Food

Chemistry
Food Chemistry 100 (2007) 1209–1216
www.elsevier.com/locate/foodchem

Effects of processing on the nutraceutical profile of quinoa

Kevin Brady, Chi-Tang Ho, Robert T. Rosen, Shengmin Sang, Mukund V. Karwe *

Department of Food Science, Rutgers, The State University of New Jersey, 65 Dudley Road, New Brunswick, NJ 08901, USA

Received 18 March 2005; received in revised form 5 November 2005; accepted 1 December 2005

Abstract

Quinoa flour was subjected to a variety of thermal processes. Both unprocessed and processed quinoa samples were subjected to suc-
cessive extractions in methanol and ethyl acetate solvents. Effects of processing were gauged via comparison of HPLC chromatograms of
the quinoa extracts. Quinoa flour subjected to processing via roasting and extrusion resulted in a significant impact on the chemical
profile when compared to unprocessed quinoa flour. Steam pre-conditioning had minimal effects on the chemical profile of quinoa flour.
Process-enhanced isolate from roasted and extruded quinoa possessed a molecular weight of 480 D, was dissolved in methanol and ethyl
acetate, and possessed a NMR spectra reminiscent of a terpenoid compound. This research suggests that thermal processing of
quinoa flour can result in degradation of saponin molecules. Saponin decomposition may influence sensory or pharmacological
properties.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Quinoa; Chenopodium quinoa; Roasting; Processing; Extrusion; Saponins

1. Introduction referred to as a pseudo-cereal since it is not a member of

Quinoa (Chenopodium quinoa) is a pseudo-cereal with
origins dating to the Incas. The pre-Colombian Andean
people used the seed as a staple food component, and at
times, replaced the animal protein in their diet with quinoa
(Koziol, 1992). Today, quinoa is mainly cultivated in
Argentina, Bolivia, Chile, Colombia, Ecuador, and Peru,
locations that mostly coincide with the limits of the Inca
Empire. Quinoa has also shown promise in tests of farm-
scale cultivation in high altitudes of Colorado, and near
sea level in Washington and Oregon, as well as in England
and in Scandinavia (National Research Council, 1989).
The quinoa plant grows from 3 to 6 ft., in height and
bears leaves that extend from the stalk. Quinoa seeds grow
in large clusters at the end of the stalk, and seed color var-
ies from pink, orange, red, purple, tan, to black. Quinoa
seeds, shape of which resembles sesame seeds, can be con-
sumed whole or ground into flour. Quinoa is commonly
*
Corresponding author. Tel.: +1 732 932 9611x224; fax: +1 732 932
6776.
E-mail address: karwe@aesop.rutgers.edu (M.V. Karwe).
the grass family, but produces seeds that can be milled into
flour and used much like a cereal crop.
The shortage of provisions in needy countries can poten-
tially be solved by cultivation of quinoa (or otherwise act
as an ideal human food supplement) since it can survive
the most barren farming conditions (potential for cultiva-
tion in marginal lands and temperate regions) while main-
taining a high nutritive value and supplying a reasonable
yield (Prakash, Nath, & Pal, 1993). Quinoa can survive
low rainfall, high altitudes, thin cold air, hot sun, sub-freez-
ing temperatures, salinity, and poor sandy alkaline soils
(Przybylski, Chauhan, & Eskin, 1994). The adaptability
of different cultivars to salt stress merits special note
(National Research Council, 1989). However, growth for
most strains is optimized with short day lengths and cool
temperatures.
Quinoa has economical potential to be realized since the
entire plant can be utilized. The leaves can be consumed in
a salad and are rich in various minerals. The seeds can be
consumed whole or ground into flour for a variety of food
applications (Coulter & Lorenz, 1990). Seeds are usually
soaked in water to remove the soapy saponins, bitter

0308-8146/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2005.12.001
1210 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216
glucoside–triterpenoid compounds that cover the surface of
the seeds. The residual water can be used to shampoo the
hair. The leaves, roots, and stalks can be used for animal
fodder.
Research efforts in quinoa have been focused on its
chemical composition, with considerable attention paid
to saponins in quinoa. There has been some work accom-
plished in processing quinoa, focused mainly on the
effects of dehulling and washing on changes in chemical
composition; specifically removal of saponins. Usually,
quinoa is processed by means of soaking, rubbing, rins-
ing, and boiling in the domestic setting. It is industrially
processed by means of wet and dry milling (Becker &
Hanners, 1990).
Today’s health conscious consumers are illustrating a
preference towards value added products, and in general,
more nutritious food items. The opportunity to supplement
or completely replace common cereal grains (corn, rice and
wheat) with a cereal of higher nutritional value (such as
quinoa) is inherently beneficial to the public interests.
In this study, quinoa flour was subjected to steam pre-
conditioning, extrusion, and roasting. Effects of these
processes on quinoa were determined via HPLC analysis.
Process enhanced compounds were characterized with
instrumental analysis.
2. Materials and methods
2.1. Materials
Quinoa. Quinoa seeds (Chenopodium quinoa Willd) were
purchased from Quinoa Corporation, Gardena, CA.
Solvents. HPLC grade methanol, HPLC grade ethyl ace-
tate, HPLC grade acetonitrile, HPLC grade water, and for-
mic acid were purchased from Sigma-Aldrich, St. Louis,
USA.
Extraction medium. Sample extraction for HPLC analy-
sis was prepared with Whatman # 42 filter paper, ashless
70 mm circles. Acrodisc 0.45 lm · 3 mm syringe filters fur-
ther refined extracts prior to direct HPLC injection.
Table 1
Screw configuration and extrusion elements
Extrusion zones*
Feed zone (84 mm) Zone 1 (196 mm) Zone 2 (210 mm)
SK 42/42 42/21 T 28/28
SK 42/42 42/42 28/28
42/42 IGEL 42
42/21 28/28
IGEL 42 28/28
28/28 KB 45/5/28
28/14
28/14
KB, kneading block; LH, left-handed (reverse) element; SK, feed element; T, transition element.
*
IGEL, mild kneading element.
2.2. Methods
Milling. Quinoa seeds were pulverized with a model D
Fitz Mill (W.J. Fitzpatrick Co., Chicago, USA) to produce
quinoa flour.
Steam pre-conditioning. Steam pre-conditioning was
performed with a 57 l Varimixer (Welbilt, South Plainfield,
NJ, USA) modified for steam injection. 11 kg of milled qui-
noa flour were mixed while being exposed to 100 C steam
for 15, 30, 45 and 60 min. Temperature of the flour was
recorded using an Omega Microprocessor Thermometer,
model HH21, complete with Omega type T thermocouple
(Omega Engineering, Stamford, CT, USA). The thermo-
couple was placed at the center of the kettle.
Extrusion. Extrusion was performed with a twin screw
extruder (model ZSK-30 from Coperion, formerly Werner
and Pfleiderer Corp., Ramsey NJ, USA). The extruder has
two co-rotating, self wiping screws (30.7 mm diameter,
4.7 mm channel depth, and 878 mm processing length; L/
D = 28.6) in a steel barrel with five zones. Each zone is
heated by resistive electric heaters and the temperature of
each zone can be controlled independently. The screw con-
figuration used in extrusion experiments consisted of for-
ward conveying elements, mild mixing elements, kneading
elements, and reverse elements (Table 1). The die had
two circular orifices (3 mm diameter, 5 mm long). Quinoa
flour was metered into the feed section of the extruder with
a volumetric feeder from K-Tron Corporation (Pitman,
NJ, USA). Water was injected into the feed section of the
extruder immediately after the feed port using a triple
action piston pump from US Electric Corporation
(Milford, CT, USA). Both the feeder and the pump were
calibrated prior to extrusion runs to determine the set
points required for desired mass flow rates of quinoa flour
and water, respectively. The total mass flow rate (flour +
water) was kept constant at 300 g/min for all experiments.
Temperatures at zones I, II, and III were set to room
temperature, 80 and 120 C, respectively, while the temper-
atures at zones IV and V were adjusted such that the
desired die temperatures could be maintained.
Zone 3 (178 mm) Zone 4 (98 mm) Zone 5 (84 mm) Die zone (28 mm)
28/14 20/10 KB 45/5/14 14/14
KB 45/5/14 20/10 KB 45/5/14 LH 14/14
20/20 KB 45/5/20 14/14
20/20 20/10 LH 14/14
20/20 20/10 14/14
20/20 20/10 14/14
20/20 14/14
KB 45/5/20 14/14
20/10 LH
20/10
20/10
 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216 1211

Extrusion was performed at a range of temperatures, ner funnel and filter paper. Liquid residues from successive
moisture contents, and screw speeds. The die temperature extractions in methanol were merged into a 500 ml flask.
was varied between 130 and 170 C, moisture content Solid residues were discarded. The methanol extracts (in
between 16% and 24%, and screw speed between 250 and the 500 ml flask) were exposed to rotary evaporation to
500 rev/min. completely remove methanol solvent, using a Buchi Rota-
Roasting. Roasting was performed with a Isotemp Vac- vapor R110 with Meyer N-Evap Analytical Evaporator
uum Oven, model 282A from Fisher Scientific (Pittsburgh, water bath, model 112.
PA, USA). Approximately 100 g of milled quinoa flour was The dried extract was then subjected to 100 ml ethyl ace-
placed uniformly in a baking sheet, 34 cm · 24 cm tate and magnetically stirred for 24 h. After exposure to ethyl
· 1.5 cm, at 200 C for 10 min. acetate, sample was vacuum filtered with Buchner funnel and
Extraction for HPLC analysis. Fifty grams of quinoa filter paper. Liquid residue was covered and set aside. Solid
sample were placed into a 500 ml Erlenmeyer flask with
150 ml of methanol and a magnetic rod. After sample
was stirred, using a Lab-Line Multi-Magnestir, and Table 2
exposed in methanol for 24 h, the slurry was subjected to HPLC program gradient
vacuum filtration with a Buchner funnel and filter paper Time (min) Solvent A (%) Solvent B (%)
(Whatman # 42). Solid sample residue from the filter was
Initial 80 20
scraped off and placed back into the original Erlenmeyer 25 10 90
flask. Filtered liquid residue was set aside and covered. 55 10 90
The solid residue was subjected to a second extraction with 60 80 20
methanol (100 ml). After sample was magnetically stirred Solvent A, HPLC grade acetonitrile; Solvent B, solution of 0.05% formic
for 24 h, vacuum filtration was performed again with Buch- acid in HPLC grade water.

Fig. 1. HPLC chromatogram of the extract of unprocessed quinoa.

1212 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216
residue in flask was exposed to a successive extract with ethyl
acetate (100 ml) and stirred for 24 h. After exposure, vacuum
filtration was performed again with Buchner funnel and filter
paper. Liquid residues from successive extractions in ethyl
acetate were merged into a 500 ml flask. The ethyl acetate
extracts (in the 500 ml flask) were exposed to rotary evapora-
tion to completely remove ethyl acetate solvent.
Residue (in flask) was exposed to methanol aliquots and
conveyed to a 25 ml graduated cylinder using a 3 ml syringe
with Acrodisc 0.45 lm filter. Process was repeated (approx-
imately seven times) until 15 ml of methanol dissolved-qui-
noa extract was obtained. The 15 ml filtered extracts were
transferred into labeled vials and placed in a 5 C freezer.
Moisture analysis. Moisture analysis was performed
with a moisture analyzer from Sartorius (model MA 30,
Edgewood, NY, USA). Approximately 2 g of pre-
conditioned sample was uniformly placed in an aluminum
dish and placed in the unit. The unit drives off moisture
using a 130 C oven and automatically calculates the
moisture content of the sample in approximately 15 min.
TLC. TLC was performed on Sigma-Aldrich silica gel
TLC plates (250 lm thickness, 2–25 lm particle size).
Fig. 2. HPLC chromatogram of the extract of quinoa flour exposed to extrusion at die temperature of 170 C, screw speed of 375 rev/min, and 16% (w/w)
moisture content.
Compounds were visualized by spraying a 10% (v/v)
H2SO4 ethanol solution under heat (Zhu et al., 2002).
HPLC. Quinoa extracts were analyzed with reverse
phase high performance liquid chromatography (HPLC).
Mobile phases, including HPLC grade acetonitrile (Solvent
A) and a solution of 0.05% formic acid in HPLC grade
water (Solvent B), were conveyed with two Waters, model
510, pumps. A Discovery C18 column by Supelco (25 cm ·
4.6 mm · 25 mm) was utilized for compound separation.
The program gradient (determines the mobile phase com-
position), encoded by a Waters Automatic Gradient Con-
troller, is available in Table 2. Twenty microliter samples
were loaded into the Waters (Milford, MA, USA) model
U6K, injector port. Analysis of samples was performed
with a Waters, model 484, Tunable Ultraviolet Absorption
Detector at 254 nm wavelength. Data was interpreted and
printed with a Spectra-Physics, model SP4290 integrator.
Nuclear magnetic resonance spectroscopy. NMR spectra
were recorded with a Varian AM-600 NMR spectrometer
(Varian, Inc., Palo Alto, CA), with TMS as an internal
standard . Isolated compounds were dissolved in methanol
for carbon NMR spectroscopy. Analyses were performed
 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216
at the City University of New York, Staten Island, New
York.
Mass spectrometry. The mass spectra of the isolated pro-
cess enhanced compound were obtained using both chemi-
cal ionization and atmospheric pressure chemical ionization.
Chemical ionization mass spectrum was recorded with a
Finnigan MAT 8230 (Finnigan, Bremen, Germany) high-
resolution, double focusing magnetic sector mass spectro-
meter using direct probe sample injection. Atmospheric
pressure chemical ionization (positive ion mode) mass spec-
trum was recorded with a Micromass Platform II LC-MS
instrument, sample was introduced through an empty col-
umn and methanol was used as a solvent.
3. Results and discussion
Pre-conditioning with steam, before extrusion, is often
used to make extruded products. Pre-conditioning with
steam increases the temperature and moisture content of
the raw materials going into the extruder, thereby reducing
the load on the extruder. It allows additional residence time
for the moisture to penetrate into the flour grains.
Fig. 3. HPLC chromatogram of the extract of quinoa flour exposed to roasting for 10 min at 200 C.
1213
Fig. 1 shows the HPLC chromatogram of raw quinoa
extract. This chromatogram illustrates the profile of com-
pounds that are eluted from the HPLC between 0 and
44.7 min. A large peak area correlates to a strong presence
of a particular compound, and a small peak area correlates
to a weak presence of a particular compound. If the detec-
tor does not register the presence of a compound, there will
be the absence of a peak on the chromatogram.
The HPLC chromatograms of extracts of quinoa flour
processed by steam for 15, 30, 45, and 60 min are nearly
identical to the chromatogram of unprocessed quinoa
extract. Therefore, it was concluded that processing by
means of steam pre-conditioning results in no significant
impact on the chemical profile of quinoa. The minimal
effect of steam pre-conditioning on the chemical profile of
quinoa can probably be attributed to the relatively mild
processing temperatures.
Processing by means of extrusion significantly impacted
the chemical profile of quinoa flour as compared to raw qui-
noa extract. Fig. 2 shows the HPLC chromatogram of
extruded quinoa extract. This chromatogram illustrates the
profile of compounds that are eluted from the HPLC
1214 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216

between 0 and 45.52 min. A duplex of peaks, with retention
times of approximately 18 min, was enhanced extensively by
extrusion. In addition, peaks with retention times between 25
and 30 min, were diminished considerably via extrusion.
This phenomenon was magnified under application of a
more intense extrusion process. Extrusion processes con-
ducted at a higher temperature, lower moisture content,
and faster screw speed produced a more dramatic escalation
of peak height and area for the duplex of peaks eluted at
18 min, and a more dramatic reduction of peak height and
area for the series of peaks eluted between 25 and 30 min.
The extrusion conditions, which revealed the most dramatic
changes in the profile, were die temperature of 170 C, screw
speed of 375 rev/min, and 16% (w/w) moisture content.

Fig. 4. APCI-LC-MS spectra for isolated process enhanced compound (retention time at 18.09 min in Fig. 2).

Fig. 5. CI-MS spectra for isolated process enhanced compound (retention time at 18.09 min in Fig. 2).
 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216
Fig. 6. NMR spectra for isolated process enhanced compound (retention time at 18.09 min in Fig. 2).
The roasting process also produced significant changes
in the chemical profile of quinoa flour, as compared to
raw quinoa flour. Fig. 3 shows the HPLC chromatogram
of roasted quinoa extract. This chromatogram illustrates
the profile of compounds that are eluted from the HPLC
between 0 and 38.38 min. A duplex of peaks, with retention
times of approximately 19 min, was enhanced extensively
by roasting. A series of peaks that elute from 10 to
14 min, and another series that elutes from 35 to 38 min,
were also enhanced via roasting.
Of significant interest was the process enhanced com-
pounds that appeared in both roasting and extrusion pro-
cesses. These processes significantly enhanced a duplex of
peaks with a retention time of 18 min. Extrusion and roasting
exposed flour to much higher temperatures, which presum-
ably resulted in the dramatic changes of profile for extruded
and roasted quinoa. These compounds were isolated and
accumulated by HPLC for characterization studies.
Mass spectra revealed the isolate to have a molecular
weight of 480 D. The molecular weight was confirmed
using APCI-LC-MS and CI-MS techniques. Spectra for
both MS techniques can be seen in Figs. 4 and 5. Previous
research (Gee, Price, Ridout, Wortley, & Hurrell, 1993),
which examined saponins in processed quinoa, identified
an aglycon compound with a molecular weight of 480 D
using FAB-MS. The compound was a mass fragment of
phytolaccagenic acid, present in the methanolic extract of
bitter quinoa.
1215
Examination of the NMR spectra (Fig. 6) reveals that
the spectral signal has characteristic of a terpenoid
structure.
TLC of the isolate displayed a white color when exposed
to UV radiation. This color is indicative of a terpenoid
compound.
4. Conclusions
Analysis of the extrusion and roasting enhanced isolate
strongly suggests characterization as terpenoid compound.
Extraction was performed using methanol and ethyl ace-
tate, solvents with the ability to dissolve terpenoids. HPLC
profile indicated a strong presence of the compound at
wavelength of 254 nm. MS data reveals a molecular weight
of 480 D, which is in the range of triterpenoid structures.
NMR data reveals a terpenoid signature in the spectra.
Saponins in quinoa are of the triterpenoid variety. The
experimentally determined molecular weight of the
enhanced compound suggests it may be the dehydration
product of phytolaccagenic acid, the aglycon of a major
saponin found in quinoa. It is hypothesized that saponins
present in raw quinoa flour are cleaved and dehydrated
when placed under the extreme heat duress produced in
extrusion and roasting processes. This could also explain
why compounds enhanced in extrusion and roasting pro-
cesses were not enhanced in quinoa pre-conditioned with
steam.
1216 K. Brady et al. / Food Chemistry 100 (2007) 1209–1216
Saponin degradation would directly alter sensory per-
ception and pharmacological properties. The bitter taste
imparted by saponins could potentially be reduced through
extrusion and roasting processes. The nutraceutical proper-
ties of saponins could also be reduced through extrusion
and roasting processes.
Acknowledgments
This paper is dedicated to our co-author Robert T.
Rosen who passed away in May 2005. The research was
supported by a Grant from the Center for Advanced Food
Technology, Rutgers University, NJ, USA. Authors thank
Mr. Joe Lech for technical support.
References
Becker, R., & Hanners, G. (1990). Compositional and nutritional
evaluation of quinoa whole grain flour and mill fractions. Lebensmit-
tel-Wissenschaft Und-Technoology, 23, 441–444.
Coulter, L., & Lorenz, K. (1990). Quinoa-composition, nutritional value,
food applications. Lebensmittel-Wissenschaft Und-Technoology, 23,
203–207.
Koziol, M. (1992). Chemical composition and nutritional evaluation of
quinoa (Chenopodium quinoa Willd.). Journal of Food Composition and
Analsis, 5, 35–68.
Gee, J., Price, K., Ridout, C., Wortley, G., & Hurrell, R. (1993). Saponins of
quinoa (Chenopodium quinoa): effects of processing on their abundance
in quinoa products and their biological effects on intestinal mucosal
tissue. Journal of the Science of Food and Agriculture, 63, 201–209.
National Research Council, 1989. Lost Crops of the Incas, Little-Known
Plants of the Andes with Promise for Worldwide Cultivation, Report
by the Advisory Committee on Technology Innovation, Board on
Science and Technology for International Development. National
Academy Press, Washington, DC.
Prakash, D., Nath, P., & Pal, M. (1993). Composition, variation of
nutritional contents in leaves, seed protein, fat and fatty acid profile of
Chenopodium species. Journal of the Science of Food and Agriculture,
62, 203–205.
Przybylski, R., Chauhan, G., & Eskin, N. (1994). Characterization of
quinoa (Chenopodium quinoa) lipids. Food Chemistry, 51, 187–192.
Zhu, N., Sheng, S., Sang, S., Jhoo, J. W., Bai, N., Karwe, M. V., et al.
(2002). Triterpene saponins from debittered quinoa (Chenopodium
quinoa) seeds. Journal of Agricultural and Food Chemistry, 50, 865–867.