ARTICLE IN PRESS

Water Research 38 (2004) 20432052

Adsorption of anionic and cationic dyes on activated carbons with different surface chemistries
! a * o, M.F.R. Pereira* P.C.C. Faria, J.J.M. Orf
! ! Laboratorio de Catalise e Materiais, Departamento de Engenharia Qu! mica, Faculdade de Engenharia da Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal Received 20 March 2003; received in revised form 20 October 2003; accepted 27 January 2004

Abstract The inuence of the surface chemical groups of an activated carbon on the removal of different classes of dyes is evaluated. Starting from the same material (NORIT GAC 1240 PLUS), the following treatments were carried out in order to produce a series of samples with different surface chemical properties but with no major differences in their textural properties: oxidation in the liquid phase with 6 M HNO3 and 10 M H2O2 (acid materials) and heat treatment at 700 C in H2 or N2 ow (basic materials). The specic micropores volume and mesopores surface area of the materials were obtained from N2 adsorption equilibrium isotherms at 77 K. The surface chemistry was characterised by temperature programmed desorption , by the determination of the point of zero charge (pHpzc) and by the evaluation of the acidity/basicity of the samples. Elemental and proximate analyses were also carried out. Equilibrium isotherms of selected dyes (an acid, a basic and a reactive dye) on the mentioned samples were obtained and the results discussed in relation to their surface chemistry. In general, the Langmuir model provided the best t for the adsorption data. It is shown that the surface chemistry of the activated carbon plays a key role in dye adsorption performance. The basic sample obtained by thermal treatment under H2 ow at 700 C is the best material for the adsorption of all the tested dyes. r 2004 Elsevier Ltd. All rights reserved.
Keywords: Activated carbon; Surface chemistry; Wastewater; Decolourisation; Dyes; Adsorption

1. Introduction The textile industry is characterised by its high water consumption and is one of the largest industrial producers of wastewaters. Water from spent dyebaths and dye rinse operations contains unxed dyes and may be highly coloured. The persisting colour and the nonbiodegradable nature of the spent dyebaths represent serious problems to the environment. However, the primary concern about efuent colour is its undesirable aesthetic impact in receiving waters [1].
*Corresponding author. Tel.: +351-22-508-1468; fax: +35122-508-1449. E-mail address: fpereira@fe.up.pt (M.F.R. Pereira).

Despite the existence of a wide range of wastewaters treatment techniques, there is no single process capable of adequate treatment for these efuents [2]. Thus, the best solution for textile wastewater treatment appears as a combination of different techniques. The conventional association of biological and physicalchemical processes results in the removal of most of the organic content of the textile wastewaters. However, the resulting efuent is still fairly coloured. Ideally, a nal rening treatment is needed to remove the colour [3]. This nal step in the treatment of textile efuents is indispensable for a potential water reuse, since total decolourisation and extensive elimination of organic content are required for water recycling in the textile industry [4]. In this context, adsorption techniques appear as

0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2004.01.034

ARTICLE IN PRESS
2044 P.C.C. Faria et al. / Water Research 38 (2004) 20432052

Nomenclature AC a.u. Ce CI d KF KL L1 n N pHpzc qe activated carbon arbitrary units adsorbate equilibrium liquid phase concentration (mg/dm3) Colour Index particle diameter parameter in the Freundlich equation mg=gAC dm3 =mg1=n parameter in the Langmuir equation (dm3/ gAC) average width of the small micropores (nm) parameter in the Freundlich equation number of points point of zero charge adsorbate equilibrium solid phase concentration (mg=gAC )

calculated adsorbate equilibrium solid phase concentration (mg=gAC ) qm maximum adsorption capacity determined by the Langmuir equation (mg=gAC ) SBET BET surface area (m2/g) Smeso mesopore surface area calculated by the tmethod (m2/g) SD standard deviation T temperature ( C) Wmicro micropore volume calculated by the t-method (cm3/g) W01 ; W02 volumes of small (1) and large (2) micropores calculated by the Dubinin method (cm3/g) Greek symbols maximum absorption wavelength (nm) lmax

qcalc e

successful efcient alternatives for the removal of dyes in wastewater. Studies involving the application of a wide range of adsorbents in the decolourisation of wastewaters are reported in literature [5,6], adsorption on activated carbon (AC) being one of the most studied techniques [79]. Activated carbons have a high degree of porosity and an extensive surface area. These materials are effective adsorbents for several organic compounds of concern, in water and wastewater treatment. The adsorption capacity of an AC is determined not only by the textural properties but also by the chemical nature of its surface, i.e., the amount and nature of oxygen-containing functional groups [10]. The nature of an activated carbon surface can be changed by different treatments, which include liquid phase oxidations with HNO3 and H2O2 and gas phase oxidations with O2 or N2O, as well as thermal treatments at high temperatures [11]. A thorough knowledge of AC surface chemistry enables the preparation of adsorbents with appropriate characteristics for specic applications [10]. Yet, little attention has been given to the role of the AC surface chemistry in the adsorption process. Karanl et al. [12] evaluated the uptake of two organic contaminants on a set of commercial and treated ACs in order to study the inuence of the surface chemistry on the adsorption of those organic pollutants. More recently, some authors started to study the effect of activated carbon surface chemistry on the removal of dyes from textile efuents [13,14]. Al-Degs et al. [13] studied the adsorption of three reactive dyes on a Filtrasorb 400 activated carbon, and attributed the high adsorption capacity of this adsorbent to its net positive surface charge during the adsorption process. In a previous work, Pereira et al.

[14] followed a different approach. Starting from the same material, a set of activated carbons differing in their chemical surface properties was prepared. In the mentioned work, a screening test was carried out in order to evaluate the performance of the treated ACs in the removal of a series of dyes. It was concluded that basic carbons were the most efcient for the removal of both cationic and anionic dyes. In this work, a NORIT GAC 1240 PLUS activated carbon is modied by selected oxidation and heat treatments and characterised by different techniques. The adsorbents prepared are subsequently tested in the removal of three dyes and the equilibrium adsorption data modelled with the Langmuir and Freundlich equations. The results obtained are discussed in relation to both textural and chemical properties of the adsorbents and to the pH vs. pHpzc of the AC samples. Our main goal is to account for the interactions between the different classes of dyes and the surface groups of a modied commercial activated carbon in order to develop optimised adsorbent materials for the decolourisation of wastewaters.

2. Experimental 2.1. Preparation of modied activated carbons A Norit GAC 1240 PLUS activated carbon was used as supplied as the starting material for this study (Sample AC1). This carbon was selected because it is an acid washed granular activated carbon, which offers good adsorption properties for water applications and a high purity level. According to the supplier

ARTICLE IN PRESS
P.C.C. Faria et al. / Water Research 38 (2004) 20432052 2045

specications, it has a neutral pH. The treatments outlined below were carried out in order to obtain materials with different surface chemistries, while maintaining the original textural properties as far as possible. All treatments were performed on the unground materials (0.42od o2 mm). 2.1.1. Oxidation treatments Oxidation with HNO3 was performed using a 100 cm3 Soxhlet extraction apparatus. Initially 200 cm3 of 6 M HNO3 were introduced into a 250 cm3 Pyrex roundbottom ask and heated to boiling temperature with a heating mantle. The Soxhlet with 9 g of activated carbon was connected to the boiling ask and to the condenser. The reux was stopped after 3 h. The activated carbon was washed with distilled water to neutral pH, dried in an air convection oven at 110 C for 24 h and stored in a desiccator for later use (Sample AC2). This sample was used as the starting material for the thermal treatments described below. Sample AC3 was prepared by mixing 1 g of the original material with 25 cm3 of 10 M H2O2 at room temperature until complete degradation of the H2O2 (when there was no further gas evolution). The sample was washed with distilled water to neutral pH, dried for 24 h in the oven at 110 C and stored in the desiccator until used. 2.1.2. Thermal treatments All the thermal treatments were performed on sample AC2 since it is important that the starting material presents a large amount of surface groups in order to produce activated carbons with a high basic character [15]. About 3 g of sample AC2 were placed in a fused silica tubular reactor, heated to 700 C at 10 C/min under a ow of H2 (50 cm3/min) and kept at this temperature for 1 h. After cooling to room temperature under the same atmosphere, a ow of air was introduced into the reactor and these conditions were maintained for 1 h. Finally the sample was collected and stored in a desiccator (Sample AC4). The same treatment was repeated under a ow of N2 at 700 C (sample AC5). Thermal treatments under a ow of H2 are expected to generate higher amounts of basic sites on the surface of the carbon than thermal treatments under a ow of N2 [16]. 2.2. Characterisation of the activated carbon samples 2.2.1. Textural characterisation The textural characterisation of the materials was based on the N2 adsorption isotherms, determined at 77 K with a Coulter Omnisorp 100 CX apparatus. The micropore volumes (Wmicro ) and mesopore surface areas (Smeso ) were calculated by the t-method using the standard isotherm proposed by Rodr! guez-Reinoso

et al. [17]. The adsorption data were also analysed with the Dubinin equation. In all cases, a type IV deviation was noted [18]. Two micropore structures were taken into account, and the corresponding volumes, W01 and W02, were calculated [18]. The Stoeckli equation [19] was used to estimate the average micropore width of the smaller pores (L1), using a value of 0.34 for the afnity coefcient of nitrogen. 2.2.2. Surface chemistry characterisation Temperature-programmed desorption (TPD) proles were obtained with a custom built set up, consisting of a U-shaped tubular micro-reactor, placed inside an electrical furnace. The mass ow rate of the helium carrier gas (69 mg/s) and the temperature programme from room temperature to 1100 C at a heating rate of 5 C/min were controlled with appropriate equipment. The amounts of CO and CO2 desorbed from the activated carbon (100 mg) were monitored with a Spectramass Dataquad quadrupole mass spectrometer. The masses monitored for all samples were 2 (H2), 16 (O), 18 (H2O), 28 (CO) and 44 (CO2). The basicity of the carbon materials was estimated mixing ca. 0.2 g of each sample with 25 cm3 of 0.025 M HCl solution in a closed ask, and maintaining the contact under agitation for 48 h at room temperature. Then the suspension was decanted and the remaining HCl in solution was determined by titration with a 0.025 M NaOH solution. The total acidity of each sample was obtained by a similar procedure where a 0.025 M NaOH solution was put in contact with the activated carbon and the titration solution was 0.025 M HCl. The determination of the pHpzc of the samples was carried out as follows [20]: 50 cm3 of 0.01 M NaCl solution was placed in a closed Erlenmeyer ask. The pH was adjusted to a value between 2 and 12 by adding HCl 0.1 M or NaOH 0.1 M solutions. Then, 0.15 g of each AC sample was added and the nal pH measured after 48 h under agitation at room temperature. The pHpzc is the point were the curve pHnal vs. pHinital crosses the line pHinitial=pHnal. The materials used were also characterised by elemental and proximate analyses, with a CARLO ERBA EA 1108 Elemental Analyser and a Mettler TA 4000 thermal analyser, respectively. 2.3. Adsorption experiments Three dyes from different classes, with known molecular structure, were used to evaluate the adsorption capacity of the prepared materials. The selected dyes and their characteristics are shown in Table 1. In order to reduce mass transfer resistances and, consequently, the time necessary to reach equilibrium, the activated carbon was nely ground and only the fraction

ARTICLE IN PRESS
2046 P.C.C. Faria et al. / Water Research 38 (2004) 20432052

with a particle diameter smaller than 50 mm was used. Preliminary kinetic tests were carried out in order to determine the equilibrium time, which was found to depend on the specic dye (3 days for the basic dye, 4 for the reactive and 5 days for the acid dye). It was assumed that equilibrium was attained when no further changes in dye uptake were observed after 24 h. Longer times are necessary when larger particles are used, usually between 2 and 3 weeks [9,13]. The acid and basic dye solutions were prepared simply by dissolving the desired amount of dye in distilled water. Taking in consideration that in textile efuents reactive dyes are in the hydrolysed state, an hydrolysis procedure was carried out as described elsewhere [14]. The initial pH of all solutions was adjusted to 67. For each dye solution the UV-Vis spectrum was previously determined and the maximum absorption wavelength identied (Table 1). Adsorption isotherms were obtained by mixing 50 cm3 of dye solution of different concentrations (501000 mg/ dm3) with 50 mg of activated carbon nely ground (d o50 mm) in closed Erlenmeyer asks. The suspensions were shaken at room temperature, until equilibrium was reached. After that, they were centrifuged and the remaining concentration of dye in solution was determined in a JASCO V-560 UV-Vis spectrophotometer.

removal of dyes. For that purpose, a set of modied activated carbons with different levels of acidity/basicity but with no major differences in their textural parameters was prepared. Both liquid phase oxidations and thermal treatments have been reported not to signicantly change these properties [11]. Oxidations with HNO3 and H2O2 are known to generate acidic materials [21]. Treatments with HNO3 originate materials with large amounts of surface acidic groups, mainly carboxylic acids and, in less extent, lactones, anhydrides and phenol groups [11], whereas treatments with H2O2 produce less acidic materials [14]. On the other hand, heat treatments under inert or H2 atmosphere at high temperatures are used to selectively remove some of the surface oxygen groups, thus originating basic materials [16].

3. Results and discussion 3.1. Characterisation of the activated carbon samples The main goal of the present work is to study the relationship between the surface chemical characteristics of activated carbons and their performance on the

3.1.1. Textural characterisation The results obtained from the N2 equilibrium adsorption isotherms at 77 K are presented in Table 2. As expected, no major differences in the textural properties of the materials were observed, particularly in the case of the micropores volume. The mesopores area, as well as the volume of micropores, slightly decreased for the sample oxidised with HNO3. This may be due to the abundant presence of oxygen-containing groups on the surface of the activated carbon, which makes part of the carbon surface inaccessible. The increase in micropores volume and mesopores area in samples AC4 and AC5, relatively to sample AC2, can be explained by the thermal decomposition of the mentioned oxygen groups. As previously discussed [14], no drastic changes in the textural properties of the prepared samples were observed, suggesting that the differences in

Table 1 Characteristics of the dyes selected for this study Class Basic Reactive Acid Commercial name Astrazon Brilliant Red 4G Rifax Red 3BN Erionyl Navy R Generic name CI basic red 14 CI reactive red 241 CI acid blue 113 Chemical class Cyanine Monoazo Diazo lmax (nm) 514 540 566

Table 2 Textural characterisation of the AC samples Sample AC1 AC2 AC3 AC4 AC5 Wmicro (cm3/g) 0.367 0.353 0.365 0.375 0.356 Smeso (m2/g) 125 91 110 103 114 W01 (cm3/g) 0.312 0.303 0.311 0.318 0.301 W02 (cm3/g) 0.058 0.048 0.054 0.058 0.050 L1 (nm) 1.1 1.1 1.2 1.2 1.2 SBET (m2/g) 972 909 949 972 946

ARTICLE IN PRESS
P.C.C. Faria et al. / Water Research 38 (2004) 20432052 2047

the adsorption capacities of the samples will be mainly due to their chemical surface properties. 3.1.2. Surface chemistry characterisation Fig. 1 shows the TPD spectra of the commercial activated carbon as received and after the different treatments. Surface oxygen groups on carbon materials decompose upon heating releasing CO and CO2 at different temperatures. According to this it is possible to identify and estimate the amount of oxygenated groups on a given carbon by TPD experiments. It has been reported [11] that CO2 peaks result from the decomposition of carboxylic acids at low temperatures or lactones at high temperatures, and anhydrides originate both CO and CO2. Groups such as phenols, carbonyls, ethers and quinones originate CO peaks. Analysing the results, it is clear that sample AC2, which was oxidised with HNO3, has the highest amount of surface oxygencontaining groups. Comparing to sample AC1, the H2O2 treated sample contains a higher amount of oxygenated surface groups, mainly carboxylic ones, which means that sample AC3 is more acidic than the starting material. Thermal treatments originate materials with quite low content of oxygen-containing groups. As can be seen in the TPD spectra of samples AC4 and AC5, the CO2 releasing groups have almost completely been
2.5
AC2

2.0

1.5
AC3

1.0 0.5
AC1 AC4 AC5

(a)

0.0 0 0.6
AC2

200

400

600

800

1000

T (C)

0.5

CO2 (a.u.)

0.4 0.3 0.2 0.1 0.0

AC3

removed, but some groups releasing CO at high temperatures still remain on the carbon surface. These ones are presumed to be part of the carbonyl groups that have not been decomposed by the treatment at 700 C, and pyrone type groups, which have basic nature and decompose at high temperatures. Thermally treated samples have a basic character which is due to this kind of oxygen-containing groups and mainly to the electron rich oxygen-free sites located on the carbon basal planes ! ndez et al. [16], heat treatments [22]. According to Mene under inert atmosphere at high temperatures are effective in removing oxygen but the resulting activated carbon has a surface with reactive sites capable of readsorbing oxygen leading to the formation of some of the previously removed groups. Treatments with H2 also remove oxygen groups but leave stable basic surfaces by forming CH bonds, thus preventing further adsorption of oxygen. This is in agreement with the results obtained, since the TPD spectrum of sample AC5 shows a higher amount of CO2 releasing groups than sample AC4. The acidity and basicity of the samples and the values of the pHpzc are shown in Table 3. These two kinds of results are complementary. As expected, the total surface acidity increased for the oxidised samples, whereas the total surface basicity decreased. These treatments yielded ACs with the lowest pHpzc. The greatest total surface acidity was obtained for sample AC2 due to the introduction of several oxygen-containing functional groups, namely carboxylic acids. Sample AC3 has a more acidic character than the starting material. However, the oxidation with H2O2 originates a smaller amount of acid groups on the surface than oxidation with HNO3, which is also conrmed by TPD results. On the other hand, sample AC4 has the highest basic character. Comparing the results obtained for samples AC4 and AC5, one can say that the treatment under H2 atmosphere leads to more basic materials than the treatment under inert atmosphere, which is in agreement with previous studies reported in the literature [16]. Table 4 shows the results of proximate and elemental analyses, which are consistent with the data from the previous characterisation techniques. It can be seen that oxidation with HNO3 leads to a substantial increase in
Table 3 Acidity/basicity of the AC samples Sample Acidity (meq/gAC) 149 716 358 89 179 Basicity (meq/gAC) 302 34 153 437 365 pHpzc 9.7 2.7 5.4 10.8 9.9

CO (a.u.)

AC1

AC5

AC4

(b)

200

400

600

800

1000

T (C)

Fig. 1. TPD spectra for AC samples: (a) CO evolution and (b) CO2 evolution.

AC1 AC2 AC3 AC4 AC5

ARTICLE IN PRESS
2048 P.C.C. Faria et al. / Water Research 38 (2004) 20432052 Table 4 Proximate and elemental analyses of the AC samples Sample Proximate analysis Moisture (wt%) AC1 AC2 AC3 AC4 AC5 3.3 4.2 3.7 1.4 2.3 Volatiles (wt%) 5.3 13.7 9.5 5.2 7.0 Cxed (wt%) 85.9 78.7 81.5 88.4 86.0 Ash (wt%) 5.5 3.4 5.3 5.0 4.7 Elemental analysis C (wt%) 94.9 78.5 90.8 98.1 93.9 H (wt%) 0.3 0.8 0.4 0.3 0.2 N (wt%) 0.2 0.3 0.3 0.4 0.2 S (wt%) 1.0 0.7 0.9 0.5 0.8 O (wt%) 3.6 19.7 7.6 0.7 4.9

volatiles content, which is also corroborated by the increase in oxygen content obtained by elemental analysis. Some of the ash content in sample AC2 may have been removed during the oxidation in liquid phase. As mentioned before, the treatment with H2O2 originates less acidic samples. This is conrmed by the results of Table 4. There is a smaller increase both in volatile matter and oxygen content in sample AC3 than in sample AC2 comparatively to the commercial carbon. Comparing samples AC4 and AC5 with the associated starting material (AC2), a decrease in both volatile matter and oxygen content is evident, which is in agreement with the previous results. Both thermal treatments result in the removal of the oxygen groups from the surface of the carbon. Even though sample AC5 has a higher amount of oxygen than the commercial carbon (AC1), it is noticeable that after the thermal treatments the activated carbon does not recover the oxygen groups initially present. Elemental analysis of sample AC4 revealed that this material is almost free of oxygen. Summarising the results obtained from the different techniques, it is clear that sample AC2 has a strong acid character while samples AC4 and AC5 have basic characteristics. 3.2. Equilibrium adsorption isotherms

400

300

qe (mg/gAC)

200

100

(a)

0 0 200 400 600 800 1000

Ce (mg/L)
300

qe (mg/gAC)

200

100

(b)
800

200

400

600

800

1000

Ce (mg/L)

qe (mg/gAC)

The adsorption isotherms at room temperature of the three selected dyes on the activated carbon samples are presented in Fig. 2. The results were analysed using Langmuir and Freundlich isotherms: KL Ce qe qm 1 KL Ce and
1=n qe KF Ce ;

600

400

200

(c)

0 0 100 200 300 400 500 600

respectively, where Ce and qe are the adsorbate equilibrium concentrations in the liquid and solid phases, qm is the maximum adsorption capacity according to the Langmuir model, KL ; KF and n are constants.

Ce (mg/L)
Fig. 2. Adsorption isotherms at room temperature of (a) acid, (b) reactive and (c) basic dyes on commercial and treated ACs: AC1 (&), AC2 (J), AC3 (n), AC4 (,) and AC5 (B).

ARTICLE IN PRESS
P.C.C. Faria et al. / Water Research 38 (2004) 20432052 2049

The model equations can be linearized, leading to Ce 1 1 Ce 3 qm KL qm qe and 1 ln qe ln KF ln Ce n 4

and tted to experimental data. The corresponding parameters are shown in Tables 5 and 6, as well as the standard deviations associated to the linear ts. The Langmuir adsorption isotherm provides the best t for the systems ACs/basic dye. In general, for acid and reactive dyes, both Langmuir and Freundlich models similarly represent the experimental data. However, the Langmuir model was preferred due to the physical meaning attributed to its constants, particularly qm ; which is used for the discussion of the results on the basis of the chemical properties of the samples. The curves associated to the Langmuir model are also represented in Fig. 2. 3.3. Discussion of adsorption results on the basis of the chemical properties of the AC samples Since no drastic changes were made in the textural properties of the adsorbents, the disparity in dye uptakes for the different activated carbons may be explained
Table 5 Parameters of the Langmuir model Sample Acid dye qm (mg/gAC) AC1 AC2 AC3 AC4 AC5 KL (dm /mg)
3

almost entirely by the differences in the chemical nature of the prepared samples. For basic dyes, no signicant correlation could be established between the textural properties and the colour removal. For instance, with samples AC1, AC2 and AC3, the dye removal increased with the decrease in both Wmicro and Smeso ; as shown in Fig. 3. Similar behaviour was observed for samples AC4 and AC5, in relation to Smeso : These observations demonstrate that the textural properties of the studied materials cannot explain the behaviour of these adsorbents towards basic dyes. Nevertheless, the inuence of the adsorbent texture in the dye uptake should not be completely discarded as for acid and reactive dyes weak correlations between the adsorption capacity of the ACs and their textural properties might be found. In contrast, strong correlations between acidity/basicity and maximum adsorption capacity for acid and reactive dyes are clearly shown in Fig. 4 and will be discussed below. Activated carbons are materials with amphoteric character; thus, depending on the pH of the solution, their surfaces might be positively or negatively charged. At pH>pHpzc the carbon surface becomes negatively charged favouring the adsorption of cationic species. On the other hand, adsorption of anionic species will be favoured at pHopHpzc [23]. In the present work the initial pH of the aqueous dyestuff solutions was set to

Reactive dye SD
a

Basic dye KL (dm /mg) 0.052 0.030 0.034 0.061 0.055

3

qm (mg/gAC)

SD

qm (mg/gAC) 546 633 568 714 680

KL (dm3/mg) 0.963 0.994 0.640 0.828 0.955

SDa 0.004 0.002 0.005 0.002 0.003

310 0.120 0.010 190 197 0.025 0.085 157 244 0.057 0.081 201 345 0.058 0.032 246 262 0.093 0.030 223 r  PN a calc 2 =N 1 ; y C =q : SD e e i 1 y y

0.074 0.156 0.157 0.050 0.081

Table 6 Parameters of the Freundlich model Sample Acid dye n AC1 AC2 AC3 AC4 AC5 KF mg=gAC dm =mg
3 1=n

Reactive dye SD
a

Basic dye
3 1=n

n 4.9 5.7 5.8 4.9 5.1

KF mg=gAC dm =mg 53 49 65 71 66

SD

n 23 17 19 13 16

KF mg=gAC dm3 =mg1=n 431 469 424 490 495

SDa 0.049 0.081 0.041 0.090 0.088

11 175 0.037 5.6 60 0.051 7.6 106 0.052 9.8 177 0.016 6.7 107 0.078 r  PN a calc 2 =N 1 ; y ln q : SD e i 1 y y

0.131 0.049 0.058 0.117 0.166

ARTICLE IN PRESS
2050 P.C.C. Faria et al. / Water Research 38 (2004) 20432052

800
AC4 AC5 AC2

400
AC4

r = 0.888
AC3

AC1 AC5

600

300
AC3

AC1

qm (mg/gAC )

qm (mg /gAC)

AC4

200

AC2 AC3 AC2 AC1

AC5

400

r = 0.899

100

200
0 0 100 200 300 400 500

(a)

0 0.350

0.355 0.360 0.365 0.370

0.375 0.380
800
AC4 AC5

Basicity (eq/gAC )

Wmicro (cm /g)

800
AC4 AC 5

AC2

600

AC1

AC3

600

AC3

qm (mg /gAC)

AC 2

r = 0.992
AC1

400

qm (mg/gAC )

400

200

200

200

400

600

800

Acidity (eq /gAC )

(b)

0 90 95 100 105 110 115 120 125 130

Smeso (m /g)
Fig. 3. Maximum adsorption capacity for the basic dye versus Wmicro (a) and Smeso (b) of the AC samples.

Fig. 4. Correlation between samples acidity/basicity and maximum adsorption capacity for acid (&), reactive (J), and basic (,) dyes.

67. If the electrostatic interactions were the main adsorption mechanism, cationic (basic) dyes would be expected to have greater afnity for samples AC2 and AC3 (pHpzcopH) and anionic (reactive and acid) dyes would preferably adsorb on samples AC1, AC4, and AC5 (pHpzc>pH). Nonetheless, two parallel adsorption mechanisms, one involving electrostatic interactions and a second one involving dispersive interactions were report to explain the adsorption of organics in activated carbons [24]; thus, one cannot only account for the role of electrostatic interactions. For the acid and reactive dyes, a strong correlation between the basicity of the AC samples and the dye uptakes is shown in Fig. 4, which is in agreement with the previous reasoning. Comparing both heat-treated samples it is evident that sample AC4 has the best adsorption capacity for both acid and reactive dyes. According to Leon y Leon et al. [22], basic carbons are characterised by a high content of electron rich sites on their basal planes and a low concentration of electron

withdrawing groups. The interaction of the molecules of dyes and the activated carbon surface is expected to occur between the delocalised p electrons of the oxygenfree Lewis basic sites and the free electrons of the dye molecule present in the aromatic rings and multiple bonds. The same authors have reported evidence for the protonation of basal plane sites on basic carbon. They state that oxygen-free carbon sites can adsorb protons from the solution, conferring a positively charged surface to the carbon. Thus, it is possible that negatively charged ions of the dyes also interact with these sites. The presence of oxygen-containing functional groups, which are electron-withdrawing groups, has a negative effect on the adsorption of anionic species, as can be observed for sample AC2. These groups reduce the electron density on the surface of the carbon [24], thus decreasing the adsorption potential for the dye molecules. Simultaneously there is an increase in repulsive electrostatic interactions between the anions of the dyes and the negatively charged surface of sample AC2. When dissolved in water, basic dyes are positively charged; thus it was expected that these dyes would preferably adsorb onto activated carbons with acid characteristics. In fact, when comparing the behaviour

ARTICLE IN PRESS
P.C.C. Faria et al. / Water Research 38 (2004) 20432052 2051

of anionic and cationic dyes, we observe that sample AC2 has a better performance for the latter ones, relatively to AC1. This was expected due to the electrostatic interactions between the cations and the negatively charged surface of acid samples. For basic dyes there is a strong correlation between the dye uptake and the increasing acidity of samples AC1, AC3 and AC2 (Fig. 4). However, samples AC4 and AC5 provide higher adsorption uptakes than the previous samples. This observation leads to the conclusion that, even though adsorption of cationic species on activated carbons is enhanced by the presence of surface acid groups, basic carbons still present better performances, indicating that dispersive interactions between the delocalised p electrons on the surface of the basic activated carbons and the free electrons of the dye molecule present in the aromatic rings and multiple bonds play a dominant role in the adsorption mechanism.

Acknowledgements The authors are indebted to NORIT N.V., Amers! Morgado and foort, The Netherlands, and to Jose ! ! Antonio Vieira from CITEVE (Centro Tecnologico das ! rio de Portugal) for # xtil e do Vestua Industrias Te providing samples. This work was carried out with the # ncia e a Tecnologia * o para a Cie support of Funda@a (FCT) under programme POCTI/FEDER (POCTI/ # ncia de Inova@a * o (ADI) under pro1181) and Age gramme POCTI/FEDER (project ANOXITRATA).

References
[1] USEPA, 1996. Best management practices for pollution prevention in the textile industry. EPA/625/R-96/004. Cincinnati, OH: US Environmental Protection Agency. [2] Cooper P. Removing colour from dyehouse wastewaters a critical review of technology available. J Soc Dyers Colour 1993;109:97100. [3] Marmagne O, Coste C. Color removal from textile plant efuents. Am Dyestuff Reporter 1996;85:1521. [4] Rott U, Minke R. Overview of wastewater treatment and recycling in the textile processing industry. Water Sci Technol 1999;40:13744. [5] Bhavani KD, Dutta PK. Physico-chemical adsorption properties on chitosan for dyehouse efuent. Am Dyestuff Reporter 1999;88:538. [6] Bousher A, Shen X, Edyvean RGJ. Removal of coloured organic matter by adsorption onto low-cost waste materials. Water Res 1997;31:208792. [7] Choy KKH, McKay G, Porter JF. Sorption of acid dyes from efuents using activated carbon. Resour Conservat Recycling 1999;27:5771. [8] Walker GM, Weatherley LR. Adsorption of acid dyes onto granular activated carbon in xed beds. Water Res 1997;31:2093101. [9] Walker GM, Weatherley LR. CDO removal from textile industry efuent: pilot plant studies. Chem Eng J 2001; 84:12531. [10] Bansal RC, Donnet JB, Stoeckli F. Active carbon. New York: Marcel Decker; 1988. ! a * o JJM. [11] Figueiredo JL, Pereira MFR, Freitas MMA, Orf Modication of the surface chemistry of activated carbons. Carbon 1999;37:137989. [12] Karanl T, Kilduff JE. Role of granular activated carbon surface chemistry on the adsorption of organic compounds. 1. Priority pollutants. Environ Sci Technol 1999;321724. [13] Al-Degs Y, Khraisheh MAM, Allen SJ, Ahmad MN. Effect of carbon surface chemistry on the removal of reactive dyes from textile efuent. Water Res 2000;34: 92735. ! a * o JJM, Figueiredo JL. [14] Pereira MFR, Soares SF, Orf Adsorption of dyes on activated carbons: inuence of surface chemical groups. Carbon 2003;41:81121. [15] Papirer E, Li S, Donnet JB. Contribution to the study of basic surface groups on carbon. Carbon 1987;25:2437.

4. Conclusions The following conclusions can be drawn from the present work: 1. The Langmuir adsorption isotherm provides the best t for the systems ACs/basic dye. The adsorption data for the remaining systems are well tted with both Langmuir and Freundlich models. 2. It is possible to tailor the surface of activated carbons in order to optimise their adsorption capacity towards dye molecules. Among the several treatments performed, thermal treatment with H2 at 700 C proved to be the most efcient for the removal of both anionic and cationic dyes. 3. The differences in the textural properties of the activated carbons cannot explain the disparity in dyes adsorption on the samples, leading to the conclusion that surface chemistry plays a key role in the adsorption process. 4. For reactive and acid dyes, a close relationship between the surface basicity of the adsorbents and dye adsorption was shown. The interaction between the oxygen-free Lewis basic sites and the free electrons of the dye molecule, as well as the electrostatic interactions between the anions and the protonated sites of the carbon are the main adsorption mechanisms. 5. For the basic dye, the acid oxygen-containing surface groups have a positive effect on the adsorption, but thermally treated samples still present better performances, showing the existence of two adsorption mechanisms involving electrostatic and dispersive interactions.

ARTICLE IN PRESS
2052 P.C.C. Faria et al. / Water Research 38 (2004) 20432052 tions by adsorption of bacteria and their effect on aqueous lead adsorption. J Chem Technol Biotechnol 2001;76: 120915. ! ! MV, Carrasco-Mar! Moreno-Castilla C, Lopez-Ram on n F. Changes in surface chemistry of activated carbons by wet oxidation. Carbon 2000;38:19952001. Leon y Leon CA, Solar JM, Calemma V, Radovic LR. Evidence for the protonation of basal plane sites on carbon. Carbon 1992;30:797811. Rodr! guez-Reinoso F. The role of carbon materials in heterogeneous catalysis. Carbon 1998;36:15975. ! ndez JA, Leon y Leon Radovic LR, Silva IF, Ume JI, Mene CA, Scaroni AW. An experimental and theoretical study of the adsorption of aromatics possessing electron-withdrawing and electron-donating functional groups by chemically modied activated carbons. Carbon 1997; 35:133948. ! ndez JA, Phillips J, Xia B, Radovic LR. On the [16] Mene modication and characterization of chemical surface properties of activated carbon: in the search of carbons with stable basic properties; Langmuir 1996;440410. [17] Rodr! guez-Reinoso F, Martin-Martinez JM, Prado-Burguete C, McEnaney B. A standard adsorption isotherm for the charaterization of activated carbons. J Phys Chem 1987;91:5156. [18] Linares-Solano A.Textural characterization of porous carbons by physical adsorption of gases. In: Carbon and coal gasication, Nato ASI Series, E105. Figueiredo JL, Moulijn JA, editors. Dordrecht: Martinus Nijhoff Publishers, 1986. [19] Stoeckli HF, Ballerini L, De Bernardini S. On the evolution of micropore widths and areas in the course of activation. Carbon 1989;27:5012. [20] Rivera-Utrilla J, Bautista-Toledo I, Ferro-Garc! a MA, Moreno-Castilla C. Activated carbon surface modica-

[21]

[22]

[23] [24]