Abstract

SIMULATION AND THERMODYNAMIC MODELING OF THE EXTRACTION OF TOCOPHEROL FROM A SYNTHETIC MIXTURE OF TOCOPHEROL, SQUALENE AND CO₂

M.F.Mendes1^* * To whom correspondence should be addressed , A.M.C.Uller¹ and F.L.P. Pessoa 2

¹ Programa de Engenharia Química/COPPE/UFRJ, Centro de Tecnologia, Bloco E, Sala 115,

CEP 21945-970, Cidade Universitária, Rio de Janeiro - RJ, Brazil

² Escola de Química (UFRJ), Centro de Tecnologia, Bloco E, Sala 209, CEP 21949-000,

Cidade Universitária, Rio de Janeiro - RJ, Brazil

(Received: October 7; Accepted: April 7, 2000)

INTRODUCTION

The oils and fatty acids are consumed all over the world. The vegetable oils represent 70% of the demand including all types of natural oils and fatty acids. The soybean oil is responsible by 20 to 30% of this market in the world. According to Eitenmiller (1997), the soybean and palm oils are responsible for 50% of this market, the first one representing 29%.

The production of the soybean oil represents an important stage of the market in Brazil. This occurrs due to the extensive utilization of the products in the metalurgical, food and cosmetic industries. This market involves the production of the seeds and of the oils (raw and refined) and the production of the soybean bran (Barbosa et al., 1998). Due to these characteristics, in 1997/1998, the Brazilian production of soybean seeds reached 30,9 millions of tons, with an increase of 18.2% compared to the last year. For 1999 this production is estimated in 38 million of tons of soybean seeds.

The soybean oil can be used in many industries such as food and cosmetic. However, to be used for human consumption, it needs to pass through a series of processing treatment steps that includes the refining, the bleaching and the deodorization step. The deodorization step produces a residue that is rich in Vitamin E, hydrocarbons and sterols, principally campesterol, stigmasterol and b -sitosterol.

The Vitamin E, commonly known as a -tocopherol, is composed by four isomers, a, b, d and g - tocopherols. The a-tocopherol is the major activity isomer (Carpenter, 1979, Speek et al., 1985 and Eitenmiller, 1997). The differences between the isomers are in its structural forms. The structural formulas of the four isomers of tocopherol are presented in Figure 1 and in Table 1.

The concentration of tocopherols in the deodorizer distillate of the soybean oil varies from 13 to 14%, according to Brunner et al. (1991) and Lee et al. (1991). Marquezan (1988) confirmed that the soybean sludge is concentrated in 9% in relation to tocopherols, 26% in sterols and 3.5% in squalene.

The extraction of Vitamin E from this raw material have been carried out using the molecular distillation process (Byung-Soo et al., 1997), but this process involves high temperatures that can deteriorate the components present in the deodorizer distillate. More recently, the separation and concentration of individual tocopherols from the enriched deodorizer distillate has been investigated using supercritical carbon dioxide (Byung-Soo et al., 1997). Supercritical fluids can be attractive solvents in difficult separations. In food processing, carbon dioxide is the solvent of first choice because of its advantages like low cost, atoxicity, lower critical temperature and high volatility.

The use of supercritical CO₂ for the extraction and purification of vegetable oils has been extensively studied for several years. It has a number of processing advantages over hexane conventional extraction methods which include the absence of residues of the organic solvent and the thermal degradation of the natural products (Byung-Soo et al., 1997).

The motivation of this study is concentrate the tocopherol and make possible its use in the human diet. This work also studies the separation and concentration of the others compounds like squalene and sterols. The sterols are important in the production of hormones and in the artificial production of other vitamins and the squalene is used in the cosmetic industry and in the biosynthesis of the cholesterol (Bondioli et al., 1993).

So, the objective of this work is to study the equilibrium between the major components present in the soybean sludge with the supercritical carbon dioxide and propose a process to concentrate tocopherol from the deodorizer distillate of the soybean oil.

LITERATURE REVIEW

The motivation for concentrating tocopherol from the deodorizer distillate using supercritical carbon dioxide initiated in the decade of 80’s, but no process had been proposed to concentrated it.

Lee et al. (1991) concentrated 40% of the tocopherols present in the soybean sludge using an extractor at temperature and pressures ranging from 35 to 70⁰C and 200 to 400 bar, respectively. Better results of the concentration of the tocopherols (40%) can be observed using a countercurrent multistage column (Brunner et al., 1991, Brunner, 1994). Brunner (1994) studied the tocopherol concentration from a synthetic mixture composed of a-tocopherol and squalene using two columns, one representing the stripping and the other representing the enrichment. The conclusion obtained was that the separation between the two components is very difficult and the difficulty increases when the squalene concentration increases. King et al. (1996) studied the tocopherol concentration by supercritical fluid extraction and chromatography using different raw materials as soyflakes, rice bran and wheat germ. The extractions were conducted at 250 bar from 40 to 80⁰C and at 700 bar at 80⁰C. The soybean flakes provide the better yields for the extraction of the tocopherols and of the oil. Another different raw material, garlic, was used to concentrate tocopherol using supercritical carbon dioxide between 80 and 160 bar and 25 to 60⁰C (Byung et al., 1997). The best results were obtained at 30 to 35⁰C and 140 bar. Corsini et al. (1997) extracted the vitamin E from different vegetable matrices using SC-CO₂ at 40⁰C and from 100 to 240 bar. The experimental procedure was made in two stages and the best results encountered for tocopherol were in the second step of the extraction at 240 bar at 60 minutes of extraction.

The studies on the extraction of tocopherol present different results according to the raw materials utilized. The operational conditions applied were similar. There is only one work involving the deodorizer distillate and simulating the extraction of the vitamin E (Lee et al., 1991).

The equilibrium data available in the literature is scarce for the main components present in the soybean sludge. The binary system CO₂-a-tocopherol has equilibrium and solubility datas in different operational conditions (Chrastil, 1982; Ohgaki et al., 1989; Pereira et al., 1993; Meier et al, 1994). The binary system CO₂-squalene has less equilibrium and solubility data compared to tocopherol (Catchpole et al., 1997, 1998). Equilibrium data for sterols and fatty acids were not found in the literature.

SIMULATION OF THE EXTRACTION PROCESS

A general supercritical extraction process was studied to extract the tocopherols from the synthetic mixture composed of fatty acids, squalene and tocopherols using supercritical carbon dioxide. The process presented in Figure 2 was proposed by Sievers (1996) and it consists of an extractor (E), separator (S), heat exchangers (W) and compressors (C). The top stream of the separator, composed mainly of carbon dioxide, is recycled to the extractor. The bottom stream of the separator is mainly consisting of tocopherol, but it still has small quantities of the other compounds and carbon dioxide.

The simulations of the process were carried out using ASPEN+ (1990) simulator with two different characterizations of the feed. Initially, the feed was represented by a binary system (a-tocopherol and squalene) and then by a ternary system (a-tocopherol, fatty acid and squalene) in different proportions. The composition of the ternary mixture was 70% for the fatty acids, 20% for the squalene and 10% for the a-tocopherol. The solvent flow was 15000kg/h and the feed flow was 1000kg/h, according to the solvent/feed ratio proposed by Brunner et al. (1991).

The simulations of the extraction of tocopherol were conducted at operational conditions that varied from 35 to 90⁰C and from 130 to 250 bar. These operational conditions were similar to those encountered in the literature.

LCVM THERMODYNAMIC MODEL

The thermodynamic model chosen to study the equilibrium between the main compounds and the supercritical solvent was the model that use the Peng-Robinson equation of state with the LCVM mixing rule with the UNIFAC method (PR-LCVM).

The equation of Peng-Robinson used to calculate the vapor-liquid equilibrium and solubilities is expressed as:

(1)

The LCVM mixing rule, proposed by Boukouvalas et al. (1994), was created by the linear combination of the Huron and Vidal (1979) (P®¥ ) and MHV1 (1990) (P®0) mixing rules. It is expressed in terms of a as:

(2)

where a_v is given by Huron and Vidal, a_M by MHV1 and l is the parameter that ranges from zero to one weighing the contribution of Huron and Vidal model and MHV1. The value of l estimated and used by Boukouvalas et al. (1994) was 0,36. The expression of a is represented by:

(3)

where AV and AM, for the Peng-Robinson equation of state, are 0,623 and –0,52 respectively. The parameter "b" is a linear mixing rule.

The vapor-liquid equilibrium is calculated using the isofugacity expression, at constant values of temperature and pressure

(4)

where x and y are the liquid and vapour compositions and f is the fugacity coefficient for the liquid and vapour phases.

The solubilities calculations with the LCVM thermodynamic model were made considering the solid-vapor equilibrium. It was made adopting a pure solid in a supercritical fluid and the solubility expression was expressed as:

(5)

where the exponential term is known as by the Poynting factor, V_i is the molar volume of the solid, P is the absolute pressure, Pi^sat is the saturation pressure, f^sat is the fugacity coefficient at saturation condition and y₂ is the solubility of the solute.

RESULTS AND DISCUSSION

The supercritical extraction of tocopherol from the deodorizer distillate was simulated using a general process proposed by Sievers (1996). The operating conditions varied from 40 to 90⁰C and from 90 to 150 bar. The results of the simulations are presented in Tables 2 and 3 for the binary and ternary systems, respectively.

The results are discussed based on the efficiency, fold and on the concentration factor. The efficiency of the process was defined as the ratio of the final mass and the initial mass of tocopherol. The fold of the process is the ratio between the final and initial masses. The concentration factors were defined as the distribution coefficients to permit the evaluation of the separation between the compounds. FC1 and FC2 are the concentration factors for a-tocopherol/squalene and a-tocopherol/fatty acid, respectively.

It was observed that the separation between tocopherol and squalene in very difficult confirming the conclusions of Brunner (1994). The indication is based on the lower values of the concentration factors. In the ternary mixture, the behaviour has changed. The separation betweeen tocopherol and squalene occurred at high values of the concentration factor but the separation of the fatty acids from the synthetic mixture did not occurr. The efficiency and the concentration factors presented opposite behaviors. The best operating region for the extraction is the one that satisfies both parameters. The separation process containing an extractor and a separator is not the best experimental set-up to concentrate tocopherol; on the other hand, large amounts of squalene can be obtained.

To understand the best region of operation to concentrate tocopherol and to understand the equilibrium behavior between the compounds, the equilibrium data of the main components was studied using the LCVM thermodynamic model. The equilibrium data was predicted through the thermodynamic model and compared to the experimental data. This study was very interesting to know the behaviour of the equilibrium due to the scarcity of experimental data.

This study was divided into two types of equilibrium calculations, one involving vapor-liquid equilibrium and the other involving solubilities of solids in supercritical carbon dioxide.

The critical properties and interaction parameters of natural compounds, squalene and a-tocopherol, can not be measured experimentally because they are thermally unstable. Particularly, many of the compounds related to supercritical fluid extraction decompose when heated before Tc is reached (Melo et al., 1996).

Due to these difficulties, the critical properties necessary for the calculation of vapor-liquid equilibrium data, were predicted by group contribution methods. The methodology used in this work was proposed by Melo et al. (1996).

The experimental vapor-liquid equilibrium data for the binary system CO₂-squalene is presented by Catchpole et al. (1997) from 313 to 333K. Figures 3, 4 and 5 show the calculated and experimental vapor-liquid data for squalene in supercritical CO₂. The figures present the vapor-liquid equilibrium curves, where X1 and Y1 are the liquid and vapor fractions of the carbon dioxide, respectively. The pressure used in the calculation of the equilibrium is in the bar unit.

The behaviour of the LCVM predictive model agrees with the experimental data, with deviations from 26.6 to 24.5% in relation to the pressure and 0.14, 0.19 and 0.15% in relation to the vapour composition for 313, 323 and 333K, respectively.

The solubility data for a-tocopherol in supercritical carbon dioxide is presented by Johannsen et al. (1997) at 40, 60 and 80⁰C and from 100 to 400 bar. The solubilities were calculated by the LCVM thermodynamic model and the results are compared with the experimental solubilities in Figures 6, 7 and 8, that present the behaviour of the solubilities in function of the pressure (bar). Y1 is the solubility in mol fraction of the tocopherol. The model can not represent the experimental data, presenting an inverse behaviour. This can be due to the critical properties calculation and sublimation pressure of a-tocopherol, which were predicted by group contribution methods. Another error source was the group contributions encountered in UNIFAC method that can not efficiently represent the a-tocopherol structural molecule.

CONCLUDING REMARKS

The concentration of tocopherol from a synthetic mixture of tocopherol, fatty acid and squalene was simulated using ASPEN+, with supercritical carbon dioxide. The operational conditions varied from 40 to 90⁰C and from 90 to 150 bar. It was observed that the separation between tocopherol and squalene is very difficult, considering the feed as a binary mixture; on the other hand, this separation occurred at elevated concentration ratios with the presence of the fatty acids. The efficiency and the concentration factors present an inverse behaviour. The efficiency of the supercritical extraction increases when the temperature also increases.

The vapor-liquid equilibrium and solubility data of the main components were also studied using the LCVM thermodynamic predictive model. The equilibrium data for tocopherol and squalene with supercritical CO₂ are scarce. The behaviour of the thermodynamic model varied according to the components. It represents very well the vapor-liquid equilibrium data for the binary system CO₂-squalene, but it was not in agreement with the experimental solubilities of the a-tocopherol. This problably occurred due to the critical properties calculation and sublimation pressure of the compound and because its structural molecule that can not be well represented with the UNIFAC method.

This study will continue in order to extract tocopherols from the deodorizer distillated of soybean oil according to the results obtained with the simulation. The critical properties of the compounds will be studied in order to identify the sensibility of the model. The LCVM thermodynamic model will continue to be studied to predict new equilibrium data that can not be encountered in the literature (Mendes et al., 2000).

ACKNOWLEDGEMENTS

To CNPq for the financial support and to Escola de Química for the use of ASPEN+.

NOMENCLATURE

REFERENCES

Publication Dates

History