Available online at www.sciencedirect.com

J. of Supercritical Fluids 44 (2008) 186–192

Extraction of green coffee oil using supercritical carbon dioxide A.B.A. de Azevedo a , T.G. Kieckbush a , A.K. Tashima a , R.S. Mohamed a , P. Mazzafera b,∗ , S.A.B. Vieira de Melo c a

Faculdade de Engenharia Qu´ımica, Universidade Estadual de Campinas, CP 6066, 13083-970 Campinas, SP, Brazil b Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, 13083-970 Campinas, SP, Brazil c Departamento de Engenharia Qu´ımica, Universidade Federal da Bahia, 40210-630 Salvador, BA, Brazil Received 1 March 2007; received in revised form 2 November 2007; accepted 9 November 2007

Abstract Supercritical CO2 extraction was used to refine coffee oil obtained by mechanical pressing. Extractions were carried out using 50–70 ◦ C and pressures ranging from 15.2 to 35.2 MPa, with a CO2 flow rate of 1 standard L/min using a semi-continuous high-pressure extraction apparatus. Green coffee oil fractions were collected at fixed time intervals and the composition of each fraction was determined by HPLC analyses. Caffeine and traces of chlorogenic acid were detected in the first fractions while waxes remained in the extraction vessel. Compared with the original oil the triglyceride composition remained almost unchanged in the fractions. The results also indicate an increase of triglyceride and caffeine extraction with pressure. An increase in extraction temperature results in a retrograde behavior over the pressure range of 15.2–28.3 MPa. At pressures higher than 30 MPa the solubility behavior of coffee oil was apparently independent of the temperature. A good correlation of the solubility data of green coffee oil was obtained using the Chrastil equation. © 2007 Elsevier B.V. All rights reserved. Keywords: Caffeine; Chlorogenic acid; Coffee; Coffee oil; Triglyceride

1. Introduction Vegetable oils are extensively used in industrial applications (especially by the cosmetic, pharmaceutical and food industries) due to their important role in product formulations. Lipids can act as emollients, emulsifiers, carriers, viscosity modifiers, spreading agents, binders and lubricants in many cosmetic products [1]. The specific applications depend on the oil characteristics, which may also vary with the seed used. Coffee beans have a lipid content that ranges from 10% to 13% on dry basis composed of: triacylglycerols (75%), terpene esters (14%), partial acylglycerols (5%), free fatty acids (1%), free sterols (1.5%), sterol esters (1%) and polar lipids (<1%) [2]. Folstar [3] reported a green coffee oil wax content of 0.2–0.3%. The major fatty acids constituents are linoleic (43.1%), linolenic (1.8%), oleic (9.6%), stearic (9.6%) and palmitic (31.1%). Beveridge et al. [4] reported some therapeutical properties of linoleic and linolenic fatty acids for the relief of chronic eczema and the cure of dermatitis. Kahweol and cafestol are the two main terpene esters

∗

Corresponding author. Tel.: +55 19 35216213. E-mail address: [email protected] (P. Mazzafera).

0896-8446/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.supflu.2007.11.004

present in coffee lipids associated to hypercholesterolaemia, anticancer effects and applications in sunscreen [5–8]. Tocopherols (␣, ␤ and ␥) are also present in coffee lipids, and together with tocotrienols, as reported by Gonz´alez et al. [9], as one of the eight constituents of vitamin E. Green coffee oil (GCO) is a mixture of several classes of compounds with triglycerides (TAGs) representing 75% in mass basis. Raw GCO has a relatively high price in the market and is currently obtained by mechanical cold pressing. The extracted oil presents in its composition caffeine and wax. Fractionation of GCO not only improve the oil quality for use in cosmetic and pharmaceutical applications but it can also be a source of other valuable products such as caffeine, sterols, terpenes and tocopherols. Raw GCO obtained by cold press has a dark brown-green color with a cloudy aspect, what is attributed to the presence of pigments and other components (e.g. chlorophyll, terpenes, sterols, wax, etc.). Supercritical CO2 extraction is among the emerging clean technologies for the processing of food, cosmetic and pharmaceutical products. CO2 extraction also presents other well-known advantages over the conventional processes. While conventional processes use organic solvents (which are toxic and difficult to handle) and severe process conditions (e.g. steam

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

187

Fig. 1. Experimental apparatus process diagram.

distillation which can result in product thermal degradation), CO2 , which is non-flammable and non-toxic compound, allows the extraction operation to be carried out at moderate conditions due its reasonable low-critical constants values (Tc = 31.8 ◦ C and Pc = 7.3 MPa). However, to concept, design and optimize supercritical fluid extraction processes of active principles from natural products; it is necessary to know the kinetic and thermodynamic behavior of the system. Therefore, the objective of this work was to obtain new experimental data on the capacity of supercritical CO2 in the extraction and fractionation of GCO, to assess the effects of thermodynamic variables (pressure and temperature) on these operations through the critical evaluation of the experimentally collected data and to correlate the data for coffee oil using the correlation of Chrastil [10].

2. Experimentation 2.1. Materials GCO obtained by mechanical pressing of Coffea arabica beans was donated by Chemyunion Qu´ımica LTDA (Campinas, Brazil). This raw GCO initially presented 2.7 g caffeine/kg oil and a dark brown-green color with a cloudy aspect. The oil was divided in 20 g portions, placed in bags and stored in a freezer (−20 ◦ C) before use. Pure carbon dioxide (99.9% in purity) was obtained from White Martins Inc. (Campinas, Brazil). Methanol, acetone, ethanol and acetonitrile, HPLC grade (Merck), were purchased from a local supplier. TAG standards were purchased from Sigma (St. Louis, USA).

2.2. Experimental apparatus A semi-continuous flow experimental apparatus (Fig. 1) with independent control of temperature and pressure was used in the extraction experiments. The apparatus was projected and assembled by the supercritical fluid process research group of the School of Chemical Engineering, State University of Campinas, Brazil. The apparatus was designed for pressures up to 41.3 MPa at 200 ◦ C. The major components of this apparatus are positive liquid displacement pumps (P-1 and P-2, Thermal Separation Products, Riveira Beach, FL, USA) for solvent delivery (46–460 mL/h), one 300 mL high-pressure extraction vessel (E-1, Autoclave Engineers, Erie, PA), two high-pressure columns (C-1 and C-2, 300 mm × 12.7 mm i.d.) and collection flasks. The extraction vessel and the columns were supplied with heating jackets and temperature controllers and were operated in series. Heating tapes were used throughout the apparatus to maintain constant temperature over the entire apparatus. In order to ensure constant and steady solvent delivery, the pump heads were cooled by a circulating fluid passing through a chiller, CFT R134a (NESLAB Instruments, Newington, NH, USA). Flow rates and accumulated gas volumes passing through the apparatus were measured using a digital flow meter device (FM-1, EG&G Instruments Flow Technology, Gaithersburg, MD, USA). Autoclave Engineers (Erie, PA, USA) micrometering valves (VM-1) were used for flow control throughout the apparatus. Heating tapes were also used around these valves to prevent freezing following depressurization. Pressure in extractor and columns were monitored with a digital transducer system (G1, G2 and G3, Heise Series 901A RTS acquired from Dresser Industries, Stratford, CT, USA) with

188

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

Fig. 2. Green coffee oil chromatogram. Peaks are 1, LLL; 2, PLLn; 3, OLL; 4, PLL; 5, OLO; 6, PLO + SLL; 7, PLP; 8, OOO; 9, POP; 10, OOS; 11, SOS. L = linoleic acid; Ln = linolenic acid; P = palmitic acid; O = oleic acid; S = stearic acid. Ex: POP = 1,3 palmitic-2-oleic-glycerol.

a precision of ±0.03 MPa. Extractor and columns temperatures were controlled to ±0.5 ◦ C. 2.3. Analyses The compositions of the coffee extract fractions obtained were determined by HPLC using Shimadzu LC-10AD chromatograph (Kyoto, Japan) and a C18 column (25 cm × 4.6 mm, 5 ␮m) purchased from Shimadzu (Kyoto, Japan). For determination of the oil content in the extracts the mobile phase used was acetone/acetonitrile (62:38, v/v) with a flow rate of 1.0 mL/min. A Shimadzu refractive index detector model RID 10A (Kyoto, Japan) was used and the peaks were identified by comparison with standards and with the chromatogram published by Gonz´alez et al. [9]. Caffeine and chlorogenic acid amounts in the extracts were determined using a SPD 10AV detector (Shimadzu) operating at 280 and 313 nm, respectively. Methanol (50%, v/v) and (40%, v/v) in aqueous sodium acetate (0.5%) at a flow rate of 0.8 mL/min was used as the mobile phase. Fig. 3. Green coffee oil extraction curves (a) 50 ◦ C, (b) 60 ◦ C and (c) 70 ◦ C.

2.4. Experimental procedures Experiments of green coffee oil fractionation with carbon dioxide were carried out at 50, 60 and 70 ◦ C and in the pressure range from 15.2 to 35.2 MPa. CO2 flow rate was maintained at 1.8 g/min, as recommended by Filippi [11] and also based on our own experience with oil extraction [12–14], which confirmed that no mass transfer limitations were observed and in the equilibrium concentration of the solute with the particular solvent in question was always achieved. A 15 ± 1.5 g sample of

GCO was mixed with glass beads (4 mm in diameter) and placed in the extraction vessel. In a typical extraction experiment, solvent (CO2 ) was delivered by the pumps as a liquid and slowly fed into the extractor until the desired extraction pressure was reached. The extractor was heated to the extraction temperature and the micrometering valve downstream was slowly opened, always keeping constant pressure in the extractor. The micrometering valve was properly heated to prevent solvent freezing

Table 1 TAG composition in crude green coffee oil TAG

LLL

PLLn

OLL

PLL

OLO

PLO + SLL

PLP

OOO

POP

OOS

SOS

Composition (%) Peak no. in Fig. 2 a

7.76 1

2.35 2

5.84 3

32.74 4

1.56 5

19.53 6

26.74 7

0.27 8

0.65 9

1.77 10

0.79 11

a 1, LLL; 2, PLLn; 3, OLL; 4, PLL; 5, OLO; 6, PLO + SLL; 7, PLP; 8, OOO; 9, POP; 10, OOS; 11, SOS. L = linoleic acid; Ln = linolenic acid; P = palmitic acid; O = oleic acid; S = stearic acid. Ex: POP = 1,3 palmitic-2-oleic-glycerol.

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

189

and plugging with precipitated solid solute. The precipitated oil following depressurization was recovered in the collection flasks immersed in a cooling bath. The last collection flask contained ethanol to assure complete recovery of the precipitated oil. In these particular experiments, fractions were collected at time intervals corresponding to a delivery of 183 g of CO2 to the extractor at constant pressure and temperature until oil depletion. The amount of solvent used was measured by the gas flow meter. The weight of extracted products was determined gravimetrically and the amount of oil was the difference between the total amount of extract obtained and the amount of caffeine, determined by HPLC. Fig. 4. Green coffee oil solubility as a function of CO2 density.

3. Discussion of the results 3.1. Green coffee oil extraction All the GCO fractions obtained by supercritical CO2 extraction were clear and transparent, with an intense yellow color. After the depressurization step, at the end of each experimental run, a highly viscous residue was collected in the extractor bottom. This residue did not contain triglycerides and was identified as coffee wax (polyalcohols of high polarity) extracted during the cold press process and not solubilized by supercritical CO2 at pressures below 40 MPa [15]. The TAG compositions of the oil fractions collected in each experimental run were determined by HPLC analysis and Fig. 2 shows a typical chromatogram obtained for raw coffee oil. In addition to the nine compounds identified by Gonz´alez et al. [9], two more TAG (OOS and SSO (see legend of Fig. 2) could be identified by comparing the retention times observed with standards. Table 1 shows the average TAG composition found for the raw GCO. The composition of the oil fractions collected in the supercritical extraction process was very similar among the different conditions used in the extractions. Considering the seven main TAGs (see Table 1 and legend of Fig. 2) the composition

Fig. 5. Green coffee oil solubility as a function of pressure.

was: LLL, 6.95%; PLLn, 2.37%; OLL, 5.89%; PLL, 32.39%; PLO + SLL, 19.82; PLP, 31.08%. The other TAGs amounted to approximately 1.5%. Each fraction obtained in the extractions were analyzed twice by HPLC and showed a maximum deviation of 3.7%. The lack of appreciable changes in the oil composition indicates that the extraction procedures were inefficient to frac-

Table 2 Caffeine initial mass and recovery in the supercritical CO2 extraction of green coffee oil Pressure (MPa)

Temperature (◦ C)

Oil mass fed (g)

Caffeine initial mass (g)

Caffeine mass extracted (g)

Caffeine recovered mass (%)

15.2

50 60 70

15.10 14.76 15.04

0.04077 0.03985 0.04061

0.03760 0.0299 0.02989

92.22 75.03 73.61

24.8

50 60 70

15.26 14.12 14.30

0.04120 0.03812 0.03861

0.03933 0.03366 0.03604

95.46 88.29 93.34

28.3

50 60 70

15.60 15.14 16.26

0.04212 0.04088 0.04390

0.03822 0.03877 0.03858

90.74 94.84 87.89

31.7

50 60 70

15.10 15.06 15.54

0.04077 0.04066 0.04196

0.03553 0.03536 0.03673

87.15 86.96 87.54

35.2

50 60 70

16.70 16.19 17.65

0.04509 0.04370 0.04766

0.03842 0.03759 0.04182

85.21 86.01 87.76

190

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

tionate the TGAs, which might be attributed to the molecular weight distribution of TAGs in the coffee oil. Coffee oil is a mixture of TAGs in a narrow range of molecular weight. Perry et al. [16] reported that TAGs of same molecular weights have approximately the same volatility and it is thus expected that the TAGs contained in GCO have the same solubility in supercritical CO2 . The same trend was observed for pigments distribution in the GCO fractions collected during the extraction and a constant extraction ratio in supercritical CO2 extraction process was verified. Based on these observations it is reasonable to assume GCO as a unique pseudo-component in thermodynamic modeling. The complex mixture that constitutes GCO could then be modeled as one component, resulting in a binary system of GCO and CO2 . It is expected that the low content of wax and caffeine in the system would not influence in a significant extent the thermodynamic equilibrium. The extraction curves obtained at 50, 60 and 70 ◦ C are shown in Fig. 3A–C. Each experimental point on the extraction curves represents the average values of two independent experiments with reproducibility within ±5.4%. As can be seen in Fig. 3, the amount of coffee oil extracted per solvent mass used remains constant over the entire extraction period, until almost all the GCO is extracted, indicated by a change in the slope of the extraction curves. The arrangements of the isobaric curves shown in Fig. 3A–C reveals that as the pressure increases at constant temperature, the extraction ratio of GCO (defined as the quotient between the oil mass extracted and the solvent mass used) also increases. Assuming that equilibrium has been reached in this dynamic method, the extraction ratios of GCO obtained are also solubility data. As previously discussed, the effect of pressure can be attributed to the increase in solvent power and by the strengthening of intermolecular physical interactions [17–19]. This is confirmed by the data shown in Fig. 4, where the amount of GCO extracted per solvent mass is presented as a function of the solvent density for each isotherm. As expected, the increase in extraction ratio correlates well with the increase in solvent density. This pressure effect is in agreement with similar trends found by other authors [20,21]. The results presented in Fig. 4 also show that at the same solvent density, GCO solubility increases with an increase in temperature, which can be attributed to the increase of the oil components vapor pressure. The temperature effect on the extraction ratio of GCO is presented in Fig. 5. The experimental data indicate that at pressures below 30 MPa, an increase in temperature reduces the coffee oil extraction ratio, an indicative of retrograde behavior. The data suggests that at pressures higher than 30 MPa coffee oil extraction ratio is temperature independent, and the three isotherms merge to the same extraction ratio curve. This extraction ratio behavior with temperature found for this system is frequently observed in supercritical fluid systems and was previously reported by Hadolin et al. [22], in the study of Silybum marianum oil extraction using supercritical CO2 as a solvent and Ill´es et al. [23]

in components extraction of Rosa Canina using supercritical CO2 or propane as solvents. These authors verified that the oil solubility isotherms obtained at 35 and 55 ◦ C and in the pressure range from 10 to 40 MPa, cross at approximately 31 MPa. The retrograde behavior observed at the lower pressure levels investigated in this work is commonly found in supercritical extraction processes with vegetable oils and these results are in agreement with those obtained by Salda˜na et al. [13], Temelli [24] and Azevedo et al. [14], who identified the same value for crossover pressure in the extraction of cupuac¸u fat using supercritical CO2 .

Fig. 6. Caffeine extraction curves (a) 50 ◦ C, (b) 60 ◦ C and (c) 70 ◦ C.

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

3.1.1. Caffeine extraction The raw coffee oil fed to the extractor vessel (obtained by cold press) initially presented 2.7 g caffeine/kg oil. As coffee oil was extracted, caffeine was also extracted (Fig. 6A–C). Table 2 shows the caffeine initial mass and the amount recovered. Comparing the coffee oil extracted mass percent (considering the mass fed to the extractor) with caffeine extracted mass percent for each extraction run, it could be seen that caffeine in GCO was extracted in the three or four initial fractions collected. Regarding the fractionation of caffeine in the oil, the best results were observed at 15.2 MPa and 70 ◦ C, since under these conditions, caffeine was removed from GCO with only 6% of the lipid mass extracted (see Figs. 3 and 6). In other conditions the same amount of caffeine was recovered with a higher amount of GCO. 3.2. Solubility correlations To correlate the solubility of coffee oil in the supercritical CO2 , the Chrastil [10] model was adopted. It revealed to be useful in correlating vegetable oils solubility [25,26]. By regression of the experimental data using the isotherms obtained at 50 and 70 ◦ C, expression (1) was obtained for the solubility of coffee oil in supercritical CO2 :   −2869.20 7.90 C=d exp − 42.63 (1) T where C is the solubility (g/L), d (g/L) is the density of pure CO2 and T (K) is the absolute temperature. The experimental values represented by points and the calculated solubilities (lines) are shown in Fig. 7. The percent errors in the adjustment to the model were 6.28, 6.72 and 9.89 for the isotherms obtained at 50, 60 and 70 ◦ C, respectively. The values found for the parameters are very close to previous ones reported by Azevedo et al. [14] correlating cupuac¸u fat solubility in supercritical CO2 , an indicative that vegetable oils composed by TAGs of the same range of molecular weight present similar solubility behavior. This assumption is in agreement with Del Valle and Aguilera [27] who presented a general density and temperature-dependent correlation for vegetable oils in supercritical CO2 .

Fig. 7. Experimental data of green coffee oil solubility (C, g/L) as a function of CO2 density (d, g/L) and results from the model of Chrastil (lines).

191

4. Conclusions In this work, GCO was extracted in a high-pressure apparatus using supercritical carbon dioxide. The extraction of GCO increased with solvent density and a retrograde behavior was observed in the pressure range of 15.2–31.7 MPa. CO2 showed higher affinity for the non-polar TAGs. Regarding the fractionation of coffee oil, the TAGs composition of GCO remained approximately constant due to the lack of selectivity of CO2 for any specific TAG present in the coffee oil. The analyses of the extraction samples indicated that caffeine was removed in the first fractions, leaving GCO fractions almost caffeine-free and also a highly viscous residue, believed to be wax, in the bottom of the extraction vessel at the end of the extraction. An effective fractionation of green coffee oil was obtained. The best results were obtained at the lower pressure, 15.2 MPa and at 70 ◦ C. The model of Chrastil was able to correlate satisfactorily the GCO solubility in CO2 . The GCO was modeled as a pseudocomponent, in order to simplify the calculations. Acknowledgements The authors wish to thank CNPq-Brasil and FAPESP, S˜ao Paulo, for financial support to this work. References [1] T. Gassenmeier, P. Busch, H. Hensen, W. Seipel, Some aspects of refatting the skin, Cosmet. Toiletries 113 (1998) 89–92. [2] B. Nikolova-Damyanova, R.E. Velikova, G.N. Jham, Lipid classes. Fatty acid composition and triacyglycerol molecular species in crude coffee beans harvested in Brazil, Food Res. Int. 31 (1998) 479–486. [3] P. Folstar, in: R.J. Clarke, R. Macrae (Eds.), Lipids in Coffee, Elsevier, London, 1985, pp. 203–221. [4] T. Beveridge, T.S.C. Li, B.D. Oomah, A. Smith, Sea Buckthorm products: manufacture and composition, J. Agric. Food Chem. 47 (1999) 3480–3488. [5] A. Van Tol, R. Urget, R. Jong-Caesar, T. van Gent, L.M. Scheek, B. Ross, M. Katan, The cholesterol-raising diterpenes from coffee beans increase serum lipid transfer protein activity levels in humans, Atherosclerosis 132 (1997) 251–254. [6] C. Cavin, D. Holzhaeuser, G. Scharf, Cafestol and kahweol, two coffee specific diterpenes with anticarcinogenic activity, Food Chem. Toxicol. 40 (2002) 1155–1163. [7] J.F. Follier, S. Plessis, Use of coffee beans oil as a sun filter. US Patent 4,793,990, (1988). [8] C. Cavin, D. Holzhauser, A. Constable, A.C. Huggett, B. Schilter, The coffee-specific diterpenes cafestol and kahweol protect against aflatoxin B1-induced genotoxicity through a dual mechanism, Carcinogenesis 19 (1998) 1369–1375. [9] A.G. Gonz´alez, F. Pablos, M.J. Mart´ın, M. Le´on-Camacho, M.S. Valdenebro, HPLC analysis of tocopherol and triglycerides in coffee and their use as authentication parameters, Food Chem. 73 (2001) 93–101. [10] J. Chrastil, Solubility of solids and liquids in supercritical gases, J. Phys. Chem. 86 (1982) 3016–3021. [11] R.P. Filippi, CO2 as a solvent: applications to fats, oils and other materials, Chem. Ind. 19 (1982) 390–394. [12] R.S. Mohamed, M.D.A. Salda˜na, F.H. Socantaype, T.G. Kieckbusch, Reduction in cholesterol content of butter oil using supercritical ethane extraction and adsorption on alumina, J. Supercrit. Fluids 16 (2000) 225–233. [13] M.D.A. Salda˜na, R.S. Mohamed, P. Mazzafera, Extraction of cocoa butter from Brazilian cocoa beans using supercritical CO2 and ethane, Fluid Phase Equilibria 194 (2002) 885–894.

192

A.B.A. de Azevedo et al. / J. of Supercritical Fluids 44 (2008) 186–192

[14] A.B.A. Azevedo, R.S. Mohamed, U. Kopcack, Supercritical fluid extraction and fractionation of cupuac¸u fat from fermented seeds with supercritical solvents, J. Supercrit. Fluids 27 (2003) 223–237. [15] L. Taylor, Supercritical Fluid Extraction, John Wiley & Sons Inc., New York, 1996. [16] E.S. Perry, W.H. Weber, B.F. Daubert, Vapor pressure of phlegmatic liquids. I. Simple and mixed triglycerides, J. Am. Chem. Soc. 71 (1949) 3720–3726. [17] A. Morita, O. Kajimoto, Solute–solvent interaction in nonpolar supercritical fluid: a clustering model and size distribution, J. Phys. Chem. 94 (1990) 6420–16420. [18] S. Bai, M.V. Craig, L.F. Liu, C.L. Mayne, R.J. Pugmire, D.M. Grant, CO2 clustering of 1-decanol and methanol in supercritical fluids by 13C nuclear spin-lattice relaxation, J. Phys. Chem. 101 (1997) 2923–2928. [19] D.S. Bulgarevicg, T. Sako, T. Sujeta, K. Otake, Y. Takebayashi, C. Kamizawa, Y. Horikawa, M. Kato, The role or general hydrogen-bonding interaction in the solvation process of organic compounds by supercritical CO2 /n-alcohol mixtures, Ind. Eng. Chem. Res. 41 (2002) 2074–2081. [20] J.P. Friedrich, G.R. List, A.J. Heakin, Petroleum-free extraction of oil from soybeans with supercritical CO2 , J. Am. Oil Chem. Soc. 59 (1982) 288–292.

[21] S. Li, S. Hartland, A new industrial process for extracting cocoa butter and xanthines with supercritical carbon dioxide, J. Am. Oil Chem. Soc. 73 (1996) 423–429. ˇ ˇ Knez, D. Bauman, High-pressure extraction [22] M. Hadolin, M. Skerget, Z. of vitamin E-rich oil from Silybum marianum, Food Chem. 74 (2001) 355–364. [23] V. Ill´es, O. Szalai, M. Then, H. Daood, S. Perneczki, Extraction of hiprose fruit by supercr´ıtical CO2 and propane, J. Supercrit. Fluids 10 (1997) 209–218. [24] F. Temelli, Extraction of triglycerides and phospholipids from canola with supercritical carbon dioxide and ethanol, J. Food Sci. 57 (1992) 440–457. [25] P. Maheshwari, Z.L. Nikolov, T.M. White, R. Hartel, Solubility of fatty acids in supercritical carbon dioxide, J. Am. Oil Chem. Soc. 69 (1992) 1069–1076. [26] H. Sovov´a, M. Zarev´ucka, M. Vacek, K. Str´ansk´y, Solubility of two vegetable oils in supercritical CO2 , J. Supercrit. Fluids 20 (2001) 15–28. [27] J.M. Del Valle, J.M. Aguilera, An improved equation for predicting the solubility of vegetable oils in supercritical CO2 , Ind. Eng. Chem. Res. 27 (1988) 1551–1553.