/Mtk Ruo- sensitive. As a consequence, the generation of form NMR. New York: Academic Press, 1971,
Intenc:icDcc, 1963, Part I, voL 4, p. 1932. nance. New York: McGraw-Hili, 1959, pp. 480-485.
and the observation time between pulses is the free induction decay signals emitted by
Pulsed NMR Spectra nuclei as they return to their ground state
approximately 1 s. Figure 14-23a depicts the
In Fourier transform or pulse NMR studies,
the sample is irradiated periodically with
time relations for the input signal.
Time-Domain Decay Spectra. The periodic
from the excited state that was induced by the
pulse of radiation. The frequencies of the
l
:; 0
brief, highly intense pulses of radio-frequency pulses serve the same function as the inter- emitted radiation are, of course, identical to a:>
the absorption frequencieS that would appear o
radiation, following which the free induction ferometer in infrared Fourier transform
decay signal-a characteristic radio-frequency spectroscopy in the respect that they produce in a conventional NMR spectrum for the var-
emission signal stimulated by the irradi- a time-domam spectrum during the period ious types of nuclei present.
Figure 14-23b shows the decay signal when Time"
ation-is recorded as a function of time. between pulses. Two such spectra are shown
Typically, pulses of 1 to 10 IJS are employed in Figure 14-23b and 14-23c. These consist of the sample contains but a 'single type of (a) Time domain decay spectrum

nucleus and the excitation frequency corre-


sponds exactly to the resonance frequency for
that nucleus. Figure 14-23c depicts the decay
signal when more than one type of nucleus is
- .., excited. Here, a more complex time-domain
response is observed because of interference
T among the radiations of differing frequencies.
T
Because the free induction decay curves owe
TK 1 to 10llsee their characteristic appearance to interference Frequency
TK 1 see
among various emitted radiations, it is to be (b) frequency domain emission signal
expected that these time-domain spectra wfll FIGURE 14-24 Spectra for Hp CCI. (a)
be unique and characteristic for every con- Free-induction decay s~rum. (b) Conven-
ceivable combination of frequencies. Thus,just tional spectrum for 13C showing spin-spin
as an infrared time-domain spectrum, obtained coupling between 13C and 'protons. [Adapted
with a Michelson interferometer (p. 244~ con- from T. C. Farrar, Anal. Chern., 42 (4), I09A,
tains all of the information necessary to derive (1970). Reprinted with permission. Copyright
a conventional infrared spectrum, so also does by the American Chemical Society.]
the free induction decay spectrum contain the
information needed to yield a normal fre-
quency-domain NMR spectrum. Conversion
of the data from the time domain to the
frequency domain can be realized with the aid ciently great to excite nuclei with different
of a digital computer programmed to perform resonance frequencies. Fortunately, a suffi-
a fast Fourier transform. Figure 14-24 demon- ciently short pulse of radiation, such as that
strates the relationship between the two types shown in Figure 14-23a, provides a band of
of spectra. The four peaks shown in the fre- frequencies from a monochromatic source. The
quency-domain spectrum for carbon-13 arise frequency range of this band is about tr Hz,
from the coupling between the carbon nucleus where T is the length in seconds of each pulse.
and the three protons. Figure 14-24a indicates Thus, by employing a pulse of 1 IJS with a
the parts of the time-domain spectrum that IS-MHz transmitter, a frequency range of
contain information about the chemical shift, 15 MHz ± 120 kHz would result. This produc-
6, and the coupling constant, J. tion of side bands by pulsing can be under-.
eel Output signII for -.I ~ of nuclei
stood by reference to Figure 4-4 (p. 96), where
FrequetICY Range of • Pulse. Radio-fre-
~IGU~E 14-23 .(a) Input signal for pulsed NMR. (b) Free quency generators, when operated continu- it is shown that a rectangular wave form is
inductIOn decay signal when a single type of nucleus is ously, produce radiation of a single frequency. made up of a series of sine or cosine functions
present. (c) Free induction decay signal when several types The NMR experiments we have just described, differiDg from one another by small frequency
of nuclei are present. , however, require a range of frequencies suffi- increments. Thus, a pulse generated by rapid
on and off switching of a radio-frequency ELECTRON SPIN _::;;C-o
oscillator will consist of an envelope of power RESONANCE SPECTROSCOPY
_:::::CH-O
having a shape somewhat similar to the posi-
tive half of the solid line in Figure 4-4b. Electron spin resonance spectroscopy or ESR CH
----- 2-0
(also called electron paramagnetic resonance -CH.-O
spectroscopy or EPR) is based upon the _:;C-N::::
Carbon-13 NMR'· absorption of microwave radiation by an un-
_::::CH-N:::
Carbon-13 NMR has sev.eral advantages over paired electron when it is exposed to a strong
magnetic field.t7 Species that contain unpaired --CH2-N~
proton NMR in terms of its power to eluci-
electrons, and can therefore be detected by CH3-N:::::
date organic and biochemical structures. First,
there is the obvious advantage that carbon-13 ESR spectroscopy, include free radicals, odd- ------::::::::::C -Hal
NMR provides information about the back- electron molecules, transition-metal com- -----::;::CH -Hal
bone of molecules rather than about the peri- plexes, rare-earth ions, and triplet-state mole-
phery. In addition, the chemical shifts for cules.
_ C = 0 quinone
carbon-13 in a majority of organic compounds
Principles of ESR _C=Oketone ---::::::;::c-cE
is about 200 ppm, compared with approxi-
mately 10 to 20 ppm for the proton; less over-
lap of peaks is the consequence. Thus, for
example, it is often possible to observe individ-
ual carbon resonance peaks for compounds
The principleS of ESR spectroscopy are simi-
lar to those of NMR spectroscopy, which
were discussed in the previous sections. The
aldehyde
C=O-
I
C=OlCid._

I
C=O_ester,amide
I I I
_-C=:N
I I
_-C=:C-
I I
----CH
----:::::cH-cE

I
----cH.-cE
2-
cE

I
electron, like the proton, has a spin quantum
ranging in molecular weight from 200 to 400. 200 160 120 80 40 o
number of one-half and thus has two energy PPM
Also, homonuclear, spilt-spin coupling be- levels that differ slightly in energy under the FIGURE 14-25 Chemical shifts forcarbon-13. (From D. E. Leyden and R. H. Cox, Analytical
tween carbon atoms is itot encountered, be- influence of a strong magnetic field. In con- Applications of NMR. New York: Wiley, 1977, p. 196. With permission.)
cause in unenriched saniples, the probability trast to the proton, however, the lower energy
of two carbon-13 atoms occurring in the same level corresponds to m = -! and the higher
molecule is small. Furthermore, heteronuclear to m = +!; this difference results from the
spin coupling between carbon-13 and carbon- ESR spectrometers often employ a field of up a fluctuating magnetic field (electromag-
IlCptive charge of the electron.
12 does not occur because the spin quantum 3400 G. Substituting this value into Equation netic radiation) in the guide. The wave guide,
Applying Equations 14-4 and 14-5 to an
number of the latter is zero. Finally, good 14-17 gives which is a rectangular metal tube, transmits
unpaired electron yields
methods exist for decoupling the interaction ',fJHo 2.0 x 927 X 10-21 x 3400 the microwave radiation to the sample, which
between carbon-13 atoms and protons. Thus, v·= -h- = 6.63 x 10-27 is generally held in a small quartz tube posi-
AE = hv = J-lfJNHo = gfJNHo (14-17) tioned between the poles of the permanent
generally, the spectrum for a particular type of I
= 9.51 X 109 Hz or 9500 MHz magnet. Helmholtz coils provide a means for
carbon consists of but a single line.
where 9 is the splitting factor and fJN is the varying the field over the small range in which
Figure 14-25 demonstrates some of the Thus, the resonance frequency for an un-
Bohr magneto", which has a value of resonance occurs.
chemical shifts which are observed for paired electron is about 9500 MHz, which lies
927 x 10-21 erg G-1• The value for g varies Generally, ESR spectra are recorded in de-
carbon-13 in various chemical environments. in the microwave region (see Figure 4-7,
with the electron's environment. For a free rivative form to enhance sensitivity and resolu-
As with proton spectra, these shifts are rela- p. 101).
electron, its value is 2,0023; for an unpaired tion. Figure 14-26 contrasts two derivative
tive to tetramethylsilane.;
electron in a molecule or ion, g lies within a spectra with their ordinary absorption counter-
Instrumentation
few percent of this number. parts.
The source of microwave radiation is a klys-
tron tube, which is operated to produce mono-
.6For a thorough discussion of carbon-13 NMR spectros-
chromatic radiation having a frequency of
ESR Spectra
copy, see: J. B. Slathers. Carbon-13 NMR Spectroscopy. 11 For further detai1son ESR spectroscopy, see: M. C. R.
New Yorlc: Academic Press. 1972; D. E. Leyden and Symons, CMrnIcal IIIIIl Bioc"-ical JUpeCfS of Electron- about 9500 MHz. The klystron tube is an elec- Most molecules fail to exhibit an ESR spec-
R. H. Cox, Analytical ApplicDtions of NMR. New Yorlc: Spill RnontJIICe S/1«froscopy. New York: Wiley, 1978; tronic oscillator in which a beam of electrons trum because they contain an even number of
Wiley.1977.ChapterS;and E.Breitmaier,and W. Voelters, and J. E. Wertz and J. R. Bolton, Electron Spin Reso- is pulsed between a cathode and a reflector. The electrons, and the number in the two spin
UC NMRSpectroscopy.2d ed. New Yorlc:VerlagChemie, nance: Elementary Theory and Practical Applications. New oscillating output ofthe klystron is transmitted states is identical (that is, the spins are paired);
1978. Yorlc: McGraw-Hill, 1972
to a wave guide by a loop of wire, which sets as a consequence, the magnetic effects of elec-
tative determination of paramagnetic ions in

1)\ 9995 1‫סס‬oo 10005 Gauss


IL\
9960 1‫סס‬oo 10040 Gauss
biological systems in concentrations as low as
a few parts per billion.
. Another important biological application
mvolves the use of spin-label reagents. These
compounds are stable molecules that contain
(al an odd electron and, in addition, react selec-
tively with certain amino acids or functional

rJ groups in a biological system. The ESR spectra


ofthe product then provide information about
the environment, such as structural features
polarity, viscosity, phase changes, and chemicai
9995 ~OOO5
FIGURE 14-27 The ESR spectrum reactivity.
for the hydrogen atom. The hyperfine
~ 00 splitting is a/gp gauss. (From R. S.
FIGURE 14-26 Comparison of spectral presentations as absorp- Drago, Physical Methods in Chemistry,
tion (a and c) and the corresponding derivative (b and d) curves. 2d ed. Philadelphia: Saunders, 1977,
Note that the shoulders in (c) never reach a maximum; consequently p. 324. With permission.)
the corresponding derivative peaks do not pass through the
abscissa. The number of peaks can be obtained by a count of the
number of maxima or minima (as shown by the asterisks). (From
R. S. Drago, Physical Methods in Chemistry, 2d ed. Philadelphia:
Saunders, 1977, p. 323. With permission.)
Hydroq'uinone and quinone are not ESR
active because neither contains an odd elec-
tron spin are canceled. Such substances are simplest of all free radicals, the hydrogen atom. tron. The postulated intermediate does, how-
diamagnetic because of the small fields induced Here, the spin of the hydrogen nucleus couples ever, and the five-peak hyperfine pattern is
from the orbital precession of the electrons with that of the unpaired electron to produce consistent with a semiquinone radical. Here,
around the nuclei; these fields act in opposition (2 x i x 1 + 1) or 2 peaks. the four ring protons are assumed to be equiv-
to the applied field. In contrast, the spin of an Figure 14-28 illustrates the application of alent with respect to the interactions of their
unpaired electron induces a field that reinforces ESR to the detection of a free radical inter- spins with that of the odd electron; five peaks
the applied field. This paramagnetic effect is mediate in a chemical process. This spectrum result. This picture is reasonable since the
much larger than the diamagnetic behavior. was obtained after the addition of base to a position of the odd electron is not confined
Splitting of the energy levels of an electron solution of hydroquinone in the presence of to the oxygen atom as shown. Indeed, several
occurs in a magnetic field; this splitting can be air. The reaction involves oxidation of the resonance structures can be written in which
observed, as DOted earlier, by microwave ab- anion of hydroquinone to give quinone. That the odd electron is associated with any of
sorption studies. is, several other atoms in the molecule.
The spin of an unpaired electron can couple Ele~ron spin has been widely applied to
with the spin of nuclei in the species to give the study of chemical, photochemical, and
splitting patterns analogous to those observed electr<><;hemicalreactions which proceed via
for nuclear spin-{ipin coupling. When an elec-
tron interacts with n equivalent nuclei, its re-
sonance peak is split into (2nl + 1) peaks,
where 1 is the spin quantum number of the
nuclei and n is the number of equivalent nuclei.
This process is termed hyperfine splitting.
o~o·~o
:0:
H
:0:
H
:B
free radical mechanisms. Often, as in the fore-
going example, the technique has provided
useful structural information about interme-
diate radicals.
ESR studies are also used to obtain struc-
tural information about the transition metals
FIGURE 14-28
(bl

The ESR spectrum for the


semiquinone radical. (a) Derivative spectrum.
(b) Absorption spectrum. [From B. Venkatara.
man and G. K. Fraenkel, J. Amer. Chern. Soc .•
Figure 14-27 shows the spectrum that results and their complexes. In addition, they have 77,2707 (1955). With permission of the Amer-
from hyperfine splitting that occurs in the been used for the detection and semiquanti- ican Chemical Society.]
1. Diagram the high-resolution NMR spectrum expected for the follo.w- 12. From the 6O-MHz spectrum given in Figure 14-40, deduce the struc-
ing compounds. Give chemical shifts, splitting patterns, and relative ture of this compound, whose empirical formula is C4HuNO. Draw
intensities of each peak. the spectrum as it would appear if obtained with a l00-MHz
instrument.
'(a) Toluene, C6H,CH3
(b) Diethyl ether, C2H,OC2H, 13. From the 6O-MHz spectrum given in Figure 14-41, deduce the struc-
(c) Propionaldehyde, CH3CH2CHO ture of this chlorine containing compound. Draw the spectrum as it
(d) Isopropyl chloride, (CH3hCHCl would appear if obtained with a l00-MHz instrument.
(e) 2-Phenylethanol, C6H,CH2CH20H 14. Shown in Figure 14-42 is the NMR spectrum of a common aprotic
2. Diagram the NMR spectrum expected Jor each of the following com- solvent having the empirical formula C3H,NO. When the spectrum of
pounds. Give chemical shifts, splitting patterns, and relative intensities this compound is obtained at a higher temperature, the two peaks at
of each peak. 2.9 6 coalesce into one sharp singlet. What is this compound? Explain
(a) l,2-Dimethoxyethane, CH30CH2CH20CH3 the high-temperature behavior.
(b) 2,3-Dibromobutane, CH3CHBrCHBrCH3 15. It is apparent from the spectra in Figure 14-43 that the sample gives a
very different NMR spectrum when DCI is added to it. What is this
(c) Cyclohexane, I I bromine-containing compound? Explain the effect of the OCt.
CH2CH2CH2CH2CH2CH2
16. From the spectrum shown in Figure 14-44, deduce the structure of
7Hr-<:~ this biologically interesting molecule with the empirical formula'
C,HuN02· Notice that the spectrum has been taken in deuterium
(d) Tetrahydrofuran, CH2 CH2
oxide (D20).
~o/ 17. The IH NMR spectrum of a pleasant-smelling liquid consists of two
singlets. Shown in Figure 14-45, is a proton decoupled 13C NMR
(e) p-Dimethylaminobenzaldehyde, (CH3hN-Q-CHO spectrum of the compound. Deduce the structure of this unknown.

3. From the spectrum given in Figure 14-29, deduce the structure of this
strong-smelling compound with an empirical formula of C3H602 .
4. From the spectrum given in Figure 14-30, deduce the structure of this
hydrocarbon.
r J L • . .. IH 3H

5. From the spectrum given in Figure 14-31, determine the structure of


this compound, which is a commonly used pain 'killer; its empirical 50~ z of et
= 1 .3
J -
formula is CloH13N02.
6. From the NMR spectrum shown in Figure 14-32, determine the struc-
ture of this pleasant-smelling compound, CaHa03 .
7. From the NMR spectrum shown in Figure 14-33, determine the struc-
ture of this easily oxidized compound.
8. Determine the structure of the basic compol!nd whose NMR spec-
trum is shown in Figure 14-34; its empirical formula is CaHuN.
9. Figures 14-35 and 14-36 show the NMR spectra of two isomers of
C4Ha02 . Both of these compounds show a strong absorption at 1735 . Mm I
I
em -1 in their IR spectrum. Identify the two compounds.
10. The NMR spectra of two isomers of C,HloO are shown in Figures 9 8 4 3
14-37 and 14-38. Identify these isomers. They both show a carbonyl I I
peak in their IR spectra. 50ll 200

11. Determine the possible structure(s) for the compound giving the spec- FIGURE 14-29 (From C. J. Pouchert and J. R. Campbell, The Aldrich
trum in Figure 14-39. Can any conclusions be drawn about the differ- Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
ent possible isomers? 1974. With permission.) See Problem 3.
-
3H
9H
'H -

I""" ~~ oM

60 Hz ill
H ').11 pm

5H
tr
.. 11£

0'0 98765432
" 0
'0 9 8 7 6 5 3
...
2_ , 0

1 1 1 1
~ ~ ~ ~ ~ ~ 0
FIGURE 14-32 (From C. J. Pouchert and J. R.Campbell, The Aldrich
FIGURE 14-30 (From C. J. Pouchert and J. R. Campbell, The Aldrich
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company, Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 4. 1974. With permission.) See.Problem 6.

r I 3H

H
( MS< -d, + CO< 13

H
,~
4H
H- ,~ i
, I I puity
/
ill. ..
f
-
J

~o 98765432 0
~o 9 8 6 5 4 3 2[ 0
1 I I I I J I
I I 1 1 I 1
~ ~ ~ ~ ~ ~ 0
~ ~ ~ ~ ~ ~ 0

FIGURE 14-31 (From C. J. Pouchert and J. R. Campbell, The Aldrich FIGURE 14-33 (From C. J. Pouchert and J. R. Campbell, The Aldrich
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission~ See Problem S. ' 1974. With permission.) See Problem 7.
3H

fl
H
H

3H

~ IJ
~I
Ml -
~0 9 8 7 6 6 4 3 2 1 0
I I I I I I I ~0 9 8 7 66432 0
~ ~ ~ ~ ~ ~ 0 I I I 1
FIGURE 14-34 (From C. J. Pouchert and J. R. Campbell, The Aldrich ~ ~ ~ ~ ~ ~ ~
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company, FIGURE 14-36 (From C. J. Pouchert and J. R. Campbell, The Aldrich
1974. With permission.) See Problem 8. Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 9.

3H
H

3fl

H
2 II
H
2
1
U
;

.~ ~
i
~0 ;9 8 7 6 6 4 3 1 0
I I I' 9 8 4 3
~ ~ ~ ~ ~ ~ 0 I I
FIGURE 14-35 (From C. J. Pouchert and J. R. Campbell, The Aldrich ~ ~.

Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company, FIGURE 14-37 (From C. J. Pouchert and J. R. Campbell, The Aldrich
1974. With permission.) See Problem 9. Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 10.
3H

21 lH

. l .-.J ~i I

~0 9 B 7 6 6 4 3 2 1 9 B 7 6 4 3
! ! ! ! I I I 1 J I
600 500 400 300 200 100 0 500 400 200
FIGURE 14-38 (From C. J. Pouchert and J. R. Campbell, The Aldrich FIGURE 14-40 (From C. J. Pouchert and J. R. Campbell, The Aldrich
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Comp,lny, Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 10. .
1974. With permission.) See Problem 12.

3H 1 H 1

lH
H

600 Hz s ift 61 2

~ .1

~o 9 B 7 6 5 4 3 2 0 ~o .9 B 6 6 4 3 2 0
! !!!. I I ! I I ! I !
600 500 400 300 200 100 0 600 500 400 300 200 ~ 0
FIGURE 14-39 (From C. J. Pouchert and J. R. Campbell, The Aldrich FIGURE 14-41 (From C. J. Pouchert and J. R. Campbell, The Aldrich
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company, Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 11. 1974. With permission.) See Problem 13.
0 2 3 4 5 6 7 6 9 10
100

90

80
D. +0 1+ SP

70

60

50

40

30
I~
20
.•.. \. IIAIc
10

0,0~L~9:--L-:!:8-L-l-7---.JL~6 ---l-5~...L-4;-...L--;3H--..l---;2;--..l--:---L~0~ 0,0 9 8 7 6 5 4 3 2 0


1 1 1 I
I 1 1 I 800 600 400 300 200 100 0
600 600 400 300 200 100 0
FIGURE 14-44 (From C. J. Pouchert and J. R. Campbell, The Aldrich
FIGURE 14-42 (From C. J. Pouchert and J. R. Campbell, The A.ldrich
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company, 1974. With permission.) See Problem 16.
1974. With permission.) See Problem 14.

Tile. o DC addd

9 8 7 6 4 3
1 I I
500 400 200
190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 06e
FIGURE 14-43 (From C. J. Pouchert and J. R. Campbell, The Aldrich
FIGURE 14-45 (From L. F. Johnson and W. C. Jankowsk~ Carbon-13 NMR Spectra. New
Library of NMR Spectra. Milwaukee, WI: The Aldrich Chemical Company,
1974. With permission.) See Problem 15. York: Wiley-Interscience, 1972. With permission.) See Problem 17.
Several types of X-ray spectroscopic methods or Bremsstrahlung (meaning radiation that
exist that are similar in many respects to arises from retardation by particles; such ra-
those encountered in optical spectroscopy. diation is generally continuous).
Thus, methods based upon emission (includ- Continuous Spectra from Electron Beam
ing fluorescent emission~ absorption, and dif- Sources. In an X-ray tube, electrons produced
fraction of X-rays now find wide use for at a heated cathode are accelerated toward an
qualitative and quantitative analysis.! anode (the target) by a potential as great as
100 kV; upon collision, part of the energy of
the electron beam is converted to X-rays.
Under some conditions, only a continuous
spectrum such as that shown in Figure 15-1
Before describing the applications of X-rays results; under others, a line spectrum is super-
to analytical problems, it is desirable to con- imposed upon the continuum (see Figure
sider certain theoretical aspects that relate to 15-2).
the emission, absorption, and diffraction of The continuous X-ray spectrum shown in
radiation in the wavelength range between the two figures is characterized by a well-
about 0.1 and 25 A. (The ingstrom. unit A, defined, short-wavelength limit (Ao), which is
which is 0.1 om, was first used because it dependent upon the accelerating voltage V
provided a convenient way of expressing the but independent of the target material. Thus,
wavelengths of X-rays.) Ao for the spectrum produced with a molyb-
denum target at 35 kV (Figure 15-2) is identi-
cal to Ao for a tungsten target at the same
Emission of X-Rays voltage (Figure 15-1~
For analytical purposes, X-rays are obtained The continuous radiation from an electron
in three ways, namely: (I) by bombardment of beam source results from collisions between
a metal target with a beam of high-energy the electrons of the beam and the atoms of
electrons, (2) by exposure of a substance to a
primary beam of X-rays in order to generate
a secondary beam of fluorescent X-rays, and
(3) by employment of a radioactive source
whose decay process results in X-ray emission.
X-Ray sources, like ultraviolet and visible
emitters, often produce both a continuous
and a discontinuous (line) spectrum; both
types are of importance in analysis. Contin-
uous radiation is also called white radiation

1 For a more eX1ensive discussion or the theory and


analytical applications or X-rays, see: E. P. Bertin, in-
troduction to X-Ray Spectrometric Analysis. New York:
Plenum Press, 1978; L S. Birks, X-Ray S~etroelJemiaU
Analysis. New York: Intcrscicnc:e, 1969; H. A. Ucbhafsky,
H. G. Prcilrcr, E. H. Winslow, and P. D. Zemany,
X-Rays, Electrons and Analytic4l CMmlstry. New York:
FIG U R E 15-1 Distribution. of continuous
Wilcy-Intcrscicnc:e, 1972; and R. O. Muller, Spectrochem- radiation from an X-ray tube with a tungsten
ical Analysis by X-Ray FlwreSCDIU. New York: Plenum target. The numbers above the curves indicate
Press, 1972 the accelerating voltages.
Upon substituting numerical values for the 0.1 to 1.0 A even at 50 kV. Characteristic K vacated orbital. As shown in Figure 15-4 the
constants and rearranging, Equation 15-1 be- lines appear at 0.18 and 0.21 A, however, if lines in the K series involve electronic t~si-
the voltage is raised to 70 kV. tions between higher energy levels and the K
comes
Figure 15-3 illustrates the linear relation- shell. The L series of lines results when an
lo = 12,398/V (15-2) ship between the square root of the frequency electron is lost from the second principal
where lo and V have units of angstroms and for a given (K or L) line and the atomic quantum leveL either as a consequence of
volts, respectively. It is of interest to note that number of the element responsible for the ejection by an electron from the cathode or
Equation 15-1 has provided a direct means radiation. This property was first discovered from the transition of an L electron to the K
for the highly accurate determination of by H. G. S. Moseley in 1914. level that accompanies the production of a
Planck's constant. X-Ray line spectra result from electronic quantum of K radiation. It is important to
Characteristic Line Spectra from Eledron transitions that involve the innermost atomic appreciate that the energy scale in Figure 15-4
Beam Sources. As shown in Figure 15-2, orbitals. The short-wavelength K series is pro- is logarithmic. Thus, the energy difference be-
bombardment of a molybdenum target pro- duced when the high-energy electrons from tween the Land K levels is significantly larger
duces intense emission lines at about 0.63 and the cathode remove electrons from those or- than that between the M and L levels. The K
0.71 A; an additional simple series of lines bitals nearest to the nucleus of the target atom lines therefore appear at shorter wavelengths.
FIGURE 15-2 Line spectrum for a tube (in X-ray terminology, the orbital of principal
occurs in the longer wavelength range of 4 to It. is; also important to note that the energy
with a molybdenum target. quantum number n = 1 is called the K shell'
6A. dtfferences between the transitions labeled OCt
The emission behavior of molybdenum is the orbital of quantum number n = 2 is called and OC2 as well as those between Pt and P2 are
typical of all elements having atomic numbers the L shell, and so forth). The collision results so small that only single lines are observed in
the target material. At each collision, an elec- larger than 23; that is, the X-ray line spectra in the formation of an excited ion, which then all but the highest resolution spectrometers
tron is decelerated and a photon of X-ray are remarkably simple when compared with loses quanta of X-radiation as electrons from (see Figure 15-2).
energy is produced. The energy of the photon ultraviolet emission and consist of two series outer orbitals undergo transitions to the 11te energy level diagram in Figure 15-4
will be equal to the difference in kinetic ener- of lines. The shorter wavelength group is would be applicable to any element with suffi-
gies of the electron before and after the colli- called the K series and the other the L series.2 cient electrons to permit the number of transi-
sion. Generally, the electrons in a beam are Elements with atomic numbers smaller than tions shown. The differences in energies be-
decelerated in a series of collisions; the result- 23 produce only a K series. Table 15-1 (p.43O) tween the levels increase regularly with atomic
ing loss of kinetic energy differs from collision presents wavelength data for the emission number because of the increasing charge on
to collision. Thus, the energies of the emitted spectra of a few elements. the nucleus; therefore, the radiation for the K
X-ray photons vary continuously over a con- A second characteristic of X-ray spectra is series appears at shorter wavelengths for the
lID
siderable range. The maximum photon energy that the minimum acceleration voltage re- heavier elements (see Table 15-1~ The effect of
generated corresponds to the instantaneous quired for the excitation of the lines for each nuclear charge is also reflected in the increase
deceleration of the electron to zero kinetic element increases with atomic number. Thus, ~.60
N
in minimum voltage required to excite the
energy in a single collision. For such an event, the line spectrum for molybdenum (atomic 11 spectra of these elements.
E
:>
we may write number = 42) disappears if the excitation volt- c
u
It is important to note that for all but the
age drops below 20 kV. As shown in Figure 'E 4.0 lightest elements, the wavelengths of charac-
he g
hvo = lo = Ve (15-1) 15-1, bombardment of tungsten (atomic num- « teristic X-ray lines are independent of chemi-
ber = 74) produces no lines in the region of cal combination because the transitions re-
where Ve, the product of the accelerating volt- 20 sponsible for these lines involve electrons that
age and the charge on the electron, is the take' no part in bonding. Thus, the position
kinetic energy of all of the electrons in the 2 For the heavier elements, additional series of lines (M, of the K. lines for molybdenum is the same
beam, h is Planck's constant, and c is the ve- N, and so forth) are found at longer wavelengths. Their regardless of whether the target is the pure
locity of light. The quantity Vo is the maxi- intensities are low, however, and little UIC is made of metaL its sulfide, or its oxide.
mum frequency of radiation that can be them. Fluorescent Line Spectra. Another conven-
The designations K and L &rOle from 1M German FIGURE 15-3 Relationship between X-ray ient way of producing a line spectrum is to
produced at voltage V, while lo is the low- words \curtz and \ang for short and long wavelengths.
wavelength limit for the radiation. This rela- emission frequency and atomic number (K t irradiate the element or one of its compounds
The additional alphabetical desipations were then added
tionship is known as the Duane-Hunt law. for lines occurring at progressively longer wavelengths. and Lot lines). • with the continuous radiation from an X-ray
Energy
tube. This process is considered further in a The resulting manganese K,. line at about
levels later section. 2.1 A has proved to be a useful source for
X-ray Quantum
designation states K
Radioactive Sources. X-Radiation occurs both fluorescence and absorption methods.
n I j series in two radioactive decay processes. Gamma
Nv4 2 "2~ rays, which are indistinguishable from X-rays,
N,v 4 2 3/2 owe their production to intranuclear reac-
NIII 4 1 3/2 tions. Electron capture or K capture also pro-
Nil 4 1 1/2 duces X-radiation. This process involves cap-
N, 4 0 1/2
ture of a K electron (less commonly, an L or When a na~row beam of X-rays is passed
an M electron) by the nucleus and formation through a thm layer of matter, its intensity or
ofan element ofthe next lower atomic number. power is generally diminished as a consequence
As a result of K capture, electronic transitions of absorption and scattering. The effect of scat-
Mv 3 2 5/2

M,v 3 2 3/2 to the vacated orbital occur, and the X-ray line tering is ordinarily small and can be neglected
Mill 3 1 3/2
spectrum of the newly formed element is ob- in those wavelength regions where appreciable
Mil 3 1 1/2 served. The half-lives (p. 462) of K-capture absorption occurs. As shown in Figure 15-5,
M, 3 0 1/2
processes range from a few minutes to several the absorption spectrum of an element like
its emission spectrum, is simple and co~sists
J.
flJ "2 ~,.
thousands of years.
Artificially produced radioactive isotopes of a few well-defined absorption peaks. Here
again, the wavelengths of the peaks are char-
Y~3 j' ~12
provide a very simple source of mono-ener-
getic radiation for certain analytical applica- acteristic of the element and are largely in-
2 1 3/2
tions. The best known example is iron-55 dependent of its chemical state.
LIII
2 1 II I I
LII 1/2
which undergoes a K-capture reaction with ~ . A peculiarity of X-ray absorption spectra
2 0 1/2

I
L,
half-life of 2.6 years: IS the appearance of sharp discontinuities,
called absorption edges, at wavelengths im-
5SFe -+ 54Mn + hv mediately beyond absorption maxima.
"2

200

FIGURE 15-4 Partial energy level diagram showing common transitions


L, LIII
leading to X-radiation. The most intense lines are indicated by the widest arrows.
160
:l.

c
.!!
TABLE 15-1 WAVELENGTHS IN ANGSTROM UNITS
OF THE MORE INTENSE EMISSION LINES FOR iu
c
120

SOME TYPICAL ELEMENTS 0


.~
K Series L Series
.•
51 80

,. ,.
.0

Atomic i!
Element Number 1I. 1I. ::;:
40
Na 11 11.909 11.617
K' 19 3.742 3.454
Cr 24 2.290 2.085 21.714 21.323
Rb 37 0.926 0.829 7.318 7.075
Cs 55 0.401 0.355 2.892 2.683 o 0.4 0.6 0.8 1.2
W 74 0.209 0.184 1.476 1.282 Wovelength. A
U 92 0.126 0.111 0.911 0.720
FIGURE 15-5 X-Ray absorption spectra for lead and silver.
'The AlJIorptioo Process. Absorption of an of the corresponding absorption edge, how- between the scattering centers are of the same
X-ray quantum causes ejection of orie of the ever, because absorption requires a complete order of magnitude as the wavelength of the
innermost electrons from an atom and the removal of the electron (that is, ionization~ radiation. Difl'raction is the result.
where x is the sample thickness in centimeters
consequent prciduction of an excited ion .. In whereas emission involves transitions of an Bragg's Law. When an X-ray beam strikes
and P and Po are the powers of the trans-
this process, the entire energy hv of the radIa- electron from a higher energy level within the a crystal surface at some angle 8, a portion is
mitte<l and incident beams. The constant III is
tion is partitioned between the kinetic energy atom. For example, the K absorption edge for scattered by the layer of atoms at the surface.
called the linear absorption coefficient and is
of the electron (the photoelectron) and the silver occurs at 0.485 A, while the K emission The unscattered portion of the beam pene-
characteristic of the element as well as the
potential energy of the excited ion. The high- lines for the element have wavelengths at 0.497 trates to the second layer of atoms where
number of its atoms in the path of the beam.
est probability for absorption arises when the and 0.559 A. When fluorescence is to be excited again a fraction is scattered, and the remain-
A more convenient form of Beer's law is by radiation from an X-ray tube, the operat- der passes on to the third layer. The cumula-
energy of the quantum is exactly equal to
the energy required to remove the electron Po ing voltage must be sufficiently great so that tive effect of this scattering from the regularly
just to the periphery of the atom (that is, as
In p = IlPX (15-3)
the cutoff wavelength Ao (Equation 15-2) is spaced centers of the crystal is a diffraction of
the kinetic energy of the ejected electron ap- shorter than the absorption edge of the ele- the beam in much the same way as visible
where P is the density of the sample and Il is ment whose spectrum is to be excited. Thus, radiation is diffracted by a reflection grating
proaches zero).
the mass absorption coefficient, a quantity that to generate the K lines for silver, the tube (p. 129). The requirements for diffraction are:
The absorption spectrum for lead, shown
is independent of the physical and chemical voltage would need to be (1) the spacing between layers of atoms must
in Figure 15-5, exhibits four peaks, the first
states of the element. Thus, the mass absorp-
occurring at 0.14 A. The energy of the quan- be roughly the same as the wavelength of the
tion coefficient for bromine has the same 12398 -3
tum corresponding to this wavelength exactly V ~ 0.485 x 10 = 25.6 kV radiation and (2) the scattering centers must
value in gaseous HBr as in solid sodium bro-
matches the energy required to just eject the be spatially distributed in a highly regular
highest energy K electron of the element; im- mate. way.
mediately beyond this wavelength, the energy Mass absorption coefficients are additive In 1912, W. L. Bragg treated the diffraction
functions of the weight fractions of elements Diffraction of X-Rays
of the radiation is insufficient to bring about of X-rays by crystals as shown in Figure 15-6.
removal of a K electron, and an abrupt de- contained in a sample. Thus, In common with other types of electro-mag- Here, a narrow beam strikes the crystal sur-
crease in absorption occurs. At wavelengths JlM= WAllA+ WBIlB+ Welle + ... (15-4) netic radiation, interaction between the electric face at angle 8; scattering occurs as a con-
lower than 0.14 A, the probability of interaction vector of X-radiation and the electrons of the sequence of interaction of the radiation with
between the electron and the radiation dimin- where JlMis the mass absorption coefficient of matter through which it passes results in scat- atoms located at 0, P, and R. If the distance
ishes and results in a smooth decrease in a sample containing the weight fractions WA, tering. When X-rays are scattered by the or·
WB, and ~ of elements A, B, and C. The AP+ PC= n),
absorption. In this region, the kinetic energy dered environment in a crystal, interference
of the ejected photoelectron increases con- terms IlA, ~, and Ile are the respective mass (both constructive and destructive) takes place where n is an integer, the scattered radiation
tinuously with the decrease in wavelength. absorption coefficients for each of the ele- among the scattered rays because the distances will be in phase at OeD, and the crystal will
The additional peaks at longer wave- ments.
lengths correspond to the removal of an elec-
X-Ray Fluorescence
tron from the L energy levels of lead. Three
sets of L levels, differing slightly in energy, The absorption of X-rays produces electron-
exist (see Figure 15-4); three peaks are, there- ically excited ions that return to their ground
fore, observed. Another set of peaks, arising state by transitions involving electrons from
from ejections of M electrons, will be located higher energy levels. Thus, an excited ion
at still longer wavelengths. with a vacant K shell is produced when lead
Figure 15-5 also shows the K absorption absorbs radiation of wavelengths shorter than
edge for silver, which occurs at 0.485 A. The 0.14 A (Figure 15-5); after a brief period, the
longer wavelength for the silver peak reflects ion returns to its ground state via a series
the lower atomic number of the element com- of electronic transitions characterized by the
pared with lead. emission of X-radiation (fluorescence) of wave-
'The Mass AlJIorptioo Coefficient. Beer's lengths identical to those that result from
law is as applicable to the absorption of X- excitation produced by electron bombardment.
radiation as to other types of electromagnetic The wavelengths of the fluorescent lines are R

radiation; thus, we may write always somewhat greater than the wavelength FIGURE 15-6 Diffraction of X-rays by a crystal.
appear to reftect the X-radiation. But it is spectrum. Here, isolation can be achieved elec- Cooling Wltar
readily seen that tronically with devices that have the power to In
tun~ten target (Figure 15-1) could be used to
discriminate between various parts of a spec- ex~te the K,. and K, lines of molybdenum
AP=PC=dsin8 trum based on the energy rather than the wave- (Figure 15-2). The resUlting fluorescence spec-
where d is the interplanar distance of the crys- length of the radiation. Thus, X-ray instru- t~m would then be similar to the spectrum in
tal. Thus, we may write that the conditions ments are often described as wavelength dis- Figure 15-2 except that the continuum would
for constructive interference of the beam at persive instruments .or energy dispersive instru- be removed.
angle 8 are ments, depending upon the method by which
they resolve spectra. Filters for X-Ray Beams
nA. = 2d sin 8 (15-6)
In many applications, it is desirable to employ
Equation 15-6 is called the Bragg equation
an X-ray beam that is restricted in its wave-
and is of fundamental importance. Note that Sources
length range. As in the visible region, both
X-rays appear to be reflected from the crystal Three types of sources are encountered in X-ray Tungsten
filament
filters and monochromators are used for this
only if the angle of incidence satisfies the con- instruments, namely, Coolidge tubes, radioiso- (cathode) purpose. .
dition that topes, and secondary fluorescent sources.
Figure 15-8 illustrates a COmmon technique
. 8 nA. The Coolidge Tube. The most common
for producing a relatively monochromatic
sm =2d source of X.rays for analytical work is the beam by use of a filter. Here, the k., line and
Coolidge tube, which can take a variety of most of the continuous radiation from the
At all other angles, destructive interference shapes and forms. Basically, however, it is a emission of a molybdenum target is removed
occurs. highly evacuated tube in which is mounted a
by a zirconium filter having a thickness of
tungsten filament cathode and a massive anode
FIGURE 15-7 Schematic diagram of the about 0.01 em. The pure K. line is;then avail-
constructed of such metals as tungsten, copper,
Coolidge tube. able for analytical purposes. Several other
molybdenum, chromium, silver, nickel, cobalt, target-filter combinations of this f type have
rhodium or iron (Figure 15-7). Separate cir- been developed, each of which serves to isolate
Absorption, emission, fluorescence, and dif- cuits are used to heat the filament and to
fraction of X-rays all find applications in ana- accelerate the electrons to the target. The
lytical chemistry. Instruments for these appli- heater circuit provides the means for control- The best radioactive sources provide simple
cations contain components that are analogous ling the intensity of the emitted X-rays while
line s~ra. Because ofthe shape of X-ray ab-
in function to the five components of instru- the accelerating potential determines their SOrptIOncurves, a given radioisotope will be Emission by
ments for optical spectroscopic measurement; energy or wavelength. The Coolidge tube is suitable for excitation of fluorescence or for molybdenum Ka
these components include a source, a device target
normally self-rectifying, and a high-voltage ac absorption studies for a range of elements. For
for restricting the wavelength range to be em- source is connected directly to the cathode to example, a source producing a line in the region
ployed, a sample holder, a radiation detector provide the accelerating potential between 0.3 and 0.47 A would be suitable for
or transducer, and a signal processor and read- The production of X-rays by electron bom- fluorescence or absorption studies involving A
out. These components differ considerably in bardment is a highly inefficient process. Less the K absorption edge for silver (see Figure II
detail from their optical counterparts. Their than one percent of the electrical power is I I
15-5). Sensitivity would, of course, improve as I I
functions, however, are the same, and the ways converted to radiant power, the remainder / Kp I
the wavelength of the source line approaches Absorption I I
in which they are combined to form instru- being degraded to' heat. As a consequence, the absorption edge. lodine-125 with a line at by / I'
ments are often similar to those shown in water cooling of the anodes of X-ray tubes is I
0.46 A would be ideal from this standpoint. zirconium ~ i

Figure 5-1 (p. 115). required. filter // I


Secondary Fluorescent Sources. In some
As with optical instruments, both X-ray Radioisotopes. A variety of radioactive sub- // I
applications, the fluorescence spectrum of an
photometers and spectrophotometers are en- stances have been employed as sources in X-ray
---'
//
element that has been excited by radiation
countered, the first employing filters and the fluorescence and absorption methods. Gen-
from a Coolidge tube serves as a sourc:'e for
second monochromators for restricting radia- erally, the radioisotope is encapsulated to
absorption or fluorescence studies. This ar.
tion from the source. In addition, however, a prevent contamination of the laboratory and rangement has the advantage of eliminating the
third method is available for obtaining infor- shielded to absorb radiation in all but certain continuous component emitted by a primary
mation about isolated portions of an X-ray directions. FIGURE 15-8 Useofafiltertopro-
source. For example, a Coolidge tube with a
duce monochromatic radiation.
collimated incident beam. From Equation 15- by pumping.
one of the intense lines of a target element. Wavelength . 6, it is evident that any given angular setting The loss of intensity is high in a mono-.
Monochromatic radiation produced in this Dispersion of the goniometer, only a few wavelengths are chromator equipped with a flat crystal be-
with Monochromators
way is widely used in X-ray diffraction studies. diffracted (l, 112,1/3,... , lfn, where 1= 2d sin cause as much as 99% of the radiation is:
The choice of wavelengths available by this 9). Thus, an X-ray monochromator does not sufficiently divergent to be absorbed in the
Figure 15-9 shows the essential components
technique is limited by the relatively small disperse an entire spectrum simultaneously as collimators. Increased intensities, by as much
of an X-ray spectrometer. The m~no-
number of target-filter combinations that are does a grating or prism; instead, a particular as a factor of ten, have been realized by em-
chromator consists of a pair of beam co~-
available. wavelength is diffracted only when the ploying a curved crystal surface which acts
tors which serve the same purpose as the sbts
Filtration of the continuous radiation from goniometer is set at an appropriate angle. not only to diffract but also to focus the
in ~ optical instrument, and a dispersing
a Coolidge tube is also feasible with. t.hin In order to derive a spectrum, it is neces- divergent beam from the source upon the exit
element. The latter is a single crystal mounted
strips of metal. As with glass filters for VISible sary that the exit beam collimator and the collimator.
on a goniometer or rota~ble table. t~t per-
radiation, relatively broad bands are prod~ced detector be mounted on a second table that As illustrated in Table IS-I, most analyt-
mits variation and prCClSCdetermmatlon of
with a significant loss in intensity of the desired rotates at twice the rate of the first; that is, as ically important X-ray lines lie in the region
the angle 9 between the crystal face and the
wavelengths. . the crystal rotates through an angle 9, the between about 0.1 and 10 A. A consideration
detector must simultaneously move through of data in Table 15-2, however, leads to the
an angle 29. Clearly, the interplanar spacing conclusion that no single crystal satisfactorily
d for the crystal must be known precisely disperses radiation over this entire range. As
(Equation 15-6). a consequence, an X-ray monochromator mUst
The collimators for X-ray monochromators be provided with at least two (and preferably
ordinarily consist of a series of closely spaced more) interchangeable crystals.
metal plates or tubes that absorb all but the The useful wavelength range for a crystal is
parallel beams of radiation. determined by its lattice spacing d and the
X-Radiation longer than about 2 A is ab- problems associated with detection of the
sorbed by constituents of the atmosphere. radiation when 29 approaches zero or ISO deg.
Therefore, provision is usually made for a con- When a monochromator is set at angles of 29
tinuous flow of helium through the sample that are much less than 10 deg, the amount of
compartment and monochromator when long- polychromatic radiation scattered from the
er wavelengths are required. Alternatively, pro- surface becomes prohibitively high. Generally,
visions may be made to evacuate these areas values of 29 greater than about 160 deg can-

TABLE 15-2 PROPERTIES OF TYPICAL DIFFRACTING


CRYSTALS

Wavelength Range-, A Dispersion d9/dl, degJA


Lattice •
Crystal Spacing d, A A.....x l".in at A.....x at l".in

TOrz 1.356 2.67 0.24 2.12 0.37


Li 2.014 3.97 0.35 1.43 025
FIGURE 15-9 An X-ray monochromator and de- NaCI 2.820 5.s5 0.49 1.02 0.18
EDD'r 4.404 8.67 0.77 0.65 0.11
tector. Note that the angle of the detector with respect 0.54
ADP< 5.325 10.50 0.93 0.09
to the beam (29) is twice that of the crystal face. For
absorption analysis, the source is an X-ray tube and
the sample is located in the beam as shown. For • Based on assumption that the measurable
• Ethylenediamine d-tartrate.
range of 28 is from 160 dell for 1-. to 10 deg for .t...
emission work, the sample becomes a fluorescent
, Ammonium dihydrogen phosphate.
source of X-rays as shown in the insert.
counting is also beginning to find use in ultra- trons (ion pairs) for each X-ray quantum. Three a consequence, the number of electrons reach-
not be measured because the location of the
violet and visible spectroscopy. types of X-radiation detectors, namely, ioniza- ing the anode is smaller than the number
source unit prohibits positioning of the detec-
PhotOll Couatmg. In contrast to the various tion chombers, proportional counters, and Geiger produced initially by the incoming radiation.
tor at such an angle (see Figure 15-9). The
minimum and maximum values for A",.,. in photoelectric detectors we have thus far con- tubes, are based upon the enhanced conduct- In the region between VI and "2, the
sidered, X-ray detectors are usually operated ivity resulting from this phenomenon. number of electrons reaching the anode is
Table 15-2 were determined from these limita-
as photon counters. In this mode, the individual A typical gas-filled detector is shown sche- reasonably constant and represents the total
tions.
It will be seen from Table 15-2 that a crystal pulse of electricity produced as a quantum of matically in Figure 15-10. Radiation enters the number formed by a single photon.
such as ammonium dihydrogen phosphate, radiation is absorbed by the transducer is chamber through a transparent window of . In the region between V3 and V4, the number
with a large lattice spacing, has a much greater counted; the power of the beam is then re- mica, beryllium, aluminum, or Mylar. Each of electrons increases rapidly with applied
wavelength range than a crystal in which this corded digitally in terms of number of counts photon of X-radiation may interact with an potential. This increase is the result of second-
parameter is small. The advantage of large per unit of time. This type of operation requires atom of argon, causing it to lose one of its ary ion-pair production caused by collisions
values of d is offset, however, by the consequent rapid response times for the detector and signal outer electrons. This photoelectron has a large between the accelerated electrons and gas
lower dispersion. This effect can be seen by processor with respect to the rate at which kinetic energy, which is equal to the difference molecules; amplification (gas amplification) of
differentiation of Equation 15-6, which leads to quanta are absorbed by the transducer; thus, between the X-ray photon energy and the the ion current results.
photon counting is applicable only to beams binding energy of the electron in the argon In the range V, to V6, amplification of the
dO D of relatively low intensity. As the beam inten- atom. The photoelectron then loses this excess electrical pulse is enormous but is limited by
dl = 2d cos:O sity increases, the pulse rate becomes greater kinetic energy by ionizing several hundred ad- the positive space charge created as the faster
than the response time of the instrument, and ditional atoms of the gas. Under the influence moving electrons migrate away from the slower
Here, dO/dl, a measure of dispersion, is seen
only a steady-state current, which represents of an applied potential, the mobile electrons positive ions. Because ofthis effect, the number
to be inversely proportional to d. Table 15-2
an average number of pulses per second, can migrate toward the central wire anode while of electrons reaching the anode is independent
provides dispersion data for the various crys-
be measured. the slower moving cations are attracted toward of the type and energy of incoming radiation
tals at their maximum and minimum wave-
For weak sources of radiation, photon the cylindrical metal cathode. and is governed instead by the geometry and
lengths. The low dispersion pf ammonium di-
counting generally provides more accurate in- Figure 15-11 shows the effect of applied gas pressure of the tube.
hydrogen phosphate prohi~its its use in the
tensity data than are obtainable by averaging potential upon the number of electrons that Figure 15-11 also illustrates that a larger
region oflow wavelengths; nere, a crystal such
the pulses and measuring the resulting current. reach the anode of a gas-filled detector for number of electrons is produced by the more
as topaz or lithium fluoride must be sub-
The improvement results from the tendency of each entering X-ray photon. Several charac- energetic, o.6-A radiation than the longer
stituted.
signal pulses to be larger than the pulses arising teristic voltage regions are indicated. At po- wavelength, 5-A X-rays. Thus, the size of the
from bal:kground noise in the source, detector, tentials less than VI' the accelerating force on pulse (the pulse height) is greater for the former
and associated electronics; separation of the the ion pairs is low, and the rate at which the than the latter.
X-R.y Detectors two can then be achieved with a pulse height positive and negative species separate is in- ne Geiger Tube. The Geiger tube is a gas-
.net Signal Processors discrimilUltor, an electronic device that will be sufficient to prevent partial recombination. As filled detector operated in the voltage region
Early X-ray equipment employed photograph- discussed in a later section.
ic emulsions for detection and measurement of Photon counting is used in X-ray work be-
radiation. For reasons of convenience, speed, cause the power of available sources is often
and accuracy, however, modern instruments low. (Photon counting has also been profit-
are generally equipped with detectors that ably applied to weak sources of ultraviolet and
convert radiant energy into an electrical signal. visible radiation such as those encountered in
Three types of transducers; are encountered: Raman spectrometry.) In addition, photon
gas-filled detectors, scintillation counters, and counting permits spectra to be obtained with-
semiconductor detectors. ~efore considering out the use of a monochromator. This property
how each of these devices fu~ctions, it is worth- is considered in the section devoted to energy
while to discuss photon counting, a signal pro- . dispersive systems.
cessing method, which is commonly employed Gas-Filled Detectors. When X-radiation pas-
with X-ray detectors as well as detectors of ses through an inert gas such as argon, xenon,
radiation from radioactive sources (Chapter or krypton, interactions occur that produce a
16). As was mentioned earlier (p. 146), photon large number of positive gaseous ions and elec-
represents an upper limit in the response capa- tion chambers are not employed in X-ray spec- pulse-height analyzer to monitor the output of
bility of the tube. Typically, the dead time of a trometry because of their lack of sensitivity. a scintillation counter forms the basis of energy
Geiger tube is in the range from SO .to 200 JIS. Sciatillatioo COUIIters.The luminescence dispersive photometers, to be discussed later.
Geiger tubes are usually filled WIth argon; produced when radiation strikes a phosphor In addition to sodium iodide crystals, a
a low concentration of an organic substance, represents one of the oldest methods of detect- number of organic scintillators such as stil-
I often alcohol or methane (a quinch gas1 is ing radioactivity and X-rays, and one of the bene, anthracene, and terphenyl have been
I
I I
also present to minimize the production of newest as well. In its earliest application, the used. In crystalline form, these compounds
I I secondary electrons when the cations strike technique involved the manual counting of have decay times of om to 0.1 JIS. Organic
I I
the chamber wall. The lifetime of a tube is flashes that resulted when individual photons liquid scintillators have also been developed
I I V.
0.6 A X·rays - __ / : limited to some 108 to 109 counts, by which or radiochemical particles struck a zinc sulfide and are used to advantage because they ex-
time the quench gas has been depleted. screen. The tedium of counting individual hibit less self-absorption of radiation than do
I I
I I With a Geiger tube, radiation intensity is flashes by eye: led Geiger to the development solids. An example of a liquid scintillator is a
~ I determined by a count of the pulses of current. of gas-filled qetectors, which were not only solution of p-terphenyl in toluene.
II I The device is applicable to all types of nuclear Semiconductor Detectors. Semiconductor
I I Proportional more convenient and reliable, but more respon-
--,- co~nter and X-radiation. However, it lacks the large sive to radiation as well. The advent of the detectors have assumed major importance as
I regIon
counting range of other detectors because of photomultiplij:r tube (p. 140) and better phos- detectors of X-radiation. These devices are
I
I its relatively long dead time; its use in X-ray phors, however, has reversed this trend, and sometimes called lithium drifted silicon or ger"
V. spectrometers is, therefore, limited. scintillation counting has again become one of manium detectors.
I Ionization Proportional Counters. The proportional the important methods for radiation detection. Figure 15-12 illustrates one form of a lith-
-.--cha.mber
Vz regton
counter is a gas-filled detector that is operat~ The most ~dely used modern scintillation ium drifted detector, which is fashioned from
in the V3 to V. voltage region of Figure 15-11. detector consists of a transparent crystal of a wafer of crystalline silicon. Three layers exist
Here, the pulse produced by a photon is ampli- sodium iodide that has been activated by the in the crystal, a p-type semiconducting layer
Applied potential. V fied by a factor of 500 to 10,000, but the introduction pf perhaps I % thallium. Often, that faces the X-ray source, a central intrinsic
FIGURE 15-11 Gas amplification for vari- number of positive ions produced is small the crystal is shaped as a cylinder that is 3 to zone, and an n-type layer. The outer surface
ous types of gas-filled detectors. enough so that the dead time is only about 4 in. in each dimension; one of the plane sur- of the p-type layer is coated with a thin layer
1 JIS. In general, the pulses from a proportional faces then faces the cathode of a photo- of gold for electrical contact; often, it is also
counter tube must be amplified before being multiplier tube. As the incoming radiation covered with a thin beryllium window which
counted. traverses the crystal, its energy is first lost to is transparent to X-rays. The signal output is
between V, and V6 in Figure 15-11; here, a gas The number of electrons per pulse (the the scintillator;. this energy is subsequently taken from an aluminum layer which coats the
amplification of greater than 109 occurs. Each ,,"be height) produced in the proportional released in the form of photons of fluorescent n-type silicon and is fed into a preamplifier
photon produces an avalanche of electrons region depends directly upon the energy of radiation. Several thousand photons with a with an amplification factor of about 10. The
and cations; the resulting currents are thus the incoming radiation. A proportional count- wavelength of about 400 nm are produced by preamplifier is frequently a field-effect tran-
large and relatively easy to detect and measure. er can be made sensitive to a restricted range each primary particle or photon over a period sistor which is made an integral part of the
The conduction of electricity through a of X-ray frequencies with a pulse-height analyz- of about 0.25 JIS (the decay time). The dead detector.
chamber operated in the Geiger region (and er, which counts a pulse only if its amplitude time of a scintillation counter is thus signifi- The detector and preamplifier must be ther-
in the proportional region as well) is not con- falls within certain limits. A pulse-height analy- cantly smaller than the dead time of a gas-filled mostated at the temperature of liquid nitrogen
tinuous because the space charge mentioned zer in effect permits electronic filtration of detector. (-196°C) to decrease electronic noise to a
earlier, terminates the flow of electrons to the radiation; its function is analogous to that of a The flasheS of light produced in the scintil- tolerable level. Furthermore, the performance
anode. The net effect is a momentary pulse of monochromator. lator crystal are transmitted to the photo- of the detector degrades badly if allowed to
current followed by an interval during which Proportional counters have been widely cathode of tq<: photomultiplier tube and are come to room temperature, owing to the ten-
the tube does not conduct. Before conduction used as detectors in X-ray spectrometers. in turn converted to electrical pulses that can dency oflithium to diffuse rapidly in the silicon.
can again occur, this space charge must be Ionization Chambers. Ionization chambers be amplified and counted. An important char- Thus, the liquid nitrogen cryostat is connected
dissipated by migration of the cations to the are operated in the voltage range from VI to acteristic of scintillators is that the number to a large ( ..•.20 liter) Dewar flask, which must
walls of the chamber. During the dead time, V2 in Figure 15-11. Here, the currents are of photons produced in each flash is approxi- be filled every few days.
when the tube is nonconducting, response to small (10-13 to 10-16 A typically) and rela- mately proportional to the energy of the in- A lithium drift detector is formed by depos-
radiation is impossible; the dead time thus tively independent of applied voltage. Ioniza- coming radiation. Thus, incorporation of a iting lithium on the surface of a p-doped sili-
Signel ProcesS •• Energy Dispersive Systems
end Reedout DeVices
p-typeSi A discriminator is an electronic circuit that re-
The signal from the prea,mplifier of an X-ray sponds only to pulses having voltage heights
r--\------------- --------, I
I
spectrometer is fed into a linear fast response
amplifier whose amplification can be varied by
above some preset minimum value. All modern
solid-state X-~y spectrometers (wavelength
II a factor up to 10,000. The result is a voltage
AI contact dispersive as well as energy dispersive) are
1 pulse as large as 10 V. equipped with discriminators that reject pulses
I
I Counters and Scalers: For low counting of.about 0.5 V or less (after amplification). In
1
rates (500 to 1000 counts/min~ a simple thIS way, detector and amplifier noise is re-
1
II electromechanical counter suffices. One or duced significantly.
Preamplifier more scalers are required for higher counting A pulse height analyzer or selector is an
I
--+ Liquid N2
cryostat
1 (77°K)
rates, however. A scaler is an electronic device
arranged so that its output terminals transmit
electronic circuit that rejects all pulses with
____________ JI
heights below some predetermined minimum
only some fixed fraction of the total number level and above a preset maximum level; that
n-typeSi of input pulses; that is, every second, every is, it rejects all pulses except those that lie
fourth, every eighth, and so on. This device within a limited channel or window of pulse
FIGURE 15-12 Vertical cross section of a lithium reduces the pulses to a sufficiently small heights. Figure 15-13 provides a schematic
drifted silicon detector for X-rays and radioactive number to be accommodated by the counter. diagram of a pulse height analyzer and its
radiation.

X-ray
con crystal. Upon heating to 400 to soooC, a current pulse accompanies the absorption photons
E3
the lithium diffuses into the crystal; because of each photon. In common with a propor- E2
this element easily loses electrons, its presence tional detector, the size of the pulse is directly E,

converts the p region to an n-type. While still proportional to the energy of the absorbed
at an elevated temperature, a dc potential is photons. In contrast to the proportional detec-
applied across the crystal to cause withdrawal tor, however, secondary amplification of the
of electrons from the lithium layer and holes pulse does not occur.
Distriblltioa of P8IBe Heights from X-Radia-

-L{~.4"_;cm.
from the p-type layer. Current passage across
the np junction requires migration (or drift- tion Detectors. To understand the properties
ing) of lithium ions into the p layer and for- of energy dispersive spectrometers, it is im- .- .
~"i
. - circuit counter
mation of the intrinsic layer where the lithium portant to appreciate that the size of current Lower
• . dlIcriIoinator
ions replace the holes lost by conduction. pulses resulting from absorption of successive
Upon cooling, this central layer has a high X-ray photons of identical energy by the detec- Window
resistance relative to the other layers because tor will not be exactly the same. Variations
the lithium ions in this medium are less
mobile than the holes they displaced.
arise because the ejection of photoelectrons
and their subsequc;nt generation of conduction
-/--- --V+AV
--V
2
3 Signals
from
upper
discriminator
The intrinsic layer of a silicon detector electrons are random processes governed by 123 2
functions in a way that is analogous to argon the probability law. Thus, a Gaussian distribu- Amplified Signal to
in the gas-filled detector. Initially, absorption tion of pulse heights around a mean is observed. signals scaler
from E,• E", & E3
of a photon results in formation of a highly The breadth of this distribution varies from Inverted signal
energetic photoelectron, which then loses its one type of detector to another, with the drift from
lower di.:riminator
kinetic energy by elevating several thousand detector providing a significantly narrower
electrons in the silicon to a conduction band; band of pulse heights. It is this property that FIG U RE 15~13 Schematic diagram of a signal heigh~ analyzer. Lower
a marked increase in conductivity results. has made lithium drift detectors so important plot shows heIght of transmitted signals upon exit from various electronic
When a potential is applied across the crystal, components.
for energy dispersive X-ray spectroscopy.
method of operation. The output pulses from X-rays from a Coolidge tube or a radioactive ments' In this application, the crystal and Both multichannel and single-channel in-
the detector and preamplifier are further source. Under these circumstances, the ele- detector are set at the proper angles (9 and 29) struments are equipped to handle samples in
amplified and appear as voltage signals (in ments in the sample are excited by absorption and counting is continued until sufficient the form of metals, powdered solids, evap-
the lo-V range). These signals are fed into the of the primary beam and emit their own counts have accumulated for an accurate orated films, pure liquids, or solutions. Where
linear pulse-height analyzer, the first stage of characteristic fluorescent X-rays. This pro- analysis. Automatic instruments are much necessary, the materials are held in a cell with
which consists of two discriminator circuits. cedure is thus properly called an X-ray fluo- more convenient for qualitative analysis, where a Mylar or cellophane window.
Each discriminator can be set to reject any rescence or emission method. X-Ray fluores- an entire spectrum must be swept. Here, the
signal below a certain voltage. As shown in cence is perhaps the most widely used of all electric drive for the crystal and detector is
the lower part of Figure 15-13, the upper analytical methods for the qualitative identi- synchronized with the motor of a recorder Energy
discriminator rejects signal 1, which is smaller fication of elements having atomic numbers while the detector output determines the posi- Dispersive Instruments
than V in voltage, but transmits signals 2 and greater than oxygen (>8); in addition, it is tion of the pen.
often employed for semiquantitative or quanti- As shown in Figure 15-14, an energy disper-
3. The lower discriminator, on the other hand, Most modern single-channel spectrometers
sive spectrometer consists of a polychromatic
is set to V + fJ. V and thus rejects all but signal tative determination of these elements. are provided with two X-ray sources; typically,
3. In addition, the lower circuit is so arranged one has a chromium target for longer wave-
that its output signal is reversed in polarity Instruments lengths and the other a tungsten target for
and thus cancels out signal 3 from the upper shorter. For wavelengths longer than 2 A. it is Hein
VarioUs combinations ofthe instrument com-
circuit in the anticoincidence circuit. As a
consequence, only signal 2, with a voltage in ponents discussed in the previous section lead
necessary to remove air between the source
and detector by pumping or by displacement ~ Sample t
the range fJ.V. reaches the counter. to several recognizable types of X-ray fluo- with a continuous flow of helium. A means
Pulse height anal~ are either single- rescence instruments. The three basic types are must also be provided for ready interchange
or multiple-channel devices. A single-channel wave~ngth dispersive, energy dispersive, and of dispersing crystals.
analyzer typically has a voltage range of per- nondispersive; the latter two can be further Recording single-channel instruments cost
haps 10 V or more with a window of 0.1 to subdiVided depending upon whether a Coo- approximately $40,000.
0.5 V. The window can be manually or auto- lidge tube or a radioactive substance serves as Multichannel instruments are large, expen- AmPlI,.•••.
matically adjusted to scan the entire voltage a spectral source. sive (- $150,(00) installations that permit the
range, thus in effect providing data for energy W.,eJenctb Dispersive 1nItnments. Wave- simultaneous detection and determination of
dispersion. length dispersive instruments always employ as many as 24 elements. Here, individual chan- ~"',i~~~
" heWlt_Ivzer "/"
Multichannel analyzers contain as few as tubes as a source because of the large energy nels consisting of an appropriate crystal and a ~l~' .,,;;~.....i;.

two or as many as several hundred separate losses suffered when an X-ray beam is colli- detector are arranged radially around an X-ray
channels, each of which acts as a single chan- mated and dispersed into ita component wave- source and sample holder. Ordinarily, the crys-
nel that is set for a different voltage span or lengths. Radioactive sources produce X-ray tals for all or most of the channels are fixed
width. Such an arrangement permits simul- photons at a rate less than 10-4 that of a at an appropriate angle for a given analyte
taneous counting and recording of an entire Coolidge tube; the added attenuation by a line; in some instruments, one or more of the
spectrum. monochromator would then result in a beam crystals can be moved to permit a spectral Radioactive
source in
that was difficult or impossible to detect and scan. annular
ring
measure accurately. Each detector in a multichannel instru-
Wavelength dispersive instruments are of ment is provided with its own amplifier, pulse
X·RAY two 'types, single-channel or sequentia~ and height selector, scaler, and counter or integra-
FLUORESCENCE METHODS multichannel or simultaneous. The spectrometer tor. These instruments are ordinarily equipped
sho~ in Figure 15-9 (p. 436) is a sequential ~
with a computer for instrument control, data MultichinneiPUk. ".
Although it is feasible to excite an X-ray instrUment that can be readily employed for processing, and display of analytical results. height _Iyur
_:·~,~".~l_~·'-··
emission spectrum by incorporating the sample X-ray fluorescence analy'sis; here, the Coolidge An analysis for 20 or more elements can be
tube and sample are arranged as shown in the (b)
into the target area of an X-ray tube, the completed in a few seconds to a few minutes.
inconvenience of this technique discourages its circular insert at the top of the figure. Single- Multichannel instruments are widely used FIGURE 15-14 Energy dispersive X-ray
application to many types of materials. Instead, channel instruments may be manual or auto- for the determination of several components fluorescence spectrometer. Excitation by
excitation is more commonly brought about matic. The former are entirely satisfactory for in materials of industry such as stee~ other X-rays from (a) a Coolidge tube and (b) a
by irradiation of the sample with a beam of the quantitative determination of a few ele- alloys, cement, ores, and petroleum products. radioactive substance.
Fel1gett advantage (see p. 242). . The absorption edge of one of the filters lies dot represents the counts collected by one of
source, which may be a Coolidge tube or a just· below 5.4 A. while that of the other is just the several hundred channeli.
The principal disadvantage of energy
radioactive material, a sample holder, a lith- above it. The difference between the two sig-
dispersive systems, ~hen compared wi~ crys- Qualitative information, such as that shown
ium-drifted silicon detector, and the various nals is proportional to the sulfur content of the
tal spectrometers, is their lower resolutIOns at in Figure 15-16,can be converted to semiquant-
electronic components required for energy dis- sample. A sulfur analysis with this instrument
wavelengths longer than about 1 A (at short- itative data by the careful measurement of peak
crimination. requires a counting time of about 1 min. Rela- heights. To obtain a rough estimate of con-
An obvious advantage of energy dispersive er wavelengths, energy dispersive systems ex-
hibit superior resolution). tive standard deviations of about 1 % are ob- centration, the following relationship is used:
systems is the simplicity amf lack of moving
tained for replicate measurements.
parts in the excitation and detection compo- p. = P, Jv. (15-7)
nents of the spectrometer. Furthermore, the
Qualitative and where p. is the relative line intensity measured
absence of collimators and a crystal diffrae- Nondi •••••.• ive Instruments
Semiquantitative Analysis in terms of number of counts for a fixed
tor, as well as the closeness of the detector to
Figure IS-IS is a cutaway view of a simple, period, and Jv. is the weight fraction of the
the sample, result in a l00-fold or more in-
commercial, nondispersive instrument which Figure 15-16 illustrates an interesting qualita- element in the sample. The term P, is the rela-
crease in energy reaching the detector. These
has been employed for the routine determi- tive application of the X-ray method.' Here, tive intensity of the line that would be observed
features permit the use of weaker sources
nation of sulfur and lead in gasoline. For a the untreated sample, which was excited in the under identical counting conditions if W. were
such as radioactive materials or low-power
sulfur analysis, the sample is irradiated with X-ray beam by irradiation, was subsequently unity. The value of p. is determined ;ith a
X-ray tubes, which are cheaper and less likely
manganese K-radiation produced by radio- recovered unchanged. Note that the abscissa sample of the pure element or a standard
to cause radiation damage to the sample.
active iron-55 (p. 431~ which results in excita- for wavelength dispersive instruments is often sample of known composition. .
In a multichannel, energy dispersive instru-
ment, all of the emitted X-ray lines are mea- tion of the sulfur line at 5.4 A. The fluorescent plotted in terms of the angle 28, which can be The use of Equation 15-7, as outlined in
sured simultaneously. Increased sensitivity and radiation passes through a pair of adjacent readily converted to wavelength with knowl- the previous paragraph, carries with it the
improved signal-to-noise ratio result from the filters and into twin proportional counters. edge of the crystal spacing of the monOChro- assumption that the emission from the species
mator (Equation 15-6). Identification ofipeaks of interest is unaffected by the presence of
is then accomplished by reference to tables of other elements in the sample. We shall see
emission lines of the elements. ' . that this assumption may not be justified; as
Figure 15·17 is a spectrum obtained with a consequence, a concentration estimate based
an energy dispersive instrument. With such upon the equation may be in error by a factor
equipment, the abscissa is generally calibrated of two or more. On the other hand, this uncer·
in channel numbers or energies in keY. Each tainty is significantly smaller than that asso-

Filters for
alternative
elements

20 (goniometer reading) .

FIGURE 16-16 X-Ray fluorescence spectrum for a genuine bank note


recorded with a wavelength dispersive spectrometer. (Taken from H. A.
FIGURE 16-16 Cutaway view ofa commercial nondispersive X-ray fluorescence Liebhafsky, H. G. Pfeiffer, E. H. Winslow, and P. D. Zemany, X-ray Absorption
instrument. (Reprinted from Amer. LDb. 6(9~ 62 (1974). Copyright 1974 by and Emission in Analytical Chemistry. New York: Wiley, 1960, p. 163. With
International Scientific Communications, Inc.) permission.)
PRINCIPLES OF INSTRUMENTAL ANALYSIS \

""" and mass absorption coefficients of the matrix be ahscnt in the original sample. The ratio of
4.95
6,40 clements as welL
V
Fe
the intensities between the clement being deter-
(Kal
4.66 (Kal Absorption effects by the matrix may cause mined and the internal standard serves as the
Fe 5.41 [ results calculated by Equation IS-7 to be analytical parameter. The assumption here is
....
1.74 escape Cr
Si
(Ka) either high or low. If, for example, the matrix that absorption and enhancement effects are
(Kal 7.06 contains a significant amount of an element
5.90 Fe
the same for the two lines and that use of
2.02 Mn • (Kill that absorbs either the incident or the emitted intensity ratios compensates for these effects..
P
(Ka)
(Ka)
+ ..t.
.. beam more strongly than the element being Dilution of Sample and Standards. Here,
4.03 Cr • .. determined, then ~ will be low, since p. was both sample and standards are diluted with a
2.31 12.70 Diffraction
..
(Kill.
:. :-.. . [7,48 evaluated with a standard in which absorp- substance that absorbs X-rays only weakly
1.49
AI
5
(Kal
Rh
(anode)
peak
/\
r ~.
....
: ... ".j :
• N·

(~al
tion was smaller. On the other hand, if the
matrix clements of the sample absorb less
(that is, a substance containing elements with
low atomic numbers). Examples of such dilu-
.
(Kal ::~. 3.48
;... :••; \t
V ..
• •• ,. • Sn·. I
• 8.05 than those in the standard, high values for ~ ents include water; organic solvents contain-
· oj:: : ILl; ., " • Cu
result.
· •. N ... (Kal ing carbon, hydrogen, oxygen, and nitrogen
A second matrix effect, called the enhance-

~..
':' \. 'of .' only; starch; lithium carbonate; alumina; and
l· ; ••>,
'l:re> "',.tf ment effect, can also yield results that are boric acid or borate glass. By employing an

Excitation conditions
"'...- .~.~.
:,"
.~ greater than expected. This behavior is en-
countered when the sample contains an cle-
excess of diluent, matrix effects become es-
sentially constant for the diluted standards
Tube .node: Rh Filter: None
Anode voltage: 10 kV
ment whose characteristic emission spectrum and samples, and adequate compensation is
Anode current: 50 imPS is excited by the incident beam, and this spec- achieved. This procedure has proved par-
Llvetime: 200 seconds trum in turn causes a secondary excitation of ticularly useful for mineral analyses, where
the analytical line. both samples and standards arc dissolved in
Absorption and enhancement effects can molten borax; after cooling, the fused mass is
FIGURE 15-17 Spectrum of an iron sample obtained with an energy disper- clearly cause the intensity of an analytical line excited in the usual way.
sive instrument with a Rh anode Coolidge tube source. The numbers above the to depend not only upon the concentration Some Quantitative Applications of X-Ray
peaks are energies in keY. (Reprinted from Amer. Lab. 8(11~ 44, (1976). of the clement of interest but also, to a lesser Fluorescence. With proper correction for ma-
Copyright 1976 by International Scientific Communications, Inc.) degree, upon the concentrations of the various trix effects, X-ray fluorescence spectrometry
other elements in the sample matrix. Several is perhaps the most powerful tool available to
techniques have been developed to compensate the chemist for the rapid quantitative deter-
for such effects. mination of all but the lightest clements iJi
Matrix Effects. It is· important to realize Calibration Against Standards. Here, the complex samples. For example, Baird and
ciatcd with a semiquantitative analysis by op-
that the X-rays produced in the fluorescence relationship between the analytical line inten- Henke3 have demonstrated that nine elements
tical emission where an order of magnitude
process are generated not only from atoms at sityand the concentration is determined empir- can be determined in samples of granitic rocks
error is not uncommon.
the surface of a sample but also from atoms ically with a set of standards that closely ap- in an elapsed time, including sample prepara-
well below the surface. Thus, a part of both proximate the samples in overall composition. tion, of about 12 min. The precision of the
Quantitative Analysis The assumption is then made that absorption method is better than wet chemical analyses
the incident beam and the resulting fluores-
Modern X-ray fluorescence instruments are cent beam traverse a significant thickness of and enhancement effects arc identical for both and averages 0.08 % relative. It is noteworthy
sample within which absorption and scatter- samples and standards, and the empirical data that one of the elements analyzed is oxygen,
capable of producing quantitative analyses of
ing can occur. The extent to which either arc employed to convert emission data to which ordinarily can be determined by differ-
complex materials with a precision that equals
or exceeds that of the classical wet chemical beam is attenuated depends upon the mass concentrations. Clearly, the degree of com· ence only.
methods or other instrumental methods. For absorption coefficient of the medium, which pensation achieved in this way depends upon X-Ray methods also find widespread
the accuracy of such analyses to approach this in turn is determined by the coefficients of all the closeness of the match between the samples application for quality control in the manu-
level, however, requires either the availability of the elements in the sample. Thus, while the and the standards.
of calibration standards that closely approach net intensity of a line reaching the detector in UseoClDteraalStaDdards. In this procedure,
the samples in overall chemical and physical an X-ray emission analysis depends upon the an clement is introduced in known and fixed
composition or suitable methods for dealing concentration of the element producing the concentration into both the calibration stand- S A. K. Baird and B. L. Henke, Anal. CIwm.,37. 727

with matrix effects. line, it is also affected by the concentration ards and the samples; the added element must (l96S~
facture of metals and alloys. Here, the speed few parts per million can be measured. More identity can be assumed. In addition, diffrac- as to fulfill the Bragg condition for reflection
of the analysis permits correction of the com- commonly, however, the concentration range tion data sometimes yield quantitative infor- from every possible interplanar spacing.
position of the alloy during its manufacture. . of the method will be from perhaps 0.01 to mation concerning a crystalline compound in Samples may be held in the beam in thin-
X-Ray emission methods are readily adapt- 100%. X-Ray fluorescence methods for the a mixture. The method may provide data that walled glass or cellophane capillary tubes. Al-
ed to liquid samples. Thus, as mentioned ear- lighter elements are inconvenient; difficulties in are difficult or impossible to obtain by other ternatively, a specimen may be mixed with a
lier, methods have been devised for the direct detection and measurement become progres- means as, for example, the percentage of graph- suitable noncrystalline binder and molded
quantitative determination of lead and bro- sively worse as atomic numbers become smal- ite in a graphite-charcoal mixture. into a suitable shape.
mine in aviation gasoline samples. Similarly, ler than 23 (vanadium), in part because a com- Photographic Recording. The classical, and
calcium, barium, and zinc have been deter- peting process, called Auger emission, reduces Identification of still widely used, method for recording powder
mined in lubricating oils by excitation' of fluo- the fluorescent intensity (p.454). Present com- Crystalline Compounds diffraction patterns is photographic. Perhaps
rescence in the liquid hydrocarbon samples. mercial instruments are limited to atomic num- by X-Ray Diffraction the most common instrument for this purpose
The method is also convenient for the direct bers of 8 (oxygen) or 9 (fluorine). Another is the Debye-Scherrer powder camera, which
determination of the pigments in paint samples. disadvantage of the X-ray emission procedure Sample Preparation. For analytical diffrac- is shown schematically in Figure 15-18. Here,
X-Ray fluorescence methods are being is the high cost of instruments, which range tion studies, the crystalline sample is ground the beam from a Coolidge tube is filtered to
widely applied to the analysis of atmosphere from about $5000 for an energy dispersive to a fine homogeneous powder. In such a produce a nearly monochromatic beam (often
pollutants. For example, one procedure for system with a radioactive source, to more than form, the enormous number of small crys- the copper or molybdenum K.line), which is
detecting and determining contaminants in- $100,000 for automated and cotttputerized tallites are oriented in every possible direc- collimated by passage through a narrow tube.
volves drawing an air sample through a stack wavelength dispersive systems. tion; thus, when an X-ray beam traverses the The undiffracted radiation then passes out ·of
consisting of a micropore filter fOf particulates material, a significant number of the particles the camera via a narrow exit tube. The camera
and three ·filter paper disks imprCgnated with can be expected to be oriented in such ways itself is cylindrical and equipped to hold a
orthotolidine, silver nitrate, and sodium hy-
droxide, respectively. The latter three retain X-RAY
chlorine, sulfides, and sulfur dioxide in that DIFFRACTION METHODS
order. The filters then serve as samples for
X-ray fluorescence analysis. Since its discovery in 1912 byvon Laue,
Summary of the Advantages and Disadvan- X-ray diffraction has provided a wealth of im-
tages of X-Ray Fluorescence Methods. X-Ray portant information to science and industry.
fluorescence offers a number of impressive For example, much that is known about the
advantages. The spectra are relatively simple; arrangement and the spacing of atoms in crys-
spectral line interference is thus unlikely. Gen- talline materials has been directly deduced
erally, the X-ray method is nondestructive and from diffraction studies. In addition, such
can be used for the analysis of paintings, studies have led to a much clearer under-
archeological specimens, jewelry, coins, and standing of the physical properties of metals,
other valuable objects without harm to the polymeric materials, and other solids. X-Ray
sample. Furthermore, analyses can be per- diffraction is currently of prime importance in
formed on samples ranging from a barely elucidating the structures of such complex

mr §ti)) '})j
visible speck to a massive object. Other ad- natural products as steroids, vitamins, and
vantages include the speed and convenience of
the procedure, which permits multielement
analyses to be completed in a few minutes.
Finally, the accuracy and precision of X-ray
antibiotics. Such applications are .beyond the
scope of this text.
X-Ray diffraction also provides a conven-
ient and practical means for thei qualitative
[( ((Q))
~ film for entrance
fluorescence methods often equal or exceed identification of crystalline comp<>unds. This and exit tubes
those of other methods. application is based upon the fact that an
X-Ray fluorescence methods are generally X-ray diffraction pattern is unique for each lbl
not so sensitive as the various optical methods crystalline substance. Thus, if an exact match FIGURE 15-18 Schematic diagram of (a) a powder camera;
that have been discussed earlier in this text. In can be found between the pattern of an un- (b) the film strip after development. D2, 01, and T indicate positions
the most favorable cases, concentrations of a known and an authentic sample, chemical of the film in the camera.
strip of film around its inside wall. The inside analyte. Further elimination of ~ble. com- nature of surfaces. It has had important
diameter of the cylinder usually is 11.46 em,
so that each lineal millimeter of film is equal
pounds is accomplished by cons~eratio~ of
the spacing for the second most. mt~ line,
then the third, and so forth. Ordinarily, three
LuJ applications to phase studies in metallurgy
and ceramics, the investigation of grain bound-
to 0.5 deg in 8. aries in alloys, the determination of diffusion
The sample is held in the center of the or four spacings serve to identify the com- rates of impurities in semiconductors, the
beam by an adjustable mount. pound unambiguously. analysis of occluded species in crystals, and the
Figure 15-18b depicts the appearance of If the sample contains two or more crystal- studies of the active sites of heterogeneous
the exposed and developed film; each set of line compounds, identification becomes more catalysts. In all of these applications, both
lines (Db D2, and so forth) represents diffrac- complex. Here, various combinations of the qualitative and quantitative information about
tion from one set of crystal planes. The Bragg more intense lines are used until a match can surfaces is obtained.
angle 8 for each line is easily evaluated from be found.
the geometry of the camera. By measuring the intensity of the diffrac-
Electronic Reconling. Diffraction patterns tion lines and comparing with standards, a
X-RAY
can also be obtained with an instrument such quantitative analysis of crystalline mixtures is
PHOTOELECTRON
as that shown in Figure 15-9. Here again, the also possible.
SPECTROSCOPY
fixed source is a filtered Coolidge tube. The
sample, however, replaces the single crystal X-Ray photoelectron spectroscopy, sometimes
on its mount. The diffraction pattern is then referred to as ESCA or electron spectroscopy
obtained by automatic scanning in the same THE ELECTRON
for chemical analysis, is one of several types
way as for an emission or an absorption MICROPROBE METHOD FIGURE 15-19 Schematic view of an
of electron spectroscopy' ; others include ultra-
spectrum. electron-microprobe instrument. (From D.
violet photoelectric Spectroscopy, electron im-
In the electron microprobe method, X-ray B. Wittry, in Treatise on Analytical Chemistry,
pact spectroscopy, an4 Auger (pronounced Oh-
emission is stimulated on the surface of the eds. I. M. Kolthoff and P. J. Elving. New
Interpretation of jay) spectroscopy. All use an electron (rather
sample by a narrow, focused beam of elec- York: Interscience, 1964,part I, vol. 5, p. 3178.
Diffraction Patterns than an X-ray) speCtrometer. "These instru-
trons. The resulting X-ray emission is detected With permission.)
ments differ in that the former sorts electrons
The identification of a species from its powder and analyzed with a wavelength or energy
according to energy while the latter resolves
diffraction pattern is based upon the position dispersive spectrometer.-
X-radiation according to energy or wavelength.
of the lines (in terms of 8 or 28) and their The inclusion of a brief discussion of ESCA
relative intensities. The diffraction angle 28 is Instruments the diameter of the beam lies between 0.1 and in this chapter is convenient because the
determined by the spacing between a par- Figure 15-19 is a schematic diagram of an 11l1l1. An associated optical microscope is used method is based upon the same processes that
ticular set of planes; with the aid of the Bragg electron microprobe system. The instrument to locate the area to be bombarded. Finally, result in X-ray absorption and emission.
equation, this distance d is readily calculated employs three integrated beams of radiation, the fluorescent X-rays produced by the electron
from the known wavelength of the source and namely, electron, light, and X-ray. In addition, beam are collimated, dispersed by a single
the measured angle. Line intensities depend Principles of ESCA
it requires a vacuum system that provides a crystaL and detected by a gas-filled detector.
upon the number and kind of atomic reflection pressure of less than 10-5 torr and a wave- Considerable design. effort is required to ar- As pointed out earlier, absorption of an X-ray
centers that exist in each set of planes. length- or energy-dispersive X-ray spectrom- range the three systems spatially so that they photon by an atom generally produces an
Identification of crystals is empirical. The eter (a wavelength dispersive system is shown do not interfere with one another. excited ion and an electron. The process can
American Society for Testing Materials in Figure 15-19). The electron beam is pro- In addition to the foregoing, the specimen be described by the equation
(ASTM) publishes file cards that provide d duced by a heated tungsten cathode and an stage is provided with a mechanism whereby
spacings and relative line intensities for pure accelerating anode (not shown). Two electro- the sample can be moved in two mutually
compounds; data for nearly 10,000 crystalline magnet lenses focus the beam on the specimen; perpendicular directions and rotated as well,
materials have been compiled. The cards are
thus permitting scanning of the surface. • For reviews or electron spectroscopy, ICe: D. M. Hcr-
arranged in order of the d spacing for the
cu1es,Anal.Chem.,4Z (1~20A (1970);T. A Carlson, Photo-
most intense line; cards are withdrawn from
Applications electron and Auger Spectroscopy. New York: Plenum Presa,
this file on the basis of a d spacing that lies 4 For a detailed discuaion or this method, ICe: L S. 1975; and Handbook of X-Ray and UlrravioWt Photo-
within a few hundredths of an angstrOm of Birks, Electron Probe MIcr_lysl&. New York: Inta- The electron microprobe provides a wealth of eleetronSp"crroscopy,ed. D. Briggs.Philadelphia: Heyden,
the d spacing of the most intense line for the ICienc:e, 1963. information about the physical and chemical 1977.
where A +. represents the excited ion formed
when the clement A interacts with the X-ray
emitted electron
spectrometer.
E. by means of an electron The binding energies of K and L electrons
are influenced to a small extent by the outer
peaks are observed, each corresponding to
the carbon atom shown above it. It is evident
photon hVl' Relaxation of the excited ion can The two relaxation pr~es shown by bonding electrons since the latter influence from this figure that ESCA can provide useful
occur in two ways. That is, Equations 15-9 and 15-10 are competitive, the force field of the nucleus. Thus, the bind- structural information.
and their relative rates depend upon· the ing energy varies slightly as a function of the Figure 15-22 indicates the position of peaks
atomic number of the clement involved. High oxidation state of an element as well as the for sulfur in its several oxidation states and
atomic numbers favor fluorescence, while kind of element or clements with which it is in various types of organic compounds. The
Auger emission predominates With atoms of combined. ~ata in the top row clearly demonstrate the
very low atomic numbers. As a consequence, . Figure 15-21 is a photoelectric spectrum effect of oxidation state. Note also in the last
X-ray fluoresc:cnce is not a very sensitive means for carbon in ethyl trifluoroacetate. Four four rows of the chart that ESCA discriminates
for detecting elements with atomic numbers between two sulfur atoms contained in a single
Here, the subscripts indicate that the energies lower than about 10. ion or molecule. Thus, two peaks are observed
of the resulting photon or electron are different Both X-ray absorption and X-ray photo- for thiosulfate ion (S20~ -), suggesting different
from their counterparts in the first equation. electron spectroscopy are based upon Equa- oxidation states for the two sulfur atoms con-
The first relaxation process, of course, is the tion 15-8. In the former, an X-ray spectrometer tained therein.
is employed to study absorption as a function F 0 H H
basis of X-ray fluoresc:cnce methods (Figure I I 1 I j It should be pointed out that the informa-
15-20&).The second (Equation 15-10) is called of wavelength or energy as the sample is ir- F-C-C-O-C-C-H tion obtained by ESCA must also be present
I I I
radiated with a polychromatic source. In the F H H
Auger emission and is the basis of another type in the absorption edge of an X-ray absorption
of electron spectroscopy, which will not be latter, a monochromatic source is employed,
• spectrum for a compound. Most X-ray spec-
discussed in this text. Auger emission is a and the energy of the emitted electron is trometers, however, do not have sufficient res-
radiation less process in which energy is lost measured with an electron spectrometer (Fig- olution to permit ready extraction of the in-
by a transfer of one electron to a lower energy ure 15-20b). The kinetic energy E; of the 500
formation.
state, with a simultaneous ejection of a second emitted electron is employed to calculate the ~ !
co
electron. For example, if the absorption process binding energy of the electron by means of the .£
E
involves ejection of a Is electron, Auger relaxa- equation
tion can occur by transition of one 2p electron
8
E. = hv - E; (15-11) ~
z
to the Is orbital accompanied by the simulta-
neous ejection of a second 2p electron (Figure The binding energy is unique for each elec-
15-2Oc~ Auger spectroscopy is based upon tron of each clement and thus is valuable for
determination of the kinetic energy of the purposes of identification.
ISO,
H RS - 0131
I--i RSO,' , (51
RSO,R'H(9)
1190 1195 eV RSO," H
Kinetic energy
S,O, I
t--t RSfSOIR••.•••.••(31
I--t RSISO,I'1(4)
I Precision of I RSSO I'

measurement 3

FIGURE 15-21 Carbon Is X-ray


(al photoelectron spectrum for ethyl tri-
X-ray
fluoracence fluoroacetate. (From K. Siegbahn, et al., FIGURE 15-22 Correlation chart for sulfur
ESCA: Atomic, Molecular, and Solid- 2s electron binding energies. The numbers in
FIGURE 15-20 Mechanisms associated with three types State Structure by Means of Electron parentheses indicate the number of compounds
ofspectroscopyvIn (a1 the energy hv of the emitted radiation Spectroscopy. Upsala: Almqist and examined. [Reprinted with permission from
is measured; in (b) and (c1 the energy ofthe emitted electron Wilcksells, 1967, p. 21. With permis- D. M. Hercules, Anal. Chem., 42 (1), 20A (1970).
e is measured. sion.) Copyright by the American Chemical Society.]
Instrumentation Applications 7.. Aluminum windows are to be employed as windows for a cell for.
X-ray absorption measurements with the Ag Ka line. The mass
X-Ray photoelectron spectrometers are large The photoelectrons produced in ESCA are in-
abso.rption cc:'Cfficientfor aluminum at this wavelength is 2.74. What
and expensive pieces of equipment The source capable -of passing through more than per-
max~um thi~kness of aluminum foil could be employed to fashion
is an ordinary Coolidge tube. Monochromatic haps 20. to 50 A of a solid. Thus, the most
the \.VUldowsIf no more than 1.0% of the radiation is to be absorbed
radiation is obtained with filters (see Figure important applications of electron spectros-
by them?
15-8) or with a crystal monochromator. The copy, like X-ray microprobe spectroscopy,
electrons produced by irradiation of the sample are for tbe accumulation of information about 8. A solution of 12 in ethano~ had a density of 0.794 gfcm3• A 1.5-cm
are then passed into an electron spectrometer. surfaces. Examples of some of its uses include layer was found to transmit 27.3% of the radiation from a Mo Ka
The most common type of spectrometer utilizes identification of active sites and poisons on source. Mass absorption coefficients for I, C, H, and 0 are 39.3,0.70,
catalytic surfaces, determination of surface 0.00, and 1.50, respectively.
one or more curved magnetic fields to focus
the electron beam on a detector. The trajectory contaminants on semiconductors, analysis of (a) Calculate the percent 12 present, neglecting absorption by the
of an electron depends upon its kinetic energy the composition of human skin, and study of alcohol.
and the strength of the field. Variation in the oxide surface layers on metals and alloys. (b) Correct the results in part (a) for the presence of alcohol.
latter permits scanning of the electron energy It is evident that the method has a sub- 9. Calculate the goniometer setting, in terms of 28, required to obserVe
spectrum. stantial potential in the elucidation of chemi- t~e Kat. lines for Fe (1.76 A~ Se (0.992 A~ and Ag (0.497 A) when the
Because the electron path in a spectrometer cal structure (see Figures 15-21 and 15-22); ddTraetmg crystal is
is intluenced by the earth's magnetic field, the information obtained appears comparable (a) topaz.
elaborate shielding and Helmholtz coils must to that from NMR or infrared spectroscopy. (b) LiF.
be employed to reduce the effect of this field to A noteworthy attribute is the ability of ESCA (c) NaCI.
essentially zero. The entire instrument must be to distinguish among oxidation states of an 10. ~alculate the goniometer setting, in terms of 28, required to observe
operated at a pressure of 10-5 to 10-8 torr element. tpe L,I lines for Br at 8.126 A when the diffracting crystal is
to prevent absorption of electrons by gaseous ~) ethylenediamine d-tartrate.
species. OJ) ammo~ium dihydrogen phosphate.
11. <::alculate the minim~m tube voltage required to excite the following
hnes. The numbers 10 parentheses are the wavelengths in A for the
1. What is the short-wavelength limit of the continuum produced by a corresponding absorption edges.
Coolidge tube having a chromium target and operated at 75 kV? (a) K lines for Ca (3.064)
2. What minimum tube voltage would be required to excite the K and (b) L. lines for As (9.370)
the L series of lines for (a) Cr, (b) Cs, (c) W, (d) U? (c) L, lines for U (0.592)
(d) K lines for Mg (0.496).
3. The Kot lines for Ca, Zn, Zr, and Sn occur at 3.36, 1.44, 0.79, and
0.49 A, respectively. Calculate an approximate wavelength for the Ka
lines of (a) T~ (b) Fe, (c) As, (d) Ag, (e) I, (f) Br.
4. The L. lines for Ca, Zn, Zr, and Sn are found at 36.3,11.9,6.07, and
3.60 A, respectively. Estimate the wavelengths for the L. lines for the
elements listed in Problem 3.
5. The mass absorption coefficient for Ni, measured with the Cu Ka line,
is 49.2 cm2jg. Calculate the thickness of a nickel foil that was found to
transmit 27.3 % of the incident power of a beam of Cu K. radiation.
6. For Mo Ka radiation (0.711 A), the mass absorption coefficients for
K, I, H, and 0 are 16.7, 39.2, 0.0, and 1.50 cm2jg, respectively.
(a) Calculate the mass absorption coefficient for a solution prepared
by mixing 5.00 g of KI with 95 g of water.
(b) The density of the solution described in (a) is 1.04 gfcm3• What
fraction of the radiation from a Mo K. source would be trans-
mitted by a 0.5O-cm layer of the solution?
The discovery and production of both natural . radioactive reagent is employed to separate
and artificial radioactive isotopes have made
completely the analyte from the bulk of the
possible the development of analytical methods sample; the activity of the isolated portion is
(radiochemical methods) that are both sensi-
measured. Alternatively, the analyte may be
tive and specific. These procedures are often
titrated with a radioactive reagent; here, an
characterized by good accuracy and wide-
end point is established by activity measure-
spread applicability; in addition, some mini- ments.
mize or eliminate chemical separations that
often precede the measurement step. 1
Radiochemical methods are of three types. THE
In activation analysis, activity is induced in RADIOACTIVE
one or more elements of the sample by irra- DECAY PROCESS
diation with suitable particles (most commonly
thermal neutrons from a nuclear reactor); the The disintegration of radioactive isotopes pro-
resulting radioactivity is then measured. In an duces energetic particles and electromagnetic
isotope dilution procedure, a pure but radio- radiation. Alteration of the nucleus accom-
active form of the substance to be determined panies these processes.
-is mixed with the sample in known amount.
After equilibration, a fraction of the component Types of RadiatiOn
of interest is isolated by suitable means; the
analysis is then based upon the activity of this Table 16-1 lists some of the types of particles
isolated fraction. In a radiometric analysis, a and radiation that are encountered in radio-
chemical studies. Several of these serve as the
basis for analysis and are thus discussed in
the sections that follow.
I For a detailed treatment of radiochemical methods, Alpha Particles. Alpha particles usually re-
see: G. Friedlander. J. W. Kennedy, and J. M. Miller. sult from the disintegration of isotopes possess_
NucleQ1' tmtI Rat/iochefflistry, 2d ed. New York: Wiley. ing high atomic numbers. An example is shown
1964; H. M. Oark. in Physical Methods o/Chemistry. eds.
by the equation
A. Weissberaer and B. W. Rossiter. New York: Wiley-
Intenc:ienlle, 1972,Part IIID, vol. 1. Chapter 9. .
2UU-+ 2:trh + ~He

TABLE 16-1 CHARACTERISTICS OF COMMON


RADIOACTIVE DECAY PRODUCTS

Particle Symbol Charge Mass Number


Alpha ex +2 4
Electron p- 1
-1
1840

p+ 1
Positron +1
1840
Gamma ray "I
0 0
X-Ray x 0 0
Neutron n 0 1
Neutrino v 0 0
The alpha particle is a helium nucleus and . original isotope. As shown earlier (p. 431~ photon energy and recoils at a corresponding with a moderating substance of low atomic
carries a positive charge of 2. Its formation relaxation of the excited ion can produce nega- angle with respect to the projected path of the weight; the result is a neutron flux containing
results from the transition of a uranium-238 tive electrons in the form of Auger electrons photon. The photon, now with diminished low energies distributed about some average
nucleus to a nucleus of thorium-234. Alpha (see p. 454). energy, suffers further energy losses; finally, it value.
radiation is generally encountered as a decay In contrast to alpha emission, beta decay is causes the photoelectric ejection of an etectron Neutrons can interact with matter in any
product of heavier nuclei only. characterized by production of particles with from a component of the medium. If the gamma of seve~1 ways. The product (or products)
Alpha particles from a particular decay pro- a continuous spectrum of energies ranging photon possesses sufficiently high energy (at depend In large measure on the energies of
cess are either monoenergetic or are distributed from nearly zero to some maximum that is least 1.02 MeV), pair production can occur. the bombarding neutrons. Irradiation of a
among relatively few discrete energies. They characteristic of each decay process. The beta Here, the photon is converted into a positron stable isotope with thermal neutrons is most
progressively lose their energy as a result of particle is not nearly as effective as the alpha and an electron in the field surrounding a li~ely to give .rise to a highly excited isotope
collisions as they pass through matter, and particle in producing ion pairs in matter be- nucleus. With an atomic mass number one unit larger
are ultimately converted into helium atoms cause of its small mass (about 1/7OCJJthat of X-Ray Emission. Two types of nuclear pro- than that of the target. The prodUct achieves
through capture of two electrons from their an alpha particle); at the same time, its pene- cesses, electron capture and internal conversion, greater stability through the prompt (within
surroundings. Their relatively large mass and trating power is substantially greater. Beta are followed by the emission of X-ray photons. - 10-12 s) emission of a gamma °ray photon,
charge render alpha particles highly effective ranges in air are difficult to evaluate because In the previous chapter (p. 454~ it was pointed 'Y. The process can be depicted by the
in producing ion pairs within the matter of the high likelihood that scattering win occur. out that the former yields an excited ion that sequence
thJ;ough which they pass; this property makes As a result, beta energies are based upon the relaxes by formation of either X-rays or Auger 10-11,
their detection and measurement easy. Because thickness of an absorber, ordinarily aluminum, electrons. Which of these processes predomi- ~X + An ..• [IA+IJX]* IA+~X + 'Y
of their high mass and charge, alpha particles required to stop the particle. This thickness is nates depends upon the atomic number of the
have a low penetrating power in matter. The the range expressed in mg/cm2• excited species. Excited state Ground state
identity of an isotope that is an alpha emitter Beta particles that carry a unit positive Internal conversion is a process in which an where the superscript refers to the mass of
can often be established by measuring the charge are called positrons. Positrons have excited nucleus (resulting from a decay event) element X having an atomic nuniber Z. Fast
length (or range) over which the emitted alpha highly transient lifetimes; their ultimate fate is loses its excitation energy by ejecting an elec- neutrons interact wit1l matter by dther mecha-
particles produce ion pairs within a particular annihilation by interaction with ordinary elec- tron from one of the orbitals near the nucleus. nisms, but these reactions are not 'pertinent to
medium (often air). trons to give two gamma photons. This alternative to gamma ray emission re- our discussion. 0

Alpha particles are relatively ineffective for Gamma Ray Emission. Many alpha and sults in a vacated K, L, or M shel~ which is
producing artificial isotopes because of their beta emission processes leave a nucleus in an then filled by an electron from a higher energy Units of Radioactivity
low penetrating power. excited state, which then returns to the ground level. An X-ray photon characteristic of the
The curie is the fundamental unit of radioactiv-
Beta Particles. Beta particles are produced state in one or more quantized steps with the element results from this transition.
ity; it is defined as that quantity of nuclide
within a nucleus by the spontaneous transfor- release of gamma rays (electromagnetic radia- It is important to note that gamma rays in which 3.7 x 1010 disintegrations occur per
mation of a neutron to a proton or a proton tion of very high energy). The gamma ray and X-rays differ only in their source, the second. Note that the curie is an enumerative
to a neutron. The particles consist of an elec- emission spectrum is characteristic for each former arising from nuclear events and the quantity only and provides no information
tron (negatron) in the former case and a posi- nucleus and is thus useful for identifying radio- latter from electronic transitions outside the concerning the products of the decay process
tive electron (a positron) in the latter. Exam- isotopes. nucleus.
or their energies.
ples of two reactions that yield fJ rays are Not surprisingly, gamma radiation is highly Neutroa The neutron (n) is a particle of The millicurie and microcurie are frequently
penetrating. Upon interaction with matter, unit mass and zero charge. As such, it is a more convenient units.
l:C ..•l~N+ e- +v gamma rays lose energy by three mechanisms; highly effective bombarding particle, being un-
the one that predominates depends upon the influenced by the electrostatic charge barrier The Decay Law
~gZn..• ~~Cu + e+ + v energy of the gamma photon. With low energy surrounding a target nucleus. In contrast to
Radioactive decay is a completely random pro-
gamma radiation, the photoelectric effect re- charged particles, which require high kinetic
cess. Thus, while no prediction Clin be made
where v represents a neutrino, a particle of no sults in the ejection of a single electron from energies to surmount such barriers, slow (or
concerning the lifetime of an individual nu-
analytical significance. A third process that a high atomic weight target atom. With rela- thermal) neutrons are generally more reactive
cleus, the behavior of a large ensemble of like
yields negatrons is electron capture; here, an tively energetic gamma rays, the Compton effect than high-energy neutrons. It is for this reason
nuclei can be described by the expression
inner electron (usuanya K electron) is captured is encountered in which a gamma photon and that the neutrons emitted from the SOurce (or-
by the nucleus, leaving an excited ion with an an electron participate in an elastic collision. dinarily a nuclear reactor) are caused to lose
_dN =IN
atomic number that is one less than that of the The electron acquires only a portion of the much of their kinetic energy through collisions dt
where N represents the number of radioactive count r for a selected period is 5; curves B
nuclei in the sample at time t and l is the and C correspond to samples having true
characteristic decay constant for a particular averages of 15 and 35. Note that the absolute
radioisotope. Upon rearranging this equation deviations become greater with increases in r,
and integrating over the interval between t = 0 but the relative deviations become smaller.
and t = t (during which the number of nuclei Note also that for the smallest number of counts
in the sample decreases from No to N~ we the distribution is distinctly not symmetric
obtain around the average; this lack of symmetry is
a consequence of the fact that a negative count
N is impossible, while a finite likelihood always
In - = -At (16-1)
No exists that a given count can exceed the average
The half-life of a radioactive~isotope is de- by a factor greater than two.
Standard Deviation of Counting Data. When
fined as the time required for t,he number of
the total number of counts becomes large (r >
atoms to decrease to one-half of its original
100), the distribution of deviations from the
quantity; that is, for N to beCome equal to
mean approaches that of a symmetrical Gauss-
No 12. Substitution of No 12 for N in Equation
ian or normal error curve. Under these circum-
16-1 gives
stances, it can be shown that
till = 0.~3 (16-2)

where N is the number of counts for any given


Counting Errors2 period and aN is the standard deviation in N
The randomness inherent in·the decay prQCCSs (see Appendix 2). The relative standard devia-
prevents an exact prediction of the number of tion (aN). is given by
disintegrations that will occur in any time in-
terval Nevertheless, given a sufficiently long
counting period, the extent of decay during the
interval is found to be reproducible within a
predetermined level of precision.3 Thus. although the absolute standard devia- Substitution of Equation 16-5 leads to rate) can be expected to be with a given degree

ft
Figure 16-1 shows the deviation from the tion increases with the number of counts, the of confidence. For a Gaussian distribution, it
true average count that would be expected if relative standard deviation decreases. can be shown that (see Appendix 2)
aK = (16-6)
1000 replicate observations were made on the In normal practice, activities of samples
same sample. Curve A gives the distribution are more conveniently expressed in terms of r= N ± zaN = N ± zJN (16-8)
In relative terms,
for a substance for which the true average counting rates R (counts per minute) than in where r is the true average count and z is a
counts for some arbitrary time. Clearly,
(aK~ =~ = fIt (16-7) constant that depends upon the confidence
level desired. The quantity ±zaN = ±zJN is
. N clearly the absolute uncertainty of the mea-
R=- Uncertainty in a Single Measurement. The
2 For a morc complctc discussion, ~: G. Friedlandcr, J. t surement; that is,
standard deviation can be used to define a
W. Kennedy, and J. M. MilIcr, H.ucrear tmd Radio-
chemistry,2d ed. New York: Wiley, l~, Chaptcr 8. where t is the time in minutes required to range about the measured count within which Absolute uncertainty = ±zJN (16-9)
. obtain N counts. The standard deviation in the true average count (or the true average
• Thc counting period should be short, howevcr, with Some values of z for various confidence levels
respect to the half·lifc so that no sipUficant change in the rate units aK can be obtained by dividing foUow:
number or radioactive atoms occurs. Further restrictions both sides of Equation 16-3 by t; thus.
includc a detector that responds to the decay or a single
isotope only and an invariant counting geometry so that 50% 90% 95%
thc dctector responds to a constant rraction or the decay
0.68 1.65 1.96
evcnts that occur.
Thus, for 95 out of 100 measurements the true materials into the earth's atmosphere. In order Substituting into Equation 16-15 yields
Thus the uncertainty associated with a single to obtain a true assay, then, it is necessary to
average count r lies in the range 624 to 726.
count N at the 50% confidence level would be correct the total count for background. The + 20/4 = 00300
From Equation 16-11, (u \ = J180/10
.J- 180 - 20 .
ZCTN= ±O.68uN = ±O.68JN
The uncertainty at the 50% confidence level
(+O.68JN) is often termed the probable error
Relative uncertainty = ± p x 100..= 7.5%
counting period required to establish the back-
ground correction frequently differs from that
for the sample; as a result, it is more convenient
At the 95 % confidence leve~
to employ counting rates. Then, Relative uncertainty = z( CT.)r
oCa count. The probable error defines the Figure 16-2 illustrates the relationship be-
interval around N within which the true mean tween total counts and tolerable levels of un- = 1.96 x 0.0300
r could be expected to be found 50 times out certainty as calculated from Equation. 16-~1.
= 0.059 or 5.9%
of 100. Note that the horizontal axis is loganthmic; where R. is the corrected counting rate and
The uncertainty in a counting measure- it is clear that a tenfold decrease in the relative R. and Rb are the 'rates for the sample and the Thus, the chances are 95 in 100 relative that
ment can also be expressed in relative terms. uncertainty requires an approximately hun- background, respectively. the uncertainty in the corrected count is smaller
Thus, dredfold increase in the number of counts. The square of ~he standard deviation of a than 5.9% relative.
Relative uncertainty = ±Z(CTN~ (16-10) The foregoing treatment of uncertainty can sum or difference is equal to the sum of the
be performed equally well in terms of count- squares of standafd deviations of the individ- This example suggests that the contribu-
and from Equation 16-4, ing rates rather than of total counts; here, Ua ual components inaking up the suni or the tion of background to the standard deviation
Relative uncertainty = ±z/JN (16-11) and (ua)" as defined by Equations 16-6 and difference (see Appendix 1). Thus, we may is minimized under circumstances where the
16-7, are employed. write that background counting rilte is small with re-
EXAMPLE Backgrouad Correctioas. The count record- spect to the sample rate. It can be shown that
Calculate the absolute and relative uncer- ed in a radiochemical analysis includes a con- the optimum division of time between back-
tainties at the 95% confidence level in the tribution from sources other than the sample. where CT. is the standard deviation in R., and ground and sample counting is given by
measurement of a sample that produced 675 Background activity can be traced t~ the ex~. CT. and CTb are the borresponding standard de-
counts in a given period. tence of minute quantities of radon ISOtOpesIn [R;, (16-16)
Absolute uncertainty = ZCTN
the atmosphere, to the materials used in con-
viations for the sample and' the background.
Substituting Equation 16-6 yields
~=
t. ..JR:
struction of the laboratory, to accidental con-
= ±1.96J675 tamination within the laboratory, to cosmic
= ±51 counts radiation, and to the release of radioactive
InatnJmentation
The relative standard deviation (u.), is given Radiation from radioactive sources can be de-
by tected and measured in essentially the same
way as X-radiation (Chapter IS, p. 438). Gas-
filled detectors,scintillation counters,and semi-
conductor detectors are all sensitive to alpha,
beta, and gamma rays because absorption of
these particles produces photoelectrons which
EXAMPLE
'can in turn produce thousands of ion pairs.
A sample yielded 1800 counts in a 10-min
A detectable electrical pulse is thus produced
period. Background was found to be 80 counts
for each particle observed.
in 4 min. Calculate the relative uncertainty in
Measurement of Alpha Particles. Samples
the corrected coupting rate at the 95 % con-
to be assayed for alpha activity should be as
fidence level. .
thin as possible to minimize self-absorption
1800 losses. Similarly, any window between the sam-
R. = - = 180 cpm ple and counter must also be extremely thin.
10 To eliminate the absorption problem, alpha
80 sources are often sealed and counted in win-
~=- =20cpm dowless gas flow proportional counters.
4
As mentioned earlier, alpha spectra consist of the low probability of such noise affecting employed for excitation of the sample' slow
of discrete energies, which are useful for identi- both systems simultaneously. neutrons, fast neutrons, gamma rays, and var-
fication. Pulse height analyzers (p. 443) can be Because beta spectra are ordinarily con- ious charged particles have been used. Most
employed for the derivation of energy spectra tinuous, pulse height analyzers are less use- activation methods are based upon thermal
of alpha emitters. ful. neutrons, although gamma excitation appears
MeuInmeat or Beta Partides. For beta Meua'emeat or Gmma Wiatioa. to be a field of growing'interest. We shall con-
sources having energies greater than about Gamma radiation, which is indistinguishable fine our discussion to irradiation by thermal
0.2 MeV, a uniform layer of the sample is or- from X-radiation, is detected and measured neutrons.
dinarily counted with thin windowed Geiger by the methods described in Chapter 15. A second variable, which is used to charac-
or proportional tube counters. For low energy Interference from IX- and P-radiation is readily terize activation methods, is the type of emis-
beta emitters, such as carbon-14, sulfur-35, and avoided by filtering the beam with a thin sion measured in the final step of the analysis.
tritium, liquid scintillation counters (p. 441) are window of aluminum or Mylar. Here, both beta and gamma radiation have
preferable. Here, the sample is dissolved in a Gamma ray spectrometers are similar to the been monitored; the former is often more sen-
solution of the scintillating compound. A vial pulse height analyzers described in the previous sitive but, on the other hand, frequently suffers
containing the solution is then placed between chapter. Figure 16-3 shows a typical gamma from being less selective since the radiation is
two photomultiplier tubes housed in a light- ray spectrum obtained with a 400-channel continuous rather than discrete.
tight container. The output from the two tubes analyzer. Finally, activation methods can be classified
is fed into a coincidence counter, an electronic Figure 16-4 is a schematic diagram of a well-
FIGURE 18-4 A well-type scintillation
as being destructive or nondestructive of the
device that records a count only when pulses type scintillation counter. Here, the sample is sample. In destructive methods, the irradiated
counter. (Courtesy of Texas Nuclear Division,
from the two detectors arrive at the same time. contained in a small vial and placed in a cylin- sample is dissolved, and the element of interest
Ramsey Engineering Co., Austin, Texas. For-
The coincidence counter reduces background drical hole or well in the scintillating crystal of is counted after it has been isolated by suitable
merly Nuclear-ehicago Corporation).
noise from the detectors and amplifiers because the counter. chemical or physical means; possible inter-
ferences from other species made radioactive
by the irradiation are thus eliminated. In the
nondestructive procedure, the activated sample
is counted without preparatory treatment;
The basis for activation analysis is the mea.
here, the ability of a gamma ray spectrometer
surement of the radioactivity induced in a
to discriminate between radiation of different
sample as a result of irradiation by nuclear
energies is called upon to provide the required
particles (usually thermal neutrons from a
selectivity. The nondestructive method offers
reactor). The single most important advan-
the advantage of great speed. On the other
tage of activation methods is high sensitivity,
hand, the resolution of a gamma ray spectrom-
which exceeds that of other methods for
eter may be insufficient to eliminate interfer-
many elements by a factor of 100 or more;
ences. Also, it does not permit the use of beta
concentration determinations in the parts-
emission for an analysis.
per-billion range are common.·

Classification of Activation Methods


Destructive Methods
Activation methods can be classified in several
ways. One is based on the type of radiation The most common activation procedure in-
300 volves solution of a known amount of the
Chlnnel number, 3.6 keVlchannel irradiated sample followed by separation of
the analyte from interferences. The isolated ma-
4 Monographs on neutron activation analysis include: D.
FIG U R E 18-3 Gamma ray spectrum of aluminum wire after terial or a fraction thereof is then counted for
. DeSocte, R. Gybcls, and J. Hoste, Neutron Aetll7otion
neutron activation. [Reprinted with permission and taken from Analysis. New York: WiJey-lnterscience, t912; p, Kruger, its beta or gamma activity.
S. G. Prussin, J. A. Harris and J. H. Hollander, Anal. Chern., Principles of Aetill<ltionAnalysis. New York: Wiley-Inter- Conventional neutron activation analysis
37,1127 (1965). Copyright by the American Chemical Society.] science, 1971.
involves irradiation of a standard containing
of the sample and the standard immediately due to scattering, absorption, and differences
a known mass w. of the analyte simultane- after irradiation. The weight of the analyte is
A.~ in geometry between sample and standard. The
ously and in the same neutron flux as ~e
sample. Insofar as the activity !hat results IS
Os = W. then calculated directly from Equation 16-17. errors from these causes can usually be reduced
Clearly, success of the nondestructive meth- to less than 10% relative; uncertainties in the
proportional to mass, and provided also that Substituting these expressions into Equation od requires that the spectrom~ter be able to range of 1 to 3 % are frequently obtainable.
" the other components of the sample do .not 16-17 yields isolate the gamma ray signal produced by the SeasitiYity. The most important character-
. produce detectable radioactivity, the Weight analyte from signals arising from the other istic of the neutron activation method is its
w of the element in the sample is given by
• components. Whether or not an adequate reso- remarkable senSitivity for many elements. Note
lution is possible depends upon the complexity in Figure 16-5, for example, that as little as
Whe~e the condition WI <C W. is not satisfied, of the sample, the presence or absence of ele- 10- 5 Jig of several elements can be detected.
a more complex equation must be employed. ments which produce gamma rays of about the Note also the wide variations in sensitivities
The Substoichiometrie MethocL5 It is ex- same energy as that of the element of interest, among the elements; thus, about 50 Jig of iron
where A. and A. are the activities of the
perimentally feasible to impose the further con- and the resolving power of the spectrometer. are required for detection, in contrast to 10-6
sample and standard. Generally, however, t?e
ditions that, at the time of assay, WI = W. and Improvements in resolving power, which have Jig for europium.
neutron flux can be expected to generate actiV-
w~ =~. Equation 16-21 then simplifies to been made in the last few years, have greatly The sensitivity of the activation method for
ity in elements other than the ~alyte: Thus,
chemical isolation of the speCIes of mte~t broadened the scope of the nondestructive an element is a function of a number of var-
from a solution of the sample ordinarily pre-
WI ~ w. ~ (16-22) method. At the present time, however, the most iables. Some of these are associated with the
cedes radioassay. When the analyte is present selective and sensitive activation methods are properties of the particular nucleus. Others are
as a trace (as it usually is in a neutron activation which forms the basis for the substoichio- still based upon isolation of the analyte. The related to the irradiation process; still others
analysis),. its separation from the major con- metric method. Where the mass of collector great advantage of the nondestructive approach have to do with the efficiency of the counting
greatly exceeds that of the radioisotope, WI is its simplicity in terms of sample handling apparatus.
stituents may be difficult and the source: of large
error. This problem is minimized by introdpc- and WI are essentially identical provided the and speed; to be sure, the required instrumen- The effect of a number of these variables
same weight of collector is added to the sample tation is more complex. on the activity A induced in a sample after
ing a known weight W. of the element to !he
and the standard. The requirement that the irradiation for a time t is given by the ex-
solution ofthe irradiated sample as a nonacnve
same total amount of species be taken for pression
carrier or collector. Separation of the carrier Application of Neutron Activation
plus the irradiated element (WI + WI) is then radioassay (that is, that w~ - ~) is analytically
accomplished by precipitation, extraction, ion
unique, in that the same quantity must be taken
from solutions of inherently dissimilar concen-
Figure 16-5 illustrates that neutron activation 0.693
A = Nue/> [ 1 - exp ( - ~ t) 1 (16-23)
exchange, or chromatographic means. .A is potentially applicable to the determination
tration. The problem is resolved by introducing of 69 elements. In addition, four of the inert
weighted quantity ~ of the isolated matenal
is counted, and the resulting ldivity a. is a suitable reagent in an amount insufficient gases form active isotopes with thermal neu- where A is given in counts per second. The
related to the total activity of the original (substoichiometric) for complete removal of the trons and thus can also be determinC:d. Finally, quantity N refers to the number of target
species of interest from either sample or stan- three additional elements (oxygen, nitrogen, nucle~ and u is their neutron capture cross
sample A. by the relationship
dard. If the same amount of this reagent is and yttrium) can be activated with fast neu- section in cm2 per nucleus. The neutron flux
w· used for both sample and standard, w~ and w~ trons. A list of types of materials to which the in units of neutrons per cm2 second is given
a.=AIW.~W. are identical because their amounts are deter- method has been applied is impressive and bYe/>,while t and tl12 are the irradiation time
mined by the amount of added reagent. includes metals, alloys, archeological objects, and half-life of the isotope formed; the two
Ordinarily, the amount of nonactive element semiconductors, biological specimens, rocks, times are expressed in the same units.
added is several orders of magnitude greater minerals, and water. Acceptance of evidence The neutron capture cross section is a
than the weight from the sample; that is, Nondestructive' Method
developed from activation analysis by courts measure of the probability that a given kind
WI <l1 WI' Under this circumstance, Equation In the nondestructive method, a gamma ray of law has led to its use in forensic chemistry. of nucleus will capture a neutron. This quan-
16-18 simplifies to ! spectrometer is used to measure the activities Most applications have involved the determi- tity depends in a complex way upon the energy
nation of traces of various elements. of the neutron; typically, one or more neutron
A.~
a.=-- Accuracy. The principal errors that arise in energies will correspond to a very high proba-
WI activation analyses are due to self-shielding,
• For a more detailed treatment, ace: J. Ruzicka and J.
bility for capture.
The standard sample is treated in an identical Stary. Substoichiometry In Radi«hemical Analysll. New unequal neutron flux at sample and standard, Figure 16-6 illustrates how the induced ac-
way; thus, an analogous expression can be York: Pergamon Press, 1968. counting uncertainties, and errors in counting tivity varies with neutron flux and with irradia-
1.1 '1'1 '1'1' t't o~ .
,,-- ae~
I _92
_00
u. 1- 0-- lD--
's ~ C)
--
xx Xx

"''''
xx
....
"'-
I I
-'xx
"''''
..,1.1. .8 ~ "i associated with counting rates the exceed the
:E~ Ii ] .
_22 >---
""'~ ~ .'" "'xx xx resolving time of the detecting system; under
'1"''''
.~ oS •.• these circumstances, a correction must be in-
«'~
'1'1
.,-- 'I' 1
--
"i ~i .1::
IL-
0 .00
«--
..,00
iD;
EOO troduced to account for the difference between
1 xx I "'xx
x xx ~ .8 "ij elapsed (clock) and live (real) counting times .
'"
1.
-Ill
"'- '" '1'1 CQ. H .~
.tt egg
i i
w--
_00 g-:~
..
Cl-- "'0 I
<n~~ xx "xx IL_
xx
,,
'0 0 Q.,
~
- ..
'"
-sd:~
"'''' "''''
n t1 8. 6 .~
'.
..c::~
,,00
Cl-- .=-- 000
:J:--
~::l
«xx
-- xx
"''''
i i
xx
Ill_

11 'I' J.t
xx
8.g~ Isotopic dilution methods, which predate acti-
N--
coo
xx B--
xx :J:'" I~
0
0--
xx
~ '2 ~
.8i:: vation procedures, have been and still are ex-
-- '1"'''' n - -'" e' E-o tensively applied to problems in all branches
'7'1
as!S! 029 ::199
'1
--
'i'i
... xx
..,00 ~ S.8
l'~
of chemistry. These methods are among the
most selective available to chemists.
-- , .."''''-
_1.1
xx «xx «xx
Ill'"
"'- c.~
::>
.!-
u •.•••
~8 .~
Both stable and radioactive isotopes are
z-- ~--
xx
,,0
a.;1tJ
I
00
xx
I
~11
xx
Cl vi.[~.~ FIGURE 16-6 The effect of neutron flux
employed in the isotopic dilution technique.
The latter are the more convenient, however;

8 xx
"''''
11 .coo
",--
'"
'11
=5!5!
Ill_

, '1
::aSS
-'"
'f1 1~~8.
uulll:..c::
e B .•..
upon the activity induced in a sample. because of the ease with which the concentra-
tion of the isotope can be determined. We
--
W xx
xx xx :t .~ shall limit this discussion to methods employ-
"'- ,-'"
rr 'I' ,"''''
'1 g 'O~
.~..c:: ~ . ing radioactive species.
~ill~
,,00 .0
0::--0- I EOO
"xx -- 1;g",~
.::: u c::('f"\
i

coo
,, xx
-'"
x
,'" '1 "''''
~ ..s.5
'"
~"C)tn
r:i.
.::1 •
tion time. It is seen that irradiation in excess
of the saturation time dOes not provide any
Principles of the
:0--
xx
.5!5!
"'xx e ~ .~
•...•..•~...-4
further increase in activity; here, the rate of
decay and the rate of formation of the active
Isotopic Dilution Procedure
Ill'"
"'- 6~ ~>
0
i i
00
'1'1
00 ,,1'& 1t 'a --
a ....~J
0 ~
species are the same. Equation 16-23 indicates Isotopic dilution methods require the prepara-
0 1- :0--
xx
iO--
xx z-- xx :l --
xx e III .,r that the irradiation time required to reach tion of a quantity of the analyte in a radio-
Ill_

.. -
-Ill
,.,. III III ~~"'g
.~.~
.~ ~
saturation increases with increasing half-life for active form. A known weight of this species is
'7'1
00
>--
xx
'1
z;:;_ ... xx
..,0 ,,00
--
I I
_00
IL --
xx :i!:; u
~-{/)

eg.::: ~ ~
O?J
the product nucleus .
The efficiency of chemical recovery, if re-
then mixed with a weighed quantity of the
sample. After treatment to assure homogene-
"'-
i e;e '" "'''' "'Ill

i i
....
I I
~.;;~= quired prior to radioassay, may limit the sensi- ity between the active and nonactive species,
00
t=;; N-- i 1- 1lS!S!
xx
~ 99
•..
xx '2EB~
••• fIl ..•..•••.•
lIlI•••r! ..c::
C
tivity of an activation analysis. Other factors
include the sensitivity of the detection equip-
a part ofthe analyte mixture is isolated chemi-
cally in the form of a purified compound of
Ill'"

'7'1 "''''
"'-I"
.5 g e..s: ment for the emitted radiation, the extent to known composition. By counting a weighed
"
cllS!S!
xx "--
00
-'xx
"'-I""
~ .8u~
c,=··
c:: 0
><
which activity in the sample decays between
irradiation and assay, the time available. for
portion of this product, the extent of dilution
of the active material can be calculated and
-'" -'" eN
i i
",00 ,,0
:0--
xx U"":ao
'1'7
00
cr i
.5!5!
U;;;;; lD
. = ....
It)::Io~
fDtxz
u
counting, and the magxtitude of the back-
ground count with respect to that for the
related to the amount of nonactive substance
in the original sample. It is important to realize
XX
~ ~ sample. A high rate of detay is desirable from
"'''' "'''' "'- wooc!= that quantitative recovery of the species is not
.1.1. 1.1.
f
z-- ~-- ",-",
xx xx
'i'
..,0
x
l3--
xx
i i
00 11:":'".::
j •• O'
Cl~SO
~ the standpoint of minimizing the duration (}f
the counting period. Concomitant with high
required. Thus, in contrast to the typical ana-
lytical separation, steps can be employed to
"'''' "'''' '" "'''' LLlib=); decay rates, however, is the need to establish assure a highly pure product on which to base
with accuracy the time lapse between the cessa- the analysis. It is this independence from the
tion of irradiation and the commencement of need for quantitative isolation that leads to the
counting. A further potential complication is high selectivity of the isotopic dilution method.
Direct Isotope DDutioa. Assume that Wo division of Equation 16-24 by Equation 16-27 can be expected, however, because of the rela- Examples include the determination of chro-
grams of a radioactive species having an activ- and rearrangement yields tive simplicity of the equipment required.' In mium by formation of active silver chromate
ity of Ao are mixed with a sample containing addition, the procedure is often applicable with radioactive silver ion, precipitation of
W. grams of the inactive substance. After se- W. = Wo(~: - 1) (16-28) where the activation' method fails. magnesium or zinc by phosphate containing
paration and purification, a weight W. or the phosphorus-32, and the determination of fluo-
species is found to have an activity of A •. We The substoichiometric procedure is advan- ride ion by "pl'eCipitation with radioactive cal-
may then write tageous when the amount recovered, W., is cium.
so small that its weight is difficult to assess. Radiometric titrations employ a radioactive
compound for preparation of the standard
One radiometric method employs a radioac- solution. Usually, the reaction between the
Application of the tive reagent of known activity to isolate the
which rearranges to analyte and the standard involves precipitate
Isotopic Dilution Method
analyte from the other components of a sample. formation with the activity of the supernatant
Ao The isotopic dilution technique has been em- After quantitative separation, the activity of the liquid being monitored as the titration pro-
W'=AW,-Wo
• ployed for the determination of about 30 ele- product is readily related to the amount of the gresses. An example is the titration of silver
ments in a variety of matrix materials. 6 Sub- species being determined. Methods for the ion with a bromide solution that is enriched
Thus, the weight of the species originally pres-
stoichiometric methods have proved useful for radiometric determination of more than 30 of with radioactive bromine. Until the equiva-
ent is obtained from the four measured quanti-
determining traces of several metallic elements. the common elements' have been described. lence point is reached, essentially no radio-
ties on the right-hand l!ide of Equation 16-25.
Where the activity of the tracer is large, the For example, fractions of a microgram of cad- activity is found in the supernatant liquid. After
mium, copper, mercury, or zinc have been equivalence, a linear increase in count as a
weight Wo added can be kept small, and Equa-
tion 16-25 simplifies to determined by a procedure in which the ele- , For a review of rldiometric methods, see: T. Brauer
function of volume is observed. Clearly, pre-
ment is isolated for counting by extraction and J. Toll)'CIIY,Radlomnrlc TIt'atto",. New York: Per- cautions are needed to assure prompt coagula-
AoW with a substoichiometric amount of dithizone pmon Press, 1967. tion and settling out of the precipitate.
W'=A • in carbon tetrachloride .
• Isotopic dilution procedures have been most
Substoichiometric Isotope DilutiolL A sub- widely used for the determination of com- 1. Potassium-42 is a p-emitter with a half-life of 12.36 hr. Calculate the
stoichiometric method, analogous to that for pounds that are of interest in organic chemistry fraction of this isotope remaining in a sample after (a) 2 hr; (b) 5 hr;
activation analysis, can also be used in is0- and biochemistry. Thus, methods have been (c) 30 hr; (d) 60 hr.
topic dilution experiments. Here, identical developed for the determination of such diverse 2. Calculate the fraction of the following isotopes that remains after
amounts Woof the tracer are added to two substances as vitamin 0, vitamin 812, sucrose, 30 hr (half-lifes are given in parentheses):
solutions that are the same in every respect insulin, penicillin, various amino acids, corti- (a) yttrium-90 (64 hr)
except that one contains the sample and the costerone, various alcohols, and thyroxine. (b) silicon-31 (2.6 hr)
other does not. A suitable reagent is then added Isotopic dilution analysis has had less wide- (c) gold-198 (2.69 days)
to isolate a quantity W. oCthe species of interest spread application since the advent of activa- (d) zirconium-95 (65 days)
from each. Care is taken to be sure that the tion methods. Continued use of the procedure
3. A BaS04 sample contains 1.20 microcurie of barium-128 (tl/2 = 2.4
amount of added reagent is, however, less than days). What storage period is needed to assure that its activity is less
that required for complete removal of the than 0.01 microcurie?
species; thus, W. is identical for the two solu-
6 11is of inlerest that tbe dilution technique has also had 4. Estimate the standard deviation and the relative standard deviation
tions. Equation 16-24 describes the activity of
nonchemical applications. One application has heen to associated with counts of (a) 50.0; (b) 500; (c) 5000; (d) 5.00 x 104•
the product from the solution containing the the estimation of the size of salmon spawning runs in
sample; for the solution having no sample, W. 5. Estimate the standard deviation (in absolute and relative terms) asso-
Alaskan coastal streams. Here. a small fraction of the
is zero, and Equation 16-24 takes the form salmon are trapped. tagged, and returned to the river. A ciated with a counting rate of 150 cpm that is observed for (a) 40 s;
second trappios then takes place perhaps 10 miles up- (b) 80 s; (c) 4.0 min; (d) 12.0 min.
stream and the I'raction of taged salmOll is determined. 6. Estimate the absolute and relative.uncertainty associated with a mea-
The total salmOllpopulation is readily calculated from this
surement involving 600 counts at the
informatiOlland from the number orisinaUy tagecl. The
assumptiOllmust, of course, be made that the fish popula. (a) 50% confidence level.
Recall, however, that conditions have been tion becomes homosen= during its travel between (b) 90% confidence level.
chosen such that W. = W;. As a consequence, statiOlls. (c) 99% confidence level.
7. Estimate the absolute uncertainty at the 90% confidence level for a and assayed for their II activity. Calculate the weight of 1- in each
measurement that involves a total count of (a) 64; (b) 200; (c) 422; (d) milliliter of the sample solution if the activity (corrected for back-
1025. ground) for the electrode from the cell containing the standard was
8. Estimate the absolute and relative uncertainty at the 90% confidence 5730 cpm, while that from the cell containing the unknown was 3960
level associated with the corrected counting rate obtained from a total cpm ..
counting rate of 250 cpm for 15 min and a background count of
(a) 9 cpm for 2 min.
(b) 9 cpm for 10 min.
(c) 18 cpm for 2 min.
(d) 40 cpm for 2 min.
9. The background activity of a laboratory when measured for 3 min
was found to be approxiniately 11 cpm. What total count should be
taken in order to keep the relative uncertainty at the 90 % confidence
level smaller than 5.0~",given a total counting rate of about (a)
120 cpm; (b) 250 cpm; (c)SOO cpm?
10. If a total of 25 min is available, calculate the best division of counting
time between the background and sample for each of the counting
rates in Problem 9. What will be the expected relative standard devia-
tion for the analysis if the only significant uncertainty lies in the
counting process? ,
11. A 2.00-ml solution containing 0.120 microcurie per milliliter of tritium
was injected into the bloOdstream of a dog. After allowing time for
homogenization, a l.00-mt sample of the blood was found to have a
count corresponding to 127 disintegrations per second. Calculate the
blood volume of the animal.
12. In order to determine the mercury content of a specimen of animal
tissue, a 0.652-g sample of the tissue and a standard solution contain-
ing 0.102 Ilg of Hg as HgClz were irradiated for 3 days in a thermal
neutron flux of IOU neutrons/cmz sec. After irradiation was complete,
20.0 mg of Hg as HgzClz was added to each. Both were digested in a
nitric acid/sulfuric acid mixture to oxidize organic material; suitable
precautions were taken to avoid loss of mercury by volatilization.
Hydrochloric acid was then added, and the HgClz formed was distilled
from the reaction mixtures. The mercury in each of the distillates was
deposited electrolytically on gold foil electrodes, resulting in an increase
in weight of 17.6 mg for the sample and 16.5 mg for the standard. The
"I activity due to 197Hg w:as then determined. The sample was found
to yield a count of 750 cpm and the standard 1009 cpm. Calculate
the ppm Hg in the sample.
I
13. Identical electrolytic cells; fitted with silver anodes and platinum cath-
odes, were arranged in series. Exactly 1.00 ml of a solution containing
3.75 x lO-z mg of KI labeled with 1311(a II-emitter with a half-life of
8.0 days) and 5.00 ml of an HOAc/OAc- buffer were introduced to
one cell. A 5.00-ml aliquot of an iodide-containing sample was added
to the acetate buffer contained in the second cell. After passing a
substoichiometric quantity of electricity, the anodes were removed
A mass spectrum is obtained by converting
.the compounds of a sample into rapidly mov-
ing ions (usually positive) and resolving them The principles of mass spectral measurements
on the basis of their mass-t<H:harge ratio. The are simple and easily understood; unfor-
utility of mass spectrometry arises from the tunately, the simplicity does not extend to the
fact that the ionization process generally pro- instrumentation. Indeed, the typical high-
duces a family of positive particles whose mass resolution mass spectrometer is a complex
distribution is characteristic of the parent electronic and mechanical device that is ex-
species. As a consequence, a mass· s~rum pensive in terms of both initial purchase as
provides information that is useful for elucidat- well as operation and maintenance costs.
ing chemical structures. Mass spectl1ll data are
easier to interpret than infrared and NMR Instrument Components
spectra in some respects, since th~y provide
information in terms of molecular mass of the Figure 17-1 is a block diagram showing the
structural components of a sample; in addition, major components of a mass spectrometer,
an accurate measure of the molecuJar weight which are similar in function to the compo-
of the analyte can usually be obtained from nents of the optical instruments shown in
the data. Figure 5-1 (p. 115). The mass analyzer is a
Mass spectra can also be employed for dispersing device analogous to the prism or
the quantitative analysis of complex mixtures. grating of an optical spectrometer; however,
Here, the magnitude of ion currelJts at var- it disperses particles rather than electromag-
ious mass settings is related to concentration. netic radiation from the sample. A characteris-
Mass spectrometry evolved froin studies, tic feature of mass spectrometry, which is not
carried out at the beginning of this century,. encountered in most optical methods, is the
that were concerned with the behavior of pos- need to maintain all of the components leading
itive ions in magnetic and electrostatic fields. up to the detector at low pressures (10-4 to
During the following two decades, the method 10- 8 torr); thus, the elaborate vacuum systems
was refined and provided important informa- are an important part of mass spectrometers.
tion concerning the isotopic abundance of vari- Figure 17-2 shows schematically the essen-
ous elements. The first true analytical applica- tial parts of a typical analytical mass spectrom-
eter. Its operation is based on the following
. tion of mass spectrometry was described in
1940 when reliable instruments became com- sequence of events. (1) A micromole (or less)
mercially available. The focus of this early of sample is volatilized and allowed to. lea~
slowly into the ionization chamber, which IS
analytical work was toward quantitative deter-
mination of the components in complex hydro- maintained at a pressure of about 10-5 torr.
carbon mixtures. The mass spectrometer has (2) The molecules of the sample are ionized
since proved to be an invaluable tool for this directly or indirectly by a stream of electrons
work; it is widely used in the .petroleum
industry. '
Beginning in about 1960, interest in mass I Rererence works on mass spectromelry include: R. W.
Kiser, Introduction to Mass S~ctrometry and Irs Appli£a-
spectrometry shifted toward its ~e for the lions. Englewood Cliffs, NJ.: Ptenlice·Hal~ 1965; J.
identification and structural analysis of com- Roboz and E. Chait, in Physi£al Methods of Chemi&try,
plex compounds. Within three or four years, eds.· A. Weissberger and B. W. Rossilor. New York:
it became recognized as a powerful and versa- Wiley, t977, Part VI, vol. I, Chapter 3; J. Roboz, M_
Spectromelry. New York: IntencieDce, t968; and PrQCfI-
tile tool for this purpose-on a par with or
col M_ Spectrometry, cd. B. S. Middlcdilch. New York:
perhaps more important than infrared or Plenum Press, 1979.For a brier review or the field, ICe:
nuclear magnetic resonance spectroscopy. I W. V. Ligon. Jr, Science,1OS. 151 (1979~
as well as upon the field strength. Particles directly into the ion source by means of a
of different mass can be focused on the exit slit sample probe, which is' inserted through a
by varying the accelerating potential or the vacuum lock. The probe consists of a holder
field strength. (5) The ions passing through the for a small capillary tube or cup that contains
exit slit fall upon a collector electrode; the ion a microgram (or less) of the sample. A principal
current that results is amplified and recorded advantage of the probe is the minute quantity
as a function of field strength or accelerating of sample required. The probe is equipped with
potential: a heater to volatilize the sample; at the low
pressure of the ion source, however,. much
lower temperatures are needed to producesuf-
ficient gaseous molecules of the sample. The use
s.nple Handling System
of a probe permits the study of such nonvolatile
The purpose of the sample handling system is materials as carbohydrates, steroids, and low-
to introduce a representative sample of the molecular-weight polymeric substances.
material to be analyzed into the ion source as GIS Chromatographic Inlet Systems. Sam-
flowing from the heated filament toward an thousand volts between A and B. A collimated a gas at a low and reproducible pressure. The ple requirements with regard to volatility
anode (both positive and negative ions are beam of positive ions enters the separation nature of the inlet system will differ depend- and quantity (,.., 1 pmol) are similar for both
formed by impact, but the former predominate; area through slit B. (4) In the analyzer tube, ing upon the physical state of the sample; gas chromatography (Chapter 26) and mass
analytical methods are generally based upon which is maintained at a pressure of about several systems must be available if all types spectrometry; thus, the effiuent from a chroma-
positive particles). (3) The positive ions are 10-7 torr, the fast-moving particles are sub- of samples are to be accommodated. tographic column can serve as a sample source.
separated from the negative by the small nega- jected to a strong magnetic field which causes The Batch Inlet. Most commonly, the sam- Enhanced sensitivity can be realized by sepa-
tive potential at slit A and are then accelerated them to describe a curved path, the radius of ple is introduced as a gas from a 1- to 5-liter rating the sample components from the large
by a potential of a few hundred to a few which depends upon their velocity and mass2 reservoir (see Figure 17-2). The pressure ofthe excess of carrier gas (usually helium) that is
sample in the reservoir is about one to two always present. This separation is readily
orders of magnitude greater than that within achieved by passing the effluent from the
the ionization chamber, so as to maintain a chromatographic column through a narrow
steady flow through a pinhole into the cham- tube of porous glass or Teflon which is perme-
ber; sample pressures of O.ot torr are typical. able to helium atoms but not to the larger
For liquids boiling below 150°C, a suitable sample molecules. The latter are then bled into
quantity can be evaporated into the evacuated the ion chamber through a pinhole.
OulpUtto reservoir at room temperature. For less volatile The excellent separation qualities of gas
omplif",r.nd
recordet" samples, the sample and reservoir may be chromatography, combined with the powerful
heated, provided the compound is thermally identification properties of mass spectrom-
stable; otherwise, the sample must be intro- etry, provides the chemist with a most useful
duced directly into the ionization chamber, a tool for analyzing complex mixtures. This ap-
procedure that requires special equipment. plication of mass spectrometry is considered
Gaseous samples are readily handled byexpan- in Chapter 26.
sion of a small volume into the reservoir.
The Direct Probe Inlet. Nonvolatile or ther-
Ion Sources
mally unstable materials are often introduced
Numerous methods exist for converting mole-
cules into gaseous ions; many find application
in mass spectrometry.3
2 More correctly, the dependence is on the ratio ollllUl

to c:harae (III/e~Generally, however, the ions or interat


bear only a ainp charae; thus, the IlWI is frequently
usecI in lieu of the more cumbersome maa-to-c:harae • For a review or ion sources, see: E. M. C1Wt, Anal.
ratio. CheIII. 44 (3~ 77A (1972~
../
------------.-,--
Shield Electron / "-
natively, it may be mixed with graphite and must always be made. Discrimination be-
pressed into an electrode. A radio-frequency tween integral mass numbers suffices for some
/ Heater ~~/_ s1~it First ac:cel.slit '\
/ -/ Focus slit :--.. voltage of about 30 kV is applied across the applications. For others, however, much higher
I Gas beam ~ Second ac:cel.
electrodes, whereupon the sample vaporizes to resolution is necessary; thus, for example, iden-
MOlec~~~
~ Fil~:
,.",__ '{I __",,:
...~ent y~. . ~ '~,
....'",~~', ~"""
r" slit,'\\ \ form a gaseous ionic plasma. 'Pie ions in the tification of the ions C2Ht, CHzN+, Nt, and
== -. '--"17.i~.., . "-;;- plasma are then accelerated into the analyzer
by a suitable dc potential.
CO+ (mass numbers 28.031, 28.019, 28.006,
and 27.995, respectively) requires a resolving
Repellers ~'
\ Ionizing region Ion ac:c;,erating II Field Ionization Sources. Several other ion power of about O.oI mass unit.
\ Electron beam region / sources have been developed and may prove The main difference among the various mass
\ . Anode \ to be important. One of these is the field ioni- spectrometers lies in their systems for sepa-
~-----------~~~/ zation source,5 which consists of a metallic rating ions.
anode in the form of a sharp blade or one or Singie-FOClIiiDg ADalyzen with Magnetic
FIGURE 17-3 An ion source. (From R. M. Silverstein, G. C.
more fine points (in one version, an array of Deflection. Separators of this kind employ a
Bassler, and T. C. Morrill, Spectrometric Identification of Organic
1000 points spaced 25 IlrJ1 apart is used) and circular beam path of 180, 90, or 60 deg; the
Compounds, 3d ed. New York: John Wiley & Sons, Inc., 1974 p. 7.
a cathode which also serves as a slit; the second is illustrated in Figure 17-2. The path
With permission.)
anode and cathode are located 0.5 to 2 mm described by any given particle represents a
apart. When a potential of 5 to 20 kV is balance between the forces that are acting
Electron Impact Source. The most common source at a pressure of perhaps 1 torr. The applied, the gas phase in contact with the fine upon it. The magnetic centripetal force FAt is
way of producing ions for mass spectrometry sample is then introduced at a concentration electrode is subjected to an electrical force given by
involves bombarding the sample with a beam . that is 10-3 to 10-4 times that of the reagent. field as large as 108 VIem. which suffices to
of energetic electrons. Figure 17-3 is a sche- . Ionization of the sample is largely a conse- cause ionization of organic compounds.
matic diagram of a typical electron ionization quence of collisions with reagen~ ions rather Field ionization is a gentle technique, in where H is the magnetic field strength, v is
source or ion gun. The positive ions produced than electrons. Interaction of the clectron beam the sense that fragmentation is minimized; the particle velocity, and e is the charge on
on electron impact are forced through the slit with the reagent gas produces a 'host of ions, the molecular ion and M + 1 ion are often the the ion. The balancing centrifugal force Fe
oCthe first accelerating plate by a small poten- such as CHt, CHt, CHt, Ht, and CzHt. major products. In structural investigations, it can be expressed as
tial difference between this plate and the repel- Several of these, such as CHt, are strong is often advantageous to obtain spectra with
both a field ionization or chemical ionization mv2
ler. The high potential between the first and proton donors and thus react with the analyte F==-
e r
second accelerators gives the particles their molecules to form ions that are one unit larger source and an electron impact source; the for-
final velocities; the third slit provides further in mass (the M + 1 ion). For example, mer provides information about the molecular where m is the particle mass and r is the radius
collimation to the beam. In most spectrom- weight of the substance and the latter provides of curvature. Finally, the kinetic energy of the
eters, the potential impressed between the' fragmentation patterns that are useful for particle E is given by
accelerator slits provides the means whereby identification.
particles of a particular mass are focused on where MH is the analyte and MHt is its M + 1 E == eV == !mvz (17-3)
the collector. ion. Generally, the chemical ionization source Mass Analyzer where V is the accelerating voltage applied in
The electron impact source is sufficiently causes less fragmentation of the analyte and the ionization chamber. Note that all par-
Several arrangements exist for the resolution
energetic to cause a good deal of fragmenta- thus leads to simpler and more easily inter- ticles of the same charge, regardless of mass,
of ions with different mass-to-charge ratios.
tion of molecules, leading to a large number pretable mass spectra. are assumed to acquire the same kinetic energy
Ideally, the analyzer should distinguish be-
of positive ions of various masses. The com- Chemical ionization requires:modification during acceleration in the electrical field. This
tween minute mass differences; in addition, it
plex mass patterns that result are useful for of the spectrometer to sustain the higher pres- assumption is only approximately valid since
should produce a high level of ion currents.
identification. sure in the ion source region. , the ions will possess a small distribution of
As with an optical monochromator, to which
Chemical Ionization Source. 4 In a chemical Spark Source. A spark sourC¢ permits ion energies before acceleration.
the separator is analogous, these two proper-
ionization source, a reagent gas such as meth- formation from nonvolatile inorganic samples A particle must fulfill the condition that
ties are incompatible; a design compromise
ane is introduced into a modified electfl~n beam such as metals, semiconductors, and minerals. F II and Fe be equal in order to traverse the
Here, a radio-frequency spark source, similar circular path to the co11ector; thus,
to that discussed in Chapter 12, is employed
in place of the electron source. Often, the sam- A""'. eM._ mvZ
4 For a review of this type of source, see: B. Munson, 5 See: M. Anbar and W. Aberth, <t6, 61A Hev •••- (17-4)
Anal. Ch ••••_ 49 {9~ 772A (1977~ ple may serve as one or both electrodes; alter- (1974~ r
Substituting Equation 17-4 into Equation 17-3
and rearranging gives
unit thus becomes possible. The ion currents
produced are extremely small, however, and
require large amplification for detection and I
I "
I \
\
\
m HZ,-2 I \
-=-- recording. Double-focusing mass spectrom-' I \
e 2V eters are available commercially; their cost is . I \
I \

I
In most spectrometers, H and r are fixed; high and their maintenance difficult. I \
Accelerltlng
therefore, the mass-to-charge ratio of the par- Figure 17-5 shows a second type of double- ~
Spark slit I 31°50' \
ticle reaching the slit is inversely proportional fOCusinginstrument, which employs Mattauch- . source "-.,1 I \
Herzog geometry. This geometry, which results I \
to the acceleration voltage. Most particles
possess a single unit of positive charge; thus, in all ions of various mass-to-charge ratios 1II1 . I
I \
\
any desired mass can be focused on the exit being focused on a single focal plane, is par- I

slit by suitable adjustment of V. ticularly useful for photographic detection. The


Double-Focusing Aualyzers. The ability of dispersion along the focal plane is linear with
\
the single-focusing instrument to discriminate respect to mass-to-charge ratio; thus, identi- \
~90"
between small mass differences (that is, its re- fication of lines is readily accomplished if the \ '\.-
\--
solving power) is limited by the small varia- mass of the particles responsible for two ofthe Photographic
tions in the kinetic energy of particles of a lines can be determined. pllte
given species as they leave the ion source. Tune-of-Flicht Aaalyzen. Ion separation
FIG U R E 17-5 Double-focusing spark -source mass spectrometer, Mattauch-Herzog type.
These variations, which arise from the initial in a time-of-flight instrument is achieved by
distribution of kinetic energies of the neutral nonmagnetic means. Here, the positive ions
molecules, cause a broadening of the ion are produced intermittently by bombardment
beam reaching the collector and a loss of re- with brief pulses of electrons. These pulses,
solving power. In a double-focusing instru- which are controlled by a grid, typically have struments employing time-of-ftight mass sepa- and ease of a~ibility to the ion source,
ment, shown schematically in Figure 17-4, the a frequency of 10,000 Hz and a lifetime of rators are less satisfactory than those based which allows t~ ready insertion of nonvolatile
beam is first passed through a radial elec- 0.25 JlS. The ions produced are then accelerated upon magnetic focusing. On the other hand, or heat-sensitive samples: The instantaneous
trostatic field. This field has the effect of fo- by an electrical field pulse that has the same several advantages partially offset these limita- display feature is also useful for the study of
cusing only particles of the same kinetic energy frequency as, but lags behind, the ionization tions.Included among these are the ruggedness short-lived species. In general, time-of-ftight
on slit 2, which then serves as the source for pulse. The accelerated particles pass into a
the magnetic separator. Resolution of par- field-free drift tube about a meter in length
ticles differing by small fractions of a mass (Figure 17-6). As noted earlier, all particles
entering the tube have the same kinetic ener-
gies; thus, their velocities in the drift tube must
vary inversely with their masses (Equation
17-3~ with the lighter particles arriving at the = i_
I I Ions
cathode
Ion

I
collector earlier than the heavier ones.
The detector in a time-of-ftight mass spec-
Simple ---... ttJ:===:==::::_-_-_-r
trometer is an electron multiplier tube, similar
:.J :-----F;;iif;-,;----~
Ionization separation L.:
in principle to a photomultiplier tube (p. 140); region . region
the output ofthis tube is fed across the vertical
deflection plates of a cathode ray oscilloscope
Acceleration
region
~ Anode
1 Electron
multiplication
Vacuum region
(p. 79~ the horizontal sweep is synchronized
From To
with the accelerator pulses, and an essentially
ion detector instantaneous display of the entire mass spec-
source
trum appears on the oscilloscope screen.
FIGURE 17-4 Design ofa double-focusing From the standpoint of resolution, repro-
separator. ducibility, and ease of mass identification, in- FIGURE 17-6 Schematic diagram of a time-of-ftight mass
spectrometer.
FIG~RE 17-7 A quadrupole mass spectrometer. [From
D. LIChtman, Res. Dev., 15 (2), 52 (1964). With permission.]

FIGURE 17-8 Mass spectra of n-butane recorded simul-


taneously by four galvanometers of different sensitivities.
instruments are smaller, more mobile, and Measurement and (From top to bottom, galvanometer sensitivities are in the
more convenient to use than their magnetic Display of Ion Currents ratio 30: 10: 3 : 1. (Courtesy of E. I. DuPont de Nemours &
focusing counterparts. Company, Wilmington, Delaware. With permission.)
The ions from the separator pass through a
Quadrupole Analyzer. The quadrupole spec-
slit and are collected on an electrode that is
trometer employs four short, parallel metal
well shielded from stray ions. In many instru-
rods arranged symmetrically around the beam
ments, the current produced is passed through
(see Figure 17-7). The opposed rods are con·
a large resistor to ground, and the resulting
nected together, one pair being attached to the amplified electron current which can then be Computerized Mass Spectrometers
potential drop is impressed on a field effect
positive side. of a dc source and the other pair further amplified electronically and recorded.
transistor or the grid of an electrometer tube. Minicomputers and microprocessors are an
to the negatIve terminal. In addition, a radio- Ion currents are ordinarily small and vary
The resulting current is then further amplified integral part of most modem mass spectrom-
fr~uency ac potential is applied to both pairs. over a range of 10-17 to 10-9 A (the former
before being recorded. Alternatively, the ions eters.7 A characteristic of a mass spectrum is
Nelt~er fi~ld acts to accelerate the positive current corresponds to about 60 ions reach-
from the separator may strike a cathode surface the wealth of structural data that it provides.
partIcles ejected from the ion source. The com- ing the detector per second). In order to
and cause electron emission. The electrons For example, a molecule with a molecular
bined fields, however, cause the particles to record peaks of such diverse size, most mass
form~ are accelerated toward a dynode and, weight of 400 to 500 may be fragmented by
oscillate about their central axis of travel' s~rometers are equipped with several pens,
upon Impact, produce several additional elec- an electron beam into as many as 200 different
only thosi: with a certain mass-to-charge rati~ WIth each responding to a different current
trons. This process is repeated several times as ions, each of which leads to a discrete spectral
can pass through the array without being re- range. Thus, a peak of convenient size can
in a photomultiplier tube, to produce a highly peak. For a structural determination, the
moved by collision with one of the rods. Mass always be found on the chart regardless of the
heights and mass numbers of each peak must
scanning is achieved by vary~g the frequency current magnitude. Some instruments incor-
be determined, stored, and ultimately dis-
of the ac supply or by varying the potentials porate a set of mirrored galvanometers which
of the two sources while keeping their ratio record the spectrum on sensitized paper; each
constant. Quadrupole instruments are com- • For a thorough treatment or quadrupole mus spectrom-
sa,tvanometer ~as a different sensitivity to per- 7 For a detailed discussion of computerized mus spec-
pact and less expensive than magnetic focusing etm, see: Quodrupole MlW Sp«trometry IUIdlIS App/IcQ- mIt the recordmg of peaks of widely different trometry, see: J. R. Chapman, Compuler. in Mass
instruments.6 lions, ed. P. H. Dawson. New York: Elsevier, 1977. sizes. A typical tracing is shown in Figure 17-8. Spectrometry. New York: Academic Press. 1978.
MASS SPECTROMETRY 487

-
RoOT lIS
played. Because the amount of data is so large, am be digitized and superimposed on the ion- O:ID .•• as leAN u DATE 1'4'71
...,RATE II leTlME I IllS'. "I
it is essential that the data acquisition and current signaL The computer is then program- KIIlASS THil_ 1
processing be rapid; the computer is ideally med to employ the resulting signal for the _ITUllATE
suited to these tasks. Moreover, in order for assignment of mass. For magnetic instruments,
mass spectral data to be useful, several instru- a Hall probe, which measures the magnetic
•• CIlGII
IGNOIIE I.
I
.•...
IlASE
I. •
I
SUBTIIT I

mental variables must be closely controlled field strength. am serve to provide mass data. 1'.5. '.I.a
I" smuEIl
•••••
1'1
"'E
during data acquisition. Computers or micro- The computer output is often displayed ~6 ~ 61 I I~ 31 IN 4 U~ ~ 166 I
processors are much more efficient than a both in digital form and as a graph. Figure 31 5 61 I 14 IS I" 14 134 I 167 I
51 II 63 5 IS ~I "I • 135
• 161 3
human operator in exercising such controls. 17-9 is an example. The odd columns in the Sf 13' 64 I 16 I
• ..~
111 II 1S6 I- 169 5

.~
41 51 61 13 IT 111 73 13T 6 i7. I
These considerations have led to the wide- digital display list mass numbers in an in- 61 364 66 13 II I II 131 T III 6
spread incorporation of mini- and microcom- creasing order. The even columns contain the ••
6S
13
311
6T
II
sa
3S
'1
'1
5
3
116
115
II
I
U•
161 II
6 183
lIS
7
I
corresponding ion currents normalized to 66 16 69 III 6 116 I 161 1.6 191 I
puters into modem commercial mass spec- 65 6 TI 67 21 117 I 142 71 193 I
'4
trometers. the largest peak (the base peak1 which is sa 5 71 19 t5 17 II' I 165
•• 195 I
•• • ••
51 T2 5 II III I 144 I 197 61
The interface between a mass spectrometer found at mass 156; the current for this par-
ticle is assigned the number 1000 and all
51
53
16
64
T3
T6 I
97
'1
55
liT
III
122

6
151
ISI
I
I
198
199
II
2
and a computer usually has provisions for
digitizing the amplified ion-c:urrent signal plus other peak heights are relative to this one.
54
55
51
161
7S
7T
5
II I"
•• ~• lIS
126
5
7
IS.
1S5 113
6 116
217 6
I

several other signals that are used for control Thus, the peak at mass 141 is 82.6% of the
56
57
31
17
TI
7t Ii
S III
103 •
I
125
126
6
II
156 II"
1ST 271
lei

"'
I
I

of instrumental variables. Examples of ·the base peak.


sa
59
II
S
I ••
II
21
II
115
116
lIT
I
••
III
lit
151
14
••
ISI
IS.
It
3
127
211
6
I
5 I' I~ 2 161 I
latter are source temperature, accelerating As with infrared spectroscopy, computer-
voltage, sam rate, and magnetic field strength. stored library files of mass spectra are avail-
The digitized ion-c:urrent signal ordinarily able; some commercial instruments are pro- 100 0 156
requires considerable processing before it is grammed to search these files for spectra that
CH CH
ready for display. First, the peaks must be match that of the analyte. HN:J
~
2 CH
(
3
21.CH3
normalized, a process whereby the height of
o N 0
each peak relative to some reference peak is H
Resolution of Ma •• Spectrometers
calculated. Often the largest peak serves as
the reference and is arbitrarily assigned a The capability of a mass spectrometer to dif-
'I
.5
peak height of 100 (or sometimes 1000). Each
peak must also be assigned a mass number.
This assignment is frequently made on the
ferentiate between masses is usually stated in
terms of its resolution mlAm. where m and
m + &II are the masses of two particles that
..
~
a::

basis of the time of its appearance and the give just separable peaks of equal size. Two
sam rate. Periodic calibration is necessary; peaks are considered to be separated if the
for this purpose, a ftuorinated hydrocarbon is height of the valley between them is no more
used. For high-resolutiotl work, the standard than 10% of their height.
may be admitted with the sample. The com- The resolution needed in a mass spectrom- Mas>

puter is then programmed to recognize and em- eter depends greatly upon its application. For R25 =~I~lion 100
ploy the peaks of the standard as references for example, discrimination between particles of 32 Seen number
mass measurement. For low-resolution instru- the same nominal·mass, such as Nt (mass
ments, the calibration must generally be ob- 28.006) and CO+ (mass 27.995), requires an FIGURE 17-9 A computer display of mass-
tained separately from the sample because of instrument with a resolution of several thou- spectral data. The compound was isolated from
the likelihood of peak overlaps. sand. On the other hand, low-molecular-weight a blood serum extract by chromatography. The
Variables other than time may be employed particles differing by a unit of mass or more spectrum showed it to be the barbituate, pento-
for mass assignments. Thus, with quadrupole (NH; and CHt, for example) am be dis- barbital. The instrument was a DuPont Model •
spectrometers, the mass of the ion reaching tinguished with an instrument having a res0- 21..()94computerized mass spectrometer. (Cour-
the detector is proportional to the ac voltage lution smaller than SO.A spectrometer with a tesy of DuPont Instruments, Wilmington, Dela-
applied to the rods. Hence, a rod voltage signal resolution of perhaps 2SOto 500 is needed for ware.)
unit separation where the sample yields large terpretation of a spectrum is seldom, if ever, For a molecule containins a large number cause of the presence of these elements, there
fralments. possible. of atoms, the number of different positive ions beyond the molccular ion peak (the M + 1
Table 17-1 lists typical resolutions and produced can be large. Their distn"bution de- peak ~ and on occasion at M + 2 as well.
mass ranges for various types of commercially pends upon the stability of the precursor ion Compounds containing chlorine or bromine
The Electron and the energy imparted to the molecule by
available mass spectrometers. The cost of will clearly have relatively large M + 2 peaks.
Impect loniZ8tion Proceu the electron beam. Fortunately, when the beam
these instruments is related to their resolution Peaks for CoUision Products. Ion-molecule
and range. Unit resolution to mass number Generation of a mass spectrum requires some energy exceeds 50 to 70 eV, the pattern of collision can produce peaks of higher mass
250 requires an instrument with a purchase minimum electron-beam energy (7 to 15 eV products from a given molecule becomes more number than the molecular ion peak (Equa-
price of roughly $30,000. A high-resolution, for most organic compounds) to initiate the or less reprodUCIble. tion 17-9). At ordinary sample pressures, the
double-focusing instrument costs over ionization proceSs The neutral fragments in Equations 17-7 to only important reaction of this type is one in
$100,000. Low-resolution instruments with 17-9 are shown as radicals, but they may also which the collision process transfers a hydro-
M+e ....•M+ +2e occur as molcculcs. Neither will reach the
limited range are priced between $5000 and gen atom to the ion; an enhanced M + 1 peak
$10,000. where M represents the molecule and M + is detector. results. The proton-transfer process is a see-
the molecular iOn, or parent ion. Small in- ond-order reaction, and the amount of product
creases above the minimum in electron-beam depends strongly upon the reactant concentra-
energy produce ta higher yield of molecular The Molecule' Ion tion. The height of an M + 1 peak of this type
ions, owing to the greater probability for ion- increases at a more rapid rate with increased
Even for relatively simple compounds, the·mass producing collisions. Large increases in beam The mass of the molecular ion M + is identi- sample pressure than docs the height of other
spectrum generally contains an array of peaks energy may result in a decrease in the molecu- cal to the molecular weight of the compound peaks; thus, detection ofthis reaction is usually
of differing heights (Figures 17-& and 17-9). lar ion peak; here, the added energy causes from which it is generated. Thus, the mass of possible.
The detailed nature of the spectrum depends bond rupture, With the formation of frag- this ion is an important parameter in the Stability of the MoI~ loa. For a given
upon the properties ofthe molecule as well as ments that have smaller masses (and occa- identification of a compound. For perhaps 80 set of conditions, the intensity of a molecular
upon the ionization potential, the sample pres- sionally larger trlasses as well) than the parent to 90% of organic substances, the molecular ion peak depends upon the stability of the
sure, and the instrument design. Complete in- molecule. TypiciI of these processes are the ion peak is readily recognizable; certain char- ionized particle; a minimum lifetime of about
following, illustrated with the hypothetical acteristics of this peak must be borne in mind, 10-5 s is needed for a particle to reach the
molecule ABCO: however. collector and to be detected. The stability of
Isotope Peaks. Table 17-2 lists the natural the ion is strongly affected by structure;
isotopic abundances of elements that the size of molecular ion peaks will thus show
frequently occur in organic compounds. Be- great variability.
ABCO + e ....•ABCO+ + 2e
ABCO+ ....•BCD' + A'+
B' + A+
/ TABLE 17-1 COMPARISON OF SOME TYPICAL
CD' + AB+ COMMERCIAL MASS SPECTROMETERS
'\.
A' + B+ Approximate Approximate
0' +C+ Type Mass Range Resolution
/
AB' +CO+ Double focusing 2-5000 10,000-20,000
'\. 1-240 1000-2500

ABCO+ ....•AOBC+ ~
c' + 0+
BC' +AO+ ! Rearrangement
followed by (17-8)
1-1400
2-700
2-150
1-700
1500
500
100
150-250
fragmentation
AO' + BC+ 0-250 130
ABCO+ + ABCO ....•(ABCO)t ....•BCD' + ABCOA + Collision (17-9) 2-100 100
followed by 2-80 20-50
fragmentation
68 OH NHCH, the M - 1 peak, which is formed from a reac-
In general, the molecular ion is stabilized a detectable parent ion peak provided the peak I I
by the presence of 7t electron systems, which is at least 1 % of the total; from some instru- . VCH-CH-CH3
tion such as
more easily accommodate the loss of an elec- ments, this limit is lowered to 0.1 %. Thus, with MH + C2H; ..• M+ + C2H6
tron. Cyclic structures also give large parent the exception of some alc;ohols and 'branched
peaks, since rupture of a bond docs not nee:es- hydrocarbons, the mass of the molecular ion, Clearly, the molecular, weight of the com-
sarily produce two fragments. In general, the and thus the molecular weight of the com- pound can be determined from this spectrum.
stability of the molecu1ar ion decreases in the pound, can be'determined.
following order: aromatics, conjugated olefins,
alicyclics, sulfides, unbranched hydrocarbons, The Base Peek
QUALITATIVE
mercaptans, ketones, amines, esters, ethers,
The largest peak in a mass spectrum is termed APPLICATIONS OF
carboxylic acids, branched hydrocarbons, and
the base peak; it is common practice to report MA$S SPECTROMETRY
alcohols. These effects are illustrated in Table
peak heights as a fraCtion of the base peak
17-3 (p. 492~ which compares the height of
height. Alternatively, intensities are reported as The:mass spectrum of a pure compound pro-
the molecular ion peak for some Cl0 com-
percentages of the total peak heights, a more vides valuable information for qualitative iden-
pounds relative to the total peak heights in
informative number, but one that is more tifi~tion purposes. In addition, mass spec-
the spectrum. .
laborious to calculate. trodletry has proved useful for identifying the
Figure 17-9 (p. 487) is a spectrum of a com-
components of simple mixtures-in particular, '
pound with an unstable molecular ion. Here,
ChemiclII lonIznon Spectra gaseous mixtures. Some of these applications
the molecular ion peak at 226 is absent and
are outlined in the paragraphs that follow.
the M + 1 is barely detectable. The base peak Figure 17-10 contrasts the electron-impact ion-
at mass 156, on the other hand, corresponds ization spectrum for ephedrine with its chcmi-
to a particle formed by the loss of the cal ionization counterpart. The electron-impact Mo~"r Weight Detennination
branched aliphatic side chain. spectrum (Figure 17-10.) consists largely of
A mass spectrometer will generally provide peaks corresponding to small fragments of the For'most compounds that can be volatilized,
the mass spectrometer is unsurpassed for the
determination of molecular weight. As we
have noted earlier, the method requires the
FIGURE 17-10 Mass spectra for ephed- identification of the molecular ion peak (or
rine (molecular weight -=
165) by (a) electron the M + 1 peak with chemical ionization~ the
TABLE 17 .•Z NATURAL ABUNDANCE OF ISOTOPES impact ionization and (b) chemical ionization mass of which gives the molecu1ar weight to
OF SOME COMMON ELEMENTS with methane as .the reagent. (From H. M. at least the nearest whole number-an accu-
Fales, in Mass Spectrometry, ed. G. W. A. racy that cannot be realized by other molecu-
MOlt Abundant Abundaoce of Other Isotopes Relative
Element" Isotope to 100 Parts of the MOlt AIMmda•••• Milne. New York: John Wiley & Sons, Inc., lar weight measurements. Caution must be
1971, p. 198. With permission.) used, however, for occasionally the molecular
Hydrogen lH 2H 0.016 ion peak may either be absent or so small
Carbon 12C UC 1.08 that it is confused with a peak caused by an
Nitrogen 14N UN 0.38
Oxygen 160 17
0 0.04 impurity (see Figure 17-10.). In addition, col-
180 0.20 lision processes may produce an M + 1 peak
Sulfur 32S 33S 0.78 parent molecule. The molecular ion peak 165 thai is more intense than the parent ion peak.
3~ 4.40 is entirely absent; as a result, the molecular No single method is available for establish-
Chlorine 35Cl 37Cl 32.5
79Br weight of the compound cannot be determined. ing 1unambiguously that the peak of highest
Bromine 81Br 98.0 maSs number (neglc:c\ing the small isotope
The base peak at 58 is due to the fragment
(CH3CH~NHCH3r . peaks) is indeed produced by the molecular
• FluoriDe ('9F). pboIpboruI (lIp). aud iodiDe (lUl) have DO ~ naturally occurrina Figure 17-1Ob is the chemical ionization ion. On the other hand, from a series of
isotopes.
spectrum for ephedrine. Here, a strong M + 1 observations, the experienced mass spectros-
• The Dumerical eatrieI iDdicate the avcrap Dumber or iIotDpic aloIIII praeal ror eadt 100 atODll
peak is observed, which arises from a reaction copist can ordinarily make this judgment
or the moat abuDdaDt iIotopc; thus, for every too lie aloIIII tbcR will be an averaac or t.08 lie
atoms. such as that shown on page 480. Note also with reasonable assurance of being correct.
As mentioned earlier, it is frequently possible It is of interest to point out that the molec- mass of the parent ion peak. This application, peak will be 6.48% of the M peak. The iso-
to identify an M + 1 peak by observing its ular weight determined by mass spectroscopy however, requires a high-resolution instru- topes of the other elements also contribute to
behavior as a function of sample size. Deter- will not be identical with that calculated from ment capable of detecting mass differences of this peak; we may tabulate their effects as
mination of whether or not the highest mass atomic weights on the chemical scale if the a few thousandths of a mass unit. Consider, follows:
peak is indeed the molecular ion is more parent compound contains certain elements. for example, the molecular weights of the
troublesome. Here, a knowledp of fragmenta- For example, methyl bromide will have a peak following compounds: purine, C,H4N4
C6H4Nz04
tion patterns for various types of compounds at mass 96 that is nearly as strong as the one (120.044); benzamidine, C,H,N2 (120.069);
is essential. For exampte. a peak at M - 3 at mass 94 because of the high isotopic abun- ethyltoluene, C,H12 (120.094); and aceto- 13C 6 x 1.08 = 6.48%
immediately casts doubt upon the peak at M dance of bromine-81 (see Table 17-2). phenone, C,H,O (120.157). If the measured zH 4 x 0.016 = 0.064%
mass of the parent ion peak is 120.069, then ISN 2 x 0.38 = 0.76%
being caused by the molecular ion. This pattern 170 4 x 0.04 = 0.16%
could only occur as the result of abstraction all but C,H,N2 are excluded as possible for-
Detennlnetlon of MoIecul., Fonnulu +
of three hydropn atoms from the molecular
ion; such a fragmentation is most unlikely. Partial or exact molecular formulas can be nations ofC, H, N, and °
mulas. Tables that list all reasonable combi-
by molecular weight
to the third decimal place have been com-
(M l)jM = 7.46%

On the other hand, a strong peak at M - 18 determined from the mass spectrum of a C12Hz4
(or M + 1 - 18) suggests that even a weak compound, provided the molecular ion peak piled. 'In addition, a table of molecular weights
peak at M may be the parent ion since, for can be identified. (to the sixth decimal place) of all of the com- 13C 12 x 1.08 = 12.96%
alcohols and aldehydes, the 10sI of water is a Moleealar Form ••• from Hiab-ReIOIutIoa pounds listed in the ninth edition ofthe Merck 2H 24 x 0.016 = 0.38%
common occUrrence; Figure 17-1Ob provides 1..vumeatL A unique formula for a com- Index is available.' (M + l)jM = 13.34%
an example of this eft'ect. pound can often be derived from the exact FOI'IlIlIIas from Isotopic Ratios. The data
from an instrument that can discriminate be-
It is seen in this example that a measure-
tween whole mass numbers provides useful
ment of the ratios of the (M + 1) to M peak
information about the formula of a com-
heights would permi~ discrimination betwee(l
pound, provided only that the molecular ion
TABLE 17-3" VARIATION IN MOLECULAR ION PEAK WITH two compounds that have identical whol~-
peak is sufficiently intense that its height and
STRUCTURE number mass weights. '
the heights of the (M + 1) and (M + 2) isotope
The use of relative isotope peak heights for
RelatlYe Peek HeIPt peaks can be determined accurately. The fol-
the determination of molecular formulas is
Compoaad F~ (perceut or total piak •••• ) lowing example illustrates this type of analysis.
greatly expedited with the tables developed
by Beynon10; a portion of a modified form of
EXAMPLE
Naphthalene
(X) 44.3 Calculate the ratios of the (M + 1) to M
peak heights for the following two compounds:
his tabulations is shown in Table 17-4. Here,
a listing for all reasonable combinations of C,
H, 0, and N is given for mass numbers 83
dinitrobenzene, C6H4N204 (M = 168) and an
o-C4H9 8.3 and 84 (the original tables extend to mass
n-Butylbenzene olefin, C12H24 (M = 168).
From Table 17-2, we see that for every 100 number 5(0); also tabulated are the heights of
12C atoms there are 1.08 13C atoms. Since the corresponding (M + 1) and (M + 2) peaks
trans-Decaline

Diamyl sulfide
(t) 82 there are six carbon atoms in nitrobenzene,
however, we would expect there to be 6.48
(6 x 1.08) molecules of nitrobenzene having
reported as percentages of the M peak. If a
reasonably accurate experimental determina-
tion ofthese percentages can be obtained from
(C,HllhS 3.7 a spectrum, a likely formula can be ascertainql.
n-Decane C1oH22 1.41 one 13C atom for every 100 molecules having
For example, a molecular ion peak at maSs
Diamylamine (C,Hll)2NH 1.14 none. Thus, from this effect alone the (M + 1)
84 and with (M + 1) and (M + 2) peaks pf
Methyl nonanoate C,H17COOCH3 1.10
5.6 and 0.3 % would suggest that the formula
Diamyl ether (C,HuhO 0.33 of the compound is C,H,O (Table 17-4).
3,3,5-Trimcthylheptane C10Hu 0.007 • J. H. 1IeyDon and A. E. williams. M __ A!nutfI<Ince
n-Decanol C1oH210H 0.002 Tabla for Ule In M_ Spectrr1IMIry. Amsterdam:
EIIevier. 1963.
'0 J. H. 1IeyDon and A. E. Williams, M __ A!nutfI<Ince
• Taken from K. Biermann, 11_ S~I. 0I'flIIIIc App/k«IoIu. New York: McGnw-HiU Book • Table uf MoI«ul4r Wrights. Rahway. N. J.: Merdt and Tabln for Uu ill M_ S~. AmIterdun:
Company. loe.. 1962, p. 52. With pennioaion. Co. loe.. 1978. Elsevier. 1963.
The isotopic ratio is particularly useful for mass number if they contain zero or an even
the detection and estimation of the number of number of nitrogen atoms; conversely, such
sulfur, chlorine, and bromine atoms in a com- fragments have an even mass number if the
pound because of the large contribution they nitrogens are odd in number. The rule is a
make to the (M + 2) peak (see Table 17-2~ direct consequence of the fact that, with the
Thus, for example, an (M + 2) peak that is exception of nitrogen, the valency and the
about 65 % of the M peak would be strong mass number of the isotopes of elements that t
evidence for a molecule containing two chlor- commonly occur in organic compounds are J
'0 50
ine atoms; an (M + 2) peak of about 4~o"on either both even or both odd. The nitrogen
the other hand, would suggest one atom of
sulfur. By examination of the heights of the'
(M + 4) and (M + 6) peaks as well, if is
rule applies to all covalent compounds con-
taining carbon, hydrogen, oxygen, sulfur, the
halogens, phosphorus, and boron.
J
sometimes feasible to identify combinations
of chlorine and bromine atoms. " Identification
The Nitrogen Rule. The nitrogen rule also of Compounds from
provides information concerning possible for- Fregment8tion Pettems
mulas of a compound whose molecular weight From Figure 17-11, it is evident that fragmen- 100
has been determined. This rule stales that all tation of even simple molecules produces a
organic compounds with an even molecular
weight lpust contain zero or an even number
large number of ions with ditTerent masses. A
complex spectrum results, which often per-
t
of nitr. atoms; all compounds with Odd mits identification of the parent molecule or J
'0 50
molecular weights must have an odd nUlJlber at least recognition of likely functional groups ~
of nitrogen atoms. The fragments formed by in the compound. Systematic studies of frag- "I!
cleavage of one bond, however, have an Odd l
mentation patterns for pure substances have

00

TABLE 17-4 ISOTOPIC ABUNDANCE RATIOS FOR


VARIOUS COMBINATIONS OF CARBON, HYDROGEN,
OXYGEN, AND NITROGEN··b

M=83 M=84 t
M+l M+2 M+l M+2
~ 50
C2HN~0 3.36 '0.24 C2H2N3O 3.38 0.24
C 2H
C3H O2
C3H3N2O
N•
3.74
3.72
4.09
0.06
0.45
0.27
C2H.N.
C3H2N02
3.75
3.73
0.06
0.45
0.27
J
C3H.N2O 4.11
CiH,N3 4.47 0.08" C3H,N3 4.48 0.81
C.H302 4.45 0.48' C4H.02 4.47 0.48
C4H,NO 4.82 0.29 C4H,NO 4.84 0.29 mI.
C4H7N2 5.20 0.11i C4H.N2 5.21 0.11 (el
C,H7O 5.55 0.33; C,H.O 5.57 0.33
C,H,N 5.93 0.15 C,H1ON· 5.95 0.15
FIGURE 17-11 Mass spectra of some simple compounds. (a) n-Heptane;
C~l1 6.66 0.19 C,H12 6.68 (b) l-pentanol; (c) n-heptanaI.
0.19

• Tlkeo from R. M. s~ G. c. Bassler. UId ·C.C. Morrill, Sp«t..-rlc Ithnt!/iaJt/oll t{


Or,.,ac
• Data
C_",.".
are liven •
3d ed. New York: John ~dey cI:.SoaI. IIIC., .974, p. 43. With permission.
perIlCD•••• or pat beilht or M.
led to rational fragmentation mechanisms and ing closely related components. As mentioned refers to the ion current at mass m for com- halogenated aromatic derivatives, and ar0-
a series of general rules that are helpful in earlier, the first commercial mass spectrom- ponent n. The partial pressure' of component matic nitriles.
interpreting Spectrall It is seldom possible eters were developed in about 1940 for deter- II in the mixture is given by P•• The value of Mass spectrometry has also been used for
(or desirable) to acx:ount for all of the peaks mining the constituents in volatile hydro- i.. is determined for each component by cali-
bration with a standard at a known partial
the characterization and analysis of high-
in a spectrum. Instead, characteristic patterns carbon samples of the kind encountered in . molecu1ar-weight polymeric materials. Here,
of fragmentation are sought. For exampJe. the the petroleum industry. Within a decade, how- pressure P.. Substituting the measured ion '.the sample is first pyrolyzed; the volatile prod-
top spectrum in Figure 17-11 is characterized ever, the method had been applied successfully currents for a mixture into the equations per- ucts are then led into the spectrometer for
by clusters of peaks differing in mass by 14. to the analysis of a large number of other mits solution of the simultaneous equation to examination. Alternatively, heating can be per-
Such a pattern is typical of straight-chain par- volatile compound types. More recently, mass give the partial pressures of each of the com· formed on the probe of a direct inlet system.
affins, in which the successive loss of a methy- spectroscopy has been adapted to the analysis ponents in the sample. Some polymers yield essentially a single frag-
lene group results in the observed mass de- of both inorganic and organic mixtures of low PreciIioa ••• Aeearaey. The precision of ment; for example, isoprene from natural ru~
creases. This same pattern is evident in the Ieft- volatility. quantitative mass spectral measurements ap- ber, styrene from polystyrene, ethylene from
hand parts of the two lower spectra as well. pears to range between 2 and S % relative. The polyethylene, and CF 2-eFCI from Kel-F.
Quite generally, the most stable hydrocarbon accuracy varies considerably depending upon Other polymers yield two or more products
Quantitative
fragments contain three or four carbon atoms the complexity of the mixture being analyzed that depend in amount and kind upon the
Analysis of Orpnic Mixtu •.••
and the corresponding peaks are thus the and the nature of its components. For gaseous pyrolysis temperature. Studies of temperat~
1a1'Flt.
Alcohols usually have a very weak or non-
•• RI""".'" The basic requirements
for a successful mass spectrometric analysis
hydrocarbon mixtures containing S to 10 effects can provide information regarding tlie
components, absolute errors of 0.2 to 0.8 mole stabilities of the various bonds, as well as the
existent parent iOn peak but lose water to are: (1) each component must exhibit at least percent appear to be typical. approximate molecular weight distribution.
give a strong peak at (M - 18) (see Figure one peak that differs markedly from the
17-11b). Cleavage of the C-e bond next to others; (2) the contnbution of each compo-
an oxygen is also common, and primary a1co- nent to a peak must be linearly additive; (3) Component Analysis Component
hols always have a strong peak at mass 31 the sensitivity (ion current per unit partial
The literature dealing with quantitative appli- Type Determination
due to the ion CHz -oH+. pressure) must be reproducible to perhaps
cations of mass spectrometry is so extensive
The interested reader should refer to refer- 1% relative; and (4) suitable standards for Because of the complex nature of petroleum
ence literature for further generalizations con- calibration must be available. as to make a summary difficult.13 The listing
products, quantitative data as to types of
cerning the identification of organic com- of typical applications assembled by Melpol-
CaIibntiou. Quantitative mass spectrom- compounds are often more useful than analysis
pounds from mass spectrometric data.12 etry is based upon empirical calibration with der and Brown demonstrates clearly the ver-
for individual components. Mass spectrometry
standards. UDder appropriate conditions, mass satility of the method. For example, some of can provide such information. For example,
the mixtures that can be analyzed without it has been found that paraffinic hydrocarbons
peak heights an: dim:tly proportional to the
QUANTITATIVE sample heating include natural gas, Ca-e,
partial pressures of the components. For com- generally give unusually strong peaks at masses
APPLICATIONS OF hydrocarbons; C,-e. saturated hydrocar-
MASS SPECTROMETRY
plex mixtures, it is seldom possible to find a 43, S7, 71, 8S, and 99. Cycloparaffins and mono-
bons; C1-e, alcohols, aldehydes, and ke-
mass peak that is unique for each component; olefins, on the other hand, exhibit character-
tones; C1-C4 chlorides and iodides; fluoro-
The mass spectrometer is a powerful tool for a set of simultaneous equations must, there- istically intense peaks at masses 41, 5S, 69, 83,
carbons; thiophenes; atmospheric polutants;
the quantitative analysis of mixtures contain- fore, 'be solved to resolve the data from and 97. Another group of peaks is attributable
mixtures. That is, exhaust gases; and many others. By employing to cycloolefins, diolefins, and acetylenes (67,
higher temperatures, successful analytical
68,81,82, 9S, and 96). Finally, alkylbenzenes
II See, for eumple: R. M. Silvcntein, G. C. Bassler, and methods have been reported for C16-eZ7
T. C. Morrill, SpeetrMnelric ltIeltt/jiciltion of Organ", illPl + i12pz + + ilop. = 11
alcohols, aromatic acids and esters, steroids,
are found to fragment to masses of 77, 78, 79,
Compounb, 3d ed. New York: Wiley, 1974, p. 16- 91, 92, lOS, 106, 119,120, 133,and 134.A matIie-
;ZlPl + ;22PZ + + iz.p. = Iz fluorinated polyphenyls, aliphatic amides, matical combination of the peak heights of a
12 R. M. Silverstein, G. C. Bassler, and T. C. Morrill,
Speet_ric ltlelttificotlollof OrganU: COIlljlOlUlils,
3d ed. set provides an analytical parameter for ~s-
New York: Wiley, 1974, Chapter 2; K. BiemaDn, M_ ing the concentration of each type of hydro-
Spe_y, 0rganU:C"-kal App/iQltlolls.New York: _ carbon. Type analyses have been used to
McGraw-Hili, 1962; H. Budzikiewicz, C. Djeruai, and D.
II See, for example: F. W. Me1po1der and R. A. Brown, characterize the properties and behavior of
H. Williams, lnUrpretilllon of M_ SpeetrG of 0rganU:
in TMiltIu 011 AllIJlytkGl C"-Istry, edI. I. M. KoltboJl' gasolines, fuel oils, lubricating oils, asphalts,
COIIl/N1fIIld&. San FraDCiIc:o: HoIdeD-Day, 1964; and F. and P. J. E1viq. New York: 1Dtenc:ieaee, 1963, Part I,
W. McLall'erty, InUrpretiIIIon of M_ Spectra, 2d eel. where I. is the measured ion current at mass and mixtures of paraffins, olefins, alcohols,
Menlo Park, CA: Benjamin, 1973. m in the spectrum of the mixture, and i.. voL 4, p. 2047; and B. J. MiDard, QutmtlttJtl~ M_ Spec-
I_ry. PbiIIdelpbia: Heyden, 1977. and ketones.
lnorpnlc T•.•• AMI,... resulted in general use of the Mattauch- 2. Calculate the (M + 2)IM and (M + 4)/M peak ratios for the follow-
Herzog geometry for the desip of the analyz- ing compounds.
The'development of the spark source (p. 480) (a) Dibromotoluenc, C7H6Br2
er (Figure 17-5~
has made possible the application of mass SesBitiYity,Acancy, ••• Precisioa. Under (b) Methylene chloride, CH2C12
spectrometry to the analysis of inorganic (c) .1-Bromo-2-ehloroethanc, C214C1Br
ideal conditions, spark source mass spectrom-
solids; in the Jait two clccadcs, this technique
etry is sensitive to concentrations in the range 3. What dcgrcc of resolution would be needed to distinguish between
has become important for the trace analysis
ofa few parts per billion. In less favorable cases, formaldehyde (CH20) and ethane (C2H6)? What type of mass spec-
of elements in metals, alloys, superconductors,
on the other hand, the sensitivity limit may be trometer would be required? How else might this distinction be
and minerals.
one or two orders of magnitude greater than made?
Spark source inass spectrometry owes its
this figure. 4. Given below are the major peaks from the mass spectra of two
growth to three inherent advantages. F:irst is
Quantitative measurements in spark source isomers of C2H60. Identify the two compounds.
its high sensitivity, which often permits detec-
mass spectrometry arc based upon measure-'
tion and semiquantitative determination of
ments of the blackening of the photographic
elements in the parts per billion range.
plate with a microdensitometer; where suitable
Second is the simplicity of its spectrum lor an standards arc available, uncertainties of 10 to (a) m/e % Intensity (b) m/e % Intensity
element, which consists of but a single p1&jor
30 % relative arc to be cxpected. Greater errors 46 16 46 61
line and a few weaker Iincs that occur at frac-
arc often encountered with Jess favorable 45 SO 45 100
tional values of the mass of the clcmcnt (the
standard samples. 31 100 31 5
Iattcr arc due to multiply charged ions ~ The
third advantaae is that line intensities for var- Isotope Abund8nce M__ .....-m
ious etcmcnts arc roughly equal (usually ~thin
S. Compounds A and B both have a molecular ion peak at 58. After
a factor of 2 to 3~ : The mass spectrometer wu developed initially
treatment in NaOO with 020, compound A has a molecular ion
The major disadvantqe of spark $ource for the study of isotOPic abundance, and the
peak at 64, and compound B at 60. What type of compounds would
mass spectrometry can be traced to thc,crrat- instrument continues to be the most important
show this behavior when treated with 020? What arc A and B1
ic nature of the source and the consequent , source for this kind of data. Information re-
IlICk of reproducibility of the measurements. garding the abundance of various isotopes is
Because of the wide fluctuation in ion cur· now employed for a variety of purposes; the
rents with time, measurements with an inte- determination of formulas of organic com-
sratinI detector over a period of time arc pounds cited earlier is one example. Other
required; pbotopapbic detection is usually important applications include analysis by is0-
CI\lploycd for this reason. tope dilution, tracer studies with isotopes, and
A second disadvantage of the spark source dating of rocks and minerals by isotopic ratio
is that it produces ions varying widely in ki- measurements. The tcchniques arc similar to
netic energy. Thus, expensive double-focusing those described in Chapter 16 for radioactive
spectrometers arc required. isotopes. Mass spectrometry, however, permits
Lwta ••••••• SCVeral commercial spark the cxtension of isotope detection to nonradio-
source mass spectrometers are available. The active species such as 13<;.110, 180, UN, 3"8,
requirement of photographic detectors has and others.
I

1. Calculate the (M + 2)IM peak ratios for the following molecules.


(a> Ethyl bromide, C2H,Br
(b) Chlorobcnzcnc, C6H,C1
(c) Dimethylsulfoxide, C2H6SOZ

Electroanalytical chemistry encompasses a Cell Components
group of quantitative analytical methods that
An electrochemical cell consists of two metallic
arc based upon the electrical properties of a
conductors called electrodes, each immersed in
solution of the ana1ytc when it is made part of
a suitable electrolyte solution. For electricity to
an electrochemical cell. Three tYP!=lof electro-
flow, it is necessary: (1) that the electrodes be
analytical methods arc encountered. One in-
connected externally by means of a metal con-
cludes methods that arc dcpcndcnt upon the
du~or and (2) that the two electrolyte solutions
direct relationship between concentration and
be m contact to pennit movement of ions from
an electrical parameter such as potential, cur-
one to the other. Figure 18-1 shows an example
rent, resistance (or conductance~ capacitance, of a galvanic celL The fritted glass disk is
or quantity of electricity. A second group em-
porous, so that 1nH, CuH, and 80:- ions as
ploys one of the foregoing electrical parameters
well as H20 molecules can move across the
to establish the end point in a titration. A third
junct~on ~tween the two electrolyte solutions;
category includes methods in Whichthe analyte the disk sunply prevents extensive mixing of
is converted to a weighable fonn by means of
the two solutions.
an e1cctrical current. Coaduetion ia an EIectroebemical CeIL
. Regardless of type, the in~lIigent applica- Electricity is conducted by three distinct
tIOn of an electroanalytical method requires processes in various parts of the galvanic ceil
an understanding of the basic theory and shown in Figure 18-1. In the copper and zinc
practical aspects of the operation of electro- electrodes, as well as in the external conduc-
chemical cells. This chapter is devoted largely tor, electrons serve as carriers, moving from
to these matters. I :
the zinc through the conductor to the copper.
Within the two solutions the flow of eleetric-
ity involves migration of both cations and
ELECfROCHEMICAL CELLS anions, the fonner away from the zinc elec-
trode toward the copper and the latter in the
Electrochemical cells can be conveniently clas- reverse direction. All ions in the two solutions
sified as galvanic if they are employed to pro- participate in this process.
duce e1cctrica1 energy and electrolytic when A third type of conduction occurs at the
they consume electricity from an external two electrode surfaces. Here, an oxidation or
source. Both find use in electroanalytical chem- a reduction process provides a mechanism
istry. It is important to appreciate that many whereby the ionic conduction of the solution
cells can be operated in either a galvanic or an is coupled with the electron conduction of the
electrolytic mode by variation of experimental eleetrode to provide a complete circuit for a
conditions. current. The two eleetrode processes are de-
scribed by the equations
Zn(s)~ZnH + 2e
I

I Some reference works on elcctfochcmistry and its CuH + 2e~Cu(s)


appliCations include: J.O'M. Brockis and A. K. N.
Reddy, Motiern ElectrochemWry, 12 vols. New York:
The net cell reaction is the sum of these two
Plenum Press, 1970; G. Kortum, 'TreatiM on Electro- half-cell reactions:
cheMUtry,2d ed. New York: EIscvicr,I965; J. J. Linpnc,
E~yt/cIIl ChemUtry, 2d ed. New York: IDter-
Zn(s) + CuH ~ZnH + Cu(s)
ICicDce,19S8; and D. T. Sawyer and J. L Roberts, Jr. Because this reaction has a strong tendency
ExperllflenUJl E/ectroclaemUtry for ChemUtJ. New York:
to proceed to the right, the cell is galvanic
W'alcy, 1974.
and produces a potential of about 1 V under Solution of the sparingly soluble precipitate
, most conditions. occurs in the first step to provide the silver
Anode •••••Cathode. By definition, the cath- ions that are reduced in the second.
ode of an electrochemical cell is the electrode The last half-reaction has been included to
at which reduction occurs, while the anode is demonstrate that a cathodic reaction can in-
the electrode where an oxidation takes place. volve anions as well as cations.
These definitions apply to both galvanic and Reactions at AneMIa Examples of typical
electrolytic cells. anodic half-reactions include
For the galvanic cell shown in Figure 18-1, Cathode
the copper eleCtrode is the cathode and the CU(S)~CUH + 2e eu2+ + 2. 10 Cu
zinc electrode is the anode. Note that this cell FeH~Fe3+ +e
could be caused to behave as an electrolytic
cell by imposing a sufficiently large potential 2CI- ~Cl2(g) + 2e
from an external source. Under these circum-
H2(g)~2W + 2e
stances, the reactions occurring at the elec-
trodes would be 2H~0~02(g) + 4H+ + 4e
ZnH + 2e~Zn(s) The first half-reaction requires a copper elec-
Cu(S)~CUH + 2e trode to supply Cu2+ ions to the solution.
The remaining four half-reactions can take
Now, the roles of the electrodes are reversed; place at any of a variety of inert metal sur- trolyte. An example of a cell without a liquid The overall cell reaction is then
the copper electrode has beCome the anode faces. To cause the fourth half-reaction to junction is shown in Figure 18-2. Here, the
and the zinc electrode the cathode. AgCl(s) + !H2(g)
occur, it is nFsarY to replenish the hydrogen reaction at the silver cathode can be written
React_ at Cadtodes. Some typical cath- in the solution by bubbling the gas across the ~Ag(s) + W(aq) + CI-(aq)
odic half-reactions are AgC1(s) + e~Ag(s) + Cqaq)
surface ofthe electrode '(usually platinum~ The
The direct reaction between hydrogen and
CuH + 2e~Cu(s) reactions can then be formulated as Hydrogen is evolved at the platinum
solid silver chloride is slow. As a consequence,
anode:
Fe3+ + t~Fe2+ a common eleetrolytecan be employed without
H2(g)~ H2(sat'd)
significant loss of cell efficiency.
2W +2e~H2(g)
H2(sat'd)~2H+(aq) + 2e
AgCl(s) + t~Ag(s) + Cl-
IO; + 2H+ + 2e#10; + H20 The final reaction, giving oxygen as a prod-
uct, is a common anodic process in aqueous
Electrons are supplied for each of these pro- solutions containing no easily oxidized species.
cesses from the external circuit via an electrode Liquid JlIIICtioI& Cells with a liquid junc-
that docs not participate directly in the chemi- tion, such as that shown at the fritted disk in
cal reactiOn: In the first process, copper is de- Figure 18-1, are ordinarily employed to avoid
posited on the electrode surface; in the second, direct reaction between the components of
only a change in oxidation state of a solution the two half-cells. If the two electrolyte solu- 0.01 FHCI
component occurs. The third reaction is fre- saturated
tions in Figure 18-1 were allowed to mix, a with AgCl
quently observed in aqueous solutions that reduction in, the cell efficiencY would occur as
contain no easily reduced species. a result of t;te direct deposition of copper on Cathode
Anode AgClls)" Ag' laq) + CI- (aql
The fourth half-reaction is of interest be- the zinc. As will be shown later"a small poten- H2(gl •• H2 (aql
Ag' (aql + ••• Ag Is)
cause it can be CQPSidcredthe result of a two- tial called ajunction potential arises at the inter- !H2 (aql •• W (aq) + •
step process; that is, face between two electrolyte solutions that
AgCl(s)~Ag+ + CI- differ in composition.
Occasionally, useful cells can be constructed
Ag+ + e~Ag(s) in which the electrodes share a common elec-
'I1te Salt Bridp. For reasons to be dis- of partial pressure data, 1.00 atm is implied);
cussed later, e1ectrochemical cells are often the indicated molar silver ion concentration
equipped with a salt bridge to separate the was computed from the solubility product
electrolytes in the anode and cathode compart- constant for silver chloride.
ments. This device takes a variety of forms. The presence of a salt bridge in a cell is
In FJgUre 18-7 (p. S14~ for example, the bridge indicated by two vertical lines, implying that
consists of a U-shaped tube filled with a satu- a potential difference is associated with each
rated solution of potassium chloride. Such a of the two interfaces. Thus, the cell shown in
cell has two liquid junctions; one is between Figure 18-7 (p. 514) would be represented as
the cathode electrolyte and one end of the
MIM2+(yM)IW(xM)IH2(p atm~ Pt
bridge while the second is between the anode
electrolyte and the other end of the brid.ge.
Scbematie Representation of Cells. To sim- DC Currents In an
plify the description of cells, chemists often Electroehemicel Cell
employ a shorthand notation. For example,
the cells shown in Figures 18-1 and 18-2 can As noted earlier, electricity is transported with-
.be described by in a cell by the mip'ation of ions. In common
with metallic conductors, Ohm's law is often
C1lhode Anode
I
Zn ZnS04(xM) ICuS04(yM) ICu obeyed (departures from Ohm's law are dis-
6W + 6 •.•• 3H2 (g) 6C1- .•• 3CI2 (g) + 6.
cussed in the section devoted to polarization
Pt, H2(p = 1 atm)l H+(O.OIM~
effects~ That is,
I
CI- (O.oIM~ AgCI(sat'd) Ag FIGURE'8·3 Changes resulting from a current made up of ;
E six electrons. i
By convention, the anode and information with 1=-
R
respect to the solution with which it is in con-
tact is listed 011 the left. Single vertical lines where I is the current in amperes, E is the
represent phase boundaries at which poten- potential difference in volts responsible for
tials may develop. Thus, in the first example, movement of the ions, and R is the resistance by a battery, resulting in formation of three suiting from electrolysis could be offset by mi-
a part of the cell potential is associated with in ohms of the electrolyte to the current The molecules ofhydrogen; three molecules of chlo- gration of the added species as well as by the
the phase boundary between the zinc elec- resistance depends upon the kinds and con- rine are also produced at the anode (see Figure hydrogen and chloride ions. Since the added
trode and the zinc sulfate solution. A small centrations of ions in the solution. 18-3b~ TheresultingchalJe imbalance brought salt would tepresentan enormous excess, es-
potential also develops at liquid junctions; It is found experimentally that under a fixed about by the removal of ions from the electrode· sentially all of the electricity would be carried
thus, another vertical line is inserted between potential, the rate at which various ions move compartments is offset by migration, with posi- within the cell by the potassium and nitrate
the zinc and copper sulfate solutions. The in a solution differs considerably. For example, tive ions moving toward the negative electrode ions rather than by the reactant ions; only
cathode is then represented symbolically with the rate of movement (or mobility) of the proton and conversely. Because the proton is about across the elCctrode surfaces would the current
another vertical line separating the electrolyte is about seven times that for the sodium ion five times more mobile than the chloride ion, result from the presence of hydrogen and chlo-
solution from the copper electrode. and five times that of the chloride ion. Thus, however, a significant difference in concentra- ride ions.
In the second cell, only two phase bound- although all of the ions in a solution partici- tion in the outer electrode compartments de-
aries exist, the electrolyte being common to pate in conducting electricity, the fraction car- velops during electrolysis. In effect, five-sixths
of the current has resulted from movement of Alternating
both electrodes. An equally correct represen- ried by one ion may differ markedly from that
the hydrogen ions and one-sixth from the trans- Current Conduction
tation of this cell would be carried by another. This fraction depends upon
port of chloride ions. When a de potential is applied[ to a cell, con-
the relative concentration of the ion as well as
I
Pt H2(sat'd~ HCl(O.OIF~ its inherent mobility. To illustrate, consider the It is important to appreciate that the current duction requires an oxidation reaction at the
cell shown in Figure 18-38, which is divided need not result from transport of the electrode anode and a reduction at the cathode. 1bc
Ag+(1.8 x 10-SM)IAg term faradaic is sometimes used to denote
into three imaginary compartments, each con- reactants exclusively. Thus, if we were to intro-
duce, say, 100 potassium and nitrate ions into such currents and processes. Both faradaic
Here, the molecular hydrogen concentration taining six hydrogen ions and six chloride ions.
each of the three compartments of the cell and nonfaradaic conduction occurs with an
is that of a saturated solution (in the absence Six electrons are then forced into the cathode
under consideration, the charge imbalance re- ae potential
current results from the periodic alteration in terns in this country were powered by one form
these processes as a consequence of the alter- of this cell. The potential of the Daniell cell is
-i +
- -,1 -+ - 11
- j- 1 nating voltage.
+ . I dependent upon the concentration of the ca-

\1,
The dielectric current depends upon the di-
,
- +1 - + I -
- , + - I +
t elcctric constant of the mediUJil and is directly
tions in the two electrolyte solutions; when
these are about equal, a voltage of 1.1 V de-
- + 1 + proportional to frequency; only at radio fre-
- -, - - +1
1 ~ , velops. As current is drawn, however, the zinc
+ 'C \J quencies (- 10' Hz) does this current become ion concentration increases while the copper
- tl- + I
-
- I
-
-,1 - i f important
ion concentration decreases by an equivalent
- 1- + + I

do
+1- +-
. I
: d,
-,
+I
d2
0
do d,
1
d2
Reversible and
Il'I'8V8nibie Cell.
amount. The cell potential undergoes a corre-
sponding decrease and eventually reaches zero;
d d at this point, equilibrium has been reached
The galvanic cell shown in Figure 18-2 would for the cell reaction
(a) (bl
develop a potential of about 0.46 V. If a bat-
FIGURE 18-4 Electric double layer formed at electrode sur- tery with a potential somewhat greater than
Zn(s) + CuH ~ZnH + Cu(s)
face as a ~ult of an applied potential 0.46 V were inserted in the circuit, with its The relationship between the cell potential
negative terminal connected to the platinum and the concentrations of the participants in
electrode, a reversal in direction of electron a cell reaction is readily derived from thermo-
flow would occur; the reactions at the two dynamic considerations. A knowledge of this
electrodes would thus become relationship, and its use, is vital to the under-
NonflU'llllaie C.-rents. ~onfaradaic cur- this ionic movement. Thus, each electrode sur-
rents involve the formation of an electrical face behaves as a capacitor, the capacitance Ag(s) + a- ~AgCI(s) + e standing of all types of electroanalytical meth-
ods.
double layer at the electrdde-solution inter- of which may be remarkably large (several
face. When a potential is applied to a ptetallic hundred to several thousand microfarads per
W + e~tH2(g) As in all thermodynamic calculations,a close
electrode immersed in an electrolyte, a momen- cm2). The capacitance current increases with Now the silver electrode is the anode and the agreement between calculated potential and
tary surge of current creates an excess (or a frequency and with electrode size; by control- platinum electrode the cathode. A cell (or an experimental findings requires the use of activ-
deficiency) of negative charge at the surface of ling these variables, it is possible to arrange electrode) for which a change in direction of ities rather than molar concentrations. Thus.
the metal. As a consequence of ionic mobility, conditions so that essentially all of the alternat- the current causes a reversal of the electro- it is worthwhile to review the relationship be-
however, the layer of solution immediately ing electricity flowing through a cell is carried chemical reaction is said to be reversible. tween these two concentration parameters be-
adjacent to the electrode acquires an opposing across the electrode interface by this non- Cells in which a current reversal results in fore considering methods for the computation
charge. This effect is shown in Figure 18-4a. faradaic process. different reactions at one or both electrodes of cell potentials.
The charged layer consists of two parts: (1) a Alternating Current in an Electrolyte Solu- are called irreversible. The cell shown in
compact inner layer, in which the potential tion. For frequencies up to a few thousand Figure 18-1 is also reversible. If, however, a Activity and
decreases linearly with distance from the elec- cycles per second, alternating current results small amount of dilute acid were introduced Activity Coefficient
trode surface; and (2) a more diffuse layer, in
which the decrease is exponential; see Figure
almost exclusively from ion movement. The
direction of this motion, of course, reverses
into the zinc electrode compartment, the reac-
tion would tend to become irreversible. Here, The relationship between the activity of a a..
18-4b. This assemblage of charge inhomoge- with each half-cycle. At very high frequencies, zine would not deposit at the cathode upon species and its molar concentration [M] is
neities is termed the electrical double layer. a significant fraction of the electricity is carried application of a potential; instead, hydrogen given by the expression
The double layer formed by a dc potential by a second mechanism that results from elec- would form by the reaction
involves the development of a momentary trical polarization of the dielectric medium.
current which then drops ~o zero (that is, the Here, the voltage gradient causes induced polar-
2H+ + 2e~H2(g)
where 1M is a dimensionless quantity ca1Ied
electrode becomes polarized) unless some far- ization and orientation polarization of the mole- Thus, the zinc electrode and the cell would be the activity coefficient. The activity coefficient,
adaic process occurs. With an alternating cur" cules in the medium. In the former, distortion termed irreversible in the presence of acid. and thus the activity, of M varies with the Ionic
rent, however, revCl'Slt1of the charge relation- of the electron cloud surrounding the nucleus strength of a solution such that the employment
ship occurs with Cach half-cycle as first negative of a molecule causes a temporary polarized of aN instead of (M] in an electrode potential
and then positive ions are attracted to the condition; in the latter, molecules with a per- calculation (and in the other equilibrium calcu-
electrode surface. Electrical energy is con- manent diode moment become aligned with the The galvanic cell shown in Figure 18-1 is called lations as well) renders the numerical value
sumed and converted to frictional heat from electrical field. Regardless of the mechanism, a Daniell cell or battery. Early telegraph sys- obtained independent of the ionic Itreqth.
Here, the ionic strength Il is defined by the The constants 0.5085 and 0.3281 are appli- strength is less than 0.01; under these circum-
cable to solutions at 25OC; other values must stances, uncertainties in lXAare of little signifi-
equation ~ 0.8 be employed at different temperatures. cance in calculating activity coefficients.
Il- i(mlZ~ + m2Zi + m3Z~+ ...) j
(18-3) j 0.6
Unfortunately, considerable uncertainty ex-
ists regarding the magnitude of itAin Equation
Kienand3 has calculated values of lXAfor
numerous ions from a variety of experimental
where mb m2' m3' ... represent the molar ~ 0.4 18-4. Its value appears to be approximately data. His ••best values ••for effective diameten
concentration of the various ions in the solu- ! 3 A for most single charged ions so that, for are given in Table 18-1. Also presented are
:l 0.2 A13•
tion and ZI, Z2' Z3" .. are their respective these species, the denominator of the Debye- activity coefficients calculated from Equation
Fe(cNII- Huckel equation reduces to approximately
charges. Note that an ionic strength calcula- 18-4 using these values for the size parameter.
tion requires taking account of all ion.ic species 0.3 0.4 (1 + ..;p). For ions with higher charge, lXAmay
in a solution, not just the reactive Ofles. "'Ionic strength be as large as 10 A. It should be noted that
Properties of Activity CoeOidents. Activity FIGURE 18-5 Effect of ionic strength on the second term of the denominator becomes
coefficients have the fonowing properties: activity coefficients. small with respect to the first when the ionic
1. The activity coefficient of a species can
be thought of as a measure of the eff~veness
with which that species influences ail equilib-
rium in which it is a participant. In very
dilute solutions, where the ionic strength is
TABLE 18·' ACTIVITY COEFFICIENTS FOR IONS AT ZS'C"
minimal, this effectiveness becomes constant,
arid the activity coefficient acquires ,. value of 4. Activity coefficients for ions of the same lXA Activity Coefficient at Indicated
unity; the activity and molar con~ntration charge are approximately the same at any Effective Ionic StrengtlB
thus become numerically identical. A~the ionic given ionic strength. The small variations that ~iameter,
do exist can be correlated with the effective Ion 0.001 0.005 0.01 0.05 0.1
strength increases, an ion loses some of its.
effectiveness, and its activity coefficient de- diameter of the hydrated ions. H3O+ 9 0.967 0.933 0.914 0.86 0.83
creases. We may summarize this behavior in 5. The product of the activity coefficient and Li+, C6H,COO- 6 0.965 0.929 0.907 0.84 0.80
terms of Equation 18-2. At moderate ionic molar concentration of a given ion describes 4-4.5
Na+,IOj, HSOj, HCOj, 0.964 0.928 0.902 0.82 0.78
strengths.! •• < 1; as the solution approaches its effective ~vior in aU equilibria in which
H2PO;, H2AsO;, OAc-
infinite dilution,! •• -+ 1 and thus a•• -+ [M). it participates. 00-, F-, SC~-, HS-, CIOj, 3.5 0.964 0.926 0.900 0.81 0.76
At high ionic strengths, the activity coeffi- EvaIatioD fIl Aceirity CoefIicieBts. In 1923,
00;, BrOj, 10;, MnO;
P. Debye and E. HUcke! derived the following
. cients for some species increase and may even K+, 0-, Br-, 1-, CN-, NOl, 3 0.964 0.925 0.899 0.80 0.76
become greater than one. The behavior of such theoretical expression, which permits the calcu-
NOj, HCOO-
solutions is difficult to interpret; we shall con- lation of activity coefficients of ions2 Rb+, Cs+, Tl+, Ag+, NHt 2.5 0.964 0.924 0.898 0.80 G,1~
fine our discussion to regions of low to moder- MgH, BeH 0.52 0.45
8 0.872 0.755 0.69
ate ionic strengths (that is, where Il < 0.1). -log/A = O.sos5Zi..;P (18-4) ea2 +, Cu2 +, Zn1 +, Sn2 +, Mn1 +, 0.40
6 0.870 0.749 0.675 0.48
The variation of typical activity coefficients 1 + 0.3281lXAh Fe2+, NiH, CoH,
as a function of ionic strength is shown in Phthalate2 -
Figure 18-5. SrH, BaH, CdH, HgH, Sl- 0.744 0.67 0.46 0.38
5 0.868
2. In dilute solutions, the activity' coefficient
IA = activity coefficient of the species A PbH, CO~-, SO~-, C20~- 4.5 0.868 0.742 0.665 0.46 0.37
for a given species is independent of the specific
ZA = charge on the species A Hgi+, SO~-, S20~-, crO~-, 4.0 0.867 0.740 0.660 0.44 0.36
nature of the electrolyte, and dePends only
Il = ionic strength of the solution HPO~-
upon the ionic strength. '
·ltA"" the effective diameter of the A13+, Fe3+, Cr3+, La3+, Ce3+ 9 0.738 0.54 0.44 024 0.18
3. For a given ionic strength, the activity
hydrated ion in Angstr6m units PO:-, Fe(CN):- 4 0.725 0.50 0.40 0.16 0.095
coefficient of an ion departs further from unity
as the charge carried by the species increases. Th4+,~+,Ce4+,Sn4+ 11 0.588 0.35 0255 0.10 0.065
This efI'ectis shown in Figure 18-5. The activity Fe(CN>:- 5 0.57 0.31 0.20 0.048 0.021
coefficient of an uncharged molecule is ap-
proximately one, regardless of ionic strength. • From J. KieUand. J. "'me,. C"""'. Soc,59, 1675 (1937).
Experimental verification of individual ac- page 524 illustrate the magnitude of uncertain- constant in their solid phases. Thus, these quan- Here, the subscript a is employed to indicate
tivity coefficients such as those shown in Table ties that can arise when concentrations rather tities are included in the constant K.5 that the bracketed terms are instantaneous
18-1 is, unfortunately, impossible; all experi- than activities are employed in the calculation It is convenient to define a second quantity concentrations and not equtlibrium concentra-
mental methods give only a mean activity coef- of cell potentials. Q such that tions. The quantity Q, therefore, is not a con-
ficient for the positively and negatively cbarged Molar concentrations will usually be used stant, but changes continuously until equilib-
ions in a solution.· It should be pointed out, in lieu of activities in the discussion that fol- rium is reached; at that point, Q becomes equal
however, that mean activity coefficients calcu- lows, for the sake of convenience. Occasionally, to K and the a subscripts are deleted.
lated from the data in Table 18-1 agree satis- however, it will be desirable or necessary to From thermodynamics, it can be shown
factorily with the experimental values. revert to the more exact concentration param- that the change in free energy AG for a cell
The Debye-Hiickel relationship, and the eter. 5 An alternative way of treatinl the eITectof pure solids reaction (that is, the maximum work obtain-
data in Table 18-1, give satisfactory activity co- IS weU IS pure liquids aui!the solvent in an equilibrium able at constant temperature and pressure) is
efficients for ionic strengths up to about 0.1. Be- system is the foUowinl: we define the standard state of a given by
pure substance IS its physieal stale at 2S"C and at a
yond this value, however, the equation fails, Effect of pressure of one atmosphere. The activity of a substance AG = RT In Q - RT In K (18-7)
and experimentally determined mean activity Concentretion in its standard stale is then assigned a value of unity.
coefficients must be employed. Unfortunately, on Cell Potential. Thus, the equilibrium constant for the reKtion in
where R is the gas constant (8.316 J mol-1
many electrochemical calculations involve s0- question can be written as deg - 1) and T is the temperature in degrees
lutions of high ionic strength for which no The effect of concentration (or activity) on K; the term In refers to the logarithm to the
K _ (H+]I[O-]'{AI] _ [W]I[O-]'(UIO)
experimental activity coefficients are available. cell potentials can be illustrated with the cell base e. It can also be shown that the cell
1'H,[AaCI] I'H,(I.IIO) potential E.ell is related to the free energy of
Concentrations must thus be employed instead shown in Figure 18-2, which can be form-
of activities; uncertainties that vary from a few ulated in more general concentration terms as For aU pses, the standard state is a pressure of one the reaction by the relationship
atmosphere; for a solute,;it is an Idivity of 1.00.
percent relative to an order of magnitude may For the reKtion AG = -nFE.en (18-8)
be expected. The data shown in the example on Pt, H2(x atm)lW(yM1 CI-(yM)IAg
Br,(I) + H.01HOBr + Br- + H+ where F is the faraday (96,491 coulombs per
Here, we may describe the electrode processes chemical equivalent) and n is the number of
by the two half-reactions, namely, equivalents of electricity (or moles of elec-
K _ [HOBrIBr-IW) trons) associated with the oxidation-reduction
2AgCl(s) + 2e¢2Ag(s) + 2CI- [Br,][H,O] process (in this example, n = 2).
• The mean .aivity of the electrolyte A.B. is dcfioed IS 2W +2e¢H2(g) Both the liquid bromine and the solvent water are in Substitution of Equations 18-6 and 18-8
foIlowa: their standard states. Thus, the equilibrium constant ex- into 18-7 yields
presaioDreduces to
f* - mean Idivity codlic:ient- (n· fa)'iIoo+- The reason for writing both reactions as re-
ductions will become apparent later. To derive K - (HOBrIBr-][W]
[H+:C[O-:e
The meaD .aivity coellic:ientcan he measured in any of -nFE.e •• = RT In <PH,). - RT In K
several ways, but it is impouible experimentally to re- a cell reaction, it is necessary to subtract the It is important to note that this last expression applies
solve this term into the individualldivity codlic:ients for anodic reaction from the cathodic, thus ob- only irthe solution is saturated with Br,; ifit were under-
(18-9)
fA and f•. For example, if A.B. is a precipitate, we can taining saturated, we would write Let us define the standard potential E.'e •• for
write
Br,(aq) + H,O~HOBr + Br- + H+ the cell as follows:
K •• - [AnBr . n.f; - [AnBr . 1":+-
K [HOBrIBr-IW] RT
By measuring the solubility of A.B. in a solution in
E:. •• = nF In K (18-10)
The equilibrium constant for this reaction t -, [Brzl
which the electrolyte eonc:entration approaches zero (that ,
is, where fA and f•..•1~ we could obtain K •. A second
is given by where the denominator is the molar concentration of Br2 Substitution of this equation into Equation
solulnlity _urement at some ionic:strenath, Il, would in the unsaturated solution. Note that 18-9 yields, upon rearrangement,
[H+)2[Cn Z
live values for [A] and (8]. These data woulclthen permit
the caIcu1alionof r:.. .
fa - I": +. for ionic:Itrenath Il,. It
K=~~---
PH,
K 1K ,[Br ,] •••.• RT [H+:e[O-:e
is important to undentand that there arc insulIicient where [Br,]... .• is the molar conc:cntratiOllor bfomine in E.en = E:.n - nF In <PH,). (I8-U)
experimental data to permit the caJculatioa of the lMllIiII- where the bracketed terms represent molar a saturated solution.
."" quantitiesfA _fa _ that there appeus to he no It should also be noted that the terms in an equilib-
additional experimeDtal inforIIIatioa that would' permit
concentrations of hydrogen and chloride ions, Note that the standard potential is a constant,
rium constant expression arc in fact ratios of the con.
evaluation or these quantiUea. This aituation is aeaeraJ; respectively, and PH, is the partial pressure of centration or the species relative to their concentrations which is equal to the cell potential when the
the experilM/lttll determination of individual adivity hydrogen in atmospheres. Note that the con- or pressures in the standard stale (1.00 M or 1.00 atm~ reactants and products are at tmlt concentration
codlic:ients appean to be impossible. centrations of silver and silver chloride are Thus, 1M equilibrium cons/ant i.s effeetil1ely unitless. (more co"ectly activity) and pressure.
is not realized; instead, what is measured is a question and the standard hydrogen electrode.
HALF-CELL OR
ELECTRODE POTENTIALS combination of the potential of interest and the It must always be borne in mind that the elec-
half-cell potential for the contact between the trode potentials are in fact relative potentials,
In electroanalytical work, it is often conven- voltage-measuring device and the solution. all being referred to the common reference
ient to think of a cell potential as being com- Our inability to measure absolute poten- electrode.
posed or two IuJlf-cell or electrode potentials, tials for haIf-cell processes is not a serious Several secondary reference electrodes are
one being associated with the cathode and the handicap because relative half-cell potentials more convenient for routine use and are exten-
other with the anode. Thus. for the cell just are just as useful. These relative potentials sively employed; some of these are described
considered, we may write can be combined to give cell potentials; in later in this chapter.
addition, they are useful for calculating equilib-
Eeell s: EAaCt - Ea.
rium constants of oxidation-reduction pro-
where EAaCt is the electrode potential for the cesses.
silver/silver chloride electrode (which is the Measurement of
Pt electrode
cathode of the cell) and Ea. is the electrode coated with Electrode Potentials
potential for the hydrogen gas electrode (here, The Standard Ptblack

the anode~ Similarly, for the Daniell cell Hydrogen Electrode Although the standard hydrogen electrode is
shown in Figure 18-1, The standard hydrogen electrode (SHE) is the the universal standard of reference, it should
universal reference for reporting relative half- be understood that the electrode, as described,
Eeell s: Ecu - EZa
cell potentials. It is a type of gas electrode. can never be realized in the laboratory; that is,
A more general statement would be ne Hy4rogen ElectrocIe. The hydrogen it is a hypothetical electrode to which experi-
electrode was widely used in early electrochem- mentally determined potentials can be referred
Ecotl = Ecalhodc - E.aodc
ical studies as a reference electrode and as an only by suitable computation. The reason that
where Ecatllode and E•..- are electrode poten- indicator electrode for the determination of the electrode, as defined, cannot be prepared is
tials for the electrodes that are acting as the pH. Its composition can be formulated as that chemists lack the knowledge to produce
cathode and anode, respectively. A rigorous a solution with a hydrogen ion activity of
definition of the term electrode potential will Pt, Hz(P atm)IH+(xM) The hydrogen electrode may act as an exactly unity; no adequate theory exists to
be forthcoming (p. 515). anode or a cathode, depending upon the half- permit evaluation of the activity coefficient of
As suggested by the terms in parentheses, the cell with which it is coupled. Hydrogen is hydrogen ions in a solution in which the ionic
potential developed at the platinum surface oxidized to hydrogen ions at an anode; the strength is as great as one. Thus, the concen-
Nature of depends upon the hydrogen ion concentra- reverse reaction takes place at a cathode. tration of HCI or another acid required to give
Electrode Potentials tion of the solution and the partial pressure of Under proper conditions, then, the hydrogen a hydrogen ion activity of ul)ity cannot be
the hydrogen employed to saturate it. electrode is electrochemically reversible. calculated. Notwithstanding, data for more
At the outset. it should be emphasized that
Figure 18-6 illustrates the components of a The potential of a hydrogen electrode de- dilute solutions, where activity coefficients are
there is no way of determining an absolute
hydrogen electrode. The conductor is con- pends upon the temperature, the hydrogen ion known, can be used to provide potentials at
value for the potential of a single electrode,
structed from platinum foil, which has been concentration (more correctly, the activity) in unit activity. Thus, for example, the activity of
since all voltage-measuring devices determine
platinized-that is, coated with a finely divided hydrogen and chloride ions in the cell shown
only differeltCeS in potential One conductor the solution, and the pressure of the hydrogen
from such a device is connected to the elec-
layer or platinum (called platinum black) by at the surface of the electrode. Values for these in Figure 18-2 can be calculated from the
rapid chemical or electrochemical reduction of parameters must be carefully defined in order Debye-Hiickel relationship; measurements can
trode in question; in order to measure a poten-
HzPtCI6• The platinum black provides a large for the half-cell process to serve as a reference. also be made at lower acid concentrations.
tial difference, however, the second conductor
must be brought in contact with the electrolyte surface area to assure that the reaction Specifications for the standard hydrogen elec- These data can then provide, by suitable ex-
solution of the half-cell in question. This latter trode call for a hydrogen activity of unity and trapolation, information about the potential of
2H+ + 2e~Hz(g) a hypothetical cell in which the hydrogen and
contact inevitably involves a solid-solution in- a partial pressure for hydrogen of exactly one
terface and hence acts as a second half-cell at proceeds rapidly and reversibly at the elec- atmosphere. By convention, the potential q this chloride ions have uni! activities. For the cell
which a chemical reaction must also take pltJce trode surface. As was pointed out earlier, the electrode is assigned the value q exactly zero in Figure 18-2, Equation 18-11 then reduces to
if electricity is to Bow. A potential will be stream of hydrogen serves simply to keep the volt at all temperatures.
associated with this second reaction. Thus, an solution adjacent to the electrode saturated Electrode potentials are defined as cell p0- Fj RT In (l.OOf(l.OOf
absolute value for the desired half-cell potential with respect to the gas. tentials for a cell consisting of the electrode in Eeen = AaCI - nF 1.00
olation will then provide a standard electrode reaction, then, is the reverse of that in the two
potential E& for the half-reaction cells considered earlier:

M2+ + 2e= M(s) E& ••. zV Cu2+ + Hz(g)~Cu(s) + 2H+


Thus, metallic copper is a much less effective
Thus, E& is the potential for the cell when the
reducing agent than either zinc, cadmium, or
activities ofM2+ and H+ are exactly one. This
hydrogen. As before, the observed potential is
measurement would be less accurate than the
a quantitative measure of this strength.
one employing a cell without a liquid junction
By additional measurements and suitable
because of the uncertainty with respect to the
extrapolations, the data from the experiment
junction potentials that exist at the two ends
just described could be made to yield computed
of the salt bridge.
potentials for the system when the activities of
If the metal M in Figure 18-7 is cadmium
MH and H+ are unity. These potentials for
and the solution is approximately O.ol M in
CdH, ZnH, and CuH have values of 0.403,
cadmium ions, the voltage indicated by the
0.763, and 0.337, respectively. Note, however,
measuring device, Y, wi~ be about 0.5 V.
that a need exists to indicate that the copper
Moreover, the cadmium Will function as the
electrode behaves as a cathode while the zinc
anode; thus, electrons pass from this electrode
and cadmium electrodes function as anOdes
Solution of • SHE to the hydrogen electrode via the external cir-
attofM
when coupled to the hydrogen electrode. Posi-
cuit. The halC-cellreactions for this galvanic cell
tive and negative signs are used to make this
FIGURE 18-7 Schematic diagram of an arrangement for the measure- can be written as
distinction, the potential for halC-cellssuch as
ment of electrode potentials against the standard hydrogen electrode. copper being provided with one sign and the
Cd(s)~CdH -ti2e Anode
other two electrodes being assigned the oppo-
2H+ + 2e~Hz(g) , Cathode site. The choice as to which potential will be
positive and which will be negative is purely
where ~tC1 is called the stondmd electrode The overall cell reaction is the sum of these, arbitrary; however, the sign convention that is
potential for the half-reaction or chosen must be used consistently.
EAeCJ ••• E'AeCJ - R: In [0-]. (18-14)
+ 2W + Hz(g)
AgCl(s) + e~Ag(s) + 0- E' == 0122 V Cd(s) ~CdH
(18-13) This equation shows how the electrode poten- Sign ConventioM
If the cadmium electrode is replaced by a zinc
We can also imagine a cell with a liquid tial EAcCI for the silver/silver chloride elec- electrode immersed in a solution that is about for Electrode Potentials
junction or a salt bridge in which the hydrogen trode varies as a function of chloride ion con- O.ol M in zinc ions, a potential of about 0.8 V It is not surprising that the arbitrariness in
and chloride ions can be varied independently. centration. will be observed. The metal electrode is again specifying signs has led to much controversy
For example, A cell with a salt bridge such as that shown the anode in this cell. The larger voltage and confusion in the course of the develop-
in Figure 18-7can be used to measure electrode developed reflects the greater tendency of zinc ment of electrochemistry. In 1953, the Inter-
Pt, Hz(l.(IO atm)IH+("",,, 1.oo)lO-(zM~ potentials for half-reactions involving the met- to be oxidized. The difference between this national Union of Pure and Applied Chemistry
AgCl(sat'd)l Ag al of the electrode and its ion. Here. the refer- potential and the one for cadmium is a quan- (IUPAC1 meeting in Stockholm, attempted to
ence electrode is shown as a hydrogen electrode titative measure of the ('elative strengths of resolve this controversy. The sign convention
The left half of this cell is again not experi- in which the acid concentration is sufficiently
mentally realizable, but its potential could be these two metals as reducing agents. adopted at this meeting is sometimes called
low to permit calculation of the hydrogen ion If the half-cell in FiguI'F 18-7 consisted of a the IUPAC or Stockholm convention; there is
computed from data for a cell in which aH+ activity by means of the Debye-Hiickel equa-
was considerably smaller than one. Neglect- copper electrode in a O.ot M solution of cop- hope for its general adoption in years to come.
tion; alternatively, one of the secondary elec- per(II) ions, a potential of about 0.3 V would We shaU always use the IUPAC sign conven-
ing junction potentials, the potential for this trodes described at the end or this chapter.
cell can be obtained by suitable substitutions develop. However, in distinct contrast to the tion.
whose potential against the standard hydrogen previous two examples, copper would tend to Any sign convention must be based upon
into Equation 18-11 electrode is known, can be substituted. The deposit, and an external electron flow, if al- half-cell processes written in a single way-
. RT [o-J:(1.oof activity of the metal ion y is then varied and
E •• II == EAeCJ'" E'AeCJ - 2F In 1.00 lowed, would be from the hydrogen electrode that is, entirely as either oxidations or as re-
the cell potential measured. Suitable extrap- to the copper electrode. The spontaneous cell ductions. According to the IUPAC convention.
-
AN INTRODUCTION TO ELECTROANALYTICAL CHEMISTRY

the term electrode potential (or more exactly, proceeds toward the right under ordinary direction of this reaction in a given table can
shall not use the subscript; the student should
relative electrode potential) is reserved exclu- conditions. The negative electrode potential then serve as a key to any changes that may be always be alert to the fact, however that the
sively for half-reactions written as reductions. for zinc, on the other band, means that :the needed to convert all data to the IUPAC con- quotients that appear in this type of' equation
There is no objection to using the term oxida- analogous reaction vention. arc not equilibrium constants, despite their sim-
tion potential to connote an elcctrode process Zn2+ + H2{g)~2H+ + Zn(s) ilarity in appearance.
written in the opposite sense, but an oxidation Effect of
To summarize the meaning of the bracketed
potential should never be called an electrode does not ordinarily occur; indeed, the equilib- Concentration on
terms in Equation 18-15, when the substance
potentiaL The sign of an oxidation potential rium favors the species on the left. Electrode Potential R is a gas,
will always be opposite to its corresponding The IUPAC convention was adopted in
Equation 18-14 shows how the electrode poten- [R] == partial pressure in atmospheres
electrode or reduction potential. 1953, but electrode potential data given in
tial for the silver/silver chloride electrode varies
The sign of the electrode potential is deter- many texts and reference works arc not always When R:is a solute
as a function of chloride concentration. Turn-
mined by the actual sign of the electrode of in accord with it For example, in a source of
ing to a more general case consider the half- [R] ~ concentration in moles per liter
interest when it is coupled with a standard hy- oxidation potential data compiled by Latimer,6
reaction
drogen electrode in a galvanic celL Thus, a zinc one finds or occasionally
or a cadmium electrode will behave as the pP+qQ + ... +ne~rR+sS
anode from which electrons flow through the Zn(s)~Zn2+ + 2e E= +0.763 V r [R] = activity of R, aR
external circuit to the standard hydrogen elec- where the capital letters represent formulas of
Cu(s)~Cu2+ + 2e E- -0.337 V reacting species (whether charged or un- When R is a pure solid or liquid in excess or
trode. These metal electrodes are thus the nega· the solvent
tive terminals of such galvanic cells, and their To convert these oxidation potentials to elec- charged~ e represents the electron, and the
electrode potentials arc assigned negative val- trode potentials as defined by the IUPAC con- lower-aise italic letters indicate the number of [R] = 1.00
ues. That is, vention, one must mentally: (1) express the moles of each species (including electrons)
participating in the half-cell reaction. Employ- Equation 18-15 is called the Nernst equa-
half-reactions as reductions; and (2) change the tion in ~onor of a nineteenth-century electro-
Zn2+ + 2e~Zn(s) EJ= -0.763 V signs of the potentials. ing the same arguments that were used in the
case of the silver/silver chloride electrode, we chemist.,Application of the Nernst equation is
Cd2+ + 2e~Cd(s) EJ = -0.403 V The sign convention employed in a table of illustrated in the following examples.
obtain
standard potentialsmay not beexplicitlystated.
The potential for the copper electrode, on the This information is readily ascertained, how- RT [R:t{St··· 1. Zn2+ + 2e~Zn(s)
other hand, is given a positive sign because ever, by referring to a half·reaction with which E = EJ - nF In [P]''[Q:J: ...
the copper behaves as a cathode in a galvanic one is familiar and notinl the direction of the
E = EJ _ 0.0591 I _1 _
At room temperature (298°K~ the collection 2 og [Zn2+]
cell constructed from this electrode and the reaction and the sip of the potentiaL What- of constants in front of the logarithm has
hydrogen electrode; electrons flow toward the ever changes, if any, arc required to convert to 2. Fe3+ + e~Fe2+
units of joules per coulomb or volt (p. 7). That
copper electrode through the exterior circuit. the IUPAC convention arc then applied to the is,
It is thus the positive terminal of the galvanic remainder of the data in the table. For example, E = EO _ 0.0591 I [Fe2 +]
RT 8.316 J mol-1 deg-1 x 298 deg 1 og [Fe3+]
cell and all one needs to remember is that strong oxidiz-
ing agents such as oxygen have large positive nF = n equiv mol-1 x 96491 C equiv-1 This electrode potential can be measured with
electrode potentials under the IUPAC conven- an inert metal electrode immersed in a solu-
·2.568 X 10-2 J C-l
It is important to emphasize that electrode tion. That is, the reaction tion containing iron(II) and iron(III). The
n
potentials and their signs apply to half·reac- potential is dependent upon the ratio between
02{g) + 4W + 4e~2H20
tions written as reductions. Both zinc and cad- Thus, upon converting from a natural to a the molar concentrations of the two ions.
mium are oxidized by hydrogen ion; the spon- EJ = + 1.23 V base ten logarithm by multiplication by 2.303, 3. 2H+ + 2e~H2(g)
the foregoing equation can be written
taneous reactions arc thus oxidations. It is evi- tends to occur spontaneously with respect to ~
dent, then, that the sign of the electrode potential the standard hydrosen eJcctrode. The sign and ; E - EJ - 0.0591 1 2!!L
0.0591 [R]1S]"· ..
will indicate whether or not the reduction is sport- E= EJ -~ log [P],[Q1 ... (18-15) - 2 og [H+]2
taneous with respect to the standtJrd hydrogen
electrode. That is, the positive sign for the cop- For convenience. we have also deleted the a In this example, PH. represents the partial pres-
per electrode potential means that the reaction • W. M. Latimer, TIw Oxiatloll Stllta of tile E'-nts subscripts, which were inserted earlier as a sure of hydrogen, expressed in atmospheres, at
tmtJ TIW POUIItitIIs ill ~ SoIulIons. 2d ed. EnaJe- reminder that the bracketed terms represented the surface of the electrode. Ordinarily, PH. will
Cu2+ + H2(g)~2H+ + Cuts) wood CIifti, N-!.: Ptmlice-HaU. 1952. nonequilibrium concentrations. Hereafter, we be very close to atmospheric pressure.
AN INTRODUCTION TO ELECTROANALYTICAL CHEMISTRY
)\

although dependent upon the concentration bottom of tl)e table bave little tendency to triiodide ions, we can predict that the latter pair
of silver ions, is the same regardless of whether take place as written. On the other hand, they will predominate. .
0.0591 [Cr3+]Z we write the half-reaction as above or as do tend to occur in the opposite sense, as
E = £0 - -6-10g [CrzO~-][H+]14 oxidations. The most cft'ectivereducing agents, Calculation of Half-Call
100 Ag+ + l00e~loo Ag(s) then, arc those species that appear in the lower Potentials from EO Values
Here, the potential depends not only on the £0= +0.799 V right-hand side of the equations in the table.
concentrations of chromium(III) and dichro- A compilation of standard potentials pro- Typical applications of the Nernst equation
mate ions but also on the pH of the solution. To be sure, the Nernst equation must be con- vides the chemist with qualitative information !o the cal~ulation of half-cell potentials are
sistent with the half-reaction as it has been regarding the extent and direction of electron- Illustrated In the following examples.
The StlIndard written. For the first ofthesc, it will be transfer reactions between the tabulated spe-
EXAMPLE
Electr0d8 Potential, EO cies. On the basis of Table 18-2, for example,
0.0591 1 Wbat is the potential for a half-cell consist-
E = 0.799 - -1- log [Ag+] we see that zinc is more easily oxidized than ing of a cadmium electrode immersed in a
An examination of Equation 18-15 reveals
cadmium, and we conclude that a piece of zinc solution that is O.OIOOF in Cd2+?
that the cOnstant £0 is equal to the half-cell
potential when the logarithmic term is zero. and for the second immersed in a solution of cadmium ions will From Table 18-2, we find
This cond;tion occurs whenever the activity cause the deposition of metallic cadmium; con-
0.0591 1 Cd2+ + 2e~Cd(s) £0 = -0.403 V
quotient is equal to unity, one such instance E = 0.799 -100 log [Ag+]I00
versely, cadmium has little tendency to reduce
zinc ions. Table 18-2 also shows that iron(I1I)
being when the activities of all reactants and Thus,
products arc unity. Thus, the standard poten- Standard electrode potentials are available is a better oxidizing agent than triiodide ion;
tial is often defined as the electrode potential for numerous balf-reactions. Many have been therefore, in a solution containing an equilib-
of a balf-cell reaction (vs. SHE) when all reac- determined directly from voltage measure- rium mixture of iron(I1I~ iodide, iron(II~ and
tants and jproducts exist at unit activity. ments of cells in which a hydrogen or other
The standard electrode potential is an im- reference electrode constituted the other half
portant physical constant tbat gives a quanti- of the cell. It is possible, however, to calculate
tative desCription of the relative driving force £0 values from equilibrium studies of oxida-
for a half-cell reaction. Several facts regarding tion-reduction systems and from thermochem-
this constant should be kept in mind. First, ical data relating to such reactions. Many of TABLE 18-2 STANDARD ELECTRODE
the c1cctrodc potential is temperature-dcpen- the values found in the literature were so ob- POTENTIALS"
dent; if it is to have sianificancc, the tempera- tained.?
ture at which it is determined must be specified. For illustrative purposes, a few standard Reactioa ~ at 25"C, V
Second, the standard electrode potential is a electrode potentials are given in Table 18-2; a Clz(g) + 2e~2Cl- + 1.359
relative quantity in the sense that it is really more comprehensive table is found in Appen-
the potential of an electrochemical cell in which
Oz(g) + 4W + 4e~2HzO + 1.229
dix 2. The species in the upper left-hand part Brz(aq) + 2e~2Br- +1.087
the anode is a carefully specified reference elec- of the equations in Table 18-2 are most easily Brz(l) + 2e~2Br- +1.065
trode-that is, the standard hydrogen elec- reduced, as indicated by the large positive £0 Ag+ + e~Ag(s) +0.799
trode-whose potential is assigned a value of values; they are therefore the most effective Fe3+ + e~Fe2+ +0.771
zero volts. Third;the sian of a standard poten- oxidizing agents. Proceeding down the left-
Ii +2e~3I- +0.536
tial is identical with that of the conductor in hand side of the table, each succeeding species
Hg2Clz(s) + 2e~2Hg(l) + 2Cl- +0.268
contact Withthe half-cell of interest in a galvan- is a less effective acceptor of electrons than AgCl(s) + e~Ag(s) + Cl- +0.222
ic cell, the other half of which is the standard the one above it. The half-cell reactions at the
Ag(SZ03)~- + e~Ag(s) + 2Sz0~- +0.010
hydrosc* electrode. Finally, the standard p0-
2W + 2e~Hz(g) 0.000
tential is a measure of th~ intensity of the
AgI(s) + e~Ag(s) + 1- -0.151
driving force for a half-reaction. As such, it is
7 Two aUlhoritative sources for standard potenlia1 da1a
PbS04(s) + 2e~Pb(s) + SO~- -0.350
independent of the notation employed to ex- Cd2+ + 2e~Cd(s)
are: L MeiteI, HtIIIIlbook of AlIIJlytlcal CltemUtry. New -0.403
press the half-ccll process. Thus, the potential Yorlt: Mc:Graw-HiI1, 1963. pp. 5-6 10 5-14; and W. M. Zn2+ + 2e~Zn(s) -0.763
for the process Latimer. TM Oxi4arlon SUItes of tM Elements ItIIII TMir
Potentials in A4UfOIIS Solutio"'. 2d ed. EJII1ewood CIiIIS,
Ag+ + e~Ag(s) £0 = +0.799 V NJ.: Prenlice-HaI1. 19S2.
Substituting the 012 + 'concentration into this Electlode Potentlel. In is unity, the potential is the sum of two con-
the , •.•• ance of Preciplt8tIon stants; it is thus the standard electrode poten-
c:<Iuation gives
end Complex-Fonnlng R•••• enta tial for the half-reaction
E _ 1.065 _ 0.0~91 log(0.0100)2
0.0591 1
E - -0.403 - -2 - log 0.0100 As shown by the following example, reagents AgI(s) + e~Ag(s) + 1-
that react with the participants of an electrode
_ 1.065 - 0.0591 (-4.00) where
process have a marked effect on the potential
_ -0.403 _ 0.0~91 (+2.0) 2
for that process. E' - +0.799 + 0.059ilog KIP (18-17)
- 1.183 V
==-0.462 V EXAMPLE The Nemst relationship for the silver electrode
Calculate the potential of a silver electrode in a solution saturated with silver iodide can
The sign for the potential indicates the direc- in a solution that is saturated with silver iodide then be written as
tion of the reaction when this half-cell is cou- EXAMPLE
and has an iodide ion activity of exactly 1.00
pled with the standard hydrogen electrode. The Calculate the potential for a pla~um elec- E ==E' - 0.0591 log (1-]
(KIP for AgI ==8.3 X 10-1)
fact that it is negative shows that the reverse trode immersed in a solution that is O.OI00F
- -0.151 - 0.0591 log [r]
reaction in KBr and 1.00 x 10-3F in Br2' Ag+ + e~Ag(s) E' == +0.799 V
Here, the half-reaction used in the preced- Thus, when in contact with a solution sat-
Cd(s) + 2H+ ~H2(g) + 012+ ing example does not apply becau3e the solu- 1 urated with sillier iodiJe, the potential of a
E == +0.799 - 0.0591 log [Ag+)
tion is no longer lQtUTatedin Br2' Table 18-2, silver electrode can be described either in terms
occurs spontaneously. Note that the calculated however, contains the half-reaction of the silver ion concentration (with the stan-
We may calculate [Ag+] from the solubility
potential is a larger negative number than the dard electrode potential for the simple silver
product constant
standard electrode potential itself. This follows Br2(aq) + 2e~2Br- E' - 1;087 V half-reaction) or in terms of the iodide ion con-
from mass-law considerations because the half-
reaction, as written, has less tendency to occur The term (aq) implies that all of thd Br 2 pres-
ent is in solution; that is, 1.087 is the electrode.
[Ag+]== ~1 centration (with the standard electrode poten-
tial for the silver/silver iodide half-reaction).
with the lower cadmium ion concentration. The latter is usually more convenient.
Substituting into the Nernst equation gives

r]
potential for the half-reaction when the Br- The potential of a silver electrode in a so-
and Br2 solution activities are 1.00 mole/liter. lution containing an ion that forms a soluble
EXAMPLE It turns out, however, that the solubility ofBr2 E == +0.799 - 0.0~1 log
IP
complex with silver ion can be treated in a
Calculate the potential for a platinum elec- in water at 25°C is only about 0.18 mole/liter. fashion analogous to the foregoing. For exam-
trode immersed in a solution prepared by sat- Therefore, the m:orded potential of 1.087 is This equation may be rewritten as ple, in a solution containing thiosulfate and
urating • O.OI00F solution of KBr with Br2' based on • ltypotMtic4l system that CQIIIIOt be silver ions, complex formation occurs:
E - +0.799 + 0.05911og KIP
Here, the half-reaction is . realized experimentally. Nevertheless, this p0-
tential is useful because it provides the means - 0.0591 log W] (18-16) Ag+ + 2S20~- ~Ag(S203n-
Br2(l) + 2e~2Br- E' - 1.065 V by which potentials for undersaturated systems
If we substitute 1.00 for [1-] and use 8.3 [Ag(S203~-]
Note that the term (I) in the equation indi- can be calculated. Thus,
x 10-17 for KIP' the solubility product for K, ==[Ag+][S20~-]2
cates that the aqueous solution is. kept sat- AgI at 25.0°C, we obtain
urated by the presence of an excess of liquiJ 0.0591 [Br-]2 where K, is the formation constant for the
E - 1.087 - -2 - log [Br2] E- -0.151 V
Br2' Thus, the overall process is the sum of complex. The half-reaction for a silver elec-
the two equilibria 0.0591 (1.00 x; 1o-2f trode in such a solution is
This example shows that the half-cell poten-
- 1.087 --2-log 1.00 x 10-3 tial for the reduction of silver ion becomes Ag(S203n- + e~Ag(s) + 2S20~-
Br2(1)~ Br2(sat'd aq)
smaller in the presence of iodide ions. Qualita- The standard electrode potential for this half-
Br2(sat'd aq) + 2e~2Br- 0.0591
- 1.087 - -2-log 0.100 tively this is the expected effect because de- reaction will be the electrode potential when
The Nernst equation for the overall proc:tss is creases in the concentration of silver ions di- both the complex and the complexing anion
== 1.117 V . minish the tendency for their reduction. are at unit activity. Using the same approach
Equation 18-16 relates the potential of a as in the previous example, we find that
E _ 1.065 _ 0.0~91 log ~~~2
Here, the Br2 activity is 1.00 x 10-3 rather silver electrode to the iodide ion concen-
than 1.00, as was the situation when the solu- tration of a solution that is also saturated 1
tion was saturated. with silver iodide. When the iodiJe ion activity
E' ==+0.799 + 0.0591 log K ,
AN INTRODUCTION TO ELECTROANALYTICAL CHEMISTRY

ity coefficients. The activity coefficients of the tiv.ely 1.0F, O.IF, and O.01F in hydrochloric gen electrode when the concentrations of reac-
Data for the potential of the silver elec- two species are less than one in this system acid. These differences are attributable to the ~ts and products are IF and the concentra-
trode in the presence of selected ions are given because of the high ionic strength imparted by diffa:ence in the degree of association of ferro- tions of any other constituents of the solution
in the tables of standard electrode potentials in ~amde and ferricyanide ions with hydrogen are carefully s~ed. Thus, for example, the
the perchloric acid and the iron salts. More
~~~2andT~IeI~2.s~winwrma- important, however, the activity coefficient of Ions. The hydroferrocyanic acids are weaker ~ormal pot~tlal for the reduction of iron(III)
tion is also provided for other eJectrode sys- than th~ hydroferricyanic acids; thus, the con- ~ +0.732Vm IF perchloricacid and +O.700V
the iron(Ill) ion is smaller than that of the
tems. Such data often simplify the calculation centration of the ferrocyanide ion is lowered m IF ~ydrochloric acid; simiwly, the formal
iron (II) ion, inasmuch as the effects of ionic
of half-cell potentials. strength on these coefficients increase with the m~re than that of the ferricyanide ion as the potential for the reduction of ferricyanide ion
~cld concentration increases. This effect tends would be .+0.71 V in IF hydrochloric acid and
charge on the ion (p. 508). As a consequence,
I~turn to shift the oxidation-reduction equilib- +0.48 V m a O.OIF solution of this acid. Use
the ratio of the activity coefficients as they
num to the right and leads to more positive of these values in place of the standard elec-
Some limItationS to the Use appear in the Nernst equation would be larger
electrode potentials. trode potential in the Nernst equation will yield
of Standard Electrode Potentials than one and the potential of the half-cell
would be smaller than the standard potential. . A somew~at analogous effect is encountered bet~er agreement between calculated and ex-
Standard electrode potentials are of great im- Activity coefficient data for ions in solutions ~ the behaVior ofthe potential of the iron{III)/ peruoental. potentials, provided the electrolyte
portance in understanding electroanalytical of the types commonly encountered in oxida- lr~(II) couple. As noted earlier, an equiformal concentration of the solution approximates
processes. There are, however, certain inhe~t mature of these two ions in IF perchloric acid that for which the formal potential was mea-
tion-reduction titrations and electrochemical
limitations to the use of these data that should h~ an electrode potential of +0.73 V. Substi- sured. :,"p~lication offormal potentials to sys-
work are fairly limited; consequently, mow
be clearly appreciated. tutIOn of hydrochloric acid of the same con- t~ dilfenng greatly as to kind and concentra-
con~trations rather than activities must be
s-.itutioB or· Co8c:eDb'lltioM(or Activi- used in many calculations. Appreciable errors centration alters the observed potential to tion of electrolyte can, however, lead to errors
ties. As a matter of convenience, molar con- +0.70 V; a value of +0.6 V is observed in IF greater than those encountered with the use of
centratiOns-rather than activities-of reaC\ive may result. phosphoric acid. These differences arise be- stand~rd potentials. The table in Appendix 2
Etrm fIl Other Equilibria. The application
species are generally employed in the Ne.;nst of standard electrode potentials is further com- ~use iron~lII) forms more stable complexes contains selected formal potentials as well as
equation. Unfortunately, the assumption that plicated by the occurrence of solvolysis. disso- ~Ith chloride and phosphate ions than does standard pot~ntials; ~nsubsequent chapters, we
these two quantities are identical is valid only aation, association, and complex-formation Iron(II). As a result, the actual concentration shall use.whichever ISthe more appropriate.
in very dilute solutions; with increasing elec- reactions involving the species of interest. The of uncomplexed iron(lII) in such solutions is R~tlOn Rates. It should be realized that
trolyte concentrations. potentials calculated on less than that of uncomplexed iron{II~ and the the exIStence of a half-reaction in a table of
equilibrium constants required to correct for
the basis of moW concentrations can be ex- l net effect is a shift in the observed potentiaL electrode potentials does not necessarily imply
these effectsare freqUC!1tlyunknown. Lingane
pected to depart from those obtained by ex- . Phenomena such as these can be taken th.at there is a real electrode whose potential
cites the ferrocyanide/ferricyanide couple as an
periment. mto account only if the equilibria involved will respond to the half-reaction. Many of the
excellent example of this problem: data in. such tables have been obtained by
To illustrate, the staDdard electrode poten- are ~nown and constants for the processes are
tial wr the half-reaction aval~able. Often, however, such information is calculationS based upon equilibrium or ther-
Fe(CN>:- +e~Fe(CN>:- lackmg; the chemist is then forced to neglect mal measurements rather than from the actual
E' = +0356 V such effects and hope that serious errors do measurement of the potential for an electrode
not flaw the calculated results. system. For some, no suitable electrode is
is +0.771 V. Neglecting activities, we would Although the hydrogen ion does not appear F~I Potentials. In order to compensate known; thus. the standard electrode potential
predict that a platinum electrode immersed in in this half-reaction, the experimentally mea- ~artlally for activity effects and errors result- for the process,
a solution that was one forrntJl in iron(II~ sured potential is markedly affected by pH. mg from side reactions, Swift' has proposed
iron(lII~ andperchloric acid would exhibit a Thus, instead of the expected value of + 0.356 t~e ~se of a quantity called the fOmull poten-
potential numerically equal to this value rela-
V, solutions containing equiformal concentra- t~l I.n pla~ of the standard electrode poten-
tive to the standard hydrogen electrode. In tions of the two species yield potentials of tial m oxidation-reduction calculations. The
fact, however, a potential of +0.732 V is ob- formal potential of a system is the potential of
+ 0.71, + 0.56, and + 0.48 V with respect to the
served experimentally. The reason for th~ dis- the half-cell with respect to the standard hydro-
standard. hydrogen electrode when the mea-
crepancy is seen if we write the Nernst equa-
surements are made in media that are respec- has been arrived at indirectly. The reaction is
tion in the form not reversible, and the rate at which carbon
~i~xide combines to give oxalic acid is neg-
E = E' - 0.0591 log [Fe:+lfFC" ligibly slow. No electrode system is known
[Fe +lfFc•• • J. J. LiJIpne, El«tr_'ytklJl Clteminry.2d eel. New
'.E. H. Swift. A Systmt o/Chemicill AMl,.u. San Fran-
CISCO: Freeman. 1939. p. so. whose potential varies with the ratio of activi-
Yorlt:lntencience. 1958. p. 59.
where fFc" and fF." are the respective activ-
&24 PRINCIPLES OF INSTRl,)MENTAL ANALYSIS

For the zinc half-reaction, The Nemst equation for the Pb electrode now
The positive sign for the cell. potential indi-
ties of the reactants and products. Nonetheless, becomes
cates that the reaction [Zn2+] == 5.00 X 10-4
the potential is useful for computational pur- and
Zn(s) + eu2+ •.• Zn2+ + CU(s) EPb = -0.350 _ 0.0591 x
poses.
Ez" == -0.763 _ 0.05911 1 2
occurs ipontaneously and that this is a gal- 2 og 5.00 X 10-4
vanic celL log 0.820 x 5.00 x 10-4
The foregoing cell, diagrammed as = -0.860 V
CALCULATION = -0250
OF CELL POTENTIALS I
Cu CuI +(acul+ = 1.(0)11 Since the Pb electrode is specified as the cath-
For the zinc electrode
FROM ELECTRODE POTENTIALS ode,
Zn2+(az,,2+ c 1.(0)\Zn
E,oll = -0252 - (-0.860) == 0.608 Ez" = -0.763 - 0.0591 x
An important use of standard electrc,xte potc:n- implies that the copper electrode is now the 2
Cell potentials at the other concentrations can
tials is the calculation of the potential obtain- anode. Thus,
be derived in the same way. Their values are log 1
able from a galvanic cell or the potential re- =
Ecell -0.763 - ( +0.337) = -1.100 V tabulated at the bottom of the page. 0.825 x 5.00 x 10-4
quired to operate an electrolytic celL These
(b) To obtain activity coefficients for Zn2+
calculated potentials (sometimes called tlaermo- The negative sign indicates the nonsponta- == -0.863
and SOi- we must first 'calculate the ionic
dynamic potentials) are theoretical in the sense neity of the reaction
strength with the aid of Equation 18-3.
that they refer to cells in which there is es- Cu(s) + Zn2+ •.• Cu2+ + Zn(s)
sentially no current; additional factors must be /l == H5.00 X 10-4 x (2)1 E.olI = -0250 - (-0.863) == 0.613 V
taken into account where a current is involved. The application of an external potential greater
than 1.100 V would be required to cause this + 5.00 x 10:4 x (2)1] Values at other concentrations are listed at the
bottom of the page.
reaction to occur. == 2.00 X 10-3;
i It is of interest to compare the calculated
Calculation of In Table 18-1, we find ,for SOi -, CXA = 4.0 cell potentials shown in the columns labeled
~ic Call Potentials
EXAMPLE and for Zn 1 +• CXA == 6.0. Substituting ihese val- (a) and (b) in the foregoing example with the
As shown earlier (Equation 18-12~ the electro- Calculate the potentials for the following ues into Equation 18-4 gives for sulfate ion experimental results shown in the last column.
motive force of a cell is obtained by combining cell employing (a) concentrations and (b) Clearly, the use of activities provides a signifi-
r 0.5085 x 21 x .J2.00 X 10-3
half.<:ell potentials as follows: activities : - Iog JSO = -;--=-:-:---~====
3
cant improvement at the higher ionic strengths.
• 1 + 0.3281 x 4.0J2.00 x 10-
ZnIZnS04(xF~ PbS04(sat'd)lPb
- 8.59 x 10-1
where x == 5.00 X 10-4, 2.00 X 10-3, 1.00 EXAMPLE
where E._ and E•••_ are the electrode p0- 1 Iso. = 0.820 Calculate the potential required to initiate
X 10-1,2.00 X 10-1, and 5.00 X 10- .
tentials for the two half-reactions constituting the deposition of copper from a solution that
(a) In a neutral solution, little HSOi will Repeating the calculations employing CXA = 6.0
the cell for Zn2+ yields is O.OIOF in CuS04 and contains sufficient
be formed; thus, we may assume that
Consider the hypothetical cell sulfuric acid to give a hydrogen ion concen-
4 fz" = 0.825 tration of 1.0 x 1O-4M.
ZnIZnS04(az,,2+ = 1.<)0)1\
[SOi-] == Fz"so. == x == 5.00 X 10-

CUS04(ac..2+ = 1.(0)\Cu The half-reactions and standard potentials are


(a) (b)
PbS04(s) + 2e~Pb(s) + soi- x E(caIc)
The overall cell process involves the oxida- Jl E(calc) E(exptl)"
tion of elemental zinc to zinc(II) and the re- E'= -0.350 V 5.00 x lO-f 2.00 x 10-3 0.608 0.613 0.611
duction of copper(Il) to the metallic state. Be- 2.00 x 10-;3 8.00 X 10-3 0.583
Zn1+ + 2e~Zn E' == -0.763 V 0.572 0.582
cause the activities of the two ions are specified 1.00 x 10-1 4.00 x 10-1 ·0.531 0.549 0.553
as unity, the standard potentials are also the The potential of the lead electrode is given by - 2.00 x 10-1 8.00 x 10-1 0.513 0.537 0.542
electrode potentials. The cell diagram also 5.00 x 10-1 2.00 X 10-1 0.490 0.521 0.529
specifies that the zinc electrode is the anode.
EPb== -.3 0.0591 Iog. 500 x 10-4
O 50 --2-
Thus, using E' data from Table 18-2, • Expcrimcotal cIala from: I. A. Cowperthwaite and V. K. LaMer, J. Amer. Chern. Soc
53, 4333 (1931~ •
Ecell = +0.337 - (-0.763) == + 1.100 V == -0252 V
driving force for this migration being propor- potential can be cal~ted from a knowledge The net effect of 1R drop is to increase the
The deposition of copper necessarily occurs
tional to the concentration difference. The rate of the mobilities of the two ions involved. potential required to operate an electrolytic
at the cathode. Because no easily oxidizable
at which various ions move under the in6uence ~owever, it is seldom that a cell of analytical cell and to decrease the measured potential of
species are present, the anode reaction will
of a fixed force varies considerably (that is, ~portance has a sufficiently simple composi- a galvanic cell Therefore, the 1R drop is always
involve oxidation of H20 to give 02·-From
their mobilities are different). In the present tion to permit'such a computation. _ subtracted from the theoretical. cell potential.
the table of standard potentials, we find
example, hydrogen ions are several times more It is an experimental fact that the magni- That is,lO
Cu2+ + 2e~Cu(s) EJ == +0.337 V mobile that chloride ions. As a consequence, tude of the liquid junction potential can be
there is a tendency for the hydrogen ions to greatly decreased by interposition of a concen- E••II == Eca._. - E.aod• -IR (18-18)
02(g) + 4H+ + 4e -+ 2H20 EJ == + 1.229 V
trated electrolyte solution (a salt bridge) be-
outstrip the chloride ions as diffusion takes
Thus, for the copper electrode place; a separation of charge is the net result tween the two solutions. The effectiveness of EXAMPLE
(see Figure 18-8). The more dilute side of the this contrivance improves as the concentration 1. Calculate the potential when 0.100 A is
0.0591 1
E == +0.337 - -2-log 0.010 == +0.278 V boundary becomes positively charged owing to of the salt in the bridge increases and as the drawn from the galvanic cell
the more rapid migration of hydrogen ions; mobilities of the ions of the salt approach one
Assuming that O2 is evolved at 1.00 atm, the the concentrated side, therefore, acquires a another in magnitude. A saturated potassium I
Cd Cd2 +(0.0IooM)IICu2 +(0.0100M) Cu I
potential for the oxygen electrode is negative charge from the slower-moving chlo- chloride solution is good from both stand- Assume a cell resistance of 4.00 n
ride ions. The charge that develops tends to points, its concentration being somewhat great- Substitution into the Nernst equation re-
E - + 1.229 _ 0.0591 counteract the differences in mobilities of the er than 4F at room temperature, and the mobil- veals that the electrode potential for the Cu
- 4 ity of its ions dilferiJig by only 4 %. With such
two ions, and as a consequence, an equilibrium electrode is 0.278 V, while for the Cd elec-
1 condition soon develops. The junction poten- a bridge, the junction potential typically trode it is -0.462 V. Thus, the cell potential
log (1.00)(1.0 x 10-4 r tial difference resulting from this charge separa- amounts to a few millivolts or less, a negligible
quantity in many, but not all, analytical mea-
is
tion may amount to 30 mV or more.
••• +0.993 V surements. E == Ec. - Ec.J
In a simple system such as that shown in
The cell potential is then Figure 18-8, the magnitude of the junction •••0.278 - (-0.462) •••0.740 V

Eed •••• +0.278 - 0.993 == -0.715 V E••II == 0.740 -IR

Thus, to initiate the reaction EFFECT OF CURRENT = 0.740 - (0.100 x 4.00) == 0.340 V
ON CELL POTENTIALS
2Cu'2+ + 2H20 .•.•02(g) + 4H+ + 2Cu(s) Note that the emf of this cell drops dramat-
ically in the presence of a current.
would require the application of a potential When electricity flows in an electrochemical 2. Calculate the potential required to gen-
greater than 0.715 V. cell, the overall potential may be in6uenced erate a current of 0.100 A in the reverse direc-
1F I : O.OIF
HC' I
I
, HC'
I
by three additional phenomena, namely, ohmic tion in the foregoing cell.
Uquid Junction Potential , I potential, concentration polarization, and ki-
E ••• Ec.J - Ec.
W Ii '
): netic polarization.
When two electrolyte solutions of different I , == -0.462 - 0.278 == -0.740 V
composition are brought in contact with one I I
another, a potential develops at the interface. CI--: ,
:
I
E••II == -0.740 - (0.100 x 4.00)
This junction potential arises from an unequal I I =-1.l4OV
distribution of cations and anions across the - ......-E --+- + To develop a current in either a galvanic or
j
boundary due to differences in the rates at an electrolytic cell, a driving force or a poten- Here, an external potential greater than 1.140V
which these species migrate. FIGURE 18-8 Sche- tial is required to overcome the resistance of would be needed to cause Cd 2 + to deposit and
Consider the liquid junction that exists in matic representation of the ions to movement toward the anode or the Cu to dissolve at a rate required for a current
the system a liquid junction showing cathode. Just as -in metallic conduction, this of 0.100 A.
the source of the junction force follows Ohm's law and is equal to the
HQ(IF) IHQ(O.OIF) potential EJ• The length product of the current in amperes and the re-
Both hydrogen ions and chloride ions tend to of the arrows correspond sistance of the cell in ohms. The force is gener-
.0 Here and in the lubsequenl dilcuaion we will.-ume
diffuse across this boundary from the more to the relative mobility ally referred to as the ohmic potential, or the the juoctioo potenlial is oesJi&iblerelative to the other
concentrated to the more dilute solution, the of the two ions. lR drop. potentials.
irrespective of the concentration of this cation forces as they relate to electrode processes.
the current; and the physical states of the spe-
PoIlH'izatlOn Effects in the bulk of the solution. The reduction of Whenever a concentration gradient devel-
cies involved in the cell reaction. Some of th~
The linear relationship between potential ~ cadmium ions is rapid and reversible; con- ops in a solution, molecules or ions diffuse from
factors are sufficiently understood to permit
the instantaneous current in a .cell (Equation quantitative statements concerning their effects sequently, the concentration of this ion in the the more concentrated to the more dilute re-
18-18) is frequently observed ~perimentally film of liquid lIm'OUIIding the .cadmium elec- gion. The rate at which transfer occurs is pro-
upon cell processes; others, however, can be
when I is small; at high currents, however, accounted for on an empirical basis only.
trode iI tletermitted at any instant by the poten- portional to the concentration difference. In an
marked departures from linearity occur. Under For purposes of discussion, polarization tial «(that electrode at that instant If the poten- electrolysis, a gradient is established as a result.
these circumstances, the cell is said to be polar- tial is changed, deposition or dissolution of orions being removed from the film ofsolution
phenomena are convenient~y divid~ ~to the
ized (see Figure 18-9~ ThUS;a po~ electro- two categories of concentratIon polarizatIon and cadmium occurs such that the cadmium ion adjacent to the cathode. Diffusion then occurs,
lytic cell requires application of potent~ ~g- kinetic polarization (also ~ overvoltage or in the surface film rapidly attains the concen- the rate being expressed by the relationship
er thari theoretical for a given current; sU11l~r- tration required by the Nerost equation.
overpotential). ; Rate of diffusion to cathode surface
ly, a polarized galvanic cell.develops potent~als ConceotratiOll Polarization. When the reac- In contrast to this substantially instantane-
that are smaller than predicted. The polanza- tion at an electrode is rapid and reversible, the ous surface process, the rate at which equilib- = k(C - Co) (18-19)
tion ofa cell may be so extreme that the current concentration of the reacting species in the riU!;Jlbetween the electrode and the bulk of the
becomes essentially independent of the voltage; layer of solution immcdu,te~y adjacent to the solution is attained can be very slow, depending where C is the reactant concentration in the
under these circumstances, polarization is said electrode is always that Which would be pre- upon the magnitude of the current as well as the bulk of the solution, Co is its equilibrium
to be complete. dicted from the Nerost equation. Thus, the cad- volume of the solution and its solute concen- concentration at the cathode surface, and k is
polarization is an electrode phenomenon; mium ion concentration Co in the immediate tration. . a proportionality constant. The value of Co is
either or both electrodes in a cell can be af- vicinity of a cadmium electrode will always be When a sufficient potential is applied to fixed by the potential of the electrode and can
fected. Included among the fadors influenqing this electrode, cadmium ions are reduced, and be calculatetl from the N ernst eqlUltion. As
given by
the extent of polarization are the size, shape, an instantaneous current is generated. For higher potentials are applied to the electrode,
and composition of the electrodes; the compo- current to continue at a level predicted by Co becomes smaller and smaller and the dif-
.,0 O.ds91 10 1
Equation 18-18, however, an additional supply fusion rate greater and greater.
sition of the electrolyte solution; the tempera- E = r.e.t --2- g Co
ture and the rate of stirring; the magnitude of of the cation must be transported at a suitable Electrostatic forces also influence the rate
rate into the surface film surrounding the elec- at which an ionic reactant migrates to or
trode. If the demand for reactant cannot be from an electrode surface. The electrostatic
Cu'+ + Cd(s)-CulsJ + Cd"+ met by this mass transfer, concentration polar- attraction (or repulsion) between a particular
CulS) + Cd'+ - Cu'+ + Cd (s)
I \ ization will set in, and loweretl currents must ionic species and the electrode becomes small-
\
I
\
result. This type of pOlarization, then, occurs er as the total electrolyte concentration of the
I
I \ when the rate of transfer of reactive species solution is increased. It may approach zero
. I
l
between the bulk of the solution and the elec- when the reactive species is but a small fraction
trode surface is inadequate to maintain the of the total concentration of ions with a given
current at the level required by Ohm's law. A charge.
departure from linearity such as that shown in Qearly, reactants can be transported to an
Figure 18-9a is the consequence. Inadequate electrode by mechanical means. Thus, stirring
material transport will also cause concentra- or agitation will aid in decreasing concen-
tion polarization in a galvanic cell such as that tration polarization. Convection currents, due
described by the curve in Figure 18-9b; here, to temperature or density differences, also tend
however, the current would be limited by the to offset concentration polarization.
transport of copper ions. To summarize, then, concentration polar-
0.6
Ions or molecules can be transported ization occurs when the forces of diffusion, elec-
Cellpotential. V trostatic attraction, and mechanical mixing are
through asolution by: (1) diffusion; (2) electro-
static attraction or repulsion; and (3) mechan- insufficient to transport the reactant to or from
(I) ElectJolyticcoil lbl Go'-ic coil
ical or convection forces. We must therefore an electrode surface at a rate demanded by the
FIGURE 18-9 Current-voltage curves for the cell: (a) CulCu2+ consider the variables that influence these theoretical current. Concentration polariza. .•.....
(1.00 M)ICd2(I.oo M)ICd, (b) CdICd2+(1.oo M)IICu2+(UIO M)ICu.
AN INTRODUCTION TO ELECTROANALYTICAL CHEMISTRY

tion causes the potential of a galvanic cell to 5. The magnitude of overvoltage in any given As with IR drop, the overvoltage is subtracted
be smaller than the value predicted on the basis situation cannot be predicted exactly, be- from the theoretical cell potential . semble and should maintain an essentiall con-
of the theoretical potential and the IR drop cause it is determined by a number of un- stant and reproducible potential in th y
of small epresence
(Figure 18-9b~ Similarly, in an electrolytic cell, controllable variables. 11 currents. Several electrode systems
a potential more neptive than theoretical is Overvoltage associated with the evolution meet these requirements.
required in order to maintain a given current of hydrogen and oxygen is of particular interest
(Figure 18-9a). to the chemist. Table 18-3 presents data that
depict the extent of the phenomenOn under
c;.lomei Electrodes
Concentration polarization is important in In ~y eleetroanalytical applications, it is
several electroanalytical methods. In some ap- specific conditions. The difference between the desirable that the half-c:ell potential of one Calomel half-c:ells may be represented as f, 1-
plications, steps are taken to eliminate it; in overvoltage of these gases on smooth and on lows: 0
~~e be known, constant, and completely
others, however, it is essential to the method, platinimd platinum surfaces is of particular msensttive to the composition of the solution
and every effort is made to promote its occur- interest. This difference is primarily due to the under study. An electrode that fits this descrip- HHg202(sat'd1 KO(xF) Hg I
rence. The degree of concentration polarization much larger surface area associated with plati-
!ion is.~ a reference electrode.12 Employed where x .represents the formal concentration
is influenced experimentally by: (1) the reactant nized electrodes. which results in a real current m conjunction with the reference electrode will
density that is significantly smaller than is ap- of POtassium chloride in the solution. The elec-
concentration; (2) the total electrolyte concen- be an indicator electrode, whose response de- trode reaction is given by the equation
tration; (3) mechanical agitation; and (4) the parent from the overall dimensions of the elec- pends upon the analyte concentration.
size of the electrodes; as the area toward which trode. A platinimd surface is always employed A reference electrode should be easy to as. Hg202(5) + 2e~2Hg(l) + 2CI-
a reactant is transported becomes greater, p0- in construction of hydrogen reference elec-
larization effects become smaller. trodes. so as to lower the current density to a The potential of this cell will vary with the
Kinetic PoiarizatioL Kinetic polarization point where the overvoltage is negligible. chloride con~tration X, and this quantity
12 For a description or commen:ially available reference
results when the rate at which the electro- The high overvQltage associated with the must be 5peciti~ in describing the electrode.
electrodes, lee: R. D. Caton, Jr~ J. CIwrn. E4uc 50 AS71
chemical reaction at one or both electrodes is formation of hydrogen permits the electro- (1973); 51, A7 (1974~ ~ , Ta~le 18-4 lists the composition and the
slow; here, an additional potential (the over- lytic deposition of several metals that require potentials for the three most commonly en-
voltQ{/e or overpotential) is required to over- potentials at which hydrogen would other-
come the energy barrier to the half-reaetion. wise be expected to interfere. For example, it
In contrast to concentration polarization, the is readily shown from their standard poten-
current is controlled by the rate of the electron-
trIIIU/er process rather than by the rate of mass
transfer.
tials that rapid formation of hydrogen should
occur well before a potential sufficient for the
deposition of zinc from a neutrailOlution can
TAF-:;-:~:~~O:~~R~~~~~~:
:~:~;::~:~~ ~~. OXYGEN
Although exceptions can be cited, some em- be realimd. Nevertheless, quantitative deposi- ~\) OYerY= OYerYoIta&.
(V)
pirical generalizations can be made regarding tion of zinc can be attained provided a mer- I A/ern2) ICy
Electrode ~~~2) . ~~) it)'
the magnitude of overvoltage. cury or copper electrode is used; because of
1. Overvoltage increases with current density the high overvoltage of hydroleD on these Composition H2 O2 H2 O2 H2 O2
(current density is defined as the amperes metals. little or no gas is evolved during the Smooth Pt 0.024 0.721 0.068 0.85
per square centimeter of electrode surface ~ electrodeposition. Platinized Pt 0.015 0.676 1.49
0.348 0.030 0.521 0.048
2. It usually decreases with increases in tem- The magnitude of overvoltage can, at best, Au 0241 0.673 0.76
Cu 0.391 0.963 0.798 1.63
perature. be only crudely approximated from empirical 0.479 0.422 0.584 0.580
Ni 0563 1269 0.793
3. Overvoltage varies with the chemical com- information available in the literature. Calcula- 0.353 0.747 0.519 1.241
Hg O.~ 1.1' 0.853
position of the electrode, often being most tion of cell potentials in which overvoltage Zn 1.14 .
0.716 0.746 1.229
pronounced with softer metals such as tin, plays a part cannot, therefore, be very accurate. Sn 0.856 1.077 1231
lead. zinc, and particularly mercury. Pb 0.52 1.090 1262
4. Overvoltage is most marked for electrode Bi 0.78 1.05 123
processes that yield gaseous prodiJets such
as hydrogen or OXYIeD;it is frequently neg- • Natioaal Academy oI'Sc:icacea, IIIUr1tiItJoIIQJ CrilIctIJ -
II OvervoJtaae data for various pICOUI species aJ clilfer- pp. 339-340. (WiJb )IeI1IIlIIioa) TlIbla. New York: McGraw-HiB. 1929, voL 6,
ligible where a metal is being deposited or
eat electrode .urr- arc to be found in: HaMbooIcof
where an ion is undergoing a change of oxi- AItalytiarl CIIemUtry, cd. L Mates. New Yark: McGraw- • 0.556 V at O.oooon A/emJ; 0.929 V at O.ooIS4 A/emJ.
< U163 V at 0.00769 A/emJ.
dation state. HiB. p. 5-184.
• 1.126 V at 1.IS3 A/emJ.
is.S to 15 em in length and 0.5 to 1.0 em in
diameter. A mereury/mereury(I) chloride paste
is contamed in an inner tube that is connected
~o the saturated potassium chloride solution
m the outer tube through a small opening. For
~lectrode (a~ contact with the second half-eell
IS made by means of a fritted disk or a porous
Solution Slltureted fiber ~led ~ the end of the outer tubing. This
with KCllIId Hg2CI2 type ofJunction has a relatively high resistance
Solid KCI (2000~o3000 0) and a limited current-carrying
capaCIty; on the other hand, contamination of
the analyte solution due to leakage is minimal.
The electrode shown in Figure 18-11b has a
Hg
much lower resistance but tends to leak small
HlIlf·l'IICtlon: amounts of saturated potassium chloride into
HII2Cl2 + 2." 2Hg + 2Cl-
the sample. Before it is used, the ground glass
~ ~ collar of this electrode is loosened and turned
FIGURE 18-:10 Two easily constructed reference electrodes. (a) A saturated calomel electrode. so that a drop or two of the KCI solution
(b) A si1ver/sil~er chloride electrode. flows from the hole and wets the entire inner
ground surface. Better electrical contact to the
analyte solution is thus established.

countered calomel electrodes. Note that each trode potentials for calomel half-eells at tem-
solution is saturated with mercury(I) chloride peratures t other than 25°C Silver/Silver Chloride Electrodes
and that the cells differ only with respect to The saturated calomel electrode (SCE) is A reference electrode system analogous to the
the potassium chloride concentration. Note most commonly used by the analytical chemist calomel electrode consists of a silver electrode
also that the potential of the normal calomel because of the ease with which it can be pre- Ground
immersed in a solution of potassium chloride
glass
eIec:trode is Jfellter than the standard poten- pared. Compared with the other two, however, sleeve that has been saturated with silver chloride
tial for the half-reaction (£0 == 0268 V) be- its temperature coefficient is somewhat larger.
cause the chloride ion activity in a IF solution A simple, easily constructed saturated cal- (b) 'AgCI(sat'd~ KC(xF)I Ag
of potassium chloride is significantly smaller omel electrode is shown in Figure 18-10a. The FIGURE 18-11 Typical commercial calo- The half-reaction is
than one. The last column in Table 18-4 gives salt bridge, a tube filled with saturated potas- mel reference electrodes.
expressions that permit the calculation of elec- sium chloride, provides electric contact with AgC(s) + e~ Ag(s) + CI-
Normally, this electrode is prepared with a
satu.rated po~iu~ chloride solution, the p0-
the solution surrounding the indicator elec- tential at 25 C being 0.197 V with respect to
trode. A fritted disk or wad of cotton at one the standard hydrogen electrode.
end of the salt bridge is often employed to A simple and easily constructed silver/silver
SPECIFICATIONS OF CALOMEL ELECTRODES prevent siphoning of the cell liquid and con- chloride electrode is shown in Figure IS-lOb.
EJectnNIe PoteaPal (V) tamination of the solutions by foreign ions' The electrode is contained in a Pyrex tube fitted
Coaceatratioll • ••.SIuUri IIydropD alternatively, the tube can be filled with a 5% with a IG-mm fritted-glass disk. A plug of apr
EIectralIe agar gel that bas been saturated with potas- gel saturated with potassium chloride is formed
Name ",202 ICO ",202(1) + 2e ~ 2H&(l) + 10- sium chloride.
on top of the disk to prevent loss of solution
Several convenient calomel electrodes are
Saturated
Normal
Saturated
Saturated
Saturated
1.0F
+0241 - 6.6 x 1O-4
4
ft -
251
+0280 - 2.8 x 10- t - 25 available commercialJy; typical are the two
from the half-eell. The plug can be prepared
by heating 4 to 6 g of pure agar in 100 m1 of
Decinormal Saturated O.IF +0.334 - 8.8 x 10- t - 25
5 illustrated in Figure 18·11. The body of each water until a clear solution is obtamed; about
electrode consists of an outer glass tube that 35 g of potassium chloride are then added. A
Cell reaction:
portion of this mixture, while still warm, is CdIHg). + H92SO." Cd'· + I" + 2)Hg + 50"- 2. Calculate the electrode potentials of the following half-cells
E-l.0183 V (a) Fez +(0.132M) IFe .
poured into the tube; upon cooling, it solidifies
to a gel with low electrical. resistance. A layer (b) BiO+(2.00 x 1O-3M), W(3.00 x lO-zM)lBi
of solid potassium chloride is placed on the gel, Saturated
(c) HCl(4.SOF) IHz (1.00 atm), Pt
and the tube is filled with a: saturated solution CdSO. (d) HOAc(O.OlOOF)IHz(1.00 atm), Pt
solution (e) IOj(O.l00M), Iz(0.235M), H+(2.00 x 10-3M)I Pt
of the salt A drop or two of IF silver nitrate
is then added and a heavy gauge (1- to 2-mm (f) AgzCr04(sat'd), crO~-(O.04OOM)lAg
diameter) silver wire is inserted in the solution. 3. Calculate the potential of a silver electrode in contact with the follow' g
Silver/silver chloride reference electrodes are (a) 1.00 x 1O-4M Ag+ In .
also available from commercial sources. Their (b) O.06OMAg+
construction is similar to that of the electrodes (c) a solut~on that ~ O.04OOMin .- and saturated with Agi
shown in Figure 18-11. (d) a solutlo~ that IS O.OOSOMin CN- and 0.0300M in Ag(CNt
z
(e) the solutIon that results from mixing 25.0 ml of O.OSOOFKBr with
20.0 ml of O.l00M Ag+
(f) the solution that results from mixing 25.0 ml ofO.050M Ag+ with
FIGURE 18-12 A Weston standard cell 20.0 ml of O.l00F KBr
(saturated ).
4. Calculate the electrode potentials for the following systems.
The accurate measurement of potentials for (a) y3+(3.00 x 1O-3M), Yoz+(1.SOM), HCI(O.lOOF)lPt
electrochemical work requires the frequent use (b) MnOi(O.200M), Mn2+(O.OSOOM),W(0.0800M)IPt
of a cell whose emf is precisely known. The 5. Calculate the electrode potentials for the following systems.
Weston cell, which is used almost universally the mercury(l) sulfate. As a result, Weston cells
remain constant in voltage for remarkably long (a) CrzO~-(5.00 x 1O-3M), Cr3+(1.00 x lO-zM), W(O.lOOM)lPt
for this purpose, can be represented as follows: (b) UO;(O.lOOM), U4+(0.200M), W(O.300M)/Pt
periods oftime, provided large currents are not
Cd(Hg)lCdS04' 8/3HzO(sat'd), drawn from them. 6. Calculate. the theoreti.cal potential of each of the following cells. Is the
The Weston cell has a temperature coeffi- cell as written galvamc or electrolytic?
HgzS04(sat'd)I Hg
cient of about -0.04 mY rC, due primarily to (a) PtICr3+(2.00 x 1O-4M), Cr2+(1.00 x 1O-IM)IIPb2+(4.00 x
changes in solubility of the cadmium(lI) and lO-ZM)IPb
mercury(I) salts. A cell with a coefficient about . (b) PtISn4+(5.00 x 1O-4M), Sn2+(3.00 x 1O-ZM)IIAg(CN)i(2.5 x
Cd(Hg)" ...•Cd2+ + xHg(l) + 2e one-fourth this magnitude is obtained by using 1O-3M), CN-(4.00 x lO-zM)lAg
a solution of cadmium sulfate that has been (c) HgIHg~+(2.00 x 1O-ZM)IIH+(3.00 x 10-zM),
HgZS04(s) + 2e ...•2Hg(l) + SO~- saturated at 4°C; no solid CdS04 . 8/3HzO is y3+(8.00 x lO-ZM), YOz+(4.00 x 1O-3M)IPt
incorporated in the cell itself. Most commer- Cd) PtIFe3+(4.00 x lO-zM),
Figure 18-12 shows a typical Weston cell.
cially available cells are of this type and are Fe2+(6.00 x 1O-SM)I\snZ+(5.00 x lO-zM),
Its theoretical potential at 25°C is 1.0183 Y.
termed unsQturated Weston cells. Their poten- Sn4+(2.00 x 1O-4M) Pt
The emf of a Weston cell is governed by
tials lie beJween 1.0185 and 1.0195 Y. (e) MnIMn2+(7.SO x 1O-3M)IIAg(CN)i(0.0800M),
the activities of the cadmium and mercury(l)
ions in the solution. At any given tempera- Weston cells may be sent to the National CW (O.0300M)I Ag
ture, these quantities are invariant, being fixed Bureau of Standards for calibration and certifi- 7. Calculate the theoretical potential of each of the following cells. Is the
by the solubilities of the cadmium sulfate and cation. cell galvanic or electrolytic as written?
(a) PtIFe(CN):-(4.00 x lO-zM),
Fe(CN)~-(8.00 x 1O-4M)W(2.00 x 1O-3M),
Iz(3.00 x lO-ZM)IPt
1. Calculate the electrode potentials of the following half-cells. (b) Ptly3+(4.00 x 1O-3M), y2+(0.800M)IIFe2+(3.00 x 1O-4M)IFe
(a) Ag+(O.0300M)IAg (c) BilBio+(o.04OOM), H+(2.00 x lO-ZM)W(O.200M),
(b) Ni2+(0.662M)INi AgI(sat'd) IAg
(c) HCl(1.SO x 1O-4F)I Hz(1.00 atm), Pt (d) ZnIZn2+(3.00 x 1O-4M)IIFe(CN>:-(4.00 x lO-zM),
(d) Fe3+(2.35 x 10-4M), Fe2+(O.l00M)lPt Fe(CN)~-(7.00 x lO-zM)lPt
(e) AgBr(sat'd), Br-(4.SOM)IAg (e) Pt, Hz(O.l00 atm)l HCI(5.00 x 1O-3F), AgCl(sat'd)l Ag
&37

8. Calculate the potential of each of the following cells. Is the cell gal- (b) Ce4+ + e~Ce3+ E = 1.44 V (in IF H SO )
vanic or electrolytic as written? (c) Tl+ + e~l1(s) E = '-0.33 V (in IF HClO )
(a) Cu ICuI(sat'd~ I-(0.250M)llr(3.50 x 1O-3M~ CuI(sat'd)I Cu 18. A current of 0.0510 A is to be drawn from the cell 4
2
(b) Pt, H2(0.800 atm)1 H+ (3.00 x 1O-6M)IIH+(2.00 x 1O- M)!
H2(0.200 atm~ Pt . PtIV3+(1.0 X 1O-5~~ VH(2.50 x 1O-lM)IIBr-(300 x lO-lM\
9. Compute ~ for the process
I
AgBr(sat'd) Ag . ."

Ni(CN)i + 2e~Ni(s) + 4CN- As a consequence of its design, the cell has an internal resistance of
4.30 Q. Calculate the potential to be expected initially
given that the formation constant for the complex is 1.0 x 1022. 19. The cell· .
10. The solubility product constant for TlBr is 3.4 x 10-6 at 25°C. Calcu-
late ~ for the process PtIV(OH)t(4.00 x 10-4M~ VOH(6.73 x 10-2M~

TlBr(s) + e~Tl(s) + Br- H+(2.75 x 10-3M)1I CuH(3.00 x 1O-2M)ICu .


11. Calculate the standard potential for the half-reaction ?as an inter:nal resistance of 1.25 Q. What will be the initial potential"
if 0.0300 A ISdrawn from this cell? .
Cu(OHh(s) + 2e~Cu(s) + 20W 20. The resistance of the galvanic cell
given that Kip for Cu(OHh has a value of 5,,9 x 10-15.
PtIFe(CN):-(3.60 x 1O-2M~ Fe(CN):-(2.70 x 10-3M)IIAg+
12. Calculate the standard potential for the half-reaction (1.65 x 10-2M)IAg
Al(C204); + 3e -+ Al(s) + 2c20i- is .2.5 n Calculate the initial potential when 0.0120 A is drawn fro
if the formation constant for the complex is 1.3 x 1013. thIS cell. m
13. Calculate the standard potential for the half-reaction 21. ~5~~~ute the chloride ion activity in a normal calomel electrode at!
FeY- + e~Fey2-
22. Compute the chloride ion activity in a decinormal calomel electrode"
if the formation constant for the EDTA complex of iron(I1I) is 1.3 at 25°C.
x 1025 and that for the iron(I1) complex is 2.1 x 1014.
14. From the standard potentials
Tl+ + e~Tl(s) ~= -0.336 V

TlI(s) + e~Tl(s) + 1- ~= -0.761 V


calculate the solubility product constant for TII.
15. From the standard potentials
Ag2Se04(S) + 2e~2Ag(s) + SeOi- ~ = 0.355 V

Ag+ + e~Ag(s) ~=0.799V

calculate the solubility product constant for Ag2Se04·


16. From the cell potentials shown on the right below, calculate the elec-
trode potentials for the half-cells coupled to the reference electrode.
(a) saturated calomel electrode IIM+"IM E = 0.561 V
(b) normal calomel electrode IIX3+, XH Pt I E = -0.642 V
(c) saturated, silver/silver chloride electrode IIMA(sat'd), A2-IM
E= -0.319 V
17. Convert each of the following electrode potentials to potentials versus
the saturated calomel electrode.
(a) CuH + 2e~Cu(s) EO = 0.334 V
In the preceding chapter, it was shown that Second-onler Electrodes for Anions. A
the potential of an electrode is determined by
metal electrode is also indirectly responsive to
the concentration (or, more correctly, the
anions that form slightly soluble precipitates
activity) of one or more species in a solution.
or stable complexes with its cation. For the
This chapter considers how this phenomenon
former, it is only necessary to saturate the
is applied to the quantitative determination of
solution under study with the sparingly sol-
ions or molecules.'
uble salt. For example, the potential of a silver
The equipment required for a potentiomet-
electrode will accurately reflect the concentra-
ric measurement includes a reference electrode,
tion of iodide ion in a solution that is saturated
an indicator electrode, and a potential measuring
with silver iodide. Here, the electrode behavior
device. The preparation and properties of some can be described by
common reference electrodes were considered
in Chapter 18; the latter two components will
be discussed in this chapter, in addition to
applications of potentiometric measurements
Application of the Nernst equation to this half-
to quantitative. analyses.
reaction gives the relationship between the
electrode potential and the anion concentra-
tion. Thus,

E = -0.151 - 0.0591log[r]
Indicator electrodes for potentiometric meas-
urements are of two basic types, namely, me- = -0.151 + 0.0591pI
tallic and membrane. The latter are also referred
to as specific or selective ton electrodes. where pI is the negative logarithm of the iodide
ion concentration.2 A silver electrode serving
as an indicator for iodide is an example of an
Metallic Indicator Electrode. electrode of the second-order because it mea-
First-onler Electrodes for Cations. A first- sures the concentration of an ion that is not
order electrode is employed to determine the directly involved in the electron transfer
cation derived from the electrode metal. process.
Several metals, such as silver, copper, mercury,
lead, and cadmium, exhibit reversible half-
reactions with their ions and are satisfactory 2 The results of potentiometric measurements are often
for construction of first-order electrodes. In expressed in terms or a parameter, the p function, that is
contrast, other metals are not very satisfactory directly proportional to the measured potential. The
as indicator electrodes because they tend to p runction then provides a measure or concentration in
terms or a convenient, small, and ordinarily positive,
develop non reproducible potentials that are
number. Thus, for a solution with a calcium ion concen-
influenced by strains or crystal deformations tration or2.00 x 10-6• we may write
in their structures and by oxide coatings on
pCa = -log(2.00 x 10-6) = 5.699
their surfaces. Metals in this category include
iron, nickel, cobalt, tungsten, and chromium. Note that as the concentration or calcium increases, its
P function decreases. Note also that because the concen-
tration was given to three significant figures, we are en-
titled to keep this number of figures to the right of the
4~ point in the computed pCa because these are the
I For a detailed review of potentiometric methods, see:
only numbers that carry information about the 0riJina1
R. P. Duct, in Physical Me1/roh of Chemi.slry, eds. A. 2.00.The 5 in the value for the pCa provides information
Weissberser and B. w. Rossiter. New York: Wiley- about the position of the decimal point in the original
Intencience, 1971, Part IIA, vol. I, Chapter 2. number only.
Membrane Indicetor Electrode.3
An important second-order electrode far
measuring the concentration of the EDT A For many years, the most convenient method
anion y4- (p. 567) is based upon the re- for determining pH has involved measure-
sponse of a mercury electrode in the presence ment of the potential that develops across a
of a small concentration of the stable EDTA thin glass membrane that separates two solu-
complex of Hg(II). The half-reaction for the tions with different hydrogen ion concentra-
electrode process can be written as tions. The phenomenon, first reported by
HgYz- + 2e~Hg(l) + y4- E'l = 0.21 V Cremer,4 has been extensively studied by many
investigators; as a result, the sensitivity and
for which selectivity of glass membranes to pH is reason-
0.0591 (y4-] ably well understood. Furthermore, membrane
E = 0.21 - -2-log [HgYz-] electrodes have now been developed for the
direct potentiometric determination of two
To employ this electrode system, it is ~eces-
dozen or more ions such as K +, Na +, Li +,
sary to introduce a small concentration of
F-, and ea2+.5
HgYz- into the analyte solution at the outset.
It is convenient to divide membrane elec-
The complex is very stable (for HgYz -, K I =
trodes into four categories based upon mem-
6.3 X lOZ1); thus, its concentration remains
brane composition. These include: (1) glass
essentially constant over a wide concentration FIGURE 19-1 Typical electrode system for measuring
electrodes; (2) liquid-m~brane electrodes; (3)
range of y4- because dissociation to form
solid-state or precipitate electrodes; and (4) pH.
HgZ + is slight. The foregoing equation can then
gas-sensing membrane electrodes. We shall
be written in the form consider the properties and behavior of the
glass electrode in particular detail both because
E=K--0.0591
2-w
g(Y4-] .0.0591
=K+--2-P
Y
of its historical and its current importance.
1ar to the ones described in Chapter 18 (p. 533). trodes, each with a potential that is constant
The glass electrode is manufactured by sealing and independent of pH; one reference electrode
where the constant K is equal to a thin, pH-sensitive glass tip to the end of a is the external calomel electrode, the other is
heavy-walled glass t~bing. The resulting bulb is the internal silver/silver chloride electrode
0.0591 1 The G•••• Electrode
K =0.21--2-10g [Hgy2-] filled with a solution of hydrochloric acid (often (p. 533~ which is a part of the glass electrode
for pH MeesurementB O.lF) that is saturated with silver chwride. A but is not the pH-sensitive component. In
This second-order electrode is useful for Figure 19-1 shows a modem cell for the mea- silver wire is immersed in the solution and is fact, it is the thin membrane at the tip of the
establishing end points of EDT A titrations. surement of pH. It consists of commercially connected via an external lead to one terminal electrode that responds to pH changes.
IDlIicators for Redox Systems. Electrodes available calomel and glass electrodes im- of a potential-measuring device; the calomel Figure 19-2 is a schematic representation
fashioned from platinum or gold serve as in- mersed in a solution whose pH is to be mea- electrode is connected to the other terminal. of the cell shown in Figure 19-1. It is found
dicator electrodes for oxidation-reduction sys- sured. The calomel reference electrode is simi- Note that the cell contains two reference elec- experimentally that at 25°C the potential of
tems. Of itself, such an electrode is inert; the
potential it develops depends solely upon the
potential of the oxidation-reduction systems of Reference
the solution in which it is immersed. For S For further information on this topic, see: lon-Selective electrode 1
example, the potential of a platinum electrode Electrodu, ed. R. A. Dunt. National Bureau of Stand- ~
ards Special Publication 314. Washington, D.C.: U.S.
in a solution containing Ce(II1) and Ce(IV) Government Printing Office. 1969; N. Lakshminarayan-
ions is given by aiah, M embrlJlle Electrodes. New York: Academic Press,
1976; J. Veae1y, D. Weis, and K. Stu!ik, Analysil with Glass
[Ce3+] membr •••
lon-Selealve"EIectrodes. New York: Wiley, 1979; and Ion
E = E'l - 0.0591 log [Ce4+] Selectloe Electrodes in AnalytlaJl C1IemUtry,ed. H.
V,
Reference electrode 2

Thus, a platinum electrode can serve as the


Freiser. New York: Plenum Press, 1978.
• M. Cremer. Z. Bioi, .7. S62 (1906~
E•••-.a
indicator electrode in a titration in which , See: G. A. Rechni1Z, Chem. Eng. News, ~ (2S~ 146 FIGURE 19-2 Schematic diagram of a glass-calomel cell for measure-
Ce(IV) serves as the standard reagent. (1967~ ment of pH.
such a cell depends on the hydrogen ion activ- widely used. This glass shows an excellent
ities al and a2 on either side of the membrane specificity toward hydrogen ions up to a pH
as follows: of about 9. At higher pH values, however, the
membrane becomes somewhat sensitive to so-
E = Q + 0.0591 log al (19-1) dium and other alkali ions.
a2 Hygroscopicity of Glass Memlnnes. It has
been shown that the surfaces of a glass mem-
The constant Q contains four components Surface sites Surface sites
brane must be hydrated in order to have pH occupied occupied
Q = EA1• AaCI - EscE + EJ + E. activity. Nonhygroscopic glasses such as Pyrex byW byW

where EA1• AaCI and EscE are the potentials of and quartz show no pH function. Even Coming
the two reference electrodes, EJ is the junction. 015 glass shows little pH response after dehy-
potential across the salt bridge, and E. is the dration by storage over a desiccant; its sen-
sitivity is restored, however, after standing for FIGURE 19-3 Schematic representation of a well-soaked glass
asymmetry potential; the source and proper- membrane. Note that the dimensions of the three inner layers are
ties of E. will be considered later. a few hours in water. Hydration involves ab-
sorption of approximately 50 mg of water per rwt to scale.
In practice, the hydrogen ion activity of the
internal solution a2 is fixed and constant. We cubic centimeter of glass.
It has also been demonstrated experimen-
may thus Write
tally that hydration of a pH-sensitive glass
E ••• L + 0.0591 log al (19-2) membrane is accompanied by a chemical reac- Electrical Resistance of Glass Membranes. Conduction within the two silicic gel layers
= L - 0.0591 pH
tion in which certain cations of the glass are The membrane in a typical commercial glass is due to migration of hydrogen and sod!um
exchanged for protons of the solution. This electrode will be between 0.03 and 0.1 mm thick ions. In the dry center region of the membrane,
where L is a new constant consisting of Q and exchange involves singly charged cations al- and will have an electrical resistance of 50 to sodium ions take over this function. I
the logarithmic function of a2 . . most exclusively, inasmuch as the di- and tri- 500 Mn Current conduction across the mem- Theory of the Glass Electrode Potential. 7
It is important to emphasize that, in princi- valent cations in the silicate structure are much brane involves migration of singly charged The potential across a glass membrane consists
pal, the potentials EscE, EA1• AaCI , EJ ' and E. more strongly bonded. The ion-exchange reac- cations. Across each solution interface, passage of a boundary potential and a diffusion potential.
remain constant during a pH measurement. tion can be written as of charge involves a transfer of protons; the Ideally, only the former is affected by pH.
Thus, the source of the pH-dependent varia- direction of migration is from glass to solution The boundary potential for the membrane
tion in E must lie across the glass membrane. H+ + Na+Ol- ~Na+ + H+OI- (19-3) at one interface and from solution to glass at of a glass electrode contains two components,
That is. when al and a2 differ, the two sur· soln solid soln solid the other. That is. each associated with one of the gel-solution
faces of the membrane must differ in potential interfaces. If VI is the component originating
H+OI- ~ 01- + H+ (19-4)
by some amount VI - V2. The only function where 01- represents a cation bonding site in at the interface between the external solution
solid solid soln
of the two reference electrodes is to make ob- the glass. The equilibrium constant for this and the gel (see Figure 19-3~ and if Jt2 is the
servation of this difference possible process favors incorporation of hydrogen ions corresponding potential at the inner surface,
Composition of Glass Membranes. Much then the boundary potential Eb for the mem-
into the silicate lattice; as a result, the surface H+ + 01- ~H+OI-
systematic investigation has been devoted to of a well-soaked membrane will ordinarily brane is given by
soln solid solid
the effects of glass composition on the sensitiv- consist of a layer of silicic acid gel which is
ity of membranes to protons and other cat- The position of these two equilibria is deter- Eb = VI - V2 (19-6)
10-4 to 10-5 mm thick. ,
ions, and a variety of compositions are now In all but highly alkaline media-where mined by the hydrogen ion concentrations in The potential VI is determined by the
used commercially.6 For many years, Coming sodium ions may occupy an appreciable num- the two solutions. When these positions differ, hydrogen ion activities both in the external
015 glass (consisting of approximately 22% ber of bonding sites-the predominant uni- the surface at which the greater dissociation solution and upon the surface of the gd; it
Na20, 6% Cao, and 72% Si02) has been valent cation in the outer surface of the gel is has occurred will be negative with respect to can be considered a measure of the driving
the proton. From the surface to the interior of the other surface. Thus, a potential develops •
the gel, there is a continuous decrease in the whose magnitude depends upon the difference
• For a summary of this work, ICC: J. O. liard, MTbe number of protons and a corresponding in- in hydrogen ion concentration on the two
DepeodCDCC of GIall-EJec:trodc Properties on Composi- sides of the membrane. It is this potential that
crease in the number of sodium ions. A sche- , For a comprehensive discussion of this topic, ICC: G.
tion, in G/Qu Electrodes for Hyrlrogen and Other CIJI-
M

matic representation of the two surfaces of a serves as the analytical parameter in a poten- Eisenman, G/Qu Electrodes for Hyrlrogen and Other Cat-
iollS. ed. G. Eisenman, New York: Marcel Dekker. 1967,
Chapter 3. glass membrane is shown in Figure 19-3. tiometric pH measurement. ions. New York: Marcel Dekker, 1967, Chapters 4--6.
fotce of the reaction shown by Equation 19-4. and the potential becomes a measure of the
hydrogen ion activity in the external solution. ~here a.1 and bl are activities of H+ and B+
In the same way, the potential Vz is related to A. Coming 015, H2SO4
the hydrogen ion activities in the internal so- In addition to the boundary potential, a B. Corning 015. HCI m solutiOn, and al and b1 are the activities of
C. Corning 015, 10M Na- the same ions in the surface of the gel. The
lution and on the corresponding gel surface; it so-called di.f/Usionpotential also develops in
O. Beckman-GP. I·M Na-
is a measure of the driving force of the reac- each of the two gel layers. Its source lies in constant K •• depends upon the composition
E. L & N Black Dot, 10M Na-
tion described by Equation 19-5. the difference between the mobilities of hydro- F. Beckman Type E. 10M Na' of the glass membrane and is ordinarily small.
It can be demonstrated from thermodynam- gen and alkali-metal ions in the membrane. Thus, the f~ion of the surface sites pop-
ic considerations8 that VI and Vzare related to The two diffusion potentials are equal and ulated by cationS other than hydrogen is small
hydrogen ion activities at each interface as opposite in sign insofar as the two solution- except when the hydrogen ion concentration
follows: gel interfaces are the same. Under these condi- is very low and the concentration ofB + is high.
tions, then, the net diffusion potential is zero, The effect of an alkali-metal ion on the emf
. RT al ~cr~ a membrane can be described in quan-
VI=h+-ln- and the emf across the membrane depends
F a'i only on the boundary potential as given by titatIve terms by rewriting Equation 19-11 in
Equation 19-10. the form
. RT az
J'2=h+-ln-
F az Asymmetry PoteatiaL If identical solutions
and identical reference electrodes are placed
pH
al _ (al + K..bd
FIGURE 19-4 Acid and alkaline error of
where R. T, and F have their usual meanings; on either side of the membrane shown in selected glass electrodes at 25·C. (From R. O.
al - (al + b1)
al and az are activities of the hydrogen ion in Figure 19-3, VI - Vzshould be zero. However, Bates, Determination of pH: Theory and Prac-
the solutions on either side of the membrane; it is found that a small potential, called the tice. New York: John Wiley & Sons, Inc., 1964, Substitution of Equation 19-12 into Equation
and' a'i and az are the hydrogen ion activities asymmetry potential, often does develop when p. 316. With permission.) 19-7 and subtraction of Equation 19-8 gives
in each of the gel layers contacting the two this experiment is performed. Moreover, the
solutions. If the two gel surfaces have the same asymmetry potential associated with a given E. = VI - Vz (19-13)
number of sites from which protons can leave, glass electrode changes slowly with time.
then the two constants il and iz will be iden- The causes for the asymmetry potential are tions. Thus, at pH 12, the alkaline error for _. . + RT I (al + K ••bl)az
-It-jz -F n (' b')a
tical; so also will be the two activities a'i and obscure; they undoubtedly include such fac- Corning 015 glass is about -0.7 pH when the al + I Z
az in the gel layers, provided that all original tors as differences in strains established within sodium ion concentration is 1M (Figure 19-4)
sodium ions on the surface have been replaced and only about -0.3 pH in solutions that are If the number of sites on each side of the
the two surfaces .during manufacture of the
by protons (that is. msofar as the equilibrium membrane, mechanical and chemical attack of O.IM in this ion. membrane is the same, the activity of hydrogen
shown in Equation 19-3 lies far to the right~ All singly charged cations cause alkaline ions on the internal surface is approximately
the surfaces, and contamination of the outer
errors; the mapitude of the error varies ac- equal to the sum of the activities of the two
With these equalities, substitution of Equations face during use. The effect of the asymmetry
cording to the kind of metallic ion and the cations 011 the external surface; that is, Qz ~
19-7 and 19-8 into 19-6 yields potential on a pH measurement is eliminated
composition of the glass. (at + b1). Further, it ~h; thus, Equation
by frequent calibration of the electrode against
RT al The alkaline error can be satisfactorily ex- 19-13 reduces to
E. = VI - Vz = -In - (19-9) a standard buffer of known pH.
F az The AIk.1ine Error. Some glass membranes plained by assuming an exchange equilibrium
respond not only to changes in hydrogen ion between the hydrogen ions on the surface of _ RT (al + Knbd
E.-- I n---~
Thus, provided the two gel surfaces are identi- concentration but also to the concentration of the glass and the cations in the solution. This F az
cal, the boundary potential E. depe!1ds only alkali-metal ions in solutions of pH 9 or process, which is the reverse of the one de- Finally, where az is constant,
upon the activities of the hydrogen i0n in the greater. The magnitude of the resulting error scribed by Equation 19-3, can be formulated as
solutions on either side of the membrane. If for four types of glass membranes is indicated RT
one of these activities az is kept tonstant, H+OI- + B+ ~B.+G1- + H+ E. = constant +F In(a, + Knbd (19-14)
on the right-hand side of Figure 19-4. In each
then Equation 19-9 further simplifies' to gel soln gel soln
of these curves, the sodium ion concentration
was niaintained at 1 M, and the pH was varied. where B+ represents a singly charged cation, It has been shown that not only the bound-
RT
E. = constant + Final (19-10) Note that the pH error is negative at high pH, such as sodium ion. The equilibrium constant ary pot~tial but also the diffusion potential
which suggests that the electrode is responding for this reaction is on one side of the membrane is changed when
to sodium ions as well as to protons. This some of the surface sites are occupied by cat-
observation is confirmed by data obtained for ions other than hydrogen. Under such cir-
solutions with different sodiuin ion concentra- cumstances, the difTusi,.n potentials do not
effect of glass composition on the magnitude -1.000
subtract out as they did previously (see p. 5441 at lower hydrogen ion c:oncentrations (pH > 6~
and E. does Dot adequately describe the total of this error. One consequence of this work -900 Thus, the surface has less affinity for lithium
emf that develops in the presence of another has been the development of glasses for which than for sodium ion.
-800
cation. It has also been shown that this effect the term K••(UB/UH)b1 in Equation 19-15 is Li+ , Studies such as these have led to the develop-
can be accounted for by modifying Equation small enough so that the alkaline error is neg- > -700 N.' ment of glass membranes suitable for direct
E
19-14 as follows9: ligible below a pH of about 12 Other studies ~~ --6Dll potentiometric measurement of the coneenira-
have been directed toward finding glass com- ..
E = constant + R: In [al + K •• (~:)bl] positions in which this term is greatly en-
hanced, in the interests of developing glass
0
..
C>

g
-500

-400
tions of such singly charged species as Na +,
K+, NHt, Rb+, Cs+, Li+, and Ag+. Some of
these glasses are reasonably selective, as may
(19-15) electrodes for the determination of cations j
w
be seen in Table 19-1. Because of the strong
-300
where U Band U H are measures of the mobili- other than hydrogen. This application requires bond between multiply charged cations and
that the hydrogen ion activity al in Equation u+
ties of B+ and H + in the gel. -200
1.7%AI203 glass, little hope exists for the development
For many glasses, the term K••(UB/UH)b1 19-15 be negligible with respect to the second K'
of glass membranes for determination of these
N.'
is small with respect to the hydrogen ion activ- term containing the activity bl of the other species.
ity a•• as long as the pH of the solution is cation;'under these circumstances, the poten-
less than about 9. Under these conditions, tial of the electrode would be independent of
pH
Equation 19-15 simplifies to Equation 19-10. pH but would vary with pB in the typical way. Liquid Membrane Electrode.
With high concentrations of singly charged A number of investigators have demon- FIGURE 19-5 Responseoftwoglassmem- Liquid membranes owe their response to the
cations and at higher pH levels, however, this strated that the presence of Al203 or B203 in branes in the presence of alkali-metal ions potential that is established across the inter-
second term plays an increasingly important glass causes the desired effect Eisenman and (0.1 M in each case). (Curves taken from G. face between the solution to be analyzed and
part in determining E. The composition of the co-workers have carried out a systematic study Eisenman, in Advances in Analytical Chemistry an immiscible liquid that selectively bonds with
glass membrane determines the magnitude of of glasses containing Na20, A1203, and Si02 and Instrumentation, ed. C. N. Reilley. New the ion being determined. Liquid membrane
K ••(UB /UH)bI; this parameter is relatively in various proportions; they have demon- York: John Wiley & Sons, Inc., 1965, vol. 4, electrodes are particularly important because
small for glasses that have been designed for strated clearly that it is practical to prepare p. 219. With permission.) they permit the direct potentiometric deter-
work in strongly basic solutions. The mem- membranes for the selective measurement of mination of the activities of several polyvalent
brane of the Beckman Type E glass electrode, several cations in the presence of others.IO cations and certain anions as well.
illustrated in Figure 19-4, is of this type. Glass electrodes for potassium ion and sodium are clearly important at low pH values, but at A liquid membrane electrode differs from a
'I1Ie Acid Error. As shown in Figure 19-4, ion are now available commercially. levels above about pH 5, the potential becomes glass electrode only in that the solution of
the typical glass electrode exhibits an error, Figure 19-5 illustrates how two types of independent of pH. Thus, al <C K..(Ua/UH)b1 known and fixed activity is separated from
opposite in sign to the alkaline error, in solu- glass electrodes respond to pH changes in so- in this region, with the result that the potential the analyte solution by a thin layer of an im-
tions of pH less than about 0.5; pH readings lutions that are O.1F in various alkali.metal can be expected to vary linearly with pNa, pI(, miscible organic liquid instead of a thin glass
tend to be too high in this region. The magni- ions. The one electrode was fabricated from or pLio Experimentally, this prediction is veri- membrane. As shown schematically in Figure
tude of the error depends upon a variety of Coming 015 glass, a composition that con- fied; such a glass can be employed to measure 19-6 a porous, hydrophobic (that is, water-
factors and is generally not very reproducible. tains no A1203; at pH levels below 9, the the concentration of these ions. repelling1 plastic disk serves to hold the or-
The causes of the acid error are not well term K ••(Ua/UH)b1 in Equation 19-15 is small Sources for the difference in behavior be- ganic layer between the two aqueous solutions.
understood. with respect to al' At higher pH values, the tween the two glasses shown in Figure 19-5 Wick action causes the pores of the disk or
second term becomes important; note that its (toward Na+, for example) include the rela- membrane to stay filled with the organic liquid
Glass Electrodes for the magnitude also depends upon the type of alkali tive affinity of anionic surface sites of the gel from the reservoir in the outer of the two con-
Determination of Other Cations ion present. for hydrogen ions as compared to sodium ions, centric tubes. The inner tube contains an
With the glass containing A1203, both and (to a lesser extent) the mobility ratios of aqueous standard solution of MCI2, where
The existence of the alkaline error in early potential-determining terms in Equation 19-15 these two ions in the two glasses. Above pH 4, M2+ is the cation whose activity is to be deter-
glass electrodes led to studies concerning the where glass containing Al203 is no longer sen- mined. This solution is also saturated with
sitive to hydrogen ions, these anionic sites are Agel to form a Ag/AgCl reference electiode
• See: G. EisenmaJI, Glass Elecuo4u for Hydrogen IJIIIl •0 For a summary of this work, see: G. EiIemban, in
occupied entirely by the metallic ion, and the with the silver lead wire .
Ollter Catioll.s. New York: Marcel Dekker. 1967. Chap- Ad_ In AlIGlyticalClaemUtry IJIIIl Irull'lllllDlt4tiorl, ed. electrode behaves as a perfect sodium elec- The organic liquid is a nonvolatile, water-
ters 4 and 5. C. E. Reilley. New York: Wiley, 1965. vol. 4. p. 113. , trode. With lithium ion, saturation occurs only immiscible, organic ion exchanger that COD-
tains acidic, basic, or chelating functional Repeated treatment converts the exchange liq-
groups. Between this liquid and an aqueous uid essentially completely to the cationic form, glass. el~rode. The ion exchanger is an ali-
solution containing a divalent cation, there is RM,,; it is this form that is employed in elec- phatic diester of phosphoric acid dissolved in
.established an equilibrium that can be repre- trodes for the determination of MH . a polar solvent. The chain lengths of the ali-
sented as In order to determine the pM of a solu- phatic groups in the ester range from 8 to 16
tion, the electrode shown in Figure 19-6 is ~~n atoms. The diester contains a single
8CldiCproton; thtts, two molecules are required
RH2" + xMH ~ RM" + 2xH+ immersed in a solution of the analyte which Liquid ion
organic aqueous organic aqueous also contains a reference electrode-usually a exchanger ~o bond a ~valent cation (here, calcium). It
(organic)
phase phase phase phase saturated calomel electrode. The potential be- IS ~e se~ective affinity of this compound for
~IClum IOnsthat imparts the selective proper-
Aqueous solution tIes ~o t~e electrode. The internal aqueous
sat'd Agel + Mel.
So!utlon m contact with the exchanger con-
[M'+) E ••

tams a fixed concentration of calcium chloride'


~ silver/silver chloride reference electrode i~
Imme~ in th!s solution. The porous disk
PriDci~ Cation Glass contammg the lon-exchange liquid separates
to Be eastnd Composition Selectivity CbaracteJisticsb Remarks Porous plmic membrane
holding liquid ion the analyte solution from the reference calcium
a.eIlInger chloride solution. The equilibrium established
Li+ ISLi20 KU+,Na+ l:::: 3, KU+.K+ > 1000 Best for Li+ in
2SAl203 presence of H+ FIGURE 19-6 Liquid membrane elec- at each interface can be represented as
6OSi02 and Na+ trode sensitive to MH. [(ROhPOO]2Ca~2(RO)2POO- + CaH
Na+ llNa20 KNa+,K+ l:::: 2800 at pH 11 Nemst type of organic organic aqueous
18Al203 KNa+,K+ l:::: 300 at pH 7 response to
-lO-'MNa+ tween the external and the internal reference ~ote the similarity of this equation to Equa-
71Si02
electrodes is proportional to the pM of the tion 19-4 for the glass electrode. The potential
10.4Li2O KNa+, K+ l:::: 10' Highly Na+ of this electrode is given by an equation anal-
analyte.
22.6AI203 selective, but very ogous to Equations 19-2 and 19-10 for the
time-dependent Figure 19-7 shows construction details of a
67Si02 glass electrode; that is,
commercial liquid-membrane electrode that is
K+ 27Na20 KK+,Na+ l:::: 20 Nemst type of selective for calcium ion and compares the
SAl203 response to structural features of this electrode with a - L + -2-
E- 0.0591 log Q.
(19-16)
68Si02 < 10-4MK+
Ag+ 28.8Na20 KAs+,H+ l:::: 10' Highly sensitive
19.1A1203 and selective to
S2.lSi02 Ag +, but poor
stability
llNa20 KAs+,Na+ > 1000 Less selective
18Al203 for Ag+ but Ion·
71Si02 more reliable exchanger
reservoir
Liquid ion-
• Reprinted, with permilsion of the copyright owner. the American Chemical Society, from the June 12, 1967 ilsue of exchange
layer
Chemical and EII(/inurill(/ News. p. 149.
• The constant K1• J. the ICleclivity ratio. is a measure of the relative responIC of the electrode to the two species i and j. Porous membrane
For example, for the Iodium electrode, K ••••.••• iDdic:ates that the electrode is 2800 times U lCDIitive to &odium ion U to saturated with
ion exchanger
potUIium. Thill, 2800 times the COIICetItration of potauium ion is required to produce the _ potential u a unit
concentration of IIodium ion. Another method for expreuing IClectivity is to define a conatanl which is the nx:iprocaI of
4
the one jUlt described; that is, K •••••••• - 1/2800 - 3.6 x 10- • Here one can lay that a 3.6 x 10-4M IOlution of &odium
ion will produce the _ efI'ect u a 1M lO1ution of potusium ions. Unfortunately. both typea of CODltantl are fouod in FIGURE 19-7 Comparison ofa liquid membrane calcium
the literature (often without definition). Thill, the data in Table 19-2 employ a constant which is the reciprocal of the one ion electrode with a glass electrode. (Courtesy of Orion Re-
employed in this table. .
search, Inc., Cambridge, MA.)
where al here is the activity of Ca2+ and L is Table 19-2 lists commercially available liq- which are selective for Pb2+ Ca2+ and Cu2+·
a constant whose value is determined by the uid membrane electrodes. The anion-sensitive in these, silver ion serves to t~spo~ electricit;
measurement of E when the electrode is im- electrodes employ a solution of an anion ex- within the solid membrane. Figure 19-8 is a schematic diagram of a.so-
mersed in a standard solution. change resin in an organic solvent. Recently, A solid-state electrode selective for fluoride called gas-sensing electrode, which consists of
The sensitivity of the electrode just described liquid membrane electrodes for Ca2+. K +, ion is also available from commercial sources. a reference electrode. a specific ion electrOde,
is reported to be SOtimes greater for calcium NO;. and BF; have become available in The membrane consists of a single crystal of and an electrolyte solution housed in a cylin-
ion than for magnesium ion. and 1000 times which the liquid is held in the form of a poly- lanthanum fluoride that has been doped with drical plastic tube. A thin, replaceable, gas_
greater than that for sodium or potassium ions. vinylchloride gel. These electrodes have the europium(II) to increase its electrical conduc- permeable membrane is attached to one end
Calcium ion activities as low as 5 x 1O-7M appearance of a solid-state electrode described tivity. The membrane, supported between a of the tube and serves to separate the internal
can be measured. The performance of the elec- in the next section. reference solution and the solution to be electrolyte solution from the analyte solution.
trode is said to be independent of pH in the measured, shows the theoretical response to The membrane is described by its manufac-
range between 5.5 and 11. A~ lower pH levels. changes in fluoride ion activity from 10° to turer as being a thin microporous film fab-
hydrogen ions undoubtedly replace some of the 1O-6M; that is, ricated from a hydrophobic plastic; water
calcium ions on the exchanger to a significant and electrolytes are prevented from entering
Considerable work has been devoted to the E = K + 0.0591 log aF-
and passing through the pores of the film,
extent; the electrode then be<:omes pH- as well
as pCa-sensitive. development of solid membranes that are selec- The electrode is reported to be selective for owing to its water-repellent properties. Thus,
The calcium ion membrane electrode has tive towards anions in the way that some fluoride over other common anions by several the pores contain only air or other gases to
proved to be a valuable tool for physiological glasses behave toward specific cations. As we orders of magnitude. Only hydroxide ion ap- which the membrane is exposed. When a so-
studies because this ion plays important roles have seen, the selectivity of a glass membrane pears to offer serious interference. lution containing a gaseous analyte such as car-
in nerve conduction, bone formation, muscle is due to the presence of anionic sites on its Table 19-3 lists some of the commercially bon dioxide is in contact with the membrane,
contraction. cardiac condu<ltion and contrac- surface that show particular affinity toward available solid-state electrodes. the CO2 distills into the pores, as shown by
tion, and renal tubular function. At least some certain positively charged ions. By analogy, a
of these processes are influenced more by membrane having similar cationic sites might
calcium ion activity than calcium concentra- be expected to respond selectively toward
tion; activity. of course, is the parameter mea- anions. To exploit this possibility, attempts
sured by the electrode. have been made to prepare membranes of
Another liquid specific-ion electrode of great salts containing the anion of interest and a Concentration Preferred
value for physiological studies is that for potas- cation that selectively precipitates that anion Analyte loa Range,M pH Selectivity Constant for Interferences·
sium, because the transport of nerve signals from aqueous solutions; for example. barium Ca2+ 10°-5 x 10-7 6-8 Zn2+, 3.2; Fe2+. 0.8; Pb2+, 0.63; Mg2+, 0.014
appears to involve movement of this ion across sulfate has been proposed for sulfate ion and CI- 10°-8 x 10-6 2-11 1-,17; NO;. 4.2; Br-, 1.6; HCO;, 0.19;
nerve membranes. Study of the process requires silver halides for the various halide ions. The
SOi -,0.14; F-. 0.10
an electrode that can detect small concentra- problem encountered in this approach has
tions of potassium ion in the presence of much been in finding methods for fabricating mem- BF; 10-1-3 x 10-6 3-10 NO;, 0.1; Br-. 0.04; OAc-, 0.004; HCO;,
larger concentrations of sodium. A number of branes from the desired salts that possess ade- 0.004; CI-, 0.001
liquid membrane electrodes show promise of quate physical strength. conductivity. and re- NO; 10°_7 x 10-6 3-10 1-,20; Br-. 0.1; NOi, 0.04; Cl-, 0.004; SO~-,
meeting these needs; one is based upon the sistance to abrasion and corrosion. 0.OOOO3;COi- ,0.0002;CIOi, 1000; F-, 0.00006
antibiotic valinomycin, a cyclic ether that has Membranes prepared from cast pellets of
CIO; 10°-2 x 16-6 3-10 1-,0.012; NO;, 0.0015; Br-, 0.00056; F-,
a strong affinity for potassium ion. Of equal silver halides have been successfully used in
0.00025; Cl-, 0.00022
importance is the observl\tion that a liquid electrodes for the selective determination of
membrane consisting of Valinomycin in di- K+ 10°_10-6 3-10 Cs+, 1.0; NHt, 0.03; H+, 0.01; Ag+, 0.001;
chloride. bromide, and iodide ions. An elec-
phenyl ether is about Ht times as responsive. trode employing a polycrystalline Ag2S mem- Na +,0.002; Li+, 0.001
to potassium ion as to sodium ion. 11 brane is offered by one manufacturer for the Ca2+ +Mg2+ 10-2-6 X 10-6 5-8 Zn2+, 3.5; Fe2+, 3.5; Cu2+, 3.1; Ni2+, 1.35;
determination of sulfide ion. Silver ions are Ba2+, 0.94; Na+, O.o1S
sufficiently mobile to conduct electricity
II M. S. Frant and J. W. Ross, Jr. Scierru. 167, 987 through the membrane. Mixtures of PbS. CdS. • AnalytlCDl Methods Guide, 9th cd. Cambridge, MA.: Orion Research, Ine., December, 1978.With permission.
(1970~ and CuS with silver sulfide provide membranes • Note that these selectivity CODStantsare the reciprocals ofthe ones in Table 19-1 (see footnote b of Table 19-1~
As a consequence of the two reactions, the
external solution rapidly (in a few seconds to
C02(aq)~ CO2 (g) minutes) equilibrates with the film of internal
external membrane solution adjacent to the mem~rane. Here, an-
solution pores other equilibrium is establishe4 that causes the
Because the pores are numerous, a state of pH of the internal surface film to change;
equilibrium is rapidly approached. The COz namely,
in the pores, however, is also in contact with
the internal solution and a second equilib- COz(aq) + 2H20~HCO; + H30+
rium reaction can readily take place; that is, A glass-reference electrode pair immersed in
CO2.<g) ~C02(aq) the film of internal solution (see Figure 19-8)
membrane internal then detects the pH change.
pores solution The overall reaction for the process just

External solution
-containing dissolved
ua-us lnalyte

Analyte Concentration Preferred


Ion Ruge,M pH Interferences
Specific
Br- 10°-5 x 10-6 2-12 Max: [SZ-] = 10-7 M; [1-] = 2 x 1O-4[Br-] ion
electrode
Cd2+ lO-t_IO-7 3-7 Max: [Ag+], [Hg2+], [Cu2+] = 10-7 M
Cl- 10°-5 x 10-5 2-11 Max: [S2-] = 10-7 M; traces of Br-,'-, CN- do not
interfere FIGURE 19-8 Schematic diagram of a gas-sensing elec-
Cu2+ 10°_10-8 3-7 Max: [SZ-], [Ag+], [Hg2+] = 10-7 M; high levels of trode. Specific ion electrode is shown as a glass electrode.
Cl-, Br-, Fe3+, Cd2+ interfere The reference electrode is a Ag-AgCl electrode. Other elec-
CN- 10-2-10-6 11-13 Max: [S2-] = 10-7 M; [.-] = 0.1 [CN-]; trode combinations are possible.
[Br-] = 5 x 103[CN-]; [Cn = lQ6[CN-]
F- 10°_10-6 5-8 Max: [OW] = 0.1[F-]
1- 10°-5 x 10-8 3-12 Max: [S2-] = 10-7 M described is obtained by adding the three chem-
Max: [Ag+], [Hg2+], [Cu2+] = 10-7 M; high levels of
[H30+] K K (19-17)
Pbz+ 10°-10-6 4-7 ical equations to give
[COz(aq)]... [HCO;] = ,
Cd2+ and Fe3+ interfere
COz(aq) + 2H20~H30+ + HCO; which may be rewritten as
Ag+ 10°-10-7 2-9 Max: [Hg2+] = 10-7 M external internal solution
S2- 10°-10-7 13-14 Max: [Hg2+] = 10-7 M solution at = [H30+] = K,[C02(aq)]u. (19-18)
Na+ 10°_10-6 9-10 Selectivity constants: Li+, 0.002; K+, 0.001; Rb+, The eqUilibrium constant for the reaction is where at is the internal hydrogen ion activity.
0.‫סס‬OO3;Cs+, 0.0015; NHt, 0.‫סס‬OO3;11+,0.0002; given by The potential of the electrode system in
Ag+, 350; H+, 100 ! the internal solution is dependent upon at as
SCN- 10°-5 x 10-6 Max: [OW] = [SCN-]; [Br-] = 0.003[SCN-]; _[H_30_+_U_H_CO_3_-]
=K described by Equation 19-2. Substitution of
2-12
[Cl-] = 2O[SCN-]; [NH3] = 0.13(SCN-]; [C02(aq)] •••· Equation 19-18 into 19-2 yields
[S20~-] = O.Ol[SCN-]; [CN-] = 0.007[SCN-]; If the concentration of HCO; in the internal E=L + 0.0591 log K,[C02(aq)] •••
[1-1 [S2-] = 10-7 M solution is made relatively high so that its
concentration is not altered significantly by
the COz which distills, then
Direct-Reading Instruments can then be realized, depending upon amplifier
changes in potential. More' important is the
Numerous direct-reading pH meters are avail- noise leveL It should be appreciated, however,
l. = L + 0.0591 log K, dependency of cell potentials on current,
able commercially. Generally, these are solid- that it is seldom, if ever, possible to measure
which: arises from the effects of I R drop and
Thus, the potential of the cell consisting of state devices employing a field e~ect transistor pH with this kind of accuracy; indeed, uncer-
polarization phenomena (p. 527). ~e int1u~ce
the internal reference and indicator electrode or a voltage follower as the first amplifier stage tainties of ±0.02 to ±0.03 pH unit are more
of IR· drop is particularly Significant WIth typical (see p. 557).
is determined by the CO2 concentration of the specific ion electrodes, which may have resis- in order to provide the needed high internal
external solution. Note that no electrode comes resistance. As shown in Figure 19-9, the circuit The price of pH meters ranges from less
tancesof 100 MO or more. With these, currents
directly in contact with the analyte solution; can be relatively simple. Here, the output of than $100 to $1000 or more.
must be limited to 10-12 A or smaller; this
for this reason, it would be better to call the limitation requires that the potential measur- the ion-selective electrode is connected to the
device a gas-sensing cell rather than a gas- ing device have an internal resistance of 10110 inverting terminal of an operational amplifier,
sensing electrode. Note also that the only spe- which employs a field effect transistor for the DIRECT
or more (see example, p. 31).
cies that will interfere· with the measurement first stage of amplification. Note that the cell POTENTIOMETRIC
Two types of instruments for voltage mea-
are dissolved gases that can pass through the is incorporated in the feedback loop. The am- MEASUREMENTS
surement are employed in potentiometry-the
membrane and can additionally affect the pH plifier current then drives a meter that is usually
potentiometer and the electronic voltmeter.
of the internal solution. calibrated in terms of bot4 pH and millivolts. Direct potentiometric measurements can be
Both instruments are referred to as pH meters
The possibility exists for increasing the selec- when their internal resistances are sufficiently The variable resistance RT is calibrated in used to complete chemical analyses of those
tivity of the gas-sensing electrode by employing terms of temperature so that correction for species for which an indicator electrode is
high to be used with glass and other mem-
an internal electrode that is sensitive to some temperature variations is possible. available. The technique is simple, requiring
brane electrodes; with the advent ofthe many
species other than hydrogen ion; for example, The readout of commercial pH meters is only a comparison of the potential developed
new specific ion electrodes, pIon or ion meters
a nitrate-sensing electrode could be used to either a digital meter or a: meter with a 5- to by the indicator electrode in the test solution
would perhapS be a more descriptive name.
provide a cell that would be sensitive to nitro- lo-in. scale that covers a range of 0 to 14 pH with its potential when immersed in a stand-
gen dioxide. Here, the equilibrium would be units. Many of the latter instruments are also ard solution of the analyte; insofar as the
Potentiometers equipped with scale expansion capabilities, response of the electrode is specific for the
2N02(aq) + H20 ~ NOi" + NO; + 2H+ which provide full-scale ranges of 0.5 to 2 pH analyte. no preliminary separation steps are
external internal solution The potentiometer described in Chapter 2
units; a precision of ±0.001 to 0.005 pH unit required. In addition, direct potentiometric
(p. 33) is the classical instrument for poten-
solution measurements are readily adapted to the con-
tial measurement and has, in the past, been
tinuo!!s and automatic monitoring of analyti-
This electrode should permit the determina- widely used in potentiometric studies. It is un-
cal parameters.
tion of N02 in the presence of gases such as satisfactory for measurements with membrane
Notwithstanding these attractive advan-
802, CO2, and NH3 which would also alter electrodes. It can, however, be adapted to such
tages, the user of direct potentiometric mea-
the pH of the internal solution. measurements by replacing the relatively low-
surements must be alert to limitations that are
Gas-sensing electrode systems are commer- resistance galvanometer with a current de-
inherent to the method. An important example
cially available for CO2, N02, and NH3• An tector having a high-input resistance. In the
is the existence in most potentiometric mea-
oxygen sensitive cell system is also on the mar- past, vacuum tube electrometers have been
surements of a liqUidjunction potential (see
ket; it, however, is based on a voltammetric used for this purpose. Alternatively, the out-
p. 526). For most electroanalytical methods,
measurement and is discussed in Chapter 21. of-balance current of the potentiometer can be
this junction potential is inconsequential and
fed into a high-resistance voltage-follower am-
can be neglected. Unfortunately. however, its
plifier and then amplified by means of a suitable
existence places a limitation on the accuracy
operational amplifier (see Figure 3-20, p. 60).
that can be attained from a direct potentio-
INSTRUMENTS FOR CELL An instrument of this type with a high- ·c metric measurement.
POTENTIAL MEASUREMENT precision slide wire is sensitive to better than 100-
±0.1 mY, which corresponds to ±0.002 pH. 50-
D- Rr
An instrument for potentiometric measure- Potentiometric pH meters appear to be dis- • Equation for Direct Potentiometry
ments should draw essentially no electricity appearing from the laboratory, being replaced
The observed potential of a cell employed for
from the galvanic cell being studied. One by more convenient and rapid direct-reading FIGURE 19-9 Simplified circuit a direct potentiometric measurement can be
reason for this is that a current causes electronic voltmeters or pH meters. diagram for a direct-reading pH meter. expressed in terms of the potentials developed
PRINCIPLES OF INSTRUMENTAL ANALYSIS

The difference between activity and con- the hydrogen ion concentration will differ ap-
Several methods for performing analyses by
by the indicator electrode, t~e refer~ce elec- centration is illustrated by Figure 19-10, where preciably from the activity measured.
direct potentiometry have been developed; .all
trode, and a junction potential. That IS the lower curve gives the change in potential Inherent Error in the Electrode Calibration
are based, directly or indirectly, upon Equation
of a calcium electrode as Ii function of calcium Procedure. A st;rious disadvantage of the elec-
Eoba = Ere( - E1nd + Ej {19-19)12 19-22,13
chloride concentration (note that the activity trode calibration method is the existence of
Typically, the junction potential Ej has two or concentration scale is logarithmic). The an inherent uncertainty that results from the
Electrode Calibration Method
components, the first being located at the nonlinearity of the curve is due to the increase assumption that K in Equation 19-22 remains
junction of the analyte solution and one end In the electrode calibration procedure, K. in in ionic strength-and the consequent decrease constant after calibration. This assumption
of a salt bridge and the second where the Equation 19-22 is determined by measunng in the activity coefficient of the calcium ion-as can seldom, if ever, be exactly true because
bridge interfaces with the reference electrode E for one or more standard solutions of the electrolyte concentration becomes larger. the electrolyte composition of the unknown
solution. The two potentials tend to cancel one k:wn pM. The assumption is then made When these concentrations are converted to will almost inevitably differ from that of the
another, but seldom do so completely. Thus, that K does not change during measurement activities, the upper curve is obtained; note solution employed for calibration. The junc-
Ej in Equation 19-19 may be as large as 1 mV of the analyte solution. Genera1\y, the. calibra- that this straight line has the theoretical slope tion potential contained in K wi1l vary slightly
or more. tion operation is performed a~ the tune. that of 0.0296 (0.0591/2). as a consequence, even when a salt bridge is
Ideally, the potential of the indicator elec- pM for the unknown is determmed;recalibra- Activity coefficients for singly charged ions used. Often, this uncertainty wi1l be of the
trode is related to the activity at of M"+ , the tion may be required if measurements extend are less affected by changes in ionic strength order of 1 mV or more; it is readily shown
ion of interest, by a Nemst-type equation over several hours. than are coefficients for species with multiple (see Problems 17 and 18 at the end of this
The electrode calibration method offers the charges. Thus, the effect shown in Figure 19-10 chapter) that the relative error in activity or
0.0591 . 0.0591
E1nd = L + -n-log at = L - -n- pM advantages of simplicity, speed, and applica- will be less pronounced for electrodes that concentration associated with a 1 mV uncer-
bility to the continuous monitoring ~f pM. respond to H +, Na +, and other univalent ions. tainty in K is about 4 % when n in Equation
(19-20)
Two important disadvantages attend its use, In potentiometric pH measurements, the 19-22 is I, and 8% when it is 2. It is important
where L is a constant. For metallic electrodes, however. One of these is that the results of an pH of the standard buffer employed for cali- to appreciate that this uncertainty is charac-
L is often the standard potential for the indi- analysis are in terms of activities rather than bration is generally based on the activity of teristic of all measurements involving cells that
cator electrode; with membrane electrodes, concentrations; the other is that the accuracy hydronium ions. Thus, the resulting hydrogen contain a salt bridge and that this error cannot
however, L may also include an unknown, of a measurement obtained by this procedure ion analysis is also on an activity scale. If the be eliminated by even the most careful measure-
time-dependent asymmetry potential (p. 544). is limited by the inherent uncertainty ca~ unknown sample has a high ionic strength, ments of cell potentials or the most sensitive
Combination of Equation 19-20 with Equa- by the junction potential; unfortu~a~ely, thIS measuring devices; nor does it appear possible
tion 19-19 and rearrangement yields uncertainty can never be totally elunmated. to devise a method for completely eliminating
Activity venus Coaceatradoa. Electrode re- the uncertainty in K that is the source of this
pM= -log at sponse is related to activity rather than to ana- error.
Eoba - (E,er + Ej - L) (19-21) Iyte concentration. Ordinarily, however, the
0.0591/n scientist is interested in concentration, and the
determination of this quantity from a poten- Calibration Curves
Eoba - K tiometric measurement requires activity coef- for Direct Potentiometry
~ +20
0.0591/n
The new constant K consists of three con-
ficient data. More often than not, activity co-
efficients will be unavailable because the ionic I An obvious way of correcting potentiometric
measurements to give results in terms of con-
strength of the solution is either unknown or ~ 0 centration is to make use of an empirical cali-
stants, of which at least one (Ej) has a magni-
tude that cannot be evaluated from theory. so high that the Debye-Hiickel equation is not ~ bration such as the lower curve in Figure 19-10.
W -20 For this approach to be successful, however,
Thus K must be determined experimentally applicable. Unfortunately, the assumption that
with'the aid of a standard solution of M activity and concentration are identical may it is essential that the ionic composition of the
before Equation 19-22 can be used for the lead to serious errors, particularly when the standard closely approximate that of the ana-
analyte is polyvalent. lyte-a condition that is difficult to realize
measurement of pM.
FIGURE 19·10 Responseofacalciumion experimentally for complex samples.
electrode to variations in the calcium ion Where electrolyte concentrations are not
U A$ written, Equation 19-19bas the reference electrode
U For additional information, see: R. A. Durst, Ion- concentration and activity of solutions pre- too great, it is often helpful to swamp both
acting as cathode and the indicator electrode acting as Select~ Electrodes. National Bureau of Standards pared from pure calcium chloride. (Courtesy the samples and the calibration standards
anode. U. in a particular ceo, the roles arc reversed, the Special Publication 314.Washington, D.C.: u.s. Govern- of Orion Research, Inc., Cambridge, MA.) with a measured excess of an inert electrolyte.
signs of the two electrodes arc likewise reversed. ment Printing Office. 1969.Chapter 11.
teins, and gases; the pH of viscous or even attainment of steady readings.
This equation rearranges to
Under these circumstances, the added eff~ . semisolid fluids can be determined. Electrodes 50 Variation in junction potentiaL It should be
of the electrolyte in the sample becomes neglI- C. V. (19-23). for special applications are available. Included reemphasized that this is a fundamental
gible,and theempiricalcalibrati~n curve.yields C" = (V" + V.)1O-·&E/O.OS91 - v,. among these are small electrodes for pH uncertainty in the measurement of pH, for
results in terms of concentration. This ~p- measurements in a drop (or less) of solution which a correction cannot be applied. Abso-
proach has been employed for. the .potent1~ Thus, C" is readily calculated from the con-
or in a cavity of a tooth, microelectrOdes which lute values more reliable than 0.01 pH unit
metric determination of fluonde m public centration of the standard, the two volumes,
permit the measurement of pH inside a living are generally unobtainable. Even reliability
water supplies. Here, both s~mp~es and s~- and the potential difference tJ:.
celL systems for insertion in a flowing liquid to 0.03 pH unit requires considerable care.
dards are diluted on a I : I basis WIth a solutIOn The standard addition method has been
stream to provide a continuous monitoring of On the other hand, it is often possible to
containing sodium chloride, a citrate buffer, applied to the determination of chloride and
pH, and a smali glass electrode that can be detect pH differences between similar solu-
and an acetate buffer; the diluent is sufficiently fluoride in samples of commercial phosphors.14
swallowed to indicate the acidity of the stom- tions or pH changes in a single solution
concentrated so that the samples and standards Here, two solid-state indicator electrodes and
ach contents (the calomel electrode is kept in that are as small as 0.001 unit. For this
do not differ significantly in ionic strength. a reference electrode were used; the added
the mouth). reason, many pH meters are designed to
The procedure permits a rapid measure~ent standard contained known quantities of the
Summary of Errors Affecting pH Measure- permit readings to less than 0.01 pH unit.
of fluoride ion in the I-ppm range, WIth a two anions. The relative standard deviation for
ments with the Glass Electrode. The ubiquity 6. Error in the pH of the standard buffer. Any
precision of about 5 % relative. the measurement of replicate samples was
. of the pH meter and the general applicability inaccuraci~ in the preparation of the buffer
found to be 0.7% for fluoride and 0.4 % for
of the glass electrode tend to lull the chemist used for calibration, or changes in its com-
Standard Addition Method chloride. When the standard addition method
into the attitude that any measurement ob- position during storage, will be propagated
was not used, relative errors for the analyses
In the standard addition method, the poten- tained with such an instrument is surely as errors in pH measurements. A common
appeared to range between 1 and 2 %.
tial of the electrOde system is measured before correct. It is well to guard against this false cause of deterioration is the action of bac-
and after addition of a small volume of a sense of security since there are distinct limi- teria on organic components of buffers.
standard to a known volume of the sample. tations to the electrode system. These have
The assumption is made that this addition Potentiometric been discussed in earlier sections and include
does not alter the ionic strength and thus the pH Measurements the following:
activity coefficient f of the analyte. It is fur- with a Glass Electrode' 5

ther assumed that the added standard does 1. The alkaline error. The ordinary glass elec-
The glass electrode is unquestionably the most
not significantly alter the junction potential.. trode becomes somewhat sensitive to alkali- The potential of a suitable indicator electrode
important indicator electrode f~r h~rogen
If the potentials before and after the addi- ion. It is convenient to use and IS subject to metal ions at pH values greater than 9. is conveniently employed to establish the
tion are EI and Ez, respectively, we may write, few of the interferences that affect other pH- 2. The acid error. At a pH less than 0.5, values equivalence point for a titration (a potentio-
with the aid of Equation 19-22 obtained with a glass electrode tend to be metric titration). A potentiometric titration
sensing electrodes.
Glass electrodes are available at relatively somewhat high. provides different information from a direct
E1-K
-log al = -log C,J= 0.0591/n low cost and in many shapes and sizes. A 3. Dehydration. Dehydration of the electrode potentiometric measurement. For example, the
common variety is illustrated in Figure 19-1 may cause unstable performance and errors. direct measurement of O.lOOF acetic and hy-
where al is activity of the analyte in the (p. 541); the reference electrode is usually a 4. Errors in unbuffered neutral solutions. Equi- drochloric acid solutions with a pH-sensitive
sample,fis its activity coefficient, and C" is its commercial saturated calomel electrode. librium between the electrode surface layer electrode would yield widely different pH
molar concentration. Addition of V. ml of a The glass/calomel electrode system is a re- and the solution is achieved only slowly in values because the former is only partially
standard with a molar concentration of C. to markably versatile tool for the measurement poorly buffered, approximately neutral so- dissociated. On the other hand, potentiometric
V" ml of the sample causes the potential to of pH under many conditions. ~e elect~Ode lutions. Errors will arise unless time (often titrations of equal volumes of the two acids
acquire a value E2; Equation 19-22 then can be used without interference m solutions several minutes) is allowed for this equilib- would require the same amount of standard
becomes containing strong oxidants, reductants, pro- rium to be established. In determining the base for neutralization.
C"V" + C.v.
-log [----
V" + V.
f] E2 - K
= ---
0.0591/n
pH of poorly buffered solutions, the glass
electrode should first be thoroughly rinsed
The poten(iometric end point is widely ap-
plicable and provides inherently more accurate
with water. Thea, if the unknown is plenti- data than the corresponding method employ-
Subtracting the first equation from the second ful, the electrodes should be placed in suc- ing indicators. It is particularly useful for
gives, upon rearrangement, 14 L. G. Bruton. Anal. Chml, 43, 579 (1971~ cessive portions until a constant pH reading titration of colored or turbid solutions and for
IS For a detailed discussion of potentiometric pH mea-
is obtained. Good stirring is also helpful; detecting the presence of unsuspected species
C"(V,, + V.) ] E2 - EI tJ: surements, see: R. G. Bates, Determination of pH, Theory
several minutes should be allowed for the in a solution. Unfortunately, it is more time-
log [C" V" + C. V. = 0.0591/n = 0.0591/n anti Practice. New York: Wiley, 1964.
The first two columns afTable 19-4 consist evaluated during the titration and recorded
consuming than a titration that makes use of
of typical potentiometric titration data ob- directly in lieu of the potential itself. Thus, in
an indicator.
tained with the apparatus illustrated in Figure Several methods can be used to determine the column 3 of Table 19-4, it is seen that the
Figure 19-11 shows a typical apparatus for
19-11. The data near the end point are plotted end point for a potentiometric titration. The maximum is located between 24.3 and 24.4
performing a potentiometric titrati~n. Or-
in Figure 19-12a. Note that. this experim~tal most straightforward involves, direct plot of ml; selection of 24.35 ml would be adequate
dinarily, the titration involves measunng ~
plot closely resembles titration curves denved potential versus reagent volume, as in Figure for most purposes.
recording a cell potential (or a pH ~~)
after each addition of reagent. The titrant 18 from theoretical considerations. 19-12a. The midpoint in the steeply rising With both techniques. the assumption is
added in large increments at the outset; as the portion of the curve is then estimated visually made that the titration curve is symmetric
end point is approached (as indicated by larger and taken as the end point. Various mechani- about the true equivalence point and that the
potential changes per addition~ the increments cal methods to aid in the establishment of the inflection in the curve therefore corresponds
midpoint have been proposed; it is doubtfu~ to that point. This assumption is valid, pro-
are made smaller. w
Sufficient time must be allowed for the .,
0 however, that these significantly improve the vided the participants in the chemical process
attainment of equilibrium after each addition ; 0.3
accuracy. react with one another in an equimolar ratio,
'ii
of reagent. Precipitation reactions may re- .~ A second approach is to calculate the change and also provided the electrode process is per-
quire several minutes for equilibration. p~r- i" in potential per unit change in volume of re-
agent (that is, AEIAV). as has been done in
fectly reversible. Where these provisions are
not met, an,asymmetric curve results (see curve
ticularly in the vicinity of the equivalence porot. ,

1
0.2
A close approach to equilibrium is indicated column 3 of Table 19-4. A plot of this pa· B, Figure i9-13). Note that the curve for the
when the measured potential ceases to drift by w rameter as a function of average volume leads oxidation of iron (II) by cerium(IV) is symmet-
more than a few millivolts. Good stirring is to a sharp maximum at the end point (see rical about the equivalence point. On the other
frequently effective in hastening the achieve- Figure 19-12b). Alternatively. the ratio can be hand,S moles of iron(II) are consumed by each
ment of equilibrium.
1.0

0.8 TABLE 19·4 POTENTIOMETRIC TITRATION DATA FOR


E
D"0 2.422 MILLIEQUIVALENTS OF CHLORIDE WITH 0.1000 F
:.
~
>
0.6

0.4
SILVER NITRATE-

Vol AgN03, Evs. SCE, AEIAV,


<I
0.2 ml V Vlml
0 5.0 0.062
20
15.0 0.085 ~.002
ft

004
20.0 0.107
0.123 O.008
22.0 O.015
23.0 0.138
23.50 0.146 ~.016
-v. 050
23.80 0.161
'; 2.0
24.00 0.174 ~.065
24.10 09
~ 020 22 24 0.183 ft'

-v.
24.20 0.194 11 1---- 2.8
<I -2.0 VolumeO.l000F O. 39 1 44
AgN03. ml 24.30 0.233 o.24
83 .
24.40 0.316 1----· - 5.9
24.50 0.340 2- 11 .1----- - 1.3
24.60 0.351 o.07,1---- -0.4
Ie) 24.70 0.358 o.osd
25.00 0.373 o.024
FIGURE 19·12 (a) Potentiomet- 25.5 0.385
ric titration curve for 2.433 meq of 26.0 0.396 2·022
015
Cl- with 0.1000 F AgN03• (b) First· 28.0 0.426
FIGURE 19-11 Apparatus for a potentio- derivative curve. (c) Second-derivative
metric titration. curve. • The electrode system consists or a silver indicator electrode and a saturated calomel electrode.
The volume or the analy1e solution was 50.00 ml.
562 PRINCIPLES OF INSTRUMENTAL ANALYSIS
For a given titration, all of the terms in this
GraD'. Plots. Gran 17has suggested a me~- equation are constant except V", and E•••.
-
-
od to permit evaluation of ~e ~d pomt
with data from a region of the titration ~here
Thus, a plot of the quantity (Va + V",)anti-
loge -16.9E.) against V", should be linear.
J
::I
4.0

Equivalence _~ the reaction is forced nearly to completion by In addition, at equivalence, the left-hand side I8'
pointS
the common ion effect. Such a procedure ~as of the equation will equal zero. Thus, an ex- 3.0

EqUiV'I~
point A
.•••..•.•• the advantage of requiring fewer data ?Omts
than a conventional plot. FU~~o~ It may
trapolation of the plot to zero on the antilog i
"i
axis will give the equivalence-point volume. A 2.0
. ld more accurate end pomts m mstances ::.
Gran's plot of the data in Table 19-4 is shown +
yteherethe rate of change in p function is small
-~ V w . in Figure 19-14. Note that the first four points ::.8 1.0

6 - ::::-
in the equivalence-point region.
As an example, we shall apply Gran s
method to the data shown in Table ~9-4 ~or
, lie on a good straight line. A departure from
linearity is observed, however, for the data
that lie less than about 2 ml from equivalence.
the potentiometric titration of chlo~de Ion For less complete reactions, the curvature
40 60 80 100 120 140
with silver nitrate. Short of the equivalence occurs earlier in the titration. Clearly, a satis-
Volume of 0.1 F Co'·. ml
I I I I I point in this titration, the silver electrode ~- factory end point can be obtained from 3 or 4 FIGURE 19-14 Gran's plot of the data
20 40 80 80 100 120 haves as a second-order electrode for chlon~e data points in contrast to the 19 that were in Table 19-4.
Volumo of 0.1 N KMnO,. ml ion, and we may describe the cell potential recorded in Table 19-4.
FIGURE 19-13 Curvesdepictingt~etit~a- E. by the equation When the concentration of the reagent is
tion of 100 ml of 0.100 N iron(II) solutIOn With E. = E1aa - 0.059110g{CI-] - Esa. made large with respect to the analyte,
(A) cerium(IV) and with (B) ~ganate. (Va + VA.) is equal to approximately Va metal from which the reacting cation is derived.
We have neglected the junct.ion potent~al be-
Note 'that the horiZontal axes are dISplaced to throughout the titration; the quantity antilog Membrane electrodes that are sensitive to one
cause it is of no significance m a poten!lome!-
permIt comparison of the curves. ( -16.9E.) can then serve as the ordinate. of the ions involved in the titration process
ric titration. It is convenient to rewnte thIS
Semiantilog paper is available from Orion may also be employed. Occasionally, an inert
equation in the form Research Inc.,18 and makes it possible to ob- platinum electrode can serve as the indicator.
Iog(Cn = -16.9E. + K' tain Gran's plots without computations. The Thus, for example, the titration of zinc ion with
where 16.9 is the reciprocal of 0.0591 ~nd K' vertical axis of this paper is skewed to eliminate ferrocyanide may be followed with a platinum
mole of permanganate and a highl~ n~n- the need to correct for volume change, pro- electrode, provided a quantity of ferricyanide
is equal to (E'AaCJ - Esa.)/O.0591. Takmg the
symmetric titration curve results. 0r:c:'manly, vided the titrant volume is less than 10% of ion is also present. The potential developed
the change in potential within the eqUivalence- antilogarithm of both sides of this equation
the total at the end point. by this electrode is determined by the half-
point region of these curyes is ~tly iarge leads to
TItratiOll to • Fixed PotentiaL Another pro- reaction
so that a negligible titration err~r.1Smtr~u~ [Cn = antilog( -16.9E. + K') cedure consists of titrating to a predetermined
if the midpoint of the steeply rISing portion IS Fe(CN)~- + e.~Fe(CN):-
Until the end point is reached, [Cn is given end point potential. The value chosen may be
chosen as the end point. Only when .unusual the theoretical equivalence-point potential cal- With the first excess of reagent, the ferrocya-
accuracy is required or where very dilute so- by
culated from formal potentials or an empirical nide concentration 'increases rapidly and
lutions are employed must account be tak~ VaFa - VA.F'" causes a corresponding change in the poten-
[CI-] Va + V", potential obtained by the titration of standards.
of this source of uncertainty. Under such. CIr- Such a method demands that the equivalence- tial of the electrode.
cumstances, a correction can be determmed where Va and V", are the meas.ured volum.es o! point behavior of the system be entirely re- The most widely used volumetric precipi-
empirically by titration of a standard. Alt~r- the analyte and the silver mtrate solutions, producible. tation reagent is silver nitrate, which is em-
nati~ely, the correct position of the eq~IV- F and F are the initial formal concentra- ployed for the determination of the halogens,
alence point can be calculated from theoretical ti~ns of th~two solutions. Combining the two Precipitation Titrations the halogenoids, mercaptans, sulfides, arse-
comliderations when the error is due to a non- equations and rearranging gives nates, phosphates, and oxalates. For all of
Electrode Systems. For a precipitation ti-
sym~etrical reaction.16. these, a silver indicator electrode is used. When
(VaFa - V",F A.) = (Va + V",)antilog tration, the indicator electrode is often the
dilute solutions are being titrated or where
. ( -16.9E. + K') highest precision is required, reference elec-
II Orion Resean:b IDe., 380 Putman Ave. Cambridge, trodes such as those shown in Figure 18-11
•• See: I. M. Ko1thoir and N. H. Furman, potmlollletrlc cannot be used directly because of leakage of
MA. 02139.
TitratlolU. 2d ed. New York: Wiley. 1931, Chapters 2 and
3.
chloride ion from the salt bridge. This problem . is in terms of pAg. As shown by the dotted latter effect is general and is encountered in It is reasonable to assume, however, that
can be avoided by immersing the calomel elec- lines, the experimental curve for iodide departs complex formation, neutralization, and oxida-
somewhat from the theoretical as a conse- tion-reduction titrations. [Ag+] ~ 8.40 x 10-4
trode in a concentrated potassium nitrate
solution; this solution is then connected to the quence of adsorption. Just before the equiva- It is evident from the two curves for chlo-
Therefore, the potential of the indicator elec-
analyte solution by means of an agar bridge lence point, iodide ions are held on the surface ride ion shown in Figure 19-15 that less sharp trode will be given by
that contains about 3 % potassium nitrate. For of the solid, thlls decreasing the iodide con- end points are observed as the analyte and
reagent and analyte concentrations of O.IF or centration in the solution and increasing the reagent solutions become more dilute. For EAa = - 0.151 - 0.0591 log 8.40 x 10-4
greater, a calomel electrode such as that shown potential slightly. Beyond equivalence, adsorp- chloride, titrations with solutions as dilute as = 0.031 V
in Figure 18-11a can be used directly in the tion acts to decrease the concentration of silver O.ooIF can be performed with a relative un-
analyte solution without incurring significant ions in the solution; lower potentials result. certainty of about 1%. Smaller concentrations If the calomel electrode is treated as the anode,
errors. This effect is less pronounced for bromide and of bromide and iodide can be titrated suc-
Titratioo Curves. Figure 19-15 shows theo- chloride ion. cessfully because their reactions with silver ion E.11I = 0.031 - 0.241 = -0.210 V
retically derived titration curves for chloride, Figure 19-15 demonstrates that the magni- are more complete. The example that follows
(b) At the equivalence point
bromide, and iodide ions with a silveri tude of change in electrode potential in the illustrates how these curves are derived.
saturated calomel electrode system. The or- equivalence-point region increases as the reac- [Ag+] = [r]
dinate on the left is calibrated in terms of the tion product becomes less soluble-that is, as
EXAMPLE
cell potential in volts while that on the right the reaction becomes more complete. This The solubility product for silver iodide is
An electrode system consisting of a silver.
indicator electrode and a saturated calomel KIP = [Ag+][I-] = 8.3 X 10-17
electrode was employed for the potentiomet-
ric titration of 100 ml of 0.02ooF sodium At equivalence, [Ag+] = [r]. Thus,
0.6 iodide with O.looF silver nitrate. Calculate
0.0 [1-] = J8.3 x 10- 17 = 9.1 X 10-9
the cell potential at: (a) 1.00 ml before the
equivalence point; (b) the equivalence point; Therefore,
2.0 and'(c) 1.00 ml beyond the equivalence point.
0.4
------ The required electrode potential data are EA1 = - 0.151 - 0.0591 log 9.1 x 10-9
4.0 = 0.324
W
AgI(s) + e~Ag(s) + 1- EO = -0.151 V
()
Ul Again treating the calomel cell as the anode,

-----:0 .."
Hg202(s) + 2e~2Hg(l) + 2CI- (sat'd KCI)
! 0.2 / 6.0
> CI- ____ E.11I = 0.324 - 0.241 = 0.083 V
7i' (0.005 FI • '.82)1. '0 E= +0.241 V
.~ CI- IC •• 6.0 «0-'" It is interesting to note the polarity of the
"
! (0.05FI (a) The titration will require 20.0 ml of electrodes has changed between the initial
0.0 A~~03' After 19.0 ml of the reagent have addition and the equivalence point. This be-
1
jij
8,-
(0.05FI
10.0 been added, the formal concentration of 1-
will be
havior is not often encountered.
(c) When the equivalence point has been
12.0 exceeded by 1.00 ml, the concentration of
-0.2 F _ 100 x 0.0200 - 19.0 x 0.100 excess Ag + is given by
N.I- 119
1- 14.0 F _ 1.00 x 0.100 _ 6 -4
1- adsorption
10.05FI AINO. - 121 - 8.2 x 10
-0.4 16.0
0 10 20 30 40 An additional source of iodide results from and
the finite solubility of the precipitate. The
Volume 0.100 F (or 0.Q1 FI AgNo•• ml
- concentration of iodide ions from this source
[Ag+] = 8.26 x 10-4 + [r] ~8.26 X 10-4
FIGURE 19-1 & Theoretical titration curves for 50.0 ml of will be just equal to the silver ion concentra- Here, it is easier to describe the behavior of
0.0500 F (or 0.005 F) sodium halide. Dotted lines show experi- tion. Thus,
mental deviations from theoretical curve for iodide due to ad- the silver electrode in terms of
sorption ofI- and Ag+ on the AgI precipitate.
I
-. .L
iodide originally prest;nt has precipitated.19 far, the most important reagents for complexo-
Thus, to this point, the titration curve will metric titrations are a group of amino carboxy-
1 Qe identical to that for iodide by itself (see lic acids, of which ethylene-ciiaminetetraacetic
E", = 0.799 - 0.0591 log [Ag+]
Figure 19-15). Precipitation of silver bromide acid is an example. This reagent, which is
then occurs; here again, it can be shown i~at commonly abbreviated as EDT A and given the
1
= 0.799 - 0.0591 log 8.26 x 10-. in theory no silver chloride forms until all but formula H. Y, has the structure
about 0.3 % of the bromide ion has been re-
moved from solution. Therefore, the titration HOOCCH2 CH COOH
= 0.617 V " / 2
curve in this region is similar to that for /N-CH2-CH2-N
Eceu = 0.617 -0.241 = 0.376 V bromide ion alone. Finally, chloride starts to HOOCCH2 'cH2COOH
precipitate, and the remainder of the curve
Titration of Mixtures. An important ad- is similar to the chloride titration curve in The four carboxylate functional groups, as
vantage of the potentiometric method is that Figure 19-15. well as the two amin~ nitrogens, participate in
it often permits discrimination among com- Figure 19-16 suggests that the individual bond formation with metal ions. Regardless
ponents of a mixture that react with a common components of a halide mixture can be deter- of the charge on the cation, the complex for-
titrant. Figure 19-16 illustrates this type of mined by a single potentiometric titration; in mation reaction can ~ formulated as
application for the determination of iodide, fact, the feasibility of such an analysis has been
bromide, and chloride ions in an approxi- demonstrated experimentally.20 For the ti-
mately equimolar mixture. During the early tration of any pair of these ions, it is generally Platinum
stages of the. titration, only silver iodide, found that the volume of silver nitrate required An important example of the application of wire sealed
in glass
the least soluble of the three silver halides, to reach the first end point is somewhat greater this reagent is the determination of calcium
precipitates. Indeed, it is readily shown that than that predicted by theory; the total volume, and magnesium ion~ in hard water. The reac- FIGURE 19-17 Atypical mer-
precipitation of silver bromide will not in however, approaches the correct amount. tion for calcium is ! cury electrode. (Electrodes of this
theory begin until all but about 0.02% of the These observations can be explained by assum- type can be obtained from Kontes
ing that coprecipitation of the more soluble Ca2+ + H. Y -+ Cay2- + 4H+ Manufacturing Corp., Vineland,
silver halide occurs during formation of the New Jersey.)
less-soluble compound; an overconsumption Indicator Electrodes. The use of mercury
of reagent in the first part of the titration as a second-order electrode for EDT A anions
results. was considered earlier (p. 540). Figure 19-17
Despite the coprecipitation error, the poten- is a schematic diagram of a typical commer- Figure 19-18 illustrates an application of
- +0.4
tiometric method is useful for the analysis of cially available mercury electrode. their procedure to the determination of bis-
w
u halide mixtures. When approximately equal Application of Potentiometric EDT A Titra- muth, cadmium, and calcium in a mixture.
'"~ quantities of halides are present, relative errors tions. Reilley and co-workers have made a Bismuth(III) is first titrated at a pH of 1.2.
; +0.2 can be kept to about 1 to 2 %. systematic theoretical and experimental study At this acidity, neither cadmium nor calcium
:! of the application of the mercury electrode to ions react to any significant extent. That is,
~ Complex Fonnation Titrations the potentiometric determination of 29 di-, for these two ions the equilibrium shown by
<> 0.0
trio, and quadrivalent cations with EDTA as Equation 19-24 lies far to the left. Bismuth,
Both metal and membrane electrodes can be
! ¥ used to detect end points for reactions that the reagent. 21 In t~e studies, 5 to 500 mg
quantities of the cation were titrated with 0.05
on the other hand, forms a complex that is
sufficiently stable to provide a satisfactory end
W -0.2 involve formation of soluble complexes. By
or 0.OO5F reagent splutions. One drop of a point. After bismuth has been titrated, the
1O-3F solution of Hgy2- was employed for solution is brought to a pH of 4 by addition
each titration . of an acetate/acetic acid buffer, and the titra-
•• See: D. A. Skoog and D. M. West, Fundament.1s of
tion is continued to an end point for cadmium.
Cheml.stry. 3d ed. New York: Holl. Rinehart
AlUJlytlaJI
and Winston, 1976, p. 176. Calcium ions do not react appreciably at this
FIG U R E 19-16 Potentiometric titration of 2. See: C. N. ReiUcyand R. w. Schmid, A,ud. CIrem.,30, pH but can be subsequently titrated in a basic
2. For further dc1ails, see: 1. M. Kollhoff and N. H.
2.50 mfw of .-, Br-, and CI- with 0.200 F Furman, Potentiometric Titrations, 2d ed. New York: 947 (1958); and C. N. Reilley. R. W. Schmid. and. D. W. solution obtained by addition of an ammonia!
AgN03• Wiley. 1931.pp. 154-158. Lamson, Ibid_ 953 (1958~ ammonium chloride buffer.
Volume 0.100 N NaOH. ml Volume NaOH. ml

(al Effect of acid strength (bl Effact of concentration

FIGURE 19-19 Titration curves for weak acids. (a) 50.0 ml of 0.100 N acid in each case.
(b) 50.0 ml of 0.100 N and 0.001 N HA with 0.100 N and 0.00100 N NaOH.

mI0.1MEOTA

FIGURE 19-18 Potentiometric EDTA titration of a mixture of three


cations. [Reprinted with permission from C. N. Reilley, R. W. Schmid, and Titration of Mixtures of Acids (or Bases).
D. W. Lamson, Anal. Chern., 30, 953 (1958). Copyright by the American Potentiometric titrations are advantageously
Chemical Society.] employed for the titration of mixtures of acids
'(or bases). If the dissociation constants differ CH3COOH
sufficiently, it is possible to determine the con- K •• 1.75 X 10-5
centrations of the individual components in
such mixtures. This application is demon-
Neutralization Titrlltions tion becomes lower, the ease of end point
strated by the titration curves shown in Figure
detection becomes poorer. Furman22 has in-
Potentiometric acid-base titration curves are 19·20. Note that the individual concentrations
dicated that in order to obtain an accuracy of
readily obtained with a glass/calomel elec- of trichloroacetic and acetic acid in a mixture
1 % relative in the potentiometric titration
trode system. End points from such curves could be derived from a potentiometric titra-
of a weak acid with a strong base, it is neces-
are particularly useful when colored or turbid tion curve. On the other hand, only the total
sary that
solutions must be titrated. acid concentration of the iodoacetic/acetic acid
Titration Curves for Weak Acids (or Weak FNaAK. ~ 10-5 mixture could be ascertained because the
Bases). Figure 19-19ashows theoretical curves strengths of these two acids are not sufficiently
where F NaA is the end point concentration of
for acids having different dissociation con- different. Generally, the concentrations of the
the sodium salt of the acid being titrated and
stants when titrated with standard sodium K. is the dissociation constant for the acid. two components of a mixture of acids (or bases)
hydroxide. Figure 19-19b shows the effect of can be determined if the ratio of their dissoci-
The foregoing considerations apply equally 20
analyte and reagent concentrations on the ation constants is 11)4or greater.
to the titration of solutions of weak bases Volume 0.100 N NaOH. ml
titration curve for one of these acids. Exper- Titration Cunes of Polyprotic: Acids or
with strong acids.
imental curves, obtained with a glass/calomel Bases. Figure 19-21 shows titration curves for FIGURE 19-20 Titration curves for 50.0
electrode system. approximate these theoreti- three common acids that have more than one ml of two mixtures. Both are 0.0500 F in
cal curves closely. Oearly, as the strength of •• N. H. Furman, in Treatise on AnalyriaJI C/Ierni$lry, titratable proton. Sulfuric acid exhibits only a acetic acid. Mixture A is also 0.0250 F in
the acid becomes less (and thus the reaction cds. I. M. KohhofJ' and P. J. Elving. New York: WiIey- single end point (curve C) because the degree iodoacetic acid; similarly B; is 0.0250 F in
with base less complete) and the concentra- Intencienee. 1967, Part I, vol. 4, p. 2287. of dissociation of the two protons is not trichloroacetic acid.
I
,...-L
what lower pH values for curve A compared Upon converting to p functions and rearrang- ible; in water, the end point for this titration
with curves B and C. ing, we obtaih is not satisfactory.
Determination of D~tion Constants The potentiometric method has proved
for Acids (or Bases). An approximate numeri- (pK.)..pt = pH = pK. + log hfiA- particularly useful for signaling end points of
. HA
cal value for the dissociation constant of a titrations in nonaqueous solvents. The ordi-
weait acid or base can be estimated from ~ Thus, the true pK. will differ from the exper- nary glass/calomel electrode system can be
tentiometric titration curves. In theory, thIS imental pK. by the logarithm of the ratio used; the electrodes must be stored in water
quantity can be obtained from any ~~t along of activities. Ordinarily, the ionic strength between titrations to prevent dehydration of
the curve; as a practical matter, It IS most during a titration will be 0.1 or greater. Thus, the glass and precipitation of potassium
easily found from the pH at the point of half- the ratio of fA- to fHA would be at least 0.75 chloride in the salt bridge. Ordinarily, the
neutralization. For example, in the titration of (see Table 18-1) if HA is uncharged. For sol- millivolt scale rather than the pH scale of the
the weak acid HA, we may ordinarily assume utes such as H2PO.; and HPO~-, this ratio potentiometer should be employed because
that at the midpoint will be even larger. the potentials in nonaqueous solvents may
A value for the equivalent weight, the num- exceed the pH scale. Furthermore, the pH scale
ber of titratable protons, and the approximate based upon aqueous buffers has no significance
dissociation constant of a pure sample of an in a nonaqueous environment. The titration
unknown acid can be obtained from a single curves are thus empirical; however, they pro-
K = [W][J?] = [H+] potentiometric titration; this information is vide a useful and satisfactory means of end
• [111\] frequently sufficient to identify the acid. point detection.
FIG U R E 19-21 Curves for the titration Titrations in Nonaqueous Solvents. Acid-
of polybasic acids. A 0.100 N NaOH solution base titrations in aqueous solvents are limited
is used to titrate 25.0 ml of: 0.100 F H3P04 to substances with acidic or basic dissociation Oxidation-Reduction Titration.
pK.= pH
(Curve A~ 0.100 F oxalic acid (Curve B~ constants greater than about 10- 9. For weaker
It is important to note that a dissociation acids or bases, the reactions with the reagent A platinum electrode responds rapidly to
0.100 F H2S04 (Curve C).
constant determined in this way may differ are so incomplete that the change in pH at the many important oxidation-reduction couples
from that shown in a table of dissociation equivalence point is not sufficient to permit and develops a potential that depends upon
constants by a factor of 2 or more because the accurate establishment of the end point. Often, the concentration (strictly, ~ctivity) ratio of
latter is based upon activities while the former however, analysis becomes feasible if a non- the reactants and the products of such half-
sufficiently different (one proton is completely
is not Thus, if we write the dissociation con- aqueous solvent is substituted.23 For example, reactions. For example, a platinum electrode
dissociated while K. for the second is 1.2 x
stant expression in its more exact form, we aqueous solutions of aniline (Kb - lO-IO) are system can be used to determine the end
lO-2). The ratio between K1 and K2 for oxalic
so weakly basic that titration with standard point in the titration of iron(lI) with a stand-
acid is about 1000; the curve for the titration obtain
perchloric acid is unsatisfactory. On the other ard solution of potassium permanganate (see
(curve B) shows an inflection corresponding
aH.O+ . [A -] . fA- hand, if glacial acetic acid solutions of aniline Figure 19-13). The platinum electrode is re-
to the first equivalence point. However, the
[HA] 'fHA and perchloric acid are employed, the reaction sponsive to both oxidation-reduction systems
magnitude of the pH change is too small to
between the two is nearly complete at the that exist in the solution throughout the ti-
permit accurate determination of this end Again, we assume that [A -] and [HA] are
point; the second is clearly well suited for a equivalence point, and a large change in pH tration. That is,
approximately equal; furthermore, the poten-
quantitative determination of the acid. occurs as a consequence. Similarly, the reaction
tial of a glass electrode gives a good approxi-
For phosphoric acid, the ratio of K1 to K2 of phenol (K. - 10-1°) with a base in a non-
mation for aH,o+' Therefore, aqueous solvent such as methyl isobutyl ketone
is 105; two well-defined end points are ob-
tained. Thus, analysis of phosphoric acid, so- K _ aH.O+ . (K"] . fA- = aH.O+ . fA- is sufficiently favorable to make titration feas-
dium dihydrogen phosphate, and mixtures of .- [~] 'fHA fHA
these two compounds would be possible by
means of a single potentiometric titration. The Taking the negative logarithm of the two sides
After each addition of reagent, interaction
third hydrogen of phosphoric acid is so slightly of the equation yields 23 For a more complete discussion or acid-base lilralions
among the species occurs until reactant and
dissociated (K3 = 4.2 x 10-13) that it does in nonaqueous solvents. see: D. A. Skoog and D. M.
fA- product concentrations are such that the elec-
not yield an end point of practical value. Its -log K. = -log aH.O+ -log fi- West, Fundamental. of Analyrieal Chemistry. 3d ed. New
York: Holl. Rinehan and Winslon, 1976. Chapter 12. trode potentials for the two half-reactions are
presence, however, is reflected in the some- HA

i
-L
identical. That is, at equilibrium, creased change in the electrode potential in the Rubber
bulb
routine analyses are required.24 Automatic ti-
end point region. trators cannot yield results that are more accu-
0.OS91 [Mn2'+] The curves in Figure 19-22 were derived
\ rate than those obtained by manual potentio-
EPl = I.S1 - -S-log [MnO;][H+]8
for reactions in which the oxidant and reduc- metric techniques; however, they do decrease
0.OS91 [FeH] tant each exhibit a one-electron change; where the operator time needed to perform the ti-
= 0.771 - -I-log [Fe3+] both reactants undergo a two-electron transfer, trations and thus may offer significant ec0-
the change in potential in the region of 24.9 nomic advantages.
Thus, the measured potential can be thought to 2S.1 ml is larger by about 0.14 V. Flow Control and Measurement of Reagent
of as arising from either of the two half-cell Solutions containing two oxidizing agents Volume. Several methods exist for automatic
systems. or two reducing agents will yield titration control of the addition of a reagent and the
- The change in the ordinate function in the curves that contain two inflection points, pro- measurement of its volume. The simplest em-
-equivalence-point region of an oxidation-re- vided the standard potentials for the two ploys an ordinary buret in which the stop-
:duction titration becomes larger as the reaction species are sufficiently different. If this differ- cock is replaced with an electromagnetic pincer
becomes more complete; this effect is identical ence is greater than about 0.2 V, the end points device. Here, an elastic plastic tube is inserted
with that encountered for other reaction types. are usually distinct enough to permit analysis between the buret body and the tip. Flow is
:Thus, in Figure 19·22, data are plotted for for each component. This situation is quite prevented by pinching the tube between a
titrations involving a hypothetical analyte that comparable to the titration, with the same spring-loaded soft iron piece and a metal
has a standard potential of 0.2 V with several reagent, of two acids with different dissociation wedge. Titrant is introduced by passage of
reagents that have standard potentials ranging constants or of two ions that form precipitates electricity through a solenoid that surrounds
from 0.4 to 12 V; the corresponding equilib- of different solubilities. the pinching device. In another type of valve,
rium constants for the reaction range from FIGURE 19-23 Apparatus for differ- a small piece of iron is sealed into a glass or
about 2 x 103 to 9 X 1016• Clearly, an increase ential potentiometric titrations. plastic tube that fits inside the outflow tube of
in completeness is accompanied by an in- a buret. The two surfaces are ground to form
a stopper. Current in a solenoid unseats the
We have seen that a derivative curve gen- stopper and allows a flow of reagent.
erated from the data of a conventional p0- The most widely used apparatus for auto-
homogenized by squeezing the rubber bulb
tentiometric titration curve (Figure 19-12b) matic reagent addition consists of a calibrated
1.2
A several times, whereupon I1E again becomes
reaches a distinct maximum in the vicinity of syringe that is activated by a motor-driven
EO" _ EOA
zero. If the volume of solution in the tube
K the equivalence point. It is also possible to micrometer screw. Such a device is shown in
1.0 B that shields the electrode is kept small (say, 1
> A 1.00 V 8.9 X 10'8 acquire titration data directly in derivative Figure 19-2S.
B 0.80 V 3.6 X 10'3 to S ml1 the error arising from failure of the
i4i 0.8 form by means of suitable apparatus. Preset EncI-Point TItr.tors. Figure 19-24 is
C 0.60 V 1.5 X 10.0 C final addition of reagent to react with this
1i
0 0.40 V 6.0 X 10·
A differential titration requires the use of a schematic diagram of the simplest and least
.~ portion of the solution can be shown to be
0.6 E 0.20 V 2.4 X 103 two identical indicator electrodes, one of which expensive type of automatic titrator. Here, a
g, D negligibly small.
is well shielded from the bulk of the solution. preset equivalence point potential is applied
For oxidation-reduction titrations, two
1
W
0.4

0.2
E
Figure 19-23 illustrates a typical arrangement.
Here, one of the electrodes is contained in a
small sidearm test tube. Contact with the bulk
platinum wires can be employed, with one
enclosed in an ordinary medicine dropper.
across the electrodes by means of a calibrated
potentiometer. If a difference exists between
this potential and that of the electrodes, an
The main advantage of a differential
of the solution is made through a small " error" signal results. This signal is amplified
method is the elimination of the need for the
(-1 mm) hole in the bottom of the tube. and closes an electronic switch that permits
reference electrode and salt bridge.
Because of this restricted acoess, the com-
position of the solution surrounding the
FIGURE 19-22 Titration of 50.0 ml of shielded electrode will not be immediately
0.0500 N A. Fj is assumed to be 0200 V. affected by an addition of titrant to the bulk
2' For a comprehensive discussion of aUlomatic titrato~
From the top, ~ for the reagent is 120, of the solution. The resulting difference in s0- In recent years, several automatic titrators
see: G. Svehla, Automatic Potmtiomnrlc TitratioM. New
1.00, 0.80, 0.60, and 0.40 V, respectively. lution composition gives rise to a difference in based on the potentiometric principle have York: Pergamon Press, 1978; and J. K. Foreman and
Both the reagent and analyte are assumed potential AE between the electrodes. After become available from commercial sources. P. B. Stockwc1~Automatic Chemical Analysis. New York:
to undergo a one-electron change. each potential measurement, the solution is These instruments are useful where many Wiley, 1975,pp, 44-62,
Indicator
lamp

,t-
-;"I\F

a flow of electricity through the solenoid- the equivalence point. This change in sign then!
,.--
I
operated valve of the buret. As the error signal causes a switching device to turn oft' the flow I

,J I
approaches zero, current to the solenoid is of titrant. _J
switched oft', and flow of titrant ceases. Rec:orcIing Tatrators. Recording titrators Potentiometer drive on;
Titrators such as the one shown in Figure carry a titration beyond the equivalence point 1 buret and chart drive off

-- r-- --
19-24 do not respond instantaneously and white recording a curve for the analyte. In
I
tend to overshoot the end point. To overcome some instruments, the rate of reagent addition
this problem, some instruments are equipped is held constant and is synchronized with the Buret and chart drive on;
potentiometer drive off
to superimpose a square-wave signal upon chart drive of a millivolt recorder. The pen ~ ~ 00
the error signal. The solenoid switch is then then rerords the amplified output potential of
FIGURE 19-25 (a) An automated, curve-recording titrator. (b) Theoretical titration curve.
set to close only when the net signal exceeds the cell as a function of time, which is propor-
(c) Recorded curve with end-point anticipation. (d) Enlarged portion of (c) with explanation.
that of the square wave. Titrant is then added tional to the reagent volume.
(From G. Svehla, Automatic Potentiometric Titrations. New York: Pergamon Press, 1978, p. 176.
in increments that can be controlled by the Often, recording titrators are equipped to With permission.)
frequency of the square-wave signal. perform a titration the way a skilled chemist
Second-Derivative Tatraton. Second-deriv- does-that is, to add titrant rapidly before
ative titrators can also be relatively simple and after the end point but in small incre-
devices; they have the advantage that no pre- ments as the end point is approached and
knowledge of the equivalence-point potential passed. Figure 19-25 is a simplified block dia- the error signal is small, the electronic switch drives are turned on. Thus, the reagent is added
is required. The signal processor of these gram of such an instrument. The switching activates the motor of the syringe-type buret in a series of steps, as shown by the recorded
devices contains two electronic derivative cir- device in this instrument operates from an and the chart drive of the recorder; simulta- titration curve in Figure 19-25c.
cuits (p. 66) in series to convert the amplified amplified error signal that develops when the . neously, the potentiometer and pen drive Fully Automatic Tatraton. A fully automat-
signal from the electrode to a voltage propor~ •potential of the electrodes differs by some motor is switched off. When sufficient reagent ic titrator is equipped with a turntable that
tional to the second derivative of the electrode minimum amount from the potential applied has been added to cause the error signal to holds a series of samples for titration. After a
potentiatofthe indicator electrode. The output by the potentiometer. The contact of this exceed some predetermined level, the switch- titration has been completed, the solution is
is then similar in form to that shown in Figure potentiometer is driven by the motor, which ing is reversed. Now, the buret and chart discarded, the vessel and electrodes are rinsed,
19-12c, where the sign of the signal changes at also ~itions the pen of the recorder. When drives are off and the potentiometer and pen the buret is refilled, the sample table is rotated,
a measured volume of a new sample is intro- analytical results. Instruments of this kind are (a) A cell with a Pb indicator electrode for the determination of pS04' ,
duced into the system, and the titration process expensive, but their costs can be easily justified (b) A cell with a Aa indicator electrode for the determination of pC03 .
is resumed. Such instruments are controlled by for laboratories that must perform large num- (Ag2C03 has a slow solubility.)
microprocessors and usually have computing bers of routine titrations regularly. (c) A ce~ with a platinum electrode for the determination of pTl(I).
facilities for calculating and printing out the 5. The' following cell was employed for the determination of pCrO 4:
SCEIIAg2Cr04(sat'd~ CrO~-(xM>I Ag
1. (a) Calculate the standard potential for the reaction
Calculate pCr04 ifthe cell potential is 0.356 V.
CuBr(s) + e~Cu(s) + Br- 6. Calculate the potential of the cell (neglecting the junction potential)
For CuBr, KIp = 5.2 X 10-9• Indicator electrodellSCE
(b) Give a schematic representation of a cell with a copper indicator
electrode as an anode and a saturated calomel electrode as a where the indicator electrode is mercury immersed in a solution that
is .
cathode that could be used for the determination of Br-.
(c) Derive an equation that relates the measured potential of the cell (a) 4.00 x 1O-6M HgH.
in (b) to pBr (assume that the junction potential is zero). (b) 4.00 x 10-6M Hg~+. .
(d) Calculate the pBr of a bromide-containing solution that is sat- (c) saturated with Hg2Br2 (KIP = 5.8 x 10-23) and is'3.00 x 1O-3M
urated with CuBr and contained in the cell described in (b) if the in Br-.
resulting potential is 0.0176 V. (d) 2.50 x 10-'F in Hg(N03h and O.04OOF in KCI.
2. (a) Calculate the standard potential for the reaction HgH. + 2CI- ~HgCI2(aq) K, = 6;1 X 1012
AgSCN(s) + e~Ag(s) + SCW (e) 2.50 x lO-'F in Hg(N03h and O.OO4OOF in KCI. :

For AgSCN, Ksp = 1.1 X 10-12. 7. The formation constant for the mercury(II) cyanide cdmplex is
(b) Give a schematic representation of a cell with a silver indicator
Hg2+ + 2CN- ~Hg(CNh(aq) K, = 5.0 X W4
electrode as the cathode and a saturated calomel electrode as an
anode that could be used for determining SCN-. Calculate the standard potential for the half-reaction
(c) Derive an equation that relates the measured potential of the cell
in (b) to pSCN (assume that the junction potential is zero).
Hg(CNh + 2e~Hg(l) + 2CW
(d) Calculate the pSCN of a solution that is saturated with AgSCN 8. The standard electrode potential for the reduction of the mercury
and contained in the cell described in (b) if the resulting potential complex of EDT ~ is given by
is 0.308 V.
Hgy2- + 2e~Hg(l) + y4- EO = 0.21 V
3. Give a schematic representation of each of the following cells and
an equation for the relationship between the cell potential and the Calculate the formation constant for the reaction
desired quantity. Assume that the junction potential is negligible, and
specify any necessary concentrations as 1.00 x 10-4M. Treat the indi-
HgH + y4- ~Hgy2-
cator electrode as the cathode in each case. 9. Calculate the potential of the cell (neglecting the junction potential)
(a) A cell with a Hg indicator electrode for the determination of pBr.
(Hg2Br2 has a low solubility.) Hg\Hgy2-(4.50 x lO-'M), y4-(xM)IISCE
(b) A cell with a Ag indicator electrode for the determination of
where y4- is the EDTA anion, and the concentratioll of y4- is
pC204' (Ag2C204 has a low solubility.) (a) 3.33 x 10-IM, (b) 3.33 x 10-3M, and (c) 3.33 x iO-'M.
(c) A cell with a platinum electrode for the determination of pSn(II~
4. Give a schematic representation of each of the following cells and Hgy2- + 2e~Hg(l) + y4- £0 = 0.21 V
an equation for the relationship between the cell potential and the
10. The following cell was found to have a potential of 0.209 V when the
desired quantity. Assume that the junction potential is negligible, and
solution in the left compartment was a buffer of pH 5.21:
specify any necessary concentration as 1.00 x 10-4M. Treat the indi-
cator electrode as the anode in each case. Glass electrode I W(a = x)IISCE
The following potentials were obtained when the buffered solution (c) What is the relative error in [H+] associated with the uncertainty
was replaced with unknowns. Calculate the pH and the hydrogen ion in 8J?
activity of each unknown. 18. The following cell was found to have a potential of 0.3674 V:
(a) 0.064 V (c) 0.510 V . Membrane electrode for Mg2+ IMgH(a = 6.87 x 10-3M)IISCE
(b) 0.329 V (d) 0.677 V
11. The following cell was found to have a potential of 0.411 V: (a) When the solution of known magnesium activity was replaced
with an unknown solution, the potential was found to be 0.4464 V.
Membrane electrode for Mg2+ IMgH(a = 1.77 x 10-3M)IISCE What was the pMg of this unknown solution?
When the solution of known magnesium activity was replaced with an (b) Assuming an uncertainty of ±0.002 V in the junction potential,
unknown solution, the potential was found to be 0.439 V. What was what is the range of MgH activities within which the true value
the pMg of this unknown solution? Neglect the junction potential. might be expected?
12. The following cell was found to have a potential of 0.893 V: (c) What is the relative error in [MgH] associated with the uncer·
tainty in 8j?
Cd ICdX2(sat'd~ X- (0.0200M)IISCE 19. The sodium ion concentration of a solution was determined by mea-
Calculate the solubility product of CdX2, neglecting the junction surements with a glass-membrane electrode. The electrode system
potential. developed a potential of 0.2331 V when immersed in 10.0 ml of the
13. The following cell was found to have a potential of 0.693 V: unknown. After addition of 1.00 mI of a standard solution that was
2.00 x 1O-2M in Na+, the potential decreased to 0.1846 V. Calculate
Pt, H2(1.00 atm)l HA(O.200F~ NaA(O.300F)IISCE the sodium ion concentration and the pNa of the original solution.
Calculate the dissociation constant of HA, neglecting the junction 20. The calcium ion concentration of a solution was determined by mea-
potential. surements with a liquid-membrane electrode. The electrode system
14. A 4O.00-ml aliquot of 0.1000N U4+ is diluted to 75.0 ml and titrated developed a potential of 0.4965 V when immersed in 25.0 ml of the
with O.OSOONCe4+. The pH of the solution is maintained at 1.00 sample. After addition of 2.00 ml of 5.45 x 10- 2F CaCI2, the poten-
throughout the titration. (Use 1.44 V for the formal potential of the tial changed to 0.4117 V. Calculate the calcium concentration and the
cerium system.) pCa of the sample.
(a) Calculate the potential of the indicator cathode with respect to a
saturated calomel reference electrode after the addition of 5.00,
10.00, 15.00,25.00,40.00,49.00, SO.OO,51.00, 55.00, and 60.00 mlof
cerium(IV~
(b) Draw a titration curve for these data.
15. Calculate the potential of a mercury cathode (vs. SCE) after the addi-
tion of 5.00, 15.00, 25.00, 30.00, 35.00, 39.00, 40.00, 41.00, 45.00, and
50.00 ml of O.OSOOF Hg2(N03)2 to SO.OOml of O.OSOOF NaCI. Con-
struct a titration curve from these data Kip Hg2C12, 1.3 X 10-18•
16. Calculate the potential (vs. SCE) of a lead anode after the addition of
0.00, 10.00, 20.00, 24.00, 24.90, 25.00, 25.10, 26.00, and 30.00 ml of
O.2000F NaI03 to 50.00 ml of O.OSOOF Pb(N03h. For Pb(I03h,
Ksp = 3.2 X 10-13•
17. A glass-calomel electrode system was found to develop a potential of
0.0620 V when used with a buffer of pH 7.00; with an unknown
solution the potential was observed to be 0.2794 V.
(a) Calculate the pH and [H+] of the unknown.
(b) Assume that K is uncertain by ±0.001 Vasa consequence of a
difference in the junction potential between standardization and
measurement. What is the range of [H+] associated with this
uncertainty?
Three related electroanalytical methods, name- values can be calculated from standard poten-
ly, electrogravimetric analysis, constant-poten- tials by means of the Nernst equation. The
tial coulometry, and coulometric titrations, are terms III and 1I2 are potentials associated
discussed in this chapter. Each involves an with concentration and kinetic polarization
electrolysis that is carried on for a sufficient (p. 5281 respectively. Both are always negative,
length of time to assure quantitative oxidation thus implying that an additional applied p0-
or reduction of the analyte. In electrogravi- tential is required to overcome their effects.
metric methods, the product of the electrolysis Furthermore, III and 1I2 are each made up of
is weighed as a deposit on one ofthe electrodes. two terms, one associated with each electrode.
In the two coulometric procedures, on the Thus,
other hand, the quantity of electricity needed
to complete the electrolysis serves as a measure
of the amount of analyte present.
The three methods generaUy have moder-
ate selectivity, sensitivity, and speed; in many
instances, they are among the most accurate
and precise methods available to the chemist, Here, the subscripts a and c again refer to
with uncertainties of a few tenths percent rel- anode and cathode.
ative being not uncommon. Finally, in contrast Of the seven potentials shown in the three
to all of the other methods discussed in this equations, only Ee and E. can be derived
text, these three require no calibration against from theory; the others must be evaluated
standards; that is, the functional relationship empirically.
between the quantity measured and the weight
of analyte can be derived from theory.
Operation of a Cell at
a Fixed Applied Potential
One method of operating an electrolytic cell
CURRENT-VOLTAGE is to hold the applied potential at some fixed
RELATIONSHIP predetermined level through the entire elec-
DURING AN ELECTROLYSIS trolysis. This method, while simple, has some
distinct limitations. In order to understand
It is useful to consider the changes in current, these limitations, it is necessary to consider
voltage, and time when an electrolytic cell is how the current and the potential of the
operated in three different modes: (I) the ap- working electrode (the electrode at which the
plied cell potential is held constant; (2) the cell analytical reaction occurs) vary as a function
current is kept constant; and (3) the potential of time.
of one of the electrodes (the working electrode, To illustrate current-voltage relationships
which involves the analyte) is held constant. during an electrolysis at fixed potentia~ con-
For all three, the behavior of the cell is sider a cell consisting of two platinum elec-
governed by the relationship trodes, each with a surface area of ISO em2,
immersed in 200 ml of a solution that is
0.0220M with respect to copper(lI) ion and
LOOM with respect to hydrogen ion. The cen
where E.pp1 is the applied potential from the has a resistance of 0.50 n. When electricity is
external source and Ee and E. are the revers- forced through the cell, the analyte copper is
ible or thermodynamic potentials associated deposited upon the cathode, and oxygen
with the cathode and anode, respectively; their is evolved at a partial pressure of 1.00 atm at
1.0 tion of trace impurities such as oxygen and will be 1.23V throughout the !'lectrolysis. Thus, where D is the diffusion coefficient in em2 Is or
Theoretical
curve I iron(III~ which are inevitably present in the the required applied potential at any time the rare at which the reactant diffuses under a
neglecting I during the electrolysis is given by (Equation
polarization I solution. In addition, however, small amounts unit concentration gradient. The quantity A is
of copper are deposited at potentials lower than 20-1) .
< I the electrode surface area in em, V is the
C I the decomposition potential. This apparent
~
0.5 I Eappl= E. - 1.23 + n1• + n~a - 0.501 volume of the solution in em3, and 15 is the
I departure from theoretical behavior arises be- thickness of the surface layer in which the
"
u I
cause it was assumed that the activity of copper
(20-4)
I concentration gradient exists. Typical values
I was unity in calculating the decomposition provided that the cell resistance remains con-
I for D and 15 are 10-5 em2/s and 2 x 10-3 em.
0.0 potential. Experiments have shown, however, stant throughout the electrolysis. (The constant 25.8 includes the factor of 60 for
0 -1.0 -2.0 -3.0
that the activity of a metal deposit that only Let us now assume that we wish to operate converting D to cm2/min, thus making k com-
partially covers a platinum surface is less than the cell initially at a current of 1.5 A, which patible with the units of t in the equation
Applied potential, V
onel; under these circumstances, then, the corresponds to a current density of 0.010 for I,.)
FIGURE 20-' Current-voltage curve for behavior of the cathode is more correctly A/em2• From Table 18-3, it is seen that n2a, When the initial applied potential is - 2.5 V,
the electrolysis of a copper(II) solution. described by the initial oxygen overvoltage, will be about it is found that concentration polarization, and
-0.85 V. Initially, the concentration polari- thus an exponential decrease in current, occurs
0.0591 [Cu2+]
zation will be zero because the concentration
E= ~U2+ --2-10g [Cu] essentially immediately after application of the
of copper ions is high. and Equation 20-4 potential. Figure 20-2a depicts this behavior;
becomes
the anode. The overall cell reaction can be where [Cu] is infinitely small at the outset the curve shown was derived for the cell under
expressed as and approaches one only after sufficient cur- Eappl= 0.29 - 1.23 - 0.00 - 0.85 consideration with the aid of the foregoing
rent has passed to coat the platinum surface two equations. After 30 min, the current has
Cu2+ + H20 -+ Cu(s) + !02(g) + 2H+ completely. .
- 1.50 x 0.50 decreased from the initial 1.5 A to 0.08 A; 'by
Decompositioo PotentiaL The reversible de- Finally, it should be noted that the expen- = -2.54 V this time, approximately 96 % of the cop~r
composition potential can be computed f~om mental curve departs from linearity at high has been deposited. ,
Thus, a reasonable estimate of the potential
standard potential data for the half-reactiOns currents owing to the onset of concentration Potential Changes During an Electrolysis at
required to produce an initial current of 1.5 A
polarization (see p. 528). is -2.5 V. COMtant Applied PotentiaL It is instructive to
Cu2+ + 2e~Cu(s) EJ = 0.34 V Cakulatioo of Requirell PotentiaL For the consider the changes in some of the potentials
ClHTent Changes During an Electrolysis at
!O2 + 2H+ + 2e~H20 EJ = 1.23 V cell under consideration, kinetic polarization in Equation 20-4 as an electrolysis proceeds.
Constant Appliecl PotentiaL It is useful to con-
occurs only at the anode where oxygen is Figure 20-2b depicts these changes during the
and has a value of -0.94. No current would sider the changes in current in the foregoing
evolved; the cathodic reaction is rapid and copper deposition under consideration.
be expected at less negative potentials; at cell whcn the potential is held constant at
reversible. Thus, Equation 20-3 simplifies As mentioned earlier, the thermodynamic
greater applied potentials, the current would - 2.5 V throughout the electrolysis. Under
to n2 = n2a. Furthermore, concentration anode potential remains substantially un-
theoretically be determined by the magnitude these circumstances, the current would be
polarization at the anode is negligible ~use changed throughout the electrolysis. The re-
of the cell resistance. found to decrease with time owing to depletion
the anodic reactant and products are m large versible cathode potential E., on the other
The dotted lines in Figure 20-1 illustrate of copper ions in the solution and an increase
excess compared with concentration changes hand, becomes smaller (more negative) as the
the theoretical current-voltage behavior of this in the degree of concentration polarization. In
brought about by the electrolysis; therefore, it copper concentration decreases. The curve for
cell; the decomposition potential is seen to be fact, with the onset of concentration polariza-
is unlikely that the surface layer will become E in Figure 20-2b was derived by substituting
the intersection of two straight lines. The linear tion, the current decrease becomes exponential
depleted in these species, and Equation 20-2 the calculated copper concentration after vari-
rise in current immediately beyond the de- in time. That is,
simplifies to nl = nl •. Finally, it is readily ous electrolysis periods into the Nernst equa-
composition potential is a reflection of Ohm's shown by suitable substitutions into the Nernst I, = Ioe-I<t tion. Note that the potential decrease is
law. The experimental current-voltage re- equation that the increase in hydrogen ion approximately linear with time over a C<)n-
lationship for this system will, however, more where I, is the current t min after the onset of
concentration due to the anodic reaction re- siderable period. ;
closely resemble the solid curve in Figure 20-1. polarization and 10 is the current immediately
sults in a negligible change in electrode p0- before polarization. Lingane2 has shown that The IR drop shown in Figure 2O-2b paral-
Here, the oxygen ovcrvoltagc at the anode has tential «0.01 V); that is, the anode potential • leis the current changes shown in Figure 20-2a.
the effect of displacing the curve to more values for the constant k can be computed
negative potentials. Moreover, a small current from the relationship
is observed with the first application of a poten- See: L. B. Rosen et aI. J. Eleetroehern. Soc. 95, 25, 33, k= 25.8DA
1 2 See: J. J. Lingane, ElectrOQ1lQylkol Chemistry, 2d ed.
tial. A part of this current is due to the reduc- 129 (1949); !l8, 441, 452, 451 (1951~ V15 New York: Interscience, 1958, pp. 223-229.
The negative sign is employed for consistency increase in the time required for an analysis.
Electrode reaction primarily
with Equation ~. At best, an electrolysis at constant cell 2H' + 2e = H2(g)
As shown by the topmost curve of Figure potential can only be employed to separate
2O-2b; the oxygen overvoltage also becomes an easily reduced cation from those that are ~ -0.5

j~~ o 10 20
less negative as the current, and thus the cur-
rent density, falls. The data for this curve
were obtained from more extensive compila-
more difficult to reduce than ·hydrogen ion.
Evolution of hydrogen can also be expected
near the end of the electrolysis unless certain ~
I
precautions are taken to prevent it.
Time,min
tions that are similar to Table 18-3.
The most significant feature of Figure 20-2b
5
is the curve representing the change in total Equilibrium potential
for Cu2• + 2e •• Cuts)
cathode potential (E. + nlc) as a function of Constent-Current Electrolysis
time. It is evident that as IR and nla become
less negative, one or more of the other poten- The analytical electrodeposition under con-
tials in Equation 20-4 must become more nega- sideration, as well as others, can be carried
FIGURE 20-3 Changes in cathode po-
tive. Because of the large excess of reactant and out by maintaining the current, rather than
tential during the deposition of copper with
product at the anode, its potential remains the applied potential, at a more or less con-
a constant current of 1.5 A. Here, the cathode
substantially constant. Thus, the only poten- stant level. Here, periodic increases in the ap-
potential is equal to (E. + 7tlc)'
tials that can change are those associated with plied potential are required as the electrolysis
the cathode; as seen from a comparison of proceeds.
the curves labeled E. and (E. + nl.~ it is evi- In the preceding section, it was shown that
dent that most of this negative drift is a con- concentration polarization at the cathode
causes a decrease in current. Initially, this effect predominates. The changes in cathode poten-
sequence of a rapid increase in concentration
tial under conditions of constant current are
polarization n1•• That is, even with vigorous can be partially offset by increasing the applied
shown in Figure 20-3.
stirring, copper ions are not brought to the potential; the enhanced electrostatic attraction
electrode surface at a sufficient rate to prevent will cause copper ions to migrate more rapidly,
polarization. The result is a rapid decrease in thus maintaining a constant current. Shortly,
however, the solution becomes sufficiently
Constent Cathode
IR and a corresponding negative drift of the
Potential Electrolysis
cathode potential . depleted in copper ions so that the forces of
The rapid shift in cathode potential that diffusion and electrostatic attraction cannot From the Nernst equation, it is seen that a
accompanies concentration polarization often keep the electrode surface supplied with suf- tenfold decrease in the concentration of an
leads to codeposition of other species and loss ficient copper ion to maintain the desired ion being deposited requires a negative shift
of selectivity. For example, points A and Bon current. When this occurs, further increases in in potential of only 0.0591/n V. Electrogravi-
the cathode potential curve indicate where E.ppl cause rapid changes in n1 and the cathode metric methods, therefore, are potentially high-
lead and cadmium ions would begin to depos- potential; codeposition of hydrogen (or other ly selective. In the present example, the copper
it if present in concentrations about equal to reducible species) then takes place. The cath- concentration is decreased from 0.02M to
that of the copper ion. Another event that . ode potential ultimately becomes stabilized at lO-6M as the thermodynamic cathode poten-
probably would occur would be the evolution a level fixed by the standard potential and the tial E. changes from an initial value of +0.29 V
of hydrogen at about point C (this process overvoltage for the new electrode reaction; to +0.16 V. In theory, then, it should be
was not taken into account in deriving the further large increases in the cell potential are feasible to separate copper from any element
i curve in Figure 20-2a). no longer necessary to maintain a constant that does not deposit within this 0.13-V poten-
FIGURE 20-2 Changesin(a)current;and The interferences just described could be current. Copper continues to deposit as tial range. Species that deposit quantitatively
(b) potentials during the electrolytic deposi- avoided by a decrease in the applied potential copper(II) ions reach the electrode surface; the at potentials more positive than + 0.29 V could
tion of CuH. Points A. and B are potentials by several tenths of a voh so that the negative contribution of this process to the total current, be eliminated with a prercduction; ions that
at which Pb and Cd would begin to codeposit drift of the cathode potential could never however, becomes smaller and smaller as the require potentials smaller than +0.16 V would
if present. Point C is the potential at which reach a level at which the interfering ions pre- deposition becomes more and more nearly not interfere with the copper deposition. Thus,
82 might begin to form at the cathode. Any cipitate. The consequence, however, is a dim- complete. The alternative process, such as re- if we are willing to accept a reduction in analyte
of these processes would distort the curve A.. inution in current and, ordinarily, an enormous duction of hydrogen or nitrate ions, soon concentration to lO-16M as a quantitative
ELECTROGRAVIMETRIC AND COULOMETRIC METHODS

separation, it follows that divalent ions differ- trode. This technique is called controlled cath-
ing in standard potentials by about 0.15 V or ode potential electrolysis. Physical Properties of
greater can, theoretically, be separated quanti- Experimental details for performing a con- Electrolytic Precipitates
tatively by electrodeposition, provided their trolled cathode potential electrolysis are pr~ Ideally, an electrOlytic deposit should be
initial concentrations are about the same. Cor- sented in a later section. For the present, it ~ s~ronglY adherent, dense, and smooth so that
respondingly, about 0.3- and O.l-V differences
are required for univalent and trivalent ions,
sufficient to note that the potential difference
between the reference electrode and the cath-
~ 1.6

~
-0.8 f ~ e processes of washing, drying, and weigh-
mg can .be pe~ormed without mechanical loss
respectively. ode is measured with a potentiometer or elec- 81.2 8- ~r reactl?n WIth the atmosphere. Good metal-
An approach to these theoretical separa- tronic voltmeter. The potential applied· be- -0.6 :~
lICdepoSIts are fine-grained and have a metallic
tion values, with a reasonable electrolysis tween the working electrodes is controlled with ~ uste~; spongy, pOWdery, or flaky precipitates
-0.4 «
period, requires a more sophisticated technique a voltage divider so that the cathode potential are hkely to be less pure and less adh t
than the ones thus far discussed because con- is maintained at a level suitable for the separa- Th " eren.
o e prmclpal factors that influence the
centration polarization at the cathode, if un- tion. Figure 20-4 is a schematic diagram of an phYSIcal characteristics of deposits include
checked, will prevent all but the crudest of apparatus that would permit deposition at a current den~ity, temperature, and the presence
separations. The change in cathode potential constant cathode potentiaL of co?,plexmg agents. Ordinarily, the best
is governed by the decrease in IR drop (Figure An apparatus of the type shown in Figure deposIts are fonned at current densities that
2O-2b). Thus, where relatively.large currents 20-4 can be operated at relatively high initial are less than 0.1 A/cm2• Stirring generally im-
~IGURE 20-6 Changes in applied poten-
are employed at the outset, the change in applied potentials to give high currents. As proves the quality of a deposit. The effects of
tial an~ current during a controlled cathode
cathode potential can ultimately be expected the electrolysis progresses, however, a lower- temper~ture are unpredictable and must ·be
potentllli electrolysis. Deposition of COpper
to be large. On the other hand, if the cell is ing of the applied potential across AC is determmed empirically.
upon a cathode maintained at -0.36 V vs
operated at low current levels so that the required. This decrease, in turn, diminishes It is also found that many metals form
SCE. (Experimental data from J. J. Lingane:
variation in cathode potential is lessened, the the current. Completion of the electrolysis smoo~her and more adherent films when
A~/: Chimo Acta, 2, ~89 (1949). With per-
time required for completion of the deposition will be indicated by the approach of the cur- mISSIon.] , de!",slt~ from solutions in which their ions
may become prohibitively long. An obvious rent to zero. The changes that occur in a typi- eXIst p~lmarily as complexes. Cyanide and
answer to this dilemma is to initiate the elec- cal constant cathode potential electrolysis are amm~D1a complexes often provide the best
trolysis with an applied cell potential that is depicted .in Figure 20-5. In contrast to the dep?slts. The reasons for this effect are not
sufficiently high to ensure a reasonable current; electr?lytlc methods described earlier, this obVIOUS.
the applied potential is then continuously de- de source
techm~ue demands constant attention during . ~od~position of hydrogen during electroly-
creased to keep the cathode potential at the -1111 + operatIon. Usually, some provision is made SIS IS hkely to cause the formation of non-
level necessary to accomplish the desired sepa- ~orauto~tic control; otherwise, the operator adher~t deposits, which are unsatisfactory for
ration. Unfortunately, it is not feasible to time requIred represents a major disadvantage analytIcal p~rposes. ~e evolution of hydrogen
predict the required changes in applied poten- to the controlled cathode potential method. can be aVOIded by mtroduction of a cathode
tial on a theoretical basis because of uncertain- depolarizer-a substance that is reduced in
ties in variables affecting the deposition, such prefe~ence to hydrogen ion. Nitrate functions
as overvoltage effects and perhaps conductivity ELECTROGRAVIMETRIC ~n thIS manner, being reduced to ammonium
changes. Nor, indeed, does it help to measure o METHODS OF ANALYSIS Ion
the potential across the working electrodes,
since such a measurement gives only the over- Electrolytic precipitati~n has been used for
all cell potential E.pp1' The altemative is to ~ver a century for the gravimetric determina-
measure the cathode potential against a third ~Ion of ~etals. In most I/-pplications, the metal Instrumentation
electrode whose potential in the solution is IS depos~ted on a. weigijed platinum cathode,
known and constant-that is, a reference elec- ~he appa~atus for an analytical electrodeposi-
and the mcre~e m weight is determined. Im-
trode. The potential impressed across the FIGURE 20-4 Apparatus for electrolysis tlOn COnsISts of a suitable cell and a direct-
portant ~XceptIO~ to this procedure include
working electrodes can then be adjusted to the at a controlled cathode potential. Contact C the an~IC deposItions of lead as lead dioxide
cUITentpowersupp~.
level that will impart the desired potential to is continuously adjusted to maintain the cath- Cells. Figure 20-6 shows a typical cell for
on ~latmum and chloride as silver chloride
the cathode with respect to the reference elec- ode potential at the desired leveL on silver. the deposition of a metal on a solid electrode.
Tall-form beakers are ordinarily employed,
analysis. For example, copper, nickel, cobalt,
silver, and cadmium are readily separated
from ions such as aluminum, titanium, the
I
alkali metals, and phosphates. The precipi- •......• /'
tated elements dissolve in the mercury; little Booster
omplifier
hydrogen evolution occurs even at high applied
potentials because of large overvoltage effects.
Ordinarily, no attempt is made to determine
the elements deposited in the mercury, the
goal being simply their removal from solution.
A cell such as that shown in Figure 20-7 is used.
Power Supplies. The apparatus shown in
Figure 20-6 is typical of that employed for
most electrolytic analyses. The dc power
supply may consist of a storage battery, a
generator, or an alternating-current rectifier
(p. 53). A rheostat is used to control the
cathode (or anode) pOtential (see also p. 63~
applied potential; an ammeter and a volt-
A voltage E1 is impressed on the noninverting
meter are provided to indicate the approxi- EI = E. - E••• -IIR1
terminal of the operational amplifier, the out-
mate current and applied voltage.
put of which is fed into a booster amplifier, where RI is the internal resistance of the cell.
Figure 20-8 shows an apparatus suitable
for carrying out a deposition at a controlled
which provides the large current needed for If the potential of the cathode E. begins to
FIGURE 20-6 Apparatus for the electrolysis. The saturated calomel elec- rise, as a consequence of concentration polar-
electrodeposition of metals. trode is in the feedback circuit to the opera- ization or an increase in the cell resistance,
tional amplifier and controls the current output the amplifier responds by decreasing its output
Wire II to the anode. The cathode is then connected current II until EI again equals E2•
.node to the common ground. The booster amplifier shown in Figure
+ In order to understand the cathode poten- 20-8 is a noninverting type that provides a
and mechanical stirring is provided to mini- tial control exercised by this system, consider much larger current than can be obtained
mize concentration polarization; frequently, first the operation of the circuit in the absence from the single operational amplifier. Its
the anode is rotated with an electric motor. of the booster amplifier. Note that the input presence has no elTecton the cathode control
Electrodes. Electrodes are usually con- EI is to the noninverting terminal of amplifier circuit. .
structed of platinum, although copper, brass, 1. On page 62, it was shown that in a non- An electronic instrument, such as that
and other metals find occasional use. Plat- inverting circuit of this type the potentials at shown in Figure 20-8 is called a potentiostat.
inum electrodes have the advantage of being the two inputs are' always equal; that is, Several instrument suppliers olTer poten-
relatively nonreactive; moreover, they can be EI = E2• Thus, the potential applied to the tiostats for sale.
ignited to remove any grease, organic matter, cell consisting of the calomel electrode and
or gases that could have a deleterious elTect the working cathode is EI. That is,
on the physical properties of the deposit. Cer- Applications of
tain metals (notably bismuth, zinc, and E2 = EI = E. - ESCE -12R2;: E. - ESCE Constant-Current Electrolysis
gallium) cannot be deposited directly onto where E. and ESCE are the reversible poten- Without control ofthe cathode potential, elec-
platinum without causing permanent damage; tials for the two electrodes. Because the inter- trolytic methods suITerfrom a lack of specific-
a protective coating of copper should always nal resistance of the amplifier R2 is high, ity. Despite this limitation, several applica-
be deposited on a platinum electrode before tions of practical importanCe make use of this
current 12 (and thus 12 R2) will be negligible,
the electrolysis of these metals is undertaken. relatively unrefined technique. In general, the
as indicated.
1be Mercury CatbolIe. A mercury cathode FIGURE 20-7 Mercury cathode for the
The potential for the cell consisting of the analyte must be the only component in the
is particularly useful in removing easily electrolytic removal of metal ions from
auxiliary electrode and the cathode is given solution that is more readily reduced than the
reduced elements as a preliminary step in an solution.
trate solution. Copper. is first reduced quanti- more convenient than the former. Finally, coulomb is that amount of electricity that flows
hydrogen ion. Any interfering species should coulometric procedures are readily adapted to
tatively by maintaining the cathode potential during the passage of a constant current of
be eliminated by a chemical precipitation or
at -0.2 V with respect to a saturated calomel automation.5 one ampere for one second. Thus, for a constant
prevented from depositing by complexation
electrode. After weighing, the copper-p'ated current of I amperes flowing for t seconds,
with a ligand that does not appreciably in-
cathode is returned to the solution, and. bis- Units for Quantity of Electricity the number of coulombs Q is given by the
fluence the electrochemical behavior of the
muth is removed at a potential of -0.4 V. expression
analyte. The quantity of electricity is measured in units
Constant-current deposition with a mer- Lead is then deposited quantitatively by in-
of the coulomn (C) and the faradoy (F). The Q = It (20-5)
cury cathode is also useful in removing easily creasing the cathode potential to -0.6 V.
Throughout these depositions, the tin is re- For a variable current, the number of cou-
reduced ions from solution prior to com-
tained in solution as a very stable tartrate lombs is given by the integral
pletion of an analysis by some other method.
The deposition of interfering heavy metals complex. Acidification of the solution after • For summaries of coulometric methods, see: H. L.
deposition of the lead is sufficient to decom- Kies, J. ElectrOQlllJI.Chem, 4. 251 (1962); J. J. Linganc, Q= fIdt
prior to the quantitative determination of Electroanalytical Chemistry. 2d ed. New York: Inter-
pose the complex by converting tartrate ion o
the alkali metals is an example of this science. 1958. Chapiers 19-21; G. w. C. Milner and G.
application. to the undissociated acid; tin can then be Phillips. Coulometry.in Analytical Chemistry. New York: Thefaraday is the quantity of electricity that
Table 20-1 lists the common elements that readily deposited at a potential of -0.65 V. Perpmon Press, 1961; and J. T. Stock, Anal. Chem.,50, will produce one equivalent of chemical change
can be determined by electrogravimetric pro- This method can be extended to include zinc lR (1918); •••• lR (1916~ at an electrode. Since the equivalent in an
cedures for which control over the cathode and cadmium as well. Here, the solution is
potential is not required. made ammoniacal after removal of the copper,
bismuth, and lead. Cadmium and zinc are then
successively deposited and weighed. Finally, TABLE 20-' COMMON ELEMENTS THAT CAN BE
the tin is determined after acidification, as ~ETERMINED BY ELECTROGRAVIMETRIC METHODS
Application of
Controlled Electrode before.
Weighed As
Potential Electrolysis3 A procedure such as this is particularly
attractive for use with a potentiostat because CdH Cd Alkaline cyanide solution
The controlled cathode potential method is a of the small operator time required for the Co2 + Co Ammoniacal sulfate solution
potent tool for the direct analysis of solutions complete analysis. CuH Cu HN03IH2S04 solution
containing a mixture of metallic elements. Table 20-2 lists some other separations Fe3+ Fe (NH4"hC204 solution
Such control permits quantitative separation that have been performed by the controlled PbH Pb02 HN03 solution
of elements with standard potentials that cathode method. NiH Ni Ammoniacal sulfate solution
differ by only a few tenths of a volt. For
example, Ungane and Jones4 developed a
Ag+ Ag Cyanide solution
SnH Sn (NH4hC204IH2C204 solution
method for the successive determination of COULOMETRIC ZnH Zn Ammoniacal or strong NaOH solution
copper, bismuth, lead, and tin. The first three METHODS OF ANALYSIS
can be deposited from a nearly neutral tar-
Coulometry encompasses a group of analyti-
cal methods which involve measuring the TABLE 20-2 SOME APPLICATIONS OF CONTROLLED
quantity of electricity (in coulombs) needed CATHODE POTENTIAL ELECTROLYSIS
3 This method was first suggested by H. J. S. Sand,
to convert the analyte quantitatively to a dif-
TraIlS. Chem. Soc~ 91. 373 (1901). For many of its appli-
cations. see: H. J. S. Sand, Electrocllmtiftry and Electro- ferent oxidation state. In common with the Element Determined Other Elements That May Be Present
chemical AlI<I1ysioS.Glasgow: Blackie '" Son, Ltd, 1940. gravimetric method, coulometry offers the
~g Cu and heavy metals
voL 2. An excellent dilc:ussionof applicatiOlll or automat- advantage that the proportionality constant
ic c:ootrol to the method can be found in J. J. Linganc,
Cu B~ Sb, Pb, So, N~ Cd, Zn
between coulombs and the weight of a,palyte Bi Cu, Pb, Zn, Sb, Cd, Sn
Electroall<l1ytical ChemJmoy, 2d ed. New York: Inter-
can be derived from known physical constants; Sb Pb,Sn
science, 1958, Chapten 13-16. See also: G. A. Rccbnitz,
thus, a calibration or standardization step is Sn Cd, Zn, Mn, Fe
Controlled-Potential AlI<I1ysis.New York: MaaniIIan, Cd, Sn, Ni, Zn, Mn, AI, Fe
not ordinarily required. Coulometric methods Pb
1963. Cd Zn
• J. J. Lingane and S. L. Jones, Ana/. C~ 13, 1198
are often as accurate as gravimetric or volu- Ni Zn, AI, Fe
(1951~
metric procedures; they are usually faster and
The current required to complete the oxida- (H3As04) at a platinum anode. Similarly, the
oxidation-reduction reaction corresponds to Types of
. tion of iron(II) would thett exceed that de- analytical conversion of iron(II) to iron(III)
the change brought about by one mole of elec-
Coulometric Methods
manded by theory. To avoid the consequent can be accomplished with suitable control of
trons, the faraday is equal to 6.02 x IOz3 elec- Two general techniques are used for C?ul~~et- error, an unmeasured excess of cerium(III) can the anode potential.
trons. One faraday is also equal to 96,491 C. ric analyses. The first involves mamta1nmg be introduced at the start of the electrolysis. AppantuL A constant electrode potential
the potential of the working electrode at a This ion is oxidized at a lower anode potential coulometric analysis requires a potentiostat
constant level such that quantitative oxida- than is water: such as that shown in Figure 20-8, an instru-
EXAMPLE tion or reduction of the analyte occurs with- ment for measuring current as a function of
A constant current of 0.800 A was passed out involvement of less reactive species in the time, and ail integrating device for deriving
through a solution for 15.2 min. Calculate the sample or solvent. Here, the current is initially The cerium(IV) produced diffuses rapidly the quantity of electricity by means of Equa-
grams of copper deposited at the cathode and high but decreases rapidly and approaches from the electrode surface, where it then oxi- tion 20-6.
the grams of Oz evolved at the anode, assum- zero as the analyte is removed from the solu- dizes an equivalent amount of iron (II): In early work, chemica/ cou/orneters, placed
ing that these are the only products .formed. tion (see Figure 20-5). The qu:mtity of e~ec- in series with the working cel~ were employed
The equivalent weights are determmed from tricity required is measured WIth a chemIcal ee4+ + Fe2+ -+ ee3+ + Fe3+
as current-tiine integrators. Here, the quantity
consideration of the two half-reactions coulometer or by integration of the current- The net effect is an electrochemical oxidation of electricity was determined by a volumetric
time curve. A second coulometric technique of iron(II) with 100% current efficiency even or gravimetpc measurement of the extent of
Cu2+ + 2e -+ Cu(s) makes use of a constant current that is con- though only a fraction of the iron(II) ions are chemical change occurring in the coulometer
tinued until an indicator signals completion
2HzO -+ 4e + Oz(g) + 4H+ directly oxidized at the electrode surface. during the electrolysis. An example is the
of the reaction. The quantity of electricity re- The coulometric determination of chloride hydrogen/oxygen coulometer, which consists
quired to attain the end point is then calcul~ted provides another example of an indirect pro- of a pair of platinum electrodes immersed in a
From Equation 20-5, we find
from the magnitude ofthe current and the tIme cess. Here, a silver electrode serVes as the anode solution of potassium sulfate; the oxygen and
Quantity of electricity = 0.800 A of its passage. The latter method has. ~joyed and silver ions are produced by the current. hydrogen eyolved at the two electrodes are
wider application than the former; It IS fre- These cations diffuse into the solution and pre- collected in ia gas buret and their combined
x 15.2 min x 60 s/min
quently called a cou/ornetric titration. cipitate the chloride. A current efficiency of volume is m'easured. .
= 729.6 A s A fundamental requirement of all coulo- 100% with respect to the chloride ion is Modern integrators are generally elec-
metric methods is that the species determined achieved even though this ion is neither oxi- tronic and employ a circuit similar to that
= 729.6 C
interact with 100% current efficiency. That is, dized nor reduced in the cell. shown in Figure 3-26 (p. 66). Often such
each faraday of electricity must bring about a integrators are a part of a recorder, wli"lCh
chemical change corresponding to one equiv- provides a plot of current as a function of
alent of the analyte. This requirement does Coulometric Methods at
time. Several controlled potential coulometrlc
729.6 C _ 7.56 X 10-3 F not, however, imply that the analyte must Constllnt Electrode Pf;ttentlal
instruments are available from commercial
96,491 C/F necessarily participate directly in the electron- In variable current coulometry, the potential sources.
transfer process at the electrode. Indeed, more of the working electrode is maintained at a ApplicatiOlL6 Controlled potential coulo-
From the definition of the faraday, often than not, the subStance being determined constant level that will cause the analyte to metric methods have been widely used for the
7.56 x 10-3 equivalent of copper is deposited is involved wholly or in part in a reaction react quantitatively without involvement of determination of various metal ions. Mercury
on the cathode; a similar quantity of oxygen that is secondary to the electrode reaction. other components in the sample. A constant appears to be favored as the cathode and
is evolved at the anode. Therefore, For example, at the outset of the oxidation electrode potential or variable current analysis methods for the deposition of two dozen
of iron(I1) at a platinum anode, all of the of this kind possesses all the advantages of an metals at this electrode have been described.
3 63.5 g Cu/mol
wt Cu = 7.56 X 10- eq Cu x 2 eq Cu/mol current results from the reaction electrogravimetric method and is not subject The method has found widespread use in the
Fe2+ ~Fe3+ +e to the limitation imposed by the need for a nuclear enq-gy field for the relatively inter-
= 0.240 g weighable product. The technique can there- ference-free*nalysisofuranium and plutonium.
As the concentration of iron(I1) decreases, fore be applied to systems that yield deposits
however, concentration polarization may with poor physical properties as well as to
cause the anode potential to rise until de- reactions that yield no solid product at aIL
3 32.0 g Oz/mol composition of water occurs as a competing For example, arsCmc may be determined • For a IlIJIIIIWy or the applicatiolll, ICe: J. E. Harrar,
wt Oz = 7.56 x 10- eq Oz x 4 eq Oz/mol process. That is, coulometrically by the electrolytic oxidation ElectrOtUUJlyticalC"-isrry, ed. A. J. Bard. New York:
of arsenous acid (H3As03) to arsenic acid Marcel Dekker, 1975, vol. 8.
= 0.0605 g 2HzO~Oz(g) + 4H+ + 4e
example of the latter is the coulometric oxida- Elec:tricaIApparatus. Coulometric titrators
The controlled· potential coulometric pro- tion of iron(II}-in part by electrolytically are available from several laboratory supply
cedure also offers possibilities for .the elec- generated cerium(IV) and in part by direct houses. In addition, they can be readily as-
trolytic determination (and synthesIS)of or- electrodereaction (p. 592).Under any cirCum- sembled from components available in most
ganic compounds. For example, Me~~ and stance, the net process must approach 100% laboratories.
Meites7 have demonstrated that ~~oro- current efficiency with respCct to a single Figure 20-10 depicts the principal compo-
acetic acid and picric acid are quantitatIvely chemicalchange in the analyte. . nents ofa typicalcoulometrictitrator. Included
reduced at a mercury cathode whose poten- The current in a coulometric titration is are a source of constant current and a switch
tial is suitably controlled: carefully maintained at a constant and ac- that simultaneously initiates the current and
CI3CCOO- + H+ + 2e = curately known level; the product of this cur- starts an electric stopclock. Also required is
rent in amperes and the time in seconds the means for accurately measuring the cur-
Cl2HCCOO- + Cl- required to reach an end point yields the rent; in Figure 20-10,the potential drop across
number of coulombs and thus the number of the standard resistor, Rotd is used for this
equivalents involved in the electrolysis. The measurement.
constant current aspect of this operation pre- Many electronic or electromechanical con-
cludes the quantitative oxidation or reduction stant current sources are described in the
FIGURE 20-9 An instrument for continu- of the unknown speciesentirely at the genera- literature. The ready availability of inexpen-
ouslyrecordingthe O2 content of a gas stream. tor electrode; as the solution is .depleted of sive operational amplifiers makes their con-
analyte, concentration polarization is inevi- struction a relatively simple matter (for
OH table. The electrode potential must then rise if example, see Figure 3-24, p. 64).
a constant current is to be maintained An ordinary electric stopclock is inade-

=
NH2-(rNH2
Y +6H20 The anode is a heavy cadmitlm sheet; here
the half-cellreaction is
(p. 585).Unless this potential rise produces a
reagent that can react with the analyte, the
current efficiencywill be less than 100%.In a
quate for determining the total electrolysis
time because the motor tends to coast when
stopped and lag when started; the result-
NH2 Cd(s) + 20U- ~Cd(OHh(s) + 2e coulometric titration, then, at least part (and ing accumulated error can be appreciable.
frequently all) of the analytical reaction Stopclocks with solenoid operated brakes or
Coulometric measurements permit the analy- Note that a galvaniccell is formed so that no occurs away from the surface of the working electronic timers eliminate this problem.
sis of these compounds with an accuracy of a external power supply is required. The elec- electrode. It is useful to point out the close analogy
tricity produced is passed through a standard A coulometric titration, like a more con· between the various components of the
few tenths of a percent
Variable current coulometrlc methods are resistor and the potential drop is recorded on ventional volumetric titration, requires some apparatus shown in Figure 20-10 and the
frequently used to monitor CC?ntinuously~ a millivolt recorder. The oxygen concent~- meansofdetectingthe point of chemicalequiv- apparatus and solutions employed in a con-
automatically the concentration of COnstitu- tion is proportional to the recorded potential, alence. Most·of the end points applicable to ventional volumetric analysis. The constant
ents in gas or liquid streams. An important and the chart paper can be made to display volumetric analysis are equally satisfactory current source of known magnitude serves
example is the determination of small concen- the instantaneous oxygen concentration di- here; color changes of indicators, potentio- the same function as the standard solution in
trations of oxygen.' A schematic diagram of rectly. The instrument is reported to provide metric, amperometric (p. 635), and conduct- a volumetric method. The electric clock and
the apparatus is shown in Figure 20-9. The oxygen concentration data in the range from ance measurements have all been successfully switch correspond closely to the buret, the
porous silver cathode serves to break up the 1 ppm to 1%. applied. switch performingthe same function as a stop-
incoming gas into small bubbles; the reduc- The analogy between a volumetric and a cock. During the early phases of a coulometric
tion of oxygen takes place quantitatively Coulometric Titration. ; coulometric titration extends well beyond the titration, the switchis kept closed for extended
within the pores. That is, A coulometric titration emp,loysan. elect~o- common requirement of an observable end periods; as the end point is approached, how-
lytically generated titrant fer reactl~n With point. In both, the amount of analyte is deter- ever, small additions of" reagent" are achieved
02(g) + 2H20 + 4e~40H- mined through evaluation of its combining by closing the switch for shorter and shorter
the analyte. In some analyses, the ~Ive elec-
trode process involves only generation of ~he capacity-in the one case, for a standard solu- intervals. The similarity to the operation of a
reagent; an example is the titra~on of.a halide tion and. in the other, for a quantity of buret is obvious.
, T. Meitel aDd L. MeiteI, AIIGI. C••.• 27, IS31 (19SS); by electrolytically generated silver lon~. In electricity. Similar demands are made of the Cells for CouIomeCricTitratioas. A typical
Zll, 103 (19S6~ other titrations, the analyte may also be direct- reactions; that is, they must be rapid, essen- coulometric titration cell is shown in Figure
• For further delaiIs, see: F. A. Keidel, lrt4. Eng. CIImI.
ly involved at the generator electrode; an tially complete, and free of side reactions. 20-11. It consists of a generator electrode at
51, 491 (1960~
Electrolyte solution
from reservoir
The cells shown in Figures 20-11 and 20-12
can be employed. A convenient alternative in-
~ Ahode reaction
volves substitution of a silver wire as the
H20-~02 + 2W + 2e
anode and the addition of chloride or bro-
mide ions to the solution of the analyte. The
anode reaction then becomes
Potentiometer
for current
measurement
Ag + Br(sr ~AgBr(s) +e
Clearly, the silver bromide will not interfere
with the neutralization reaction as would the
hydrogen ions that are formed at most
J
Source
anodes.
Both potentiometric and indicator end
olOW
points can be· employed for these titrations.
FIGURE 20-12 A cell for external genera- The problems associated with the estimation
tion of acid and base. of the equivalence point are identical with
those encount~red in a conventional volumet-
ric analysis. A real advantage to the coulo-
the reagent externally. The apparatus is so metric method, however, is that interference
arranged that flow of the electrolyte continues by carbonate ion is far less troublesome; it is
briefly after the current is discOntinued, thus only necessary to eliminate carbon dioxide
FIGURE 20-10 Schematic diagram of a coulometric flushing the residual reagent into the titration from the solu~ion containing the analyte by
titration apparatus. vessel. Note that the apparatus shown in aeration with a carbon-dioxide-free gas before
Figure 20-12 provides either hydrogen or beginning the analysis. .
hydroxide ions depending upon which arm is The coulometric titration of strong and
used. The apparatus has also been used for weak bases can be performed with hydrogen
which the reagent is formed and an auxiliary generation of other reagents such as iodine ions generated at a platinum anode.
electrode to complete the circuit. The genera- produced by oxidation of iodide at the anode.
tor electrode, which should have a relatively
H20~!02(g) + 2H+ + 2e
large surface area, is often a rectangular strip Application. of Here, the cathode must be isolated from the
or a wire coil of platinum; a gauze electrode Coulometric Titrations9• solution or external generation must be em-
such as that shown on page 588 can also be Coulometric titrations have been developed ployed to prevent interference from the hy-
employed. for all types of volumetric reactions. Selected droxide ions produced at that electrode.
The products formed at the second elec- applications are described in the following Precipitation and Complex-Formation Ti-
trode frequently represent potential sources of paragraphs. trations. A variety of coulometric titrations
interference. For example, the anodic genera- Neutralization TitratiOl& Both weak and involving anodically generated silver ions have
tion of oxidizing agents is often accompanied strong acids can be titrated with a high been developed (see Table 20-3). A cell, such
by the evolution of hydrogen from the cath- degree of accuracy using hydroxide ions gen- as that shown ip Figure 20-11, can be employed
ode; unless this gas is allowed to escape from erated by the reaction with a generator electrode constructed from a
the solution, reaction with the oxidizing agent length of heavy silver wire. End points are
becomes a likelihood. To eliminate this type 2H20 + 2e ..• 20H- + H2(g) detected potentiometrically or with chemical
of difficulty, the second electrode is isolated indicators. Similar analyses, based upon the
by a sintered disk or some other porous generation of mercury(I) ion at a mercury
• Applications or the coulometric procedure are sum- anode, have been described.
medium. mariud in J. J. Lingane, Eleetroanalyrical Chemistry. 2d
As an alternative to isolation of the auxil- ed. New York: Interscience, 1958. pp. 53~13; H. L
An interesting coulometric titration makes
iary electrode, a device such as that shown in Kies. J. Eleetroanal. Chem .•4.257 (t962); and J. T. Stock. use of a solution of the ammine mercury(lI)
FIGURE 20-11
Figure 20-12 may be employed to generate Anal. Chem.5O. tR (1978); •••• lR (t976~ complex of ethylenediaminetetraacetic acid
titration cell.
(14y).10 The complexing agent is re~eased to positive manganese, and the chloride complex reduction, or neutralization reagents. Further- magnitude of uncertainty associated with the
the solution as a result of the followmg reac- of unipositive copper. more, the coulometric method is readily adapt- electrode process are difficult, current effi-·
tion at a mercury cathode: Comparison of Coulometrie •••• Volumetric ed to automatic titrations, because current ciencies of 99.5 to better than 99.9% are often
Titratiom. Some real advantages can be control is easily accomplished. reported in the literature: Currents are readily
HgNH3Yz- + 2e
+ NHt claimed for a coulometric titration in compar- Coulometric titrations are subject to five measured to ±0.1 % relative or better. Simi-
~Hg + 2NH3 + HY3- (20-7) ison with the classical volumetric process. potential sOurces of error: (lfvariation in the larly, a good quality timer permits measure-
Principal among these is the elimination of current during electrolysis; (2) departure of ment of time to within ±0.1 % relative.
Because the mercury chelate is more stable problems associated with the preparation, the process from 100% current efficiency; (3) To summarize, then, the current-time mea-
than the corresponding complexes with cal- standardization, and storage of standard solu- error in the measurement of current; (4) error surements required for a coulometric titration
cium, zinc, lead, or copper, complexation of tions. This advantage is particularly impor- in the measurement of time; and (5) titration are inherently as accurate or more accurate
these ions will not occur until the electrode tant with labile reagents such as chlorine, error due to the difference between the equiv- than the comparable volume-normality mea-
process frees -the ligand. bromine, or titanium(III) ion; owing to their alence point and the end point. The last of surements of a classical volumetric analysis,
Oxidation-Reduction Titratiom. Table 20-4 these difficulties is common to volumetric particularly where small quantities of reagent
instability, these species are inconvenient as
indicates the" variety of reagents that can be volumetric reagents. Their utilization in cou- methods as well; where the indicator error is are involved. Often, however, the accuracy of
generated coulometrically and the analyses lometric analysis is straightforward, however, the limiting factor, the two methods are likely a titration is not limited by these measurements
to which they have been applied. Electro- because they undergo reaction with the to have comparable reliability. but by the sensitivity of the end point; in this
generated bromine has proved to be partic- analyte immediately after being generated. With simple instrumentation, currents con- respect, the two procedures are equivalent.
ularly useful among the oxidizing agents and Where small quantities of reagent are stant to 0.2% relative are easily achieved;
forms the basis for a host of methods. Of required, a coulometric titration olTers a con- with somewhat more sophisticated apparatus,
interest also are some of the unusual reagents Automatic Coulometric Titrators
siderable advantage. By proper choice of control to 0.01 % is obtainable. In general,
not ordinarjly encountered in volumetric then, errors due to current fluctuations are A number of instrument manufacturers olTer
current, micro quantities of a substance can
analysis because of the instability of their so- seldom of importance. automatic coulometric titrators. Most of these
be introduced with ease and accuracy; the
lutions; these include dipositive silver ion, tri- Although generalizations concerning the employ the potentiometric end point and are
equivalent volumetric process requires small
volumes of very dilute solutions, a recourse
that is always difficult.
10 C. N. Reilley and W. W. Porterfield, Anal.
443 (19S6~
e,,-. %8, A single constant-current source can be
employed to generate precipitation, oxidation- TABLE 20-4 SUMMARY OF APPLICATIONS OF COULOMETRIC
TITRATIONS INVOLVING OXIDATION-REDUCTION REACTIONS

Generator Electrode Reaction

2Br- ~Brz + 2e As(III~ Sb(III~ U(IV), Tl(I~


TABLE 20-3 SUMMARY OF APPLICATIONS OF COULOMETRIC
1-, SCN-, NH3, NzH4, NHzOH,
TITRATIONS INVOLVING NEUTRALIZATION, PRECIPITATION,
phenol, aniline, mustard gas;
AND COMPLEX-FORMATION REACTIONS
8-hydroxyquinoline
Species DetenniDed Generator-Electrode Reaction Secondary Analytical Reaction Clz 2Cl- ~Clz + 2e As(III~ r
2HzO + 2e~20H- + Hz OH- +H+~HzO Iz 2r~lz +2e As(III~ Sb(III~ SzO~ -, HzS
Acids
Ce4+ Ce3+ ~Ce4+ + e Fe(lI~ Ti(III~ U(IV~ As(III~
Bases HzO~2H+ +!<>z + 2e H+ +OH-~HzO
r, Fe(CN):-
Cl-, Br-,r Ag~Ag+ +e Ag+ + Cl- ~AgCl(sb etc.
Mn3+ Mn2+ ~ Mn3+ + e HZCZ04, Fe(lI~ As(III)
Mercap~ Ag~Ag+ +e Ag+ + RSH~~ + H+
Agz+ Ag+~Ag2+ + e Ce(III~ V(IV~ HZCZ04, As(III)
CI-, Br-,'- 2Hg~H~+ +2e H~+ + 2Cl- ~HgzClz(s1 etc.
Fe2+ Fe3+ + e~Fez+ Cr(VI~ Mn(VI11 V(v1 Ce(IV)
Znz+ Fe(CN):- + e~Fe(CN):- 3Zn2+ + 2K + + 2Fe(CN):- ~
Ti3+ Ti02+ + 2H+ + e~Ti3+ + HzO Fe(III1 V(v1 Ce(IV1 U(VI)
KZZn3[Fe(CN)Jz(s)
Caz+, Cuz+, See Equation 20-7 Hy3- + Ca2+ ~CayZ- + H+, etc. CuCl~- eu2+ + 3Cl- + e~CuCl~- V(v1 Cr(V11 10;
Zn2+ and Pb2+ U4+ UO~+ + 4H+ + 2e~U4+ + 2HzO Cr(VI1 Ce(IV)
similar in construction to the automatic ti- are designed for a single analysis. Examples of (a) Using 1.(0)( 1O-6M as the criterion for quantitative removal,
trators discussed in the' previous chapter the latter include chloride titrators in which determine whether or not separation of the three species is feasible
(p. 573). Here, however, the error signal con- silver ion is generated coulometrically, sulfur by controlled cathode potential electrolysis.
trols a flow of electricity rather than a flow dioxide monitors, where anodically generated (b) If separations are feasible, evaluate the range (vs. SCE) within
of liquid reagent_, Some of the commercial in- bromine oxidizes the analyte to sulfate ions, , which the cathode potential should be controlled for the deposi-
struments are multipurpose and can be used for and water titrators in which Karl Fischer re- tion of each.
the determination of a variety of species. Others agent is generated electrolytically. (c) Calculate the potential needed to deposit the third ion quantita-
tively after removal of the first two.
5. Halide ions can be deposited at a silver anode, the reaction being
Ag(s) + X- ...•AgX(s) +e
(a) Determine whether or not it is theoretically feasible to separate'-
1. Bismuth is to be deposited at a: cathode from a solution that is and Br- ions from a solution that is O.04OOMin each ion by
0.15OM in BiO+ and O.600F in HCI04• Oxygen is evolved at a pres- controlling the potential of the silver anode. Take 1.00 )( 1O-6M
sure of 0.800 atm at a 2O-cm2 platinum anode. The cell has a resist- as the criterion for quantitative removal of one ion.
ance of 1.30 n. (b) Is a separation of Cl- and I - theoretically feasible?
(a) Calculate the thermodynamic potential of the cell. (c) If a separation is feasible in either (a) or (b~ what range of anode
(b) Calculate the I R drop ifa current of 0.200 A is to be used. potentials (vs. SCE) should be employed?
(c) Estimate the O2 overvoltage. 6. What cathode potential (vs. SCE) would be required to lower the
(d) Estimate the total applied pOtential required to begin electro- total nickel concentration of the following solutions to 1 )( 10-'F:
deposition at the specified conditions. (a) a perchloric acid solution of Ni2+?
(e) What potential will be requirbd when the BiO+ concentration is (b) a solution having an equilibrium CN- concentration of 0.0 100M ?
O.OSOOM? '
Ni(CN)i- + 2e~Ni(s) + 4CW EO = -0.82
2. Nickel is to be deposited at a cathode from a solution that is 0.200M
in Ni2+ and O.400F in HCI04• Oxygen is evolved at a pressure of 7. What cathode potential (vs. SCE) would be required to lower the
0.800 atm at a 15-cm2 platinum anode. The cell has a resistance of Hg2+ concentration of the following solutions to 1.00 )( 10-6F:
2.10 n. (a) an aqueous solution of Hg2+?
(a) Calculate the thermodynamic potential of the cell. (b) a solution with an equilibrium SCN- concentration of 0.100M?
(b) Calculate the IR drop if a current of 0.150 A is to be used.
Hg2+ + 2SCN- ~Hg(SCNh(aq) K1 = 1.8)( 107
(c) Estimate the O2 overvoltage.
(d) Estimate the total applied potential required to begin electro- (c) a solution with an equilibrium Br- concentration of 0.250M?
deposition at the specified conditions.
(e) What potential will be required when the NiH concentration is
HgBr; + 2e~Hg(I) + 4Br- EO = 0.223 V
0.100M? 8. The cadmium and zinc in a '1.06-g sample was dissolved and sub-
3. It is desired to separate and determine bismuth and lead in a solution sequently deposited from an ammoniacal solution with a mercury
that is O.OSOOMin BiO+, O.OSOOMin Pb2+, and I.00F in HCI04• cathode. When the cathode potential was maintained at -0.95 V (vs.
(a) Using 1.00 x 1O-6M as the criterion for quantitative removal, SCE~ only the cadmium deposited. When the current ceased at this
determine whether or not a separation is feasible by controlled potential, a hydrogen/oxygen coulometer in series with the cell was
cathode potential electrolysi~ found to have evolved 44.6 ml of hydrogen plus oxygen (corrected for
(b) If a separation is feasible, c!valuate the range (vs. SCE) within water vapor) at a temperature of 21.0·C and a barometric pressure of
which the cathode potential should be controlled. 173 torr. The potential was raised to about -1.3 V, whereupon zinc
(c) Calculate the potential needed to deposit the second ion quantita- ion was reduced. Upon completion of this electrolysis, an additional
tively after removal of the first. 31.3 ml of gas had been produced under the same conditions.
4. It is desired to separate and determine bismuth, copper, and silver in Calculate the percent Cd and Zn in the ore.
a solution that is O.0800M in BiO+, 0.242M in eu2+, O.l06M in Ag+, 9. A 1.74-g sample of a solid containing BaBr2' KI, and an inert species
and I.00F in HCl04• was dissolved, made ammoniacal, and placed in a cell equipped with a
silver anode. By maintaining the potential at -0.06 V (vs. SCE~ r 14. An apparatus similar to that shown in Figure 20-9 was employed to
was quantitatively precipitated as AgI without interference from Br-. determine the oxygen content of a light hydrocarbon gas stream
The volume of H2 and O2 formed in a gas coulometer in series with having a density of 0.00140 glml. A W.o-Iiter sample of the gas
the cell was 39.7 mI (corrected for water vapor) at 21.7°C and at a consumed 3.13 C of electricity. Calculate the parts per million O2 in
pressure of 748 torr. the sample on a weight basis.
After precipitation of iodide was complete, the solution was acidi- 15. The odorant concentration of household gas can be monitored by
fied, and the Br- was removed from solution as AgBr at a potential passage of a fraction of the gas stream through a solution containing
of 0.016 V. The volume of gas formed under the same conditions was an excess of. bromide ion. Electrogenerated bromine reacts rapidly
23.4 ml. Calculate the percent BaBr 2 and KI in the sample. with the mercaptan odorant:
10. The nitrobenzene in a 210-mg sample of an organic mixture was
reduced to phenylhydroxylamine at a constant potential of -0.96 V
(vs. SCE) applied to a Hg cathode:
Continuous analysis is made possible by means of a potentiometric
C6HsN02 + 4H+ + 4e -+ C6HsNHOH + H20 electrode system that signals the need for additional Br2 to oxidize the
mercaptan. Ordinarily, the current required to react with the odorant
The sample was dissolved in 100 ml of methanol; after electrolysis for is automati~lly plotted as a function of time. Calculate the average
30 miD, the reaction was judged complete. An electronic coulometer odorant conCentration (as ppm C2HsSH) from the following infor-
in series with the cell indicated that the reduction required 26.74 C.
mation:
Calculate the percent nitrobenzene in the sample.
11. At a potential of -1.0 V (vs. SCE), carbon tetrachloride in methanol Average density of the gas 0.00185 glmI
is reduced to chloroform at a Hg cathode: Rate of gas;ftow 9.4Iiters/min
Average current during a 10.o-min sampling period 1.35 mA
2CCl4 + 2H+ + 2e + 2Hg(l) -+ 2CHCl3 + Hg2Ch(s) ,
16. The phenol Content of. water downstream from a coking furnace was
At -1.80 V, the chloroform further reacts to give methane:
determined by coulometric analysis. A 100-ml sample was rendered
2CHCl3 + 6H+ + 6e + 6Hg(l) -+ 2CH4 + 3Hg2C12(s) slightly acidic and an excess of KBr was introduced. To produce Br2
for the reaction
A 0.750-g sample containing CC14• CHCI3, and inert organic
species was dissolved in methanol and electrolyzed at -1.0 V until
the current approached zero. A coulometer indicated that 11.63 Chad
been used. The reduction was then continued at - 1.80 V; an addi- a steady current of 0.0313 A for 7 min and 33 s was required. Express
tional 44.24 C were required to complete the reaction. Calculate the the results of this analysis in terms of parts phenol per million parts of
percent CCl4 and CHCl3 in the mixture. water (assume that the density of water = 1.00 glml).
12. A 0.13OO-gsample containing only CHCl3 and CH2C12 was dissolved 17. The calcium content of a water sample was determined by adding an
in methanol and electrolyzed in a cell equipped with a Hg cathode; excess of an HgNH3 y2- solution to a 25.0-ml sample. The anion of
the potential of the cathode was held constant at -1.80 V (vs. SCE). EDT A was then generated at a mercury cathode (see Equation 20-7,
Both compounds were reduced to CH4 (see Problem 11 for the reac- (p. 598). A constant current of 20.1 mA was needed to reach an end
tion type). Calculate the percent CHCl3 and CH2Cl2 if 306.7 C were point after ~ min and 56 s. Calculate the milligrams of CaC03 per
required to complete the reduction. liter of ~ample.
13. The iron in a 0.854-g sample of ore was converted to the + 2 state by 18. The cyanide concentration in a 10.0-ml sample of a plating solution
suitable treatment and then oxidized quantitatively at a Pt anode was determined by titration with electrogenerated hydrogen ions to a
maintained at -1.0 V (vs. SCE). The quantity of electricity required methyl orange end point. A color change occurred after 3 min and
to complete the oxidation was determined with a chemical coulometer 22 s with a current of 43.4 mA. Calculate the number of grams of
equipped with a platinum anode immersed in an excess of iodide ion. NaCN per liter of solution.
The iodine liberated by the passage of current required 26.3 mI of 19. A 6.39-g sample of an ant-control preparation .w~ decomJ?OSed by
0.0197-N sodium thiosulfate to reach a starch end point. What was wet-ashing with H2SO4 and HN03. The arsemc 1D the residue was
the percentage of Fe304 in the sample? reduced to the trivalent state with hydrazine. After the excess reducing
agent had been removed, the arscnic(III) was oxi~ with electrolyt-
ically generated 12 in a faintly alkaline medium:
HAsO~- + 12 + 2Hoo; :-+ HAsO:- + 21- + 2002 + H20
The titration was complete after a constant current of 98.3 mA had '.
been passed for 13 min and 12 s. Express the results of this analysis in -
terms of the percentage As20~ in the original sample.
20. The H28 content of a water sample was assayed with electrolytically
generated iodine. After 3.00 g of potassium iodide had been introduced
to a 25.Q-ml portion of the water, the titration required a constant
current of 66.4 mA for a total of 7.25 min. Reaction:
H28 + 12 -+ 8(s) + 2H+ + 2.-
Express the concentration of H28 in terms of mg,lliter of sample.
21. A 0.0145-g sample of a purified organic acid was dissolved in an
alcohol/water mixture and titrated with coulomctrically generated
hydroxide ions. With a current of 0.0324 A, 251 s were required to
reach a phenolphthalein end point. Calculate the equivalent weight of
the acid.
22. The chromium deposited on one surface of a 10.o-cm2 test plate was
dissolved by treatment with acid and oxidized to the +6 state with
ammonium peroxodisulfate:
3820;- + 2Cr3+ + 7H20 = Cr20~- + 14H+ + 680:-
The solution was boiled to remove the excess peroxodisulfate, cooled,
and then subjected to coulometric titration with Cu(l) generated from
50 ml of O.lOF CuH. Calculate the weight of chromium that was
deposited on each square centimeter of the test plate if the titration
required a steady current of 325 mA for a period of 7 min and 33 s.
Voltammetry comprises a group of electro- polarographic method are first considered; A Brief Description of
analytical methods in which information about subsequently, the modifications to the original Polarographic Measurements
the analyte is derived from current-voltage method, which have led to the enhancement
curves-that is, plots of current as a function of its usefulness, are presented. Polarographic data are obtained by measur-
of applied potential-obtained under con- ing current as a function ·of the potential
ditions that encourage polarization of the indi- applied to a special type of electrolytic cell. A
cator or working electrode. Generally, the plot of the data gives current-voltage curves,
working electrodes in voltammetry are charac- called polarograms, which provide both quali-
terized by their small surface area (usually a Virtually every element, in one form or another, tative and quantitative information about the
few square millimeters~ which enhances polar- is amenable to polarographic analysis. In ad- composition of the solution in which the elec-
ization. Such electrodes are generally referred dition, the method can be extended to the trodes are immersed.
to as microelectrodes. determination of several organic functional Polarographic Cells. A polarographic cell
Historically, the field of volt am metry devel- groups. Because the polarographic behavior of consists of a small easily polarized microelec-
oped from the discovery of polarography by any species is unique for a given set of experi- trode, a large nonpolarizable reference elec-
the Czechoslovakian chemist Jaroslav Hey- mental conditions, the technique offers attrac- trode, and the solution to be analyzed. The
rovskyl in the early 19205. Polarography, tive possibilities for selective analysis. microelectrode, at which the analytical reac-
which is still the most widely used of all vol~- Most polarographic analyses are per- tion occurs, is a mercury surface with an area
metric methods, differs from the others in the formed in aqueous solution, but other solvent of a few square millimeters; the metal is forced
respect that a dropping mercury electrode systems may be substituted if necessary. For by gravity through a very fine capillary to pro-
Agar plug
(DME) serves as the microelectrode. The quantitative analyses, the optimum concen- vide a continuous flow of identical droplets saturated
unique properties of this electrode are dis- tration range lies between 1O-~ and 10-4M; with a maximum diameter between 0.5 and 1 withKCI

cussed in a later section. through suitable modifications, however, con- mm. The lifetime of a drop is typically 2 to 6 s.
By 1950, voltammetry appeared to be a centration determinations in the parts-per- We shall see that the dropping mercury elec-
mature and fully developed technique. The de- billion range become possible. An analysis trode has properties that make it particularly
cade from 1955 to 1965, however, was marked can be easily performed on 1 to 2 ml of solu- well suited for voltammetric studies. FIGURE 21-1 A dropping mercury elec-
by the appearance of several major modifica- tion; with a little effort, a volume as small as The reference electrode in a polarographic trode and cell. Reprinted with permission from
tioflS of the original method, which served to one drop is sufficient. The polarographic cell is massive relative to the microelectrode J. J. Lingane and H. A. Laitinen, Ind. Eng.
overcome many of its limitations. Following method is thus particularly useful for the so that its behavior remains essentially con- Chem., Anal. Ed., 11, 504 (1939). (Copyright by
this, the advent of low-cost operational ampli- determination of quantities in the milligram stant with the passage of small currents; that the American Chemical Society.)
fiers in the early 19605 made possible the to microgram range. is, it remains unpolarized during the analysis.
development of relatively inexpensive, com- Relative errors ranging between 2 and 3 % A saturated calomel electrode and salt bridge,
mercially available instruments that incor- are to be expected in routine polarographic arranged in the manner shown in Figure 21-1,
porated many of these important modifica- work. is frequently employed; other common refer- Figure 21-2 shows two polarograms. The
tions. The result has been a recent resurgence ence electrodes consist of a large pool of mer- lower one is for a solution that is O.lF in
of interest· in the applications of voltammetric cury or a silver/silver chloride electrode. potassium chloride; the upper one is for a
methods to the qualitative and quantitative • The principles and applications or polarography are Polarograms. A polarogram is a plot of cur- solution that is additionally 1 x 10- 3 F in
determination of a host of organic and in- considered in detail in a number or monographs; see, Cor rent as a function of the potential applied to cadmium chloride. A step-shaped current-
organic species.2 example: t. M. Kalthoff and J. J. Lingane, Polarography, voltage curve, called a polarographic wave, is
a polarographic cell. The microelectrode is or-
2d ed. New York: Interscience, 1952; L. Meites. PolDro-
In the sections that follow, the theory, graphic Techniques, 2d ed. New York: Interscience, 1965; dinarily connected to the negative terminal of produced as a result of the reaction
practice, and limitations of the conventional J. Heyrovsky and J. Kuta, Principles oj PolDrography. the power supply; by convention the applied
New York: Academic Press, 1966; P. Zuman, Topics in
CdH + 2e + Hg~Cd(Hg)
potential is given a negative sign under these
Organic Polarography. New York: Pergamon Press, 1970; circumstances. By convention also, the currents where Cd(Hg) represents elementary cadmium
I J. Hcyrovsky, Cltem. LIMy, 16, 2S6 (1922~ Hcyrovsky
EkctroalllJlytiall ChtmIMry, ed. H. W. Numberg. New dissolved in mercury. Tfie sharp increase in
was awarded the 1959 Nobel Prize in chemistry Corhis are designated as positive when the flow of
York: Wiley, 1974,Chapters 1-5; and O. H MOller, in
discovery and development oC polarOlJ'&Phy. electrons is from the power supply into the current at about - 2 V in both plots is associ-
Physiall Methods oj ChtmIMry, eds. A. Weissberger and
2 For a brieC summary or the new vo\tammetric B. W. Rossitor. New York: Wiley-lnterscience, 1971,Part microelectrode-that is, when that electrode ated with reduction of potassium ions to give
C"-~
techniques, see: J. B. Flalo, AIIIJI. 44, 75A (1972~ IIA, vol. I, Chapter 5. behaves as a cathode. a potassium amalgam.
,-/

diffusion-controlled limiting current is given a Here, the bracketed terms are concentrations We have noted (p. 529) that the ions or
special name, the diffusion cu"ent, and is of the reactant and product, £0 is the stan· molecules in a cell migrate under the influence
assigned the symbol i4• Ordinarily, the dif- dard potential for the half-reaction shown by of diffusion, convection (thermal or mechani-
fusion current is directly proportional to the Equation 21-1, and EDME is the potential of cal1 and electrostatic attraction. In polaro-
concentration of the reactive constituent and is the dropping mercury electrode (vs. SHE~ graphy, every effort is made to eliminate the
0( 10 Diffusion thus of prime importance from the standpoint The subscript zeros are employed to denote latter two effects. Thus, vibration or stirring of
current. of analysis. As shown in Figure 21-2, the dif- that the concentration terms apply to the sur- the solution is avoided, and an excess of a non-
""
~~ 8 id
l!! fusion current is the·difference between the face film only; the concentrations in the bulk reactive electrolyte is employed; if the concen-
~ 6 Decomposition
U potential limiting and the residual currents. of the solution will ordinarily be substantially tration of this supporting electrolyte is 50
One other important quantity, the half-wave different from these surface values. times (or more) greater than that of the reac-
potential, is the potential at which the current If the anode is a saturated calomel elec- tant, the attractive (or repulsive) force between
is equal to one-half the diffusion current. The trode, the applied potential is given by the the electrode and the charged reactant becomes
half-wave potential is usually given the symbol equation negligible.
E1i2; it may permit qualitative identification With the effects of mechanical mixing and
of the reactant. Eapp1 = EDME - ESCE - IR electrostatic attraction eliminated, the only
-0.2 -D.6 -1.0 -1.4 -1.8 -2.2
Applied potential. V In general, polarographic currents are so small force responsible for the transport of Ox to
that IR is negligible with respect to the other the electrode surface is diffusion. Since the
FIGURE 21-2 Polarogram for cadmium two potential terms; so also is the junction rate of diffusion is directly proportional to the
ion. The upper curve is for a solution that is Interpretation of
potential. Under these conditi"ons, substitution concentration (strictly activity) difference be-
1 x 10- 3 F with respect to Cd2 + and·O.1 F Polarographic Waves
of Equation 21-2 yields tween two parts of a solution, we may write
with respect to KCI. The lower curve is for a To account for the s~urce of the typical po-
solution that is 0.1 F in KCI only. ro. = k"([Ox] - [Ox]o)
larographic wave, corisider the following gen- E EO 0.0591 I [Redlo E
eralized half-reaction. at a droppjng electrode: appl = - -n- og [Oxlo - !iCE
where [Ox] is the concentration in the bulk of
(21-3) the solution from which ions are diffusing and
Ox(aq) + ne¢Red(aq) (21-1)
[Ox]o is the concentration in the aqueous film
For reasons to be considered presently, a Consider now what occurs when Eappl is
where Ox and Red are the formulas for the surrounding the electrode. If diffusion is the
polarographic wave suitable for analysis is made more negative than the decomposition
oxidized and reduced form of an electro reactive only process bringing Ox particles to the
obtained only in the presence of a large excess potential (Figure 21-2). Because the half-
substance. In contrast to the previous example, surface, it follows that
of a supporting electrolyte; potassium chloride reaction is reversible, an essentially instanta-
note that Red is not an amalgam; the treatment
serves. this function in the present. example. that follows is readily extended to such a
neous decrease in [Ox]o and a corresponding i = k'ro. = k'k"([Ox] - [Ox]o)
Exammation of the polarogram for the sup- increase in [Red]o takes place until the rela-
reduction as well. = k([Ox] - [Ox]o) (21-4)
porting electrolyte alone reveals that, a small tionship shown by Equation 21-3 is again
It will be assumed that the half-reaction
current, called the residual current, passes satisfied; a momentary surge of current re- Note that [Ox]o becomes smaller as Eappi is
shown by Equation 21-1 is reversible. In the
throdgh the cell even in the absence of cad- sults. This current would rapidly decay to zero made more negative. Thus, the rate of diffu-
context of polarography, reversibility implies
mium ions. were it not for the fact that Ox particles are sion as well as the current increases with in-
that the electron transfer process is suffi-
A characteristic feature of a polarographic mobile in the aqueous medium and migrate to creases in applied potential. Ultimately, the
ciently rapid that the concentrations (more
wave is the region in which the current, after the surface of the mercury electrode as a con- applied potential becomes so negative that
exactly, the activities) of reactants and products
increasing sharply, becomes essentially inde- sequence. Because the reduction reaction is in- the concentration of Ox particles in the sur-
in the liquid film at the interface between the
pendent of the applied voltage; this constant stantaneous, the magnitude of the current de- face film approaches zero with respect to the
solution and the merCury electrode are deter-
current is called a limiting current. The limiting pends upon the rate at which the Ox particles concentration in the bulk of the solution;
mined by the electr04e potential alone; thus,
current is the result of a restriction in the rate nwvefrom the bulk of the solution to the surface under this circumstance, the rate of diffusion
for the reversible reduction of Ox, it may be
at which the participant in tJte electrode pro- where reaction occurs. That is, and thus the current becomes constant. That
assumed that the concentrations of reactant
cess can be brought to the surface of the micro- is, when
and product in the film are given at any instant
electrode; with proper control over experimen- by [Ox]o <c [Ox]
tal conditions, this rate is determined ex-
where i is the current at an applied potential the expression for current becomes
clusively for all points on the wave by the EDME = £0 _ 0.0591 log [Red]o EapPI' ro. is the rate of migration of Ox par-
velocity at which the reactant diffuses. A n [Ox]o (21-2) i.= k[Ox]
ticles, and k' is a proportionality constant.
where i4 is the potential-independent diffusion reactant concentration but directly related to determine the formula and the formation con-
current. Note that the magnitude of the diffu- the standard potential for the half-reaction. In stant for a complex, provided that the cation
i4- i
sion current is directly proportional to the con- [Ox]o=-k- practice, the half-wave potential can be a use- involved reacts reversibly at the dropping elec-
centration of the reactant in the bulk of the ful quantity for identification of the species trode. Thus, fOr the reactions
solution. Quantitative polarography is based The concentration of Red in the surface film responsible for a given polarographic wave.
upon this fact. will be proportional to the current. That is, It is important to note that the half-wave
When the current in a cell is limited by the potential may vary considerably with concen-
. rate at which a reactant can be brought to the i = kR[Red]o tration for electrode reactions that are slow
surface of an electrode, a state of complete con- and thus irreversible; in fact, Equation 21-8
centration polarization is said to exist. With a i no longer describes such waves.
[Red]o = k he derived the relationship
: microelectrode, the current required to achieve R Effect of Complex Formation on Polaro-
this condition is small-typically 3 to 10 pA graphic Waves. We have already seen (Chap- 0.0591
Substitution of these quantities into Equation (EI/2)c - E1/2 = - --log K1
: (microampere) for a 10-3M solution. Such 21-3 gives, upon rearrangement, ter 18) that the potential for the oxidation or n
current levels do not significantly alter the reduction of a metallic ion is greatly affected
reactant concentration, as shown by the follow- 0.0591 k by the presence of species that form complexes 0.0591x I F
EapPI == £0 - ESCE - -- log -k ---n- og A
: ing example. n R with that ion. It is not surprising, therefore,
that similar effects are observed with polaro-
0.0591 i where (EI/2)c is the half-wave potential when
EXAMPLE --- 1 og-- graphic half-wave potentials. The datcdn Table
The diffusion current for a 1.00 x 10- 3M n i4 - i 21-1 indicate that the half-wave potential for the formal concentration of A is FA, while
solution of Zn2 + was found to be 8.4 pA. Cal- By definition, Eappi = E1/2 when the reduction of a metal complex is generally E1/2 is the half-wave potential in the absence
culate the percent decrease in the ZnH con- more negative than that for reduction of the of complexing agent; K 1 is the formation con-
centration after passage of this current for 8.0 . i4 corresponding simple metal ion. stant of the complex.
min throllgh 10.0 ml of the solution.
1=2" Lingane4 has shown that the shift in half- Equation 21-9 makes it possible to evalu-
wave potential as a function of the concentra- ate the formula for the complex. Thus, a plot
Substituting this relationship in Equation 21-6 of the half-wave potential against log FA for
q = 8.0 min x 60 ~ x 8.4 X 10-6 A tion of complexing agent can be employed to
reveals that
mm several formal ligand coneentrations gives-a
= 4.03 x 10-3 C o . 0.0591 k straight line, the slope of which is 0.0591xln.
E 1/2 = E - £seE- -n-Iog ka (21-7) If n is known, the combining ratio of ligand
2+ ed 4.03 X 1O-3C to metal ion is thus obtained. Equation 21-9
N o. 0f meq Z n consum == ----- Thus, the half-wave potential is a constant that
96,491 CIF can then be employed to calculate K I'
is related to the standard potential for the half-
3
10 meq reaction, k, and kR. The latter two terms are
x---
F constants that depend upon the diffusion coef-
ficients for the two species and certain charac-
= 4.18 X 10-5
teristics of the dropping electrode, which will
TABLE 21-1 EFFECT OF COMPLEXING AGENTS ON
No. of millimoles Zn2 + consumed POLAROGRAPHIC HALF-WAVE POTENTIALS AT THE
be discussed later. Substitution of Equation
DROPPING MERCURY ELECTRODE
21-7 into 21-6 yields
= 2.09 x 10-5
Noncomplexing IFNH3,
% decrease in Zn2+ concentration 0.0591 i IFNH .•O
EapPI = E1/2 ---log -. -. (21-8) Ion Media IFKCN IFKCI
n 14-1
2.09 x 10-5
= 1.00 X 10-3 X 10 x 100 = 0.21 CdH -0.59 -1.18 -0.64 -0.81
As will be shown subsequently, i4 can be cal- -1.00 -1.35
ZnH -1.00 NRa
culated from the characteristics of the particu- PbH -0.40 -0.72 -0.44 -0.67
lar electrode and the diffusion coefficient for NiH -1.36 -1.20 -1.10
Ox. Thus the current i can be calculated at Co2 + -1.45 -1.20 -1.29
Equation for the -0.24
any desired value of EapPI• Cu2+ +0.02 NR" +0.04
Polarographic Wave -0.22 -0.51
Examination of Equation 21-7 reveals that
To obtain an equation relating current and the half-wave potential is a reference point on
potential, let us subtract Equation 21-4 from a polarographic wave; it is independent of the • No reduction occurs berore involvement or the supporting electrolyte.
LJ
Because of the effect of complexing reagents I determined, provided the drop rate t is repro-
on half-wave potentials, the electrolyte content , ducible. Note the effect of irregular drops in
of the solution should be carefully controlled the center of the limiting current region, prob-
whenever polarographic data are used for the ably aiused by vibration of the apparatus.
qualitative identification of the constituents of Advaatages aad Limitatioas of the Drop-
a solution. pi.. Mercury Electrode. The dropping mer-
Polarogralm· for Irreversible Reactions. cury electrode offers several advantages over
Many polarographic electrode processes, par- other types of microelectrodes. The first is the
ticularly those associated with organic systems, large overvoltage for the formation of hydro-
are irreversible; drawn-out and less well- gen from hydrogen ions; as a consequence, the
defined waves result. The quantitative descrip- reduction of many substances from acidic solu-
tion of such waves requires an additional term tions can be studied without interference.
in Equation 21-8 (involving the activation Second, because a new metal surface is con-
energy of the reaction) to account for the tinuously generated, the behavior of the elec-
kinetics of the electrode process. Although trode is independent of its past history; thus,
half-wave potentials for irreversible reactions reproducible current-voltage curves are ob-
ordinarily show a dependence upon concen- tained regardless of how the electrode has been
tration, diffusion currents remain linearly re- used previously. A third unique feature of the
lated to this variable, and such processes are dropping electrode is that reproducible average
readily adapted to quantitative analysis. currents are inmiediately achieved at any given
applied potential.
The most serious limitation to the drop-
The Dropping
ping electrode is the ease with which mercury
Mercury Electrode is oxidized; this property severely restricts the
The dropping mercury electrode possesses use of the electrode as an anode. At applied 10.0
unique properties that make it particularly potentials much above +0.4 V formation of c( i•••
useful for voltammetric studies. mercury(I) occurs; the resulting current masks
ClIITeIItVariations Duriag the Lifetime of. the polarographic waves of other oxidizable "-
~
:>
Drop. The current in a cell containing a drop- species in the solution. Thus, the dropping 0
id
ping electrode undergoes periodic fluctuations mercury electrode can be employed only for 5.0
corresponding in frequency to the drop rate. the analysis of reducible or very easily oxi-
As a drop breaks, the current falls to zero (see .' dizable substances. Other disadvantages of the
Figure 21-3a); it then increases rapidly as the' dropping mercury electrode are that it is cum-
electrode area grows because ofthe greater sur- bersome to use and tends to malfunction as a
face to which diffusion can occur. In order to result of clogging.
determine the average current, it is necessary
to reduce the large fluctuations in the current
by a damping device or by sampling the current lbl
Polarographic
near the end of each drop, where the change FIG URE 21-3 Effect of drop growth on polarographic
Diffusion Currents
in current with time is relatively small. currents. (a) Current-time relationship during the lifetime
Damping is ordinarily achieved by employing The I1kovic Equation. In 1934, D. Ilkovic5 t of drops. The shaded area represents the microcoulombs
a well-damped galvanometer or a low-pass deriTed a fundamental equation relating the of electricity associated with each drop. Note that the
filter (p. 23). As shown in Figure 21-3b, current increases by only about 20 % during the second
damping limits the oscillations to a reasonable half of the drop lifetime. (b) A recorded polarogram in
magnitude; the average current (or, alterna- • D. Ilkovic, Col~cr. CzechoslOll. Chem. Commun, 6, 498. which the current fluctuations have been damped, thus
tively, the maximum current) is then readily (1934~ permitting the measurement of the average current i. y •
various parameters that determine the magni- diffusion current is directly proportional to
tude of the diffusion currents obtained with a the square root of the column height. The
dropping mercury electrode. He showed that drop time t of a given electrode is also af-
at 2SoC, fected by the applied potential since the inter-
facial tension between the mercury and the
solution varies with the charge on the drop.
Generally, t passes through a ~imum at
Here, i4 is the time-average diffusion current about -0.4 V (vs. SCE) and then falls off
in microamperes (see Figure 21-3) during the fairly rapidly; at -2.0 V, t may be only one-
lifetime of a drop; n is the number offaradays half of its maximum value. Fortunately, the
per mole of reactant, D is the diffusion coef- diffusion current varies only as the one-sixth
ficient for the reactive species (expressed in power of the drop time so that, over small
units of square centimeters per second), m is potential ranges, the decrease in current due
the rate of mercury flow in milligrams per to this variation is negligibly small.
second, t is the drop time in seconds, and C is Temperature. Temperature affects several
the concentration of the reactant in millimoles of the variables that govern the diffusion cur-
per liter. The quantity 607 represents the com- rent for a given species, and its overall influence
bination of several constants (for the maximum is thus complex. The most temperature-sensi-
current, this constant takes a value of 709). ."." (, tive factor in the Ilkovic equation is the dif-
The Ilkovic equation is based upon certain fusion coefficient, which ordinarily can be ex-
assumptions that cause discrepancies of a few pected to change by about 2.S % per degree.
percent between calculated and experimental As a consequence, temperature control to a few
diffusion currents. Corrections to the equa- tenths of a degree is needed for accurate
tion have been derived that yield better corre- polarographic analysis.
spondence6; for most purposes, however, the FIGURE 21-4 Polarograms of A: approximately 0.1 mM each of
simple equation gives a satisfactory accounting silver(n thallium(I1 cadmium(II1 nickel(II1 and zinc(I11 listed in
of the factors that influence the current. the order in which their waves appear, in IF amonia/IF ammonium
C.piUary Characteristics. The product Polarogram. for chloride containing 0.002% Triton X-loo; B: the supporting electro-
m2/3t1/6 in the Ilkovic equation, called the Mixture. of Reactants lyte alone. (From L. Meites, Polarographic Techniques,2d ed. New
capillary constant, describes the influence of York: John Wil~y & Sons, Inc., 1967, p. 164. With permission.)
dropping electrode characteristics upon the dif- Ordinarily, the reactants of a mixture will be-
fusion current; since both m and t are readily have independently of one another at a micro-
evaluated experimentally, comparison of dif- electrode; a polarogram for a mixture is thus
fusion currents from different capillaries is thus simply the summation of the waves for the Anodic Wave. and
possible. individual components. Figure 21-4 shows the Mixed Anodic-Cathodic Wave.
Two factors, other than the geometry of polarogram of a five-component cation mix-
the capillary itself, playa part in determining ture. Clearly, a single polarogram may permit Anodic waves as well as cathodic waves are
the magnitude of the capillary constant. The the quantitative determination of several ele- encountered in polarography. The former are As the potential is made more negative, a de-
head that forces the mercury through the ca- ments. Success depends upon the existence of less common because of the relatively small crease in the anodic current occurs; at about
pillary influences both m and t such that the a sufficient difference between succeeding half- range of anodic potentials that can be covered ? -0.02 V, the cur(ent becomes zero because the
wave potentials to permit evaluation of indi- with the dropping mercury electrode before . o"'Xidition of irori(I1) ion has ceased.
vidual diffusion currents. Approximately 0.2 V oxidation of the electrode itself commences. Curve C represents the polarogtam for a
is required if the more reducible species under- An example of an anodiC wave is illustrated in solution of iron(III) in the same medium.
goes a two-electron reduction; a minimum of curve A of Figure 21-S, where the electrode Here, a cathodic wave results from reduction
about 0.3 V is needed if the first reduction is reaction involves the oxidation of iron(I1) to of the iron(III) to the divalent state. The half-
6 See: J. J. Linpne and B. A. Loveridge, J. AllIn. Chem. a one-electron process. The analysis of mix- iron(III) in the presence of citrate ion. A dif- wave potential is identical with that for the
Soc, n, 438 (19SO~ tures is considered in a later section. fusion current is obtained at about +0.1 V, anodic wave, indicating that the oxidation
molecular weight substances as gelatin, Triton each new drop is charged as it forms, a small
X-loo (a commercial surface-active agent), but steady current results. At applied poten-
methyl red, or other dyes will cause a maximum tials smaller than about-0.4 V, the mercury
to disappear. Care must be taken to avoid tends to be positive with respect to the solu-
large amounts of these reagents, however, be- tion; thus, as each drop is formed, electrons
cause the excess may reduce the magnitude of are repelled from the surface toward the bulk
the diffusion current. The proper amount of of mercury, and a negative current is the
suppressor must be determined by trial and result. At about -0.4 V, the mercury surface
error; the amount required varies widely from is uncharged, and the condenser current is
species to species. zero (see Figure 21-4, curve B). The charging
current is classed as a nonfarac1aic current in
+0.4 +0.2 0.0 -0.2 -0.4 -0.6 -O.B Residual Current the sense that electricity is carried across an
Applied emf w. SCE, V electrode-solution interface without an accom- o -0.4 -0.8 -1.2 -1.6 -2.0
The typical residual current curve, such as panying oxidation-reduction process.
FIGURE 21-5 Polarographic be- curve B in Figure 21-4, has two sources. The Applied potential w. SCE, V
Ultimately, the accuracy and sensitivity of
havior of iron(II) and iron(III) in a first of these is the reduction of trace impuri- FIGURE 21-7 Polarogram for the
the polarographic method depend upon the
citrate medium. Curve A : anodic wave ties that are almost inevitably present in the reduction of oxygen in an air-saturated
magnitude of the residual current and the
for a solution in which [FeH] = blank solution; contributors here include small 0.1 F KCI solution. The lower curve is
accuracy with which a correction for its effect
1 x 10-3. Curve B: anodic-cathodic amounts of dissolved oxygen, heavy metal ions for oxygen-free 0.1 F KCI.
can be determined.
wave for a solution in which (FeH] = from the distilled water, and impurities present
[Fe3+] = 0.5 x 10-3. Curve C: cath- in the salt used as the supporting electrolyte.
odic wave for a solution in which A second component of the residual cur-
Supporting Electrolyte
[Fe3+] = 1 x 10-3. rent is the so-called charging or condenser cur- An electrode process will be diffusion-con- The second corresponds to the further reduc-
rent resulting from a flow of electrons that trolled only ifthe solution contains a sufficient tion of the hydrogen peroxide:
charges the mercury droplets with respect to concentration of a supporting electrolyte to H202 + 2H+ + 2e¢:2H20
the solution; this current may be either nega- diminish the attractive or repulsive force be-
and reduction of the two iron species are per- tive or positive. At potentials more negative tween the electrode and the analyte to a value As would be expected from stoichiometric con-
fectly reversible at the dropping electrode. than about -0.4 V, an excess of electrons that approaches zero. Generally, this condition siderations, the two waves are of equal height.
Curve B is the polarogram of an equifor- provides the surface of each droplet with a is realized when the supporting electrolyte con- These polarographic waves are convenient
mal mixture of iron(II) and iron(III). The por- negative charge. These excess electrons are centration exceeds that of the reactive species for determining the concentration of dissolved
tion of the curve below the zero-current line carried down with the drop as it breaks; since by a factor of 50 to 100. Under these circum- oxygen; however, the presence of this element
corresponds to the oxidation of the iron(lI); stances, the fraction of charge carried through often interferes with the accurate determina-
this reaction ceases at an applied potential the solution by the species of interest is neg- tion of other species. Thus, oxygen removal is
equal to the half-wave potential. The upper ligibly small (p. 504) because of the large excess ordinarily the first step in a polarographic
portion of the curve is due to the reduction of of other ions. The limiting current then be- analysis. Aeration of the solution for several
iron(III). comes a diffusion current, which is independent minutes with an inert gas accomplishes this
of electrolyte concentration. end; a stream of the same gas, usually nitrogen,
Current Maxima is passed over the surface during the analysis
to prevent reabsorption.
The shapes of polarograms are frequently dis- Oxygen Waves
torted by so-called current maxima (see Figure
Dissolved oxygen is readily reduced at the
.21-6), which are troublesome because they in-
dropping mercury electrode; an aqueous solu-
terfere with the accurate evaluation of diffusion APPLICATIONS
tion saturated with air exhibits two distinct
currents and half-wave potentials. Although OF POLAROGRAPHY
waves attributable to this element (see Figure
the cause or causes of maxima are not fully
21-7). The first results from the reduction of Apparatus
understood, there is considerable empirical
oxygen to peroxide:
knowledge of methods for eliminating them. FIGURE 21-6 Typical current Cells. A typical general-purpose cell for
Generally, the addition of traces of such high maxima. 02(g) + 2H+ + 2e¢:H202 polarographic analysis is shown in Figure 21-1
double-pole, double-throw switch is in posi- A~~is of Mixtures. The quantitative de-
(p. 607). The solution to be analyzed is sepa-
tion 2. The current is measured by deter- termmatlon of each species in a multicom-
rated from the calomel electrode by a sintered .
mining the potential drop across the precision ponent mixture from a single polarogram is
disk backed by an agar plug that has been
rendered conducting by the addition. of potas-
10,000-0 resistance R2 with the same poten- theoretically feasible, provided the half-wave
sium chloride. Such a bridge is readily prepared tiometer and the switch in position L The potentials of the various species are sufficiently
current oscillations associated with the drop- d~erent .(see Figure 21-4). Where a major con-
and lasts for extended periods, provided a
ping electrode require that the null-detecting stituent IS more easily reduced than a minor
potassium chloride solution is kept in the reac-
galvanometer be damped with a suitable one, however, the accuracy with which the
tion compartment when the cell is not in use.
resistance. latter can be determined may be poor because
A capillary sidearm permits the passage of
nitrogen or other gas through the solution. The apparatus shown in Figure 21-8 pro- its diffusion current can occupy only a small
Provision is also made for blanketing the solu- vides satisfactory data for routine quantitative fraction of the current scale of the instrument.
work; in such applications, current measure- Under such circumstances, a small error in
tion with the gas 4uring the analysis.
ments at only two voltages are needed to define current measurement leads to a large relative

J
In a simpler arrangement, a mercury pool
in the bottom of the sample container can be ;4 (one preceding the decomposition potential error in the analysis.
used as the nonpolarizable electrode. Here, and one in the limiting-current region). On the Several methods exist for treating mixtures
the observed half-wave potentials differ some- FIGURE 21-8 A simple circuit for polaro- other hand, where the entire polarogram is containing species in unfavorable concentra-
what from published values, which are based graphic measurements. [Reprinted with per- required, a point-by-point determination of the tion ratios. One method is to alter the sup-
upon a saturated calomel reference electrode. mission from J. J. Lingane, Anal. Chern., 21, curve with a manual instrument is tedious and porting electrolyte so that the wave for the
Dropping Electrodes. A dropping electrode, 45 (1949). Copyright by the American Chemi- time-consuming; for this type of work, auto- minor constituent appears first; with the va-
such as that shown in Figure 21-1, can be pur- cal Society.] matic recording is of great value. Several simple riety of complexing agents available, this tech-
chased from commercial sources. A .1Q-cm recording instruments are available in the nique is often feasible. Alternatively, a prelim-
capillary ordinarily has a drop time of 3 to 6 s range of $2000 to $3000. In such instruments, inary chemical separation can be employed.
under a mercury head of about 50 em. The tip use, the head is increased, the liP is immersed the contact of the potential divider R1 in Figure Finally, the current due to the major con-
of the capillary should be as nearly perpen- in 1 : 1 nitric acid for a minute or so, and ·then 21-8 is driven by a motor synchronized with stituent can be decreased to zero (or a very
dicular to the length of the tube as possible, rinsed with distilled water. the paper drive motor of a millivolt recorder; small value) by application of a counter emf
and care should be taken to assure a vertical Electrical Apparatus. The electrical appara- the voltage drop across resistance R2 is then in the current-measuring circuit. The current
mounting of the electrode; otherwise, erratic tus for polarographic measurements can be recorded. sensitivity can then be increased to give a satis-
and nonreproducible drop times and sizes will relatively simple. A voltage divider that per- factory signal for the reduction of the minor
be observed. mits continuous variation of an applied voltage Treatment of Data component. Most modem polarographs are
With reasonable care, a capillary can ~ from 0 to about t25 V is needed; the applied equipped with such compensating devices.
used for several months or even years. Such voltage must be known to about ±0.01 V. In It is common practice to use either the aver- Concentration DeterminatiolL The best and
performance, however, requires the use of addition, it must be possible to measure cur- age or the maximum value ofthe galvanometer most straightforward method for quantitative
scrupulously clean mercury and the mainte- rents over the range between om and perhaps or recorder oscillations. polarographic analysis involves preliminary
nance of a mercury head, no matter how slight, 100 JlA with a precision of tOm JlA. A manual Determination of Diffusion Currents. For calibration with a series of standard solutions;
at all times. If solution comes in contact with instrument that meets these requirements is quantitative work, limiting currents must as nearly as possible, these standards should
the inner surface of the tip, malfunction of the easily constructed from equipment available always be corrected for the residual current. be identical with the samples being analyzed,
electrode is to be expected. For this reason, in most laboratories. More elaborate devices One method involves the use of a blank to and should cover a concentration range within
the head of mercury should always be in- for recording polarograms aUtomatically are derive an experimental residual-current curve. which the unknown samples will likely fall. The
creased to provide a good flow before the tip commercially available. The diffusion current is then evaluated from linearity of the current-concentration relation-
is immersed in a solution. Figure 21-8 shows a circuIt diagram of a the difference between the two at some poten- ship can be assessed from such data; if the
Storage of an electrode always presents a simple instrument for polarographic work. tial in the limiting-current region. relationship is nonlinear, the analysis can be
problem. One method is to rinse the. capillary Two 15-V batteries provide a voltage across Because the residual current usually in- based upon the calibration curve.
thoroughly with water, dry it, and then care- the 100-0 voltage divider Rio by means of creases nearly linearly with applied voltage, it Another useful technique is the standard-
fully decrease the head until the flow of mer· which the potential applied to the cell can be is often possible to make the correction by addition method. Here, the diffusion current
cury in air just ceases. Care must be taken to varied. The magnitude of this voltage can be extrapolation; this technique is illustrated in for an accurately known volume of sample
avoid lowering the mercury too far. Before determined with the potentiometer when the Figure 21-3 (p. 613). solution is measured. Then, a known amount
of the species of interest is introduced (or- The polarographic method is also appli· fruitful for the detennination of structure, the Solyents for Organic Polarography. Solubil-
dinarily as a known volume of a standard cable to the analysis of such inorsanic anions qualitative identification of compounds, and ity considerations frequently dictate the use
solution~ and the diffusion current is again as bromate, iodate, dichromate, vanadate, the quantitative analysis of mixtures. of some solvent other than pure water for
evaluated. Provided the relationship between selenite, and nitrite. In general, polarograms Etrect of pH on Polarograms. Organic organic polarography; aqueous mixtures con-
current and concentration is linear, the in- for these substances are affected by the pH of electrode processes ordinarily involve hydro- taining varying amounts of such miscible sol- .
crease in wave height will permit calculation the solution because the hydrogen ion is a gen ions, the most common reaction being vents as glycols, dioxane, alcohols, cellosolve,
of the concentration in the original solution. participant in the reduction process. As a represented as or acetic acid have been employed. Anhy-
The standard-addition method is particularly consequence, strong buffering to some fixed drous media such as acetic acid fonnamide
R + nH + + ne = RH.
effective when the diffusion current is sensi- pH is necessary to obtain reproducible data diethylamine, and ethylene glyc~1 have als~
tive to other components of the solution that for these species. where Rand RH. are the oxidized and reduced been investigated. Supporting electrolytes are
are introduced with the sample. For further applications of polarography fonns of the organic molecule. Half-wave often: lithium salts or tetraalkyl ammonium
to inorganic analysis, the reader is referred to potentials for organic compounds are therefore salts.:
the monographs by Kolthoff and Lingane markedly pH-<lependent. Furthennore, altera- Reactive Functional Groups. Organic com-
Inorganic and by Meites.8 tion of the pH often results in a change in the pounds containing any of the following func-
Polarographic Analysis reaction products. For example, when benz- tional groups can be expected to produce one
aldehyde is reduced in a basic solution, a wave or more polarographic waves.
The polarographic method is generally ap-
Organic is obtained at about -1.4 V, attributable to
plicable to the analysis of inorganic substances.
Polarographic Analysis the fonnation of benzyl alcohol: 1. The carbonyl group, including aldehydes,
Most metallic cations, for example, are reduced
ketones, and quinones, produce polaro-
at the dropping electrode to fonn a metal Almost from its inception, the polarographic C6HsCHO + 2H+ + 2e~C6HsCH20H graphic waves. In general, aldehydes are
amalgam or an ion of lower oxidation state. method has been used for the study and
If the pH is less than 2, however, a wave reduced at lower potentials than ketones;
Even the alkali metals and alkaline-earth analysis of organic compounds, and many
occurs at about - 1.0 V that is just half the co.njugation of the carbonyl double bond
metals are reducible, provided the supporting papers have been devoted to this subject.
size of the foregoing one; here, the reaction al~o results in lower half-wave potentials.
electrolyte does not react at the high potentials Several common functional groups are oxi-
involves the production of hydrobenzoin: 2. Certain carboxylic acids are reduced polar-
required; here, the tetraalkyl ammonium ha- dized or reduced at the dropping electrode;
ographically, although simple aliphatic and
lides are useful. compounds containing these groups are thus 2C6HsCHO + 2H+ + 2e aromatic monocarboxylic acids are not.
The successful polarographic analysis of subject to polarographic analysis.'
:;=C6HsCHOHCHOHC6Hs Dicarboxylic acids such as fumaric, maleic,
cations frequently depends upon the support- In general, the reactions of organic com-
or phthalic acid, in which the carboxyl
ing electrolyte that is used. To aid in this pounds at a microelectrode are slower and
At intennediate pH values, two waves are groups are conjugated with one another,
selection, tabular compilations of half-wave more complex than those for inorganic ca-
observed, indicating the occurrence of both give characteristic polarograms; the same is
potential data should be consulted.7 The judi- tions. Thus, theoretical interpretation of the
reactions. true of certain keto and aldehydo acids.
cious choice of anion often enhances the selec- data is more difficult and perhaps impossible;
It should be emphasized that an electrode 3. Most peroxides arid epoxides yield polar-
tivity ofthe method. For example, with potas- moreover, a much strieter adherence to detail
process that consumes or produces hydrogen ograms.
sium chloride as a supporting electrolyte, the is required for quantitative work. Despite these
ions tends to alter the pH of the solution at 4. Nitro, nitroso, amine oxide, and azo groups
waves for iron(II1) and copper(II) interfere handicaps, organic polarography has proved
the electrode surface; unless the solution is are generally reduced at the dropping
with one another; in a fluoride medium, how-
well-buffered, marked changes in pH can occur electrode.
ever, the half-wave potential of the fonner is
in the surface film during the electrolysis. These S. Most organic halogen groups produce a po-
shifted by about -0.5 V, while that for the
changes affect the reduction potential of the larographic wave which results from re-
latter is altered by only a few hundredths of a • I. M. Kolthoff and J. J. Lingane, Polarography, 2d ed.
New York: Interscience, 19S2, vol. 2; and L. Meites, Po- reaction and cause drawnout, poorly defined placement of the halogen group with an
volt. The presence of fluoride thus results in the
larographic Teclmiques, 2d ed. New York: Interscience, waves. Moreover, where the electrode process atom of hydrogen.
appearance of separate waves for the two ions. 1965. is altered by the pH, nonlinearity in the dif- 6. T#le carbon/carbon double bond is reduced
, For a detailed dilcu.aion of orpnic poIarosraPhic fusion current-concentration relationship must when it is conjugated with another double
analysis, ICe: P. Zuman, Organic PaImograpltic Analysis.
also be exPected. Thus, in organic polarog- bond, an aromatic ring, or an unsaturated
Oxford: Pergamon Pras, 1964;Polarography of Molecules
1 See, for example, the ret'ereoca cited on page 606; of Biological Significance, ed. W. F. Smyth. New York: raphy good buffering is vital for the genera- group.
another extensive source is Handboolc of Analyrical cliem- AClIdemicPras, 1979; and Toplu in Orgtmk Polarog· tion of reproducible half-wave potentials and 7. Hydroquinones and mercaptans produce an-
ist", ed. L. Meites. New YorIt: McGraw·HilI, 1963. raphy, ed. P. Zuman. New York: Plenum Pras, 1970. diffusion currents. odic waves.
cell becomes larger. oxide film formation. These processes are height of the maximum is proportional to
In addition, a number of other organic
Most of the substitutes for the dropping time-dependent; thus, the electrode behavior concentration. The position of the peak along
groups cause catalytic hydrogen waves that
mercury electrode are solid electrodes fab- often depends upon the history of its use. the potential axis is characteristic of the
can be used for analysis. These include
ricated from platinum, gold, or graphite. I I Currents with Stationary Electrocles. When electroactive species.
amines, mercaptans, acids, and hete~~clic
Two operational modes are employed. In one, a potential sufficient to cause an electrode The cause of the initial current increases
nitrogen compounds. Numerous applications
to biological systems have been reported!O the electrode is stationary, and every effort is reaction is applied to a solid inicroelectrode, observed in Figure 21-9 is the same as that
made tQ avoid vibrational and convectional a large initial current is observed, which then already described for a conventional polaro-
stirring. In the second, the electrode is rotated decays to a constant and reproducible level graphic wave. At. the peak, the concentration
or vibrated at a constant rate; alternatively, after several minutes. The initial current is as- of the ions or molecules of the reactive species
VOLTAMMETRY AT the solution is stirred in a reproducible man- sociated with the formation of a concentra- at the electrode surface approaches zero, and
SOLlI) ELECTRODES ner. In either case, the electrode and solution tion gradient that ultimately extends several the current becomes limited by their diffusion
are in motion with respect to one another. tenths of a millimeter from the electrode sur- rate. As the sweep is continued, the time of
Two limitations to the dropping mercury elec- face. Within this gradient, the reactant concen- electrolysis is longer, and the distance over
trode h~ve led chemists to search for substi- tration increases continuously from the equi- which the reactive particles must move to
tutes. One undesirable property is the ease with librium value at the electrode surface to that reach the surface is greater. Thus, the rate of
which mercury is oxidized; as a consequence, Voltammetry at of the bulk of the solution. Constant currents, their arrival at the surface is lower; so also is
oxidatiOn processes that occur at potentials Stationary Electrodes
which are applicable for analysis, are obtained the current.
greater than about +0.4 V vs. SCE cannot be Stationary microelectrodes have been em- only when the slope ofthe relationship between The peak height with a stationary elec-
studied (in the presence of halide and certain ployed for the analysis of a variety of organic concentration and distance from the electrode trode is proportional to the square root of the
other ions, the electrode is restricted to poten- compounds, usually by anodic oxidation at the surface becomes essentially constant; ordi- voltage sweep rate and depends upon the di-
tials more negative than 0.0 V); nor can the indicator electrode. In addition, such elec- narily, 2 to 5 min are required to reach this rection of the sweep.
reduction of strong oxidizing agents be in- trodes have made possible the determination condition. These same current changes are also The Oxygen Electrode. An important appli-
vestigated because these substances oxidize of strong inorganic oxidizing agents; the most observed with a dropping electrode.12 cation of stationary solid electrodes is to the
mercury. . important application of this type is to the Generally, no attempt is made to reach determination of oxygen in water, sewage
The'second limitation of a dropping elec- determination of dissolved oxygen in water. steady currents for each new voltage setting effluents, blood, and other fluids.13 The elec-
trode is that of sensitivity, which prevents the Electrocle COIBtruction. A platinum elec- that is applied to a stationary electrode. In- trode consists of a platinum or gold surface
analysis of solutions more dilute than about trode is formed by sealing a wire into the end stead, a linear voltage sweep is employed in on the end of a support rod that is covered
10- 5M. At lower concentrations, the diffusion of a soft glass tube. The platinum may be filed which the applied potential is varied at a con- with an oxygen-permeable membrane of Teflon
current is of the same magnitude or smaller and polished flush with the glass to provide a stant rate, and the instantaneous current is or polyurethane. The arrangement is often
than the residual current; corrections for the planar rather than a cylindrical surface; such recorded. Current-voltage curves such as those similar to the gas-sensing electrode shown in
background current then lead to large errors electrodes, as well as graphite and gold elec- shown in Figure 21-9 are obtained. Here, the Figure 19-8 (p. 553) with the exception that
that ultimately limit the sensitivity. Most of trodes, are available from commercial sources. voltage sweep was begun at about +0.4 V the glass electrode is replaced with a micro
the residual currents are nonfaradaic (p. 506) No solid electrode yields as reproducible platinum electrode. The reference electrode is
(vs. SCE) and extended to +0.7 to 1.0 V; the
and cannot be decreased to less than a few current-voltage behavior as does the drop- ordinarily a silver/silver chloride or other silver
sweep rate was 74 mY/min. The voltam-
hundredths of a microampere. Their main ping mercury electrode, which continuously mograms result from the oxidation of N,N- based electrode. A potential of about -0.8 V
sources are the charging current (p. 616) and provides a new and fresh electrode surface. In (vs. SCE) is applied across the pair of elec-
tetramethylbenzidine at a micro platinum wire
noise currents, which originate in the cell cir- contrast, the behavior of a solid electrode is trodes, whereupon reduction of oxygen occurs
electrode in an aqueous medium. In contrast·'
cuit of the instrument itself and from other often affected by adsorption, deposition, or in the film of liquid adjacent to the platinum
to a polarogram, a current maximum instead
"electriCal sources as well (see p. 71); the effects
of a constant limiting current is observed. The surface. The electrode process is:
of the.iatter increase as the resistance of the
!

11 For discussions o( voltammetry at solid electrodes,


see: R. N. Adams, in Treatise on Analyrk:a1 C1leIIIistry, 12 With a dropping electrode, however, the solution be-
•• M. Brezine and P. Zuman, Po/Qrography In M"'k:ine, eds. I. M. Kolthotr and P. J. Elving. New York: Inter- comes homogenized by local Itirring u a drop breaks,
Biochemistry and PlulnruJcy. New York: Intencienc:e, science, 1963,Part I, vol. 4, Chapter 47; and S. Piekanki and (or each new drop, conditions are identical to those
1958; and Polarography of Moleadn of BioIogk:a1 and R. N. Adams, in Physk:al Merlloth of Chemistry, eds. encountered by the previous drop. Thus, the current U For a thorough treatment o(the oxygen electrode, see:

Significance. ed. W. F. Smyth. New York: Academic A. Weissberger and B. W. Rossitor. New York: Wiley- fluctuations are perfectly reproducible, and the average I. Fatt, Polarographic Oxygen Sensors. Cleveland: CRC
Press, 1979. Interscienc:e,1971,Part IIA, vol. I, Chapter 7. current is independent ofthe time when it is measured. Press. 1976.
The current is proportional to the concentra- In addition, the high currents obtained cause
tion of oxygen in the internal solution, which the electrode to be particularly sensitive to
in turn is proportional to the oxygen concen- traces of oxygen in the solution. These two
tration in the extemal solution adjacent to factors have largely confined its employment
+J the electrode. Transport of oxygen occurs by to anodic reactions. Limiting currents are often
Direction of sweep the mechanism described earlier for CO2 influenced by the previous history of the
Potential ¥s. SCE. V (p.551). rotating platinum electrode and are seldom as
0.8 0.6 reproducible as the diffusion currents obtained
1.0 0.9 0 Oxygen electrodes are available commer-
cially. They could more aptly be termed with a dropping electrode.
oxygen sensitive voltarnmetric cells. Rotating electrodes have been used exten-
-2 sively in studying the oxidation of organic
compounds and as indicator electrodes for
Voltammetry
amperometric tit rations (p. 636).
-4
at Rotating
Platinum Electrode.
Most voltarnmetric applications of solid elec- POLAROGRAPHY WITH
-6 trodes are based upon the rotating platinum POTENTIOSTATIC CONTROL
electrode, an example of which is shown in
« Figure 21-21 (p. 637). The microelectrode Figure 21-10 illustrates the effect of cell resist-
-8 "- consists of a short length of platinum wire ance on the polarographic wave for.a revers-
¥ sealed into the side of a glass tube; mercury ible system. At 100 n, the IR drop is so small
"
u
within the tube provides electrical contact be- that it does not influence the shape of the wave
-10 tween the wire and the lead to the polaro- significantly. With high resistances, on the
graph. The tube is held in the hollow chuck of other hand, the waves become drawn out and
a synchronous motor and rotated at a con- ultimately are so poorly defined as to be oflittle
-12 stant speed in excess of 600 rpm. Commercial use. The cause of this effect is easily understood
models of the rotating electrode are available. by reference to the equation
Voltammetric waves, which are similar to
E.PPI = Eind - EIiCE -IR
-14 those observed with the dropping electrode,
can be obtained with the rotating platinum Here, E.ppl is varied in a regular way while
electrode. Here, however, the reactive species EIiCE is constant. If IRis smaiL the indicator
-16 is brought to the electrode surface not only
by diffusion but also by. mechanical mixing.
As a consequence, the limiting currents are as
-18 much as 20 times larger than those obtained
with a microelectrode that is supplied by dif-
FIGURE 21-9 Linear sweep voltammograms for N,N'-tetramethyl- fusion only; thus, the rotating electrode has
benzidine with a platinum wir!: microelectrode. Curve A: 0.4 x 10-4 F; the potential for enhancing sensitivity. Rotat-
Curve B: 0.8 x 10-4 F; Curve'C: 1.2 x 10-4 F; Curve D: 1.6 x 10-4 F; ing electrodes, like their dropping electrode
E
e
Curve E: 2.0 x 10-4 F. (Adapted from R. N. Adams, in Treatise on
Analytical Chemistry, cds. I. flit. Kolthoff and P. J. Elving. New York:
counterparts, provide steady currents instan- a 5
taneously. This behavior is in distinct con-
John Wiley & Sons, Ine., 1963, Part I, vol. 4, p. 2407. With permission.) trast to that of a solid microelectrode in the
absence of stirring.
Several factors restrict the widespread ap-
plication of the rotating platinum electrode for
voltammetry. The low hydrogen overvoltage FIGURE 21-10 Effect of cell resist-
prevents its use as a cathode in acidic solutions. ance on a reversible polarographic wave.
electrode then faithfully reflects the variation to control the applied ~tential EOPPI s~ch that Thus, the output potential is determined by marized in Figure 21-12 and are described
in Eoppl' On the other hand, if R is large, a the variation of E1nd is linear with time. Thus; E, and R, and at any instant is directly pro- briefly in the paragraphs that follow.14
greater and greater fraction of EappI is required tpe abscissa of the polarographic wave be- portional to the time that has elapsed after
to overcome the IR drop with increasing I. comes E1nd rather than Eoppl' The effect is to closure of the switch S. A fraction of Eo is
applied to the noninverting tepninal of the Differential
Thus, Eind is no longer directly proportional produce a wave form similar to that for the
operational amplifier associated with the po- Pulsed Polarography
to EoPPI' A drawn-out wave results. lowest resistance curve in Figure 21-10. '

r The use of potentiostatic control has per-


mitted the extension of polarography to sol-
vents with high electrical resistances. Here,
three electrodes are employed: a dropping
Figure 21-11 is a circuit diagram of an all-
electronic instrument for three-electrode polar-
ography. It consists of four components: a
ramp generator, a potentiostat, a cell containing
tentiostat.
The potentiostat shown in Figure 21-11 is
similar in design to' the one shown in Figure
20-8 and described on page 589.
In differential pulsed polarography, a dc poten-
tial, which is increased linearly with time, is
applied to the polarographic cell. As in classical
polarography, the rate of increase is perhaps
electrode (or other microelectrode 1a counter three electrodes, and a current amplifier and The cell consists of a dropping mercury 5 mVIs. In contrast, however, a dc pulse of an
or auxiliary electrode, and a reference elec- recorder. The purpose of the ramp generator electrode, a saturated calomel electrode (which additional 20 to 100 mV is applied at regular
trode. The counter electrode and the micro- is to provide a continuously varying potential is part of the control circuit), and an auxiliary intervals of about 1 to 3 s; the length of the
electrode serve the same purpose as the two or linear sweep across the indicator-saturated or working electrode. Essentially no current pulse is about 60 ms and terminates with de-
electrodes in ordinary polarography. The refer- calomel electrode system. Note that the genera- develops in the circuit that includes the refer- tachment of the mercury drop from the elec-
ence electrode, on the other hand, is employed tor circuit is similar to the integration circuit ence and indicator ,electrodes because of the trode. To synchronize the pulse with the drop,
to measure and control the potential of the shown in Figure 3-26 (p. 66). The relationship high internal resistance of the operational the latter is detached by an appropriately timed
microelectrode. The arrangement here is anal- between the constant input potential E, and the amplifier. The measured current is thus the mechanical tap or movement of the electrode.
ogous to that described for controlled potential resulting output potential Eo can be obtained output current 1 from the operational ampli- The voltage program is shown in Figure 21-13.
electrolysis (see Figure 20-4 or 20-8). The from Equation 3-20. 'That is, fier, which is determined by the potential of Two current measurements are made al-
reference electrode is positioned as close as
possible to the dropping electrode, and the
potential between the two is then employed
E
o
=- E,
R,
r
0
dt =_ E't
R,
the dropping electrOde relative to the saturated
calomel electrode. :
The auxiliary eiectrode in polarographic
studies with three 'electrodes· is frequently a
tematively--Qne just prior to the dc pulse and
one near the end of the pulse (see Figure
21-13). The difference in current per pulse (6i)
is recorded as a function of the linearly in-
pool of mercury in the bottom of the cell. In creasing voltage. A differential curve results
Figure 21-11, it is depicted as a platinum elec- consisting of a current peak (see Figure 21-
trode separated from the analyte solution by . 14); the height of this peak is directly propor-
a fritted disk. tional to concentration.
The capacitor C 1 in the current amplifier One advantage of the derivative-type po-
circuit is employed to damp the current fluctu- larogram is that individual peak maxima can
ations associated with the dropping electrode. be observed for substances with half-wave po-
tentials differing by as little as 0.04 to 0.05 V;
in contrast, classical polarography requires a
potential difference of at least 0.2 V for resolu-
MODIFIED tion of waves. More important, however, dif-
VOLTAMMETRIC METHODS ferential pulse polarography increases the sen-
sitivity of the polarographic method by about
As mentioned earli,er (p. 622), classical polar- three orders of magn.itude. This enhancement
ography with a dropping mercury electrode is is illustrated in Figure 21-14. Note that a
limited to solutions more concentrated than classical polarogram for a solution containing
1O-5M; in additioh, potentials more positive 180 ppm of the antibiotic tetracycline gives
than 0.4 V versus the SeE are not accessible barely discernible waves; pulsed polarography,
because of the ready oxidation of mercury.
RIII1P PotlntiollBt Cell Amplifier
generotor end
Several modifications of the classical polaro-
recorder graphic procedure overcome these limitations
FIGURE 21-11 Schematic diagram showing the components of a modem, three-electrode to various degrees. Some of these modem '4 For odditional details, see: J. B. F1ato, Arrol. Chem~ ••••.
polarograph and cell. adaptations of the classical method are sum- (11~ 7SA (1972~
VoluueSWHP AE'I2 for Curront·voltoge
Sensitivity curve Current surge of current is not observed because the
Nome progrlm resolution
m..,rement time scale of the measurement is long relative

:LL-
(II CII•• icolor 10-oM 0.2 V int""",ls
to the lifetime of the momentary current. On
0.02 see
Iinear-lCln

l~~
pollrogrlphy
the other hand, in pulse polarography, the
E~ current measurement is made before the surge
Timo -E- has completely decay.ed. Thus, the current
measured contains both a diffusion-controlled
II
w]A
10-8M 0.05 V 1
(b) Diflorentill II l-----l 0.06 see
component and a component that has to do
pulsed
pollrogrlphy I I I with reducing the surface layer to the concen-
j+-Drop time'--j
I , 1-2 see tration demanded by the Nernst expression;
Timo -E- the total current is typically several times
larger than the diffusion current.

]A
(c) Ropid•••• n 10-7 M 0.05 V When the potential pulse is first applied to
pollrogrlphy FIGURE 21-13 Voltage program for differ-
the electrode, a nonfaradaic current surge also
,~ ential-pulsed polarography.
occurs, which increases the charge on the
Timo -E-

(d) Cyclic
voltlmmetry

'ML Time
:~ ;t=E==

(e) Altornlting·current
pollrography
E

Time
10-6M

10-oM
0.05 V

<~ -E(dc)-

t
~t~
(f) Stripping 0.1 V
voltommetry
,~ 20 mV/see j
<J)
1 ~
:>
(J
-E- 20nA
Time

FIGURE 21-12 Modified polarographic procedures. T

in contrast, provides two well~efined peaks at must occur in the surface layer around an
a concentration level that is 2 x 10- 3 that for electrode as the potential is suddenly in- -1.0 -1.1 -1.2 -1.3 -1.4 -.8 -.9 -1.0 -1.1 -1.2 -1.3 -1.4 -1.5
the classic wave, or 0.36 ppm. Note also that creased by 20 to 100 mY. If a reactive species Potential vs. SCE. V Potential vs. SCE. V
the current scale for l1i is in nA (nanoamperes) is present in this layer, there will be a surge of W W
or 10-3 pA. current that lowers the reactant concentration FIGURE 21-14 (a) Differential-pulsed polarogram. 0.36 ppm tetra-
The greater sensitivity of pulse polarog- to that demanded by the new potential. As cycline . HCI in 0.1 M acetate buffer, pH 4, PAR Model 174 polarographic
raphy can be attributed to two sources. The the equilibrium concentration for that poten- analyzer, dropping mercury electrode, SQ-mV pulse amplitude, I-see
first is an enhancement of the faradaic tial is approached, however, the current drop. (b) DC Polarogram. ISO ppm tetracycline' HCI in 0.1 M acetate
current, and the second is a decrease in the decays to a level just sufficient to counteract buffer, pH 4, similar conditions. [Reprinted with permission from J. B.
nonfaradaic charging current. To account for diffusion; that is, to the diffusion-controlled Flato, Anal. Chem .• 44 (11), TSA (1972). Copyright by the American
the former, let us consider the events that current. In classical polarography, the initial Chemical Society.]
drop (p. 616). This current, however, decays is considerably more favorable than in normal
exponentially with time and approaches zero polarography. Quantitative analysis thus be-
near the end of the life of a drop when its surface comes possible for solutions as dilute as 10-6
area is changing only slightly (see Figure to 10-7 M. In addition, the shapes of rapid- r
21-3a). Thus, by measuring currents attbis time scan polarograms are such th!lt accurate cur- 1-
only, the residual current is virtually eliminated rent measurements are possible even when the t +2
half-wave potentials of two species differ by as
and the signal-to-noise ratio is larger. En-
little as 0.05 V.
a
t:

hanced sensitivity results.


Reliable instruments for pulse polarography Not surprisingly, rapid-scan equipment is
are now available commercially at reasonable more complex and more expensive than that
cost. The method has thus become an electro- required for ordinary polarography.
analytical tool of considerable importance.

Cyclic Voltammetry
Rapid-Scan The apparatus and techniques of cyclic polar-
(Oscillographic) Polarography FIGURE 21-16 Rapid-span po- ography or voltammetry are similar to those
larogram of a two~mponent mix- for rapid-scan polarography, which were con-
In rapid-scan polarography, the applied
potential is swept over a range of perhaps ture. (Taken from H. Schmidt and M. sidered in the previous section. The difference
0.5 V during the lifetime (or part ofthe lifetime) von Stackelberg, Modern Polarog- between the two lies in the voltage sweep pro-
of a single mercury drop. An oscilloscope is
raphic Methods, New York: Academic gram (coinpare Figures 21-12c and 21-12d).
Press, 1963, p. 27. With permission.) As the electrode reaction becomes slower or
required to display the current-voltage rela- In cyclic yoltammetry, the sweep takes a tri-
less reversible, this difference becomes greater.
tionship; for a permanent record, the image on angular form; that is, the potential is first in-
The effect of a slow electrode reaction rate is
the oscilloscope screen may be photographed. creased lihearly to a peak and then decreased
shown in the other two curves in Figure 21-16.
Alternatively, the data can be stored in a com- to its starting point at the same rate. Generally,
Cyclic voltammetry has become an impor-
puter for subsequent plotting. the cycle is completed in a fraction of a second
tant tool for the study of mechanisms and
The most successful rapid-scan procedure to a few seconds. With a dropping electrode,
rates of oxidation-reduction processes, partic-
appears to be one in which the voltage scan RT 0.0282 the sweep is timed to occur near the end of
E. = E1/2 - 1.1- = E1/2 --- ularly in organic and metal-i>rganic systems.
occurs over the final 2 to 3 s of a drop that n F n the lifetime of the drop to minimize the non-
Usually platinum is used for fabrication of
has a total lifetime of 6 to 7 s. As mentioned (2.1-16) faradaic current.
the microelectrode.
in the previous section, the residual current Figure 21-16 shows three cyclic voltammet-
under these circumstances is small when ric curves. The solid curve is for a reversible
compared with those encountered in ordinary reaction. The cause of cathodic peak A is the
polarography. same as that for the peaks encountered in Alternating-Current Polarography
As will be seen from Figure 21-15, a rapid- rapid-scan polarography. When the potential
In this modification, a constant ac potential
scan polarogram also takes the form of a peak where A is the area of the electrode in cm2, v is first reversed (point B), the current remains
(sine wave) of a few millivolts is superimposed
or a summit. At the summit, the current is is the rate of scan in volts per second, and the positive and is largely due to the diffusion-
upon the dc potential normally employed in
made up of two components. One is the initial remaining terms are the same as in the Ilko- controlled reduction of the analyte. Ulti-
polarography. As in the conventional experi-
current surge required to adjust the concentra- vic equation (p. 612). Generally, a summit mately, however, potential C is reached at
ment, the dc potential is varied over its usual
tion of the surface film to an equilibrium state current exceeds a diffusion current by a factor which thF analyte is no longer reduced; the
range; here, however, the alternating current
demanded by the Nernst equation. The second of ten or more. current here is zero. With further positive
is measured.
is the normal diffusion~ntrolled current. The Perhaps the most important advantage changes in the potential, oxidation of the pre- Consider the application of this technique
first current then decays as equilibrium is that can be cited for rapid-scan Polarography viously reduced species begins and proceeds
to a solution containing a single, easily reduced
approached and ultimately only the diffusion- is its enhanced sensitivity. Not only are until its concentration reaches zero. The anodic species Ox, which, at a suitable potential, reacts
controlled current remains. summit currents larger than diffusion cur- peak D results. For a reversible reaction, the
rapidly and reversibly as follows:
It can be shown that at 25°C the summit rents, but the residual currents are also small- theoretical potential difference between peaks
potential E. for a reversible process is related er; therefore, the signal-to-background ratio A and D is (2 x 0.0282/n) V (Equation 21-16). Ox + ne~Red
Assume also that an excess of supporting potentials slightly greater than the half-wave layer surrounding. an electrode. Thus, ac po- of the current at the end of each half-cycle
electrolyte is present. potential, larographic peaks are superimposed upon a . only, the base line current becomes vanishingly
We must first focus our attention on the base line of nonfaradaic current. The double small.
[Red]o > [Ox]o > 0 layer structure is, however, affected !?y the dc Alternating-current polarography has the
faradaic alternating current in this cell; that
is, the current that results from the oxidation- and here, the magnitude of the alternating component; as a consequence, the b~ line of lU1Yantageof enhanced sensitivity, and in con-
reduction process at the interface. In order for current is controlled by [Ox]o. At the half- an ac polarogram is not horizont8I. For- trast to classical polarography, permits the
such a current to exist, it is clearly necessary wave potential, tunately, over a small potential range, the ready determination of a trace of a reducible
to have both an oxidizable and a reducible nonfaradaic current is nearly linear, and a sat- species in the presence of a much larger con-
[Red]o = [Ox]o
species at the microelectrode surface. Signifi- isfactory correction can be made by linear ex- centration of a substance that gives a wave at
cant concentrations of Red occur, however, and the faradaic current will have reached its trapolation between the base lines on either lower potentials. The technique is also useful
only when the applied dc potential corresponds maximum value. When a potential corre- side ora peak (see Figure 21-17). for the study of electrode kineti~. The disad-
to the steeply rising portion of the dc po- sponding to the diffusion current region is The height of an ac polarographic peak is vantage of the procedure is, of course, the
larogram. Thus, no faradaic alternating current reached, [Ox]o --+0, and the alternating cur- proportional to bulk concentration of the complex nature of the equipment required.
will be observed before the decomposition rent will, for lack of a reducible species, again anaIyte and, for a reversible reaction, to the
potential has been reached. Beyond the decom- approach zero. Thus, a typical ac polarogram square root of frequency. For reversible elec-
position potential, but short of the half-wave consists of one or more peaks, the maxima trode reactions, ac polarography is more sen- Stripping Method.
potential, the following relationship exists occurring at the half-wave potentials for each sitive than the classical procedure and Stripping analysis encompasses a variety of
between the surface concentration of Ox and of the reactive species present. The relation- permits analysis of solutions as dilute as electrochemical methods having a common,
Red ship between ac and classical dc polarograms 10-'M. characteristic initial step.15 In all of these pro-
is shown in Figure 21-17. No ac peak is observed for a totally non- cedures, the analyte is first collected by electro-
o < [Red]o < [Ox]o As noted in Chapter 18 (p. 506), alternat- reversible half-reaction (that is, where the rate deposition at a mercury or a solid electrode;
and the magnitude of the alternating current ing currents also result from a nonfaradaic of the half-reaction is very slow in one direc- it is then redissolved (stripped) [rom the elec-
will be determined by [Red]o. At applied dc process that involves charging of the double tion~ since the product generated by the de trode to produce a more concentrated solution
current does not provide sufficient reactant than originally existed. The analysis may then
for transport of the alternating current during be based either upon electrical measurements
the oxidation half-cycle. For electrode pro- made during the actual stripping process or
cesses where the rate of one or the other half- upon some electroanalytical measurement of
reaction is less than instantaneous, but still the more concentrated solution.
appreciable. the magnitude of the alternating Stripping methods are of prime impor-
current becomes dependent upon the frequency. tance in trace work because the concentration
of the ac source. As the frequency is increased, aspects of the electrolysis permit the deter-
a condition is ultimately reached in which the mination of minute amounts of an analyte
time lapse of a half-cycle is insufficient for with reasonable accuracy. Thus, the analysis
completion of the slow half-reaction; the alter- of solutions in the 10-6 to 10-9M range be-
nating current will then be limited by the comes feasible by methods that are both
reaction rate, and the current maximum will
fir"-
I

Classical
be smaller than for an entirely reversible
process. Thus, ac polarography is useful for
simple and rapid.
The stripping method that has had the

kinetic studies of electrode reactions.


~-L_ac -0.6
Alternating-current polarography
current having a square wave rather than a
with
" For detailed discussions of stripping methods, ICe: F.
sinusoidal form has also been inve$tigated Vydra, K. Stulik, and E.1ulakova, Electroclwmkal Srrip-
de potent;'l. v ping AlI4I1ysls. New York: Halsted, 1971; I. Shain, Strip-
(square-wave pokJrography). The advantage of
ping AlI4I1ysls. in Tnatl# 011 All<llyt/cQ1 Chernlstry, eda. I.
FIGURE 21-17 Comparison of ac and classical polarograms. the square wave form arises from the expo-
M. Kolthoff and P. 1. Elving. New York: Intcnc:iencc,
[11te right-hand plot is taken from B. Breyer, F. Gutmann, and nential decay of the nonfaradaic current with 1963, Part I, vol. 4, Chapter SO;W. D. Ellis, J. C"--
S. Hacobian, Australian J. Sci. Research, AJ, 567 (1950). With time and its approach to zero near the end of £due, 50, A131 (1913); and T. R. Copeland and R. K.
permission.] each half-cycle. By intermittent measurement Skogerboe. Anal. Chern, <t6(14~ 12S7A(1974~
most widespread application makes use of a methods have been devised to produce a micro- from a 1 x to-8M solution by application of
micro mercury electrode for the deposition electrode of reproducible dimensions, which is a potential of about -0.9 V (Vs.SCE~ which
process, followed by anodic voltammetric mea- essential for quantitative work. Among these is about 0.3 V more negative than the half-
is the hanging drop electrode, which is illus- ~.O2 wave potential for this ion. After 15 min of
surement of the analyte; discussion in the para-
graphs that follow will be largely confined to trated in Figure 21-18. Here, an ordinary <:a. electrolysis, stirring was discontinued; 30 s
this particular application. dropping mercury capillary provides a means i later, the potential was decreased at a rate of
Electrodeposition Step. Ordinarily, only a of transferring a reproducible quantity of mer- ~ ~.04
21 mY/so A rapid increase in anodic current
(3
fraction of the analyte is deposited during the cury (usually 1 to 3 drops) to a Teflon scoop. occurred about -0.65 Vasa result of the
electrodeposition step; hence, quantitative re- Note that the dropping mercury capillary does reaction
sults depend not only upon control of elec- not serve as an electrode in this application. . Cd(Hg) ....•Cd2+ + Hg + 2e
trode potential but also upon stjch factors as The hanging drop electrode is then formed by
electrode size, length of deposition, and stir- rotating the scoop and bringing the mercury The current decayed after reaching a maxi-
ring rate for both the sample and a standard into contact with a platinum wire sealed in a mum, owing to depletion of elemental cad-
solution employed for calibration. glass tube. The drop adheres strongly enough FIGURE 21-19 Curve A: Current- mium in the hanging drop. The peak current,
Mercury electrodes of various forms have so that the solution can be stirred; it can, voltage curve for anodic stripping of after correction for the residual current (curve
been most widely used for the electrolysis, al- however, be dislodged by tapping the electrode cadmium. Curve B: Residual current B in Figure 21-19~ was directly proportional
though platinum or other inert metals can also at the completion of the electrolysis. curve for blank. [Reprinted with per- to the concentration of cadmium ions over a
be employed. Generally, it is desirable to mini- To use this apparatus, the drop is formed, mission and adapted from R. D. DeMars range of 10-6 to 1O-9M and inversely pro-
mize the volume of the mercury to enhance the stirring is begun, and a potential is applied and I. Shain, Anal. Chern., 29, 1825 (1957). portional to deposition time. This analysis was
. concentration of the deposited s~ies. Several that is a few tenths of a volt more negative Copyright by the American Chemical based on calibration with standard solutions
Society.] of cadmium ion. With reasonable care, an
analytical precision of about 2 % relative can
be obtained. Mixtures can be resolved and
analyzed by controlling the potential of
deposition.
than the half-wave potential for the ion of Many other variations of stripping analysis
interest. Deposition is allowed to occur for a have been proposed. For example, a number
carefully measured period; 5 min usually suffice of metals have been determined by electro-
for solutions that are to-7M or greater, 15 deposition on a platinum cathode. The quan-
min for to-8M solutions, and 60 min for those tity of electricity required to remove the deposit
Dropping
mercury that are 1O-9M.1t should be emphasized that is then measured coulometrically. Here again,
electrode these times seldom result in complete removal the method is particularly advantageous for
without electrical
contact of the ion. The electrolysis period is determined trace analysis.
by the sensitivity of the method ultimately em-
ployed for completion of the analysis.
Voitammetric Completion of the Analysis.
Teflon The analyte collected in the hanging drop
scoop
for transfer
electrode can be determined by a voltammet-
of drops ric measurement. Here, stirring is discon- Voltammetric methods can be employed to
tinued for perhaps 30 s after termination of estimate the equivalence point of titrations,
the deposition. The voltage is then decreased provided at least one of the participants or
at a fixed rate from its original cathodic value products of the reaction involved is oxidized
toward the anodic, and the resulting anodic or reduced at a microelectrode. Here, the cur·
current is recorded as a function of the rent at some fixed potential is measured as a
applied voltage. This linear scan produces a function of the reagent volume (or of time if
curve of the type shown in Figure 21-19. In the reagent is generated by a constant-current
this experiment, cadmium was first deposited coulometric process). Plots of the data on
lead ions are removed from the solution by ing oxidizing agents that attack mercury
precipitation. The curvature near the equiv- [bromine, silver ion, cerium(IV~ among
alence point reflects the incompleteness of the others l, a rotating platinum electrode is
precipitation reaction in this region. The end preferable.
point is obtained by extrapolation ofthe linear
portions, as shown.
Figure 21-20b is typical of a titration in Application of
which the reagent reacts at the microelectrode Salt bridge . Amperometric Titrations
connection
and the analyte does not; an example is the . :0 reference
titration of magnesium with 8-hydroxy- electrode As shown in Table 21-2, the amperometric
FIGURE 21-20 Typical amperometric ti- . end point has been largely confined to titra-
quinoline. A diffusion current for the latter is
tration curves. (a) Analyte is reduced, reagent : tions in which a precipitate or a stable com-
is not. (b) Reagent is reduced, analyte is not. obtained at - 1.6 V, whereas magnesium ion
is inert at this potential. plex is the product. A notable exception is the'
(c) Both reagent and analyte are reduced. • application of the rotating platinum electrode
Figure 21-2Oc corresponds to the titration
to titrations with bromate ion in the presence
of lead ion with a chromate solution at an
applied potential greater than -1.0 V. Both i of bromide and hydrogen ions. For example,
: styrene in an acidic 75% methanol solution
either side of the equivalence point are straight lead and chromate ions give diffusion cur-
can be titrated with a standard solution of
lines with differing slopes; the end point is rents, and a minimum in the curve signals the
KBrO] that contains an excess of potassium
established by extrapolation to their inter- end point. This system would yield a curve
bromide. The reactions are
section. resembling that in Figure 21-20b at zero FIGURE 21~21 Typical cell arrange-
applied potential, since only chromate ions ment for amperometric titrations with a
An amperometric titration is inherently
are reduced under these conditions. rotating platinum electrode.
BrO) + 5Br- + 6H+ -+ 3Br2 + 3H20
more accurate than voltammetric methods and
less dependent upon the characteristics of the Amperometric titrations in which the micro- C6HsCH=CH2 + Br2 -+C6HsCHBrCH2Br
microelectrode and the supporting electrolyte. electrode serves as anode, or perhaps as anode
Furthermore, the temperature need not be for one reactant and cathode for another, have A titration curve similar to that in Figure
fixed accurately, although it must be kept also been reported in the literature. The curves 21-20b is observed.
constant during the titration. Finally, the sub- for such titrations are interpreted in the same (V + v)/V, where V is the original volume and
stance being determined need not be reactive way as the examples cited in the previous v is the titrant volume. An alternative, which
at the electrode; a reactive reagent or product paragraphs. It should be recalled that an is often satisfactory, is to use a reagent that is
is equally satisfactory.16 anodic current is provided with a minus sign 20 (or more) times as concentrated as the so- AMPEROMETRIC
by convention. lution being' titrated. Under these circum- TITRATIONS WITH TWO
stances, v is so small with respect to V that POLARIZED MICROELECTRODES
Titration Curve. the correction is negligible. This approach re-
Apparatu. and Techniques
Amperometric titration curves typically take quires the use of a microburet so that a total A convenient modification of the amperomet-
one of the forms shown in Figure 21-20. Accurate amperometric titrations can be . reagent volume of 1 or 2 ml can be measured ric method involves the use of two identical
Figure 21-20a represents a titration in which achieved with relatively simple apparatus. . with a suitable accuracy. The microburet stationary microelectrodes immersed in a well-
the analyte reacts at the electrode while the Cells. Figure 21-21 shows a typical cell for should be so arranged that its tip can be stirred solution oft he sample. A small potential
reagent does not. The titration of lead with an amperometric titration. A calomel half-cell touched to the surface of the solution after (say, 0.1 to 0.2 V) is applied between these
sulfate or oxalate ions may be cited as an is usually employed as the nonpolarizable elec- each addition of reagent to permit removal of electrodes, and the resulting current is followed
example. The potential that is applied is t~ode; the indicator electrode may be a drop- the fraction of a drop that tends to remain as a function of the volume of added reagent.
sufficient to give a diffusion current for lead; pmg mercury electrode or a rotating microwire attached. The end point is marked by a sudden current
a linear decrease in current is observed as electrode, as shown. The cell should have a E}ectricaI MeulB'emenes. A simple manual rise from zero, a decrease in the current to zero,
capacity of 7S to 100 mi. polarograph is entirely adequate for amper- or a minimum (at zero) in a V-shaped curve.
. Volmne Meauremenes. For linear plots, it ometric titrations. Although the use of two polarized electrodes
16 For a detailed ditc:uuioD of amperometric titration, IS necessary to correct for volume changes Electrodes. Many amperometric titrations for end point detection was first proposed
see: J. T. Stoclc, Amperomnrlc Tltratiolu. New York: due to the added titrant. The measured cur- can be carried out conveniently with a drop- before 1900, almost 30 years passed before
tnterscienee, 1965. rents can be multiplied by the quantity ping mercury electrode. For reactions involv- chemists came to appreciate the potentialities
of the method. I 7 The name dead-stop end point
l/\,~t I ~~t
/ 1/\ ~~t
Li-uJLi
and (c) are obtained when only one of the
was used to descnbe the technique, and this reactants exhibits reversible behavior.
term is still occasionally encountered. It was Titration Curves When Both Systems Are
not until about 1950 that a clear interpretation Reversible. The titration of iron(I1) with
of dead-stop titration curves was made.18 cerium(IV) is an example of a fully reversible
system. That is, the equilibrium half-reaction

Oxidation-Reduction Titration.
FIGURE 21-22 End points for amperometric oxidation-
·Twin polarized platinum microelectrodes are is established at a platinum electrode essen- reduction titrations with twin-polarized electrodes. (a) Bolh
·conveniently used for end point detection for tially instantaneously; so also is the equilib- reactants behave reversibly at the electrode. (b) Only reagent
•oxidation-reduction titrations. Figure 21-22 rium described by the equation behaves reversibly. (c) Only analyte behaves reversibly.
illustrates three types of titration curves that
are commonly encountered. Curve (a) is ob-
: served when both reactant systems are revers- When a potential of 0.1 V is applied to a
ible with respect to the electrodes. CUl'Ves(b) pair of platinum electrodes immersed in a so-
lution containing both iron(I1) and iron(III~ a taneously in the surface film surrounding the system. Here, the electrode processes are
current is observed, owing to the following electrodes; that is, overvoltage effects are
electrode reactions: cathode Ce4+ + e ~ Ce3+
negligible. Since the electrodes are also small,
17 C. W. Foulk and A. T. Bawdeu, J. Alliei'. CIrem. Soc. Anode Ce3+ .•• Ce4+ + e
Cathode Fe3+ + e~FeH concentration polarization will surely occur
.ai, 2045 (1926~
at the potential that has been applied; the
IIFor an excellent analysis of this type of end point, see: Anode FeH ~Fe3+ + e magnitude of the current will thus be deter- The rates of these reactions are also large,
J. J. Linpile, Electroanalytical Chemistry, 2d ed. New
mined by the rate at which the reactant in lesser and the magnitude of the observed currents
York: lnterscience, 1958, pp. 280-294. depends upon the concentration of the cerium
concentration is brought to the electrode sur-
face. For example, if the concentration of species that is present in lesser amount.
iron(lI) is smaller than that for iron(III), con- Let us now consider the curve for the titra-
centration polarization will occur at the anode, tion of iron(I1) with cerium(IV) as shown in
and the magnitude of the current will be deter- Figure 21-22a. At the outset, no current is
mined by the concentration of the former. If observed because no suitable cathode reac-
iron(lI) is in excess, cathodic polarization will tant is available. With addition of cerium(IV~
Reaction Type
Reagent Product Electrode" occur, and the current will depend upon the a mixture of iron (III) and iron(ll) is produced,
iron(III) concentration. If the iron(III) concen- which permits the passage of current. Initially,
K2Cr04 Precipitate DME PbH, BaH tration is zero (as it would be at the initial the magnitude of the current is determined by
Pb(N03h Precipitate DME SO~-,MoO~-,F-, Cl- point in the titration), no current will be the iron(I1I) concentration. Beyond the mid-
8-Hydroxyquinoline Precipitate DME Mg2+, Zn2+, Cu2+, Cd2+, point in the titration, the concentrati6n of .
observed unless some alternative cathode pro-
AI3+, Bi3+, Fe3+ cess can occur at the small applied potential. this species exceeds that ofiron(I1); the current
Cupferron Precipitate DME CuH, Fe3+ A possible reaction would be is then regulated by the decreasing iron(I1)
Dimethylglyoxime Precipitate DME NiH concentration. At the equivalence point, the
cc-Nitroso-p-naphthol Precipitate DME COH, CuH, PdH concentrations of both iron(I1) and cerium(IV)
K4Fe(CN)6 Precipitate DME Zn2+ are vanishingly small. Here, possible electrode
AgN03 . Precipitate RP Cl-, Br-, 1-, CN-, RSH reactions are
The overvoltage of hydrogen on platinum is
EDTA Complex DME Bil+, CdH, Cu2+, ca2+, etc. sufficiently great, however, to rule out this
KBr03, KBr Substitution, cathode Fe3+ + e~FeH
RP Certain phenols,· aromatic prospect. Thus, no current is observed be-
addition, amines, olefins; N2H4, cause of complete polarization of the cathode. H+ +e~iH2
or oxidation As(III~ Sb(III) The behavior of the twin electrode system Anode Ce3+ ~Ce4+ + e
in a solution containing both cerium(III) and
cerium(IV) is analogous to that in the iron H20~i02 + 2H+ + 2e
A potential considerably greater than 0.1 V is Here, the- current depends upon the concen- 1. Calculate the Ni concentration (mg/liter) on the basis of the following
required tp cause any possible combination of tration of excess iodine. polarographic data:
these reactions to occur. Thus, the system is Figure 21-22c shows the titration of a
completely polarized and the current is zero. .dilute solution of iodine with thiosulfate ion.
Current at
Beyond the equivalence point, depolariza- in the initial stages, iodine and iodide are Solution -1.1 V, JlA
tion occurs because of the excess cerium (IV) present, and both react reversibly at the elec-
ions. Here, the electrode processes are trodes; the current depends upon the concen~ 25.0 ml of O.2F NaCI diluted to 50.0 ml 8.4
tration of the species present in lesser amount. 25.0 ml of O.2F NaCI plus 10.0 ml of sample 46.3
Cathode ee4+ + e¢ee3+ Thiosulfate does not behave reversibly at the diluted to 50.0 ml
electrodes; therefore, the current remains at
Anode ee3+ ¢ Ce4+ + e 25.0 ml of O.2F NaCI, 10.0 ml of sample,
zero beyond the equivalence point. 5.00 ml of 2.30 x 10-zM NiH diluted
and the current increases as more and more to 50.0 ml
cerium(IV) is added to the solution.
Precipitation Titrations
Titration Curves When Only One System is
Reversible. Let us now consider the behavior Twin silver microelectrodes have been em- 2. Calculate the concentration of Al (mg/liter) on the basis of the follow-
of a twin electrode system in the titration of ployed for end point detection in titrations ing polarographic data:
an analyte that reacts only slowly at the elec- involving silver ion. For example, in the titra-
trodes (that is, a nonreversible system). An tion of silver ion with a standard solution of
Current at
example is encountered in the titration of chloride ion, currents proportional to th:e -1.7 V, JlA
arsenic(I1I) with iodine. In the early stages of metal ion concentration would result from
this titration, the solution contains appreciable the reactions 20.0 ml of O.20F HCI plus 20.0 ml of HzO 10.2
concentrations ofH3As03, H3As04, and 1-. 20.0 ml of 0.20F HCI plus 10.0 ml of sample 33.3
Possible electrode processes that might occur Cathode Ag+ + e¢Ag(s) plus 10.0 ml of H20
upon application of a potential are Anode Ag(s)¢Ag+ + e 20.0 ml of 0.20F HCI plus 10.0 ml of sample
plus 10.0 ml of 6.32 x 10- 3M AP +
Cathode With the effective removal of silver ion by
the analytical reaction, cathodic polarization
H3As04 + 2H+ + 2e¢H3As03 + HzO
would occur, and the current would approach
3. The mercury from a dropping electrode was collected for 100 s and
Anode zero at ·the end point.
found to weigh 0.196 g. The time required for 10 drops of mercury to
H3As03 + HzO¢H3As04 + 2H+ + 2e form was 43.2 s. When this electrode was employed with a standard
1.00 x 10- 3M solution of Pb2 +, a current of 8.76 JlA was observed.
2.- ¢12 + 2e Applications
Subsequently, an unknown lead solution produced a current of
The amperometric method with two micro- 16.311lA with a new electrode which had a drop time of 6.13 s and a
The reactions involving the two arsenic
electrodes has been widely applied to titrations flow rate of 3.85 mg,ls. Calculate the molar concentration of Pbz + in
species are slow at a platinum surface,
involving iodine;it is also useful with reage"ts the unknown.
however, and can be made to occur only by
applying an overpotential of several tenths of such as bromine, titanium(I1I), and cerium(IV). 4. The following data were collected for three dropping electrodes. Com-
An important use is in the titration of water plete the data for electrodes A and C.
a volt. Thus, until the equivalence point is
passed, no significant current is observed (see with the Karl Fischer reagent. The technique
has also been applied to end point detection
Figure 21-22b). A B C
Beyond the equivalence point, depolariza- • for coulometric titrations.
The principal advantage of the twin micro- Flow rate, mg,ls 1.89 4.24 3.11
tion of the cell can occur at 0.1 V by virtue of
electrode procedure is its simplicity. No refer- Drop time, s 2.12 5.84 3.87
the reactions
ence electrode is required, and the only intru- il./C, JlA liter mmol- 1 4.86
Cathode Iz + 2e -+ 2.- mentation needed is a simple voltage divider,
powered by a dry cell, and a galvanometer or
Anode 21- -+ Iz + 2e microammeter for current detection.
5. Electrode C in Problem 4 was employed to study the reduction of an
organic compound known to have a diffusion coefficient of 7.3 x
10-6 cm2/s. A 5.00 x 1O-4M solution of the compound yielded a diffu-
sion current of 8.6 pA. Calculate n for the reaction.
6. An organic compound underwent atwo-e1ectron reduction at dropping
electrode A in Problem 4. A diffusion current of 9.6 pA was produced
by a 1.17 x 1O-3M solution of the compound. Calculate the diffusion
coefficient for the compound.
7. The half-wave potential for Ni(II) in a 0.10F NaCI04 solution was
found to be -1.02 V (vs. SCE~ In a solution that was 0.10F in
NaCI04 and O.1ooF in ethylenediamine (en~ the half-wave potential
was -1.60 V; calculate the formation constant of the complex between
the two species assuming its formula is Ni(en)~+.
8. The following polarographic data were obtained for the reduction of
Pb2+ to its amalgam from solutions that were 2.00 x 1O-3M in Pb2+,
O.1ooF in KN03, and that also had the following concentrations of the
anion A -. From the half-wave potentials, derive the formula of the
complex as well as its formation constant.

Conen A-,M

0.‫סס‬OO -0.405
0.0200 -0.473
0.0600 -0.507
0.1007 -0.516
0.300 -0.547
0.500 -0.558

9. A 1.00 x 1O-3M solution of europium(III) in O.looF KN03 has a


polarographic wave for the reversible reduction of Eu3+ to the amal-
gam. The following data show the effect of increasing concentration of
the anion X2 - on the half-wave potential.

E1/2 vs. SeE, V


0.000 -0.692
0.0200 -1.083
0.0600 -1.113
0.100 -1.128
0.300 -1.152
0.500 -1.170
tween electrodes spaced one centimeter apart.2 Equivalent Conductance at Infinite DilutiolL
electrolyte concentration. Indeed, the change
Conduction of electricity through an electro- Neither the volume of the solution nor the area The mobility of an ion in solution is governed
in conductance due to the addition of titrant
lyte solution involves migration of positively of the electrodes is specified; these vary to by four forces. An electrical force, equal to the
can become largely masked in solutions with
charged species toward the cathode and nega- satisfy the conditions of the definition. For product of the potential of the electrode and
high salt concentrations; under these circum-
tively charged ones toward the anode. The example, a 1.0N solution (1.0 gram equivalent the charge of the ion, tends to move the par-
stances, the method cannot be used.
conductance, which is a measure ofthe current per liter) would require electrod"es with indi- ticle toward one of the electrodes. This effect
that results from the application of a unit elec- vidual surface areas of 1000 em2; a O.lN solu- is partially balanced by a frictional force that
trical force, depends directly upon the number tion would need 10,000-cm2 electrodes. The is a characteristic property for each ion. These
of charged particles in the solution. All ions direct measurement of equivalent conductance are the only two effects that playa significant
contribute to the conduction process, but the is thus seldom, if ever, undertaken because of role in determining the conductivity of a dilute
fraction of current carried by any given species Upon application of a potential, ions in a so- the experimental inconvenience associated solution; under these circumstances the
is determined by its relative concentration and lution are essentially instantly accelerated to- with such relatively large electrodes. Instead, equivalent conductance of a salt is inde~ndent
its inherent mobility in the medium. ward the electrode of opposite charge. The rate this quantity is determined indirectly from of its concentration.
The application of direct conductance mea- at which they migrate, however, is limited by specific conductance data. By definition, A will At finite concentrations, two other factors,
surements to analysis is limited because of the the frictional forces generated by their motion. be equal to G when one gram equivalent of the electrophoretic effect and the relaxation ef-
nonselective nature of this property. The As in a metallic conductor, the velocity of the solute is contained between electrodes spaced fect, become important and cause the equiv-
principal uses of direct meaSurements have particles is linearly related to the applied field; one centimeter apart. The volume V of the alent conductance of a substance to decrease
been confined to the analysis of binary water/ Ohm's law is thus obeyed by electrolyte solution (em3) that will contain one gram as its concentration increases. The behavior
electrolyte mixtures and to the determination equivalent of solute is given by of a sodium chloride solution is typical and is
solutions.
of total electrolyte concentration. The latter shown in Table 22-1.
measurement is particularly useful as a crite- V= 1000 The electrophoretic effect stems from the
rion of purity for distilled wa~er. Some Important Relationships C motion of the oppositely charged ions sur-
On the other hand, conductqnetric titrations, Conductance G. The conductance of a so- rounding the ion of interest. These ions carry
where C is the concentration in equivalents
in which conductance measurements are used lution is the reciprocal of the electrical resist- with them molecules of solvent; the motion of
per liter. This volume can also be expressed in
for end-point detection, can be applied to the the primary particle is thus retarded by the
ance and has the units of ohm - t. That is, terms of the dimensions of the cell
determination of numerous substances! flow of solvent in the opposite direction. The
The principal advantage to the conduc- G=.!.. V= IA relaxation effect also owes its genesis to move-
tometric end point is its applicability to the R With I fixed by definition at one centimeter, ment of the ionic atmosphere surrounding a
titration of very dilute solutions and to systems given particle. Here, however, the ion is slowed
where R is the resistance in ohms.
in which the reaction between the titrant and V= A = 1000 by the charge of opposite sign that builds up
Specific Conductance k. Conductance is di-
analyte is relatively incomplete. Thus, for C behind the moving particle.
rectly proportional to the cross-sectional area
example, the conductometric titration of an For strong electrolytes, a linear relation-
A and inversely proportional to the length I Substitution into Equation 22-2 thus gives
aqueous phenol (K.:::: to-10) solution is ship exists between equivalent conductance
of a uniform conductor; thus,
feasible, even though the change in pH at the A = l000k
equivalence point is insufficient for either a A C
potentiometric or an indicator end point. G=k-
I
Conductometric titrations become less accu- Equation 22-3 permits calculation of the equiv-
rate and less satisfactory wit~ increasing total where k is a proportionality constant called alent conductance from the experimental value TABLE 22-1 EFFECT OF
the specific conductance. Clearly, it is the con- of k for a solution of known concentration. CONCENTRATION ON
ductance when A and I are numerically equal. EQUIVALENT CONDUCTANCE
If these parameters are based upon the centi-
;
meter, k is the conductance of a cube of liquid Concentration of Nan eqfIiter A
1 For further discussion of conduetometric methods, see:
J. W. Loveland, in Treatlu 011 AlIGlytlc4l/ CMmIstry, eds. one centimeter on a side. The dimensions of 2 In the context of electrolytic conductance, the equiva- 0.1 106.7
I. M. Kolthoff and P. J. ElviJI&. New York: Intcncience, specific conductance are then ohm - t em - 1. lent is defined in terms of the number of cbarses carried 0.01 118.5
1963, Part I, vol. 4, Chapter SI; and T. SbedIoVlky and Equivalent Conductance. The equivalent by an ion and not upon its fate in a specific reaction. 0.001 123.7
L. Sbedlovsky, in Physkal Methods ofClrem&try, eds. A. Thus, the weight of one gram equivalent of Ba2 + is equal Infinite dilution 126.4(Ao)
conductance A is defined as the conductance
Weissberger and B. W. ROIIitor. New York: Wiley-inter- to the atomic weight of barium divided by two.
science, 1971, vol. I, Part llA, Chapter 3. of one gram equivalent of solute contained be-
The equivalent ionic conductance is a mea- conductance measurements are made at suf- power source S provides an alternating cur-
and square root of the concentration. Extrap-
sure of the mobility of an ion under the in- ficiently low frequencies so that the latter mech- rent in the frequency range of 60 to 1000 Hz
olation of this straight-line relationship to
fluence of an electric forCe field and is thus anisms are not important. Oscillometry, on the at a potential of 6 to 10 V. The resistances
zero concentration yields a value for the equiv-
a gauge of its capacity for transporting elec- other hand, makes use of radio frequencies; RAe and RBC can be calculated from the posi-
alent conductance at infinite dilution Ao. A
similar plot for a weak electrolyte is nonlinear, tricity. For example, the ionic conductance of here, the dielectric current becomes important. tion of C. The cell, of unknown resistance R",
a potassium ion is nearly the same as that of Oscillometric measurements are discussed at is placed in the upper left arm of the bridge
and direct evaluation ot Ao is difficult.
a chloride ion; therefore, electricity passing the end of this chapter. and a precision variable resistance R. is placed
At infinite dilution, interionic attractions
become nil; the overall conductance of the through a potassium chloride solution is car- in the right-hand side. A null detector N D is
solution then consists of the sum of the indi- ried nearly equaliy by the two species. The used to indicate the absence of current between
vidual equivalent ionic conductances situation is quite different with hydrochloric D and C. The detector may consist of a pair
acid; because of the greater mobility of the THE of ordinary headphones, since the ear is re-
Ao=.l.~ +.l.~ hydrogen ion, a larger fraction of the electric- MEASUREMENT sponsive to frequencies in the 1000 Hz range;
ity [3SO/(350 + 7~) = 0.82] is carried by that OF CONDUCTANCE alternatively, it may be a "magic eye" tube, a
where .l.~ and .l.~ are the equivalent ionic species in an electrolysis. cathode ray tube, or a microammeter.
conductances of the cation and the anion of Ionic conductance data permit comparison A conductance measurement requires a source Capacitance effects within R" cause an al-
the salt at infinite dilution. Individual ionic of the relative conductivity of various solutes. of electrical power, a cell to contain the solu- ternating-current device such as this to suffer
conductances can be determined from other Thus, we are juStified in saying that O.OIF tion, and a suitable bridge to measure the . a loss in sensitivity when very high resistances
electrolytic measurements; values for a number hydrochloric acid will have a greater conduc- resistance of the solution. are measured; in practice, a variable capaCi-
of common ions are given in Table 22-2. Note tivity than O.OIF sodium chloride because of tor across R. compensates for this effect.
that symbols such as tMgH, iFe3+, arid
the very large ionic conductance of the hy- Power Sources Conductance can also be determined by
tSOi - are used to emphasize that the con-
drogen ion. Such' conclusions are important means of a simple electronic circuit such as
centration units are in equivalents per liter. in predicting the pourse of a conductometric Use of an alternating current source elimi- that shown in Figure 3-21a (p. 62). Here, the
The differences that exist among the equiv- nates the effect of faradaic currents. There are,
titration. ' conductance is read directly from a dc meter
alent ionic conductances of various species however, both upper and lower limits to the fre- or a recorder.
(Table 22-2) arise primarily from differences quencies that can be employed; audio oscil-
in their size and the degree of their hydration. Alternating Currents in a Cell lators that produce signals of about 1000 Hz
are the most satisfactory. Cells
As was pointed out earlier (p. S05~ conduc-
tion of dc electricity across a solution/electrode For less refined work, an ordinary 6O-cycle Figure 22-1 depicts three common types of
interface is a faradaic process in which oxida- current, stepped down from 110 to perhaps cells for the measurement of conductivity. Each
tion and reduction must occur at the two 10 V, is also used. With such a source, how- contains a pair of electrodes firmly fixed in a
electrodes. An alternating current, on the other ever, faradaic processes often limit the accu- constant geometry with respect to one another.
TABLE 22-2 EQUIVALENT IONIC racy of the conductance measurements. On
hand, requires no electrochemical reaction at The electrodes are ordinarily platinized to
CONDUCTANCES AT 25°C
the electrodes (p. 506); here, electricity flows the other hand, the convenience and ready increase their effective surface and thus their
>..~ Anion >..!. availability of 6O-cycle power justify its use
Cation as a consequence of nonfaradaic processes. capacitances; faradaic currents are minimized
Because the changes associated with faradaic for many purposes. as a result.
H3O+ 349.8 ow 199.0
conduction can materially alter the electrical Power sources with frequencies much great- Determination of the Ceil Constant. Ac-
Li+ 38.7 Cl- 76.3
characteristics of a cell, conductometric mea- er than 1000 Hz create problems in conduc- cording to Equation 22-2, the specific conduc-
Na+ SO.1 Br- 78.1
surements are 8dvantageously based upon tance measurements with a bridge. Here, the tance k is related to the measured conductance
K+ 73.5 1- 76.8
nonfaradaic processes. cell capacitance and stray capacitances in other G by the ratio of the distance separating the
NHt 73.4 NO; 71.4
At low frequepcies, conduction of ac elec- parts of the circuit cause phase changes in the electrodes to their surface area. This ratio has
Ag+ 61.9 CIO; 67.3 tricity through ari electrolyte involves periodic current for which compensation is difficult. a fixed and constant value in any given cell
tMg2+ 53.1 C2H3Oi 40.9
motion of the ions toward and away from and is known as the cell constant. Its value is
teaH 59.5 -tSO~- 80.0 seldom determined directly. Instead, the <:on-
the electrodes, as noted earlier (p. 506);
tBaH 63.6 too~- 69.3
however, at radio frequencies, a significant
Resistance Bridges ductance of a solution whose specific conduc-
tPbH 69.5 te20~- 742 fraction of ac electricity is transported in the The Wheatstone bridge arrangement, shown tance is reliably known is measured; the cell
iFe3+ 68.0 iFe(CN):- 110.5
form of a dielectric current which results from in Figure 2-24 (p. 35), is typical of the appara- constant can then be calculated. Solutions of
iLa3+ 69.6 potassium chloride are commonly chosen for
induced and orientation polarization. Most tus used for conductance measurements. The
The two linear portions are then extrapolated, methods, which require observations under
the point of intersection being taken as the conditions where reaction is least complete, a
equivalence point. conductometric analysis can be employed suc-
Because reactions fail to proceed to ab- cessfully for titrations based upon relatively
solute completion, conductonietric titration unfavorable equilibria.
curves invariably show departures from strict The conductometric end point is completely
linearity in the region of the equivalence point. nonspecific. Although it is potentially adapt-
Curved regions become more pronounced as able to all types of volumetric reactions, the
the reaction in question becomes less favorable number of its useful applications to oxidation-
and as the solution becomes more dilute. The reduction systems is limited; the substantial
linear portions of the curve are best defined excess of hydrogen ions typically needed for
by measurements sufficiently removed from such reactions tends to mask conductivity
the equivalence point so that the common ion changes associated with the volumetric reac-
effect forces the reaction more nearly to com- tion.
pletion. It is in this respect thllt the conducto-
metric technique appears to 'best advantage;
Acid-Ba.e Titration.
in contrast to potentiometric or indicator
Neutralization titrations are particularly well
cell calibration.3 Typical data are shown in some specific level is not required for a suc-
adapted to the conductometric end point be-
Table 22-3. cessful conductometric titration. For many
cause of the large ionic conductances of
Once the value of the cell constant has purposes, it is sufficient to immerse the cell in
hydrogen and hydroxide ions compared with
been determined, conductivity data can be a reasonably large bath of water or oil, which \. i the conductances of the species that replace
is at room temperature.
easily converted to terms of specific conduct- \ them in solution.
ance with the aid of Equation 22-2. Titration of Strong Acids or Bases. The
Temperature Control The temperature
\.
\ / solid line in Figure 22-2 represents a curve
coefficient for conductance measurements is
about 2"1oiC; as a consequence, some tem-
\ ~ / (corrected for volume change) obtained when

perature control is ordinarily required during Conductometric measurements provide a con-


\. \ /
hydrochloric- acid is titrated with sodium
hydroxide. Also plotted are the calculated
a conductometric measurement. Although a
constant temperature is necessary, control at
venient means for locating end points in titra-
tions. Sufficient measurements (three to four
\ 1\ / contributions of the individual ions to the
conductance of the solution. During neutral-
before and after the equivalence point) are
H.O·' r--.. / ,
OH-:/
ization, hydrogen ions are replaced by an
needed to define the titration curve. After being
\ \V // equivalent number ofless mobile sodium ions;
• G. Jones and B. C. Bradshaw, J. A"",r. C/oem. Soc, ss. corrected for volume change, the conductance
' , the conductance changes to lower values as a
\
-- .-
1780 (1933). data are plotted as a function of titrant volume.
CI-
\

,
~7
/ -- .-~,-
result of this substitution. At the equivalence
point, the concentrations of hydrogen and
1\-
No' I--~1-- hydroxide ions are at a minimum and the
'\
-'
1/'';
1-" solution exhibits its lowest conductance. A
reversal of slope occurs past the end point as
TABLE 22-3 CONDUCTANCE OF SOLUTIONS the sodium ion and hydroxide ion concentra-
FOR CELL CALIBRATION FIGURE 22-2 Conductonietrictitration of tions increase. With the exception of the im-
a strong acid with a strong base. The solid mediate equivalence-point region, an excellent
Grams KCI per Specific CODlIactance at ZSOC, line represents the titration curve, corrected linearity exists between conductance and the
1000I or Solutionin Vaa.n oIun-1 em-I
for volume change. Broken lines indicate the volume of base added; as a result, only three
71.1352 0.111342 contribution of the individual species, also or four observations on each side of the
7.41913 0.0128560 corrected for volume change, to the con- equivalence point are needed for an analysis.
0.745263 0.00140877 ductance of the solution. The percentage change in conductivity dur-
(K. = 6 x 10- 10) with strong base. This reac- species is so slight, however, that little or no what less mobile nitrate ions of the reagent· a
tion is so incomplete that a potentiometric or curvature occurs with the establishment of slight decrease in conductance results. After the
visual indicator end point is unsatisfactory. In the butTer region (see Figure 22-380 for exam- reaction is complete a rapid increase occurs
the early stages of the titration, a butTer is ple). As the strength ofthe acid (or base) be- owing to the addition of excess silver nitrate:
rapidly established that imparts to the solu- comes greater, so also does the extent of the . ?<>~uctometric methods based upon pre-
-C.10N " tion a relatively small and nearly constant hy- curvature in the early portions of the titration CipItatIon or complex-formation reactions are
VolumeNaOH drogen ion concentration. The added hydrox- curve. For weak acids or bases with dissocia- not so useful as those involving neutralization
lb)
-ide ions are consumed by this buffer and thus tion constants greater than about 10- 5, the processes. Conductance changes during these
do not directly contribute to the conductivity. curvature becomes so pronounced that an titrations are seldom as large as those observed

tbd]
.A gradual increase in conductance does result, end point cannot be distinguished . with acid-base reactions because no other ion
:however, owing to the increase in concentra- Figure 22-3c illustrates the titration of the approaches the conductance of either the
tion of sodium and of borate ions. With attain- same weak acid as in Figure 22-3b, but with hydrogen or the hydroxide ion. Such factors
:ment of the equivalence point, no further aqueous ammonia instead of sodium hydrox- as slowness of reaction and coprecipitation
borate is produced; additions of base then ide. Here, because the titrant is a weak elec- represent further sources of difficulty with
VolumeHel . cause a more rapid increase in conductance trolyte, the curve is essentially horizontal past precipitation reactions.
Cd) : due to the highly mobile hydroxide ions. the equivalence point. Use of ammonia as
Figure 22·3b illustrates the titration of a titrant actually provides a curve that can be
moderately weak acid, such as acetic acid extrapolated with less uncertainty than the APPLICATIONS OF DIRECT
(K. ~ 10-51 with sodium hydroxide. Non- corresponding curve based upon titration CONDUCTANCE MEASUREMENTS
linearity in the early portions of the titration with sodium hydroxide.
curve creates problems in establishing the end Figure 22-3d represents the titration curve Direct conductometric measurements sutTer
point; with concentrated solutions, however, for a weak base, such as acetate ion, with a from a lack of selectivity, since any charged
Volume AgN03 the titration is feasible. As before, we can standard solution of hydrochloric acid. The species contributes to the total conductance
(fl interpret this curve in light of the changes in addition of strong acid results in formation of of a solution. On the other hand, the high
comJMiition that occur. Here, the solution sodium chloride and undissociated acetic acid. sensitivity of the procedure makes it an im-
FIGURE 22-3 Typicalconductometricti-
initially has a moderate concentration of hy- The net effect is a slight rise in conductance portant analytical tool for certain applica-
tration curves. Titration of (a) a very weak
acid (K. ::::10-1°) with sodium hydroxide; drogen ions (-10 - 3 M). Addition of base re- due to the greater mobility of the chloride ion tions. As we have noted, an important use of
(b) a weak acid (K.:::: 10-5) with sodium sults in the establishment of a buffer system over that of the acetate ion it replaces. After the method has been for estimating the purity
and a consequent diminution in the hydrogen the end point has been passed, a sharp rise in of distilled or deionized water. The specific
hydroxide (Note that for O.ot N solutions,
conductance'x 10 is plotted); (c) a weak acid ion concentration. Concurrent with this de- conductance attends' the addition of excess conductance of pure water is only about
(K. ::::10-5) with aqueous ammonia; (d) the crease is an increase in the concentration of hydrogen ions. The conductometric method is S x 10- 8 ohm -I em -I; traces of an ionic im-
salt of a weak acid; (e) a mixture of hydro- sodium ions as well as the conjugate base convenient for the titration of salts whose purity will increase the conductance by an
chloric and acetic acids with sodium hy- of the acid. These two factors act in opposi- acidic or basic character is too weak to give order of magnitude or more.
droxide; and (I) chloride ion with silver tion to one another. At first, the decrease in satisfactory end points with indicators. Conductance measurements are also em-
hydrogen ion concentration predominates and Figure 22-3e is typical of the titration of a ployed for determining the concentration of
nitrate.
a decrease in conductance is observed. As the mixture of two acids that ditTer in degree of solutions containing a single strong electrolyte,
titration progresses, however, the pH becomes dissociation. The conductometric titration of such as solutions of the common alkalis or
stabilized (in the buffer region); the increase such mixtures frequently leads to more accu- acids. A nearly linear increase in conductance
ing the course of the titration of a strong acid in salt content then becomes the more im- rate results than those obtained with a poten- with concentration is observed for solutions
or base is the same regardless of the concen- portant factor, and a linear increase in con- tiometric method. containing as much as 20 % by weight of solute.
tration of the solution. Thus, very dilute solu- ductance finally' results. Beyond the equiv- Precipitation and Complex-Formation Analyses are based upon calibration curves.
tionS can be analyzed with an accuracy com- alence point, the curve steepens because of the TitratiolB. Figure 22-3fillustrates the con- Conductance measurements are also widely
parable to more concentrated ones. greater ionic conductance of hydroxide ion. ductance changes that occur during the titra- used to measure the salinity of sea water in
Titration of Weak Adds or Bases. Figure In principle, all titration curves for weak tion of sodium chloride with silver nitrate. oceanographic work.
22-3a illustrates application of the conducto- acids or bases contain the general features of The iriitial additions of reagent in effect cause Finally, conductance measurements yield
metric end point to the titration of boric acid Figure 22-3b. The ionization of very weak a substitution of chloride ions by the some- valuable information about association and
dissociation equilibria in aqueous solutions- Applications
are similar to those for conventional low-
provided, of course, that one or more of the
Analysis of Binary Mixtures. Oscillometric frequency condu~o~etric titrations. The ap-
reacting species is ionic.
measurements have been employed for the ~rance of the titration curves is often quite
determination of binary mixtures of nonionic ~ifferent, however; both V-shaped curves and
species. Such analyses are b8sed upon the Inverted V-type end points are encountered.
dielectric behavior of the mixtures and re- Curvature is also frequently observed; data
As mentioned earlier (p. 506~ the conduction quire that the dielectric constants for the two must, therefore, be collected in the immediate
of ac radio-frequency (10' to 107Hz) electricity co~ponents ~iffer significantly. For this appli- end-point region. For a given reaction the
through a solution is a complex process that cation, a~ Instrument based .on frequency titration curve depends upon the concentr~tion
involves not only ionic motion but also induced changes IS employed. Empirical calibration of the substance titrated as well as upon the
and orientation polarization of the molecules curves permit conversion of inStrument read- parameter measured (that is, the change in
in the medium. As a result of these various ings to concentration ratios. Examples of this frequency or the oscillator current).
conduction mechanisms, the impedance of a type of application include the analysis of The high-frequency method does not ap-
solution to the flow of current varies in an mIXtures of ethanol and nitrobenzene, benzene pear to yield titration data of higher accuracy
intricate way with the dielectric constant, the and chlorobenzene, alcohol and water and 0- than the classical method, and suffers the draw-
electrolyte concentration, and the frequency. and p-xylene. " backs of greater empiricism and of requiring
Interpretation ofthe relationships among these Titratio... Oscillometric measurements measurements near the end point. Thus, except
variables is difficult, and the use of high- ~ave ~ employed to locate the end point where the presence of electrodes interferes with
frequency currents could hardly be justified FIGURE 22-4 Sample cell and elec- In vanous types of titrations. The methods a titration, oscillometry offers little advantage.
were it not for the fact that the experimental trodes for high. frequency measurements.
measurement does not require direct contact
between the electrodes and the solution.
the sample and cell. In the former, one or more
variable capacitors are wired in parallel with For each ~f the' following problems, derive an approximate titration
Instruments the cell and the circuit is tuned to its resonance curve o~ specIfic conductance versus volume of reagent, employing the
assumptIon that
As shown in Figure 22-4, a typical cell for frequency in the absence of sample. The sample
is then introduced and the capacitance that (a) 100.0 ml of 1.00 x 1O-3N 50lution is being titrated with
high-frequency measurements consists of a 1.00 x 1O-2N reagent;
glass container designed to fit snugly between must be subtracted (by means of a calibrated
capacitor) to restore resonance is determined. (b) the volume change occurring as the titration proceeds can be
a pair of cylindrical metal electrodes. The elec- neglected;
trode and cell are then made part of the reso- For instruments based on the resistive im-
pedance of the sample, the change in oscillator (c) t~e .equivalen~ ionic conductances of the various ions are not
nant circuit of a sine-wave oscillator (p. 24); slgmficantly different from their equivalent ionic conductances at
the behavior of this resonant circuit can be current resulting from the introduction of
infinite dilution (Table 22-2).
interpreted by assuming that the solution in sample is measured.
Because of the limited scope of oscillometry, ~alcula~~ theoretical specific conductances of the mixtures after the follow-
the cell acts like a parallel capacitor and resis- Ing addItIons of reagent: 0.00, 2.00, 4.00, 6.00, 8.00, 10.0, 12.0, 14.0, 16.0 ml.
tance. The capacitive reactance depends upon we shall not dwell further on the somewhat
complicated instrumentation and complex in- 1. Ba(OHh with HCI
frequency and the dielectric constant of the
terrelations associated with this technique; the 2. Phenpl with KOH (use K. = 1.0 X 10-10 and All. = 30 for C6H,O-)
solution; the resistive impedance is also fre-
quency-dependent and is related to the number reader is referred to the treatment of the subject 3. Sodium acetate with HCl04 (assume K. for acetic acid = 1.8 x 10-')
by Reilley4 and Pungor.' 4. P~oPllnoic acid (All. = 35.8 for CH3CH2COO- and K. = 1.3 x 10-')
and kinds of ions in the solution as well. At
high frequencies, the resistive impedance is WIth NaOH
generally lower than that measured with 5. Propanoic acid with NH3
currents in the l000-Hz range. 6. Pb(N03h with Naa
4 C. N. Rci11ey, in New lrutnunental Methods in Electro-
Instruments are available commercially that 7. Pb(N03h with LiCI
cltntlsrry, ed. P. Delahay. New York: Intenciencc, 19S4,
are based upon either the change in frequency Chapter IS.
or the change in oscillator circuit current (at • E. Punlor, Oscillometry and CondllCtometry. New York:
constant frequency) caused by the presence of Pergamon Press, 1965.
Thermal methods are based upon the mea. balances ~f this type are offered, one with a
surement of the dynamic relationship between maximum load of 2.5 g and the other 100 g;
temperature and some property of a system the former is the more common. Typically,
such as mass, heat of reaction, or volume. .mass changes of 200 mg, or some fraction
Wendlandt lists twelve thermal'methods; four thereo~ can be followed with a precision of
of the most important of these will be con- about ±O.I % relative.
sidered in this chapter,l namely, thermogravi- Figure 23-1 is a schematic diagram of the
metric methods, differential thermal methods, Mettler thermobalance. The sample holder is"
differential scanning calorimetry, and thermo- housed in a furnace that is thermally isolated
metric titrations. from the remainder of the balance. A change.
in the sample mass causes a deflection of the'
beam, which interposes a light shutter be-
tween a lamp and one of two photodiodes.:
The resulting imbalance in the photodiode
In a thermogravimetric analysis, the mass of current is amplified and fed into coil E, which
sample is recorded continuously as its tem- is situated between the poles of the perma-'
perature is increased linearly from ambient to nent magnet F. The magnetic field generated by
as high as 1200°C. A plot of mass as a func- the current in the coil restores the beam to its
tion of temperature (a thermogram) provides original position. The amplified photodiode
both qualitative and quantitative information. current also determines the position of the;
pen of a recorder. The Mettler instrument has:
Apparatus several weight ranges (I, 10, 100, and 10001
The apparatus required for a thermogravi- mg) and has a reproducibility of ± 10 JIg. :
metric analysis includes: (I) a sensitive ana- Furnaces. The furnace of a thermogravi-
lytical balance; (2) a furnace; (3) a furnace metric apparatus is generally programmed to
temperature controller and programmer; and increase the temperature linearly at predeter-
(4) a recorder that provides a plot of sample
mass as a function of temperature. Often. aux-
iliary equipment is needed to provide an inert
atmosphere for the sample.
The Balance. The Cahn electromagnetic bal-
ance is used in the thermogravimetric systems
sold by several manufacturers. The sample
holder of this balance is suspended from a lever"
. arm attached to a D'Arsonval galvanometer
coil (p. 28). Deflection of the beam from its rest
position by a change in mass is detected photo-
electrically. The resulting photocurrent is then
amplified and fed into the galvanometer coil in
such a direction as to restore the beam to its
original position. The amplified current also FIGURE 23-1 Components of a thermal
determines the position of a recorder pen. Two balance. A, Beam; B, sample cup and holder;
C, counterweight; D, lamp and photodiodes;
E, coil; F, magnet; G, control amplifier; H,
I For a thorough treatment or thermal methods, see:
tare calculator; I, amplifier; and J, recorder.
W. W. Wendlandt, Thermol Methods of Analysis, 2d cd. (Courtesy Mettler Instrument Corp., Hights-
New York: Wiley, 1974. town, New Jersey.)
mined rates (typically, from 0.5 to 25°C/min~ SrC204 ,and BaC204 , while the mass between increased at a constant rate.2 In differential
The temperature range for most instruments about 560 and 620"C corresponds to the weight scanning calorimetry, the sample and a refer-
is from ambient to 1200"C. Temperatures are of the three carbonates. The weight change in ence substance are also subjected to a continu-
determined by a thermocouple located as close the next two steps results from the loss of ously increasing temperature; here, however,
as possible to the sample. Insulation and cool- carbon dioxide, as first CaO and then SrO are heat is added to the sample or to the reference
ing of the exterior of the furnace is required formed. Clearly, sufficient data are available in as necessary to maintain the two at identical
to avoid heat transfer to the balance. the thermogram to calculate the weight of each temperatures. The added heat, which is re-
of the three elements present in the sample. corded, compensates for that lost or gained as
Figure 23-3a is the derivative of the thermo- a consequence of endothermic or exothermic
Applications
gram shown in (b). Many modem instruments reactions occurring in the sample.
Figure 23-2 is a recorded thermogram ob- are equipped with electronic circuitry to pro-
tained by increasing the temperature of pure vide such a curve as well as the thermogram
General Principles
CaC204' H20 at a rate of 5"C/min. The clearly itself. The derivative curve may reveal informa-
defined horizontal regions correspond to tem- tion that is not detectable in the ordinary Figure 23-4 is a differential thermogram ob-
perature ranges in which the indicated ca,lcium thermogram. For example, the three peaks at tained by heating calcium oxalate monohy-
compounds are stable. This figure illustrates 140, 180, and 205"C suggest that the three drate in a flowing stream of air. The two mini-
one of the important applications of thermo- hydrates lose moisture at different tempera- ma indicate that the sample becomes cooler
gravimetry, namely, that of defining thermal tures. However, all appear to lose carbon than the reference material as a consequence
conditions necessary to produce a pure weigh- monoxide simultaneously and thus yield a of the heat absorbed by two endothermic pro-
ing foim for the gravimetric determinatjon of single sharp peak at 45O"C. cesses; equations for these decomposition re-
a species. Perhaps the most important applications actions are shown below the minima. The
Figure 23-3b illustrates an application of of thermogravimetric methods are found in single maximum indicates that the reaction to
thermogravimetry to the quantitative analysis , the study of polymers. Thermograms provide give calcium carbonate and carbon dioxide is
of a mixture of calcium, strontium, and barium. information about decomposition mechanisms exothermic. It is noteworthy that when the dif-
The three ions are first precipitated as the for various polymeric preparations. In addi- ferential thermogram is obtained in an inert
monohydrated oxalates. The mass in the tem- tion, the decomposition patterns are character- atmosphere, all three reactions are endother-
perature range between 250 and 260"C is that istic for each kind of polymer and can be used mic, and the maximum is replaced by a mini-
of the three anhydrous compounds, CaC204, for identification purposes. mum; here, the reaction product from the
decomposition of calcium oxalate is carbon
tb) Thermogram
monoxide rather than carbon dioxide (see
FIGURE 23-3 Decomposition of CaC204 • Figure 23-2).
H20, SrC204 • H20, and BaC204 • H20. Sources of Differential Thermogram Peaks.
The maxima and minima, such as those ap-
pearing in Figure 23-4, are termed peaks. Those
appearing above zero on the ordinate scale are
the consequence of exothermic processes; those
DIFFERENTIAL ,lIP. below zero are for endothermic ones. These
THERMAL ANALYSIS heat changes may be the result of physical or
AND DIFFERENTIAL chemical phenomena. Physical processes that
SCANNING CALORIMETRY are endothermic include fusion, vaporization,
sublimation, absorption, and desorption. Ad-
In differential thermal analysis, the heat ab- sorption is generally an exothermic physical
sorbed or emitted by a chemical system is
observed by measuring the temperature dif-
ference between that system and an inert refer-
2 For a concise review or this method and its applica-
FIGURE 23-2 A thermogram for decomposition of ence compound (often alumina, silicon carbide, tions, see: M. I. Pope and M. D. Judd, Different"" Ther-
CaC204• H20. or glass beads) as the temperatures of both are mal Analysu. Philadelphia: Heyden, 1977.
lar to those employed in thermogravimetry. which may be at room or ice-bath temperature.

.. Indeed, several commercial instruments are


designed to permit all three types of thermal
The output of this circuit provides a measure of
the sample temperature at any instant.
~ analysis. Generally, the sample and reference cham-
e::l iJ Figure 23-5 shows a commercial differen- ber in a differential thermal apparatus is de-
; w"
C> tial thermal-analyzer system. Weighed quanti- signed to permit the circulation of inert or reo
! 0
ties of the sample and reference material are active gases. Some systems also have the capa-
!c: held in the small pans labeled Sand R. The bility for operation at high or low pressures.
thermocouple on the right (labeled control For differential scanning calorimetry, indi-
~ TC) controls the rate at which the furnace vidual heaters, located as close as possible to
0 ~
iJ" must be heated in order to provide a linear the sample and reference vessels, are provided.
"0
c: temperature increase. the sample and refer- When the thermocouples indicate a tempera-
W
ence thermocouples are connected in series. ture difference, heat is added to the cooler of
Any current due to a temperature difference the two until temperature equality is restored.
0 between the two is amplified and used to deter- The rate of heating required to keep the
Sample lemperoture. ·c mine the position of a rt:corder pen. With the temperatures equal is recorded as a function
switch in position 1$, the sample thermo- of sample temperature. The ordinate of the
FIGURE 23-4
presence of °
Differential thermogram of CaC204' H20 in the
2; the rate of temperature increase was gOC/min. (From
Handbook of Analytical Chemistry, ed. L. Meites. New York: McGraw-
couple is connected not only to the reference
thermocouple but also to a reference junction,
differential thermogram can then be expressed
in units of calories or millicalories per second.
Hil~ 1963, p. 8-14. With permission.)

change, while crystalline transitions may be an exothermic reaction and a positive one for
either exothermic or endothermic. an endothermic process.
Chemical reactions also produce differen- For a given species, k' remains constant
tial peaks; both exothermic and endothermic provided that a number of variables such as
processes are, of course, possible (for exam- heating rate, particle size of the sample, and
ple, see Figure 23-4~ placement of the thermocouples are closely
Peak Areas. The peak areas for differential controlled. Under these circumstances, Equa-
thermograms depend upon the mass of the tion 23-1 can be employed to calculate the
sample m, the heat or enthalpy Mf of the chem- mass of the analyte from peak areas; here,
ical or physical process, and certain geometric k' AH can be determined by calibration. Al-
and heat conductivity factors. These variables ternatively, Equation 23-1 permits the deter-
are related by the equation3 mination of AH for species when k' and m
have been measured.
Gm AH ,
A = - -k- = -k m AH (23-1) The thermograms obtained by differential
scanning calorimetry are similar in appear-
where A is the peak area (AT x time), G is ance to differential thermograms. In this in-
a calibration factor that depends upon the stance, however, the constant k' in Equation Cooling
sample geometry, and k is a constant related 23-1 becomes independent of the temperature jacket

to the thermal conductivity of the sample. at which the reaction occurs. Sample TC
The enthalpy Mf is given a negative sign for Reference TC

Apparatus FIGURE 23-5 Schematic diagram of the Fisher Series 300 differential
• See: W. W. Wendlandt, Thermal Methods of Analysis, The furnace, heating programs, and recording thermal analyzer system. (Courtesy of Fisher Scientific Co., Pittsburgh,
2d cd. New York: Wiley, 1974, p. 172. devices used for differential methods are simi- Pennsylvania. )
Differential scanning calorimetry thermograms Card indexes, useful fQr identification of
are similar in appearance to Figure 23-4. organic compounds by differential thermal
methods, are available from several commer-
cial sources. The Sadtler Research Labora-.
Applications torieS collection' includes thermograms for.
Differential thermal methods find widespread 1000 pure organic compounds, 450 commer- .
use in determining the composition of nat- cial compounds, ISO pharmaceutical and ster-
urally occurring and manufactured products. oidalsubstances, and 360 pure inorganic com-
The number of applications is impressive and pounds.
can be appreciated by an examination of a + + Polymen. Differential thermal methods
108119 172
two-volume monograph4 devoted largely to have been widely applied to the study and char-
100 200
uses of the two methods. A few illustrative acterization of polymeric materials. Figure
applications follow. Temperature, ·c 23-8 is an idealized thermogram that illustrates
Inorganic Substances. Differential thermal FIGURE 23-6 Differential thermogram for the various types of transitions that may be
measurements have been widely used for stud- sulfur. [Reprinted with permission from J. encountered during heating of a polymer.
FIGURE 23-8 Schematic differential;ther-
ies involving the thermal behavior of such in- Chiu, Anal. Chem.,35, 933 (1963). Copyright Figure 23-9 is a thermogram of a physical
mogram showing types of changes encourttered
organic compounds as silicates, ferrites, clays, by The American Chemical Society.] mixture of seven commercial polymers. Each
with polymeric materials. [From R. M.
oxides, ceramics, and glasses. Information is peak corresponds to the characteristic melting
Schulken, Jr., R. E. Boy, Jr., and R. H. Cox,
provided about such processes as fusion, de- point of one of the components. Poly tetra-
J. Polymer Sci., Part C, 6, 1717 (1964). Re-
solvation, dehydration, oxidation, reduction, fluoroethylene (PTFE) has an additional low-
printed by permission of John Wiley & Sons,
adsorption, degradation, and solid-state reac- temperature peak which arises from a crystal- Inc.]
tions. One of the most important applications acid at atmospheric pressure and at 200 psi.
is to the generation of phase diagrams and the The first peak corresponds to the melting
study of phase transitions. An example is point and the second to the boiling point of
line transltJon. Clearly, differential thermal
shown in Figure 23-6, which is a differential the acid.
methods can be useful for qualitative analysis
thermogram for pure sulfur. Here, the peak at of polymer mixtures.
113°C corresponds to a solid-phase change
from the rhombic to the monoclinic form,
while the peak at 124°C corresponds to the
melting point of the element. Liquid sulfur is
known to exist in at least three forms, and the
peak at 179°C apparently involves a transition
lutc
among these. The peak at 446°C corresponds 8

to the boiling point of sulfur.


Organic: Compounds. The differential ther-
mal method provides a simple and accurate
way of determining the melting, boiling, and
decomposition points for organic compounds.
Generally, the data appear to be more con-
sistent and reproducible than those obtained
with a hot stage, oil bath, or capillary tube.
Figure 23-7A shows thermograms for benzoic Tem~rature, ·c
FIG U R E 23-7 Differential thermogram for
FIGURE 23-9 Differential thermogram for a mixture of
benzoic acid. Curve A: at atmospheric pres-
seven polymers. PTFE = polytetrafluoroethyJene; HIPPE =
sure; Curve B: at 200 Ibs/in2• [From P. F. Levy,
• Diffl!Tential Thermal AlI<Ilysis, ed. R. C. Maclccnzie. New
high-pressure polyethylene; LPPE = low-pressure polyethyJ-
G. Nieuweboer, and L. C. Semanski, Thermo-
York: Academic Press, 1970, vols. 1 and 2. ene;PP = polypropylene;POM = polyoxymethylene.[FromJ.
chim. Acta, 1, 429 (1970~ With permission.]
Chiu, DuPont Thermogram,2, (3),9 (1965). With permission.]
Principles of Thennometric Titration
quire that AG in the foregoing equation be a
The temperature changes observed during a large negative number; only then will the equi-
The end point in a thermometric titration is
thermometric titration are the consequence of librium between the analyte and reagent lie
obtained from a plot of the temperature of
the heat evolved or absorbed by the reaction sufficiently far to the right. This condition is
the solution being titrated as a function of the
between the analyte and the reagent. The heat necessary to make the potential Changes in the
volume of standard reagent added.6
or enthalpy of a reaction Ml is given by the equivalence-point region large enough for ac-
familiar thermodynamic expression curate location of chemical equivalence.
In contrast, the success of a thermometric
AH=AG+ T AS
titration depends not only upon the magni-
• For more details on thermometric titrations, see: J. where T is the temperature, AG is the free tude of AG but of T AS as well. Thus, if T AS
Jordan, in Treatise on Analytical Chemistry, eds. :1. M. energy change, and AS is the entropy change is a large negative number, successful end
Kolthoff and P. J. Elving. New York: Interscicncc.I968,
Part I, vol. 8, Chapter 91; and H. V. Tyrell and A E.
for the reaction. points may still be realized even though AG is
Beczer, Thermometric Titrimetry. London: Chapman and Most titrimetric end points, such as the po- zero or even positive. A classic example of
Hall,I968. tentiometric ones discussed in Chapter 19, re- this situation is illustrated by a comparison of
the potentiometric and thermometric end
points for boric and hydrochloric acids shown
in Figure 23-10. For boric acid, AG is so small
that neutralization is relatively incomplete at
r,"- End
point chemical equivalence, and no detectable po-
I C tentiometric end point is observed. On the
I D
I other hand, AH for the neutralization of boric
I acid is about -10.2 kcal/mol compared with
I - 13.5 for hydrochloric. As a consequence,
O.OION I sharp thermometric end points are realized
H3B03
I for both acids (Figure 23-lOb).
I
11.0 ;> I The change in temperature AT during a
e r O.OION thermometric titration is given by FIGURE 23-11 Schematic diagram of an
9.0
! T
& O.02"c
I
I
HCI

nMl
apparatus for thermometric titrations. (Adapt-
ed from J. Jordan, in Treatise on Analytical
i
O.OION
Hel
~ 1 I
I
I
AT= ---
k Chemistry, cds. I. M. Kolthoff and P. J. Elving.
New York: Wiley-Interscience, 1968, Part I,
:I:
c. I 1?·50m~1 where n is the number of moles of reactant Volume 8, p. 5201. With permission.)
I and k is the heat capacity of the system. Thus,
5.0
I
the total change in temperature is directly pro-
I
I Titration . portional to the number of moles of analyte.
3.0
I begins

A 50 to 100 times more concentrated than the


8
1.0 Apparatus sample, so that titrant volumes are, at the most,
0 20 40 60 0.0 0.0
a milliliter or two. Under these circumstances,
Vol 0.01 N NaOH. ml Vol 1.0 N NaOH. ml Reproducible thermometric titration curves corrections for dilution or temperature differ-
w ~ require the use of an automatic instrument. ence between the analyte solution and the re-
Figure 23-11 shows a typical apparatus. agent are unnecessary.
FIGURE 23-10 Comparison oqa) potentiometric and (b) thermometric end
Reagent Delivery System. Reagent is added Titration VesseL Titrations must be carried
points. In each case, 50.0 ml of acid were titrated. [Thermometric titration
by means of a screw-driven syringe powered out under conditions that approach adiabatic;
curves from J. Jordan and W. H. Dumbaugh, Jr., Anal. Chern., 31, 212 (1959).
Reprinted with permission. Copyright by the American Chemical Society.] by a motor that is synchronized with the paper that is, without gain or loss of heat from the
drive of the recorder. Generally, the reagent is surroundings. Thus, the titration vessel is
generally a Dewar flask or a beaker surrounded 1. Describe what quantity is measured and how the measurement is per-
by a thick layer of styrofoam insulation. formed for each of the following techniques: (a) thermogravimetric anal-
Efficient stirring must be provided. In addi- Figure. 23-10b reveals that a typical thermo- ysis, (b) differential thermal analysis, (c) differential scanning calorim-
tion, the titration must be completed in a brief metric titration curve is made up of three etry, and (d) thermometric titration.
period (less than 5 min ~ parts. The region from A to B gives the tem- 2. A 562:46 mg sample of a mixture containing the monohydrates of
Temperature MeMUI'emeaIL Thermistors perature of the system before addition of rea- calcium, strontium, and barium oxalate only was subjected to a ther-
are generally employed as temperature sen- gent. Theoretically, the slope here should be mogravimetric analysis. At the plateau near 260°C (Figure 23-3) the
sors for thermometric titrations because of but seldom is zero because of small gains or weight was 509.80 mg. The weight was reduced to 427.93 mg at the
their large temperature coefficients (about 10 losses of heat due to stirring and imperfec~ plateau near 600°C. Find the weight percent of strontium in the orig-
times greater than a thermocouple 1their small tions in the insulation. At point B, reagent inal mixture.
size, and their rapid response to temperature addition is begun at a controlled rate of 3. Why are the two low temperature endotherms in Figure 23-7 coinci-
changes. A thermistor is a sintered metal oxide several microliters per minute. Point C corre- dent whereas the high temperature peaks are displaced from each
semiconductor that has, in contrast to other sponds to the end point in the titration. Or- 'other?
temperature sensors, a negative temperature dinarily, conditions are arranged so that point
4. The reaction of ea2+ with ethylenediaminetetraacetic acid (EDTA) is
coefficient of resistance. D is reached in less than 5 min in order to
iexothermic by 5.7 kcal/moI. The reaction of Mg2+ with EDTA is endo-
As shown in Figure 23-11, changes in re- avoid excess heat leakage to the surroundings.
Beyond C, a straight line is observed. Its slope thermic by 5.5 kcal/mol. The stability constants for the EDT A com-
sistance appear as a voltage difference across
plexes of ea2+ and Mg2+ are 1011.0 and 109.1, respectively. Draw'a
the bridge circuit; the difference is then re- may be negative or positive depending upon
schematic diagram showing the expected thermometric titration curve
corded on a miIlivolt recorder. For a thermistor whether the dilution process is exothermic or
.for the reaction of an equimolar mixture of Ca2+ and Mg2+ titrated
having a resistance of about 20 kCl, the output endothermic.
'with EDTA. Show how you would locate the two end points.
voltage would correspond to about 0.16 mV Table 23-1 shows some typical applica-
per 0.01°C. tions of thermometric titrations. 5. 1The iron-transport protein, transferrin, found in human blood serum,
;has a molecular weight of 81,000 and is capable of binding 0, I, or 2
iron(III) ions. When subjected to differential scanning calorimetry,
transferrin (and many other proteins) exhibits endotherms correspond-
ing to thermal denaturation (unfolding) of part or all of the protein. In
the absence of iron, a solution of transferrin exhibited two endotherms
at 6r and 72°C. When one mole of iron per mole of transferrin was
TABLE 23-1 SOME TYPICAL APPLICATIONS OF THERMOMETRIC added in the form of ferric nitrilotriacetate at pH 7.5, the thermogram
TITRATIONS· of this solution exhibited endotherms at 76° and 88°C, with only a
minor endotherm at 6rc. When a second mole of iron was added, the
Minimum Titratable Precision, All" sample produced just a single endotherm at 89°C. Suggest some pos-
Analyte Titrant Concentration, M Relative % kcalfmol
sible interpretations of these results in terms of structural equivalence
Acids, K. ~ 10- 10
Strong base 0.005 1 -10 to -IS of the sites and the sequence of iron binding to the sites.
Bases, Kb ~ 10- 10 Strong acid 0.005 1
Divalent cations EDTA 0.001 to 0.01 0.1 to 1 -13 to +5
Ag+ HCI 0.1 0.3
Ca2+ (NH4hC204 0.01 1 -6.1
Fe2+ Ce4+ 0.001 1 -24
Cr20~- 0.006 0.5 -27
MnO; 0.003 1 -28
Fe(CN):- Ce4+ 0.001 1 -10
Ti3+ Ce4+ 0.002 2 -30

• Data taken from J. Jordan &Del G. J. EwiD•• in HIJIIIIbooIcof AMlyrktll CMmistry, cd. L. Mcitcs. New York: McGraw-
Hin. 1963, pp. 11-5to 11-7; &Del J. Jordan. in Tretlllse 01\ AMlytictll Chemistry, cds. I. M. Kolthoff and P. J. Elvin8- New
York: Wilcy-Intcrscicncc, 1968, Part I, voL 8, p. 5198.
The physical and chemical properties upon permit the scientist to separate, isolate, and
which analytical methods are based are seldom, identify closely related components of com-
if ever, entirely specific. Instead, these proper- plex mixtures; many of these separations are
ties are far more likely to be shared by numer- impossible by other means.'
ous species; as a consequence, the elimination The term chromatography is difficult to
of interferences in quantitative analyses is more define rigorously, owing to the variety of
often the rule than the exception. systems and techniques to which it has been
The most general method for dealing with applied. All of these methods, however, employ
an interference involves its physical separa- a stationary phase and a mobile phase. Com-
tion from the analyte. Well-known methods ponents of a mixture are carried through the
for accomplishing separations include distilla- stationary phase by the flow ofthe mobile one;
tion, crystallization, solvent extraction, and separations are based on differences in migra-
chemical or electrolytic precipitation. With- tion rates among the sample components.
out question, however, the most widely used
means of elimmating interferences is by chro-
matography, a separation procedure that finds Types of Stationary Phases
application to all branches of science.
In chromatography, the components to be
Chromatography was invented and named
separated must be soluble in the mobile phase.
by the Russian botanist Mikhail Tswett shortly
They must also be capable of interacting with
after the turn of the century. He employed
the stationary phase either by dissolving in it,
the technique to separate various plant pig-
being adsorbed by it, or reacting with it chemi-
ments such as chlorophylls and xanthophylls
cally. As a consequence, during the separations,
by passing a solution of these compounds
the components become distributed between
through a glass column packed with finely
the two phases.
divided calcium carbonate. The separated spe-
In some applications, the stationary phase
cies appeared as colored bands on the column,
is a finely divided solid held in a narrow glass
which accounts for the name he chose for the
or metal tube. The mobile phase, which may
method.
be a liquid or a gas, is then forced through
The applications of chromatography have
the solid under pressure or allowed to perco-
grown explosively in the last four decades,
late through it by gravity. This type of method
owing not only to the development of several
is termed column chromatography. In planar
new types of chromatographic techniques, but
chromatography, the stationary phase may be
also to the growing need by scientists for porous paper or a finely ground solid that has
better methods for separating complex mix- been spread on a glass plate; here. the mobile
tllres. The tremendous impact of these methods phase moves through the solid either by capil-
on science is attested by the 1952 Nobel prize
that was awarded to A. J. P. Martin and R. L.
M. Synge for their discoveries in the field.

1 General references on chromatography include: E.


Heftmann, C/rromDtography, 3d cd. New York: Van
A GENERAL Nostrand Reinhold, 1975;B. L KarJer. L R. Snyder. and
C. Horvath, All llllrotluetioll to Sl!paratlOII Sdl!flU. New
DESCRIPTION
York: Wiley, 1973; J. M. Miller. Sl!paratiOll Mnhods ill
OF CHROMATOGRAPHY CMmical Analysis. New York: Wiley, 1975; R. Stock and
C. B. F. Rice, Chromatographic Mnhods, 3d cd. London:
Chromatography encompasses a diverse and Chapman & Hall, 1974; and Chromatographic tur4 AIIiM
important group of separation methods that Methods, cd. O. MikeS. New York: Wiley. 1979.
Unear Chromatography ent of the solute concentration. Curves of this
Iary action or under the influence of gravity.
type can be described by the relationship
The stationary phase can also be an immobi- All chromatographic separations are based
lized liquid which is immiscible with the mobile upon differences in the extent to which sol-
phase. Several procedures are employed to fix Cs = kC"M
utes are partitioned between the mobile and
the stationary liquid in place. For example, a the stationary phase. The equilibria involved.
finely divided solid, coated with a thin layer of where k and n are constants.
can be described quantitatively by means of a The assumption of a constant K will fre-
liquid, may be held in a glass or metal tube temperature-dependent constant, the partition
through which the mobile phase flows or per- quently be made in the derivation of working
coefficient K: equations for fractionation processes. The plots
colates. Ordinarily the solid plays no direct
part in the separation, functioning only to hold K = Cs (24-1) in Figure 24-1 suggest that this approximation
the stationary liquid phase in place by adsorp- CM will not cause serious errors in limited regions
tion. Alternatively, the inner walls of a capil- where Cs is the analytical concentration of a of the lower concentration ranges, where all
lary tube cim be coated with a thin layer of solute in the stationary phase and CM is its curves are roughly linear. For wide concentra-
liquid; a gaseous mobile phase is then caused concentration in the mobile phase. In the ideal tion ranges, however, the derived expressions
FIGURE 24-1 Typical distri- must be modified to take into account the
to flow thrQugh the tube. Finally, the station- case, the partition ratio is constant over a wide
bution curves. Cs is the concen- variation in K with solute concentration. For-
ary liquid phase can be held in place on the range of solute concentrations; that is, Cs is
tration of solute in the stationary tunately, the concentrations in a chromato-
fibers of paper or on the surface of finely ground directly proportional to CM• More often than
phase and C M its concentration graphic column are ordinarily low. Chroma-
particles held on a glass plate. not, however, nonlinear relationships occur.
in the mobile phase. tography carried out under conditions such
Table 24-1 classifies common chromato- Some typical distribution curves are shown in
graphic m~hods according to the nature of Figure 24-1. that K is constant is called linear chromatog-
the stationary and mobile phases. Most of the The ideal relationship, shown by curve C raphy. We shall treat only this type.
theoretical discussions in this chapter will be in this figure, is often approximated by dis-
concerned with partit~on and gas-liquid chro- tribution equilibria between two immiscible similar to B or D is more likely. For example, if
Linear Elution Chromatography
matography. With suitable modification, these liquids, provided association or dissociation the stationary phase is water and the mobile
concepts can be adapted to the other type as reactions do not occur in one of the solvents. phase is benzene, a curve of type B is observed In elution chromatography, a single portion
well. Where such equilibria do exist, a relationship for a solute consisting of a weak organic acid. of the sample, dissolved in the mobile phase,
Only the undissociated acid is soluble in ben- is introduced at the head of a column (see
zene. In the aqueous solution, however, appre- Figure 24-2), whereupon the components of
ciable amounts of both the undissociated acid the sample distribute themselves between the
and its conjugate base are present; moreover, two phases. Introduction of additional mobile
the ratio between these species is concentra- phase (the eluent) forces the solvent contain-
TABLE 24-1 CLASSIFICATION OF CHROMATOGRAPHIC SEPARATIONS tion-dependent. Thus, at low concentrations, a ing a part of the sample down the column,
Type Type smaller fraction of the total acid is available where further partition between the mobile
Mobile Stationary for distribution between the two solvents, and phase and fresh portions of the stationary
Name Phase Phase Method of Fixing the Stationary Phase the partition ratio is greater. If, on the other phase occurs. Simultaneously, partitioning be-
nand, water represented the mobile phase, a tween the fresh solvent and the stationary
Gas-liquid Gas Liquid Adsorbed on a porous solid held in a tube
or adsorbed on the inner surface of a curve such as D would be expected. Curves phase takes place at the site of the original
capillary tube of type B are also commonly encountered in sample. Continued additions of solvent carry
Gas-solid Gas Solid Held in a tubular column vapor-liquid equilibria, where the vapor is the solute molecules down the column in a con-
Partition! Liquid Liquid Adsorbed on a porous solid held in a mobile phase. tinuous series of transitions between the mobile
; tubular column Curve A is a typical adsorption isotherm and the stationary phases. Because solute
Adsorption Liquid . Solid Held in a tubular column
Paper Liquid Liquid Held in the pores of a thick paper which relates the amount of solute adsorbed movement can only occur in the mobile phase,
Thm layer Liquid Liquid or solid Fin~ly ~ivided solid held on a glass plate; on the surface of a solid to the solute concen- however, the average rate at which a solute
liql;lid may be adsorbed on particles tration in the solution that contacts the solid. migrates depends upon the fraction of time it
Gel Liquid Li~uid Held m the mterstices of a polymeric solid At high solute concentrations, all adsorption spends in that phase. This fraction is small for
Ion exchange Liquid Said Finely divided ion-exchange resin held in sites on the solid surface are occupied; the solutes with partition ratios that favor reten-
a tubular column
extent of adsorption then becomes independ- tion in the stationary phase and large where
sponds to solute con~ntration is placed at
the end of the column and its signal is plotted
as a function of time (or of volume of the
added mobile phase~ a series of symmetric
peaks is obtajned, as shown in the lower part
of Figure 24-2. Such a plot, called a chromato-
gram, is useful for both qualitative and quan-
titative analysis. The positions of peaks may
serve to Identify the components of the sample;
the areas under the peaks can be related to Distance migrated. L
concentration. FIGURE 24-3 Concentration profiles for solutes A and B
at different points in their migration down a column.
Theories of 'Elution Chromatography
Figure 24-3 shows concentration profiles for
solutes A and B on a chromatographic column
at an early and a late state of elution. The
partition coefficient for A is the larger of the terms employed in the kinetic theory. The Note that H decreases as the efficiency of a
two; thus, A lags behind during the migration plate theory of chromatography, which was column becomes greater. That is, as H be-
• process. It is apparent that movement down originally developed by Martin and Synge,2 comes smaller, the number of equilibrations
the column increases the distance between the envisages a chromatographic column as being that occur in a given length of column be-
two peaks. At the same time, however, broad- composed of a series of discrete but contig- comes larger. It is important to note that H
ening of both bands takes place, which lowers uous, narrow, horizontal layers called theo- and N are retained as efficiency parameters in
the efficiency of the column as a separating retical plates. At each plate, equilibration of the the rate theory and that Equation 24-2
device. Zone broadening is unavoidable; for- solute between the mobile and the stationary
FIGURE 24-2 Schematic diagram of continues to apply. It should be appreciated,
tunately, however, it occurs more slowly than phase is assumed to take place. Movement of
an elution chromatographic separation however, that a plate as a physical entity does
zone separation. Thus, as shown in Figure 24-3, the solute and solvent is then viewed as a series
of a two-component mixture. not exist in a column. Thus, the plate and the
a clean separation of species is possible pro-
of stepwise transfers from one plate to the next. plate height should be viewed as criteria for
vided the column is sufficiently long. The efficiency of a chromatographic column column efficiency.
It is important to .appreciate that a useful as a separation device improves as the number
theory of chromatography must be able to of equilibrations increases-that is, as the
account not only for the rate at which solutes number of theoretical plates increases. Thus,
retention in the mobile phase is more likely. migrate but also the rate at which the zones the number of theoretical plates N is used as a THE RATE THEORY
Ideally, the resulting differences in rates cause broaden during migration, since the cleanness measure of column efficiency. A second term, OF CHROMATOGRAPHY
the components in a mixture to separate into of separation is equally dependent upon the the height equivalent of a theoretical plate H,
bands located along the length of the column two phenomena. The original theory of chro- also serves this purpose. The relationship be- The rate theory of chromatography success-
(see Figures 24-2 and 24-3). Isolation can then matography, called the plate theory, was able to tween these two parameters is fully describes the effects of variables which
be accomplished by passing a sufficient quan- describe migration rates in quantitative terms. affect the width of an elution band as well as
tity of mobile phase through the column to Unfortunately, however, its utility was limited its time of appearance at the end of a column.3
cause these various bands to pass out the end, because it failed to describe the effects of N=~
H
where they can be collected. Alternatively, the the numerous variables responsible for zone
column packing can be removed and divided broadening. As a consequence, the plate theory J For a detailed presentation of the rate theory, ICe: J. C.
into portions containing the various com- has now been·supplanted by the kinetic or rate Giddings, J. Chern. Educ. "' 704 (1967); J. C. Giddinp,
DY/lQnliC5 of Chromatography. New York: Marcel
ponents of the mixture. theory, which is capable of accounting for these
Dekker, 1965, Part I; and J. C. Giddinp, in Chromatog-
The process whereby the solute is washed variables. raphy, 3d ed., ed. E. Heflmann. New York: Van Nostrand
2 A. J. P. Martin and R. L. M. Synge, Biochem. J. 35,
through the column by addition of fresh sol- It is useful to consider the plate theory m8 (1941). Reinhold, 1975, Chapter 3.
vent is called elution. If a detector that re- briefly in order to indicate the genesis of two
have been incorporated in the stationary phase
for a greater-than-average time. The con-
Examination of a typical chromatogram (Fig- sequence of these random individual processes
ures 24-2 and 24-3) reveals a similarity to is a symmetric spread of velocities around the
normal error or Gaussian curves (see Figure mean value, which represents the behavior of
A-I, Appendix I) that are obtained when rep- the average and most common particle.
licate values of a measurement are plotted as The breadth increases as the zone moves
a function of the frequency of their occurrence. down the column because more time is allowed
Such plots can be rationalized by assuming that for migration to occur. Thus, the zone breadth
the uncertainty associated with any single is directly related to residence time in the
measurement is the summation of a much column and inversely related to the velocity
larger number of small, individually undetect- at which the mobile phase flows.
able, and random uncertainties, each of which Standard Deviation IS a Measure of Zone
may be positive or negative in sign. The most Breadth. The breadth of a Gaussian curve is FIGURE 24-~ Determination of the standard deviation t from a
common occurrence is for these uncertainties conveniently related to a single parameter chro~atog~aphlc peak. Here, W = 4t; tR is the retention time for a solute
to cancel one another, thus leading to the mean called the standard deviation (J (Figure A-I, !hat IS retamed by the column packing, and tM is the time for one that
value. With less likelihood, the summation may Appendix I); approximately 96% of the area IS not. Thus, tM is equal approximately to the time required for a molecule
lead to· results that are greater or smaller than under such a curve lies within plus or minus of the mobile phase to pass through the column.
the mean. The consequence is a symmetric dis- two standard deviations (±2(J) of its maxi-
tribution of data around the mean value. In a mum. Thus, a standard deviation derived
similar way, the typical Gaussian shape of a from a chromatogram serves as a convenient conveniently expressed in terms of the square abscissa. Thus, the standard deviation derived
chromatogram can be attributed to the addi- quantitative measure of zone broadening. of .the standard deviation (the variance) per from such a curve will be in units of time
tive combination of the random motions of the It is important to understand that the stand- umt length of column. By definition, rather than length. From Figure 24-4, it js seen
myriad solute particles in the chromatographic ard deviation for a chromatographic peak
that tR, the retention time, is the time required
band or zone. can be expressed in units of time or, alterna-
after sample injection for the solute peak to
Let us first consider the behavior of an in- tively, in distance along the length of the
appear at the end of the chromatographic
dividual solute particle, which, during migra- column. Thus, a standard deviation derived
where H, the plate height in centimeters, is column. The average rate at which the solute
tion, undergoes many thousands of transfers from the zone profiles shown in Figure 24-3
the measure of efficiency and L is the length pa.rticles travel is then L/tR• Thus, the relation-
between the stationary and the mobile phase. would be in terms of length (usually in em).
of column, also in centimeters; note that (J ship between the standard deviation in units of
The time it spends in either phase after a trans- More commonly, the abscissa of a chromato-
must carry units of length (em) to be consis- I~gth and units of time is given by the quotient
fer is highly irregular and depends upon its gram is in units of seconds or minutes (see
tent with Hand L. Substitution of Equation of the standard deviation in centimeters and
accidentally gaining sufficient thermal energy Figure 24-4), and a standard deviation derived
24-2 and rearrangement provides an alterna- the rate of migration in centimeters per second,
from its environment to accomplish a reverse from such a curve would carry time units. It
tive way of describing the separation charac- or
transfer. Thus, in some instances, the residence is useful to employ a symbol to indicate which
time in a given phase may be transitory; in units are being employed; thus, we shall use (J teristics of a column. That is, (J
t=--
others, the period may be relatively long. Recall for a standard deviation in units of length and L/tR
that the particle can move only during residence t when the units are time.
in the mobile phase; as a result, its migration Figure 24-4 illustrates a simple means for Here t is the standard deviation in units of
down the column is also highly irregular. Be- approximating t or (J from an experimentally time and tR is the time required for 1(J 'of the
cause of variability in the residence time, derived chromatogram. Tangents to the two where N is the number of plates contained in zone to emerge from the column. As noted
the average rate at which individual particles sides of the Gaussian curve are extended to ~ column oflength L.Clearly, column efficiency earlier, however, for a Gaussian curve s~ch as
move relative to the mobile phase varies con- form a triangle with the abscissa. The inter- mcreases as the number of plates increases and that in Figure 24-4, W = 4t. Substitution of
siderably. Certain individuals travel rapidly by cepts then occur at approximately ±2t from the plate height decreases. this relationship into (24-5) yields, upon re-
virtue of their accidental inclusion in the arrangement,
the maximum; that is, W = 4t. Experimental Evaluation of N and H. Most
mobile phase for a majority of the time. Others, Cakulation of H and N from Zone Breadth. experimental chromatograms, such as that in LW
in contrast, may lag because they happen to (J=-
The broadening of a chromatographic zone is Figure 24-4, are obtained with time as the 4tR
Substitution into Equation 24-3 gives Flow possessing a limited range of sizes. Particular the stationary phase encounter fresh mobile
direction
. care is needed to avoid open channels. With phases. Again, the rate of transfer of solute
LW2
H = 16t~ (24-6) ! carefully packed columns, broadening due to
eddy diffusion is believed to be minimal.
molecules is not instantaneous; thus, the tail
of the zone is more drawn out than it would
Longitudinal Diffusion. Longitudinal diffu- be if time existed for equilibration. The net
To obtain N, we substitute into Equation 24-2
sion results from the tendency of molecules to effect is a broadening at both ends of the

~r
and rearrange, giving
migrate from the concentrated center part of solute band.
N = 16( (24-7)
a band toward more dilute regions. on either
side. This type of diffusion, which can occur
The effects of nonequilibrium mass transfer
become smaller as the flow rate is decreased
in both the mobile and the stationary phase, (Equation 24-8) because more time is avail-
Thus, N can be calculated from the two time
causes further band broadening. Longitudinal able for equilibrium to be approached. Fur-
measurements tR and W; to obtain H, the
diffusion is most important where tbe mobile thermore, a closer approach to true equilib-
length of the column must also be known. B------- --- phase is a gas, because diffusion rates in the rium is to be expected if the channels through
FIGURE 24-5 Typical gas phase are several orders of magnitude which the mobile phase flows are narrow so
Sources of Zone Broadening
pathways of two solute mol- greater than those in liquids. The amount of that solute molecules do not have far to diffuse
Chromatographic peaks are generally broad- ecules during elution. Note diffusion increases with time; thus, tthe extent in order to reach the stationary phase. For the
ened by three kinetically controlled processes, that distance traveled by of broadening increases as the flow rate de- same reason, the layers of immobilized liquid
eddy diffusion, longitudinal diffusion, and non- molecule 2 is greater than creases. on a stationary phase should be as thin as
equilibrium mass transfer. The magnitudes of that traveled by molecule 1. The second term (the longitudinal diffu- possible.
these effects are determined by such control- Thus, molecule 2 would ar- sipn term) in Equation 24-8 is seen to be The last term in Equation 24-8, which is of
lable variables as flow rate, particle size of rive at B later than mole- inversely proportional to flow rate.' The con- considerable importance at the high flow rates
packing, diffusion rates, and thickness of the cule 1. stant B is related to the diffusion fOefficient of a gaseous mobile phase, describes the effect
stationary phase. A number of equations have DM of the solute in the mobile phllse by the of nonequilibriuln on band broadening. The
been developed that relate the efficiency of the lengths of these pathways differ; thus, the equation value for C is made up of two terms, the first
chromatographic columns to the extent to residence times in the column for molecules having to do with the rate of mass transfer in
B= 2y,DM
which these three processes occur.4 The earliest of the same species are also variable. Solute the stationary phase and the second with the
and simplest of these, which is known as the molecules then reach the end of the column where y, is the obstruction factor, a measure rate in the mobile phase. Thus,
van Deemter equation, was derived for gas- over a time interval that tends to broaden of the hindrance to free molecular diffusion
liquid chromatography; it provides an ap- the elution band. caused by the presence of the particles of the
proximate relationship between the flow rate u The quantity A in Equation 24-8 describes stationary phase; typical values for y, range
and the plate height. the effect of eddy diffusion and can be related from 0.5 to 0.9, depending upon the type of
to particle size, geometry, and tightness of packing. Broadening due to longitudinal dif- where d1 is the thickness of the film or sta-
B
H =A + - + Cu (24-8) packing of the stationary phase. As a first fusion can be reduced by decreasing the tem- tionary particle having a diameter of dR' The
u quantity R is the retention ratio and is equal
approximation, A is independent of flow rate, perature (thus reducing DM) and increasing
Here, the quantity A is associated with eddy and its value is given by the velocity of flow. to t M It R (see Figure 24-4). The terms Ds and
diffusion, B with longitudinal diffusion, and C . Nonequilibrium Mass Transfer. Chromato- DM are diffusion coefficients for the solutes in
A = 2NlR the stationary and mobile phases, respectively.
with nonequilibrium mass transfer. graphic bands are also broadened because the
Eddy Diffusion. Zone broadening from eddy where dR is the average particle diameter and flow of the mobile phase is ordinarily so rapid The quantities q and (JJ are constants whose
diffusion arises from the multitude of pathways l is the packing factor. Values of l are deter- that true equilibrium between phases cannot be values are determined by the nature and con-
by which a molecule can find its way through mined by the range of particle sizes in the realized. For example, at the front of the zone, figuration of the packing. The latter has a value
a packed column. As shown in Figure 24-5, stationary phase and how they are packed. A where the mobile phase encounters[ fresh sta- of about 1.3. For a uniform film of liquid on
typical value for 20- to 4O-mesh particles is tionary phase, equilibrium is not instantly a solid support, a typical value of q is about 0.7;
about 1; for 200- to 400-mesh solids, l has a achieved, and solute is therefore carried some- for a liquid immobilized by paper, q is about
value of about 8. Band broadening due to what farther down the column than would be 0.5. The most important of the variables affect-
• For a summary of these various equations, see: J. M.
Miller, Separation Methods in Clrnnlall Analysis. New eddy diffusion is minimized by careful packing expected under true equilibrium conditions. ing Cis d I' A marked improvement in column
York: Wiley, t97S, pp. 120-126. of a column with small spherical particles Similarly, at the end of the zone, solutes in efficiency is obtained with very thin layers of
the number of plates in a column and the where CM and Cs are the concentrations of
time required for any given separation. ~here .Wx and Wy are the widths of the peaks
the solute in the mobile and stationary phases (m UDits of time) at their bases and AZ is the
respectively; similarly, VM and Ys are the total difference in time of their arrival at the detec-
:z: , Relte. of Solute Migration volumes of the two phases contained in the tor (see Figure 24-7), If W is taken as the
\
f " "",'"
;~
column. average width of the two peaks, Equation 24-
For the chromatogram shown in Figure 24-4, Substitution of Equation 24-1 into this ex- 15 reduces to
~ \
~...
.,
.,
/' zero on the time axis corresponds to the mo- pression gives
ir: \
__ ~c::. , "."
,,"
A
ment the sample was injected onto the column
and elution was started. The peak arising at
tM is for a species (often air) that is not re-
_
v-u
-
( --~~ 1
1 + KYs/YM
)
AZ
R ,=-W

",; .•••.•..•...••. Blu


------ tained by the column packing; its rate of mo-
tion will be the same as the average rate for
For peaks that are close together (as they will
t>c: when there is interest in R,), Wx and Wy
wdl often be sufficiently alike that only one
FIGURE 24-6 Effect of variables in the molecules of the mobile phase. The reten-
needs to be measured.
Equation 24-8 on plate height. tion time tR for the solute responsible for the
Figu~e 24-7 indica.tes that a resolution (R,)
second peak is the time required for that peak
where k', the capacity factor, is related to the of 1.5 gives an essentially complete separation
to reach the detector at the end of a column.
partition coefficient for the solute. That is, of X and Y, whereas a resolution of 0.75 does
The average rate of migration ii of the
not. At a resolution of 1.0, zone X contains
solute is given by
k' = KYs about 4% of Y, and conversely; at a resolu-
liquid phase. Equilibrium is also more closely _ L tion of 1.5, the overlap is about 0.3%. For the
VM
approached at high temperatures and with low v=- same kind of packing, the resolution can be
tR
solvent viscosities. Substitution of Equations 24-9 and 24-10 into improved by lengthening the column and thus
Net Effect of the Zone Broadening Proces- Similarly, the average rate of movement u of 24-11 yields, upon rearrangement, increasing the number of theoretical plates.
ses. Figure 24-6 shows the contribution of molecules of the mobile phase will be Relationship Between R, and Column Prop-
each term in the van Deemter equation as a erties. It is useful to be able to relate R to
function of mobile-phase velocity (broken
tR = tM(l + k') = tM( 1 + ~~) (24-13)
certain properties of a chromatographic ~I-
lines) as well as their net effect (solid line) on umn. The following derivation shows how this
H. Clearly, the optimum efficiency is realized Note that the retention time for a solute in-
creases with 'the partition coefficient of the is possible.
at a flow rate corresponding to the minimum The rate of migration of the solute can also
in the solid curve. From experimental curves be expressed as a fraction of the velocity of s~lute and the volume of the stationary phase
of this type, A, B, and C are readily evaluated the mobile. phase. That is, with respect to the mobile. EXAMPLE
for any column. Such data provide hints for Rearrangement of Equation 24-13 provides Derive a relationship between R, and (tR)x,
improving the performance of a given type of _ (fraction of time the solute ) an expression that permits calculation of the and (tR)y where the latter two terms are the
v=u capacity factor for a solute from the experi- measured retention times shown in Figure 24-7.
packing. spends in the mobile phase
The van Deemter equation provides only a mental parameters, tR and tM• Thus, As we have mentioned, usually
first approximation of plate height, and several This fraction, however, equals the average
W=Wy~WX
modifications have been developed that give a number of moles of the solute in the mobile
more precise description of the variables affect- phase at any instant compared with the total Thus, we may write Equation 24-15 in the
ing column efficiency (see footnote 3, p. 671). number of moles in the column. That is, form
Column Resolution
_ no. moles solute in mobile phase R, = (tR)Y- (tR>x
v = u x ---------~-- The ability of a column to resolve two solutes Wy
total number moles solute
~ of prime interest in chromatography. A
Equation 24-7 permits expression of Wy in
quantitative term for expressing the resolving
terms of (tR)y and N. Thus,
The discussion thus far has focused upon the power of a column is the column resolution
efficiency of a chromatographic column-that
R" which is defined as R = (tR)Y- (tR>x x IN
is, how many plates it contains. We must now , (tR)y 4
R _ 2 AZ _ 2[(tR)y - (tR>xl (24-15)
give consideration to the relationship between '-Wx+Wy - Wx+Wy Substitution of Equation 24-13 permits ex-
seen by substitution of Equation 24-14 into term, which is the quotient containing ky; this
24-16, which yields term depends on the properties of both the
solute and the column.
IX = (tllh - tll . (24-19) CollDllDFJIiciency. It is evident from Equa-
(tll)x - tll
tion 24-17 that resolution is proportional to
Furthermore, substitution of Equation 24-12 the square root of the number of plates in the
into Equation. 24-16 reveals that column. Increasing the number of plates to
achieve a separation can be expensive in terms
Ky of time, however, unless the increase is accom-
IX =- (24-20)
Kx plished by a reduction in H (Equation 24-21)
Resolution -and Sepantion Time. One im- and not by lengthening the column. The earlier
portant performance characteristic is absent section on rate theory described the variables
from Equations 24-17 and 24-18, namely, the that can be controlled to maximize efficiency
time required to complete the separation. In or minimize H.
order to include this important variable, Equa- Variation in the Capacity Factor. The ca-
tion 24-9 is employed to obtain the velocity pacity factor ky for a solute can frequently be
of the slower-moving solute Y. That is, varied over a considerable range by altering
the composition of the mobile or the station-
_ L ary phase. The former is usually easier and is
Vy = (tllh the one that is often changed to optimize a
given separation. From Equation 24-12, it is
Combination lof this relationship with Equa-
evident that ky can also be increased or de-
I
tions 24-11 al)d 24-2 yie~ds
t.---Wy--.l
creased by a change in the ratio of the volume
I I (tllh = NH(1 + ky) of the stationary phase to the mobile. An in-
I
u crease in ky accompanies a reduction in par-
ticle size of the stationary support because the
where (tllh is the time required to elute Y consequent increase in surface area results in
FIGURE 24-7 Separations at three resolutions. Here, R. = when the velocity of the mobile phase is u. an increase in Vs.
2.1Z/(Wx + R\-). Substitution of Equation 24-18 into the fore- Increases in ky enhance resolution but at
going expression gives the expense of elution time. In order to under-
stand the two effects, it is convenient to write
Often, it is desirable to calculate the number
(tllh = 16R;H(_IX_)2 (I + ky)J (24-21)
Equations 24-17 and 24-21 in the form
pression of R. in terms of capacity factors for U IX - 1 (ky)2
X and Y. Thus, of plates required to achieve a desired resolu- ky
tion. Such an expression is obtained by rear-
R _ (ky - k,,) -IN rangement of Equation 24-17, which gives
Optimization of Column Perfonnance R·=Ql+kY
'-(I+ky) x 4 Equations 24-17 and 24-21 are of considerable
Let us now define ky /k" as the selectivity importance in column chromatography be-
t_Q,(I+ky)J
cause they serVeas guides to the choice of con- (tillY - (k )2
factor IX.That is, y
ditions that will lead to a desired degree of
ky resolution with a minimum expenditure of where Q and Q' contain the rest of the terms
IX=-

k" time. An examination of these e.quations re- in the equations. Figure 24-8 is a plot of R./Q
Employing this relationship to eliminate kit
Properties or the Selectivity FadQl'IX. Before veals that each is made up of three parts. The and (tllh /Q' as a function of ky, assuming
from the preceding equation gives, with re- examining the application of Equations 24-17 first describes the efficiency of the column in that Q and Q' remain approximately constant.
arrangement, and 24-18, it is worthwhile noting the proper- terms of .IN or H. The second, which is the It is clear that values of ky greater than about
ties of the selectivity factor IX.First, it is readily quotient containing IX,is a selectivity term and 10 are to be avoided because they provide
R =

.jN(~) (~)+ y
4 IX 1 k
(24-17) derived from a chromatogram such as that
shown in Figure 24-7. This property can be
depends solely on the properties of the two
solutes. The third component is the capacity
little increase in resolution but markedly in-
crease the time required for separations. The
,
SUMMARY used tool for identifying the components of
(c) the plate height. OF IMPORTANT mixtures containing a limited number of pos-
(d) the length of column required to achieve RELATIONSHIPS sible species whose identities are known. For
a resolution of 1.5. FOR CHROMATOGRAPHY example, the presence or absence of 30 or
(e) the time required to elute substance B more amino acids in a protein hydrolysate
on the longer column. The number of quantities, terms, and relation- can be ascertained with a relatively high degree
Employing Equation 24-15, we find ships employed in chromatography is large and of certainty from their positions after develop-
(a) R. = 2(17.63 - 16.40)/(1.11 + 1.21) often confusing. Table 24-2 serves to sum- ment on a thin layer chromatographic plate.
I = 1.06 marize the most important definitions and Even here, however, confirmation of identity
(b) Equation 24-7 permits computation equations used in this text. would require spectral or chemical investiga-
of N. Thus, tion of the isolated components. .
It is important to add that positive spec-
Capacity factor, kx N = 16(16.40)2 and . N = 16(17.63)2 troscopic identification would ordinarily be
1.11 . 1.21 QUALITATIVE impossible on as complex a sample as the fore-
FIGURE 24-8 Effect of capacity factor k~
on resolution R. and elution time (tR)y. It is AND QUANTITATIVE going without a preliminary chromatographic
= 3493 = 3397
assumed that Q and Q' remain constant with
;
ANALYSIS BY CHROMATOGRAPHY separation. Thus, chromatography is often a
variation in k~. Na• = (3493 + 3397)/2 = 3445 = 3.44 x W vital precursor to a qualitative spectroscopic
Chromatography has grown to be the premiere analyses.
(c) H = L/N = 30.0/3445 = 8.708 x 10-3
method for separating closely related chemical
em = 8.71 x 10-3 em species. In addition, it can be employed for
Quantitative Analysis
minimum in the elution time curve occurs at qualitative identification and quantitative de-
(d) k' and ex do not change with increasing
values of k~ between 2 and 3. Often, then, the termination of separated species. This section Chromatography owes its precipitous growth
Nand L. Thus, substituting N I and N z into
optimal value of k~ lies in the range of 2 to 5. considers some of the general characteristics during the past two decades in part to its
Equation 24-18 and dividing one of the result-
Variation in the Selectivity Factor. It is evi- of chromatography as a tool for completion speed, simplicity, relatively low cost, and wide
ing equations by the other yields .
dent from Equation 24-17 that resolution is of an analysis. applicability as a separating tool. It is doubt-
improved as the selectivity factor becomes larg- N1 (R.)f ful, however, if its use would have become as
er. When this parameter is greater than 2, a Nz = (R.H widespread had it not been for the fact that it
Qualitative Analysis
clean separation is generally possible in a mini- can also provide useful quantitative informa-
where subscripts 1 and 2 refer to the original
mal time. On the other hand, as ex approaches A chromatogram provides only a single piece tion about the separated species. It is impor-
and the longer columns, respectively. Substi-
I, resolution is lost unless a very long column of qualitative information about each species tant, therefore, to discuss some of the general
tuting the appropriate values for N 10 (R.)10
is employed. As shown in Equation 24-21, how- in a sample, namely, its retention time or its aspects of quantitative chromatography.
and (R.h gives Quantitative chromatography is based upon
ever, ex values that approach 1 increase enor- position on the stationary phase after a cer-
mously the time required for separation. For 3445 (1.06)Z tain elution period. Additional data can, of a comparison of either the height or the area
example, an 84-fold increase in time is required Nz = (1.5)2 course, be derived from chromatograms in- of the chromatographic peak of the analyte
for the same resolution when ex is equal to 1.01 volving different mobile and stationary phases with that of one or more standards. If con-
as compared with 1.1. N 2 = 3445 x 2.25/1.124 . and various elution temperatures. Still, the ditions are properly controlled, both of these
= 6.90 x 103 number of data points for a species obtainable parameters vary linearly with concentration.
EXAMPLE by chromatography is small compared with the Analyses Based on Peak Height. The height
Substances A and B were found to have But L = 6.90 X 103 x 8.71 X 10-3 number provided by a single IR, NMR, or mass of a chromatographic peak is obtained by con-
retention times of 16.40 and 17.63 min, re- = 60.1 em i spectrum. Furthermore, spectral abscissa data necting the base lines on either side of the peak
spectively, on a 30.O-cmcolumn. An unretained (l, lJ, or m/e) can be determined with a much by a straight line and measuring the perpen-
species passed through the column in 1.30 min. (e) From Equation 24-2i, we may write
higher accuracy than can their chromato- dicular distance from this line to the peak. This
The peak widths were 1.11 and 121 mm. Cal- (tllh (R.>t 17.63 (1.06)2 graphic counterpart (tll). measurement can ordinarily be made with
culate (tll)z = (R.~ = (tllh = (1.5)2 The foregoing should not be interpreted to reasonably high precision and yields accurate
(a) the column resolution. mean that chromatography lacks important results, provided variations in column condi-
(b) the average number of plates in the qualitative applications. Indeed, it is a widely tions do not alter the peak widths during the
column.
period required to obtain chromatograms for approximate the composition of the unknown.
sample and standards. The variables that must Chromatograms for the standards are then ob-
be controlled closely are column temperature, tained and peak heights or areas are plotted
the eluent flow. rate, and the rate of sample as a function of concentration. A plot of the.
TABLE 24-2 SUMMARY OF THE MOST IMPORTANT
injection. In ad4ition, care must" be taken to data should yield a straight line passing
CHROMATOGRAPHIC QUANTITIES AND RELATIONSHIPS
avoid overloadmg the column. The effect of through the origin; analyses are based upon
EXPERIMENTAL .QUANTITIES sample injection rate is particularly critical for this plot. Frequent restandardization is neces-
Name Symbol Determined from the early peaks of a chromatogram. Here, rela- sary for highest accuracy.
tive errorsof5 to 10% due to this cause are The most important source of error in anal-
Retention time, mobile phase tAl Chromatogram (see Figure 24-7) not unusual with syringe injection. yses by the method just described is usually
Retention times, species X and Y (tJl)x, (tJl)v Chromatogram (see Figure 24-7) Analyses Rued on Peak Areas. Peak areas the uncertainty in the volume of sample; occa-
Peak widths, species X and Y Wx, Wy Chromatogram (see Figure 24-7) are independent of broadening effects due to sionally the rate of injection is also a factor.
Length of column packing L Direct measurement the variables mentioned in the previous para- Ordinarily. samples are small (-1 jlliter), and
Flow rate F Direct measurement graph. From this standpoint, therefore, areas the uncertainties associated with injection of a
Volume of stationary phase Ys Packing preparation data are a more satisfactory analytical parameter reproducible volume of this size with a micro-
Concentration of solute in CAI,CS Analysis and preparation data than peak heights. On the other hand, peak syringe may amount to several percent relative.
mobile and stationary phases heights are more easily melisured and for nar· The situation is exacerbated in gas-liquid
row peaks, more accurately determined. chromatography, where the sample must be
Many modern chromatographic instru- injected into a heated sample port; here, evap-
ments are equipped with ball and disc or elec- oration from the needle tip may lead to large
DERIVED QUANTITIES tronic integrators (p. 66), which permit precise variations'in the volume injected.
Calcul~tion of estimation of peak areas. When such equip- Errors in sample volume can be reduced to
Deriyed Quantities ment is not available, a manual estimate must perhaps Ilo 2 % relative by means of a rotary
be made. A simple method, which works well gas sampie valve such as that shown in
Mobile phase velocity u= L/tAi for symmetric peaks of reasonable widths, is Figure 24-9. Here, the sample loop ACB in
Volume of mobile phase VAl = tAiF to multiply the height of the peak by its width (a) is filled with sample; rotation of the valve
at one-half the peak height. Other methods by 45 deg then introduces a reprodu~ible
k'= KYs involve the use of a planimeter or cutting out volume of sample (the volume originally con-
VAl the peak and determining its weight relative tained in ACB) into the mobile-phase stream.
to the weight of a known area of recorder The Internal Standard Method. The highest
K=k'VAI K= Cs precision for quantitative chromatography is
paper. McNair and Bonelli measured the pre-
Ys CM
cision of these various techniques on chro- obtained by use of an internal standard be-
(tR)v - tAl ky Ky matograms for ten replicate samples.5 They cause the uncertainties introduced by sample
IX=---
(tR)x - tAl IX=-=-
kx Kx reported the following standard deviations: injection are avoided. In this procedure, a
electronic .integration, 0.44 %; ball and disc carefully measured quantity of an internal
R _ 2[(tR)v - (tR)x]
,- Wx+Wy R
,
=IN(~)(~)
4 l+ky
IX
.integration, 1.3%; weight of paper, 1.7%;
height times width at one-half height 2.6%;
standard substance is introduced irito each
standard and sample, and the ratio of analyte
to internal standard peak areas (or heights)
N = 16(tW)2 R N = ~H = 16R2(_1X
'IX-l
)2(1 +~ ky)2 and planimeter, 4.1 %.
Calibration with Standards. The most serves as the analytical parameter. For this
straightforward method for quantitative method t9 be successfu~ it is necessary that
_ 16R:H( IX )2 (1 (kyf
(tR)v - -u- IX- 1
+ ky)3 chromatographic analyses involves the prep- the interool standard peak be well separated
aration of a series of standard solutions that from the peaks of all other components of the
sample (R, > 1.25); the standard peak should,
on the other hand, appear close to the analyte
• H. M. McNair and E. J. Bonelli, Basic Gas ChrOfflG-
Walnut Creek, CA: Varian Aerograph, 1968,
tograph}'.
peak. With a suitable internal standard, preci-
p.158. sions of 0.5 to 1% relative are reported.
Eluent end
semple
to
column

FIGURE 24-9 A rotary sample valve. (a) Valve position for filling
sample loop ACB and (b) for introduction of sample into column.

The Area Normalization Metbod. Another Each entry in column 4 is the product of the
approach that avoids the uncertainties asso- data in columns 2 and 3. To normalize,
ciated with sample injection is the area normal- % n-butyl alcohol = 1.652 x 100/8.943 = 18.5
ization method. Complete elution of all com- % i-butyl alcohol = 4.033 x 100/8.943 - 45.1
ponents of the sample is required, a restriction % s-butyl alcohol = 2.128 x 100/8.943 - 23.8
that hasi limited the procedure to gas-liquid % t-butyl alcohol = 1.130 x 100/8.943 = 12.6
chromatography. In the normalization me-
100.0
thod, the areas of all eluted peaks are com-
puted; after correcting these areas for differ-
ences in the detector response to different com- Computerized Chromatography
pound types, the concentration of the analyte
Most modem instruments for chromatography
is found from the ratio of its area to the total
are equipped with microprocessors fot control-

tt
area of all peaks. The following example
ling such operating parameters as column
illustrates the procedure.
temperature, eluent flow rate, sample injection
EXAMPLE time, and sample temperature. As a con-
The following area data were obtained from sequence, chromatograms can be obtained
a chromatogram of a mixture of butyl alcohols with little or no human control. The most TTY

(the detector sensitivity corrections were ob- sophisticated instruments are completely auto-
tained in separate experiments with known matic. Here, several dozen samples are con-
tained in a sample turntable. After each ana-
® Computing integrator
amounts of pure alcohols).
lysis, the column is returned automatically to FIGURE 24-10 A computing integrator for chromatography. (From 1. M. Gill, Chromato-
Peak Detector Corrected its initial condition, the sample turntable is graphy. Greens Farms, CT: International Scientific Communications, Inc., 1974, Series I,
Area, Response Areas, rotated, a new sample is withdrawn and in- vol. I, p. 188. With permission.)
cm2 Factor cm2 jected into the column, and the chromatogram
0.6l>3 1.652 is obtained and stored in a computer memory.
n-butyl 2.74
0.530 4.033 The data are presented as chromatogram and
i-butyl 7.61
0.667 2.128 in tabular .form with retention times and rela-
s-butyl 3.19
1.66 0.681 1.130 tive peak areas printed out. Some instruments
t-butyl
also print out the names of possible com-
8.943 pounds giving each peak.
Several instrument manufacturers olTer AUTOLA8 A chromatogram of a mixture of species A, B, C, and D provided the
SYSTEM· 4
small, modular computers that are specifically following data:
4 CH
designed for processing the output signals from 2185 ID Retention Time, Width of Peak Base (W~
chromatographic instruments. Figure 24-10 is 21 xo min
min
a schematic diagram showing the components
PARAMT~5
of one of these units, which is capable of Nonretained 4.2
------
handling data from four chromatographic col- 23 T1 A 6.4 0.45
umns. The analog signal, after digitilization, 2567 T2 B 14.4 1.07
passes into an ROM (read only memory) 3000 T3 C 15.4 1.16
10 PW 1.45
operating program, which constantly monitors D 20.7
150 55
and operates on the data. Here, peaks are de- 1000 MA
tected and their areas and retention times
determined. The arithmetic operations are per- AREA Calculate
formed by the CPU. These processed data are ------ (a) the number of plates from each peak.
26 .37 %/ (b) the mean and the standard deviation for N.
then transferred to the RAM (random access 142 64.72 %/
memory) module for storage until the end of (c) the plate height for the column.
563 1.60 %/
the analysis. The RMM (read mostly memory) 1269 21.12 %J. 2. From the data in Problem 1, calculate for A, B, C, and D
module contains programs for optimizing the 2122 7.39 %/ (a) the capacity factor.
2968 4.80 %/
data treatment. Instructions as to the choice of (b) the partition coefficient.
programs are entered by means of the front FIGURE 24-11 Printer 3. From the data in Problem 1 for species B and C, calculate
panel keyboard. Each of the three memories output from the computing (a) the resolution.
consists of a single chip. The contents of the integrator shown in Figure (b) the selectivity factor oc.
ROM are fixed and cannot be changed. The 24-10. (From J. M. Gill, (c) the length of column necessary to give a resolution of 1.5.
contents of the RMM are also protected but Chromatography. Green (d) the time required to separate Band C with a resolution of 1.5.
can be changed under certain circumstances. Farms, cr: International
4. From the data in Problem 1 for species C and D, calculate
Data can be readily entered or removed from Scientific Communications,
(a) the resolution.
the random access memory. Inc., 1974, Series I, vol. I,
(b) the length of column required to give a resolution of 1.5.
After an analysis is complete, the stored p. 188. With permission.)
(c) the time required to separate C and D with a resolution of 1.5.
data are printed out as shown in Figure 24-11. 5. The following data were obtained by gas-liquid chromatography on a
The first set of data gives the channel or column 4O-an packed column:
number and the sample identification. These at the outset by means of the keyboard. Reten-
entries are followed by a coded description of tion times in seconds and relative areas for
the operating conditions, which were entered peaks then follow. Compound tll,min W,min

Air 2.5
Methylcyclohexane 10.7 1.3
Methylcyclohexene 11.6 1.4
Toluene 14.0 1.8

Calculate
(a) an average number of plates from the data.
(b) the standard deviation for the average in (a).
(c) an average plate height for the column.
Length of packing 22.6 em 6. Referring to Problem 5, calculate the resolution for
Flow rate 0.287 ml/min (a) methylcyclohexene and methylcyclohexane.
VM 1.26 ml (b) methylcyclohexene and toluene.
l's 0.148 ml (c) methylcyclohexane and toluene.
7. If a resolution of 1.5 was required to resolve methylcyclohexane and
methylcyclohexene in Problem 5,
(a) how many plates would be required?
(b) how long would the column have to be if the same packing were
employed?
(c) what would be the retention times for the three compounds on
the column in Problem 7b?
(d) what would be the resolution for the last two compounds on the
column in Problem 7b?
8. If Ys and VAl for the column in Problem 5 were 20.6 and 64.2 ml,
respectively, and a nonretained air peak appeared after 2.5 min, calcu-
late the
(a) capacity factor for each of the three compounds.
(b) partition coefficient for each of the three compounds.
(c) selectivity factor for methylcyclohexane and methylcyclohexene.
(d) selectivity factor for methylcyclohexene and toluene .
.9. A column was found to have the foUowing constants for Equation
24-8 A = O.ot em, B = 0.30 em2/s, and C = O.ot5 s. What is the opti-
mum flow rate and the corresponding minimum plate height?
10. What would be the effect on a chromatographic peak of introducing
the sample at too slow a rate?
11. From distribution studies species M and N are known to have partition
coefficients between water and hexane of 6.50 and 6.31 (K = [M]H,O/
[M]h ••). The two species are to be separated by elution with hexane
in a column packed with silica gel with water on its surface. The ratio
Vs/VAI for the packing is known to be 0.422.
(a) Calculate the capacity factor ratio for each of the solutes.
(b) Calculate the selectivity factor.
(c) How many plates will be needed to provide a resolution of 1.51
(d) How long a column is needed if the plate height of the packing is
1.02 x 10-2 em?
(e) If a flow rate of 7.10 em/s is employed, what time will be required
to elute each of the species?
12. Repeat the calculations in Problem II assuming the two constants have
values of 6.50 and 6.11.
Five ofthe methods listed in Table 24-1 fall in by application of vacuum or by pumping were
the category of liquid chromatography because not effective because within the range of con-
5.0
a liquid mobile phase is employed in each. ditions that were feasible, plate heights were k'·'.2
Included in this group are partition (or liquid- well beyond the minimum in the curve shown 44.7_
liquid~ adsorption (or liquid-solid~ ion-ex- in Figure 24-6 (p. 676). Thus, increased flow 4.0
change, paper, and thin-layer chromatography. rates were accompanied by marked increases
E
The first three employ a glass or metal tube in plate heights and decreased efficiencies. E
Early in the history of liquid chromatog-
~. 3.0
to hold the stationary phase and are thus E
further classified as column procedures: The raphy, it was realized that large increases in ·r~
latter two are carried out on a flat liurface and
are sometimes referred to as plane or planar
column efficiency could be expected to accom-
pany decreases in the particle size of packings.
.
!
Ii:
2.0

chromatographic procedures as a consequence. It was not until the late 19605,however, that the 22.6 I'm
technology of producing and using packings • 0
This chapter treats the five liquid chroma- 13.2 I'm
tographic methodsl and, in additi~n, presents with particle diameters as small as 10 pm was
8.8 I'm
a brief discussion of electrochronuitography, a developed. This technology required sophis-
2.0
procedure in which chromatographic separa- ticated instruments that contrasted markedly
Lineer velocity, cm/ •••
tions are enhanced by application of an elec- with the simple devices employed in classical
trical field. liquid chromatography. The name high-perfor- FIGURE 25-1 Effect of particle size of packing and flow rate
mance liquid chromatography (HPLC) is often upon plate height H. Column dimensions: 30 em x 2.4 mm. Solute:
employed to distinguish these newer proce- N,N-diethyl-p-aminoazobenzene. Mobile phase: mixture of hexane
COLUMN dures from the classical methods, which still methylene chloride, and isopropyl alcohol. The packing labeled
CHROMATOGRAPHY find considerable use for preparative purposes. Corosil®II is a pellicular silica gel (p. 694). [Reproduced from R.
E. Majors, J. Chromatogr. Sci., 11, 88 (1973). With permission.]
The classical liquid chromatographic method
employed a glass tube with a diameter of per-
High-Performance
haps 10 to SO mm to hold a SO- to 5OO-crn
Liquid Chromatography
column of solid particles of the stationary packings consisting of such particles can only
one or more glass or stainless steel reservoirs
phase. To assure reasonable flow rates, the Figure 25-1 shows plots of experimental liquid be realized, however, by pumping at high pres-
each of which contains 1 to 2 liters of a solvent:
particles of the solid were kept larger than chromatographic data relating plate height to sures. As a consequence, the equipment re- Ordinarily, the reservoirs are equipped with a
ISO to 200 JIlT1 in diameter. Ordinarily, the flow rate and particle diameter of packing quired for HPLC is considerably more elabo- means of removing dissolved gases-USUally
head of liquid above the packing sufficed to materials. In none of the plots is the minimum rate than the simple gravity-fed column used in oxygen and nitrogen-that interfere by form-
force the mobile phase down the column. Flow shown by Figure 24-6 reached; generally, in classical liquid chromatography. Figure 25-2 is ing bubbles in the column and the detector
rates were, at best, a few tenths of a milliliter liquid chromatography, this minimum is ob- a block diagram showing the components of systems. These bubbles cause band spreading'
per minute; thus, separations tended to be time served only at prohibitively low flow rates. such equipment. Each of these components is in addition, they often interfere with the per~
consuming. Furthermore, Equation 24-8 does not ade- . discussed in the paragraphs that follow.3 formance of the detector. Degassers may con-
Attempts to speed up the clas~ical proced ure quately describe the relationship between ef- Solvent Reservoir and Degassing System. A sist of a vaccum pumping system, a distillation
ficiency and velocity ofthe mobile phase; here, modern HPLC apparatus is equipped with system, or a device for heating and stirring the
a considerably more complex equation, devel- solvent.
oped by Giddings, must be employed.2 A separation that employs but a single sol-
I A large number of boob on liquid chromalo ••.• phy are It is apparent from Figure 25-1 that sepa- , For a detailed discussion of HPLC systems, see: 1••.. vent is termed an isocratic elution. Frequently,
available. Amona tbc:Ie are: B. L. Karacr, L. R. Snyder, ration efficiency is greatly enhanced at low SII'lDIInUariOll for HiglJ-Pnforrrvmce lJquId Chr_og-
and C. Horvath, A" Introduction to Separlltion ScielJu.
separation efficiency is greatly enhanced by
particle diameters. Reasonable flow rates with rap/ly, eel. J. F. K. Huber. New York: Elsevier. 1975; H.
New York: Wiley, 1973; PrllCtiall HiglJ-Per}'onrItma Enac1bardt, High Per/orMance L/qIIItl C1IroIrtatograp/Jy. gradient elution. Here two (and sometimes
liquid ClrrofIlatograp/ly, ed. C. F. Simpson. London: Chmlkal lAborlltory PrIlCti«. New York: Spriqer- more) solvents, which dift'er significantly in
Heyden, 1977; and R. P. W. Scott, Co_porary Liquid Verlag, 1979; H. Veening, J. Cn. E4uc.50, A429, A481, polarity, are employed. After elution is begun,
Chr_ography, in TeclJniqws of Cllemistry, ed. A. AS29 (1973); and H. Kern and K. Imho~ Amer. Lab. 10 the ratio of the two solvents is varied in a
Wcissberger. New York: Wiley-Interscienc:e, 1976, voL XI. (2) 131 (1978}
programmed way, sometimes continuously
be subsequently damped. The advantage of the
reciprocating pump is that it can be employed
with unlimited volumes of solvent; further-
more, its low internal volume makes it ideal
for gradient elution.
Pneumatic pumps are also employed. In
Peak identity the simplest of these, the mobile phase is con-
1. Benzene tained in a collapsible container housed in a
2. Monochlorobenzene vessel that can be pressurized by a compressed
Mixing 3. Orthodichlorobenzene
_I Pre- 4. l,2,3·trichlorobenzene gas. Pumps ofthis type are simple, inexpensive,
column 5. 1,3,5-trichlorobenzene and pulse-free; they suffer from limited capac-
6. 1,2,4·trichlorobenzene
7. 1,2,3,4·tetrachlorobenzene ity and pressure output as well as a dependence
8. l,2,4,5·tetrachlorobenzene of flow rate on solvent viscosity. In addition,
9. Pentachlorobenzene
10. Hexachlorobenzene they are not amenable to gradient elution.
It should be noted that the high pressures
generated by the pumping devices do not con-
stitute a hazard because liquids are not very
compressible. Thus, rupture of a component of

.-----------
--.----------
the system results only in solvent leakage.
Precolmnns. Some HPLC instruments are
equipped with a so-called precolumn, which
contains a packing chemically identical to that
Waste Waste
in the analytical column. The particle size is
or much larger, however, so that the pressure
fraction drop across the precolumn is negligible with
collector
respect to the rest of the system. The purpose
FIG U R E 25-2 Schematic diagram showing liquid flow in a typical of the precolumn is to remove impurities from
liquid chromatographic instrument. Many instruments do not employ the solvent, and thus prevent contamination of
a flow splitter and produce only a single output stream. the analytical column. In addition, the pre-
column saturates the mobile phase with the
liquid making up the stationary phase; in this
way, stripping of the stationary phase from the
and sometimes in a series of steps. Modem dient elution shortened the time of separation packing of the analytical column is eliminated.
HPLC equipment is often equipped with de- significantly without sacrifice in resolution of Sample Injection Systems. Rotary sampling
vices that introduce solvents from two or more the early peaks. valves, similar in construction to that shown
reservoirs into a mixing chamber at rates that Pumps. Most HPLC pumps have outputs in Figure 24-9, are often employed in HPLC.
vary continuously; the volume ratio of the of at least 1000 psi (lbs/in21 and preferably Slider valves, such as the one shown in Figure
solvents may then be altered linearly or expo- 4000 to 6000 ps~ with a flow delivery rate of FIGURE 25-3 Improvement in separation 25-4, are also common. The slider is manu-
nentially with time. at least 3 ml/min. The flow rate should be con- efficiency by gradient elution. Column: I m x factured from Kel-F and moves in a plane
Figure 25-3 illustrates the improvement in stant to ±2 %. Two types of mechanical pumps 2.1 mm id, precision-bore stainless; packing: between two pieces of polytetrafluoroethylene.
separation brought about by gradient elution. are employed; one is a screw-driven syringe 1% PermaphaseliY ODS. Sample: 5 pi of The sample is drawn into the valve by means
For Figure 25·3b, elution was carried out type and the other a reciprocating pump. The chlorinated benzenes in isopropanol. Detector: of a syringe. The slider is then moved from left
with a 50: 50 (v/v) methanol/water solution. former produces a pulse-free delivery that is UV photometer (254 nm). Conditions: tem- to right so that the sample is placed in the
In the case of Figure 25-380 elution was ini- readily controlled; it suffers from lack of capac- perature,60°C, pressure, 1200 psi. (From J. J. solvent stream. Valves holding various vol-
tiated with a 40:60 ratio of the two solvents; ity and inconvenience in changing solvents. Kirkland, Modern Practice of Liquid Chro- umes from 2 to 100 pi are available.
the methanol concentration was then in- Reciprocating pumps, which are much more matography. New York: Wiley-Interscience, Sample injection can also be accomplished
creased at the rate of 8 ~Jmin. Note that gra- widely used, produce a pulsed flow which must 1971, p. 88. With permission.) by means of a syringe and a self-sealing
casionally,alumina and Celite (a diatomaceous
earth) are employed. Uniform particle sizes in
the 5- to lO-pm range'are desirable.
A second type of packing material, called
pellicular particles, consists of small beads
(often of glass and having diameters of about UV
40 pm 1coated with a 1- to 3-pm layer of a filter

porous material such as silica gel, alumina, or

a- I .
t=
an ion exchange resin. The rate at which equi-
librium between phases is approached is high UVCX l
in these thin layers (p. 675); thus, improved source .•.•.•..• Phototubes
.........
efficiency results. Pellicular packings have .j
the disadvantage of limited sample capacity
Quartz---
(roughly one-tenth that of porous packings). window
Temperature ControL Most liquid chroma-
tography is performed at room temperature
and without thermostating. Water jacketed
columns are available commercially when pre-
cise temperature control is desired.
Detectors. No highly sensitive, universal Inlet t! Outlet
detector system, such as those found in gas Solvent
FIGURE 25-4 A sample injection valve. chromatography (p. 720), is available for
(From R. P. W. Scott, Contemporary Liquid HPLC. Thus, various systems must be em-
Chromatography. New York: Wiley, 1976, ployed, depending upon the nature of the sam-
p. 131. With permission.) ple. Table 25-1 lists the common detectors and
some of their properties.
Perhaps the most common detectors for
HPLC are based upon absorption of ultra-
violet or visible radiation. Both photometers
septum of silicone, neoprene, or Teflon. Alter- and spectrophotometers are available com- TABLE 25-' CHARACTERISTICS OF LIQUID CHROMATOGRAPHY
natively, flow through the column can be mercially. The former usually employ the 254- DETECTORS·
stopped momentarily, a cap on the column and 280-nm lines from a mercury source; at
removed, and the sample injected directly these wavelengths many organic functional Maximum Flow Rate Temperature Useful with Available
onto the head of the column with a syringe. groups absorb (see pp. 172 to 175). Spectro- Detector Basis Type· Sensitivity' Sensitive? Sensitivity Gradient? Commercially?
CoIwn... Columns for HPLC are manu- photometric detectors are considerably more 5 x 10-10
UV absorption S No Low Yes Yes
factured from heavy-walled glass tubing or versatile than photometers because the radia- IR absorption S 10-6 No Low Yes Yes
precision-bore stainless steel; the former are tion can be tailored to the anticipated absorp- Fluorometry S 10-10 No Low Yes Yes
limited to pressures below about 600 psi. The tion peaks of the sample components. Ob- Refractive index G 5 x 10-7 No ± 10-4 °C No Yes
Conductometric S 10-8 Yes ±loC No Yes
typical column ranges from 15 to 150 em in viously, photometric detectors require that the
Moving wire G 10-8 Yes None Yes Yes
length with an inside diameter of about 2 to sample components absorb and the solvent 10-10
Mass spectrometry S No None Yes
3 mm. Longer columns (up to a meter or transmit radiation within the wavelength range Polarography S 10-10 Yes ±loC No
more) are usually prepared by connecting sec- of the instrument. Radioactivity S No None Yes No
tions of shorter columns. Coiled columns are Figure 25-5 is a schematic diagram of an
also employed, although some loss in efficiency ultraviolet photometric detector. • Most of these data were taken from: L. R. Snyder and J. J. Kirkland, llltroduetion to Modern Lf4u1dChr_ogrtlphy.
results from this configuration. Infrared detectors are also available; these New York: Wiley·lnterscience, 1964, p. 165. With permission.
The most common type of packing for are similar in construction to the infrared in- • G - genera1; S - se1ective.
HPLC columns is finitely divided silica gel. 0<:- strument shown in Figure 8-8 (p' 226). , Sensitivity for a favorable sample in gr8IDS per milliliter.
W:'
-==-
able from several commercial sources; typical strengths for a number of solvents (relative to

I- ..
Mirror •...•••• "
==A-~~::-;}f~~
' ..~.... for HPLC is a 10 JlID silica gel packing in n-pentane) when alumina is the adsorbent. The

t . ==t7t~
which 80% of the particles lie in the range of eO values for silica gel are equal to about 0.77
8 to 12 JlID. One company offers columns that for alumina. A solvent with a high eluent
0__ , packed with specially prepared spherical silica strength will elute adsorbed species more
Amplifier & Recorder
power supply gel particles with 95 % of these particles having rapidly than one with a low value.
Reference zero Zero diameters within ± 1 Jlm of the nominal Often, gradient elution is useful in adsorp-
adjust particle size of 6 Jlm. tion chromatography. Here, the easily eluted
FIG URE 25-6 Schematic diagram of a differential refractive-index detector. (Cour- The choice of mobile phase is all- compounds are removed with a solvent having
important for successful liquid-solid chroma- a low eO. Then, an increasing amount of a
tesy of Waters Associates, Inc., Milford, Massachusetts 01757.)
tography; by varying the solvent, the capacity solvent with a high eO is added to remove the
factor (k') for solutes can be varied until they more tightly bound solutes from the column.
fall in the ideal range of I to 10 (p. 680). The Applications. Large differences exist in the
effect of a given solvent on k' depends upon tendencies of compounds to be adsorbed, and
Figure 25-6 is a schematic diagram of a ion source. Before reaching the source, the belt its eluent strength eO, which can be measured these differences serve as the basis of adsorp-
differential refractive index detector. Here, the passes under an infrared evaporator to remove relative to some reference solvent. Thus, for a tion chromatography. For example, a positive
solvent and analyte solutions are separated most of the solvent, through two differential solute X; it has been shown that' correlation can be discerned between adsorp-
by a glass plate mounted at an angle such vacuum chambers, and into a quadrupole type tion properties and the number of hydroxyl
Iog (Kx)1 A ( ° 0)
that bending of the incident beam occurs if mass spectrometer. Here, the sample is flash (Kxh = x e2 - el groups in an organic molecule. A similar cor-
the two solutions differ in refractive index. As vaporized into the ion chamber. Detection relation exists with double bonds. ~ompounds
a result, the position of the beam changes limits as low as 0.2 to 1 ng are reported. where (Kx)1 and (Kxh are the partition containing certain functional groups are more
with respect to the photoelectric surface; the coefficients for X on the solid adsorbent when strongly held than others. The tendency to be
output signal varies as a consequence. The Adsorption Chromatography solvents 1 and 2 are employed. The quantity adsorbed decreases in the following order:
electrical signal is then amplified and recorded Ax is the molecular size of the solute and e~ acid > alcohol > carbonyl > ester > hydro-
All of the early work in chromatography was
to provide the chromatogram. In contrast to and eg are the eluent strengths of solvent 1 carbon. The nature of the adsorbent is also
of the adsorption type, in which the station-
absorbance detectors, the refractive index indi- and 2, respectively. Table 25-2 shows the eluent influential in determining the order of adsorp-
ary phase was the surface of a finely divided
cator is general ra~her than selective and tion. Much of the available knowledge in this
solid. In such a packing, the solute competes
responds to the presence of all solutes. field is empirical; the choice of adsorbent and
with the eluting solvent for sites on the sur-
Another general type detector employs a • L. R. Snyder, Prlnclplu of Athorptlon ChromatO(lraphy. solvent for a given separation frequency must
face of the solid; retention is the result of ad-
continuous moving wire loop to transport a New York: Dekker, 1968. be made on a trial-and-error basis.
sorption forces. Currently, liquid-solid chro-
portion of the eluate to a flame ionization
matography is the most widely used of the
detector (p. 721). The wire first passes through
HPLC techniques for the separation of neutral
the eluate and carries it into an oven where
organic compounds.
the solvent is evaporated. The sample is then
Ideally, the distribution of a solute be-
passed into a furnace and is pyrolyzed. The
tween a solvent and an adsorbing solid is
volatile prod ucts are detected by the ionization TABLE 25·2 ELUENT STRENGTH OF SOME COMMON SOLVENTS
described by an adsorption isotherm such as
detector. WITH ALUMINA ADSORBENTS
that shown as curve A in Figure 24-1 (p.669).
Mass spectrometry has also been employed
Note that the curve is nearly a straight line
as a sensitive specific detector.4 One commer- Eluent Eluent
at low concentrations; under these circum- Strength, Strength,
cial instrument employs a stainless-steel or Solvent £8 Solvent £8
stances, elution is linear and peak distortion
polyimide belt to transport the eluate to the
is minimal. Fluoroalkanes -015 Chloroform 0.40
Stationary and Mobile Pbaes. Silica gel is n-Pentane 0.00 Methylethyl ketone 0.51
Petroleum ether 0.01 Acetone 0.56
by far the most widely employed adsorbent 0.63
Cyclohexane 0.04 Diethylamine
• W. H. McFadden, D. C. Bradford, D. E. Games, and for liquid-solid chromatography. Alumina also Carbon tetrachloride 0.18 Pyridme 0.71
J. L. Gower, Amer. Lab. 9 (10~ 55 (1977); and P. J. iinds considerable use. Packing materials of n-Propyl chloride 0.30 n-Propanol 0.82
Arpino and G. Guiochon, AMI. Chem. 51. 682A (1979~ specified particle diameter ranges are avail- Benzene 0.32 Methanol 0.95
Solid Supports. The most widely used solid The behavior of the chemically bonded silica
supp6rt for partition chromatography has been surface appears to be intermediate between a
silicic acid or silica gel. This material adsorbs solid surface at which adsorption occurs and
water strongly; the stationary phase is often an immobilized liquid at which a liquid-liquid
aqueous as a consequence. For some separa- equilibrium exists. Chemically bonded surfaces
tions, the inclusion of a buffer or a strong acid offer a considerable advantage over oidinary
(or base) in the water film has proved helpful. solid-supported liquids in that the stationary
Polar solvents such as aliphatic alcohols, ·gly- phase cannot be stripped of its liquid by the
cols. or nitromethane, alone or mixed with mobile phase. On the other hand, chemically
water, have also served as the stationary phase bonded surfaces suffer from limited loading
on silica gel. Other "Support media include capacities.
alumina, diatomaceous earth, starch, cellulose, Applications. Partition chromatography has
and powdered glass; water and a variety of become a powerful tool for the separation of
organic liquids have been used to coat these closely related substances. Typical examples
solids. include the resolution of the numerous amino
The mobile phase may be a pure solvent or acids formed in the hydrolysis of a protein,
a mixture of solvents; its polarity must be the separation and analysis of closely related
markedly different from the stationary liquid aliphatic alcohols, and the separation of sugar
FIGURE 25-7 A liquid-solid chromato- so that the two are immiscible. Ordinarily, derivatives. Figure 25-8 illustrates an applica- FIGURE 25-8 Separation of ami-
gram for adrenal steroid-cortisones. Peaks: the mOre polar of the two solvents is incor- tion of the procedure to the separation of no acids by partition chromatog-
(A) cortisol; (B) cortisone; (C) corticosterone. porated on the solid and serves as the station- amino acids. Here, the chemically bonded raphy. Peaks: A, proline; B, leucine;
Column: 25 em x 4.6 mm packed with 10 Jim ary phase; reverse-phase chromatography ~m- silica gel described in the previous section C, isoleucine; D, valine; E, phenyl-
silica gel. Pressure: 225 psi. Flow rate: 59.8 ploys a nonpolar stationary phase and polar was employed as the support. Note that the alanine; F, methionine. Column:
ml/hr. Mobile phase: 75 % heptane/25 % etha- mobile one. The choice of liquid pairs is largely separation required less than 2 min. 10 cm x 4.6 mm; packed with 20 Jim
nol. (Courtesy of Whatman, Inc., Clifton, NJ.) empirical. As mentioned earlier, gradient elu- chemically bonded silica gel. Mobile
tion is frequently employed to enhance separa- phase: 50-50 heptane/chloroform.
lon-Exchange
tion efficiency. Pressure: 80 psi. Flow rate: 1.9cm3/
Chrolriatography
Bonded Phase Packings. A type of pack- min. Sample size: 1 Jll. (Courtesy of
Figure 25-7 shows a typical liquid-solid
ing that is becoming increasingly popular Ion-cxchange resins are among the most widely Gow-Mac Instrument Co., Bridge-
separation by HPLC.
for HPLC consists of pure silica gel particles used packings in column chromatography. In water, New Jersey 08807. With
onto' which an organic group has been at- contrast to the other column methods, the permission.)
Partition Chromatography tached chemically. As an example, a hydro- solvent is generally water and the species to
Partition or liquid-liquid chromatography carbon surface can be formed by the reaction be separated are ions. Thus, ion-exchange
was originated in 1941 in the Nobel prize- of chlorooctadecyl silane with the OH groups chromatography finds wide use in inorganic Ion exchange is a process involving an inter-
winning work of Martin and Synge.6 In their on the surface of silica gel. That is, chemistry. It also has been used extensively for change of ions of like sign between a solution
studies, they were able to demonstrate that a the separation of amino acids and other and an essentially insoluble solid in contact
R R
properly prepared column can have a plate I I·
organic acids and bases. 7 with the solution. Many substances, both
height as small as 0.002 em. Thus, a 1O-cm SiOH + CI-Si-R Si-O-Si-R natural and synthetic, act as ion exchangers.
I I
column of this type may contain as many R R The ion-cxchange properties of clays and
as SOOOtheoretical plates. High separation . . zeolites have been recognized and studied for
efficiencies are to be expected even with rela- where R is the octodecyl group and the Si in 1 For a detailed treatment oC ion-exchange chromatog- more than a century. Synthetic ion-cxchange
tively short columns. the circle represents one of many SiOH groups raphy. see: W. Rieman. III and H. F. Walton, Ion Ex- resins were first produced in 1935 and have
cha1lfle in Analytical Chnnistry. New York: Pergamon since found widespread laboratory and indus-
- on the surface of the gel particle. Other groups
Press. 1970; P. R. Brown and A. Krstulovic, Separation
that have been bonded to silica gel include anti Purification, in Technlqua of Chnnistry. eels. E. S. trial application for water softening, water de-
6A. J. P. Martin and R. L. M. SynJC, Biochnn. J, 35. aliphatic amines, ethers, and nitrates· as well Perry and A. Weissberger. New York: Wiley-Inter- ionization, solution purification, and ion
1358 (1941~ as aromatic hydrocarbons. science. 1978.vol. 12. Chapter 4. separation.
Synthetic ion-cxchange res~ are high. ~o- lisped for each of the alkali ions:
charged species. Within a given charge group, original solution was O.OIM in each cation.
lecular weight polymeric matenals contammg however, differences appear that are related to Figure 25-9b illustrates the use of an anion-
. RH+B+~RB+H+
large numbers of an ionic functiQ11al!P'0up ~r the size of the hydrated ion as well as other exchange column for separating the anions in
molecule. For cation exchange, there IS a ChOlce where B + represents either Na + or K +. properties. Thus for a typical sulfonated cation- a boiler blow-down water; here, the ion con-
between strong acid resins containing sulfonic Equilibrium constants for ion-cxchange re- exchange resin, values for KD decrease in the centrations were in the low parts-per-million
acid groups (RSOi"H+) and weak acid resins actions take the form order Cs+ > Rb+ > K+ > NHt > Na+ range. In both separations, quantitative data
containing carboxylic acid groups (RCOOH); > H+ > Li +. For divalent cations, the order is were obtained by comparison of peak heights
the former have wider application. Anion- Ba2+ > Pb2+ > Sr2+ > Ca2+ > Cd2+ > to heights for appropriate standards. It is of
exchange resins contain basic functional Cu2+ > Zn2+ > Mg2+. interest to note that a conductivity detector
groups, generally amines, attached to the poly- .The techniques for fractionation of ions was employed in both determinations. In order
mer molecule. Strong base exchangers are where (RB] and (RH] are the concentrations with KD values that are relatively close to one to prevent the conductivity of the eluent ions
quaternary amines (RN(CH3)tOW); ,,:eak (strictly the activities) of t~e alkali and hyd!o- another are analogous to those described for from swamping that of the analyte, the ana-
base types contain secondary or tertiary gen ions in the solid resm phase. Equation adsorption and partition chromatography. lytical column was followed by a stripping
amines. 25-2 can be rewritten in the form For example, Figure 25-9a shows an HPLC column. For the cation separation. the stripper
A cation-cxchange process is illustrated by separation of the alkali-metal ions on a pellic- was packed with the hydroxide form of an
the equilibrium (RB] K(RH]_
uIar sulfonic acid cation-cxchange resin; the anion-cxchange resin that absorbed the HCI
(B+] = (H+] - KD
xRSOi"H+ + M"+ ~
where KD is defined as the distribution co-
solid solution
efficient. Under the conditions extant during
(RSOi")xM"+ + xH+ elution with hydrochloric acid, K~ is approxi-
solid solution mately constant because the hydrogen ion con-
centration of the eluate is large relative to the
where M"+ represents a cation and R repre- concentration of potassium and sodium ions.
Also, the resin has an enormous number of
~
c
sents a part of a resin molecule. The analo- 0 K'
gous process involving a typical anion- exchange sites relative to the number of alkali- ~
exchange resin can be written metal ions in the sample. Thus, the overall 5
concentrations (H+] and (RH] are not affected ~ Ab'
c
significantly by shifts in the equilibrium (Equa- c '0
xRN(CH3)tOW + A"- ~ 0.

tion 25-1). Therefore, under conditions where .§


solid solution
[RH] and [H+] are large with tespect to [RB] ~
[RN(CH3);-],.A"- + XOW and [B+], Equation 25-3 can be employed in E:
solid solution the same way as Equation 24-1 (p. 6681 and
the general theory of chromatography, which 5 10
where A"- is an anion. we have already. described, can be applied to Min
Ion-cxchange resins have been successfully a stationary phase consisting of an ion- (a)
employed as the stationary phase in elution exchange resin.
chromatography. For example, Beukenkamp Note that KD in Equation 25-3 represents
FIGURE 25-9 Separation of inorganic ions by ion-exchange chromatog-
and Reiman separated sodium and potassium the affinity of the resin for the ion B+ relative
raphy. (al Column: 250 x 9 mm with a sulfonic acid. Pellicular packing.
ions on a column packed with a sulfonic acid to another ion (here, H+). Where KD is large,
Eluent: O.ot N HCI. Sample size: 0.1 ml. [Reprinted with permission from
resin in its acidic form.s When the sample a strong tendency exists for the solid phase to
H. Small, T. S. Stevens, and W. C. Bauman, Anal. Chern., 47, 1803 (1975).
was introduced at the top of the column, the retain ion B; where KD is small, the reverse
Copyright by the American Chemical Society.] (b) Column: 500 x 2.8 mm.
following exchange equilibrium was estab- obtains. By selecting a common reference ion
Flow rate: 118 ml/hr. Packing: pellicular anion-exchange resin. Eluent:
such as H +, distribution ratios for different
0.(XH5 M NaHC03 for glycolate + CI- and 0.003 M NaHC03 + 0.0024 M
ions on a given type of resin can be compared.
Na2C03 for other anions. Sample size: 100 Illiters. [Reprinted with permission
• J. Beukenkamp aDd W. Reiman III. Anal. Chmt. 1Z, Such experiments reveal that polyvalent ions
from T. M. Stevens, V. T. Turkelson, and W. A. Albe, Anal. Chern., 49, 1176
582 (19SO~ are much more strongly held than singly
(1977). Copyright by the American Chemical Society.]
the sample.9 Several names have been given extent depending upon their size, their shape, Theory of Gel Chromatography. The total
eluent without affecting the cations. The 'reac-
to this procedure, including gel-permeation and sometimes, upon their tendency to be ad- volume of a column packed with a gel that
tion in the stripJlCl: is
chromatography, exclusion chromatography, sorbed by the gel. has been swelled by water (or other solvent)
H+ + Cl- + ROH ~ RCl + HzO and moltcular-sieve chromatography. Column Packing. The stationary phase in is given by
For the anion separations, the stripper reac- Ordinarily, gel chromatography is per- gel chromatography consists of beads of a p0-
v,=v,+v,+Y"
tion can be written as formed on a column by the elution method. rous polymeric material that readily absorbs
Here, the degree of retardation depends upon water (and in some instances, other solvents) where V, is the volume occupied by the solid
Na+ + HCOi + RH ~ RNa + HzC03 and swells as a consequence. The resulting matrix of the gel, V, is the volume of solvent
the extent to which the solute molecules or
Here, the HzC03 has no significant effect on ions can penetrate that part of the solution solid contains a large volume of solvent held held in its interstices, and Y"is the free volume
the conductivity of the solvent. phase which is held within the pores of the in the interstices of the polymeric network. The outside the gel particles. Assuming no mixing
Ion-exchange chromatography finds wide- highly porous gel-like packing material. Mole- average size of the resulting pores or interstices or diffusion, Y"also represents the volume of
spread use in amino acid analyzers. For ex- cules or ions larger than the pores of the gel is directly related to the quantity of solvent solvent required to carry through the column
ample, Figure 25-10 shows a chromatogram of are completely excluded from the interior, adsorbed; this in turn is determined by the those components too large to enter the pores
a mixture containing 1.0 x 10- 8 mol each of while smaller species are more or less free to amount of cross-linking in the polymer of the gel. In fact, however, some mixing and
17 amino acids. Separation time was only 42 enter these regions. Thus, larger molecules molecules. diffusion will occur, and as a consequence the
min. Gradient elution with buffers of different pass quickly through the column, while small- One of the most widely used polymers is components will appear in a Gaussian-shaped
pH was employed. er ones are retarded to a greater or lesser prepared by cross-linking the polysaccharide band with a concentration maximum at Y,,:
dextran with epichlorhydrin. By varying the For components small enough to enter freely
Gel Chromatography
amount of the latter, a series of resins with into the interstices of the ge~ band maxima
Gel chromatography is a technique in which • For further information: see: S. D. Abbott, Amer.lAb, differing pore sizes is obtained. These gels sell will appear at the end of the column at a
fractionation is based, at least in part, upon 9 (9~ 8 (19n); and H. Determann, Gel Chromatography. under the trade name of Sephadex®. Table volume corresponding to (V, + y"). Generally,
the molecular size and shape of the species in New York: Springer-Verlag, 1965. 25-3 describes the properties of some of these V" v., and V, are of the same order of mag-
.Sephadex gels. Another group of commercial nitude; thus, a gel column permits separation
resins consists of polyacrylamide cross-linked of the large molecules of a sample from the
with methylene bisacrylamide. small with a minimal volume of wash liquid.

TABLE 25-3 PROPERTIES OF COMMERCIAL


Sephedex B
GELS

VollBDeRelatioll'lhips:
mllg original gel
Approximate Excll.ion
Gel Designation JI. JI, JI. Lunit, mol WI
G-I0 0.6 1.0 0.9 700
G-15 0.6 1.5 0.9 1,500
G-25 0.5 2.5 2 5,000
G-50 1 5 4 10,000
G-75 1 7 5 50,000
20 25
G-l00 I 10 6 100,000
t(min) G-I50 1 15 8 150,000
G-200 1 20 9 200,000
FIGURE 26-10 Separation of amino acids on an ion-exchange
column. Packing: cation exchange with particle size of 8 lUll·
• ~hannacia Fine Chemk:als,lnc, Piscataway, NJ.
Pressure: 2700 psi. [Reprinted J. R. Benson, from American Labora- • V, a volume occupied by solid gel matrix.
tory, 4 (10) 53, 1972. Copyright 1972 by International Scientific V. a volume of solvent held interstitially.
Communications, Inc., Fairfield, CT.] V. = liquid volume between the gel beads.
Molecules of intermediate size are able to example, it is evident from Table 25-4 that the charides. For example. Sephadex G-75 permits stable materials whereas gas-liquid chroma.
the clean separation of ribonuclease, trypsin, tography is not. Often, the two methods are
transfer into some fraction K1 of the inter- G-25 gel will effectively separate salts and
amino acids from most proteins. Here the pro- pepsin, and cytochrome C from higher molec- complementary.
stitially held solvent; for these, the elution
volume V. is teins, all with K1 values of zero, are completely ular weight proteins such as serum albumin,
excluded from the gel interior, while the low hemoglobin, and fibrinogen (see Table ~5-4).
v. = Yo + K1 l'I (25-4)
molecular weight species have K1 values ap- Finally, it should be noted that gel-perme- PLANAR
Equation 25-4 describes the behavior of a gel proaching unity. As a result, clean separation ation chromatography has been used by poly- CHROMATOGRAPHY
column with respect to all solutes. For mole- can be achieved even with large sample mer chemists and biochemists for estimation
of the molecular weights of large molecules. Planar chromatography takes two forms. In -
cules too large to enter the gel pores, K1 = 0 volumes (as much as 20 to 30% of the total
Here, the elution volume of the unknown is thin-layer chromatography, separation takes
and V. = Yo; for molecules that can enter the volume of the bed). This type of separation
compared with elution volumes for a series of place on a layer of a finely divided solid that _
pores unhindered, K1 = 1 and V. = (Yo + l'I). performs the same function as a dialysis
standard compounds that possess the same is fixed on a flat surface. I 0 A sheet or strip of ;
In arriving at Equation 25-4, the assumption through a cellophane membrane, but is more
chemical characteristics. heavy filter paper serves as the separatory _
was made that no interaction, such as adsorp- rapid and convenient to perform.
Sephadex G-25 and G-SO gels have also medium for paper chromatography. I I Histor- -
tion, occurs between the solute molecules and
ically, paper chromatography was first used
the gel surfaces. If such interaction does occur, proved useful for the fractionation of peptides Comparison of
the amount of interstitially held solute will be having a size range intermediate between the for separations in the middle of the nineteenth;
High-Performance century. It was not until the late 19405, how· .
increased; for interacting solutes that can freely proteins shown in Table 25-4 and the low Liquid Chromatography ever, that the usefulness of the technique be--
enter the poreS, K1 will be greater than unity. molecular weight species. These compounds with Ga.-Liquid Chromatography came fully appreciated by scientists. Thin-layer
Table 25-4 gives experimentally determined have K1 values greater than zero but less than
values of K1 for some dextrant-type gels. one. Elution techniques for small samples are Table 25-5 provides a comparison between
ApplicatiOlL'lof Gel OJromatography. Tight- similar to those used in other types of chroma- high-performance liquid chromatography
(HPLC) and gas-liquid chromatography, I. References:J. G. Kirchner, in Tltin Layer Chromatog- :
ly cross-linked gels with small pores, such as tography. raphy. ed. A. Weissberger. New York: Wiley-Interscience,!
Sephadex G-25 and G-50, have found wide The more porous gels shown in Table 25-4 which is discussed in the next chapter. Where
1978;and J. T. Touchstone and M. R. Dobbins, PrtM:tice ,
application to the desalting or removal of low have found wide application to the fractiona- either is applicable, gas-liquid chromatog- of Thin Layer Chromatography. New York: Wiley, 1978.-
molecular weight molecules from high molecu- tion and the purification of macromolecules raphy offers the advantage of speed and sim-
II References:R.Stock and C. B. F. Rice,Chromatograpltic
lar weight natural-product molecules. For such as proteins, nucleic acids, and polysac- plicity of equipment. On the other hand,
Methods, 3d ed. London: Chapman and Hal~ 1974,
HPLC is applicable to nonvolatile substances Chapter 3; and J. Shenna and G. Zweig. Paper Chroma-
(including inorganic ions) and thermally un- tography. New York: Academic Press, 1971.

TABLE 25-4 Kd VALUES WITH SEPHADEX GELSs

K~
TABLE 25-5 COMPARISON OF HIGH-PERFORMANCE
APr:ximate Sephadex Sephadex Sephadex Sephadex LIQUID AND GAS-LIQUID CHROMATOGRAPHY
Substance mo 1ft G-25 G-7S G-lOO G-200
Ammonium sulfate 132 0.9 Characteristics ~essed by both methods:
Potassium chloride 74 1.0 Efficient, highly selective, and widely applicable
Tryptophan 204 2.2 1.2 Only small sample required
Glycine 76 0.9 1.0 Ordinarily nondestructive of sample
Ribonuclease 13,000 0 0.4 Readily adapted to quantitative analysis
Trypsin 24,000 0 0.3 0.5 0.7 Particular advantages of high-performance liquid chromatography:
Serum albumin 75,000 0 0 0.2 0.4 Can accommodate nonvolatile and thermally unstable samples
Fibrinogen 330,000 0 0 0.0 0.0 Generally applicable to inorganic ions
Particular advantages of ps-Iiquid chromatography:
• ~annacia F~e ~em~ts. Inc. ~ta~y. NJ. Fr~m B. Ge1011e,-Fractionation of Proteins, Peptides. and Amino Sim{lle and inexpensive equipment
Acids by Gel FiltratiOn, In New BIOChemical SeparatJOlU, eds. A. T. James and L. J. Morris. Princeton: Van Nostrand Rapid
1964.See Table 25-3 for the properties of these gels. •
chromatography was developed in the late factor RF, which is defined as particle sizes and a uniform layer thickness. penciL After the sample solvent has evap-
19~ and by now has become the more im- Several spreading devices have been developed orated, the plate is placed in a closed con-
R = distance of solute motion
portant of the two planar methods. F for forming a layer of constant thickness; some tainer saturated with vapors of the developer.
Both paper and thin-layer chromatography distance of solvent motion of these are available commercially. Prepared One end of the plate is then wetted with the
provide remarkably simple and inexpensive To compensate for uncontrolled variables, plates can also be purchased from a number developing solvent by means of one of the
means for separating and identifying the com- the distance traveled by a solute is frequently of sources. arrangements shown in Figure 25-11. (Note
~nents of small samples of complex inorganic, compared with that for a standard substance Plate De\'eIopment. Typical arrangements that the Sample is not immersed in the de-
organic, and biochemical substances. Further- under identical conditions. The ratio of these for the development of thin-layer plates are veloper.) After the developer has traversed the
more, the methods, particularly thiri-Iayer distances is designated as RIId • shown in Figure 25-11. A drop of the sample length of the plate, the latter is removed, dried,
chromatography, permit reasonably accurate is placed near one end of the plate (most and the position of the components determined
quantitative determination of the concentra- plates have dimensions of 5 x 20 or 20 x in any of several ways.
tions of the components of such mixtures. Thin-Layer Chromatography
20 cm~ and its position is marked with a Figure 25-12 illustrates the separation of
Stationary Phase. The solid absorbents (o~
General Considerations sometimes adsorbents) employed in thin:lay lr
chromatography are similar in chemical com-
In planar chromatography, a drop of a solu- position and particle size to those already de-
tion containing the sample is introdu~ at scribed in the discussion of the various types
some point on the planar surface of the sta- of column chromatography. The most widely
tionary phase. After evaporation of the solvent, used substance is silica geL which often func-
the chromatogram is developed by the flow of tions as a support for water or other polar
a mobile phase (the developer) acros$ the solvents for liquid-liquid separations. If, how-
surface. Movement of the developer is caused ever, the silica layer is oven-dried after prepara-
by capillary forces. In ascending development, tion, much of the moisture is lost and the
the motion of the mobile phase is upward; in .surface becomes a predominately solid one that
radial development, it is outward from a central serves as an adsorbent for liquid-solid separa-
spot. Gravity also contributes to solvent flow tions. In this latter mode, care must be taken
in descending development, where movement is to avoid exposing the surface to the atmo-
in a downward direction. sphere, for it has been demonstrated that water
The equilibrium upon which planar chro- adsorption from air occurs in a few minutes;
matography is based is most often of the Iiquid- the surface then reverts to a support for a liquid.
Iiqqid type, although liquid-solid and ion- Ion-exchange resins and Sephadex® gels
exchange equilibria are also encountered. Usu- are also employed as the stationary phase in
ally, the stationary phase is water or some other thin-layer chromatography.
polar liquid; reverse-phase separations also Preparation of Thin-Layer Plates. A thin-
find application, however. layer plate can be prepared by spreading an
The general principles developed earlier for aqueous slurry of the finely ground solid onto
column chromatography appear to apply the clean surface of a glass or mylar plate or
equally well to paper and thin-layer media. microscope slide. A binder often is incor-
Within these media, repeated transfer of solute porated into the slurry to enhance adhesion
between the stationary and mobile pl;lases of the solid particles to the glass and to one
occurs. The rate of movement of the s!>lute another. The plate is then allowed to stand
depends upon its partition coefficient . Ie)
until the. layer has set up and adheres tightly
K = CS/CII to the surface; for some purposes, it may be FIGURE 25-11 Three types of apparatus for thin-layer
heated in an oven for several hours. chromatography. (a) Ascending flow; (b) descending ftow;
It is customary to describe the travel of a and (c) horizontal ftow. S: Initial position of sample; D:
The cleanness of separations on the thin
particular solute in terms of its retardation layer depends upon having a narrow range of developer; C: chromatographic surface; W: cotton wick.
I

PRINCIPLES OF INSTRUMENTAL ANALYSIS

methods, which can be applied to most organic an adsorbent or an ion-exchange resin, thus Many of the experimental difficulties asso-

• •
mixtures, "involve spraying with solutions of permitting adsorption and ion-exchange paper ciated with free-solution electrophoresis are
iodine or sulfuric acid; both react with organic chromatography. avoided if the separations are carried out in a
10
compounds to yield dark reaction products . stabilizing medium such as a paper, a layer of

••
Several specific reagents (such as ninhydrin) finely divided solid, or a column packed with
9 are also useful for locating sc!parated species. ELECTROPHORESIS AND a suitable solid. Here, the experimental tech-

'6• •
"(
7 8
Fluorescent substances can be detected with ELECTROCHROMATOGRAPHV niques closely resemble the various chroma-
an ultraviolet lamp. Alternatively, a fluorescent tographic methods discussed earlier, with the
i material is incorporated into the stationary Electrophoresis is defined as the migration of additional parameter ofthe superimposed elec-
~ phase. After development, the plate is examined particles through a solution under the influence trical field. Depending upon the properties of
5

• 4 with ultraviolet light. Here, all of the plate of an electrical field. Historically, the particles the medium, the separations may result pri-

.-
fluoresces except where the nonfluorescing referred to were colloidal and owed their marily from the electrophoretic effect or from
3
sample components are located. charge to adsorbed ions. This definition has a combination of electrophoresis and adsorp-
Quantitative Analysis. A semiquantitative now become too restrictive, and the term elec- tion, ion exchange, or other distribution equi-
1 2
estimate of the amount of a component pres- trophoresis is presently applied both to the libria. Methods based upon electrophoresis in
ent can be obtained by comparing the area migratio, of individual ions and to colloidal a stabilizing medium bear a variety of names,
of a spot with that of a standard. Better data aggregates. Electrophoretic methods provide including electrochromatography, zone electro-
can be obtained by scraping the spot from the a powerful means of fractionating the com- phoresis, electromigration, and ionophoresis.·
FIGURE 25-12 Two-dimensionalthin-Iayer
plate, extracting the analyte from the result- ponents of a mixture, be they aggregated or Our discussion is limited to electrochromatog-
chromatogram (silica gel) of some amino acids. monodispersed. raphy.u
ing solid, and measuring the analyte by a .suit-
Solvent A: toluene-2-chloroethanol-pyridine. Electrophoretic methods can be subdivided
able physical or chemical method. Finally, a
Solvent B: chloroform-benzyl alcohol-acetic into two :categories, depending upon whether Experimental Methods
scanning densitometer can be employed to
acid. Amino acids: (1) aspartic acid; (2) gluta- the sepatation is carried out in the absence of Electrochromatography
measure radiation emitted from the spot by
mic acid; (3) serine; (4) p-alanine; (5) glycine;
fluorescence or reflection.12 ) or presence of a supporting or stabilizing
(6) alanine; (7) methionine; (8) valine; (9) iso- The solid media employed in electrochroma-
medium. In thefree-solutionmethod, the sample
leucine; and (10) cysteine. / solution is introduced as a band at the bottom tography are as numerous and varied as those
P•••• r Chromatography
encountered in the other chromatographic
of a V-tube filled with a buffered liquid. A
Separations by paper chromatography are per- methods. Examples include paper, cellulose
field is applied by means of electrodes located
formed in the same way as those on thin-layer acetate membranes, cellulose powders, starch
near the tube ends; the differential movement
plates. Usually, special papers, which are highly gels, ion-exchange resins, glass powders, and
amino acids in a mixture by development in of the charged particles toward one or the other
purified and reproducible as to porosity and agar gels. Depending upon the physical nature
two directions (two-dimensional planar chro- electrode is observed. Separations occur as a
thickness, are employed. Such papers 'contain of the solid, separations are performed on
matography). The sample was placed in one result of differences in migration rates; these
strips of paper or membranes, in columns, in
sufficient adsorbed water that paper chroma-
corner of a square plate and development was rates in turn are related to the charge-to-mass
tography can be classified as a liquid-liquid trays, or in thin layers supported by glass or
performed in the ascending direction with sol- ratios and the inherent mobilities of the species
type. Other liquids can be made to displace the plastic. Despite certain disadvantages, filter pa-
vent A. This solvent was then removed by in.the medium. The free-solution method was
water, however, thus providing a different type per and cellulose acetate are the most widely
evaporation, the plate was rotated 90 del. fol- perfected and applied to the separation of
of stationary phase. For example, paper treated used stabilizing materials; we will focus our
lowing which ascending development with proteins by Tiselius; for this work, he received
with silicone or paraffin oil permits reverse- attention on paper electrochromatography.
solvent B was carried out. After solvent remov- the 1948 Nobel prize. This procedure has been
phase paper chromatography, in which the Figure 25-13 is a schematic diagram illus-
a~ the positions ofthe amino acids were deter- of signal' importance in the development of
mobile phase is a polar solvent. Also available trating three of the many ways for performing
mined by spraying with ninhydrin, a reagent biochemi,stry; nevertheless, its widespread ap- a paper electrophoresis. In Figure 25-13a, a
commercially are special papers that contain
that forms a pink to purple product with plicationihas been hindered by such experimen-
amino acids. The spots were identified by com- tal problems as the tendency of the separated
parison of their positions with those of components to mix by convection, as a con-
n For a detailed discussion, see: E. Heftmann, Chr_
standards. sequence of thermal and density gradients, as rography, 3d ed. New York: Reinhold. 1974,Chapters 10
12 For a description or c:ommerciaDyavailable densi-
Identification of Species. Several methods well as mechanical vibrations. In addition, and 11; and J. R. Sargent and S. G. George, M~rltoh ill
tometen u well u other thin-layer equipment, see: P. F.
are employed to locate the various sample Lou and R. J. Hurtubise, J. Chern. Educ. 48, A437 elaborate optical systems are often needed to Zone Elecrrophoresm, 3d ed. Poole, Ens1and: BDH
components after separation. Two common (1971); and D. Rogers, Amer. Lab. II (5~ 77 (1979~ detect the bands of the separated species. Chemicals. Ltd., 1975.
strip of paper is held horizontally between The technique illustrated in Figure 25-13c
two containers filled with a buffer mixture; has been employed for both analytical and pre-
the paper is well soaked with buffer and evap- paratory purposes. In the former, an amount
oration is prevented by housing the appa- of buffer just equal to that retained by the
ratus in an air-tight container. The sample is paper is allowed to pass before the experiment Cd Fe

introduced as a band at the center of the


strip, and a dc potential of 100 to 1000 V is
applied across the two electrodes. The latter
is diScontinued; the paper is then removed,
dried, and the detection reagents are applied.
In preparatory applications, the buffer flow is
Ag,-"Cu Co

are sufficiently isolated from the paper to pre- continued and the various fractions are col-
vent the electrode reactions from altering the lected as shown in Figure 25-13c.
composition of the buffer on the paper. Cur-
rents in the milliampere range are ordinarily
observed. After a suitable electrolysis period, Applications of FIGURE 25-14 Electrochromatograms for
the paper is removed, dried, and the compo- Electrochromatography :inorganic ions. (a) Cations (0.05 M) in 0.1 M
Sample
application nent bands are detected by applying suitable tartaric acid; wash liquid was 0.01 M in
point Electrochromatographic methods are indis- ammonium tartrate, 0.005 M in dimethylgly-
colorimetric reagents.
pensible to the clinical chemist and the bio- ;oxime, and 4 M in ammonia. Spots were
+ The inverted V arrangement shown in Fig-
chemist, who use them for fractionating an detected with hydrogen sulfide: 160 V, 100
Electrode ure 25-13b is also widely used. Here, the sample
amazing variety of biological materials. The mamp, 20 min. (b) Continuous separation.
is introduced at the apex of the V; cationic
widest application has perhaps been in clini- Solution 0.005 M each in As(III~ Sb(III~ and
species migrate down one arm of the V and
cal diagnosis; here, electrochromatography Sn(II); solution was also 0.02 M in tarta~c
anionic constituents down the other.
makes possible the separation of proteins and acid, 0.04 M in lactic acid, and 0.04 M m
Many other modifications of the apparatus other large molecules contained in serum,
shown in Figures 25-13a and 25-13b have dl-alanine. Paths were detected with hydrogen
urine, spinal fluid, gastric juices, and other sulfide. 300 V, 95 mamp. [Reprinted with per-
been described, and several are available com- body fluids.
mercially. In some, the paper is supported mission f~om H. H. Strain and J. C. Sullivan,
Electrochromatography has also been ap- Anal. Chern., 23, 816 (1951). Copyright by the
(either horizontally or at some suitable angle) plied widely by biochemists to fractionate
between two glass or plastic plates; the plates American Chemical Society.]
smaller molecules such as alkaloids, antibio-
may be cooled to dissipate the heat generated
tics, nucleic acids, vitamins, natural pigments,
by the current. In all of these devices, the rate steroids, amino acids, carbohydrates, and or-
at which components migrate is controlled pri-
ganic acids.
marily by their charge-ta-mass ratios and their It has also' been shown that electrochroma- ion has been retarded by the precipitant
mobilities, with adsorption or other equilibria dimethylglyoxime.
tography provides a convenient means for
having only a secondary influence. Figure 25-14b illustrates the path traveled
separating inorganic ions. Figure 25-14a illus-
Figure 25-13c illustrates an apparatus for a by the components of a sample when sub-
trates the separation of six metallic ions from
type of two-dimensional electrochromatog- jected to continuous electrophoresis and ~Iu-
a complexing medium. Note that both anodic
raphy. Here, the paper is held vertically and and cathodic migration have occurred de- tion. By collecting the eluate at appropnate
the sample components are carried down the pending' upon the charge of the complex ion. places on the bottom edge of the paper, solu-
sheet by the flow of a buffered solvent. Sepa- Note also that the movement of the nickel(lI) ; tions of the separated species can be obtained.

00000000
ration then occurs in the vertical direction as a
consequence of differences in distribution ra-

=
tios between the mobile and the fixed phases.
SImple cOllection In addition, however, a field is applied at right
angles to the solution flow, and differential
leI 1. Define the following terms: (a) isocratic elution, ~b) grndadien(f)t
electromigration along the horizontal axis oc-
curs as a consequence. Thus, the various species (c) pellicular beads, (d) eluent strength, (e) desaltmg, a -
FIGURE 25-13 Some types of apparatus in the sample describe a radial path from the tion factor (RF).
for paper electrochromatography. point at which the sample is introduced. 2. What is the purpose of the precolumn in Figure 25-27
3. What are the relative advantages and disadvantages of pellicular beads dye run in the same gel. It is found that a graph of log(molecular
as compared with porous packing material? weight) vs. relative mobility is a reasonably linear fot most proteins.
4. In preparing a benzene/acetone gradient for an alumina HPLC (a) From the data below,14 estimate the molecular weight of the
column, is it desirable to increase or decrease the proportion of ben- . unknown.
zene as the column is eluted?
5. Would a strong acid resin or weak acid resin be more appropriate for
cation exchange if the solvent is O.01M HCI?
6. Referring to Table 25-4, what can be deduced about the interaction of Lysozyme 14,400 0.90
tryptophan with Sephadex G-25? Trypsin 23,300 0.73
7. Explain the trend in K4 values for trypsin across Table 25-4. Lactate dehydrogenase 36,000 0.57
8. Gel permeation chromatography is widely used as an approximate Ovalbumin 43,000 0.54
means of finding the molecular weight of globular proteins. It is Catalase 62,000 0.38
frequently found that a graph of K4 vs. log(molecular weight) is rea- Bovine serum albumin 68,000 0.40
sonably linear. To find K4, it is necessary to know Yo, v., and J-j. The Unknown ? 0.68
quantity Yo is measured with a very large molecule (such as the dye
blue dextran) which is completely excluded from the gel. To find J-j, V,
is calculated from the column dimensions and V, is usually assumed (b) Estimate the relative mobility expected for the dimer of the un-
to be negligible. A column of Sephadex G- 75 with a diameter of known protein. The dimer can be prepared by chemically cross-
2.5 em and a length of 36 em gave the following results: linking two molecules.
10. A cation exchange resin tends to exclude anions and an anion ex-
change resin tends to exclude cations. As a consequence, the concen-
tration of electrolyte inside of the resin is lower than the concentration
outside of the resin. Consider a cation exchange resin in the K + form
53.1 immersed in a solution of KCI. Let [K +]; and [CI-J be the concen-
Blue dextran 2,000,000
trations of species inside the resin and let [K +]. and [CI-]. be the
Aldolase 158,000 55.0
71.8 concentrations outside of the resin. From thermodynamic considera-
Ovalbumin 45,000
tions it can be shown that the condition which must be satisfied at
Chymotrypsinogen A 25,000 90.3
Ribonuclease A 13,700 109.5 equilibrium is

From these data, estimate the molecular weight of a protein with


V. = 76.5 ml.
9. The most common way to estimate the molecular weight of a protein
is by electrophoresis in a polyacrylamide gel containing sodium dodecyl
sulfate (SOS). The latter reagent coats polypeptide chains and de-
natures proteins such that they adopt a random coil conformation. where [R -] is the concentration of anionic sites in the resin.
The negatively charged SOS also imparts a nearly constant charge-to- (a) Derive an expression relating [CI-]. to [CI-]i'
mass ratio to the coated protein. As a result, the electrophoretic mobil- (b) If the concentration of anionic sites in the resin is 10M and
ity in this medium depends mainly on the molecular weight of the [CI-J = O.OOIM,calculate [CI"] •. Find [CI-]. if [CI-J = O.lM.
protein. Proteins with subunits which are held together by noncova- (c) Does the resin exclude electrolyte more effectively at high or low
lent bonds break apart into their subunits in SOS. Proteins with electrolyte concentration?
subunits held together by disulfide bonds may be cleaved into sub- 11. Use the principle of Problem 10 to explain the mechanism of ion
units by treatment with the reducing agent, 2-mercaptoethanol. The exclusion chromatography. In this technique an electrolyte (such as
relative electrophoretic mobility of a protein is measured by compar-
ing its distance of migration to that of a low molecular weight ionic
NaCI) can be separated from a nonelectrolyte (such as glycerol) by
passage through an ion exchange column. Hint: Consider the analogy
to gel permeation chromatography.
12. Suggest a procedure for measuring the exchange capacity (ion ex-
change sites per unit mass or volume of the resin) of a cation ex-
change resin using standard solutions of HQ and NaOH and any
other reagents you wish.
In gas chromatogi-aphy, the components of a PRINCIPLES Relationship Between V, and K
vaporized sample are fractionatcd as a con- OF GAs-LIQUID is then given by the expression It is of interest to relate V, to the partition
sequenceof partition between a mobile gaseous CHROMATOGRAPHY T P- PH,o. coefficient K. To do so, we substitute the ex-
phase and a stationary phase held in a column. F= F,. x Tc x P (26-1) pression relating tll and tM to k' (Equation
Gas-~olid chromatography employs a solid sta- In principle, gas-liquid and liquid-liquid par_ 24-14 p. 677) into 26-4, which gives
tionary phase; the partition process, then, in- tition chromatography differ only in that the where Tc is the column temperature in degrees
volves gaseous adsorption equilibria. In gas- mobile phase of the former is a carrier gas K, T is the temperature of the gas at the V. =jFtMk' x 273 = J1,k' x 273
liquid chromatography (GLC1 the stationary rather than a liquid. The sample is introduced bubblemeter,P is the pressure at the end of the , W Tc W Tc
phase is a liquid, supported on an inert solid as a vapor at the head of the column, where- column, and PH20 is the vapor pressure of
matrix; here, gas-liquid equilibria are impor- upon those components having a finite solu-
Substituting Equation 24-12 yields (here, J1,
water. The quantities P and T are normally and VM are identical) -
tant. bility in the stationary liquid phase distribute the pressure and temperature of the room.
Gas-solid chromatography has had some- themselves between this phase and the gas The retention volume VII for a species is then V. = Kl's x 27~ (26-5)
what limited application owing to tailing of according to the equilibrium law. Elution is given by , W Tc
elution peaks (a direct result of the nonlinear accomplished by forcing an inert carrier gas, VII=tIlF
character of adsorption isotherms) and semi- such as nitrogen or helium, through the col- The density Ps of the liquid o~ the stationary
permanent retention of active gases on the umn. The rates at which the various com- where tll is the retention time for the species phase is given by
stationary phase. For a decade or more, gas- ponents move along the column depends (p. 673). For a component that is not retained
W
liquid chromatography has been the most im- upon their tendency to dissolve in the station- on the column (usually air in gas chromatog- Ps=-
raphy1 we may also write l's
portant and widely used of all of the column ary liquid phase. A distribution coefficient
chromatographic methods. Recently, how- favoring the immobilized liquid results in a VM= tMF
ever, HPLC has been growing rapidly in im- low rate; on the other hand, those compo- K 273i
portance and now rivals gas-liquidchromatog- nents whose solubility in the liquid phase is where tM is the time required for it to pass
V, = Ps x 1(;; (26-6)
raphy in numbers of applications. negligible move rapidly through the column. through the column.
The concept of gas-liquid chromatography Ideally, bell-shaped (Gaussian) elution curves Both VII and VM depend upon the average
Note that V, at 273°K depends only upon the
was first described in 1941 by Martin and are obtained. pressurewithin the column-a quantity that lies
partition coefficientof the solute and the den-
Synge, who were also responsible for liquid- The various relationships that were devel- between the inlet pressure Pi and the outlet
sity of the liquid making up the stationary
partition chromatography. More than a decade oped in Chapter 24 and are summarized in pressure P (atmospheric pressure). Corrected
phase. As such, it should be a useful param-
was to elapse, however,before the value of this Table 24-2 are all applicable to gas chroma- retention volumes J1 and ~, which corre-
eter for identifyingspecies.
method was demonstratcd experimentally.! tography and can be used to calculate column spond to volumes at the average column pres-
sure are obtained from the relationships . The literature contains large numbers of
Since that time, the growth in applications efficiency,resolution, and elution time. specificretention volumes; unfortunately these
of the procedure has been phenomenaI.2 J1 = jtllF and J1, = jtMF (26-2) data are widelyscattered and often unreliable.
The quantity j is readily derived from the
Specific Retention Volume relationship
The specific retention volume V, for a speciesis 3[(p,;p)2 - 1]
a parameter that, unlike retention time, is inde- 2[(p,;p)3 - 1] Numerous instruments, varyillg in sophistica-
pendent of many of the column variables but is The specific retention volume is then defined tion and ranging in price fromseveral hundreds
I A. J. James and A. J. P. Martin, Analyst, 77, 915 (1952);
Biochem. J, SO,679 (1952~ dependent upon the nature of the solute and to several thousands of dolla~, are available
as
2 For detailed discussions or GLe, see: Modern Practia the stationary phase. In addition, V, is readily for gas chromatography. The basic compo-
of Gas Chromatography, ed. R. L. Grob. New York: derived from easily obtained experimental v.=J1-~ x-= 273 jF(tll - tM) 273 nents of these instruments are shown in Figure
x-
Wiley·lnterJcience, 1977; A. B. Uttlcwnod, Gas Chroma- data. Thus, V, is useful for qualitative analysis. , W 1(; W - Tc 26-1. A description of each component follows.
tography, 2d ed. New York: Academic Press, 1970; J. Q.
Walker, M. T. Jacbon, Jr, and J. B. Maynard, Chroma-
To obtain the specificretention volume for (26-4)
tographk: Systmu, 2d ed. New York: Academic Press, a compound, the average flow rate F within Carrier Gas Supply
1977, Part II; and H. M. McNair and E}- Bonelli, Bask: the column must be known. The flow rate at where W is the weight of the mobile phase, a
Gas Chromatography. Walnut Creek, CA: Varian Acro- the end of the column F,. is readily obtained quantity determined at the time of column Carrier gases, which must be chemicallyinert,
graph, 1971. by means of a soap-bubble meter (p. 719); !' preparation. includehelium,argon, nitrogen, and hydrogen;
oxide, or a silicate. The added surface area
increases the amount of liquid retained by the
tube, and thus the capacity of the column. .
Packed columns are fabricated from glass
or metal tubes 1 to 8 mm in inside diameter;
they are designed to hold solid packings that
range from 2 to 20 m in length. The tubes
ordinarily are folded or coiled so that they
can be conveniently fitted into a thermostat.
Typically, columns have Vs fVM ratios of IS to
Gas
from ---;> 20 and contain 100 to 1000 theoretical plates
column per foot. The best packed columns have a
total of 20,000 or more theoretical plates.
Rubber bulb
Carrier Solid Support for Packed Colwol& The ideal
gas containing
supplV
soap solution solid support would consist of smaIl, uniform,
spherical particles with good mechanical
FIGURE 26-2 Soap- strength and a specific surface area of at least
bubble meter. I m2/g. In addition, the material should be
inert at elevated temperatures and be readily
wetted by the liquid phase to give a uniform
coating. No substance that meets all of these
Gas samples are best introduced by means criteria perfectly is yet available.
of a sampling valve such as that shown in The most common supports are derived
Figure 24-9 (p. 684). Solid samples are in- from diatomaceous earths. Two types are avail-
troduced as solutions, or alternatively, are able. Firebrick. which is sold under trade
helium is the most widely used. AIl of these this film to move between two graduations on scaled into thin-waIled vials that can be in- names such as Chromosorb P, C 22, and
gases are available commerciaIly in pressurized the buret is measured and converted to a flow serted at the head of the column and crushed Sterchamo~ has the better strength and the
tanks. Associated with the gas supply are rate. from the outside. larger specific surface area (- 4 m2/g); its dis-
pressure regulators, gauges, and flowmeters. advantage lies in the fact that it is more active
In addition, the carrier gas system often con- and, therefore, cannot be employed with polar
Sample Injection System
tains a molecular sieve to remove water or compounds. Kieselguhr is more fragile and has
other impurities. Column efficiency requires that the sample be Two types of columns are employed in gas- a smaIler specific surface area (-1 m2/g) but
Flow rates are controIled by a pressure of suitable size and be introduced as a ••plug" liquid chromatography. One type is fabricated is less reactive; it is sold under such trade
regulator. Inlet pressures usuaIly range from of vapor; slow injection or oversized samples from capillary tubing (03 to 0.5 mm i.d.), the names as Chromosorb W. Celite, Embacel,
10 to SO psi (above room pressure1 which cause band spreading and poor resolution. A bore of which is coated with a very thin film and Celatom.
lead to flow rates of 2S to SO ml/min. Gen- microsyringe is used to inject liquid samples (-1pm)oftheliquid phase. Capillary columns liquid Phase. Desirable properties for the
eraIly; it is assumed that the flow rate will be through a rubber or silicone diaphragm into have a negligible pressure drop and can thus immobilized. liquid phase in a gas-liquid chro-
constant if the inlet pressure remtins constant. a heated sample port located at the head of be of great length (10 to 100 m or more); the matographic column include: (1) low volatility;
Flow rates can be established by a rotameter the column (the sample port is ordinarily ratio of Vs/VM (Equation 24-12) for these ideaIly, the boiling point of the liquid should
at the column head; this device. however, is about SO·C above the boiling point of the columns ranges from 100 to 300, which in part be at least 200· higher than the maximum
not so accurate as a simple soap-bubble meter, analyte). For ordinary analytical columns, accounts for their high efficiencies. Columns operating temperature for the column; (2)
which is located at the end of the column. sample sizes vary from a few tenths to 10 of several hundred thousand theoretical plates thermal stability; (3) chemical inertness; and
Figure 26-2 is a schematic diagram of a pliter. Capillary columns require much smaIler have been described. These columns, however, (4) solvent characteristics such that cc and k'
soap-bubble flow meter. A soap film is formed samples (_10-3 pliter);here, a sample sp.litter have very low sample capacities ( < 0.0 I pliter). (p. 682) values for the solutes to be resolved
in the path of the gas when a ru~bcr bulb system is employed to deliver only a smaIl The capacity of capillary columns can be in- faIl within a suitable range.
containing an aqueous solution of soap or fraction of the injected sample to the column creased by coating the inside of the tube with No single liquid meets all of these require-
detergent is squeezed; the time required for head, the remainder going to waste. a porous material such as graphite, a metal ments, the last in particular. As a consequence,
it is common practice to have available several in elution time and therefore the time required
interchangeable columns, each with a different to complete an analysis. Figure 26-3 illustrates
stationary phase. The choice among these is· this principle.
often a matter of trial and error, although
some qualitative guidelines exist; factors affect-
Detection Systems
ing this choice are considered in the section
devoted to applications. Detection devices for a gas-liquid chromato-
Column Preparation. The support material graph must respond rapidly and reproducibly
is first screened to limit the particle size range. to the low concentrations of solutes emitted
It is then made into a slurry with a volatile from the column. The solute concentration in
solvent that contains an amount of the station- a carrier gas at any instant is only a few parts
ary liquid calculated to produce a thin coating per thousand at most; frequently, the detector
(5 to 10 JlIIl) on all of the particles. After the is called upon to respond to concentrations
solvent has been evaporated, the particles that are smaller than this by one or two orders
appear dry and are free-flowing. of magnitude (or more). In addition, the inter-
Columns are fabricated from glass, stainless val during which a peak passes the detector is
Relative time, min
stee~ copper, or aluminum. They are filled by usually one second or less; therefore, the
slowly pouring the coated support into the detector must be capable of exhibiting its full FIGURE 26-3 Effect of temperature on the separation of
straight tube with gentle tapping or shaking to response during this brief period. . hexane isomers: A, 2,2-dimethylbutane; B, 2,3-dimethylbutane;
provide a uniform packing. Care must be taken Other desirable properties of the detector : C, 2-methylpentane; D, 3-methylpentane; E, n-hexane. (From
to avoid channeling. After the column has include linear response, good stability over ex- ~C. E. Bennett, S. Dal Nogare, and L. W. Safransk~ in Treatise
been packed, it is bent or coiled into an appro- tended periods, and uniform response for a ! on Analytical Chemistry, eds. I. M. Kolthoff and P. J. Elving,

priate shape to fit the oven (see Figure 26-1). wide variety of compounds. No single detec- . Part I, vol 3. New York: Wiley-Interscience, 1961, p. 1690. With
A properly prepared column may be em- tor meets all of these requirements, although permission.)
ployed for several hundred analyses. Numer- more than a dozen different types have been
ous types of packed columns are available proposed. We shall describe the most widely
commercially. used of these. injection chamber and the other immediately tive when these substances are used as carrier
Column ThermcJstating. Column tempera- Thermal CoaductiYity Detectors. A relative- beyond the column. In this way, the thermal gases.
ture is an important variable that must be ly simple and broadly applicable detection conductivity of the carrier gas is canceled, Thermal conductivity detectors are simple,
controlled to a few tenths of a degree for system is based upon changes in the thermal and the effects of variation in flow rate, pres- rugged, inexpensive, nonselective, accurate, and
precise work. Thus, the column is ordinarily conductivity of the gas stream; an instrument sure, and electrical power are minimized. The nondestructive of the sample. They are not so
housed in a thermostated oven. The optimum employed for this purpose is sometimes called resistances of the twin detectors are usually sensitive, however, as some of the other devices
column temperature depends upon the boil- a katharometer. The sensing element of this compared by incorporating them into two to be described.
ing point of the sample and the degree of sep- device is an electrically heated source whose arms of a simple Wheatstone bridge circuit Flame Ionization Detectors. Most organic
aration required. Roughly, a temperature equal temperature at constant electrical power de- such as that shown in Figure 26-4. compounds, when pyrolyzed at the tempera-
to or slightly above the average boiling point pends upon the thermal conductivity of the The thermal conductivities of hydrogen and ture of a hydrogen/air flame, produce ionic
of a sample results in an elution period of surrounding gas. The heated element may con- helium are roughly six fo ten times greater than intermediates that provide a mechanism by
reasonable time (2 to 30 min). For samples sist of a fine platinum or tungsten wire or, those of most organic compounds. Thus, the which electricity can be carried through the
with a broad boiling range, it is often desirable alternatively, a semiconducting thermistor. The presence of ev~n small amounts of organic flame. By employing an apparatus such as
to employ temperature programming, whereby resistance of the wire or thermistor gives a materials causeS a relatively large. decrease in that shown in Figure 26-5, these ions can be
the column temperature is increased either measure of the thermal conductivity Qfthe gas; the thermal conductivity of the column efflu- collected and the resulting ion current mea-
continuously or in steps as the separation in contrast to the wire detector, the thermistor ent; the detector undergoes a marked rise in sured. The electrical resistance of a flame is
proceeds. has a negative temperature coefficient. temperature as a result. The conductivities of very high (perhaps 1012 0), and the resulting
In genera~ optimum resolution is asso- In chromatographic applications, a double nitrogen and carbon dioxide more closely re- currents are therefore minuscule; an electrom-
ciated with minimal temperature; the cost of detector is always employed, one element being semble those of organic constituents; thus, eter must be employed for their measurement.
lowered temperature, however, is an increase placed in the gas stream ahead of the sample detection by thermal conductivity is less sensi- The ionization of carbon compounds in a
X-radiation (p. 440). Here, the effluent from Exhaust
the column is passed over a p-emitter silch as
nickel-63or tritium (adsorbed on platinum or t
titanium foil). An electron from the emitter
causes ionization of the carrier gas (often
nitrogen) and the production of a burst of +1111-
electrons. In the absence of organic species.a
constant standing current between a pair of
electrodes results from this ionization process. Amplifier
Ind
The current decreases,however,in the presenCe recorder
of organic molecules that tend to capture elec-
trons. The response is nonlinear unless the
potential across the detector is pulsed.
The electron-capture detector is selectivein
its response and is highly sensitivetoward elec-
tronegative functional groups such as halogens,
peroxides, quinones, and nitro groups. It is
insensitive to compounds such as amines,
alcohols, and hydrocarbons. An important
application of the election-capture detector
has been for the detection and determination
Amplifier
Ind of chlorinated insecticides.
recorder Electron-capture detectors are highly sen-
sitive and possess the advantage of not alter-
FIGURE 26-4 Schematicdiagram of ather- ing the sample significantly (in contrast to the
mal conductivity detector-recorder system. flame detector).

As a consequence, an important trend in the useful in gas-liquid chrC)llllllOgaphy the im-


APPLICATIONS OF field appears to be in the direction of combin- mobilized liquid must ~ ditTen:~tparti-
GAB-LIQUID CHROMATOGRAPHY ing the remarkable fractionation qualities of tion coefficients among Sou~; addl~lonally,
flameis a poorly understood process, although gas-liquid chromatography with the superior however, these coefficients :n:;..•
~t be neIther ex-
it is known that the number of ions produced In evaluating the importance of gas-liquid analytical properties of such instruments as tremely small nor extr~ large. Solutes
is roughly proportional to the number of re- chromatography, it is necessary to distinguish mass, ultraviolet, infrared, and NMR spec- with small coefficientspass C1rQ!1gh t.hecolumn
duced carbon atoms in the flame. Functional between the two roles the method plays in its trometers. so rapidly that no significant' ~ratto? occurs.
groups such as carbony~ alcohol, and amine application to chemical problems. The first On the other hand, when ~ coeffiCIentsare
produce fewer ions or none at all. is as a tool for performing separations; in large, the time required to r~e solutes from
Choice of the the columns becomes inorl't!t::Zte·
The hydrogen flame detector currently is this capacity, it is unsurpassed when applied
Stationary Liquid Pha.e To have a reasonable r~~ce time in the
one of the most popular and most sensitive to complex organic, metal-organic, and bio-
detectors. It is more- complicated and more chemical systems. The second, and distinctly Several hundred liquids have been suggested column, a solute must shr.... at least some
expensivethan the thermal conductivity detec- different function, is that of providing the for use as the stationary phase in gas-liquid degree of compatability Is..-A;.:?iljt~~with the
tor, but has the advantage of higher sensitivity. means for completion of an analysis. Here, chromatography. The successfulseparation of stationary phase. Thus, me polaritIes of the
In addition, it has a wide range of linear retention times or volumes are employed for closelyrelated compounds isoften critically de- two substances should be •. least somewhat
reponse. It is, of course, destructive of the qualitative identification, while peak heights pendent upon the proper choice among these. alike. For example, a statj~' .liquid such as
sample. or areas provide quantitative information. In The retention time for a solute depends squalane (a high molecular" ..eight satura~
Electroa-Capture Detectors.\ Electron-cap- its analytical role, gas-liquid chromato,graphy directly upon its partition coefficient (Equa- hydrocarbon) might be ch~ for separat~n
ture detectors operate in much the same way is considerably more limited than some ofthe tion 24-13~ which, in turn, is related to the of members of a nonpolar 'V'.t.::'Jologoussenes
as a proportional counter for me,asurementof other methods considered in earlier chapters. nature of the stationary phase. Clearly, to be such as hydrocarbons or ~:~. On the other
hand, a more polar liquid such as polyethylene hand, is moderately polar and capable of in- squalene.3 The resulting data pro..vide a quan- raphy, provides the scientist with a powerful
glycol would probably be more effective for ducing polarization in the olefins; seleCtive titative means for classifying stationary media method for separating complex mixtures. It is
separating alcohols or amines. For aromatic and predicting the performance of individual much less useful, however, as a tool for iden-
retention results. Note also that the retention
hydrocarbons, benzyldiphenyl might prove times for the nonpolarizable paraffins are sig- packings for various types of separations. Un- tifying separated components.
appropriate. .. nificantly less in the polar solvent than in the fortunately, the effort required to obtain the Retention Tune or Retention Volume. A sin-
Among solutes of similar polarity, the elu- nonpolar one. The selectivity of adduct forma- required data is usually greater than that neces- gle retention-time measurement is not a very
tion order usually follows the order of boiling tion between olefins and silver nitrate permits sary to determine empirically which of several satisfactory parameter upon which to base a
points; where. these differ sufficiently, clean some useful separation as well; for example, columns will bring about the required separa- qualitative analysis because of its dependence
separations are feasible. Solutes with nearly isobutene and butene-l are not separated on tion. Thus, most chemists have several columns on such variables as column temperature, flow
identical boiling points but different polarities the first two columns but are resolved in the available. and, by trial and error, determine rate, pressure, and composition of the station-
frequently require a liquid phase that will selec- presence ofsilver nitrate; improved separation which is the best for solving a particular ary phase. To be sure, if the retention times
tively retain one (or more) of the components of cis- andtrans-butene-2 also occurs. problem. for an unknown and a standard are found to
by dipole interaction or by adduct formation. Another important interaction that often Table 26-2 lists a few of the most widely vary similarly as these operating conditions
The data in Table 26-1 illustrate some of these enhances selectivity is hydrogen bond forma- used stationary liquid phases. are varied, the probability of identity of the two
effects. Here, the retention times (relative to tion. For this effect to operate. the solute must is enhanced. Thus, if similar retention times for
ethane) for a series of paraffinic and olefinic have a polar hydrogen atom and the solvent the standard and analyte are observed on two
Qualitative Analyais
hydrocarbons are compared for three liquid must have an electronegative group (oxygen, or three columns at two or three temperatures,
phases. The first, triisobutylene. is nonpolar; fluorine, or nitrogen)--or the converse must Gas chromatography, like liquid chromatog- it is ordinarily reasonable to assume that the
the other two are polar. Thl} third solvent, in be true. analyte and standard are the same. Usually,
addition, contains silver nitrate. which selec- Several: systems have been developed to however, the separated component is more
tively forms loose adducts with specific olefins. optimize tpe choice of a stationary phase for • w. R. Supina and L. P. Rose, J. Chromatog. Sci. 8, 214 easily identified by spectroscopic measure-
Note that the retention times on triisobutylene a separati~n within any given group of com- (1970~ ments than through several chromatograms.
correlate closely with the boiling points of the pounds. One characterizes stationary phases
liquids and that little or no separation is by comparing retention indices (p. 726) for
realized among the compounds with similar benzene, ethanol, methylethyl ketone, nitro-
volatilities. Acetonyl acetone, on the other methane, and pyridine relative to those on
TABLE 26-2 SOME COMMON STATIONARY PHASES
MaximlBll
Name Chemical Composition Temperature, °C Polarity" Type of Separation

Squalene C.oH6z 150 NP Hydrocarbons


TABLE 26-1 SEPARATION OF HYDROCARBONS OV-l Polymethyl siloxane 3SO NP General purpose
WITH VARIOUS STATIONARY LIQUIDS. nonpolar
DC 710 Polymethylphenyl 300 NP Aromatics
Relative Retention Time On siloxane
BoiIi Triiso- Acetonyl QF-l Polytriftuoropropyl 250 P Amino acids,
AaN03 in methyl siloxane steroids,
Compound Point':'ic butylene Acetone Glycol nitrogen compounds
Ethane -104 1.0 1.0 1.0 XE-JO Polycyanomethyl 275 P Alkaloids,
lsobutane -11.7 5.6 2.2 i siloxane halogenated
0.75 compounds
Isobutene -6.9 6.7 4.75 ' 3.25
Butene-l -6.3 6.7 4.75 6.25 Carbowax 20M Polyethylene glycol 250 P Alcohols, esters,
Buautiene -4.4 essential oils
6.7 10.0 10.0
n-Butane -0.5 7.5 3.0 0.75 DEG adipate Diethylene glycol 200 SP Fatty acids, esters
trans-Butene-2 0.88 8.0 5.85 1.75 adipate
cis-Butene-2 3.72 Dinonyl phthalate ISO SP Ketones, ethers,
8.9 6.8 5.5 sulfur compounds

• ~ta rr~ ~.w. Bradford, D. Harvey and D. E. Chalkley, J. Insr. Petrokum, 41 80 (19SS\
.•With penDlSSlon. . .
:.,.
, ,.
i • NP = nonpolar; SP = semipolar; P = polar.

J
As was noted earlier, the specific retention index for a series of n-paraffins will be a measure of the relative polarity of different 1000 Cycloalkanes
volume is a better parameter upon which to straight line. stationary phases.5 Esters
base a qualitative analysis than is the reten- The retention index for species X is then Retention TlDle Plots. The linear relation- Aldehydes
tion time, because it depends only upon the defined as
:s 100
ship between log t.and the n~ber of car~n
partition coefficient of the analyte, the tem- atoms for members of a homologous senes
,gE
perature of the column, and the density of the Ix = 100 [ log(t.)x - 10g(t'.).} + lOOn (p. 726) can be useful for qualitative work, c'
'E 10
stationary phase (Equation 26-6~ log(t'.).+ I - 10g(tR). since it is often possible to separate one homol- II'

Relative Retention time. The selectivity fac- E


ogous series from another by chromatography .;:;

tor (X is a measure of relative retention time. (26-7) c


(if the two have different polarities) or by other 0;
0

As shown in Chapter 2~, (X is defined as physical or chemical means W


1.0
where (t'.) is the corrected retention time for
(t.)y - tM each species; that is An even more useful relationship can some-
(X=---- times be obtained by injecting portions of the
(t.)x - tM 0.1
sample into two columns with stationary 0.1
Thus, by introducing an internal standard Y phases that differ in polarity. For a homol- Elution time. min, from tricresyl phosphate
into a sample containing the anaIyte X, the ogous series a plot of the retention time on one
In Equation 26-7, the subscript X refers to FIGURE 26-6 Behavior of several homol-
readily measured analytical parameter (X is ob- column versus that on the other (employing a
the analyte and subscripts nand (n + 1) refer ogous series on two columns. Points from left
tained. For a given column, this parameter log scale) yields a straight line (see Figure 26-6);
to normal paraffins containing n and (n + 1) to right correspond to the following:
should be independent of the column variables, more important, the intercepts of the plots
carbon atoms, respectively. Generally, the hy- n-alkanes: butane, pentane. hexane, heptane
provided that X and Yare chemically similar. differ from one series to another. Thus, resolu-
drocarbons are chosen so that their retention cycloalkanes: cyclopropane, cyclopentane,
Clearly, the peaks for X and Y must not overlap tion of species with similar boiling points but
times bracket that of the analyte. cyclohexane
and the two species must be chemically com- differing polar characteristics frequently can
patible with each other and the mobile phase. esters: methyl-, n-propyl-, n-butyl- acetates
EXAMPLE be achieved by obtaining chromatograms for
The selectivity factor has the advantage of aldehydes: ethanal, n-propanal, n-butanol
From the following data, calculate the reten- the sample on two columns. In addition, the
requiring only three simple measurements ketones: acetone, 2-butanone, 2-pentanone
tion index for benzene on a squalene column peaks may be identified with considerably less
(t.)x, (t.)y, and tM. alcohols: methanol, ethanol, n-propanol, n-
operated at l00·C. ambiguity. This technique is quite analogous
11te Retention Index. The retention index butanol
to two-dimensional paper chromatography.
[Reprinted with permission from J. S. Lewis,
(I) waS proposed by Kovats4 as a qualitative Summary. A single gas chromatogram
Substance (t. - tM), mia / H. W. Patton and W. L. Kaye, Anal. Chem.,
parameter for general use in reporting chro- does not provide sufficient data to permit the 28, 1370 (1956). Copyright by the American
matographicdata. The retention index is based unambiguous identification of a compound.
n-hexane 3.43 600 Chemical Society.]
upon a comparison between the position of an benzene On the other hand, identical retention index
4.72 ?
analyte peak and the peaks for two or more n-heptane data on three columns of differing polarity
6.96 700
normal paraffins. By definition, the retention would provide a good basis for assuming a
index for a normal paraffin is equal to 100 times standard and an analyte were the same species.
the number of carbon atoms in the compound, Quantitative Analysis
Gas chromatograms are widely used as cri-
regardless of the columns used or the chroma- teria of purity for volatile compounds. Con- The detector signal from a gas-liquid chroma-
tographic conditions. Thus, I for n-pentane is 1= 100 [lOg 4.72 -log 3.43} + 100 x 6 taminants, if present, are revealed by the ap- tographic column has had wide use for quan- .
always 500; n-oetane it is 800. log 6.96 - log 3.43 . pearance of additional peaks; the areas under titative and semiquantitative analyses. Under
It has been observed that, within a homol- these peaks provide rough estimates of the carefully controlled conditions, an a~uracy of
ogous series, a plot relating the logarithm of = 45.1 + 600 = 645 extent of contamination. The technique is also 1 to 3 % relative is attainable. As With most
retention time (or volume) to the number of useful for evaluating the effectiveness of puri- analytical tools, reliability is directly related
carbon atoms is linear, provided the lowest The retention index system has the advan- fication processes. to the amount of effort spept in calibration
member of the series is excluded. As a conse- tage of having readily available reference ma- and control of variables; the nature of the
quence, a plot of (t. - tM) versus the retention terials that cover a wide boiling range. In addi- sample also plays a part in determining the
tion, the temperature dependence of retention potential accuracy. . .
indices is relatively small. Furthermore, the The general discussion of quantitatIVe chro-
• For details. see: W. R. Supina and L. P. Rose. J. Chr~ matographic analysis given in Chapter 24 ap-
change in retention index AI between a polar _ag. Sci, 8, 214 (1970); and W. O. McReynolds. J.
• E. Kovats, He/v. Chim. Acta, 41. 1915 (19S8~ plies to gas chromatography as well as to other
... ..',
and a nonpolar stationary phase provides a Chromatag. Sci, 8, 68S (1970~
~
types; therefore, no further considera~ion of procedure has many ofthe advantages ofgra- while peaks for the highest molecular weight other spectroscopic techniques. The main limi-
this topic is needed here. dient elution in liquid chromatography. It iI components have not appeared even after tation to this approach lies in the very small
used to simplify, improve, and accelerate sepa- 95 min of elution. Furthermore, the shapes of (usually micromolar) quantities of solute con-
rations. the higher molecular weight. peaks that do tained in a fraction; nonetheless, the general
Programmed
Figure 26-7 illustrates the effectiveness of appear are not very satilfactory for quantita- procedure has proved useful in the qualitative
Tampereture Gas Chromatography
the programmed temperature approach to the tive analysis. analyses of components of many complex mix-
In programmed temperature gas chromatog- separation and determination of a mixture of In contrast, the programmed temperature tures.
raphy, the column temperature is increased normal paraffins. With the isothermal method, chromatogram provides well-spaced peaks A second general method involves the use
(usualJy linearly) as elution proceeds. Such a the low molecular weight peaks are bunched, with heights that are readily measured. of a selective detector to monitor the column
Programmed temperature work requires a effluent continuously. Some examples will illus-
device for varying the temperature linearly and trate this approach.
reproducibly at a rate varying from about 0.25 Automatic Titration Detectors. In their first
Co
C CIl la) Isothermal to 20°C per minute. Furthermore, the proce- publication on gas chromatography, James
C'O C'2 C'3
dure requires separate heaters for the injection and Martin employed an automatic titration
port and the detector in order to keep these cell as a detector for volatile fatty acids; the
components at constant temperature while the same device was later applied to mixtures of
Co
column temperature is being varied. In order aromatic and aliphatic amines. In this method,
C,. the effluent from the column was led directly
to achieve a constant flow rate, a flow con-
troller is needed because a temperature rise into a cell containing an acid-base indicator
causes an increase in gas viscosity and column solution. The absorbance of the solution was
resistance; both of these variables tend to monitored photometrically, the output from
Inject

I
0
I
5
I
10
I
15
I
20
I
25 30
I \(1
45
I
50
I
55
Ai
II I
90
I
95
reduce the flow. The result of changes in flow
rate is an unstable base line and changes in
detector response.
the cell being employed to control continu-
ously the addition of titrant; the volume of the
latter was recorded on a chart to produce an
integral curve of volume versus time. Both
Minutes
qualitative and quantitative information could
Gas-Liquid
be retrieved from the recorded curve.
Chromatography
C,O CIl (b) Temperature programmed A coulometric cell has been employed as the
C'2 C'3 C14 with Selective Detectors
detector in the analysis of the various chlori-
Most of the detection systems that we have nated compounds contained in pesticide resi-
considered thus far are nonselective in the sense dues.
that they respond in more or less the same way Coupling of Gas Chromatography to M ••
C'5 C
16 C to all of the solutes emerging from a column Spectroscopy.6 Chromatographic columns
n
(ideally, the nonselective detector would re- have been directly interfaced to rapid-scan
spond identically to every compound). The mass spectrometers, thus permitting the in-
employment of selective or specific detectors, stantaneous display of the spectrum of each
however, makes it possible to combine the tre- species as it leaves the column. Generally,
I I
mendous separatory power of gas-liquid chro- these instruments are also interfaced with a
32 i 36 computer so that each spectrum is digitalized
matography with the superior analytical quali-
ties of the instruments we have considered in
FIGURE 26-7. Comparison of (a) isothermal and (b) temperature- earlier chapters.
programmed chromatograms for some normal paraffins. Columns: Two general approaches are possible. In
20 ft by 1/16 in with 3 % Apiezon on 100-120 mesh VarAport 30. the first, the solute vapors are collected as
• For additional information, see: F. W. Karasek, AIIDl.
Flow rate: 10 ml/m~ with He. Temperature: (a) lSOoC and (b) SO separate fractions in a cold trap, a nondes- Chern.44(4~ 32A (t972);and C. Fcnselau, Anal. Chern. 49,
to 2SOoC at 8°/min.· (From H. M. McNair and E. J. Bonelli, Basic tructive and nonselective detector being em- S63A (t 977). For information on liquid chromatography I
Gas Chromatography. Walnut Creek, CA: Varian Aerograph, 1969, ployed to indicate their appearance. The frac- mass spectroscopy. see: W. H. McFadden, D. C. Bradford,
p. 190. With permission.)' , tions are then identified by NMR, IR, mass, or D. E. Ganes, and J. L. Gower, Amer. Lab. 9 (tO~ SS (t977~
gaseous effluents from a column with library Porous Polymers
spectra that have been obtained with liquid
Porous polymer beads of uniform size are
or solid samples. Gaseous spectra contain
manufactured from styrene cross-linked with
rotational fine structure while liquid or solid
divinylbenzene. These materials are sold under
Into spectra do not; significant differences in ap-
ion the trade name of Porapak.8 The pore size of
source pearance are' the result.
these materials is uniform and is controlled by
the method of manufacture. The application of
GAS-SOLID Porapaks to gas-solid chromatography is anal-
CHROMATOGRAPHY ogous to the application of Sephadex beads in
EXit/
constriction
Porous gel-permeation chromatography (p. 703).
glass
tube Gas-solid chromatography is based upon ad- The Porapaks have found considerable use
sorption of gaseous substances on solid sur- in the separation of gaseous polar species such
FIG U R E 26-8 Pressure reduction system, the link between gas chromatograph and masS spec-
faces. Distribution coefficients are generally as hydrogen sulfide, oxides of nitrogen, water,
trometer. [Reprinted with permission from J. T. Watson and K. Biemann, Anal. Chern., 37, 844 (1965).
much larger than those for gas-liquid chro- carbon dioxide, methanoL and vinyl chloride.
. Copyright by the American Chemical Society.]
matography. Thus, gas-solid chromatography
is useful for the separation of species that are
not retained in a gas-liquid column, such as EXAMPLES OF
the components of air, hydrogen sulfide, car- APPLICATIONS OF
and stored for later reproduction on paper. eluted sample; these are then led directly into
bon disulfide, nitrogen. oxides, carbon mon- GAS; CHROMATOGRAPHY
Instruments of this type make possible the the ion source of the mass spectrometer. Both
oxide, carbon dioxide, and the rare gases.
identification of the hundreds of components quadrupole and small double-focusing spec-
The most common adsorbents are molecu- Figurb 26-9 illustrates some typical applica-
that may be present in natural and biological trometers have been employed for analysis of
lar sieves and certain porous polymers. tions ,of gas chromatography to analytical
systems. For example, the interfacing of chro- the resulting ionic beam.
matography with spectroscopy has permitted Infrared Spectroscopy/Chromatography.7 In- problems. All are based upon gas-liquid equi-
Molecular Sieve. libria except the one shown in Figure 26-9a.
characterization of the odor and flavor com- struments have also been developed that will
ponents of foods, identification of pollutants, provide infrared spectra of chromatigraphic Molecular sieves are zeolites of carefully con- Here, a gas-liquid column is followed by a
medical diagnosis based on breath compo- peaks. Both rapid-scan and Fourier transform trolled pore size. These pores or holes are gas-solid molecular sieve. The former retains
nents, and studies of drug metabolites. instruments have been employed. Generally, of molecular dimensions; molecules that are only the carbon dioxide and passes the re-
All major manufacturers of mass spectrom- the effluent from the column is led into a long smaller than the holes penetrate into the in- maining gases at rates corresponding to the
eters offer units for interfacing their equipment narrow tube equipped with infrared-transpar- terior of the sieve particles where they can carrier rate. When the carbon dioxide is eluted
with gas chromatography. These units range ent windows. Scanning is triggered by the out- be adsorbed. For such molecules, the surface from the first column, a switch directs the flow
from $20,000 to $125,000, with computeriza- put from a nondestructive chromatographic area is enormous when compared with the around the second column so that the gas is
tion adding 50 to 100% to these costs. peak detector and begins after a brief delay area available to molecules that cannot reach not permanently adsorbed in the molecular
A major problem in interfacing of a gas chro- the interior surfaces. Thus, sieves can be used sieve. After the carbon dioxide signal has
to allow the component to travel from the
matograph with a mass spectrometer arises to separate small molecules from larger ones. returned to zero, the flow is switched back
detector region to the infrared cell. The spectral
from the presence of the carrier gas, which Partition coefficients among small mole- through the second column, thereby permitting
data are digitalized and stored in a computer
dilutes the eluted components enormously and cules on molecular sieves differ considerably, eluti~n of the remainder of the sample compo-
from which printed spectra are ultimately de-
also tends to swamp the pumping system of thus making possible the chromatographic nents:
rived.
the spectrometer. Several methods have been separation of such species as oxygen, nitrogen, D~ta on the conditions employed for ob-
Difficulty is sometimes encountered when
developed for overcoming this problem; one is carbon monoxide, hydrogen, and methanoL taining the various chromatograms in Figure
attempts are made to compare spectra for the
shown in Figure 26-8. Here, the exit gases flow some of which are not retained to any signif- 26-9 are found in Table 16-3.
through a fritted glass tube situated in an icant extent on liquid columns. One problem
evacuated chamber. The smaller atoms or 7 For lIddilional informatioD, ICe: K. L. Kizer, Amer. with molecular sieves is that polar compounds,
molecules of the carrier gas (He or Hz) diffuse Lob. 5 (6~ 40 (1973); F. D. Mcrcaldo, Amer. Lob. 6, 63
such as carbon dioxide, may become more or
(1974); R. H. Shaps, W. Simons, and A. Varano, Amer.
readily through the walls of the t.be and are less permanently adsorbed, thus blocking the
Lob. 9 (3~ 95 (1977); and V. Rossiter, Amer. Lob. II (5~
pumped away, leaving the molecules of the 59 (1979~ surface to other species.
(a) Exhaust mixture: A, 35% H2; B, 26% C02;
C, I'll 02: D, I'll N2; E, I'll C2; F, 30% CH4;
G, 3% CO; H, I'll C3; I, I'll C4;J, I'll l-e.;
K,I%n-e •.

TABLE 26-3 CONDITIONS FOR CHROMATOGRAMS SHOWN


IN FIGURE 26-8
.Temperature, Flow,
CIIromatocram CoilUDa" PackiJla Det~ °C Carrier m1fmin

1111111111111
(a) 10' x." S Cbromosorb 102 TCD 65-200 He 30
o 4 8 12 16 20 24
5' x." S Molecular sieve SA
(b) 6' xi" G 1.5% OY-17 on ECD 22
Chromosorb G
(bl Pesticidas: A, lindane; B, Heptachlor;
C, Aldrin; D, Dieldrin; E, DOT.
(c) 6' x." S Chromosorb W TCD 190 He 30
A-D, 0.3 09; E, 3.0 09. (d) 6' xi" G 1.5% OY-17 on FID 180-230
HP Chromosorb G
(e) 6' x 3.4mm G 5% OY-210 + ECD 260 A
2.5% OY-17 on
Supelcoport
Ie) Fatty acid methyl eIterS: (f) 6' x 3.4mm 2% OY-17 on FID 105-325 N2 40
A, ethyl benz_; B, caprylate; Chromosorb W
C, laurate; D, myrlstate;
E, palmitate; F,stearate;
G, oleate; H, Iinoleate.
• S - stainlesssteel;G - glass.
• TCD - thermalconductivity;ECD- electroncapture;FID = flameionization.
(d) Sedative mixture: A, butalbital;
B, amobarbital; C, pentobarbital;
D, secobarbital; E, glutethimide;
F, phenobarbital. All at 600 09. 1. A OLC column was operated under the following conditions:
Column: 110 m x 2.0 mm packed with Chromosorb P; weight of
stationary liquid added, 1.35 g; density of liquid, 1.12 gfml
I I I I I Pressures: inlet, 25.1 psi above room; room, 748 torr
o 4 8 12 16 Time Measured outlet flow rate: 27.3 ml/min
lSO° 2120 2300 Holding temp.
196" 228" Temperature: room, 211°C; column, 102.0°C
Retention times: air, 24.0 s; methyl acetate, 2.12 min; methyl pro-
pionate, 4.32 min; methyl n-butyrate, 8.63 min
Peak widths at base: 0.21, 0.42, and 0.85 min, respectively
Calculate
(a) the average flow rate in the column.
(f) Carbohydrates: (b) the corrected retention volumes for air and the three esters.
(e) Steroids: A, DHA;
B, estraionol E2;
A, fructose; (c) the specific retention volumes for the three components.
B, dextrose;
C, estrone E1 ; C, phenyl beta (d) partition coefficients for each of the esters.
D, EPI·testospherone;
E, estriol E3.
D-glucopyranoside; (e) a retention volume and a retention time for methyl n-hexanoate.
D, sucrose;
E,lactose; 2. From the data in Problem 1, calculate
F, maltose. (a) k' for each compound.
(b) oc values for each adjacent pair of compounds.
(c) the average number of theoretical plates and plate height for the
column.
(d) the resolution for each adjacent pair of compounds.
3. For the chromatogram described in Problem 1, the areas of the three
peaks were found to be 18.1, 43.6, and 29.9 in the order of increasing
retention time. Calculate the percent of each compound if the relative
FIGURE 26-9 Some ex~ples of gas-chromatographic separations.
. 1
detector response for the three species was 0.60, 0.78, and 0.88, 10. What would be the effect of the variations listed in Problem 9 on the
respectively. . retention time for a chromatographic peak? Explain.
4. The stationary liquid in the column described in Problem 1 was 11. For a gas chromatographic column, the values of A, B, and C in the
didecylphthalate, a solvent of intermediate polarity. If a nonpolar van Deemter equation were 0.060em, 0.57 em2 S-I, and 0.13 s. Calcu-
solvent such as a silicone oil had been used instead, would the reten- late the minimum plate height and the optimum flow rate.
tion times for the three compounds be larger or smaller? Why? 12. Calculate the retention index for each of the following compounds:
5. Corrected retention times for ethyl, n-propyl, and n-butyl alcohols on
a column employing a packing coated with silicone oil are 0.69,1.51, Compound (tR - tM)
and 3.57. Predict retention times for the next two members of the
homologous series. (a) Propane 1.29
6. A chromatographic column was operated under the following condi- (b) n-Butane 2.21
tions: (c) n-Pentane 4.10
Column: 1.00 m x 2.0 mm packed with Chromosorb P; weight of (d) n-Hexane 7.61
paraffin stationary liquid, 2.10 g; density, 0.796g/ml (e) n-Heptane 14.08
Pressures: inlet, 18.3Psi above room; room, 768 torr (f) n-Qctane 25.11
Outlet flow rate: 33.6ml/min (g) Toluene 16.32
Temperatures: room, 19.6°C;column, 9OJoC (h) Butane-2 2.67
Retention time air: 0.42 min (i) n-Propanol 7.60
Retention times and· peak widths: i-propylamine, 4.59 and 0.36 (j) Methylethyl ketone 8.40
min; n-propylamin~,4.91 and 0.39 min (k) Cyclohexane 6.94
Calculate (I) n-Butanol 9.83
(a) the average flow rate in the column.
(b) the corrected retention volumes for air and the two amines. 13. Predict the corrected retention times for ethanol and n-hexanol on the
(c) the specific retention volumes for the two amines.
column in Problem 12.
(d) partition coefficientsfor the two amines.
7. From the data in Problem 6, calculate
(a) k' for each amine.
(b) an ex value.
(c) an average number of theoretical plates in the column.
(d) the plate height for the column.
(e) the resolution for the two amines.
(f) the length of column required to achieve a resolution of 1.5.
(g) the time required to achieve a resolution of 1.5, assuming the
original linear flow rate.
8. For a gas-chromatographic column, the values for A, B, and C in the
van Deemter equation were 0.15 em, 0.36 em2 S-I, and 4.3 x 10-2 s.
Calculate the minimum plate height and the best flow rate.
9. What would be the effect of the following on the plate height of a
column? Explain. ;
(a) ~creas.ing the weight of the stationary phase relative to the pack-
mg weight.
(b) decreasing the rate of sample injection.
(c) increasing the injection port temperature.
(d) inCreasingthe flow rate.
(e) reducing the particle size of the packing.
(f) d~easing the column temperature.
APPENDIX 1:
PROPAGATION OF
UNCERTAINTIES IN
PHYSICAL
MEASUREMENTS
Whenever measurements are made with an parameters; that is, useful because they permit statements to be and 1.70ppm. Calculate the standard devia-
instrument having a sensitivity that is not e-~- ••)2/1 ••2 e-.2/1 "made about the probable magnitude of the tion for the measurement.
limited by its readout device, small variations Y:;= (J.j2ic = (JJii (A-I) uncertainty for any given measurement, pro-
among replicate measurements are observed. vided the standard deviation for the output of (Xl - x)l
These variations represent the accumulation In this equation, XI represents values of indi- an instrument is known. Thus, if (J for an in- 1.67 0.003 0.‫סס‬OO1
of a large number of random extraneous sig- vidual measurements, and /l is the arithmetic strument is available, one can say that the 1.63 -0.037 0.00137
nals (noise) that develop in various instru- mean for an infinite number of such measure- chances are 68.3 out of 100 that the uncer- 1.70 0.033 0.00109
ment components. In some instances,the direc- ments. The quantity (Xi - /l) is thus the devia- tainty associated with any single measure- 315.00 D-l (Xl - W = 0.00247
tion of the noise from most of the components tion from the mean; y is the frequency of ment is smaller than ± 1(J, that the chances X = 1.667
will, by chance, be positive, and a higher than occurrence for each value of (XI - /l). The sym- are 9S.5 out of 100 that the error is less than
average reading will result; at other times, boln has its usual meaning, and e is the base ±2a, and so forth. Oearly, the standard de-
negative noise signals may predominate, lead- for Napierian logarithms. The parameter (J is viation for a method of measurement is a s = JO.00247 = 0.351 = 0.04
3 -1
ing to a result that is less than average. The called the standard deviation and is a constant useful quantity for estimating and reporting
most probable event, however, is for the that has a unique value for any set of data the probable size of instrumental uncertain- Generally, in a calculation ofthis type, it is
number and size of the negative and positive comprising a large number of measurements. ties arising from random noise. • wise to avoid rounding until the final result
noise signals to be about the same, thus giving The breadth of the normal error curve, which For a very large set of data the standard has been obtained.
a reading that approaches an average or mean increases as the precision of a measurement deviation is given by Another precision term widely employed
decreases, is directly related to (J. Thus, (J is
value.
For a large number of replicate measure-
ments, the accumulated uncertainties due to
widely employed to define the precision of in-
struments.
(J = Jr:-1 ~ -/l)Z (A-3)
by statisticians is the variance, which is equal
to (J1. Most experimental scientists prefer to
employ (J rather than (Jz because the units of
random noise cause the results to be dis- The exponential term in Equation A-I can Here, the sum of the squares of t~ individual the standard deviation are the same as those
tributed in the symmetrical way shown in be simplified by introducing the variable deviations from the mean (XI - p) is divided of the quantity measured.
Figure A-I. This distribution, which is termed by the total number of measurements in the
Xl-/l
Gaussian, can be described in terms of three z=-- set, N. Extraction of the square root of this
(J
quotient gives (J.
which then gives the deviation from the mean Equations A-I and A-3 apply exactly only
in units of standard deviations. as the number of measurements approaches
infinity. When N is small, a better estimate The true mean value (Il) of a measurement is
for the standard deviation is given by a constant that must always remain unknown.
With the aid of statistical theory, however,
limits may be set about the experimentally
determined mean (x) within which we may
Equation A-I indicates that a unique distri-
expectto findthe true mean with a givendegree
bution curve exists for each value of the stand-
of probability; the limits obtained in th~ man-
ard deviation. Regardless of the size of (J, Note that the symbol s is used to distinguish ner are called confidence limits. The mterval
however, it can be shown that 68.3% of the the experimentally realizable estimate of the defined by these limits is known as the confi-
area beneath the curve lies within one stand- standard deviation from the theoretical (J. dence interval. The confidence interval can be
ard deviation (± 1(J) of the mean, /l (shaded Furthermore, the experimental mean of this determined from the value of z in Equation
region, Figure A-I). Thus, 68.3% of the re- set x can only approximate the true mean for A-2. For example, if z is given a value of + 1,
sults from replicated measurements would be an infinite number of measurements /l. When ~henEquation A·2 takes the form
expected to lie within these boundaries. Ap- IV is greater than 20 to 30, it is usually safe to
proximately 95.5% of all values should lie assume that s -> (J. Xl -/l = (J
Devilltion from meen within ±2(J; 99.7"10 should appear within ±3(J.
l>ri -Ill Values of (Xl -/l)correspondingto ± 1(J,±2(J, EXAMPLE
and ± 3(J are indicated by broken vertical lines The following data were obtained for the Xl-/l = -(J
FIGURE A-1 A normal or Gaussian distri- in Figure A-I. determination of mercury in the flesh of a
bution curve. These properties of a Gaussian curve are trout by a fluorescence method: 1.67, 1.63, As mentioned earlier, approximately 68% of
the area under a normal distribution curve For the 50% confidence leve~ z = ±0.67, the purpose of such an analysis, let us assume never cancel. In contrast, type 2 terms (called
lies within the limits of ± 10'. Thus, there and that the measured quantity x is dependent cross tenns) may be either positive or negative
upon the variables p, q, r, ... , which fluctuate in sign. Examples of these terms are
exists a 68 % probability that the deviation
from the mean (XI - Jl) for any single measure- in a random and indepe¢ent manner. That
ment will be within this interval. Similarly,
50% confidence limit = 1.67 + 0.67$10.07 is, x is a function f of p, q, r, : .. , so that we
- 3
when z is equal to ±2, may write .
= 1.67 ±0.03 If dp, dq, and dr are independent and random,
XI- Jl = ±20'
x = f(P, q, r, ... ) (A-6)
Similarly, some of the cross terms will be negative and
The uncertainty dXI (that is, the deviation others positive. Thus, the summation of all
Here, 96% of the area under the curve lies from the mean) in the ith measurement of x
within these limits; thus there exists a 96%
95 % confidence limit = 1.67 ± 1.96:fi0.07 such tenns should approach zero.
will depend upon the size and signs of the As a consequence of the canceling tendency
probability that the deviation from the mean corresponding uncertainties dph dqjo drjo ....
= 1.67 ± 0.08 of type 2 terms, the square of Equation A-8
of an individual measurement will lie within That is, can be readily obtained and summed from
this range. Areas in terms of percentages are
Thus, the chances are 50 in 100 that the true = f(dp/o dq/o = (XI - Jl) i = 1 to N. Thus,
termed confidence levels. Confidence levels for dXI dri, ...)

a few values of z are given below. mean Jl will lie in the interval between 1.64 2 2
The variation in dx as a function of the uncer-
T L (dpi)2 + T
L (dXI)2 = (bX)
and 1.70 ppm Hg; there is a 95% chance that N (bX) iN
tainties in p, q, r, ... , can be derived by taking 1-1 up uq 1-1
the true mean lies between 1.59 and 1.75 ppm.
Confidence Level 50% 68% 90% the total differential of Equation A-6. That is,
It should be noted that the standard devia-
z ±0.67 ±1.00 ±1.64
tion from the three measurements themselves L (dql)2 + (bX)2
N
- L (dri)2 + ... N

dx = (bX) dp + (bX) dq i=1 br i=1


Confidence Level 95% 96% 99% was not employed in calculating the confidence
bp 4. r.... bq p. r•...
±1.96 limit in the foregoing example. With such a
z ±2.00 ±2.58

The confidence limit for a single measure-


small number of measurements, there is no
assurance that s -+ 0'; indeed, s for the three
data in question is considerably smaller (0.04)
+ (~:L4"" dr+'" (A-7)

ment is obtained by rearranging Equation In order to relate the various terms in Equa-
than the value of 0.07 that was obtained from
A-2 to tion A-7 to the standard deviation of x, p, q,
25 replicate measurements. Without this sure
and r as given by Equation A-3, it is neces-
knowledge of the standard deviation, the con-
sary to square the foregoing equation. Thus,
fidence limit would have had to be widened. 1

When Xi is replaced by X, the mean for several (dX)2 = [(i5X) dp + (bX) . ~q From Equation A-3, however, we see that
measurements, the confidence limit is de-
bP 4. r. ... bq p. r•...

creased by )N. Thus, a more general expres-


sion is
PROPAGATION OF
UNCERTAINTIES IN + (~:L 4 •.•• dr + .. -r (A-8)
(dXI)2 = (Xi - Jl)2
N N
=~
x

confidence limit for Jl = x- IN (A-5)


A CALCULATED RESULT

The result of an analysis is typically cal-


Then, the resulting equation must be summed
between the limits of i = 1 to i = N, where N
where O'~ is the variance of x. Similarly,

(dpi)2 = (12
culated from two or more experimental data,
again is the total number of replicate measure- ;N p

The example that follows illustrates the calcu- each of which has associated with it an uncer- ments (Equation A-3).
In squaring Equation A-7, two types of and so forth. Thus, Equation A-9 can be writ-
lation of confidence limits. tainty due to random noise. It is worthwhile
terms from the right-hand side of the equa- ten in terms ofthdvariances of the quantities;
exploring the ways in which these various un- that is, ,
tion become evident. The first are type 1 terms,
certainties accumulate in the final result. For
EXAMPLE such as 2 2 2

From 25 replicate measurements, the stan-


(12
x
= (bX)
bP ~
p
+ (bX)
bq ~
4
+ (bX)
br ~
r
+ ...
dard deviation for the fluorescence method (A-10)
I For methods or lreating data when t1 is not known, see
used in the earlier example was found to be
D. A. Skoog and D. M. West, Fundamentals of Analytkal The example that follows demonstrates an
0,07 ppm Hg. Calculate the 50 and 95% confi- Chemistry, 3d ed. New York: Holt. Rinehart and Winston,
Because they are squares, type 1 terms will
dence limits for the mean of the J~ee analyses. 1976. p. 64. always be positive. Therefore, these tenns can application of Equation A-10.
EXAMPLE Dividing this equation by the square of Equa- The foregoing example illustrates two im-
The chloride in a 0.1200-g (± 0.0002) sample tion A-ll gives - portant, general statistical relationships.
was determined by couJometric titration with (1) The absolute variance ora sum or dif-
silver ions (p. 597). An end point was reached
after 167.4 ( ± 0.3) s with a current of 20.00 mA
(±0.04). Titration of a blank required 13.2
t:r = (T~ Tor+ (- T~·TJ2
ference is the sum of the individual absolute
variances.
(2) The relative variance of a product or
Note in the example that it was first neces-
sary to add the absolute variances of T and
(±0.3)s with the same current. The numbers in
parentheses are-absolute standard deviations
+(~lr+(-~r quotient is the sum of the individual relative
variances. Thus, for the relation y = a + b or
To to give the absolute variance of the difference
between the two numbers. After this the relative
associated with each measurement. What is the y=a-b, variance was calculated. It was then combined
with the relative variances of the other two
relative and absolute standard deviation for the
percent chloride?
u:=cr.+~ numbers (I and W) making up the quotient
The percent chloride is given by and the standard deviation of y is to give the relative variance of the quotient.

(1,=~
x = % CI = k(T - To)1
W In contrast, when y = a . b or y = a/b,
;

where T and T~are the times (in seconds) for Substitution of the numerical data into Equa-
titration of the sample and blank, respec- tion A-ll gives
tively, 1 is the current in mA, W is the sample
weight in g, and k is a constant whose value X=%O
(3.6742 x 10-') is known with a high degree
_ 3.6742 x 10-'(167.4 - 13.2) x 20.00
of precision. ;
- 0.1200
To employ Equation A-I0, we first take the
partial derivatiVe of percent chloride with re- = 0.94427
spect to T, holding the other variables con-
stant. Thus, Substitution of numerical values into Equa-
tion A-12 yields
(bX) kl
bT To,I.W =W
= +(:o~r
(;r (~~~;'4+~~3O;:
(bX) kI 0.0002)2
bTo T.I,W = - W +(- 0.1200

= 7.57 x 10-6 + 4.00 X 10-6

(bX)
M T,T •• W
= k(T - To)
W
+ 2.78 X 10-6

(1" = J1.435 X 10-5 = 3.8 X 10-3

(!!-.) ,
X
= _ k(T - To)1
bW T;.TOoI ~ Thus, the relative standard deviation for the
analysis would be expected to be about 4 ppt.
Applying Equation A-I0 gives To obtain the absolute standard deviation, we
write

(1" = 0.9943 x 3.8 x 10-3 = 0.0038% 0


and the result could be reported as
% CI = 0.994(±0.004)
· APPENDIX 2:
SOME STANDARD ANC
FORMAL ELECTRODE
POTENTIALS
Half-reaetion" EJ, V ~ormaI potential, V
E!', v Formal potential, V
F2(g) + 2H+ + 2e~2HF(aq) +3.06
Ag+ + e~Ag(s) +0.799 0.228, 1 F HCI; 0.792, Fe2+ + 2e~Fe(s) -:-0.440
1 F HCIO.; 0.77, 1 F H2SO. Fe3+ + e~Fe2+ +..0.771 0.700, 1 F HCI; 0.732,
AgBr(s) + e~Ag(s) + Br- +0.073 1 F HCIO.; 0.68, 1 F H2SO.
AgC1(s)+ e~Ag(s) + a- +0.222 0.228, 1 F KCI Fe(CN)~- + e~Fe(CN)~- +0.36 0.71,1 FHa; 0.72,1 F HCIO.,
Ag(CN)i + e~Ag(s) + 2CN- -0.31. H2SO.
Ag2CrO.(s) + 2e~2Ag(s) + crO~- +0.446 2H+ + 2e~H2(g) 0.000 -0.005, 1 F HCI, HCIO.
AgI(s) + e~Ag(s) + 1- -0.151 Hgi+ + 2e~2Hg(l) +0.789 0.274, 1 F HCI; 0.776,
Ag(S203W + e~Ag(s) + 2S20~- +0.01 1 F Hao.; 0.674, 1 F H2SO.
AI3+ + 3e~AI(s) -1.66 2Hg2+ + 2e~ Hgi+ +0.920 0.907, 1 F HCIO.
H3AsO. + 2H+ + 2e~H3As03 + H20 +0.559 0.577, 1 F HC~ HCIO. Hg2+ + 2e~Hg(l) +0.854
Ba2+ + 2e~Ba(s) -2.90 Hg2CI2(s)+ 2e~2Hg(l) + 2CI- +0.268 0.242,sat'd KCI; 0.282,
BiO+ + 2H+ + 3e~Bi(s) + H20 +0.32 1 F KCI; 0.334,0.1 F KCI
BiCli + 3e~Bi(s) + 40- +0.16 Hg2SO.(s) + 2e~2Hg(l) + soI- +0.615
Br2(1)+ 2e~2Br- + 1.065 1.05,4 F Ha HOi + H20 + 2e~30H- +0.88
Br2(aq) + 2e~2Br- +1.087· 12(5) + 2e~21- +0.5355
BrO; + 6H+ + 5e~!Br2(1) + 3H20 +1.52 12(aq)+ 2e~21- +0.6N
Ca2+ + 2e~Ca(s) -2.87 I; + 2e~31- +0.536
C6H.02 (quinone) + 2H+ + 2e~C6H.(OHh +0.699 0.696,1 F HCI, HCIO., H2SO. ICli + e~!12(s) + 2CI- +1.06
2C02(g) + 2H+ + 2e~H2C20. -0.49 103" + 6H+ + 5e~!12(s) + 3H2O + 1.195
Cd2+ + 2e~Cd(s) -0.403 10; + 6H+ + 5e~!12(aq) + 3H2O + 1.178b
Ce·+ + e~Ce3+ 1.70, 1 F HCIO.; 1.61, 103 + 2CI- + 6H+ + 4e~ICli + 3H2O + 1.24
1 F HN03; 1.44, 1 F H2SO.; HsI06 + H+ + 2e~10; + 3H2O +1.60
1.28,1 F HCI K+ +e~K(s) • -2.925
CI2(g) + 2e~2CI- +1.359 Li+ + e~Li(s) -3.045
HCIO + H+ + e~!C12(g) + H20 +1.63 Mg2+ + 2e~Mg(s) -2.37
ao; + 6H+ + 5e~!a2(g) + 3H20 +1.47 Mn2+ + 2e~Mn(s) -1.18
Co2+ + 2e~Co(s) -0.277 Mn3+ + e~Mn2+ 1.51,7.5 F H2SO.
Co3+ + e~Co2+ +1.842 Mn02(S) + 4H+ + 2e~Mn2+ + 2HzO + 1.23 1.24,1 F HCIO.
Cr3+ + e~Cr2+ -0.41 MnO; + 8H+ + 5e~Mn2+ + 4HzO + 1.51
Cr3+ + 3e~Cr(s) -0.74 MnO; + 4H+ + 3e~MnOz(s) + 2HzO +1.695
Cr2o~-I+ 14H+ + 6e~2Cr3+ + 7H20 +1.33 MnO; + e~MnO~- +0.564
Cu2+ + 2e~Cu(s) +0.337 Nz(g) + 5H+ + 4e~NzH; -0.23
Cu2+ + e~Cu+ +0.153 HNOz + H+ + e~NO(g) + HzO +1.00
Cu+ + e~Cu(s) +0.521 NO; + 3H+ + 2e~HNOz + HzO +0.94 0.92, IF HN03
Cu2+ + 1- + e~Cul(s) +0.86 Na+ + e~Na(s) -2.714
Cul(s) + e~Cu(s) + .- -0.185 Ni2+ + 2e~Ni(s) -0.250
HzOz + 2H+ + 2e~2HzO + 1.776
• Sources for E' values: A. J. deBcthune and N. A. S. Loud, Standard Aqwous Electrode Potentials and Temperature Oz(g) + 4H+ + 4e~2HzO +1.229
Coe~s at 25'C. Skokie, In.: CJifI'ordA. Hampel, 1964. Source of formal potentials: E. H. Swift and E. A. Butler, Oz(g) + 2H+ + 2e~HzOz "+0.682
~ltatJW' l/.-_s and .ChmI/cQl Equilibr/Q. San Francisco: W. H. Freeman and Company. Copyright C 1972 03(g) + 2H+ + 2e~Oz(g) + HzO +2.07
These potentials are hypothetical because they correspond to solutions that are 1.00 M in Brs or Is. The solubilities of
Pb2+ + 2e~ Pb(s) -0.126 -0.14, 1 F HaO.;
these two compounds at 2S'C are 0.21 M and 0.0133 M, respectively. In saturated solutions containing an excess of -0.29, 1 F H2SO.
Brs(l) or Is(s~ the standard potentials for the baIf-reactions Brs(J) + 2e~2Br- or Is(s) + 2e~21- should be used. On
the other band, at Brs and Is concentrations Jessthan saturation, these hypothetical elearode potentials should be employed. Pb02(s) + 4H+ + 2e~Pb2+ + 2HzO + 1.455
PbSO.(s) + 2e~Pb(s) + SO~- -0.350

j
HaIf-reactioa" £0, V Formal potential, V
INDEX
Pt~- + 2e~Pt(s) + 4Cl- +0.73
. PtCl~- + 2e~Pt~- + 2Cl- +0.68
-·PdH + 2e~Pd(s) +0.987
S(s) + 2H+ + 2e~H2S(g) +0.141
H2S03 + 4H+ + 4e~S(s) + 3H2O +0.45 Abbe refractometer(s), 355 Active computer lntcraction(s), 85, 86
S40~- + 2e~2S20j- +p.08 Absolute refractive index. n.,.." 353 Activity, a, S07; effect on electrode potential calculations,
SOi- + 4H+ + 2e~H2S03 + H2O +0.17 Absorbance, A, 149, ISOl; measurement of, 151, 237; 522; response of electrodes to, 556
S20i- + 2e~2S0i- variables afl'ecting, 190 Activity coeIIicient(s),.r. S07-508, 5091
+2.01 Absorption analysis, 151, 187 (see also Absorption of Addition of input signals, with operational amplifiers, 65
Sb20,(s) + 6H+ + 4e~2SbO+ + 3H2O +0.581 electromagnetic radiation, listinp under spectral Address(es), computer, 81
H2Se03 + 4H+ + 4e~Se(s) + 3H2O +0.740 regions); atomic, 3021 (see also Atomic absorption Adsorption isotherm(s), 669
SeOi- + 4H+ + 2e~H2Se03 + H2O +1.15 spectroscopy); automated, 199-202; cleaning of cells Adsorption chromatography, 690, 696-698
SnH + 2e~Sn(s) -0.136 -0.16, 1 F HCI04 for, 190;derivative, 192;dilTercntia\,194-197;effect,of Agar ge!(s), 533
Sn4+ + 2e~SnH +0.154 0.14,1 F HCl
noise on precision of, 157(see also Noise), of slit width Air, analysis for contaminants in, 239, 240/
Ti3+ + e~TiH upon, 163;erronin, 153-166,194;metbodsof,195/(see Air gap(s), of spark source, 339
-0.37 also DilTercntia1absorption 1JIB1yIis);photoacoustic, Alizarin garnet R, as f1uorometric reagent, 294, 2951
TiOH + 2H+ + e~Ti3+ + H2O +0.1 0.04, 1 F H2SO4 202; precision, 194-197; wavelength se1ectionfor, 190 Alltaline error, g\ass electrode, 542, S44-S46
Tl+ + e~Tl(s) -0.336 -0.551, 1 F HCl; -0.33, Absorption edge(s), 431 Alltane(s), as solvents in visible-ultraviolet regions, 189/
1 F HCl04, H2SO4 Absorption of electromagnetic radiation, 105; analytical Alkene(s), absorption characteristics of, 173/, 174/
Tl3+ + 2e~Tl+ + 1.25 0.77,1 F HCl
applications involving, 149 (see also under specific A\ltyne(s), absorption characteristics of, 1731
uoi+ + 4H+ + 2e~U4+ + 2H2O types, spectral regions); by atoms, 106(see also Atomic Alloy junction transistor(s), 45
+0.334 absorption); by inorganic species, 17S-ISO; instru- Alpha particle(s),cr, measurement of, 465; properties of, 459
V3+ +e~VH -0.255 -0.21, 1 F HCI04 ments for measurement of, 180-186 (see also under Alternating currcnt(s), 11-14; conduction of, in elec-
VOH + 2H+ + e~V3+ + H2O +0.361 specific types, spectral regions); in Iasen, 120;laws of, trochemical cells, S05, 506, 646
V(OH)t + 2H+ + e~VOH + 3H2O +1.00 1.02, 1 F HCl, HClO4 151 (see also Beer's law); in a magnetic field, 108,(see Alternating current polarography, 631-633
ZnH + 2e~Zn(s) -0.763 also Nuclear magnetic resonance, Electron spin Alumina, as stationary phase, liquid-solid chromatog-
J'CSOnance);mechanismof,105,171;bymixtures, 152; raphy, 697/
by molecules, 106, 169,210; nomenclature relating to, American Society for Testing Materials, catalog of d-
ISOr,by organic species, 172, 173/, 174/ spacings, X-ray diffraction, 452; spectroscopic nomen-
Absorption filtcr(s), for electromagnetic radiation, 136 clature recommendations of, ISOl
Absorption signal(s), measurement of, nuclear magnetic Amici prism(s), 355
J'CSOnancespectroscopy, 384 Amido group(s), as chromophore, 1731
Absorption spectra, 105(see also specifictypes);methods of Amine oxide group(s), polarographic behavior of, 621
plotting, 18'1;solvent effects upon, 172 Ammeter(s),28
Absorptivity, a, 149, ISO/ (see also Molar absorptivity) Ammonia, gas sensing electrode for, 554
Be Circuit(s), reactance in, 22 Ammonium dihydrogen phosphate, diffraction of X-rays
ac Polarography, 632 (see also Alternating current by, 437/, 438
polarography) Ampere, A, 7
Acetone, as solvent in visible-ultraviolet regions, 189/ Amperometric titration(s), 635-640; applications, 637,
Acetonyl acetone, as liquid phase, gas-liquid chromatog- 638/, 640; with two polarized electrodes, 637
raphy, 7241 Amperostat(s), 64
Acetylene f1ame(s), temperature of, 309/ Amplification, of analytical signals, 4; gas, 439; 'of noisy
Acid(s), conductometric titration of, 649; coulometric signals, 68; of transducer signals, 59; with transiston,
titration of, 597; potentiometric titration of, 568; 45,46
thermometric titration of, 663, 6641 Amplifier(s), based upon transistors, 48-52; chopper, 74;
Acid-base titration(s), S68 (see also Neutralization lock-in, 75
titrations) Amplifier excessnoise, effectupon measurement of T, 159'
Acid dissociation CODstant(s), evaluation from poten- Amplitude, A, of electromagnetic radiation, 93; of sine-
tiometric data, 570 wave current, !2
Acid error, glass electrode, S46 Analog-to-digital converter(s), for computers, 83
Actinide ion(s), absorption by, 175 Analytical gap, for spark source, 339
Activation analysis, 467 (see also Neutron activation Analytical signa1(s),2, 31; effect of noise upon, 70
analysis) Analyzer(s), centrifugal photometric, 200;continuous flow,
199; for mass spcctromeler, 477, 481; multiclwme1, atomi2lers for, 312-315; classification of, 3021; detec- Ben2Jene, eft'ect of substituents upon ftuorescence of, 286/; Carbon disulfide, as solvent for infrared spectrophotom-
142 tion limits for, 319/, 332/, 350/ . as solvent in visible-ultraviolet regions, 189/; for etry,227/
Analyzer prism, of polarimeter, 368 (.r« also Nicol prism) AttlIDuation, of analytical signals, 5; of reference beam, infrared absorption spectrophotometry, 227/ Carbon tetraeh1oride, as solvent, for absorption spectrom-
Angstrllm unit, A, 93 ftuorescence spectrometry, 289 Benzoin, as ftuorometric reagent, 294, 295/ etry, 189/, 227/; for nuclear magnetic resonance
Angu\ar dispersion, of optical materials, 1261; of prisms, Attenuator(s), for infrared spectrophotometer(s), 220 • Benzyldiphenyl, as liquid phase, gas-Iiquid chromatog- spectroscopy, 390
127 Aqer eIectron(s), production of. 454, 460 raphy, 724 Carbonyl bond(s), eft'ect of magnetic fields upon, 395
Angular velocity, 01, of etectromagneti<: radiation, 93 Aqer spectroscoPY. 453 Beta particle(s), fl, measurement of, 466; properties of, 459/, Carbonyl group(s), as chromophore, 173/, 174/; influence
Anharmonic oscillator(s), 217 Automatic titrator(s), 573-576; as selective detectors, gas- 460 upon fluorescence. 287; polarographic behavior of, 621
Anion interference(s), atomic absorption spectroscopy, 321 liquid chromatography, 729 Biasing, of semiconductor diode(s), 44 Carboxylic acid group(s), as chromophore, 173/; influence
Anisotropic crystal(s), absorption of radiation by, 105 Automation, of analyses, with computers, 85; of iiIStrU- Binary numbering, 80 upon ftuorescence, 287; polarographic behavior of,
Anisotropy, 357; magnetic, 395 ments, 79; for spectrophotometric and photometric Bipolar transistor(s), 45-48 621
Anode, of e1ectrocbemical cell, 502; eIectrogravimetric analysis, 199-202, 223, 327 Bit(s), in binary numbering, 81· Carrier(s), for trace constituents, radiochemical analysis,
deposits on, 587 Auxochrome(s), 174 Black-body radiation, 109 468
Anodic peak(s), cyclic vo1tammetry, 631 Averaging, of input signals, 65 Blue shift(s), of absorption spectra, 172 Carrier gas(es), for gas-liquid chromatography, 717
Anodic stripping voltammetry, 633 Azide ion(s), absorption in the ultraviolet by, 175 Bohr magneton, flR' 412 Catalog(s), of infrared spectra, 237; of Raman spectra, 271;
Anodic wave(s), polarographic, 615 Azo group(s), as chromophore, 173/; polarographic Bolometer(s), 144 of thermograms, 661
Anomalous dispersion, 102 (see also Dispersion) behavior of, 621 Bonded phase pacldng(s), for partition chromatography, Cathode. of electrochemical cell. 502; electrolytic deposits
Anthracene, use in scintillation counters, 441 ~8 : upon, 587
Antibonding molecular orbital(s), and absorption of Bonding e1ectron(s), absorption of radiation by, 170 Cathode depoIarizer(s), 587
radiation, 170 Background correction method(s), atomic absorption Boundary potential, glass electrode, 543, 545 Cathode-ray tube(s). n
Anticoincidence circuit, of pulse heigbt analyzer, 444 spectroscopy, 320; atomic emission spectroscopy, 328; Bragg equation, for X-ray dift'raction, 434 Cation interference(s), atomic absorption spectroscopy,
Antilogarithm(s), of input signals, 67 in radiochemical analysis, 464 Breakdown voltage, for reverse-biased semiconductor 322
Antimony, doping of semiconductors with, 42 Background radiation, of high-energy sources, 337 diode, 44 (see also Zener breakdown voltage) Cell(s), for absorption analysis, cleaning and handling of.
Anti-Stokes shift(s), 264 Ba1ance(s), electromagnetic, 655 Breathing vibration(s), of cycl6paramns, 272 190; for conduetance measurements, 647; for coulom-
APDC, as protective reagent, atomic absorption spectros- Band spectra, 337; correction for, atomic emission Brehmsstrahlung, 427 f etric titrations, 595; electrochemical, 8, 50 I (see also
copy, 322 spectroscopy, 328; interference from, in atomic Bromide ion, solid-state electrOde for, 552/ Electrochemical cells, Cell potentials); for elec-
Arcsource(s), for emission spectroscopy, 336, 337, 338, 347 absorption spectroscopy, 320 Bromine, isotopic abundance of, 490/ trogravimetric analysis, 587; for fluorescence analysis,
Arc spectra, emission spectroscopy, 338 Bandwidth, of Wters, 136, of monochromators, 132 Buft'er(s), radiation, 320; spectroscopic, 350 291; gas-sensing, 554; for infrared absorption analysis,
Area normalization method, chromatographic analysis, Barium titantate, use in pyroelectric detectors, 145 Bunsen monochromator(s), 126 227; for optical spectroscopy, 137; photovoltaic, 138;
684 Barrier-layer ce1l(s), 138 Bumer(s), for atomic spectroscopy, 313 for polarimetry, 369; for polarographic analysis, 601,
Argon ion laser(s), 123; as Raman source, 268 Base(s), conductometric titration of, 649; coulometric .617; forspectrophotometric analysis, 190; Weston, 34,
Argon plasma source(s), emission spectroscopy, 336, 340 titration of, 597; potentiometric titration of, 568; 534
Arithmetic unit, of computer, 82 thermometric titration of, 664/ Cadmium ion, solid-state electrode for, 552/ Cell constant(s), for conduetance measurements, 647
Aromatic compound(s), absorption by. in infrared, 238; in Base Ilia, in amplifier circuits, 49 Calcite, double refl'llction of, 359; refractive index of, 359/ Cell in<ell out method, for quantitative infrared absorp-
the ultraviolet, 174, 1751; ftuorescence by, 286/ Base layer, of bipolar transistor, 45 Calcium fluoride, optical characteristics of, 126 tion spectroscopy, 238
Arsenic, doping of semiconductors with, 42 Base line method, quantitative infrared absorption spec- Calcium ion, liquid-membrane electrode for, 549, 551/ Cellosolve, as solvent in visible-ultraviolet regions, 189/
Arytron shunt(s), 29 troscopy, 238 Calculation(s), propagation of uncertainties in, 738-744 Cell positioning uncertainty, effect upon measurement of T,
Aspiration rate(s), eft'ect upon ftame spectroscopy, 312 Base peak, of mass spectrum, 490 Calibration curve(s), for absorption analysis, 190 (see 159/,162
Ascending development, planar chromatography, 706 Base region, in flame, 311 also Beer's law); for atomic absorption spectroscopy, Cell potential(s), 507; calculation of, 524-527
Assembly 1anguage(s), for computers, 84 Basic, computer language, 84 324; for direct potentiometry, 557; for emission Central processing unit, of computer, 82
Asymmetry potential, eft'ect upon direct potentiometric Basic inverting circuit(s), incorporating operational am- spectroscopy, 345, 349; for polarographic analysis, Centrifugal fast analyzer(s), photometric, 200
measurements, 5S6; glass electrode, 542, S44 plifiers, 57 Cesi um iodide, optical characteristics of, 126/
619
Atomic absorption, 106 Batch inlet(s), for mass spectrometers, 479 Calibration standard(s), for chromatographic analysis, Channel(s), in FET transistors, 48
Atomic absorption spectra, 307 Bathochromic shift(s), of absorption spectra, 172 Charge density, of molecular orbitals, 170
683; for refractometers, 356; for X-ray analysis, 449
Atomic absorption spectroscopy, 304, 315-326; analytical Bausch and Lomb Spectronic 20 spectrophotometer, 182 Calomel electrode(s), 531-533 Charge-transfer absorption, 179
techniques for, 324; applications, 318; detection limits B Band, in ultraviolet spectra of aromatic compounds, 174, Capacitance, C, 14, 16 ; Charging, of a capacitor, 15
for, 350/; instruments for, 317; interferences aft'ecting,
320-324
175/ Capacitative reactance, Xc, 2i Charging current(s), polarographic wave, 616
Beam splilter(s), for Fourier transform spectrometer, 250 Capacitor(s), 14 Chemical bond(s), polarizability of, 266
Atomic emission spectra, 110,306 Beat, in time domain spectrum, 243 Capacity factor, k', chromatographic, 6n, 679, 682/ Chemical couIometer(s), 592, 593
Atomic emission spectrolCopy. 302/. 326; analytical Becbnan AccuLab infrared spectrophotometer. 222 Capillary column(s), gas-Iiquid chromatography, 719 Chemical deviation(s), to Beer's law, 153
teehniqla for, 330; detection limits for, 3191. 3501; IIecbnan DU-2 spectrophotomet. 183 Capillary conslant(s), for dropping mercury electrode, 614 Chemical exc:haqe, eft'ect, upon nuclear magnetk: res0n-
interferences in, 328-330 ance spectra, 40 I
Beer's law, 151-157; deviations from, in infrared region, Capture cross section, for neutrons, 469; for photons, 152
Atomic ftuorescence spectroscopy, 303, 332; detection 237; elFect, of polychromatic radiation upon, 155, of Carbon, isotopic abundance of, 490/ Chemical interference(s), atomic absorption spectroscopy,
limits for, 350/ I 320, 321; atomic emission spectroscopy, 329
refractive index changes upon, 153; limitations to, Carbon-13 nuclear magnetic resonance, 412
Atomic ftuorescent spectra, 307 Chemical ionization mass spectra, 490
153-157; and X-ray absorption, 432 Carbonate ion(s), absorption in tbe ultraviolet by, 175
Atomic spectroscopy, 302-333 (oree ~. spec::i1ic types); Bending vibration(s), 212 Carbon dioxide, gas-sensing electrode for, 551 Chemical ionization source(s), for mass spectrometers, 480
Chemical shift(s), carbon-13 nuclear mapetic resonance Complex formation, effect, upon e1ectrodepotentials, 521, Con-.ion, externa\, 284; internal, 283 (_ also Internal Cut-off lllter(s). 137
spectroscopy, 412; correlation with stnJeture, 396; con-.ion) CuYette(s), 137
upon polarosrapbic waves. 611
proton nuclear magnetic resonance spectroscopy, 391, Complex-formation titration(s), coulometric, 597, 5981; Coolidge tube(s), 434 Cyanide ion, solid-state electrode for, 552/
394-398; soun::csof, 391 potentiometric, 566; thermometric, 6641 Copper(II), solid-state electrode for, 5521 Cyanoaen band, emission spectroscopy, 337
Chemical shift ~s), nuclear mapetic resonance Compton effect, 460 Cornu prlsm(s), 126, 344 . Cyanoaen tlame(s). temperature of, '3091
spectroscopy, 392 Computer(s), application to chromatOJl'llpbic analysis, Correlation table(s), for 2.rbindilll energies of sulfur, 455; Cyclic voltammetry, 631
Chemical sbift reagent(s), 420 684; components of, 82-84; interfacing of instruments for chemical shifts, nuclear magnetic resonance CydohexaDe, as solvent, for absorption spectroscopy, 1891,
Chiral molecule(s), 366 to, 79-86 spectroscopy, 3971;infrared spectra, 230. 231t;Raman 227
Chloride ion, coulometric determination of, 593; liquid Computer interaction(s), 85 spectra, 271 Cydoparatlins, breathing vibrations of, 272
membrane electrode for, 55I I; solid-state electrode for, Computer language(s), 84 Cotton effect, 370 Czemey-Turner monochromator(s), 318
552/ Computer programming, 84 Coulomb; C, 7. 591
Chlorine, isotopic abundance of, 490/ Computer search system(s), for infrared spectra, 223 Coulombic repulsion(s). during molecular vibrations, 217 Damping, of potarosrapbic currents, 612
Chloroform, as solvent for infrared spectrophotometry, Concentration, direct potentiometric measurement of, 5S6; Coulometer(s). chemical, 592 (s« also Chemica1 Daniell cell(s), S07
2271 effect of, upon absorbance, 151 (see also Beer's law), coulometer) Dark current(s), 138, 140;compensation for, 151
Choke(s), 14 upon electrochemical cells, 510 (see also Nemst Coulometric analysis. 590-600 (see also Coulometric Dark current noise, effect on measurement of T, 1591(see
Chopper amplifier(s), 73 equation), upon fluorescent intensity, 288 (see also titrations); with constant electrode potential, 592, 593 also Noise)
Chopping, of radiation, 73, 221 Fluorescence spectrometry), upon scattering, 296; Coulometric titration(s), 581. 592, 594-600; applications, Dark current llignal(s),for fast analyzers, 202
Chromatic aberration, 103 evaluation of, by chromatosrapbic analysis, 681, 597-599; externa1 generation of reagent, 596; soun::cs d'Arsonval meter(s), 28; use, in electromagnetic balances,
Chromatogram(s), 670; zone shapes of; 672 nuclear magnetic resonance spectroscopy, 407 of error in, 599 655. as signa1processor, optical spectroscopy, 146
ChrolJl&t0Jl'llpbic analysis, applications, 681-686; com- Concentration sradient(s), in solution, 529, 609 Coulometric titrator(s), 595; automatic, 599 Data storage, in computer memories, 86
puter assisted, 684; errors in, 683 Concentration polarization, 528-530, 581; at dropping Counter(s), for X-ray spectrometers, 443 de An:(s), for excitation of emission spectra, 347
ChromatoJl'llpbic separation(s), on columns, 676-680 (see mercury electrode, 610 Counter e1ectrode(s),626 de Circuit(s), 10-11
also under specific types) Condenser current(s), polaroJl'llPbic wave, 616 Counting, photon, 438 de Current(s), in electrochemical cells, 504; in electrolytic
ChromatoJl'llphy, 667-732 (see also under specific types); Conductance, G, 7, 644 (see also Equivalent conductance); Counting error(s), in radioactive assays, 462-465 cells, with fixed applied potential, 583 .
linear elution, 669; types of, 668/; zone shapes in, 672 applications of direct measurements, 65I; measure- Coupling, of molecular vibrations. 212, 218 Deactivation, of excited molecular states, 282-284
Chromophore(s), 170; organic, 172, 173/, 174/ ment of, 61, 647; in a parallel circuit, II Coupling constant, J, 398 Dead-stop end point(s), 638
Circuit(s), amplifier, 48; equivalent, II; parallel, 10; Conduction, of alternating current, 42 (see also Alternating Critical angle, 9•• 354 Dead time, of Geiger tube, 440
reactance in, 14(see also Reactance); series, 8 (see also current); of electricity, 7-8; in electrochemical cells, Critical angle refractometer(s), 354 Debye-HUckel equation, 50S
Series circuits); RC, 14; RL, 16; RLC, 24 SOl; across glass membranes, 543; in semiconductors, Crysta1(s),ditrraction of X-rays by, 4371 (_ also Bragg Debye-Scherrer powder camera, 45I
Circular birefringence, (n, - n.), 366; wavelength depen- 42,43 equation) Decay. radioactive, 459 (see also Radioactive decay)
i:Ienceof, 370 Conductometric method(s), 644-653 Crystal field theory, 176 Decay constant, .i, 462
Circular dichroism, 370-374 Conductometric titration(s), 644, 648, 65I Curie, 461 Decay time, of scintillation counter, 441
Circular double refraction, 364 Confidence interva1(s), s, 463. 739 Current(s). in anodic stripping voltammetry, 635; control Dccibel,68
Circularly polarized radiation, 361-364 Conjugation, effect upon absorption, by chromophores, of. with operational ampIilIcn, 63; dark, 138 <_ also DecinormaI calomel c1ectrode,potential of. 5321
Coherence, of electromagnetic radiation, 99, 361 173,174/ Dark currents); diamagnetic, 394; ditrusion, I., 608 Decomposition potential(s), 582
Coherent anti-Stokes spectroscopy (CARS), 273 Constantan, 63 (SH also Diffusion current); in electrochemical cells, Decoupling, of proton spins, 40I
Coherent emission, from laser sources, 120 Constant cathode-potential electrolysis, 585-587 527-531; errors in measurement of, 60; faradaic, S05; Dedicated computer(s), 85
Coincidence counter(s), for beta assay, 466 Constant-<:Urrent coulometric analysis, 581 (see also ion, mass spectrometry, 484; measurement of, 28, 30, Degassing system(s), for HPLC, 691
Collector(s), for trace constituents, radiochemical analysis, Coulometric titrations) 35, 59, effect of meter resistance upon, 30, with a null Degeneracy, of molecular vibrations, 219
468 Constant-<:UrreDtelectrolysis, 585, 589 instrument, 33; nonfaradaie, S05; in a parallel circuit, Degrees of freedom, in molecular motions, 217
Collector layer, of bipolar transistor, 45 Constant current source(s), 64 10; residual, polarOJl'llpbic wave, 608; in a series Delocalization, of electrons, 170, 173
Collimating lens(es), for monochromators, 124 Constant-potential coulometry, 581 circuit, 9; sine-wave, 12;SlJJ!UDit,
630;voltammetric, at Demodulation, 73
Collimation, of X-rays, 436 Constant voltage source(s), 63 (see also Potentiostats) stationary electrodes, 623 Depletion layer, in semiconductor diode(s), 44
Column(s), for gas-liquid chromatoJl'llphy, 719; for Constructive interference, 94 Current amplification, with transistors, 44 Depolari=(s), 587
HPLC, 694; for liquid chromatOJl'llphy, 690 Continuous background radiation, of bigh-energy sources, Current control, with operational amplifiers, 64 Derivative curve(s), potentiometric, 561; thermoJl'llvimet-
Column chromatOJl'llphy, 667, 690; (see also Adsorption 337 Current density. effect, on electrolytic deposits, 587, on ric,657
chromatoJl'llphy, Gas-Iiquid chromatOJl'llphy, Gas- Continuous ftowanalyzer(s), 199 kinetic polarization, 530 Derivative polarogram(s), 627
solid chromatosraphy, Gel chromatosraphy, High Continuous radiation, soun::csof. 110, 117 Current elliciency, 592; for coulometric titrations, 595 Derivative spectrophotometry, 192
performance liquid chromatography, Ion exchange Continuous soun:c correction method, atomic absorption Current gain, ofbipotar transistor, 46 Desa1ting, by aeJ dIromatoJl'llphy, 704
chromat0Jl'llphy, Liquid-solid chromatosraphy, Current maxima, linear sweep voltammetry, 623; in Descending development, planar cliromatosraphy, 706
~,321
Partition chromatoJl'llphy) Continuous spectra, 109; X-ray, 427 poIarosrapbic waves, 616 Destabili7.ation of d-orbitats, 177
Column performance, chromatoJl'llphic, 679 Control gaP. for spark source, 339 Current rectifier(s), semiconductor, 44 Destructive interference. 94
Column resolution, R•• chromatograpiuc, 677 Controlled electrode-potential electrolysis, 586, 590, 5911; Current-time intesrator(s), coulometric ana1ysis, 593 Destructive method(s), of neutron activation analysis, 467
Common terminal, of bipolar transistor, 46 pOwer supply for, 589 (see also Potentiostat) Current-voltage curve(s), po1arOJl'llpbie, 607 (see also Detection limit(s),atomic spectroscopy, 3191.3321.3SOt; de
Compensator(s), for refractometers, 355 Control unit, of computer, 82 PolaroJl'llpbic waves); for semiconductor diodes, 44 arc emission spectroscopy, 347t; neutron activation
Competing~uilibria, effectupon e1ectrock potential,521,522 Convection, material transport in solution by, 529 Current-voltage relationsbips, during electrolysis, 581-587 analysis, 470
~
titrations with, S66; as protective reagent, atomic Electrolytic cell(s), SOl (see also Electrochemical cells);
Detector(s), for electromagnetic radiation, 137-145; for Dipole change(s), during molecular rotation and vibration, absorption spcctroscopy, 322 operation, at fixed applied potential, 581
ftuorometers, 290; for Fourier traDsform spectrom- 211 . EDT A titration(s), mercury e1ectrode for, S40 EIectromagnet(s) for nuclear magnetic resonance spec-
eters, 251; for gas-Iiquid chromatOJll'llphy, 720-722; Direct current(s), 7 (see also de Currents) Eddy diffusion, zone broadening by, 674 trometers, 387
for HPLC, 694; for nuclear magnetic resonance Direct potentiometric measurements, 55S-559; limitations Elfective bandwidth, of fi1ters, 136; of monochromators, ElectrOlllllgDetic ba1anoe(s), 655
spectrometers, 388, 389; for radioactivity, 465; selec- to, 557, 559; of pH, 558 132-134 . EIectromagnetic radiation, absorption of, 105 (see also
tive, gas-Iiquid chromatography, 729; thermal, Direct probe inlet(s), for mass spcctrometen, 479 EfIiciency, column, chromatngrapbic, 679; quantum, 284 Absorption of electromagnetic radiation), in a mag-
143-145; for X-rays, 438-442 Discriminator(s), pulse height, 146; for X-ray spec:trom.. Electrical conductance, 7 (see also Conductance) netic field, 381. in X-ray region, 431; coherent, 99;
Detector noise, effect on measurement of T, 1591 (see also eters, 443 Electrical current(s), 7..:s (see also Current) detectors for, 137-145; diffraction of, 95-99; disper-
Noise) Dispersing element(s), for monochromators, 124 (see also Electrical double layer, nonfaradaic currents and, 506 sion of, 102; emission of, 109; energy of,IOO; filters for,
Deuterated chloroform, as solvent, nuclear magnetic Difl'raction gratings, Prisms) Electrical force(s), on ions, 645 135; interactions with matter, 101-109; monochro-
resonance spectroscopy, 390 Dispersion of electrotnagnetic radiation, 102, 354; by Electrical potential, V. 7 (see also Potential) matic, 93; particle properties of, 100; polarization of,
Deuteration, to aid in resolution of complex· nuclear gratings, 130; by prisms, 127; wavelength dependence Electrical power, P, 7 104, 359; reflection of, 103; refraction of, 103;
/Dagnetic resonance spectra, 402 ' of. 354 Electrical signal(s), filters for, 23 scattering of, 104 (see also Raman spectroscopy,
Deuterium discharge lamp(s), 117 Dissociation, of excited molecules, 284 Electrical vector, of radiation, 92 Scattering); transmission of, 102, through anisotropic
Dexlrorolalory component, of plane-polarized nidiation, Dissociation processes, in ftames, 322, 330; and ftuores- Electricity, laws of, 8 media, 357; velocity of, 93, 353; wave properties of,
364 cence,284 Electroana1ytical chemistry, SOI~53 (see also specific 92-99
Diatnagnetic current(s), 394 Distribution oocflicient(s), KD, ion exchange, 700 methods) Electromagnetic spectrum, 100
Diamagnetism, 281 Distribution constant(s), x., gel chro/Datography, 704 Electrochemica1 cell(s). 8, SOI-S07; alternating currents in, Electrometer(s), 61; in ftame ionization detecton, 721
Diatomaceous earth(s), as solid support, gas-liquid Divider(s), voltqe, 9 S05 (see also Alternating current); conduction in, SOl; Electron(s). pairing of. 281; thenna1 excitation of, 42;
chro/Datography, 719 Division, with basic inverting amplifier circuits, 65 de currents in, 504; effect of concentration upon, 517; thermal motion of. in e1ectrica1 circuits, 71 (see also
Dielectric(s), 14 Doping, of semiconductor crystals, 42 irreversible, S07; with liquid junctions, S02; material 10hnson noise)
Diethylene glycol adipate, as stationary phase, gas-Iiquid Doppler broadening, of atomic emission lines, 329; in transport in, 529; nonfaradaic currents in, S05, 506; Electron beam source(s), for X-rays, 427
chrematography, 7251 hoUow cathode lamps, 317; of spectral lines, 307 polarization of, 528-531; potential of, S07, calcu- Electron capture, 431; nuclear. 460, 461
Diethyl ether, as solvent in visible-ultraviolet regio'ns, 1891 d- Orbital(s), absorption of radiation involving, 175; lations involving, 524; revcrsibility of, S07; schematic Electron-capture detector(s), gas-Iiquidchro/Datography,
Difference amplifier(s), 51; minimization of tn¥tsducer splitting of energies of, In representation of, 504 721
noise with, 73 • d-spacing(s), of X-ray diffraction patterns, 452 Electrochromatography, 690, 709-711 (see also Electronic circuit(s), 42-416
Differential absorption analysis, 194-197, 1951 Double-beam instrument(s), for absorption spectroscopy, E1ectrophoresis) • Electronic energy level(s), m01ecular, 107, splitting of, 414
Differential amplifier(s), 61 ISO, 185; foratomic absorption spectroscopy, 317; for Electrode(s), counter, 626; dropping mercury, 606 (set also Electronic noise, 68 (see also Noise)
Differential pulsed polarography, 627~3O Fourier transform spectroscopy, 251 Dropping mercury electrode); for e1ectrochemical Electronic transition(s), induced by absorption of radi-
Differential refractive index detector(s), for HPLC, 696 Double bond(s), as chromophore, 1731 (see also Alkenes); cells, SOl; for electrogravimetric analysis, 588; for ation, 169-180
Differential scanning calorimetry, 655, 657, 659 effect of tnagnetic fjelds upon, 395; molecular orbitals emission spectrost:opy, 338; gas-sensing, 551; glass, Electronic voltmeter(s), for rneasuranent of p-values, 554
Differential thcrtnal analysis, 655, 657~1 of, 170; polarographic behavior of, 621 S40 (see also G1us electrode); indicator, 539; liquid Electron impact ionization, for mass spectrometry. 480,
Differential thcrmogram(s), 657 Double bond rqion, infrared spectrum, 232 membrane, 547; membrane, S4O; for polarographic 488,490
Difl'erential titration(s), 572 Double focusing mass spectrometer(s), 482; resolving analysis, 606 (see also Dropping mercury electrode); Electron microprobe method, of X-ray analysis, 452
Diffe.rentiation, of input signals, 66 power of, 4891 for potentiometric titrations, 563, 567, 571; reference, Electron multiplier tube(s), for tirne-of-ftight mass spec-
Diffraction, of electromagnetic radiation, 95; by gratings, Double-monochromator system(s), 134 531 (see also Reference electrode); rotating platinum, trometer, 482
129; of X-rays, 433, by crystals, 4371 Double refraction, by anisotropic crystals, 359; circular, 625 (see also Rotating platinum electrode); solid, Electron paramagnetic resonance spectroscopy, 412 (see
Diffraction grating(s), 128 (see also .Gratings) 364 voltammetrywith, 622; standard hydrogen (SHE), 512 also Electron spin resonance spectroscopy)
Diffusion, eddy, 674; longitudinal, 675; /Daterial transport Doublet state(s), 281 Electrode calibration method, for direct potentiometric Electron spectroscopy, 453
in solution by, 529, 609 Drain lead(s), for FET transistors, 48 measurement, 556 Electron spin resonance (ESR) spectroscopy, 377. 412-415
Diffusion coeflicient(s), chro/Datographic. 675; polaro- Drift tube, of tirne-of-ftight mass spectrometer, 482 Electrode composition, effect on Idnetic polarization, 530 Electron volt, eV, 100
graphic, 614 Drive mechanism, Fourier transform spectrometer, 249 Electrodeless discharge lamp(s), for atomic ftuorescence Electrophoresis, 709
Diffusion current(s), i., 608; equation for, 612; evaluation Dropping mercury electrode(s), DME, 606, 607, 612, 618; analysis, 333 Electrophoretic effect, upon ionic mobility, 645
of,619 capillary constant for, 614 Electrode potential(s), 512-524; calculation of, 519-527; Electrostatic force(s), material transport in solution by, 529
Diffusion potential, glass electrode, 543, S44 Drop time, I, for dropping mercury electrode, 614 effect, of competing equilibria upon, 521, of concen- (see also Mobility, de Currents)
Digital-to-analog convcrtcr(s), for computers, 83 i Dual wavelength spectrophotometer(s), 193 tration upon, 517, of temperature upon, 518; IUPAC Elemental analysis, with nuclear magnetic resonance
Dilution method(s), radiochemical, 459 (see also ~sotope Duane-Hunt law, 428 definition of, 516; limitations to use of, 522-524; spectroscopy, 408
dilution); for quantitative X-ray analysis, 449 Dye IaseJis), 123 measurement of. 513; sign convention for, 515 Elliptica11y polarized radiation, 361. ~3
Dimethylformamide, as solvent for infrared spectropho- Dynode(s), of photomultiplier tube, 140 Electrogravimetric analysis, 581, 587-590; apparatus for, Ellipticity, 372
tometry, 227 587; applications, 5911; with constant cathode poten- Eluent(s), linear chromatography. 669
Dinonyl phthalate, as stationary phase, gas-Iiquid· chro- Eagle mounting, for emission spectrograph, 344 Eluent strength, liquid-101id chromatosraPhY. fIJ7
tial, 585, 589; with constant current, 589; variables
matography, 7251 E. band, in ultraviolet spectra of aro/Datic compounds, affecting, 587 Elution, 670; isocnItic, 691; of solutes, linear ~
Diode(s), 43 174,1751 graphy, 670
Electrolysis, current-voltage relationships during, 581-587
Diolefin(s), absorption by, 1741 (see also Alkenes) Ebert mounting, of reftection grating, 183 Elution chromatography, 669; ion excllanp, 699
Electrolyte(s), supporting, polarographic analysis, 608 (see
Dioxane, as solvent, for infrared absorption spectropho- Echellette grating(s), 129 Emission of electromagnetic radiation, 109 (see also
also Supporting electrolyte)
tometry, 227; in visible-ultraviolet regi9ns, 1891 EDT A, coulometric generation of, 598; potentiometric
Atomic emission apcctroICOpY, Emission tpeCIroI- Field clfect (PET) transistor(s), 48 Formal potential(s), 523, 746 Gas streams, monitoring of, 224
copy, Fluorescence spectroscopy) Field ionization sourcc(s), for _spectrometers; 481 Formation constant(s), and electrode potentials, 521 Gatc(s), for FET transistors, 48
l!Jnission profile(s), 311 F'JC1dsweep generator, nuclear magnetic resonance spec- Fortran, computer language, 84 Gaussian distribution(s), 738; of solutes in elution bands,
Emission spectra, 109, 337; encrsY level diasrams for, 304; trometer, 388 Forward biasing, 44 672, 703; of X-ray photon energies, 442
ofpses, 110 F'Jltcr(s), for current Iluctuations, 53; for electrical signals, Forward current transfer ratio, 46 Geiger tubc(s), for detection of X-rays, 439
Emission spectroscopy, 3021, 336, 351; (see also Atomic 23; for cIcctromapetic: radiation, 135, 290 (_ also Fourier transformation(s), 94; of interferograms, 247 Gelatin, as maximum suppressor, polarographic analysis,
emission spectroscopy); appli<:atiolll, 347-351; Interference filters); forelimination of electronic noise, Fourier transform spectroscopy, infrared, 241-254; instru- 616
Fourier transform, 253; instruments for, 342 73; series resonant, 24; for X-rays, 435 ments for, 249-252, resolution of, 248; nuclear Gel chromatography, 668t, 702-705
Emitter-foUower amplifier(s), SO Filter photomctcr(s), infrared, 224 magnetic resonance, 409-412 Gel layer(s), glass membrane, 542, 543
Emitter layer, of bipolar transistor, 45 Filter wedge(s), 220 Four-Icvellaser systcm(s), 122 Germanium detcctor(s), lithium drifted, 441
Enantiomer(s), 366 Fingerprint region, infrared spectrum. 232 Fragmentation pattcm(s), mass spectrometry, 488, 494 Germanium scmiconductor(s), properties of, 42
IlIId point(s), amperomctric, 636, 639; conductomctric,· F'lI'Cbrick, as solid support, gu-Iiquid chromatography, 719 Free energy, change, fiG, in electrochemical cells, 511 G18II, optical characteristics of, 126t
648; coulometric, 595; potentiometriC, 561-563 . First-order indicator elcetrode(s), 539 Free induction decay signal, Fourier transform nuclear Glass elcetrodc(s), 540-547; acid error, 546; alkaline error,
Energy, of electromagnetic radiation, 100 First-order nuclear magnetic resonance spectra, 399 magnetic resonance spectroscopy, 410 544; asymmetry potential of, 542, 544; membrane
Eflergy dispersive X-ray spectrometers, 434, 443, 444, 445: F1amc(s), factors affecting, 308-312; temperature of, 309t Free radical(s), observation by ESR spectroscopy, 414 composition of, 542; pH measurement with, 558;
Energy level(s), in a magnetic field, 378; molecular, 105, 107 Flame absorption spectroscopy, 304 (SI!I! also Atomic Free solution method, electrophoresis, 709 potential of, 543; resistance of, 543; role of water in
Energy level diagram(s), for emission spectra, 304 absorption spcctroscopy) Frequency, ., effect of, upon Ilicker noise, 71, upon functioning of, 542; for univalent cations, 546, S48t
Energy shift(s), Raman spectroscopy, 265 Flame emission spectroscopy, 303 (SI!/! also Atomic Johnson noise, 71; of electromagnetic radiation, 93; Globar(s), 118
I!Ijvironmcntal noise, 70, 71; elimination with filters, 73 emission spcctroscopy) Larmor, 381; of sine-wave current, 12 Golay dctcctor(s), 145
~ analysis, with ccntrifupl fast analyzers, 202 Flame Iluorcsccncc lpectroICOpy, 304 (_ also Atomic . Frequency lock system(s), for nuclear magnetic resonance Goniometer, 436
Epoxidc(s), polarographic behavior of, 621 Iluorescence spcctroscopy) spectrometer, 388 Gradient elution, 691
Equivalent circuit(s), 11 Flame ionization detcctor(s), gu-Iiquid chromatography, Frequency response, of inverting circuits, 58 Gran plot(s), for end point detection, 562
Equivalent conductance, A, 644; clfect of concentration. 721 Fresnel rhomb, 373 Grating monochromator(s), 128, 220; advantages over
upon, 645, 646t Flame photometry, 302t (SI!I! also Atomic emission Frictional forcc(s), affecting ionic mobility, 645 prism monochromators, 131; diffraction by, 129;
Error(s), in analytical chromatography, 683; in mcasurc-i spectroscopy) Fucl(s), for atomic spectroscopy, 3091 dispersion of, 130; resolving power of, 131
ment of current, 60; propagation, in calculations, i Flame plasma(s), 309 Fuel regulator(s), for Ilame spectroscopy, 313 Grating normal, 129
737-743 F1ame.profile(s), absorbance, 312; emission, 311 Functional group analysis, in ultraviolet-visible regions, Grazing angle, 354
ESCA spectroscopy, 453 Flame spectra, 304-308 187; polarographic, 621 Ground connection(s), in amplifier circuits, 49
JlC,IIanol,as solvent in visible-ultraviolet regions, 189t Flavanol, as Iluoromctric reagent, 294, 295t Furnacc(s), for differential tfcrma1 analysis, 6S8; for Grounding, minimization of environmental noise by, 73
Ethylenediamine d-tartrate, dilI'raction of X-rays by, 437t n
Flicker (1/ noise, 70, 71; clfect upon measurement of T, thermogravimetric analysis, 655 Ground slllte, IOS
Europium complcx(es), as chemical shift rcaacnts, proton 159t, 161; climination with filters, 73 Fused silica, optical characteristics of, 126t Ground, virtual, 58
nuclear magnetic resonance spectroscopy, 402 Flint glass, optical characteristic. of, 126t Group frequencies, infrared, 230
&chanac rcaction(s), on glau membrane, 542 F1uoborate ion,liquid membrane dectrodc for, 551t Gain, a, of operational amplifier, 57 Gyromagnetic ratio, 381 (s« also Magnetogyric ratio)
IlIdtation spectra, Iluorcsccncc spectrOmetry, 290 F111OfC1CC11CC, 110, 280; cOm1alioa, with structure, 286, Gallium, doping of semiconductors with, 42
IlIdted molecular ltate(s), 102, 105; deactivation of, 282; with transition types, 285; factors affecting, 284; Gallium arsenide, as lasing medium, 119 Hadamard transform spectroscopy, 254
lifetime of, 281; relaxation of, 109, 169 mechanism of, 283; use of, in planar chromatography, Galvanic cell(s), SO I (SI!/!also Electrochemical cells) Half-<:ell potential(s), 512 (see also Electrode potentials)
~usion chromatography, 702 (SI!I! also Gel 708; X-ray, 432 Galvanometer(s), 28; electrical measurements with, 30-35 Half-<:eUreaction(s), SOl
, chromatography) F1uorcsccncc spectrometry, 280-294; applications, Gamma ray(s), y, 431,460; measurement of, 466; properties Half-life, of a radioactive isotope, 462
Exposure, E, of photographic emulsions, 345 293-294; atomic, 3021 (SI!I! also Atomic Iluorescencc of,459t Half-shadow dcvice(s), for polarimeter, 368
External conversion(s), deactivation by, 284 spectroscopy); instruments for, 289-293 Gamma ray spectrometer(s), 466 Half-wave potential, E1/2' polarographic wave, 608, 610
External lock system, nuclear magnetic resonance spec- Fluorescent intensity, F; clfect of concentration upon, 288 Gas(es), emission of electromagnetic radiation by, 110; HaU probe, for mass spectrometers, 486
trometer, 388 (_ also Fluorescence spectrometry) infrared analysis of, 226 Halogen(s), substituted. effect upon Iluorcsccnce, 287
Extinction, E, I SOt Fluorescent spectra, 290; X-ray, 427, 429 Gas amplification, 439 Hanging drop electrodc(s), 634
Extinction coefficient, k, "I50t Fluoride ion, solid-state electrode for, 551, 552t Gas chromatographic inlet system(s), for mass spectrom- Harmonic motion, 213
~ordinary ray, lie, 359 Fluorine, nuclear magnetic resonance of, 408 eters,479 Harmonic osciUator(s) 212-217
Fluorite, optical characteristics of, 126t Gaseous discharge tubes, for atomic absorption spectros- Hartmann diaphragm, 344
Find, F, 16 F1uorolubc, 228 copy, 317 Hcat detcctor(s), for infrared radiation, 143-145
Faradaic current(s), S05 (sl!e also de Currents); ac F1uoromctcr(s), 289, 291 Gas-fi1lcd detector(s), for X-rays, 438 Height equivalent of a theoretical plate, H, 671; evaluation
polarography, 632 F1uoromctric rcaient(s), 293, 295t Gas lascr(s), 119, 122 of, 673
Faiaday, F, 7, 591 Fluorometry, 280 (_ also F1uorcsccnce spectrometry) Gas-Iiquid chromatography, 6681, 716-733; appli<:ations, Helium, as carrier, gu-Iiquid chromatography, 718
Far infrared region, 210t; Fourier transform spectrometry Flow ratc(s), F, chromatographic, 682t, 716; rcgulation of, Heliumfncon lascr(s), as Raman source, 268
722-731; comparison with high performance liquid
in, 252 718 Helmholtz coil(s), 413
chromatography, 705t; interfacing of, with infrared
Fast analyzcr(s), photomctrie, 200 Focal plane, of monochromator, 124 Henry, H, 16
spectroscopy, 730; with mass spectroscopy, 729;
Fatigue, in banier-layer ceUs, 139 Focal-p1anc dispersion, of prism monochromators, 127 Hertz, Hz, 12, 93
temperature programmed, 728
Feedback, in amplifier circuits, 5 I, 55 /-<>rbital(s), absorption of radiation involving, 175 Hexane, as solvent in visible-ultraviolet regions, 1891
Gas-sensing membrane electrodc(s), 540, 551-554
FeUgett advantage, 242 Force constant, simple harmonic motion, 213 Hcyrovsky, J., 606
Gas-solid chromatography, 6681, 716, 731
High absorbance method, di1ferentia1 absorption analysis, Inductor(s), 14 Inverting amplifier circuit(s), mathematical operations on Kinetic po1arization, 530, 581, 582
InfiDite dilution, equivalent con4uction at, 645 (see "also input signa1s with, 6~7 Kinetic theory, elution chromatography, 670 (see also Rate
195
High-pass filtcr(s), 23; applications of, 46; elimination of Equivalent conductance) Inverting circuit(s), incorporating operational amplifiers, theory)
noise with, 73 Infrared absorption spectroscopy, 2IG--254; Fourier trans- 57 Kirkholl"s laws, 8, 10
High performance liquid chromatography (HPLC). form, 241-254; Hadamard transform, 254; instru- Inverting terminal(s). of operational amplifiers, 56 Klystron tube(s), 413
690-696; columns for, 694; comparison with ps-Iiquid ments for, 219-226; interfacing of, with ps-Iiquid Iodide ion, solid-state electrode for, 552t KRS-5 (TIBr-Tl\). optical characteristics of, 126/
chromatography, 7051 chromatography, 730; qualitative applications, Iodine, charge-transfer complex with iodide, 179 K series, of X-rays, 428
High-resolution nuclear magnetic resonance spectrom- 23G--237; quantitative applications, 237-241; sample Ion current(s), in mass spectrometers, 484
etcr(s), 386 preparation for, 226-228 Ion exchange, on glass membranes, 540; in liquid mem- Laboratory recorder(s), 77
Hitachi Modell~ spectrophotometer, 185 Infrared radiation, condition for absorption' of, 211; branes,547 Laminar flow burner(s), for atomic spectroscopy, 313
Hole(s), in semiconductors, 42 detectors for, 140, 143-145; sources for, 118 Ion exchange chromatography, 668/, 690,699-702 Lanthanide ion(s), absorption by, 175
HoUow cathode lamp(s), 316; as line sources, 118; pulsed, Infrared spectra, comparison with Raman spectra, 265; Ion exchange resin(s), as stationary: phase, thin layer Larmor frequency, 381
333 computer search systems for, 223 (see also Catalogs); chromatography, 706 Laser(s), 118-123; as source, fluorescence spectrometry.
Hooke's law, 213 overtone lines in, 217 Ion gun(s), for mass spectrometers, 480 290
HPLC, 690 (see also High performance liquid Infrared spectrum, regions of, 210/, 23G--233 Ionic conduction, 8 . Laser excitation, emission spectroscopy, 339
chromatography) Inlet system(s), for mass spectrometers, 479 Ionic spectia, 337 Laser-fringe reference system, Fourier transform spec-
Hundred percent T adjustment, absorption spectroscopy, IDJIer cone, of flame, 311 Ionic strength, p, S07 trometer,2SO
151 (see also One hundred percent T adjustment) Inorpnic analysis, f1uorometric, 293; mass spectrometric, Ionization, elI'ect on atomic emissi6n intensities, 330; Laser microprobe(s), 339
Hydrocarbon(s), retention times for, 7241 498; polarographic, 620 electron impact, 488; in ftames, 323, 3241 Lasing media, 119
Hydrogen, isotopic abundance of, 490/; overvoltage elI'ects InoIpnic species, absorption in the ultraviolet by, 175; Ionization chamber(s), for detection of X-rays, 439, 440 Lead(l\), solid-state electrode for, 552/
with,531/ Raman spectra for, 270 Ionization suppressor(s), atomic absorption spectroscopy, Lead semiconductordetector(s), forinfraredradiation, 141
Hydrogen bonding, effect on retention times, ps-Iiquid Input impedance, of operational amplifiers, 55, 57 324 Levoro/alory component, of plane-po1arized radiation, 364
chromatography, 724 inPUt/output system(s), for computers, 83 Ion source(s), for mass spectrometry, 479 Ligand(s), effect on absorption maxima for transition-
Hydrogen discharge Iamp(s), 117 Input signals, mathematical operations on, with oper- IR drop, 527 " metal ions, 178/
Hydrogen electrode(s), 512 ational amplifiers, ~7 Iron-55, as source for X-rays. 431 Ligand field strength, 179
Hydrogen emission line(s), as source for refractive index Input transducer(s), 4 Iron(lI), charge-transfer complex with o-phenanthroline, Light amplification, in lasers, 120
measurements, 354 Instrumental deviation(s), to Beer's law, 153, 157-163 179; coulometric oxidation of, 59i . Limiting current(s), polarographic wave, 608
Hydrogen f1ame(s), temperature of, 309/ Instrumental noise, influence of, upon spectrophotometric Iroo(lII), charge-transfer complexes of, 179 Linear absorption coetIicient(s), X-ray, 432
Hydrogen flame detector(s), ps-Iiquid chromatography, measurements, 157 (see also Noise) Iron spectrum, as reference for emidon spectroscopy, 348 Linear elution chromatography, 669
721 Integrated circuit memory, of computer, 83 Irreversible reaction(s), polarographic waves for, 612 Linear sweep oscillator(s), for nuclear magnetic resonance
Hydrogen ion(s), coulometric generation of, 597, 5981 Integration, of input signals, 65; photographic, 342 Irreversibility, electrochemical, S07 spectrometers, 385
Hydrogen/oxygen coulometer(s), 593 Integrator(s), for nuclear magnetic resonance spectrom- lsocratic elution, 691 Linear voltage divider(s), 34
Hydrogen stretching region, infrared spectrum, 230 eter,389 lsotherm(s), adsorption, 669 Linear voltage sweep voltammetry, 623
HydroquinODe(s), polarographic behavior of, 621 Intensity, of eIectromqnetic radiation, 93; of Raman Isotope abundanee(s), measurement with mass spectrom- Line broadening, elI'ect of temperature UPOn. 311; of
Hydroxide ion(s), coulometric generation of, 597, 5981 peaks, 268; relative, of spectral lines, 345 etry, 498; naturally occurring, 490/ nuclear magnetic resonance spectra, 384
2-(o-Hydroxyphenyl)-benzoxazole, as f1uorometric re- Interfacing, of instruments to computers, 79-86 Isotope dilution, 459, 471 Line source(s), for optical spectroscopy, 118-123
agent, 2951 Interference, of electromagnetic radiation, 94; path-length Isotope peak(s), in mass spectra, 489 Line spectra, 106. 109, 337; characteristic X-ray, 428
8-Hydroxyquinoline, fluorescence of, 287; as f1uorometriC determination with, 227; of plane-polarized radiation, Isotope ratio(s), determination of molecular formulas Line width(s), of atomic spectra, 307
reagent, 294, 295/; as protective reagent, atomic 361 from, mass spectrometry, 493 Lippich prism(s), 369
absorption spectroscopy, 322 Interference filter(s), 135; infrared, 220 Isotropy, 357 Liquid(s), infrared spectrophotometry of, 228
Hyperfine splitting, in ESR spectra, 414 Interference wedge(s), for electromagnetic radiation, 136" IUPAC convention, for electrode potentials, 515 Liquid chromatograpby, 69G--711; detectors for, 694-696
Hypsochromic shift(s), of absorption spectra, 172 Interferogram(s), 246; Fourier transformation of, 247; Liquid junction potential(s), S02, 526; effect upon direct
sampling of, 249 Johnson noise, 70, 71, 158; elI'ect upon measurement of T, potentiometric measurements, 555, 557
Ice, refractive index of, 359/ Interferometcr(s), 245, 354 158; elimination with filters, 73; hi infrared spectro- Liquid membrane electrode(s), 540, 547, 551/
IIkovic equation, 612 Internal conversion, deactivation by, 283; nuclear, 461 photometers, 163; in photoelectric detectors, 138 Liquid phase(s), for ps-Iiquid chromatography, 719
Image displacement refractometer(s), 356 Internal lock system, nuclear magnetic resonance spec- Joule, 7 i Liquid scintillation counter(s), 441
Impedance, Z, 22; in operational amplifiers, SO, 59 trometer, 388 Junction(s), transfer of panicles across! 71 (see also Shot Liquid-solid chromatography, 690 (see also Adsorption
Incoherent radiation, e1ectromagnetic, 99; from spon- Internal negative feedback, in operational amplifiers, 55 noise) chromatography)
taneous emission of 1asing media, 120 Internal reflectance spectroscopy, 228-230 Junction potential(s), S02; (see also Liquid junction Lithium, as internal standard for alkali-metal ions, 327
Indicator e1ectrode(s), 531, 539-554 Internal standard(s), for chromatographic analysis, 683, potentials) Lithium drifted detector(s), for X-rays, 441
Indium, doping of semiconductors with, 42 726; for emission spectroscopy, 327, 348; for nuclear Lithium fluoride, di1fraction of X-rays by, 4371; optical
Induced po1arization, of molecules in solution, S06, 647 mapetic resonance spectroscopy, 392; for X-ray Katharometer(s), 720 characteristics of, 1261
Inductance, 14, 16 i analysis, 449 K-<:apture, radioactive decay by, 431 Lithium ion, glass membrane electrode for, 547, S48/
Inductively coupled plasma (lCP) emission spectroscopy, Intersystem crossing(s), 284 Ketone(s), absorption by, 173/ (see also Carbonyl groups) Lithium tantalate, use in pyroelectric detectors, 145
340; detection limits for, 3SO/ Intrinsic zone, of lithium drifted silicon detector, 441 Kieselguhr, as solid suppon, gas-liquid chromatography, Littrow prism(s), 126, 183, 344
Inductive reactance, XL' 22 Inversion, of sucrose, 369 719 Lock-in amplifier(s), 75
Meter resistance, effect, on current measurements, 30, on MoIe!:uIarspectra, in flames, 308 Nodal plane, of lr-bonc!,170
Logarithmic transformation, of input signals, 67
Longitudinal diffusion, zone broadening by, 675 voltage measurements, 31 Molecular structure, and fluorescence, 287 Noise, effect on precision of spectrophotometric ~easure-
Longitudinal relaxation, 383 (see also Spin-lattice Method of ultimate precision, differential absorption Molecular vibration(s), 212; quantum treatment 0(, 215; ments, 157-162; in electronic circuits, 68-77
relaxation) analysis, 1951,196 degenerate, 219; normal modes of, 218
Nonaqueous solvent(s), potentiomet1jc titrations involv-
Low absorbance method, differential absorption analysis, Methylcyclobexane, as calibration standard for refractom- Molecular weight, determination by ~ spectrometry, ing,S71
1951,196 etry,356 491
Nonbonding(lI)eIectron(s), 170,171;tlansitionsinvolving,
Low-pass filter(s), 23 Methyl red, as maximum suppressor, polarographic Monochromatic radiation, 93; dependence on Beer's law 171, 172, 175,285
L section filter(s), 53 analysis, 616 upon, 155
Nondestructive method(s), of neutron activation analysis
L series, of X-rays, 428 Michelson interferometer, modulation of infrared radi- Monochromator(s), 123-135; Czerney-Turner, 318; effec- 467,468 '
ation with, 245 tive band width of, 132-134; for emission spectros- Nondispersive spectrometer(s), infrared, 224; X-ray, 444,
M + I peak(s), in mass spectra, 489 Microammeter(s), 28 copy, 343; for spectrofluorometers, 290; stray radi- 446
Magnesium ion, liquid membrane electrode for, 5511 Microcomputer(s), BO, 85 ation in, 134; for X-rays, 436, 456 Nonequilibrium mass transfer, zone broadening by, 675
Magnet(s), for nuclear magnetic resonance spectrometers, Microcurie, 461 Mull(s),228
Nonfaradaic conduction, in electrochemicalcells, 506, 646
387 Microelectrode(s), 606 Multichannelanalyzer(s), optical, 142; X-ray, 444 Nonfaradaic current(s), 506, 517; in ac polarography, '632
Magnetic anisotropy, 395 Microfarad, 16 Multielement analysis, by emission spectroscopy, 336; Nonflame absorption spectroscopy, 314; detection limits
Magnetic core memory, of computer, 83 Microhenry,I6 instruments for, 327, 342, 345 for, 3191
Magnetic field(s), nuclear energy levels in, 378 Micrometer, pm, 93 Multiplet(s), first-order, relative intensities of, 4001 Noninverting terminal(s), of operational amplifiers, 56
Magnetic field strength, effecton chemical shift(s), 391, 402 Microprobe(s), laser, 339 Multiplex metbod(s), 242 (see also Fourier transform Nonflame atomizer(s), for atomic spectroscopy, 314
Magnetic moment(s), of elementary particles, 378 Microprocessor(s), 85; in mass spectrometers, 485 spectroscopy)
Nonradiative relaxation, 109(see also External conversion)
Magnetic quantum level(s),transitions among, 377 (see also Microwave radiation, absorption of, by electrons, 412 (see Multiplication, with basic inverting amplifier circuits, 65 Nonresonance fluorescence, 110
Nuclear magnetic resonance) also Electron spin resonance spectroscopy) Multiplicity, of spin-spin splitting, 398, 399 Normal calomel electrode, 5321
Magnetic vector, of electromagnetic radiation, 92 Microwave region, sources for. 413 Normal dispenion, 103 (see also Dispersion)
Magnetogyric ratio, 381 Middle infrared region, 2101;Fourier transform spectros- Nanometer, nm, 93 Normal mode(s), of molecular vibration, 218
Majority carrier(s), of current, in semiconductors, 42 copy in, 253 National Bureau of Standards, as source of standards, for IIJ1II Transistor(s), 45
Mass absorption coefficient(s), X-ray, 432 Migration rate(s), in chromatographic columns, 676 emission spectroscopy, 349,'for refractometry, 356 11- Type semiconductor(s), 42
Mass spectra, 488-491; generation of, 479 Milliammeter(s), 28 Natural frequency, v., of mechanical oscillator, 215 Nuclear magnetic resonance spectra, 385; chemical shifts
Mass spectrometer(s), 477-488; computerized, 485; as Millicurie, 461 Natural gas flame(s), temperature of, 3091 in, 391;chemical shift reagents for, 402;environmental
detector, liquid chromatography, 696;resolving power Minicomputer(s), 80, 84; in mass spectrometers, 485 Near infrared region, 2101 effects upon, 390-404; first order, 399; interpretation
of, 486, 4891 Minority carrier(s), of current, in semiconductors, 42 Nebulizer(s), for atomic spectroscopy, 313 of, 399-409; peak areas of, 389; pulsed, 410; spin-spin
Mass spectrometry, 477-498; interfacing of, with gas-liquid Mixture(s), analysis of, spectrophotometric, 152, 191, Negative feedback amplifier(s), 51, 55 • splitting in, 391, 398
cJ.u"omatography,729; qualitative, 491-496; quantita- polarographic, 619 Negatron(s), 4(fJ (see also Beta particles) Nuclear magnetic resonance (NMR) spectroscopy,
tive, 496-498 Mobile pbase(s), chromatographic, 667 Nephelometric analysis, 104, 294-298; applications, 2971 377-412; applications, 404-409; carbon-I 3, 412;
Material transport in solution, 529, 609 Mobility, of ions in solution, 504, 526, 645 Nernst equation, 517-524
chemical shift parameters for, 392;Fourier transform,
Mathematical operation(s), with binary numbers, 80 Modulation, for elimination of noise, 73; ofhigb frequency Nernst a\ower(s), 118
409-412; instrumentation for, 387-390; measurement
Matrix effect(s), in quantitative X-ray analysis, 448 signals, 244; of output, from hollow cathode lamp, Neutralization titration(s), conductometric, 649;coulomet- of absorption signal, 384; with nuclei other than the
Mattauch-Herzog mass spectrometer, 482 317, from infrared source, 221 ric, 597, 5981; nonaqueous, 571; potentiometric, proton, 408; solvents for, 390
Mean activity coefficient(s),.r. 510 Molar absorptivity, t, 149, 1591; of optically active 568-571 Nuclear magneton, 378
Mechanical characteristics, of optical materials, 1261 substances, 370 Neutrino(s), v, properties of, 4591 Nuclear relaxation, in a magnetic field, 382
Mechanical oscillator(s), properties of, 213 Molar extinction coefficient, ISOl Neutron(s), II, 461; properties of, 4591 Nuclear spin, 377
Membrane indicator electrode(s), S40-554; classification Molecular absorption, 106 Neutron activation analysis, 467-471; sensitivity of, 469 Nuclei, magnetic properties of, 3861
of, S40 Molecular band interference(s), atomic emission spectros- Neutron capture cross sec!ion(s), 469 Nujol,228
Memory unit, of computer, 83 copy, 328 N"lChromewire(s), as source of infrared radiation, 118 Null measurement(s), 33; for absorption analysis, 180
Mercaptan(s), polarographic behavior of, 621 Molecular band spectra, in emission analysis, 312 Nicolprism(~,359,368 Number of theoretical plates, N, 671, 6821;evaluation of,
Mercury electrode(s), dropping, 007 (see a/so Dropping Molecular ellipticity, [9], 372 Nitrate group(s), as chromophore, 1731 673
mercury electrode); for EDTA titrations, 540, 567; for Molecular emission spectra, 110 Nitrate ion(s), absorption in the ultraviolet by, 175; as
electrodeposition, 588; hanging drop, 634 Molecular energy state(s), 107 cathode depolafizer, 587; Liquid membrane electrode Obstruction factor, longitudinal diffusion zone broaden-
Mercury emission lines as source, fluorescence spectrom- Molecular fluorescence spectrometry, 280 (see a/so for, 5511 ing,675
etry, 290; refractive index measurements, 354 Fluorescence spectrometry) Nitrite ion(s), absorption in the ultraviolet by, 175 Octahedral configuration, splitting of d-orbitals in, 177
Mercury pool. as reference electrode, polarographic Molecular formula(s), determination by mass spectrom- Nitrogen, isotopic abundance of, 4901 OH radical, band spectrum of, 337
analysis, 007, 627 etry,492 Nitrogen dioxide, gas-sensing e1ectrode for, 554 Ohm, n, 7
Metallic indicator electrode(s), 539 Molecular ion(s), of mass spectrum, 488, 489 Nitrogen rule, mass spectrometry, 494 Ohmic potential, of electrocbemical cell, 527 (see also /R
Metal-ligand bond(s), infrared activity of, 252 Molecular orbital(s), and absorption of radiation, 170-180 Nitro group(s), as cbromophore, 1731; polarographic drop)
Metal-oxide semiconductor field effect (MOSFET) Molecular rotation [M], 367 behavior of, 621 Ohm's law, 8
transistor(s), 48 Molecular sieve(s), 731 Nitroso group(s), as chromophore, 1731; polarographic Olefin(s), absorption by, 1741(see also A1kenes)
Metal-oxygen bond(s), Raman activity of, 271 Molecular sieve chromatography, 702 (see a/so Gel behavior of, 621 One hundred percent T adjustment, absorption spectros-
Metal vapor lamp(s), as line sources, 118 chromatography) Nitrous oxide, as oxidant, atomic spectroscopy, 3091 copy, 151; effect of noise upon, 157
INDEX 763

One-over-f(llf) noise, 71 (see also Flicker noise) Packing factor, eddy diffusion zone broadening, 674 Phot~Ilb!c detection, of emission spectra, 3"44;of X-ray
Open-loop gain, of amplifien, 58 Padd1e(s), in sample probe, nuclear magnetic resonance clitrractiondata, 45I 61~ clitrusion, ~ (.re~ also Diffusion current);
Operational amplifier(s), 5H7; mathematical operations spectrometer, 390 Photoluminescence, 280 (see also Fluorescence) ~ual, 608, 616, SUDlD1Jt,630; variations in during
upon input signals with, 64-67; terminals of, 56 Pair production, 461 life of drop, 612 '
Ph~~s), 59,125, 180, 326; as detector, HPLC, 694;
Optical activity, 357 (see also Polarimetry, Optical rotatory Paper Chromatopphy, 6681,690, 705, 708 infrared, 224;nonclispersive, 224 , Polaroaraphic wave(s), 607-612; alternating current polar-
dispersion, Circular dichroism) Parallel circuit(s), 10 Photometric ~ysis, 151 (see also Absorption analysis); ~h!, 633; anoclic, 615; current maxima in, 616;
Optical density, D, ISOI;of spectroscopic lines, 345 Parallel resonant filters, 26 filter selection for, 181; instruments for 180' of derivative, 627; effect, of complex formation upon,
Optical material(s), characteristics of, 1261 Paramagnetism, of triplet states, 281 mixtures, 152 ' , 611, of temperature upon, 614; equations for 610'
Optical multichannel analyzer(s), 142 Parent ion(s), of mass spectrum, 488 (see also Molecular Photometric error(s), 153-162; in limited readout instru- half-wave potential for, 608; interpretatio~ of'
Optical rotation, 364; factors affecting, 367; mechanism of, ion) ments, 194-197 608--612;for irreversible reactions, 612; organic, effecd
367 Particle size, effect on scattering of electromagnetic Photometric titration(s), 197-199 of pH, 621; for oxygen, 617
Optical rotatory dispersion, 37G-374; applications, 373; radiation, 104, 296 Photomultiplier tube(s), 138, 140, 290 Polychromatic radiation, effect of Beer's law 155
curves for, 371; instrumentation for, 372 Partition chromatopphy, 6681,690, 698 (see also Liquid Photon(s), capture cross section for, 152; of electromag- Polyeyanomethylsiloxane, as stationary p~, gas-Iiquid
Optical rotatory power, 357 chromatopphy) chromatopphy,7251
netic radiation, 92, 100
Optical spectroscopy, components of instruments for, Partition coefIicient(s), K, 668, 677, 6821; in gas-Iiquid Photon counter(s), optical spectroscopy, 146 Polyethylene glycol, as stationary phase, gas-liquid chro-
114-146 chromatopphy, 717 Photon counting, 438 matopphy, 724, 7251
Optic axis, of anisotropic crystal, 357 Passive computer interactions, 85 Photon dctector(s), 138-143 POlyf:onal acids and bases, potentiometric titration of,
Orbital energies, of d-electrons, 177 Path length, determination from interference patterns, 227; Phototube(s), 138, 139; amplification of signals from, 59
Orbital precession, of electrons, 414 of electromagnetic radiation, b, ISOl Photovoltaic ceII(s), 138 Polymcthylphenylsiloxane, as stationary phase, gas-liquid
Order, of interference, 98, 135; of output from ptinBS, Peak(s), analytical uses of, chromatopphic analysis, 681, chromatopphy, 7251
~ (ll) e1ectr~n(s),effect of magnetic field upon, 395
130 682/; nuclear magnetic resonance spectroscopy, 389, Polymcthylsi1oxane, as stationary phase, gas-Iiquid chro-
Pi (~) or~tal(s), charge density of, 170; transitions
Ordinary ray, no, 359 402; thermal analysis, 658 Involvmg, 172, 174, 175,285 matopphy, 7251
Organic analysis, ftuorometric, 294; mass spectrometric, Peak amplitude, of sine-wave current, 12 Picofarad, 16 Polytriftuorpropyl~thylsiloxane, as stationary phase gas-
496; polaropphic, 620 Peak-to-peak current(s), 12 liquid chromatopphy, 7151 '
Planar chromatopphy, 667, 690, 705 (see also Thin-layer
Organic chromophore(s), 172, 1731,1741 Pellicular packing(s), for column chromatopphy, 694, Population inversion, in lasers, 120
chromatopphy, Paper chromaiopphy)
Organic functional group(s), adsorption characteristics of, 701 Plane polarization, of electromagnetic radiation, 104 Po~pak(s), use in gas-solid chromatopphy, 731
697 Peptide(s), fractionation of, by gel chromatopphy, 704 Planck constant, h, 100 Posltron(s), p+ , properties of, 4591,460
Organic halogen group(s), polaropphic behavior of, Perchlorate ion, liquid membrane electrode for, 55II Plasma, 303 Potassium brOmide,.optical characteristics of 1261
621 Period, p, of electromagnetic radiation, 93; of sine-wave Plasma methods, emission spectroscopJf, 336 Potassium chloride, conductance of solutions' of 6481
Organic solvent(s), effect upon atomic absorption and current, II Potassium ion, glass membrane electrode for 547 5481'
Plasma source(s), 340 (see also Inductively coupled plasma
emission spectra, 325 Periodicity, in time domain spectrum, 243 source) liquid membrane electrode for, 5SO, 5511' , ,
Orientation polarization, of molecules in solution, 506, 647 Permanent magnet(s), for nuclear magnetic resonance Plasma lemperature(s), emission Spectroscopy, 338 Potential(s), cathode, changes during electrolysis, 583-585;
Oscillator(s), anharmonic, 217; harmonic, 212; for nuclear spectrometers, 387 P1ate(s),theoretical, 671 control. ?f, with operational amplifiers, 63; de-
magnetic resonance spectrometers, 385 Peroxide(s), polaropphic behavior of, 621 Plate calibration curve(s), emission Spectroscopy, 345 comJlOS1tiO~582; electrical, V, 7; electrode, 515; of
Oscillopphic polaropphy, 630 p-Function(s), 539 Plate development, thin layer chromatography 707 electrochemical cells, S07; effect of concentration
Oscillometry, 647, 652 pH, direct potentiometric measurement of, 558; effect, P1a~ theory, elution chromatography, 670 ' upon, in electrochemical cells, 510, 517;formal, 523;of
Oscilloscope(s), as readout devices, 77 upon ftuorescence, 288, upon organic polaropphy, glass electrode, 543; (see also Glass electrode); half-
Platmum electrode(s), for potentiometric titrations, 540
Outer mantle, offtame,311 621; measurement of, with glass electrode, 540 pn ]unction(s), 45 cell, 512; (see also Electrode potentials); half-wave
Output impedence, of operational amplifiers, 55 pH meter(s), 554 pnp Transistor(s), 45 polaropphic analysis, 608; measurement of 60 554:
Overtone line(s), in infrared spectra, 217 Phase angle,~, of electromagnetic radiation, 93; of sine- Polarimeter(s), 368 ohmic, 527 (see also IR drop); satura;ion: fo;
Overvoltage, 530 (see also Kinetic polarization) wave current, 12 Polarimetry, 357-369; applications, 369 phototube, 139; standard electrode, 514 (see also
Oxidant(s), for atomic spectroscopy, 3091 Phase relationship(s), in emitter-follower amplifier, SO; of Standard electrode potential)
Polarizability, of chemical bonds, 266
Oxidant regulator(s), for ftame spectroscopy, 313 operational amplifier outputs, 57;in series RC circuits, Potential energy, of harmonic oscillator, 213
Polarization, in electrochemical cells, 506 (see also
Oxidation, in electrochemical cells, S02 (see also Anode) 18-20; in series RL circuits, 20 Potentiometer(s), 9, 33-35; for measurement of p-values
Concentration polarization, Kinetic polarization, 554 '
Oxidation potential(s), 516 Phase-sensitive motor(s), in laboratory recorders, 78 Voltammctry); of electromagnetic radiation, 104; of
Oxidation-reduction titration(s), amperometric, 638; o-Phenanthroline, charge-transfer complex with iron(II), Poten,tiomet,ricmethod(s), 539-576 (see also Direct poten-
~olc:eutes, 102; virtual, Raman spectroscopy, 265
coulometric, 598, 5991;indicator electrodes for, 540; 179 tiometric measuremenlS, Potentiometric titrations)
Polanzatlon effect(s), upon bathochromic shifts 172
potentiometric, 571; thermometric, 6641 Phosphorescence, 110, 280; deactivation by, 284 Potentiometric titration(s), 559-573; automatic, 573-576;
Polarizer prism, of polarimeter, 368 (see also Ni~1 prism)
Oxinate(s), ftuorescence of, 287 (see also 8- Phosphorus, nuclear magnetic resonance of, 409 complex formation, 566; clitrerential, 572; end point
Polaropm(s), 607 (see also Polaropphic waves)
Hydroxyquinoline) Photoacoustic spectroscopy, 202 determination for, 561-563; of mixtures S66 568'
Polaro~p~c analysis, 606-635; alternating current, 631;
Oxygen, isotopic abundance of, 49O/;coulometric analysis Photochemical reaction(s), 169 neutralization, 568; in nonaqueous n:ema: 571:
a~plications, 617-622; clitrerential pu1Jed, 627; of
for, 594; overvoltage effects with, 531/; as oxidant, Photoelectric dctector(s), 138 oxidation-reduction, 571; precipitation, 563 '
IDIxtures, 614, 619; oxygen waves in, 617' with
atomic spectroscopy, 309; quenching of ftuorescence Photoelectric effect, 100, 460 Potentiostat(s), 63, 589; for constant electrode potential
potentiostatic control, 62H27; rapid-scan: 627;
by, 288 ' Photoelectric spectromcter(s), for emission analysis, 342 coulomctric analysis, 593; for polaropphic analysis
square wave,633;supporting electrolytes for 608 617 627 '
Oxygen electrode(s), 623 Photoelectron(s), 432, 439 620 ' , ,
Oxygen wave(s), polaropphic analysis, 617 Photoemissive surface(s), for phototubes, 139 Power, electrical, 7 (see also Electrical power); of elec-
Polaropphic current(s), 608; charging, 616; damping of, tromagnetic radiation, 93
INDEX 765

Power supplies, for conduc:tanc:cmeasurements, 647; for Quarter-wave plates, 363 ~out resolution, eft'ecton measurement of ~ 159t
elcetrogra vimetric analysis, 588 (lee also Quartz, optical characteristics of, 125, 126t; refraetive Reciprocal ohm, 644 ' Reten~on time plot(s), gas-Iiquid chromatography 72'1
Potentiostats); employing transistors, 52 index of, 359t Rectification, with semiconductor diodes 44 Retention volume(s), V, chromatographic 682t" gas' r "d
Power transformer(s), 53 Quench gas, for Geiger tube, 440 Rectifier(~,53 ' chromatography, 717 " -Iqw
Praseodymium complex(es), as chcmic:aI shift reagents, Quinhydrone, charge-transfer in, 179 ~ shift(s), of absorption spectra, 172 Reverse biasing, 44; eft'ect ~n FET transistor 48
proton nuclear magnetic resonanc:cspectroscopy, 402 Reduced mass, Jl, 215 Revcrsc-phase chromatography 698 '
Precession, of particles in a magnetic field, 380 Reduet!0n, in electrocbernicaloeUs, S02 (IU also Cathode) R~bi1ity, e1ectrochemica1,s07
Precipitate electrode(s), S40(IU also Solid-state elcctrodcs) Radial development, planar chromatography, 706 RedUCtionpotentia1(s), 516 (lee also Electrode potential) Ring turrent(s), 395
Precipitating reagent(s), eft'ect upon electrode potentials, Radiant intensity, I, of electromagnetic radiation, 93, ISOt Reference electrode(s), 531-534; for controlled cathode- Ripple, of rectified currents 53
521 Radiant power, P, of electromagnetic radiation, 93, ISOt; potential electrolysis, 586 RL circuit(s), 16 '
Precipitation titration(s), amperometric, 636, 640; con- measurement of, 1SO Referenc:cjunction, for thermocouple, 144 RLC circuit(s), 24
ductometric, 65I; coulometric, 597, S98t; potentiomet- Radiation, background, 337; coherent, 99; electromag- ReI1ectancespectroscopy, infrared, 228 Rocking vibration(s), molecular, 212
ric, 563-566; thermometric, 664t netic, 92 (IU also Electromagnetic radiation) RcI1ection,of electromagnetic radiation 103 Root-~n-~uare (nns) current(s), 12
Precision, of spectrophotometric measurements, eft'ect of Radiation buft'er(s),320 Rellecti~n grating(s), 128;for emission ;pectroscopy 344" Rota~ng platinum electrode(s), 637; voltammetry at, 625
noise upon, 157-162 Radiation detector(s), 137-145 for Infrared radiation, 220 ' , Rotat~Onalenergy, of a molecule, 107
Precision Abbe refractometer(s), 356 Radiation souree(s), atomic absorption spectroscopy, 314; Rellecti~n Ioss(es), at air-glass interfaces, ISO Rotational spectra observation of by F '
Precision:abSorption analysis, 194-197 for nuclear magnetic resonanc:cspectrometry, 384; for spectrometry, '253 ' ouner transfonn
Refraction, of electromagnetic radiation 103
PrecoIUDin(s),for HPCC, 693 optical spectroscopy, 115-123 Refractive index, n, 102; eft'ect on Beer's law 153' Rotational transition(s), molecUlar,212
Predissociation, deactivation by, 284 RadicaI(s), band spectra of, 301, 337 Ruby 1ascr(s), 119, 122
m~urement of, 353; of optical materials' 126t:
Premix bumer(s), for atomic spectroscopy, 313 Radioactive decay process(es), 459-461; products of, 459t; vanables aft'ecting, 353 ' ,
Pressure, eft'ect of, on refraetive index, 354 as source of X-rays, 427, 431, 434 Refractometer(s), 354-356
Pressure broadening, of spectral lines, 307
0 Radioactivity, decay law for, 461; detectors for, 465; errors Ref~ctometry, 353-357; applications of, 356 Sadlier collection, of thermograms 661
Primary :coil(s), of power transformers, 53 in measurement of, 462-465; types of, 459; units of, Relative error(s), in measurement of current, 60 Saba equation, 323 '
Principl~ of superposition, 94 461 Relative intensity, of spectral lines, 345 348 Salt bridge(s), 504, 527
Prism(s)! Amici, 355; for infrared radiation, 220; Lippich, Radiochemical analysis, 459-473; isotope dilution, 459, Relativ~ retention time(s), gas-liquid Chr~matography, 726 ' Sample holder(s), for optical spectroscopy 137
369; Nicol, 359; useful ranges for, 126t 471; neutron activation, 459, 467; substoichiometric, Relaxa~lon,of excited molecular states, 109, 169;of excited Sample i~l~mi~tion, for Raman spectrosc~py, 268
Prism angle, 0<, 127 " 468 SPI? statc:s,382 (lee auo Spin-lattice relaxation) Sample injection, gas-liquid chromatography, 718; for
Prism monochromator(s), 125-128 Radio frequency generator(s), pulse range of, 411 Relaxation, VIbrational, 283 HPLC, 693; to ICP source, emission spectroscopy
Probable error, of radioactive counting data, 464 Radio-frequency souree(s), nuclear magnetic resonance Relaxation eft'ect,upon ionic mobility 645 341; mass spectrometry, 479 '
Programmed temperature gas chromatography, 728 spectrometer, 388 Relaxation pr~'-
~--, nuclear magnetic" reIOnance spec- Sample preparation, for infrared absorption analysis
Programming, computer, 84 Raman scattering, 104 troscopy, 382 22~228 '
Proportional counter(s), for detection of X-rays, 439, 440 Raman spectra, comparison with infrared, 265; intensity of Rel~ing a~nt(s), atomic absorption spectroscopy 322 Sample probe, nuclear magnetic Spectrometer 389
Protective agent(s), atomic absorption spectroscopy, 322 peaks, 268; transitions responsible for, 265 Replica grating(s), 128 ' Saturated calomel electrode (SCE) 532 '
Proton nuclear magnetic resonance spectroscopy, 390 (IU Raman spectroscopy, 263-273; applications, 270; in- Residual ~~~), POlarographic wave, 608, 616; eft'ect Saturation eft'ect(s),nuclear maanetlc resonance spectros-
al,o Nuclear magnetic resonance spectroscopy) strumentation for, 268, 269 upon SCDSltiVlty, 617, 622 copy, 408
Prussian-blue, Charge-transfer in, 179 Ramp generator(s), 626 Resistance, electrical, 7"' , , Uower amplifier,50;of
' Saturation potential, for phototube, 139
,In emJtter-IO
Pulsed input(s), eft'ectupon RC circuits, 27 Random noise, in instrument signals, 70 glass m~mbrane electrode, 543; measurement of 32 Saturation timc(s), neutron activation analysis, 471
Pulsed laser(s), 122 Range, of beta particles, 460 "35,61; In a parallel circuit, 10; in a series ci °t' 9 ' Scale-expansion method(s), photometric, 194
Pulsed nuclear magnetic resonance spectra, 410 Rapid-scan polarography, 630 Resistance brid e() , rew , Scaler(s), for X-ray spectrometers 443
, g s, lor conductance measurements, 647
Pulse height(s), from X-radiation detectors, 442 Rate theory, elution chromatography, 670, 671-676 Re59lutlon, chromatographic, 677 682t Scattering of electromagnetic radia;ion, 102, 103;analytical
Pulse height analyzer(s), 348,440, 443; for determination of Rayleigh scattering, 104, 263 0
Resolu,tion element(s), of a spect~, 241 meth~s based on, 294 (lee allo Nephelometry,
alJlha energies, 466 RC cireuit(s), 14;changes of current and potential in, 17; ResolVIngpower, R, of gratings, 131;of prism monochro- ~urbldlmetry); interference from, in atomic absorp-
Pulse Jieight discriminator(s), 146, 438 eft'ect of pulsed inputs upon, 27; response of, to mators,l28 tl~n spectroscopy, 320; in spectrophotometric anal-
Pulse rimge, of radio frequency generators, 411 sinusoidal inputs, 18; time constants for, 17 Resonance lIuorescence, III, 280 " ~SIS,o 164; mechanism of, 266; Rayleigh, 104,263
Pump(s), for HPLC, 692 RC filter(s), in chopper amplifiers, 75; for elimination of Resonance Iinc:<s~, of atomic absorption spectra, 307 Scintillation counter(s), for detection of gamma rays, 466
Pumpihg, of lasing media, 119, 120 Johnson and environmental noise, 73 Resonance radiatiOn, 280(lee allOResonance fluorescence) of X-rays, 441 '
Pyroelectric detector(s), 145 RL cireuit(s), 16, 18; phase relationships in, 20 Resonance Raman spectroscopy, 272 Scissoring vibration(s), molecular, 212
Reactanc:c, in electrical circuits, 14, 22 (IU also Resonant circuit(s), 24 Secondary coil(s), of power transformers, 53
Capacitanc:c, lnduc:tance) Resonant filter(s), for elimination of noise, 73 Secondary lIuoreseent X-ray source(s), 428
Quadrupole mass spectrometer(s), analyzer for, 484; Reaction rate(s), eft'ecton electrode potentials, 523 (IU also Resonant frequency, f., 25 Second derivative titrator(s), POtentiometric, 574
resolving power of, 489t Kinetic polarization) Retardation, 6, in interferometer 246 Second-order indicator electrode(s), 539
Quanta, of electromagnetic radiation, 100 Reaction zone, of 11ame,311 Retardation factor, R" 706 ' Secon~-ordcr nuclear magnetic resonance spectra, 401
Qualitative analysis (lee utUler listinglfor ipecijic methods) Reactive circuit(s), vector cfiasrams for, 22 Rcten~on in~ex, I, gas-liquid chromatography, 726 Selection rule(s), for vibrational transitions 217
Quantitative analysis (IU utUIer listinglfor lpecijic methods) Readout dev:ice(s),5; for output signals, 77 Retention ratlo(s), R, 675 Sel~ve d~tector(s), gas-liquid chromato~phy, 729
Quantity of electricity, Q, 591 Readout limited instrument(s), photometric error in, 160, Retention time, tI<, chromatographic, 673, 676, 6821;for Selective Ion electrode(s), 539 (lee also Membrane
Quantum yield, tit for lIuorescent processes, 284 194 electrodes)
hYdrocarbons, gas-liquid chromatography, 724
Selectivity factor, 0<, chromatographic, 678, 680, 726
INDEX'

Simultaneous multielemental analysis, emission spectros- ESR, 413; name emission, 304; mass, 488 (see also Square planar configuration, splitting of d-orbitals in. 179
. .' pectroscopy 329' of
Self-absorption, atormc emiSSion s "317 copy, 327, 345 Mass spectra); NMR, 385 (see also Nuclear magnetic Square-wave polarography. 633
fluorescent species, 289; in hollow cathode lamps, Sine-wave current(s), 12-14 resonance spectra); time domain, 410 (see aI.,o Fourier Standard addition method, for atomic absorption spectros-
Self-quenching, of fluorescent species, 289 Sine-wave voltage(s), response of RC and RL circuits to, 18 transform spectroscopy) copy, 325; for emission spectroscopy. 349; for direct
Self-reversal, of atomic emiSSion hnes, 330 Single-beam photometer(s), 180; advantages of, 185 Spectral interference(s). atomic absorption spectroscopy, potentiometric measurement. 558; for polarographic
Semiconductor(s), 42-48 . .' Single beam spectrophotometer(s). atom,c absorption. 317; 320; atomic emission spectroscopy, 328 analysis, 620
Semiconductor detector(s), for electromagnetic radiation, visible-ultraviolet region. 182 Spectronuorometer(s), 289. 292; spectral information Standard deviation, <1, 738; as measure of zone breadth,
138, 140; for X-rays, 441 Single-focusing mass spectrometer(s), 481: resolving power provided by. 290 672; of radioactive counting data. 462; of transmit-
Semiconductor diode(s), 43-45 of,4891 Spectrogram(s). 125 tance measurements. 158, 159
Semiquantitative method(s), emission spectroscopy, 349 Spectrograph(s). 125; for emission analysis. 342 Standard electrode potential(s), E". 514. 518. 746--748
Singlet state(s). 281
Sensitivity. of atomic absorptIon methods. 319 Singlet-triplet transition(s). 281 Spectrometer(s), 125; gamma-ray, 466; sequential, 327. Standard hydrogen electrode (SHE). 512
Separation(s). chromatographic. 676 SiO, band spectrum of, 337 345; X-ray. 443 Standard sample(s), for quantitative emission spectros-
Separation time. chromatographic, 679 . Slave microcomputer(s). 85 Spectrometry. nuorescence, 280 (see also Fluorescence copy, 349
Sephadex®}gels, 7031, 7041; as stationary phase, thm-Iayer spectrometry); mass, 477 (see also Mass spectrometry); Standard state(s). 511
Slit system(s), for monochromators, 124, 131
chromatography. 706 - Slit width(s), effect upon absorption measurements. 163 time domain, 242 (see aIm Fourier transform Standard Weston cell(s). 34 (sl'e also Weston cells)
Sequential spectrometer(~). 327. 345 Soap-bubble meter. 718 spectroscopy) Stationary electrode(s), voltammetry at. 622---625
Sequential wavelength disperSive X-ray spectrometer(s), Sodium chloride, as cell material, infrared spectrophotom- Spectrophotometer(s), 125; atomic.absorption. 317; atomic Stationary phase(s), chromatographic, 667; gas-liquid
etry, 227; diffraction of X-rays by, 4371; opllcal emission, 326; dual wavelength, 193; infrared, chromatography, 7251; selection of. 723; for gel
Serie~ircuit(S) 8-10; phase relationships in. 18-21; 221-226: Raman, 269; visible-ultraviolet, 180---186
characteristics of, 1261 chromatography, 703; thin-layer chromatography,
response of, to sinusoidal inputs,: 18 Sodium D line, as source for refractive index measure- Spectrophotometric analysis, 151 (.fee also Absorption of 706
Series resonant filter(s), 24 electromagnetic radiation); derivative, 192; differen- Statistical treatment of data, 738-744
ments,354
Servosystem(s). 77 . .. Sodium iodide crystal(s), for scintillation counters, 441. tial, 194-197; effect, of noise on precision of, 157 (see Stilbene, use in scintillation counters. 441
Shielding, to minimize environmental nOIse. 73; of nucleI m Sodium ion, glass membrane electrode for, 547. 5481; sohd- also Noise), of scattered radiation upon. 157, 164. of Stimulated emission, of laser, 120
a magnetic field. 394 state electrode for, 5521 slit width upon, 163; of mixtures, 191 Stokes shift(s), III, 264. 280
Shielding parameter, <1, 391 Solid(s), infrared spectrophotometry of, 228 Spectrophotometric detection. for HPLC, 694 Stray radiation, effect on Beer's law. 156; in monochro-
Shift(s) Stokes/anti-Stokes. 264 Solid-state electrode(s), 540, 550. 5521 Spectropolarimeter(s), 372 mators. 134
Shot n~ise, 70, 71; effect upon measurement ofT. 1591, 160:
Solid support(s), for electrochromatography. 709; for gas- Spectroscope(s), 125 Stretching, absorption resulting from. 213
in photoelectric detectors, 138 ' liquid chromatography. 719; for partition chromatog- Spectroscopic buffer(s), 350 Stretching vibration(s), molecular. 212; spectral regions
Shunt, Ayrtron. 29 Spectroscopy, 114; atomic absorption. 304 hee also Atomic involving. 230---233
raphy, 698
Side band(s). radio frequency. 411 Solution(s), use, in emission spectroscopy, 338 absorption spectroscopy); atomic emission, 303 (see Stripping analysis, 633---635
Siemens, S. 7 Solvent(s). effects of, upon absorption spectra, 172; upon also Atomic emission spectroscopy); atomic fluores- Stripping column(s), ion exchange chromatography. 701
Sieve(s). molecular. 731 atomic absorption spectroscopy, 325; upon fluores- cence, 304 (see 01.'" Atomic nuorescence spectros- Structure, polarimetric elucidation of. 369
Sigma (<1) orbital(s). transitions involving, 171. 285: charge
cence 287' for infrared absorption analySIS. 227; for copy); emission. 336 (see also Emission spectroscopy); Substitution, effect on ultraviolet spectra of aromatic
density of. 170 . . orga~ic ~larographY, 621; for ultraviolet-visible infrared, 210 (see also Infrared absorption spectros- compounds, 174
Signal(s), analytical. 2 (see also AnalytIcal Signals)
regions, 187, 1891 copy): Raman, 263 (see 01.'" Raman spectroscopy); Substoichiometric method(s). isotope dilution, 472; neut-
Signal averaging, 76, 242; Fourier transform spectrometry,
Solvent reservoir(s), for HPLC, 691 reflectance, 228: visible-ultraviolet, 168 (see also ron activation analysis, 468
249; of output from fast analyzer, 202 Source(s), for atomic absorption spectroscopy. 315; for Visible-ultraviolet region) Subtraction of input signals. with operational amplifiers, 65
Signal chopping. 74 atomic fluorescence spectroscopy. 332; for emission Spectrum, absorption, 105: electromagnetic. 100 Sucrose. inversion of. 369
Signal generator(s). 4 spectroscopy. 336; for fluorescence spectrometry, 290; Spin, nuclear, 377 Sulfide ion. solid-state electrode for, 5521
Signal processor(s). 4; for optical spectroscopy. 145; for X-
for Fourier transform spectrometer. 251; for mass Spin decoupling, chemical exchange and. 401; of nuclear Sulfur. correlation chart for 2, binding energies. 455;
ray spectrometers, 443 spectrometers. 479-481; for optical spectroscopy, 115 isotopic abundance of. 4901
magnetic resonance spectra, 402
Signal shot noise. effect upon measurement of T. 160 (see
("ee 01,,0 under specific types); for polarimetry. 368; for Spin-label reagent(s), 415 Summing circuit(s), 65
also Shot noise) Raman spectroscopy, 268; for X-rays. 427, 434 Spin-lattice relaxation. 383 Summit current(s), rapid-scan polarography, 630
Signal-to-noise (S/N) ratio(s). 68-70; improvement of,
Source flicker noise, effect upon measurement of T. 1591 Spin-lattice relaxation time. T,. 383 Summit potential, E., 630
72-77; influence on spectral detail. 242
(."e 01,,0 Flicker noise) Spin quantum number, I, 378. 3791 Superconducting magnet(s). for nuclear magnetic reson-
Silica gel. as solid support. partition chromatography. 698;
Source lead(s), for FET transistors, 48 Spin-spin relaxation. 384 ance spectrometers. 387
as stationary phase,liquid-solid chromatography. 696
Spark source(s). for emission spectroscopy. 336. 337, 339; Spin-spin relaxation time. T,. 384 Superposition. principle of. 94
Silicon detector(s). lithium drifted. 441 Supporting electrolyte(s). for polarographic analysis. 608.
for mass spectrometers. 480 Spin-spin splitting. in nuclear magnetic resonance spectra.
Silicon diode(s). as photon detectors, 138. 141
Specific conductance. k. 644 391,398--401; sources of. 392 609,617.620
Silicon semiconductor(s). properties ~f. 42 Sweep source(s), nuclear magnetic resonance spectrometer,
Specific ion electrode(s). 539 (."e also Membrane Splitting, of d-orbital energies. 177; of electronic energy
Silver ion. coulometric generation of. 597; glass mem-
electrodes) levels. 414 388
brane electrode for, 547, 5481; solid-state electrode for,
Specific retention volume, V,. gas-liquid chromatography, Splitting factor, for unpaired electron. g. 412 Synchronous demodulator(s), 75
5521 Spontaneous emission, of laser, 120
716.717,726
Silver nitrate/glycol. as liquid phase. gas-liquid chromatog-
Specific rotation, [aJf, 367 . . Sputtering. in hollow cathode lamp, 317
raphy.7241 Tank cireuit(s). 27
Spectra. absorption. 105 (see also Absorpt~on spectra): Squalene. as liquid phase, gas·liquid chromatography, 723.
Silver/silver chloride electrode(s). 533 band. 337; continuous X-ray. 427; emiSSIOn, 109-111, Technicon AutoAnalyzer(s) @. 199
7251
Simple harmonic motion. 213
INDEX 769

Toluene. as calibration standard for refractometry. 356 Unpaired electron, behavior in a magnetic field 414 in functioning of glass electrod 542' '
Temperature. effect of. on atomic spectra. 309-312. on Topaz. diffraction of X-rays by. 431/ unsacturated ketone(s) absorption by, 114/ '(see also visible-ultraviolet regions. 189/ e, , as solvent 10
calomel electrodes. 5321. on conductance measure- Total consumption burner(s). for atomic spectroscopy, 313 arbonyl groups) Watt, W. 1
ments. 648. upon electrode potentials. 518. on Transducer(s). 4; amplification of signals from. 59; (see also Unsaturated Weston cell(s). 534 Wave guide(s), 413
electrogravimetric analysis. 581. upon fluorescence. Wavelength, A, of electromagnetic radiation 93' d .
specific types)
281. on gas-liquid chromatographic analysis. 720. on vali~~~Ycin. as liquid ion exchanger for potassium ion,
Transducer output(s), comparison of. 61 ~;~~';'f" and t for optically active sUbs;an~ ~:::-
half-ccll potentials. 518, on kinetic polarization. 530.
Transformer(s), 52 ind;x. 3~ of. upon critical angle. 355, on refractiv;
on polarographic wave, 614. upon refractive index. Transistor(s). 45-48; amplifiers based on. 48-52 van, Deemter equation. 614-616
354. on specific rotation. 361. on Weston cells, 534; of Transition(s). involved in absorption. of infrared radiation, Vanan Cary Model 219 spectrophotometer 185 Wavelength dispersive X-ray spectrometer(s) 434 444
flamcs. 308. 3091; measurement of. thermomelric 212-219. of visible-ultraviolet radiation. 169-180 Vanance, u'. 613. 141 • Wavelength
190,220 sel ec t'Ion. ~or absorption analysis.
, •
123-131,
titrations 664 Transition-metal ion(s). absorption of radiation by. 116 Vector diagram(s). for reactive circuits 22
Temperature profile(s). of flames. 309 Transmission of electromagnetic radiation. 102; through Velocity of electromagnetic radiation ~ 93 353 ( I :avenumber" u, of electromagnetic radiation. 93
Terpheny\' use in scintillation counters. 441 tmnsparent media. 351 . Refractive index); angular, w, 93'" see a so ave properties. of electromagnetic radiation, 92-99
Tetrachloroethylene. as solvent for infrared spectrometry, Vlbration(s) degenerate, 219; molecular. 212
Transmission grating(s). 128 Wave t~m(s), of electromagnetic radiation 99
221 Transmittance. T. 149. )501; measurement of. 151; effect of Vibrational absorption. 108 Weak aCids and base s, po t'entlOmetnc.,' titration of, 568
Tetrahedral configuration, splitting of d-orbitals in. 119 instrumental noise upon. 157 Vibrational coupling: 212. 218 Wedge(s). filter. 220; IOterference, 136
Thallium bromide/thallium iodide. optical characteristics Transmittance range(s). of optical materials. 126/ Vibrational energy, of a molecule 101 Weston cell(s). 34. 534; use in potentiostats 63
of. 1261; use in internal reflectance spectroscopy. 229 Transverse relaxation. 384 (see also Spin-spin relaxation) V~brational frequency, of mechanical oscillator. 213 Wheatstone bridge(s). 35. 641 •
Theoretical plate(s). of chromatographic column. 611 Triglycine sulfate. use in pyroelectric detectors. 145 Vlbr~~~nal mode(s),' molecular. 211;, and Raman effect, White light interferometer system. Fourier t ~
Thermal conductivity detector(s). gas-liquid chromatog- Triisobutylene. as liquid phase. gas-liquid chromatog- spectrometer. 250 rans orm
raphy.720 raphy.7241 Vibrational quantum number. v, 215 W!"te ~oise, 11 (see also Johnson noise) '..,
Thermal detector(s). 143-145 Trimethylsilane (TMS), as internal lock. nuclear magnetic V~brational relaxation. deactivation by, 283 W~de-hne nuclear magnetic resonance spcctrometer(s) ;86
Thermal method(s). 655~ resonance spectroscopy, 388. 392 Vlbratlon-~otallo~al transition(s), absorption of infrared WlOdow(s). of monochromator, 124 '
Thermal neutron(s). 461; activation analysis with. 461 Triolefin(s). absorption by. 114/ (see also Alkenes) radIatIon dunng 212 Word(s), computer. 81
Thermal noise. 10 (see also Johnson noise) Triple bond(s). as chromophore. 173/ (see also Alkynes); Vidicon "tUbe(s) • 142", .cmp Ioyment m ' multIelement
, atomic Work function. w, 100
Thermal radiation, emission of. 109 effect of magnetic fields upon. 396 emIssIon analysis. 321
Thermistor(s). 144 Triple bond region. infrared spectrum. 232 Virtual ground. in arrtplifier circuit, 58
Thermobalance(s). 655 Xenon arc lamp(s) . 111". as source, fluorescence spcctrom
Triplet state(s), 281 Vlltual state(s). 265 '
etry.29O -
Thermocouple(s). 60. 144 Trithiocarbonate ion(s). absorption in the ultraviolet by, Visible-ultraviolet reg·ion .. 101' absorptl'o n ana I"YSIS In
Thermocouple junction(s). 63. 144 X-Ray() s • ,ab sorptIOn
' of, 431; analytical uses of. 421 ...••56·
115 168-204; detectors for. 138-143; instruments for'
Thermodynamic cell potential(s). 524 (see also Cell Triton X-loo, as maximum suppressor. polarographic 180-186; solvents for. 181, 189/' sources for 111 123' emISSion of., 110, 421, 461; properties of. 4591 "
potential) Volt. V, 1 " - X-Ray analySIS. mstrumentation for, 434-447' quart ti
analysis. 616 441; quantitative, 448 ' I ave,
Thermogram(s). 655; differential. 651; Sadtler collection. Voltage, control of. with operational amplifiers 63' Ze
Tube(s). photomultiplier. 140; vidicon. 142
661 breakdown. 44 • , ner X-Ray diffraction. 433. 450
Tunable laser(s), 123; as source. fluorescence spectrometry.
Thermogravimetric analysis. 655 X-Ray fluorescence, 432. 444. 449
290; for atomic fluorescence analysis, 333 Voltage divider(s). 9, 34, 618
Thermometric titration(s). 655. 662~ X-Ray photoelectron (ESCA) spectroscopy, 453
Tuning. of dye lasers, 119 Voltage follower(s), 58; in voltage measuring devices 60
Thin layer chromatography. 6681, 690. 105. 106 Tungsten filament lamp(s), spectral characteristics of, 111 Voltage measurement(s). 30, 34; effect of meter resist~nce
Thiocyanate ion. charge-transfer complexes involving. 119; Turbidimetric analysis, 294-296; applications. 291/ upon, 31; m a p~rallel circuit, 10; with a potentiometer,
solid-state electrode for. 5521 Turbidity coefficient. T. variation with wavelength. 296 34. m a senes cllcuit, 10: uncertainties in 35
Thrce-electrode polarography. 626 Turbulent flow burner(s), for atomic spectroscopy. 313 Voltage regulator(s), 53-55 . Zeeman effect correction method, atomic absorpti
Three-level laser system(s). 122 Turner 110 fluorometer. 291 Voltammetry, 606 (see also Polarography. Amperometric spectroscopy. 321 on
Time. measurement of. for coulometric titrations. 59S~ lit ratIOns); cychc. 631; rapid-scan. 630; with solid
Turner 350 spectrophotometer. 183 Zener b~akdown voltage, for semiconductor diodes 44
retention. I•. chromatographic analysis. 673 (see also
Twisting vibration(s). molecular, 212 Volu~:ctrodes. 622-625; square wave. 633: stripping, 633 Zener dlode(s). use of, in laboratory recorders 18 '
Relention lime) Two-dimensional electrochromatography, 710 c~ange. corr~tlon for, in amperometric tit rations voltage regulators. 53 • ,10
Time constant(s). in RC circuit. 11; in RL circuit. 18 Two-dimensional planar chromatography. 101 636. m conducto~etric tit rations. 648. in photometri~ Zero percent Tadjuslment. absorption spectroscopy, 151;
Time domain spectra. 242 (see also Fourier transform Two-line correction method. atomic absorption spectros- IItrallons. 191 effect of nOise upon. 151' uncertainties in 158
spectroscopy); acquisition of. 244 Ze " ,
copy, 321 Volume error(s). in chromatographic analysis. 683 ro percent T nOise, effect upon measurement of T 1-'9
Time-of-flight mass spectrometerts). 482: resolving power Zone breadth. elution chromatography, 672 • I
Tyndall effect. 104
of. 4891 Wagging \'ibration(s); molecular, 212 Zone broadening, elution chromatography. 610. 614-616
Time-sharing. of computers, 84 Water. as calibration standard for refractometry. 356; role Zone shape{s), chromatographic, 612
Titration(s). amperometric. 635; conductometric. 648, Ultimate precision method. differential-absorption anal-
coulometric. 581; differential. 572; oscillometric. 652;
ysis. 195/. 196
photometric. 191; potentiometric. 559; radiometric.
Ultraviolet photoelectric spectroscopy. 453
473; thermometric. 655 (see also under specific listings) Ultraviolet region. 101 (see also Visible-ultraviolet region)
Titration curvets). amperometric. 636; conduetometric. Uncertainties. propagation of. in physical measurements.
649; photometric, 191; potentiometric. 561-572
138-144
Titrator(s), automatic. 513
· Symbols for Units MUltiplicative Prefixes

Prefix Symbol Factor


A ampere hr hour pico p 10-12
A angstrom Hz herz nano n 10-'
atm atmosphere micro Jl 10-6
1 joule milli
C oK m 10-'
coulomb degree Kelvin centi 10-2
c
°C -degree Celsius M mole/liter deci d 10-1
eV electron volt kilo k 10'
m meter mega
F farad M 106
min minute giga G 10'
F faraday N equivatent/liter
Conversion Factors for F
G
formula weight/liter
gauss
s
V
second
volt
Electromagnetic Radiation g
H
gram
henry
W
n
watt
(omega) ohm
(To convert data in units !,f x shown in th.e first c:o!umn to the units
indicated in the remaining columns. multiply or divide as shown.).

Wave- Wave-
Units Frequency. number. length.
of x Hz cm-1 Energy cm

kcal/mol erg eV

3.00 X 1010
Hz 1.00 x 3.34 x 10- 11x 9.54 X 10-14 x 6.63 x 10-2' x 4.14 X 10-" x
x
1.00
em-I 3.00 x 1010X 1.00x 2.86 X 10-' x 1.99 X 10-16 x 1.24 X 10-4 x
x absorbance, area signal to noise
2.86 X 10-' absorptivity, amplifier gain, activity transmittance, temperature
kcal/mol 1.05 x 1013x 3.50 X 102 x 1.00 x 6.95 x 10- 14X 4.34 X 10-2 x
x capacitance, concentration time
1.99 X 10-16 diffusion coefficient dc voltage
erg 1.51 x 1026x 5.04 X 10" x 1.44 X 1013x 1.00 x 6.24 X 10" x
diameter, spacing
x volume
1.24 X 10-4 angular degree ac voltage
eV 2.42 x 1014x 8.07 X 10' x 2.31 X 101 x 1.60 X 10-12 x 1.00 x electrical potential, energy velocity
x
3.00 X 1010 1.00 2.86 X 10-' 1.99 X 10-16 1.24 X 10-4 electron reactance
em 1.00 x frequency, .activity coefficient
x x x x x mean
1.00 X 10' 2.86 104 1.99 x 10-' 1.24 X 10' conductance, free energy impedance
3.00 X 10" X
nm ---- x
1.00 X 10-' x magnetic field, enthalpy
x x x x
dc current
ac current
equilibrium constant
inductance (epsilon) molar absorptivity
refractive index, number of (lambda) wavelength
equivalents (mu) mean
spectral order (nu) frequency
radiant or electrical power (rho) density
quantity of dc electricity (sigma) standard deviation, wavenumber
quantity of ac electricity (tau) period
electrical resistance, gas constant (phi) phase angle
standard deviation

....••••.............
(omega) angular velocity

---------------------_._.'~--_.._-----------

You might also like