See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/240696197

Characterisation and Process Development of Supercritical Fluid Extraction of

Soybean Oil

Article  in  Food Science and Technology International · December 2002

DOI: 10.1177/108201302129174900

CITATIONS READS
31 419

3 authors, including:

Enrique J. Martinez de la Ossa

Universidad de Cádiz
446 PUBLICATIONS   3,406 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

SILICA PARTICLES IMPREGNATION WITH NANO CAPSULES OF NATURAL ANTIOXIDANTS USING SUPERCRITICAL TECHNOLOG View project

Energy production from wet biomass wastes by hydrothermal processes View project

All content following this page was uploaded by Enrique J. Martinez de la Ossa on 17 March 2015.

The user has requested enhancement of the downloaded file.

 Characterisation and Process Development of Supercritical Fluid
Extraction of Soybean Oil
M. Dobarganes Nodar, A. Molero Gómez* and E. Martı́nez de la Ossa
Department of Chemical Engineering, Food Technology and Environmental Technology, University of Cádiz,
Apdo. 40, E-11510 Puerto Real, Cádiz, Spain

This work describes and analyses the extraction of soybean oil using supercritical carbon dioxide as a
solvent from the point of view of both operative method and pre-treatment of the raw material. The best
conditions for soybean oil extraction were obtained at a pressure of 300 bar, a temperature of 40
C and a
solvent flow rate of 1.8 L/min at STP. The yields obtained using the supercritical fluid under the best
operating conditions were similar to those obtained by conventional extraction methods using hexane as
the solvent (19.9 %w/w.); furthermore, the quality of oil extracted by the supercritical fluid was higher
(acidity, 0.8). The fatty acid composition of the soybean seed oil extracted by supercritical fluid was par-
ticularly rich in unsaturated fatty acids, particularly linoleic acid (51.8%). For these reasons, the soybean
seed oil extracted using supercritical carbon dioxide could compete with that obtained by the conventional
process, since the oil refinement stages are simplified significantly and the solvent distillation stage is
completely removed, which are the two most costly processing stages in terms of energy consumption.

Key Words: supercritical fluids, extraction, soybean oil, linoleic acid

INTRODUCTION natural products, which require mild treatments to

preserve the nutritive and sensory characteristics.
Supercritical fluid extraction (SFE), particularly Traditionally, soybean oil has been obtained by a
involving the use of carbon dioxide, has been applied conventional extraction process using hexane as solvent.
in numerous different fields. The decaffeination of Nevertheless, after the extraction, a further stage is
coffee, production of hop flavours for use in beer required to remove the solvent to levels below the
making (Sthal et al., 1988; Bork and Körner, 1991), minimum concentrations allowed by the relevant regula-
extraction of aromas and flavours (Bunzenberger et al., tions that depend on the final destination of the product
1984; Raasch and Knorr, 1990), extraction of spices (Fernández, 1992; Domı́nguez et al., 1995). The applica-
(Quirin and Gerard, 1990) and extraction and refining tion of SFE to obtain vegetable oils leads to a high
of edible oils (Sthal et al., 1980; Eggers, 1996) are some quality product that can be used directly, thus preclud-
of the many industrial applications. ing the removal and recovery of solvent by distillation.
This separation technique offers extraction yields that It has been found in previous studies that the oil
are very similar to those obtained by conventional obtained from soybean with supercritical carbon dioxide
extraction processes using liquid solvents, but it requires (SC–CO2) has some inherent advantages, including low
a pre-determined combination of operating temperature phosphorus and iron contents due to the low solubility
and pressure. The advantages of this technique, of phospholipids in SC–CO2 (Cocero and Calvo, 1995).
compared to organic solvents used in the conventional In addition, the colour of the oil obtained is lighter due
extraction methods, are that carbon dioxide is non- to the low solubility in SC–CO2 of carotenes and other
toxic, non-flammable, non-corrosive and is also cheap pigments (Eshtiaghi et al., 1991).
and readily available in bulk quantities with a high The composition of soybeans is around 40% protein,
degree of purity (Molero et al., 1996). Carbon dioxide 20% lipids, 17% cellulose and hemicellulose, 7% sugar,
also has a relatively low critical pressure and tempera- 5% crude fibre and 6% ash. The beans have a high oil
ture (73.8 bar and 31.1
C, respectively) and, as such, it content, at approximately 35–40% by weight, and this
can be considered as the perfect solvent to extract oil has a unique composition in comparison with most
of the other common vegetable oils as well as having a
higher content of linoleic acid. The fatty acid composi-
*To whom correspondence should be sent tion of soybean oil is somewhat different but, in other
(e-mail: [email protected]). respects it is similar to other vegetable oils. It is a
Received 21 March 2002; revised 10 October 2002. polyunsaturated oil and contains around 54% of linoleic
acid (Fernández, 1992; Domı́nguez et al., 1995).
Food Sci Tech Int 2002;8(6):0337–6 The work described in this paper involved the

2002 Sage Publications
ISSN: 1082-0132
extraction of soybean oil using supercritical carbon
DOI: 10.1106/108201302031651 dioxide as solvent. Working conditions for the process
337
338 M. DOBARGANES NODAR ET AL.

(operating pressure, temperature and solvent flow opening valves V-1 and V-2. When the desired pressure
rate) and the pre-treatment of the raw material (grain was attained, the heating jacket was switched on to give
size) were studied, and the fatty acid composition of the the required operating temperature. When both the
oil obtained under the best extraction conditions were desired pressure and temperature had been reached, the
determined. extraction was started by opening valve V-3. The flow
rate of carbon dioxide through the extractor vessel was
regulated by the flow micrometer valve. The oil
dissolved in the supercritical CO2 was separated from
MATERIAL AND METHODS the carbon dioxide and collected in the separator
(100 mL capacity, 316 SS) at ambient temperature and
The soybean seeds (William’s variety) used for the pressure. The CO2 was passed through an in-line
processes were supplied by the Instituto de la Grasa volumetric flow meter, which controlled the quantity
(Sevilla, Spain). The soybean seeds were crushed in a and flow rate of CO2 used. The flow meter used was
coffee mill of 2 kg capacity (Futurmatmodel FP, a model FC-70 supplied by EG&G Flow Technology
Barcelona, Spain) prior to extraction. Dehydration of (Phoenix, AZ, USA).
the seeds was carried out by heating to 65
C until In order to compare the SFE method studied here
constant weight was achieved. The equipment used for with the conventional extraction process using liquid
the SFE processes with carbon dioxide was the SCE solvents, the use of a ‘‘Soxhlet’’-type apparatus with
screening system model manufactured by Autoclave hexane as solvent (250 mL of hexane and 30 g of
Engineers (Erie, PA, USA). A schematic diagram of this soybean seed) was also studied. The extraction time
equipment is shown in Figure 1. was 16 h, which guaranteed full depletion of the seed
Liquid CO2 was cooled to prevent its gasification and and allowed the maximum possible extractive yield to be
introduced into an HPLC pump of 46–460 mL/h determined. These values are considered very important
capacity and fitted with a cooled head. The extraction to establish a sound basis for direct comparison to the
vessel was a 316 SS high pressure cylinder (40 mL SFE process.
volume) that was capable of operating up to 400 bar and The amount of oil obtained by the two extraction
340
C. The vessel was fitted with an electrical heating techniques was determined gravimetrically. In all cases,
jacket. A sample (about 20 g) was placed in the extractor the yield of the extraction has been expressed as the
vessel. The pressure of the CO2 cylinder was maintained grams of oil extracted per grams of soybean seeds.
throughout the system by opening valves V-1 and V-2 The extraction process was carried out within the
and closing the valve V-3. Before the pump was switched following ranges of operating conditions: pressure, from
on, the valves V-1 and V-2 were closed. The pump was 100 to 380 bar; temperature, 40 and 60
C; solvent flow
then switched on and its output pressure increased using rate, 0.8 and 1.8 L/min at STP.
the pressure regulating valve until the required proces- The physico-chemical characterisation of the extrac-
sing pressure was reached. The compressed CO2 was ted oils was performed according to the indexes for the
subsequently introduced into the extractor vessel by analysis of oils for human consumption (AOCS, 1983).
The indexes determined were acidity (% of oleic acid),
iodine (Hanus method), saponification and peroxides.
The fatty acid composition of the oil was determined,
after methylation, by gas chromatography using a model
5890 Hewlett Packard chromatograph (Pittsburgh, PA,
USA) fitted with a Carbowax 20M (Supelco, Bellefonte,
PA, USA) capillary column and an FID detector.
All the products and chemical reagents used were
of analytical quality. The carbon dioxide used was of
99.95% purity from Carburos Metálicos, S. A. (Sevilla,
Spain).

Figure 1. Schematic diagram of supercritical CO2 RESULTS AND DISCUSSION

extraction apparatus: B, liquid CO2 cylinder; J; cooler,
C, check valve; P, pump; PRV, pressure regulating Extraction of Soybean Oil with Supercritical
valve; E, extraction vessel; S, separator; EHJ, elec- Carbon Dioxide
trical heating jacket; RD, rupture disk; FMV, flow
micrometric valve; T/C-1,2, thermocouples; V-1 to 5, The yield of soybean oil obtained with supercritical
shut-off valves; PG-1 to 3, pressure gauges and FM, CO2 increased with extraction pressure for both
flowmeter. temperatures (Figure 2). Furthermore, at pressures
 Supercritical Fluid Extraction of Soybean Oil 339

Figure 2. Effect of pressure on extraction yield of

Figure 3. Effect of supercritical carbon dioxide flow
soybean oil using supercritical carbon dioxide at two
rate on extraction yield of soybean oil (pressure,
different temperatures (extraction time 8 h; STP
200 bar; temperature, 40
C; soybean seed size,
solvent flow rate, 1.8 L/min; soybean seed size,
0.3 mm). (œ) 0.8 L/min, (4) 1.8 L/min.
0.3 mm). (œ) 40
C, (4) 60
C.

above 250 bar, high increments in pressure produce only

a small increment in the yield.
The differences in yield for the extraction of different
oils can be explained in terms of the nature of the seeds
(Sthal et al., 1980; Molero et al., 1996). For example,
an appropriate combination of pressure and tempera-
ture is needed to obtain the best solvent capacity for
the CO2 (which is closely related with the density) in
the extraction, in which the solvent must be able to
penetrate and overcome the mass transfer resistance to
extract the oil from the seed.
On the other hand, temperature was found to have a Figure 4. Effect of extraction time on extraction yield
relevant influence on the yield of the extraction process. of soybean oil by supercritical carbon dioxide
At 60
C, a pressure of 380 bar was needed to reach a (pressure, 200 bar; temperature, 40
C; soybean seed
similar yield to that obtained at 250 bar and 40
C. size, 0.3 mm). (œ) 0.8 L/min, (4) 1.8 L/min.
The amount of soybean oil extracted per unit of
carbon dioxide depended on the flow rate of solvent
(Figure 3). The flow rate used in the extraction process is
important because the residence time in the extractor
increases as the flow rate decreases. The lower the flow
rates the better the solubility equilibrium is maintained
in the later stages of the extraction (Friedrich and List,
1982).
Higher flow rates lead to a larger quantity of oil, but
this is also associated with a higher solvent consumption.
Lower flow rates require smaller amounts of solvent, but
the process time is longer to obtain the same extraction
yield. The maximum extraction yield was obtained at
solvent flow rates of around 1.8 L/min at STP. Figure 5. Effect of particle size on extraction yield of
When these data are plotted against the extraction soybean oil (pressure, 380 bar; temperature, 40
C;
time (Figure 4) the lower solvent flow rate needed more STP solvent flow rate, 1.8 L/min). (œ) 0.30 mm; (4)
extraction time to achieve a satisfactory extraction yield.

0.75 mm; ( ) 1.00 mm.
For the solvent flow rate of 1.8 L/min, the extraction
time was 5 h. The maximum yield obtained for the particles with
The quantity of oil extracted increased as the particle the biggest size (1 mm) was 9.73%, which represented
size decreased (Figure 5). It is reasonable to suppose almost 6% less than the yields obtained for particles of
that the oil is localised in the interior of the seed and, in 0.3 mm. This reduction in yield is a result of the change
this case, the internal mass transfer resistance is lower in the internal mass transfer resistance. From the results
when the seed is crushed. obtained for the three particle sizes studied, it seems
340 M. DOBARGANES NODAR ET AL.

clear that the resistance to internal mass transfer will on this subject are available in the literature. For this
decrease as particle size decreases. For particle sizes reason, the comparison of some relevant parameters for
larger than 0.3 mm, a higher extraction time is needed to other vegetable oils extracted with supercritical carbon
reach the same yields as those obtained with smaller dioxide (Table 2) in order to establish a better
particle sizes. comparison of the quality of the soybean oil obtained
A similar effect was observed in previous studies, in this study is interesting.
which confirmed that the seeds have to be crushed to The free fatty acid and peroxide indexes are the most
ensure a satisfactory extraction yield (Molero et al., important parameters to evaluate the degree of oil
1996). However, wheat germ oil does not need pre- degradation. The first of these values provides informa-
treatment prior its extraction (Taniguchi et al., 1986; tion about the hydrolytic deterioration that may have
Molero and Martı́nez, 2000). occurred in the oil; and, the peroxide index approaches
On the basis of the above discussion, the best yields the extent to which certain components of the oil have
for the extraction of soybean oil by supercritical carbon suffered oxidation (tocopherols and polyphenols). As
dioxide are obtained at a pressure of 300 bar, a one would expect, low values for these indexes indicate
temperature of 40
C, a solvent flow rate of 1.8 L/min that oil degradation is minimal.
(at STP) and a milled seed size of 0.3 mm. Under these The values obtained for the soybean oil extracted by
conditions the extraction time was 5 h. CO2 supercritical was very low in comparison with
the other vegetables oils shown in Table 2. Only the
Comparison of SFE Versus Other References acidity of grape seed oil is close to the value in this work
for the soybean oil. Furthermore, the value for the free
Extraction Yields fatty acids is very close to that permitted for human
consumption. This particular aspect is very important
The data in Table 1 provide a comparison of the SFE because the subsequent neutralisation stage in the
and conventional hexane extraction yields from soybean refining process must be applied on a small scale.
seeds in this study with the results obtained by other The higher values obtained for the saponifiable
authors. Those data show that pressures greater than matter and the low values for the free fatty acids,
200 bar are required in order to obtain a good yield in indicated that that the soybean oil extracted by super-
the process. The yields obtained were, in all cases, lower critical CO2 had a high triglyceride content. This high
than those obtained using the conventional solvent content is only exceeded by the grape seed oil. In
extraction technique with hexane as a solvent. addition, the iodine index also has a high value due to
The fact that the SC–CO2 has a high selectivity is well the higher percentage of unsaturated fatty acids present
known (Taniguchi et al., 1986; Zhao et al., 1987; Molero in the oil (Table 2).
and Martı́nez, 2000) and it is of interest to analyse the
differences between the oils extracted using the two
methods. Only by establishing the differences between
the extracted compounds can the use of a new method of
extraction be justified. The soybean oil extracted by SFE Table 2. Physico-chemical characterisation of
using carbon dioxide is solvent free. For this reason, the vegetables oils extracted using supercritical
distillation process usually required to recover the carbon dioxide.
solvent is not needed.
Peroxide
Iodine Free Fatty Value
Oil Composition Saponification Value Acids (meq
Oil Value (Hanus) (% Oleic) O2/kg oil)
Comparison of the results obtained in this work is not Soybean (this work) 194 127 0.8 7.2
straightforward due to the fact that only a few references (300 bar, 40
C)
Grape seed 259 98 3.4 383.0
(Molero et al., 1996)
(350 bar, 40
C)
Table 1. Comparison of soybean oil extraction yields
Soybean (Friedrich 0.5 <0.1
using supercritical carbon dioxide and hexane as and List, 1982)
solvents. (350 bar, 50
C)
Wheat germ
Yield (wt.%) (Molero and 185 102 14.0 75.0
Martı́nez, 2000)
SFE Using Extraction
(150 bar, 40
C)
Soybean CO2 Using Hexane
Borage seed (Molero 163 198 11.0 33.5
This work (300 bar, 40
C) 19.5 25.0 and Martı́nez, 2001)
Sthal et al., 1980 (300 bar, 40
C) 16.4 19.9 (300 bar, 40
C)
Friedrich and List, 1982 (350 bar, 50
C) 19.9 20.0 Soybean (FAO, 2001) 189–195 124–139 0.6 <10.0
 Supercritical Fluid Extraction of Soybean Oil 341

Table 3. Percentage of fatty acids of soybean oil extracted by supercritical carbon dioxide.
C16:0 C16:1 C18:0 C18:1 C18:2 C18:3
Soybean (Palmitic) (Palmitoleic) (Stearic) (Oleic) (Linoleic) (Linolenic)
This work 11.8 0.2 4.2 23.7 51.8 6.5
(300 bar, 40
C)
Friedrich and List, 1982 11.0 3.5 27.5 52.0 6.0
(350 bar, 50
C)
Soybean (FAO, 2001) 5.0–7.6 <0.3 2.7–6.5 14.0–39.4 48.3–74.0 <0.3

In summary, the quality of soybean oil obtained using Bunzenberger G., Lack E. and Marr R. (1984). CO2-
supercritical CO2 is similar to that of oil refined and extraction: comparison of super and subcritical extrac-
extracted with an organic solvent, consequently the tion. German Chemistry and Engineering 7: 25–31.
refining process may be used on a small scale. Cocero M.J. and Calvo L. (1995). Planta piloto-laboratorio
de extracción con CO2 supercrı́tico para la obtención de
aceites de semillas. Alimentación, Equipos y Tecnologı´a
Fatty Acid Composition 14(3): 67–75.
Domı́nguez H., Sineiro J., Núñez M.J. and Lema J.M.
Fatty acid composition of the soybean oil obtained (1995). Incorporación de un tratamiento enzimático para
using supercritical CO2 as solvent were compared with mejorar la extractabilidad del aceite de semillas en
those obtained by Friedrich and List and the values distintos procesos. Alimentación, Equipos y Tecnologı´a
regulated for human consumption (Table 3). The 14(3): 59–66.
content in fatty acids is very similar in both studies. Eggers R. (1996). Supercritical fluid extraction of oilseeds/
When these values are compared with those given by lipids in natural products. In: King J.W. and List G.R.
FAO, it can be seen that they are within the established (eds), Supercritical Fluid Technology in Oil and Lipid
limits. This aspect is of vital importance in terms of the Chemistry. Illinois: AOCS Press.
subsequent refining process. Eshtiaghi M., Messerschmidt K. and Knorr D. (1991).
The high content in unsaturated fatty acids (82%) and Supercritical fluid extraction of pigments from peels of
the lower proportion of saturated fatty acids (12%) orange fruits. ZFL, -Internationale Zeitschrift fuer
justifies the above discussion regarding the values of the Lebensmittel Technik, Marketing, Verpackung und
physico-chemical indexes and makes it particularly Analytik 42(11): 61–64.
appropriate to consider using such oil in the alimentary FAO (2001). The Codex Alimentarius. Vol. 8. Rome: FAO.
industry. Fernández P.M. (1992). Aceites y grasas vegetales comes-
Therefore, despite the high costs of the process plant, tibles. Definiciones, disposiciones legales y aspectos
SFE could prove to be more economical than nutricionales. control de calidad y pruebas de pureza
(i). Alimentación, Equipos y Tecnologı´a 11(3): 47–55.
conventional organic extraction. However, further
studies regarding costs are needed to confirm these Friedrich J.P. and List G.R. (1982). Characterization of
soybean oil extracted by supercritical carbon dioxide and
latter conclusions.
hexane. Journal of Agricultural and Food Chemistry
30: 192–193.
Molero Gómez A., Pereyra López C. and Martı́nez de la
ACKNOWLEDGEMENTS Ossa E. (1996). Recovery of grape seed oil by liquid and
supercritical carbon dioxide: a comparison with conven-
The authors wish to express their gratitude to the tional solvent extraction. The Chemical Engineering
Instituto de la Grasa (Sevilla, Spain) for valuable Journal 61: 227–231.
discussions, help in obtaining technical data, and for Molero Gómez A. and Martı́nez de la Ossa E. (2000).
providing the material required for the experiments. Quality of wheat germ oil extracted by liquid and
supercritical carbon dioxide. Journal of the American
Oil Chemists’ Society 77(9): 969–974.
REFERENCES Molero Gómez A. and Martı́nez de la Ossa E. (2001).
Quality of borage seed oil extracted by liquid and
AOCS (1983). Official and Tentative Methods of the supercritical carbon dioxide. The Chemical Engineering
American Oil Chemists Society. 3rd Edn. Illinois: AOCS. Journal 86: 1–7.
Bork M. and Körner J.P. (1991). Extraction by use of Quirin K.W. and Gerard D. (1990). High pressure
supercritical solvents. ZFL, -Internationale Zeitschrift extraction of spices. ZFL, -Internationale Zeitschrift
fuer Lebensmittel Technik, Marketing, Verpackung und fuer Lebensmittel Technik, Marketing, Verpackung und
Analytik 42(12): 732–738. Analytik 41(11): 731–734.
342 M. DOBARGANES NODAR ET AL.

Raasch A. and Knorr D. (1990). Supercritical-fluid-extrac- Taniguchi M., Tsuji T., Shibata M. and Kobayashi T.
tion of flavours compounds from peels of passion fruits. (1986). Extraction of oils from wheat germ with
ZFL, -Internationale Zeitschrift fuer Lebensmittel supercritical carbon dioxide. Agricultural and Biological
Technik, Marketing, Verpackung und Analytik 41(3): 8–12. Chemistry 49(8): 2367–2372.
Stahl E., Schütz E. and Mangold H.K. (1980). Extraction of Zhao W., Shisshikura A., Fujimoto K., Arai K. and Saito
seed oils with liquid and supercritical carbon dioxide. S. (1987). Fractional extraction of rice bran oil with
Journal of Agricultural and Food Chemistry 28: 1153–1157. supercritical carbon dioxide. Agricultural and Biological
Stahl E., Quirin K.W. and Gerard D. (1988). Dense Gases for Chemistry 51(7): 1773–1777.
Extraction and Refining. New York: Springer Verlag.

View publication stats