Ind. Eng. Chem. Res.

2002, 41, 6751-6758

6751

Extraction of Caffeine, Theobromine, and Cocoa Butter from

Brazilian Cocoa Beans Using Supercritical CO2 and Ethane
Rahoma S. Mohamed,* Marleny D. A. Saldan
a, and Paulo Mazzafera
School of Chemical Engineering, State University of Campinas-Unicamp, C.P. 6066, CEP:
13083-970, Campinas, SP, Brazil

Carsten Zetzl and Gerd Brunner

Arbeitsbereich Termische Verfahrenstechnik, Technical University Hamburg-Harburg,
Eisendorfer Strasse 38, D-21071 Hamburg, Germany

Supercritical extraction using ethane and CO2, acceptable solvents for food products, was explored
for the recovery of the methylxanthines caffeine and theobromine and cocoa butter from cocoa
beans using a high-pressure apparatus. Continuous extraction of cocoa beans was performed at
343.2 K using CO2 at pressures of 20 and 40 MPa and ethane at pressures of 15.2, 24.8, and
28.3 MPa. The extraction yields of cocoa butter obtained with ethane were much higher than
those obtained with CO2 because of the higher solubility of this fat in ethane. A pronounced
effect of pressure on the extraction of methylxanthines and cocoa butter was observed for both
solvents. Extraction curves revealed the greater facility of these solvents to extract cocoa butter
followed by caffeine and theobromine. This behavior suggests a range of possible conditions
under which the extraction and isolation of cocoa butter, caffeine, and theobromine from cocoa
beans can be achieved. The methylxanthines in cocoa beans were slightly more soluble in ethane
than in CO2 probably because of co-solvency effects of cocoa butter, which was extracted more
easily using supercritical ethane. Despite the higher cost of ethane, its critical pressure is lower
than that of CO2, and the higher butter solubility could render ethane a viable solvent through
lower energy costs.
Table 1. Composition of Cocoa Beans2,4,5

Introduction
Cocoa beans, Theobroma cacao, a source of a variety
of products such as cocoa powder, chocolate, etc., contain
mainly cocoa butter (45-54%), proteins (11.5%), a
significant amount of theobromine (1.2-1.8 wt %), and
caffeine (0.26 wt %) (Table 1).1-3
The presence of these two alkaloids, theobromine and
caffeine in common beverages such as coffee, tea, and
cocoa is of concern to some consumers because of their
potentially adverse health effects at certain levels of
consumption. The U.S. Food and Drug Administration
has recently warned pregnant women to avoid or
minimize their intake of caffeine, as studies in animals
have suggested a relationship between birth defects and
caffeine intake.4 Theobromine, present in large amounts
in chocolate, has been reported to cause physiological
effects similar to those observed for caffeine, namely,
strong diuresis and cardiac stimulation as well as
arterial dilation.5,6 Cocoa butter is not only an important
product for the food industry but also an ingredient of
many cosmetic and pharmaceutical products.
The extraction of cocoa butter and methylxanthines
(caffeine and theobromine) from natural plants is a
potentially attractive process for the recovery of alkaloids as ingredients in the formulation of different
pharmaceutical products as well as for the production
of high-value methylxanthine-free fat products for human consumption and cosmetic formulations.
* To whom correspondence should be addressed. Phone:
+55 19 3788-3932. Fax: +55 19 3788-3922. E-mail:
mohamed@feq.unicamp.br.

component

wt %

moisture
fat
theobromine
caffeine
protein
starch
cellulose
ash
others

5.0
45-54
1.2-1.8
0.26
11.5
6.0
9.0
2.6
18.5

Dimethyl chloride, chloroform, and water have been

employed for the removal of methylxanthines from
natural plants,7,8 with hexane and petroleum ether
reserved for oil recovery.9 Chemical solvents, however,
require a long time for a complete extraction and almost
always carry with them the risk of toxic residue in the
extracted products.10 Water, although is an excellent
solvent for methylxanthines, is not very selective and
is immiscible with fats and oils under normal conditions.11
Mechanical expulsion has also been employed for
obtaining fats and oils from natural products. This
technique often introduces waste solid contaminants
into the produced butter, thus necessitating the use of
other processes for their removal. Low critical temperature, nontoxicity, and low cost have rendered supercritical CO2 a suitable and environmentally benign
solvent for food products,12-14 that has been successfully
used for the extraction of oils and fats.11,15-21
Supercritical CO2 extraction is successfully used on
a commercial scale for the decaffeination of coffee
beans.21 The application of supercritical fluid technology

10.1021/ie0203936 CCC: $22.00 2002 American Chemical Society

Published on Web 11/19/2002

6752

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Figure 1. Microextraction plant: V, shutoff valves; MV1-MV4, micrometering valves; EC1-EC4, extraction columns; GV1-GV4, glass
vials; R1-R4, rotameters

in biotechnology and the prediction of the solubility of

biomolecules in supercritical solvents have been
reported.22-24 Sambarato25 reported on the potential
extractability of theobromine from cocoa beans using dry
supercritical CO2, whereas water-saturated supercritical
CO2 has been shown to be more effective than dry CO2
in the extraction of caffeine from coffee beans26 and
nicotine from tobacco leaves.27 In coffee beans, caffeine
is chemically complexed with chlorogenic acid, and
water acts to free the caffeine from its bonded form in
the coffee plant matrix.13
Ethane, also an acceptable solvent for food products,
has approximately the same critical temperature as
CO2. Despite its higher cost, it has a lower critical
pressure than CO2 and could, therefore, be used as a
supercritical solvent at lower operating pressures and
consequently lower energy costs. In this work, we
present new experimental data on the extraction of
cocoa butter and methylxanthines from Brazilian cocoa
beans using dry and water-saturated supercritical CO2
as well as ethane.
Experimental Section
Materials. Caffeine and theobromine, 99.9% in purity, were purchased from Aldrich-Chemie (Steinheim,
Germany). CO2, 99.95% in purity, was purchased from
KWD (Bad Honninghen, Germany). Ethane, 99% in
purity, was purchased from White Martins (SP, Brazil).
Cocoa beans were supplied in dried form by the Cardill
Company S.A. (Bahia, Brazil) and were kept in separate
and sealed plastic bags until being ground in a mill
before use. For all experiments, ground cocoa beans
particles ranging in size between 0.3 and 2 mm were
used.
Experimental Apparatus and Procedure. Two
experimental apparatuses that employ the dynamic

method to obtain the equilibrium solubility of solid and

liquid solutes in supercritical solvents were employed
in this study. A description of the two apparatuses and
experimental procedures employed is presented herein.
The procedures assured that the solubility data obtained
from either apparatus were at equilibrium. Extraction
yields are presented on a dimensionless basis, which
allows for a direct comparison of the data obtained with
the two different apparatuses.
The supercritical extraction plant (Figure 1) used for
obtaining data with supercritical CO2 was purchased
from Applied Separation Inc. (Allentown, PA) and is
designed for working pressures and temperatures up
to 69 MPa and 523.2 K, respectively. This microextraction apparatus was previously used by Saldana et al.28
and allows for simple and efficient extractions at supercritical conditions. The plant consists of three modules: an oven, a pump, and a collection module. The
oven module has four 50-cm3 extraction columns. The
pump module is equipped with a compressed air-driven
pump with a constant flow capacity of 90 g/min of liquid
CO2. The collection module is formed of a micrometering
valve for pressure reduction joined to a system of 25mL glass vials for the recovery of extracted products
precipitated upon pressure reduction and consequent
loss of solvent power and a flowmeter for the determination of the amount of supercritical solvent, now in the
gaseous state at ambient conditions. Pressure and
temperature are controlled in the three modules to
within (0.3 MPa and (0.5 K, respectively. The CO2 flow
rate is controlled to within (0.47 gmin-1.
The oil was quantified by weighing, and the methylxanthines (caffeine and theobromine) contents were
determined by HPLC using a Shimadzu LC-6A chromatograph (Kyoto, Japan) with a C18 column (4.6 250
mm, 5 m) ODS-Inertsil. Isocratic solvents composed

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6753

Figure 2. Experimental apparatus: BH1 and BH2, solvent; BH3 and BH4, cosolvent; G1-G4, pressure indicator; P1-P4, pumps; F1F4, filters; microm. V., micrometering valve; CV1-CV8, valves; TC1-TC7, thermocouples; extractor 2, extractor with stirring and a
window; separ. flasks, separator flasks; amst., sample; FM, flow measurement.

of 40% methanol in 0.5% of acetic acid in water were

used at a flow rate of 1 mL/min.11 The methylxanthines
were identified using a UV detector at 280 nm, and their
quantities were determined by a Chromjet integrator.
Two chromatograms for each sample were collected, and
the results of the chromatographic analysis were reproducible to within (2%.
For each experiment, an extraction column (EC3) was
packed with a determined amount of ground particles
of cocoa beans mixed with 3-mm glass beads. Glass wool
was placed at the two ends of the column to avoid any
possible carry over of solid material. In another extraction column (EC4), located upstream, was placed 40 mL
of distilled water to allow for the saturation of the
incoming CO2. When the desired temperature, as controlled by the oven, was reached in both columns, CO2
was pumped to the extraction column until the desired
pressure was reached. Once equilibrium was established, the mixture of CO2 and solute was allowed to
flow to the micrometering valve where the separation/
precipitation of the desired solute occurred as a result
of pressure reduction. The oil content was determined
by weighing, and the alkaloids content by HPLC as
described elsewhere.16,28 The first fraction was collected
with 114 g of CO2 passing through the extractor, and
subsequently, five other fractions were each collected
following the passage of 342 g of CO2 through the
extractor. After each experiment, the transfer lines were
cleaned with hot alcohol and water to ensure the total
collection of all extracted material. The material collected after this cleaning step was analyzed similarly
to that collected earlier, and the total amounts of
extracted methyxanthines and oil were corrected accordingly.
The experimental apparatus used to obtain extraction
data with supercritical ethane was the one previously

used and described by Mohamed et al.14 and Saldana

et al.11 The semicontinuous-flow high-pressure system
used was purchased from Autoclave Engineers (Erie,
PA). The major components of the apparatus included
positive liquid displacement pumps for solvent delivery,
high-pressure extraction vessels, and three separator
flasks in series (Figure 2). Flow rates and accumulated
gas volumes passing through the apparatus were controlled with micrometering valves and measured with
a flow computer-measuring device from EG&G Instrument Flow Technology (Phoenix, AZ). Heating tapes
were used to maintain constant temperature in the
extraction section and in the valves to prevent freezing
of solvents or precipitation of solid solute following
depressurization. The pressures in the two extractors
were monitored with a digital transducer system, Heise
Series 901A RTS, acquired from Dresser Industries
(Stratford, CT) with a precision of (0.03 MPa. Extractor
temperatures were controlled to within (0.5 K.
In extraction experiments with this apparatus, liquid
ethane was pumped into the 30-g cocoa bean sample
mixed with 3-mm glass beads and placed in the extractor until the specified extraction pressure was reached
at the specified temperature. Following a 3-h period, the
saturated supercritical fluid was depressurized when
passed through the micrometring valve, and the precipitated fraction was collected in the tarred separator
flasks placed in a cooling bath maintained at 273.2 K.
The three separator flasks were replaced at different
intervals, and extracted fractions were collected for each
specified amount of 31 g of ethane passing through the
extractor. A total of six fractions were continuously
collected during the extractions carried out at 343.2 K
and each of the following pressures: 15.2, 24.8, and 28.3
MPa. Tubings and valves throughout the apparatus
were cleaned with heated ethanol at the end of each

6754

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Figure 3. (A) Cocoa butter and (B) theobromine and caffeine

extracted using dry CO2 and water-saturated supercritical CO2
(w) at 343.2 K and 20 MPa.

Figure 4. (A) Cocoa butter and (B) theobromine and caffeine

extracted using dry CO2 and water-saturated supercritical CO2
(w) at 343.2 K and 40 MPa.

experiment. Collected samples were weighed and analyzed for cocoa butter and methylxanthines using HPLC,
as described earlier.
Results and Discussions
Figures 3 and 4 show the extraction curves obtained
for cocoa butter, theobromine, and caffeine from a 3-g
sample of cocoa beans using 5.7 g/min of dry and watersaturated supercritical CO2 stream at 343.2 K and
pressures of 20 and 40 MPa, respectively. The extracted
amount is represented as the extracted fraction (weight
percent) of each component with respect to the amount
of the component in the cocoa beans before the extraction. To assess the reliability of the extraction data
obtained with this apparatus, two independent experiments were carried out at 344.2 K and 20 MPa, and
the results were found to be reproducible to within (5%.
The fractions obtained in this work were solids at
room temperature, were light yellow in color (typical of
cocoa butter), and presented a pleasant flavor. In
Figures 3 and 4, it is possible to observe, for all
isotherms, a strong effect of pressure on the extraction
of cocoa butter, theobromine, and caffeine. Increasing
the pressure from 20 to 40 MPa for extractions at 343.2
K using 1.3 kg of CO2 resulted in increases in the
amounts of recovered cocoa butter, theobromine, and
caffeine from 27 to 48%, from 13.9 to 29.9%, and from
37.4 to 66.4% of their contents in the original sample,
respectively.

Figure 5. Extraction of cocoa butter with dry supercritical CO2

at 40 MPa and a temperature of (A) 323.2 and (B) 343.2 K.

Figure 5 shows the extraction curves for the extraction of cocoa butter from cocoa beans using supercritical
carbon dioxide at 40 MPa and the temperatures of 323.2
and 343.2 K. Under these conditions, using a cumulative
amount of 2.2 kg of CO2, which corresponds to 722 kg
of CO2/kg of cocoa bean, 85 and 72%, respectively, of
the cocoa butter (459 and 386 g of cocoa butter/kg of
cocoa beans) were extracted. The better extraction yields
obtained at the lower temperature are attributed to
what is known as retrograde behavior. This behavior
was observed and reported for the extraction of pure
cocoa butter using supercritical ethane by Saldana et

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6755

al.16 and also in the extraction of other oils using

supercritical fluids.18-20
The experimental solubilities of cocoa butter in supercritical CO2 obtained with the cocoa butter/dry CO2
binary system at 333.2 K in the pressure range from
10 to 30 MPa are reported to range from 0.05 to 0.8 wt
%.15 These values are somewhat higher than the 0.06
and 0.17 wt % that can be obtained from the slopes of
the linear portions of our extraction curves at 343.2 K
and pressures of 20 and 40 MPa, respectively (Figures
3 and 4). Similar qualitative trends have already been
observed when comparing solubilities in binary and
ternary systems. Differences of 1-2 orders of magnitude
have been found when comparing solubility values in
binary and quaternary systems.29 This descrepancy in
solubility could be attributed to possible component
interactions present in multicomponent mixtures and
demonstrates the difficulties encountered in attempting
to use binary data to predict the extraction of substances
from natural and complex plant structures.
The results also point clearly to the higher recoveries
of cocoa butter and caffeine from cocoa beans in comparison to theobromine. The solubility data reported by
Saldana et al.11 for the binary model systems caffeine/
CO2 and theobromine/CO2 showed that the solubility
of theobromine in CO2 is about 2 orders of magnitude
lower than that of caffeine, which could explain the
relatively lower recovery efficiencies for the extraction
of theobromine from cocoa beans even at higher pressures than those in the reported binary systems. The
large difference in the solubilities of caffeine and theobromine in CO2 can be attributed to the self-association
of theobromine molecules through hydrogen bonding,
which does not happen with caffeine molecules.30,31 This
is probably the reason for the higher melting temperature (624-630 K) and enthalpy (2346.7 J mol-1) of
theobromine, as compared to those for caffeine (511 K
and 1205.5 J mol-1, respectively). The solubility of a
solid solute in a liquid solvent can be expressed by the
mathematical equation32

ln(2x2) ) -(hf/RT)(1 - T/Tm)

where 2 is the activity coefficient of a solute, x2 is its
solubility in the solvent (mole fraction) at the system
temperature T, Tm is its melting point, hf is its enthalpy
of fusion, and R is the gas constant. It is then clear that,
at constant temperature, the higher the melting temperature and enthalpy, the lower the solute solubility
x2. Simplified calculations using this equation and
representing the supercritical solvent as a highly expanded liquid reveal that the ideal solubility of theobromine should be about 2 orders of magnitude less than
that of caffeine.
From the data shown in Figures 3 and 4, one can also
observe little effect of moisture, within the ranges
explored in this study, on the extracted quantities of
cocoa butter. The same result was also observed by
Friedrich et al.10 Lentz et al.,33 however, reported large
effects of moisture on the extraction of caffeine in binary
systems. The smaller effects found in this study with
regard to the extraction of methylxanthines could be
attributed to the lack of enough moisture to modify the
nature of the solvent or to modify the possible cosolvency effects of cocoa butter once extracted cocoa
butter was not affected by the solvent moisture content.
Figure 6 shows the extraction curves obtained for
cocoa butter, theobromine, and caffeine from a 30-g

Figure 6. (A) Cocoa butter and (B) theobromine and caffeine

extracted using supercritical ethane at 343.2 K and 15.2, 24.8, and
28.3 MPa.

sample of cocoa beans at 15.2, 24.8, and 28.3 MPa and

a temperature of 343.2 K using a flow rate of 0.86
gmin-1 of supercritical ethane. The behaviors of the
extraction curves with respect to pressure are qualitatively similar to those obtained for the extraction with
CO2. Extractions were only carried out with dry supercritical ethane, once the moisture effect was found to
be small in extractions with CO2. Two independent
extractions were carried out at 343.2 K and 15.2 MPa,
and the extracted amounts were found to be reproducible to within (4%, in agreement with the reproduc-

6756

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Figure 7. Cocoa butter extracted using (A) CO2 and (B) ethane
at 343.2 K.

ibility reported earlier by Mohamed et al.14 with this

same experimental apparatus.
At 343.2 K, using 310 g of ethane, cocoa butter yields
of 9, 62, and 72% were obtained at 15.2, 24.8, and 28.3
MPa, respectively. Caffeine and theobromine yields of
3.73 and 0.45%, 10.56 and 1.68%, and 11.82 and 2.07%,
respectively, were obtained for the same quantity of
solvent at the same three extraction pressures. Although an increase in pressure resulted in an increase
in extraction yield, as expected because of the resultant
increase in solvent density, the proportional increase
is much smaller in the second pressure interval (24.828.3 MPa) because of the smaller density increase
compared to that in the 15.2-24.8 MPa pressure
interval.
In Figure 6, one can also observe the relatively larger
quantity of extracted cocoa butter in comparison with
the amounts of methylxanthines, similarly to the difference observed in Figures 3 and 4 for supercritical
carbon dioxide.The yields of caffeine were also higher
than those of theobromine despite the larger quantity
of theobromine in the bean matrix (see Table 1).
The quantities of cocoa butter extracted by ethane
were much larger than the amounts obtained with
supercritical CO2 at the same temperature and similar
pressure, as shown by the extraction curves represented
on a dimensionless basis for the two solvents (Figure
7). Despite the different pressure conditions used, the
higher solubilities of cocoa butter in ethane are evident
by the steeper inclinations of the linear portions of the
extraction curves (Figure 7).
Ethane and carbon dioxide have molecular weights
of 30 and 44, respectively; critical temperatures of 305.4
K and 304.2 K, respectively; and critical pressures of
4.82 MPa and 7.28 MPa, respectively. The density of
ethane is approximately one-half that of CO2 at the
same pressure and temperature. Consequently, the
larger amounts of CO2 used to obtain a certain amount
of extracted cocoa butter indicate that the number of
CO2 molecules is much higher than the number of
ethane molecules. The difference in the extraction
efficiencies of cocoa butter using supercritical CO2 and
ethane (Figures 3, 4, and 6) cannot, therefore, be
attributed to density effects (or larger numbers of
ethane molecules). The fact is that CO2 is a nonpolar
and lipophilic solvent with polarizability smaller than
that of any hydrocarbon except methane.34 The polarizability of ethane as reported by Mendes et al.35 is

Figure 8. (A) Theobromine and (B) caffeine extracted using

supercritical fluids at different pressures and 343.2 K.

larger than that of CO2, which would suggest stronger

molecular dispersion interactions between ethane and
butter components than between these components and
CO2. This could well be the reason for the higher
solubilities of cocoa butter in ethane than in CO2. This
high efficiency of ethane is demonstrated by the ability
of ethane to extract most of the cocoa butter with a much
lower solvent quantity and therefore at much shorter
extraction times. The higher cost of ethane than CO2
would need to be compensated for by the lower ethane
consumption and lower operating cost due to the lower
operating pressures for ethane to be a more attractive
solvent than CO2.
Figure 8 shows, on a dimensionless basis, the yields
of caffeine and theobromine obtained using both supercritical CO2 and ethane at different pressures and at a
temperature of 343.2 K. From the slopes of the extraction curves, we can see that the solubilities of caffeine
and theobromine in carbon dioxide are comparable to
those in ethane. Extractions with ethane were obtained,
however, at lower pressures. As the solubility increases
with pressure at constant temperature because of the
increase in solvent density, it is not unreasonable to
expect the solubilities of these two alkaloids to be
greater in ethane than in CO2 at the higher extraction
pressures used when carbon dioxide was the solvent. It
is important to note, however, as kindly pointed out by
one of the reviewers of this work, that the data
presented for ethane extractions, which were carried out
for shorter times because of the more rapid extraction
of cocoa, do not provide any assurance that the ultimate
extraction yields of caffeine and theobromine with
ethane at high solvent/cocoa bean ratios will be equal
to or higher than those obtained with carbon dioxide.

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6757

Martin et al.30 reported some data on the solubilities

of caffeine and theobromine in hexane and concluded
that theobromine was more soluble than caffeine. One
might conjecture, then, that this same trend would also
be followed with ethane as the solvent. Although our
data would also suggest this same trend, the limited
extent of these data does not allow a definitive conclusion to be reached. No solubility data for caffeine or
theobromine in ethane, which would certainly allow a
more decisive conclusion, could be found in the literature.
IR spectroscopic studies reported by Kazarian et al.36
also confirm the presence of specific interactions between CO2 and carbonyl groups, which supports the
suggestion raised by one of the reviewers of this work
that Lewis acid/Lewis base interactions might be responsible for the high solubility of caffeine in carbon
dioxide.
Conclusions
Using supercritical ethane, it was possible to extract
successfully almost all of the cocoa butter contained in
a cocoa bean sample in a more efficient way than is
possible using CO2. This is due to the greater cocoa
butter solubility in ethane, which can be attributed to
the strong dispersion interactions caused by the larger
polarizability of ethane.
For either solvent, supercritical ethane or CO2, the
results revealed a large influence of pressure on the
extraction yields of cocoa butter and methylxanthines
from cocoa beans resulting from the relatively large
changes in solvent isothermal compressibility with
pressure at the conditions investigated in this study.
Relatively shorter extraction times were observed for
the removal of caffeine than theobromine, probably
because of the formation of theobromine complexes
through strong hydrogen bonding as demonstrated by
the higher melting temperature and enthalpy of theobromine in comparison with those of caffeine.
Supercritical ethane was found to be more selective
for cocoa butter than for methyxanthines, making the
first extracted fractions rich in cocoa butter and allowing
the possible production of methylxanthine-free cocoa
butter fractions.
Acknowledgment
The authors express their appreciation for the financial support received from the Sao Paulo State Research
Funding Agency (FAPESP) and the Brazilian National
Research Agency (CNPq) and are also grateful to Cardill
S.A. for donating the cocoa beans.
Literature Cited
(1) Timbie, D. J. Studies on the Proteins and Purine Alkaloids
of Cocoa Beans. Ph.D. Dissertation, Pennsylvania State University, University Park, PA, 1977.
(2) Paiva, M.; Janick, J. Variability of alkaloid production in
Theobroma cacao L. [Theobromine]. Rev. Theobroma 1983, 13, 249.
(3) Sotelo, A.; Alvarez, R. G. J. Chemical composition of wild
Theobroma species and their comparison to the cacao bean. J.
Agric. Food Chem. 1991, 39, 1940.
(4) Sun, M. FDA caffeine decision too early, some day. Science
1980, 209, 1500.

(5) James, J. E. Caffeine and Health; Academic Press: San

Diego, CA, 1991.
(6) Baltassa, F.; Carbour, N.; Ferry, S. E
tude du contenu
purique de drogues a cafeine: I.-Le Mate: Ilex paraguaryensis
Lamb. Plantes me d. phytothe r. 1984, 18, 195.
(7) Saldana, M. D. A. Extraction of caffeine and trigonelline
and chlorogenic acids from coffee beans with supercritical. M.Sc.
Thesis, State University of Campinas, Campinas, SP, Brazil, 1997.
(8) Spoladore, D. S., Teixeira, J. P. F., Morais, R. M.; Zullo, M.
A. T. Chemical Composition of Fermented Cocoa Beans. Bragantia
1983, 42 (4), 249-253.
(9) Hulbert, G. J.; Biswal, R. N.; Walker, T. H.; Mehr, C. B.;
Collins, J. L. Solid/liquid extraction of caffeine from guarana with
methylene chloride. Food Sci. Technol. Int. 1998, 4, 53.
(10) Friedrich, J. P.; Snyder, J. M.; Christianson, D. D. Supercritical CO2 extraction of lipid-bearing materials and characterization of the products. J. Am. Oil Chem. Soc. 1984, 61, 1851.
(11) Saldana, M. D. A.; Mazzafera, P.; Mohamed, R. S. Extraction of purine alkaloids from mate (Ilex paraguariensis) using
supercritical CO2. J. Agric. Food Chem. 1999, 47, 3804.
(12) Brunner, G. Mass transfer from solid material in gas
extraction. Ber. Bunsen-Ges. Phys. Chem. 1984, 88, 887.
(13) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction Principles and Practice; Butterworth Publishers: Boston,
1994.
(14) Mohamed, R. S.; Saldana, M. D. A.; Socantaype, F. H.;
Kieckbusch, T. G. Reduction of Cholesterol Content of Butter Oil
Using Supercritical Ethane Extraction and Adsorption on Alumina. J. Supercrit. Fluids 2000, 16, 225.
(15) Li, S.; Hartland, S. Influence of cosolvents on solubility
and selectivity in extraction of xanthines and cocoa butter from
cocoa beans with supercritical carbon CO2. J. Supercrit. Fluids
1992, 5, 7.
(16) Saldana, M. D. A.; Mazzafera, P.; Mohamed, R. S. Extraction of Cocoa Butter from Cocoa Beans Using Supercritical CO2
and Ethane. Fluid Phase Equilib. 2002, 194-197, 885.
(17) Reverchon, E.; Daghero, J.; Marrone, C.; Mattea, M.;
Poletto, M. Supercritical Fractionational Extraction of Fennel Seed
Oil and Essential Oil: Experiment and Mathematical Modeling.
Ind. Eng. Chem. Res. 1999, 38, 3069.
(18) Stahl, E.; Schutz, E.; Mangold, H. K. Extraction of Seed
Oils with Liquid and Supercritical Carbon Dioxide. J. Agric. Food
Chem. 1980, 28, 1153.
(19) Friedrich, J. P.; List G. R.; Heakin, A. J. Petroleum-Free
Extraction of Oil from Soybean with Supercritical CO2. J. Am. Oil
Chem. Soc. 1982, 59, 282.
(20) Friedrich, J. P.; Pryde, E. H. Supercritical CO2 Extraction
of Lipid-Bearing Materials and Characterization of Products. J.
Am. Oil Chem. Soc. 1984, 61, 223.
(21) Brennecke, J. F. New Applications of Supercritical Fluids.
Chem. Ind. 1996, 831.
(22) Schulz, K.; Mansoori, G. A. Supercritical Fluid Extraction/
Retrograde Condensation (SFE/RC) with applications in Biotechnology. In Supercritical Fluid Technology; Bruno, T. J.; Ely, J. F.,
Eds.; CRC Press: Boca Raton, FL, 1991; pp 451-478.
(23) Mansoori, G. A.; Schulz; K.; Martinelli, E. Bioseparation
using supercritical fluid extraction/retrograde condensation. Bio/
Technology 1988, 6, 393-396.
(24) Hartono, R.; Mansoori, G. A.; Suwono, A. Prediction of
solubility of biomolecules in supercritical solvents. Chem. Eng. Sci.
2001, 56, 6949-6958.
(25) Sambarato, D. Esperienze di estrazione della teobromina
da cacao con anidride carbonica supercritica, Ph.D. Dissertation,
University of Milan, Milan, Italy, 1984.
(26) Zosel, K. Decaffeinating Coffee. French Patent 2.079.261,
1971.
(27) Hubert, P.; Vitzthum, O. G. Fluid extraction of hops, spices,
and tobacco with supercritical gases. Angew. Chem., Int. Ed. Engl.
1978, 17, 710.
(28) Saldana, M. D. A.; Zetzl, C.; Mohamed, R. S.; Brunner, G.
Decaffeination of Guarana Seeds in a Microextraction Column
Using Water-Saturated CO2. J. Supercrit. Fluids 2002, 22, 119.
(29) Bamberger, T.; Erickson, J. C.; Cooney, C. L.; Kumar, S.
K. Measurement and Model Prediction of Solubilities of Pure Fatty
Acids, Pure Triglycerides, and Mixtures of Triglycerides in Supercritical Carbon Dioxide. J. Chem. Eng. Data 1988, 33, 327.
(30) Martin, A.; Paruta, A. N.; Adjei, A. Extended Hildebrand
Solubility Approach: Methyxanthines in Mixed Solvents. J.
Pharm. Sci. 1981, 70, 1115.

6758

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

(31) Ford, K. A.; Ebisuzaki, Y.; Boyle, P. D. Methylxanthines

II: Anhydrous Theobromine. Acta Crystallogr. 1998, C54, 1980.
(32) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties
of Gases and Liquids, 4th ed.; McGraw-Hill: New York, 1988.
(33) Lentz, H.; Gehrig, M.; Schulmeyer, J. Dynamic Solubility
Measurements of Caffeine in Carbon Dioxide and in Carbon
Dioxide Saturated with Water. Physica B 1986, 139 & 140, 70.
(34) Dobbs, J. M.; Johnston, K. P. Selectivities in pure and
mixed supercritical fluid solvents. Ind. Eng. Chem. Res. 1987, 26,
1476.

(35) Mendes, R. L.; Nobre, B. P.; Coelho, J. P.; Palavra, A. P.

Solubility of -carotene in Supercritical Carbon Dioxide and
Ethane. J. Supercrit. Fluids 1999, 16, 99.
(36) Kazarian, S. G.; Vincent, M. F., Bright, F. V.; Liotta, C.
L.; Eckert, C. A. J. Am. Chem. Soc. 1996, 118, 1729-1736.

Received for review May 28, 2002

Revised manuscript received October 9, 2002
Accepted October 17, 2002
IE0203936