Supercritical Carbon Dioxide Extraction of High-Value Substances from Soybean Oil Deodorizer Distillate

Ind. Eng. Chem. Res. 2000, 39, 4521-4525

4521

Supercritical Carbon Dioxide Extraction of High-Value Substances from Soybean Oil Deodorizer Distillate Chiehming J. Chang,* Yu-Fang Chang, Hong-zhi Lee, Jia-qun Lin, and Po-Wen Yang† Department of Chemical Engineering, National Chung-Hsing University, #250, Kuo-Kuang Road, Tai-chung, Taiwan 402, Republic of China

This work attempted to recover high-value substances such as free fatty acids, tocopherols, sterols, and squalene by investigating a supercritical fluid CO2 (SF CO2) distillation-extraction processing of soybean oil deodorizer distillate. Experimental conditions varied between 50 and 90 °C, at 24.1 and 31.0 MPa. Reverse-phase high-performance liquid chromatography analysis was performed to analyze 14 components in the extracts after SF CO2 treatment. Experimental results indicated that when tocopherols were extracted at 31.0 MPa, 90 (top) to 70 (bottom) °C, and 1000-L STP CO2 usage, their recovery reached 83.6% and the average value of the concentration factor was 1.38. Furthermore, the concentration factor of tocopherols grew to 1.70 when the extracts collected in the top exceeded 400-L STP CO2. The concentration factors of fatty acids, squalene, and stereols were 1.37, 1.26, and 0.60, respectively. This investigation also found that the percentage of total fatty acids decreased, while polyunsaturated fatty acids increased, with an increase in the CO2 volume given the same extraction condition. Introduction The refining process of soybean oil byproduces soybean oil deodorization distillate (SODD), which generally comprises 1-20% vitamin E, 30-60% fatty acid, 10-35% sterols, 10-30% hydrocarbons, and 7% other substances.1,2 A commerical vaccum-distillation process of SODD could extract vitamin E, phosphatides, sterols, and squalene for nutritional and medical utilization.2 Previous literature contains a few methods for recovering vitamin E such as molecular distillation, reducedpressure steam distillation, fractionation,3-5 and esterification, which uses methanol as the solvent and HCl as the catalyst, reacting at 65 °C for 5 h.6 Additionally, the adsorption method uses silica gel to eliminate fatty acids and glycerides and then increases the concentration of vitamin E by urea.7,8 However, those methods not only complicate the process but also introduce environmental pollution. Supercritical fluid carbon dioxide (SF CO2) is a recently developed technology. For example, Moyler (1993) employed CO2 to extract herbal medicines from sources such as celery seed, hop, clove bud, juniper berry, mace, and ginger.9 Meanwhile, the paper industry has applied the technology to separate and purify chemical materials in wood, with the technology removing the lignin and also bleaching the wood.10 Carbon dioxide has also been applied to fragrances, cosmetics, food, and various chemical industries.11-13 Performing important research relating to edible oil, List and Friedrich (1985) compared two processes of soybean oil extraction, from soy flakes by hexane and from SF CO2, finding that the oil extracted by SF CO2 contained more tocopherols than that extracted by hexane. However, the

content of phosphatides in the extract was relatively lower, reducing the anti-oxidation properties of the oil and thus also its shelf life.14 Lee et al. (1991) first attempted to concentrate tocopherols from soybean sludge with the high-performance liquid chromatography (HPLC) method. Following esterification, the solubility of soybean oil in SF CO2 increased 4-6 times, especially at 35 °C and 4500 psig, and the concentration of tocopherol increased from 13-14% to 40%.15 Birtigh et al. (1995) treated wastes from palm oil manufacturing with SF CO2 to extract carotene and tocopherols from palm oil tree leaves.16 However, those results are insufficient to permit an economical industrial scaleup. Besides, King et al. (1996) extracted tocopherols from soybean powder with SF CO2 at 25 MPa and 80 °C.17 They found that the recovery of tocopherols was 60% and the enrichment factor was 1.83-4.33. Finally, Montanari et al. (1996) showed the extraction of oil and phospholipid from soybean flakes in two steps with SF CO2 and cosolvent. The first step was SF CO2 extraction of soybean oil, while the second step was SF CO2/ cosolvent (ethanol) extraction of defatted soybean flakes for a phospholipid-enriched fraction. This observation resembled traditional ethanol extraction, in which increasing the molar fraction of the cosolvent increases the recovery of the phospholipid.18 The above literature review demonstrates that SF CO2 is a promising alternative to traditional solvent extraction. This investigation examines the feasibility of SF CO2 extraction as an alternative to traditional vacuum distillation for enriching tocopherols effectively from SODD. Meanwhile, other valuable substances such as sterols, fatty acids, and squalene are investigated. Experimental Section

* To whom the correspondence should be addressed. Tel: 886-4-2852592. Fax: 886-4-2854734. E-mail: cmchang@ dragon.nchu.edu.tw. † Department of Products & Process Development, Taiwan Sugar Co., Taipei, Taiwan, Republic of China.

Materials. SODD was donated from an oil plant in southern Taiwan. Samples were divided among small bottles, which were wrapped in aluminum foil and stored at 4 °C. SODD is a deep brown viscous liquid.

10.1021/ie0003537 CCC: $19.00 © 2000 American Chemical Society Published on Web 10/27/2000

4522

Ind. Eng. Chem. Res., Vol. 39, No. 12, 2000

Figure 1. Apparatus of the SF CO2 extraction system used in this investigation.

99.5% food-grade carbon dioxide acted as the supercritical extraction solvent (Lien-Hwa, Taiwan). HPLC-grade acetonitrile (J.T. Baker, Phillipsburg, NJ) and acetic acid (Shimaku, Osaka, Japan) served as a mobile-phase solvent. Methanol (99.8%) was obtained from Mallinckrodt (Seelze, Germany) for HPLC sample preparation. Meanwhile, acetone (99.5%) from Merck (Seelze, Germany) was used to wash the extraction unit. Four-type standards used in the analysis are free fatty acids (Sigma, St. Louis, MO), including linolenic acid, linoleic acid, palmitic acid, oleic acid, stearic acid, eicosenoic acid, and erucic acid; R-, β-, γ-, and δ-tocopherol (Merck, Darmstadt, Germany); sterols (Sigma, St. Louis, MO), including campesterol, β-sitosterol, and stigmasterol; and squalene (Sigma, St. Louis, MO). Apparatus and Extraction. Figure 1 schematically depicts the experimental apparatus. The equipment consisted primarily of a 100-mL syringe pump (100DX, ISCO, Lincoln, NE), a preheating coil, a 170-mL extraction vessel with a 50-mL reboiler fitted in the bottom, a temperature and pressure controlling unit, a 130-mL

sight glass-gauge precipitator (separator), and a backpressure regulator controlling its pressure. The extraction vessel was packed with 160 pieces of a 200-mesh stainless steel filling (0.074 cm o.d. × 2.5 cm length) to increase the plate quantity and mass-transfer efficiency. Liquid CO2 was delivered at a constant 5 mL/min in each experiment. The back-pressure of the separator was set at 650 psig. A sample was taken after each 200-L period of CO2 in STP condition flow through the wet gas meter. Table 1 lists the experimental condition applied for each run. All experiments were conducted using 1000-L STP CO2 extraction. Finally, sample analysis determined the influence of each operating condition on the extraction efficiency. HPLC Analysis. This study simultaneously analyzed the content of seven fatty acids, three tocopherol homologues, three sterols, and squalene in each sample. The SF CO2 extracted samples were analyzed using an HPLC isocratic program, consisting of a modular chromarographic unit (Hitachi L-7100 pump and L-4200H UV/vis detector, Tokyo, Japan) equipped with an injec-

Ind. Eng. Chem. Res., Vol. 39, No. 12, 2000 4523 Table 1. Experimental Data of SODD in SF CO2a P (MPa)

top

24.1 24.1 31.0 31.0 31.0 31.0 31.0

90 90 50 70 90 90 90

T (°C) bottom 90 70 50 70 90 50 70

oil yield (%)

∑RVitE (%)

Rg400 (%)

βR400

RR400

32.7 36.2 66.9 62.5 68.4 62.3 65.5

25.9 26.9 64.0 70.8 82.4 64.9 83.6

13.7 16.7 41.1 43.3 51.7 38.8 49.6

0.9 0.9 1.5 1.6 1.5 1.3 1.7

0.6 0.7 1.8 2.0 1.9 1.4 1.7

a ∑R g400 (%): VitE (%): accumulative recovery of tocopherols. R accumulative recovery of tocopherols when over 400 L of CO2 was collected. RR400: separation factor of tocopherols when over 400 L of CO2 was collected. βR400: concentration factor of tocopherols when over 400 L of CO2 was collected. top: top of the extractor. bottom: bottom of the extractor.

tion valve with a 20-µL sample loop and a data integrator. The extracts of unesterified plant oils were directly analyzed in the HPLC system. Each 4-mg sample was accurately weighed, then dissolved in 3 g of methanol, and filtered through a 0.22-µm syringe-membrane filter. Samples were analyzed with a reverse-phase octylbonded silica column (Inertsil C8-3; 5-µm particle diameter, 250 mm length × 4.0 mm i.d.; Tokyo, Japan) fitted with a guard column (Inertsil; 5-µm particle diameter, 35 mm length × 4.6 mm i.d.) monitored at wavelength of 210 nm and eluting isocratically with 85% acetonitrile, 5% methanol, and 10% H2O containing 1% acetic acid at a flow rate of 1.3 mL/min for the first 25 min and 2.5 mL/min for the subsequent 20 min. The analysis was conducted at 40 °C. The recovery, precision, minimum detection limit, and linearity range of this HPLC system is 95%, (1 mg/kg, (0.1 mg/kg, and 0.999, respectively. Samples were quantified by an external standard method using the peak area. Thereafter, the recovery of tocopherols, yield, separation factor (R), and concentration factor (β) were defined as the following equations:

of VitE in the extract (weight weight of VitE in the feed ) weight of the extract yield ) ( weight of the feed ) weight of VitE in the extract ( weight of VitE in the feed ) R) of other types in the extract (weight weight of other types in the feed ) VitE recovery β)( ) yield

VitE recovery )

(1) (2)

(3)

(4)

Results and Discussion Five temperatures, of 50, 70, 90, 90 (top) to 50 (bottom), 90 (top) to 70 (bottom) °C, and pressures of 24.13 and 31.03 MPa were used in this study. Table 1 lists the extraction conditions. The temperature was controlled in two portions, including the extractor (top) and the reboiler (bottom). The original SODD feed contains 30.91% free fatty acids, 12.19% tocopherols, 11.02% sterols, 1.99% squalene, and 43.89% low-volatile substances. The SODD and extracts from SF CO2 extraction were analyzed by RP-HPLC. Fourteen peaks were identified according to a standard mixture chromatogram, as Figure 2 displays. Peaks of fatty acids

Figure 2. Standard chromatogram (1, linolenic acid; 2, linoleic acid; 3, palmitic acid; 4, oleic acid; 5, stearic acid; 6, eicosenoic acid; 7, erucic acid; 8, δ-tocopherol; 9, β,γ-tocopherol; 10, campesterol; 11, stigmasterol; 12, R-tocopherol; 13, β-sitosterol; 14, squalene). Chart speed: 5 mm/min.

appeared at 3-14 min, tocopherols at 16-24 min, sterols at 20-25 min, and squalene at 44 min. Experimental results indicated that extracted oil and the recovery of tocopherols obviously increased with extraction at higher temperature and pressure. Extraction at 31.0 MPa yields twice the oil of extraction at 24.1 MPa. Notably, during extraction at 31.0 MPa with a twin temperature control (extractor at 90 °C and reboiler at 70 °C), roughly, 65.5% yield of oil, 83.6% recovery of tocopherols, and a 1.38 concentration factor (β factor) were obtained. This observation implies that tocopherols were more soluble at high temperature and high pressure. Results of the high-pressure experiment in this study resemble those of Pessoa et al. (2000) and King et al. (1996), in which when soybean was processed at 25 MPa and 80 °C, the recovery of tocopherols was 60%.17,19 Fatty acids, tocopherols, sterols, and squalene were collected at five fractions, namely, 200, 400, 600, 800, and 1000-L STP CO2. The accumulative recoveries of these four groups were plotted against the volume of CO2 used. Figure 3 represents the recovery rates. According to this figure, 50% of fatty acids, sterols, and squalene were extracted at fractions 1 and 2, but tocopherols were mostly extracted at fractions 4 and 5. Meanwhile, with 600-1000 L of STP CO2 used, the concentration factor of tocopherols in the extracts could increase to 1.7 (average value). Figure 4 presents the separation factor (R factor) of tocopherols against the other three types, which increased with CO2 volume. This phenomenon implies that the separation efficiency of tocopherols with respect to the other three types improves with CO2 volume. Therefore, it is possible to separate tocopherols preliminarily within fractions 4 and 5. Figure 5 presents concentration factors in the extracts at various conditions. The optimal extractive condition was at 31.03 MPa and 90 (top) to 70 (bottom) °C. Figure 6 depicts the change in the concentrations of saturated fatty acid (SFA, namely, palmitic acid/stearic acid), monounsaturated fatty acid (MUFA, namely, oleic acid/eicosenoic acid/erucic acid), and polyunsaturated fatty acid (PUFA, namely, linolenic acid/linoleic acid) in total fatty acids (TFA). It indicated that the concen-

4524

Ind. Eng. Chem. Res., Vol. 39, No. 12, 2000

Figure 3. Differential recovery of four groups at 31.0 MPa and 90-70 °C.

Figure 4. Separation factor of tocopherols (0, 90 °C, 24.1 MPa; +, 90-70 °C, 24.1 MPa; b, 50 °C, 31.0 MPa; 0, 70 °C, 31.0 MPa; 4, 90 °C, 31.0 MPa; O, 90-50 °C, 31.0 MPa; [, 90-70 °C, 31.0 MPa).

tration of PUFA increased when in excess of 400 L of CO2 was collected at 31.0 MPa and 90-70 °C extraction. These results indicate that PUFA was rudimentarily concentrated in the latter collection. Conclusions This investigation showed that it was possible to recover tocopherols from SODD by using a batch-type SF CO2 distillation-extraction processing with the stepwise collection. It is the first trial to analyze 14 components in the extracts after SF CO2 treatment, by

Figure 5. Concentration factor of tocopherols (with the symbols as those in Figure 4).

Figure 6. Composition of fatty acids in extracts at 31.0 MPa and 90-70 °C [(9) TFA ) total fatty acids in the extraction; (O) SFA ) saturated fatty acids in TFA; (2) MUFA ) monounsaturated fatty acids in TFA; (]) PUFA ) polyunsaturated fatty acids in TFA].

using the reverse-phase HPLC analysis. Experimental results indicated that when tocopherols were extracted at 31.0 MPa, 90 (top) to 70 (bottom) °C, and 1000-L STP CO2 usage, their recovery reached 83.6% and the average value of the concentration factor was 1.38. Furthermore, the concentration factor of tocopherols grew to 1.70 when the extracts collected in the top exceeded 400-L STP CO2. Finally, free fatty acids and squalene also concentrated in the extracts, while stereols concentrated in the raffinates. This investigation found the percentage of TFA decreased, while PUFA increased, with an increase in the STP CO2 volume given the same extraction condition.

Ind. Eng. Chem. Res., Vol. 39, No. 12, 2000 4525

Acknowledgment The authors thank Taiwan Sugar Co. for their valuable assistance and corporation. The authors also thank the National Science Council of the Republic of China for financially supporting this research under Contract NSC88-TSC-B005-006. Literature Cited (1) Takagi, Y.; Kai, Y. Process for Preparation of Tocopherol Concentrates. U.S. Patent 4,454,329, 1984. (2) Woerfel, J. B. Processing and Utilization of By-Products from Soy Oil Processing. J. Am. Oil Chem. Sci. 1981, Mar, 188191. (3) Yoshino, Y.; Kondo, K. Process for the Preparation of DLR-tocopherol of High Purity. U.S. Patent 4,217,285, 1980. (4) Sampathkumar, L. Process for Recovering Tocopherols from Deodorizer Sludge. U.S. Patent 4,594,437, 1986. (5) Thies, M. C.; Mullins, J. C.; Briones, J. A. Solvent Extraction of Fatty Acid Stream with Liquid Water and Elevated Temperatures and Pressures. U.S. Patent 5,097,012, 1992. (6) Mulholland, T. P. C. Ester of R-tocopherol. U.S. Patent 5,487,817, 1992. (7) Christian, F. Process for Tocopherols and Sterols from Natural Sources. U.S. Patent 5,487,817, 1996. (8) Barnicki, S. D.; Sumner, C. E.; Williams, H. C. Process for the Production of Tocopherol Concentrates. U.S. Patent 5,512,691, 1996. (9) Moyler, D. A. Extraction of Essential Oils with Carbon Dioxide. Presented at the 23th International Symposium on Essential Oils, Auchincruive, Ayr, Scotland, Sept 9-12, 1992. (10) Kiran, E.; Balkan, H. High-pressure Extraction and Delignifiction of Red Spruce with Binary and Ternary Mixtures of Acetic Acid, Water, and Supercritical Carbon Dioxide. J. Supercrit. Fluids 1994, 7 (2), 75-86. (11) Garcia, A.; Lucas, A. D.; Rincon, J.; Alvarez, A.; Gracia, I.; Garcia, M. A. Supercritical Carbon Dioxide Extraction of Fatty

and Waxy Material from Rice Bran. J. Am. Oil Chem. Sci. 1996, 73 (9), 1127-1131. (12) Palmer, M. V.; Ting, S. S. T. Applications for Supercritical Fluid Technology in Food Processing. Food Chem. 1995, 52, 345352. (13) Turkay, S.; Burford, M. D.; Sangum, M. K. Deacidifiction of Black Cumin Seed Oil by Selective Supercritical Carbon Dioxide Extraction. J. Am. Oil Chem. Sci. 1996, 73 (10), 1265-1270. (14) List, G. R.; Friedrich, J. P. Processing Characteristics and Oxidative Stability of Soybean Oil Extracted with Supercritical Carbon Dioxide at 50 °C and 8000 psi. J. Am. Oil Chem. Sci. 1985, 62 (1), 82-84. (15) Lee, H.; Chung, B. H.; Park, Y. H. Concentration of Tocopherols from Soybean Sludge by Supercritical Carbon Dioxide. J. Am. Oil Chem. Sci. 1991, 68 (8), 571-573. (16) Birtigh, A.; Johannsen, M.; Brunner, G. Supercritical-fluid Extraction of Oil-palm Components. J. Supercrit. Fluids 1995, 8 (1), 46-50. (17) King, J. W.; Favati, F.; Taylor, S. L. Production of Tocopherol Concentrates by Supercritical Fluid Extraction and Chromatography. Sep. Sci. Technol. 1996, 31 (13), 1845-1857. (18) Montanari, L.; King, J. W.; List, G. R.; Rennick, K. A. Selective Extraction of Phospholipid Mixtures by Supercritical Carbon Dioxide and Cosolvents. J. Food Sci. 1996, 61 (6), 12301233. (19) Pessoa, F. L. P.; Mendes, M. F.; Uller, A. M. C. Extraction of Tocopherol from a Deodorized Distillated in a Countercurrent Column Using Supercritical Carbon Dioxide. In CD-ROM Proceedings of the 5th International Symposium on Supercritical Fluids, Atlanta, GA, April 8-12, 2000; Eckert, C. A., Teja, A. S., Eds., 23-29.

Received for review March 30, 2000 Revised manuscript received June 30, 2000 Accepted July 4, 2000 IE0003537