In recent years, supercritical fluid extraction using carbon dioxide (SC-CO2) has been intensively investigated to some traditional separation techniques, such as vacuum distillation or organic solvent extraction, as an alternative.

Supercritical carbon dioxide extraction is a process where carbon dioxide passes through a mixture of interest at a certain temperature and pressure until it reaches an extractor. This process is used because supercritical carbon dioxide has a low viscosity, a high diffusivity and a low surface tension that provides selective extraction, fractionation and purification, allowing its penetration in micro — and macro-porous materials (Dumont and Narine, 2007).

The major advantage of this method is the easy post-reaction separation of the components by depressurization. Another advantage is the low temperatures used for the majority of the experimentations because carbon dioxide has a critical temperature of 31°C. However, the use of high pressure conditions makes the system energetically expensive but can be economically viable at a rate of production superior to 25% using conditions of approximately 90 atm and 40°C (Mendes et al., 2002). At these specific conditions, only fatty acids are separated from tocopherol (Mendes et al., 2005).

The phase equilibrium data can provide fundamental and necessary information for designing a SC-CO2 separation process. A number of studies are available for this purpose (Stoldt and Brunner, 1998; Stoldt and Brunner, 1999; Chia-Cheng et al., 2000; Mojca et al., 2003; Pereira et al., 2004).

Simulation and thermodynamic modeling of the supercritical fluid extraction was reported by different authors (Vazquez et al., 2007; Fornari et al., 2009; Martinez-Correa et al., 2010).

Vazquez et al. (2007) described a process for the purification of squalene by using CC-SCCO2, a by-product obtained after the distillation and ethylation of olive oil deodorizer distillate (OODD), as raw material. The Group Contribution Equation of State was employed to simulate the separation process and to design the experimental extractions. As satisfactory agreement was found between the experimental and the calculated yields and phase compositions, a raffinate with a squalene concentration of up to 90% was obtained. Finally, the thermodynamic model was employed to develop optimal process conditions to enhance squalene recovery, including partial reflux of the extract product and recirculation of the supercritical solvent in a continuous countercurrent extraction column.

Several authors have studied the concentration of tocopherols directly from the DD, without carrying out any modification pretreatment of the raw material, namely the separation of tocopherols from FFA (Lee et al., 1991; King and Dunford, 2002).

However, the application of pretreatment like esterification leads to two advantageous results for the continuous process, one is that methyl esterified DD (ME-DD) has a higher solubility in SC-CO2 than DD. The other is that the viscosity is greatly reduced after removing most of the sterols.

The chemical modification of the DD combined with SC-CO2 has been reported by different authors (Bondioli et al., 1993; Nagesha et al., 2003; Liu et al., 2006; Vazquez et al., 2006; Fang et al., 2007; Vazquez et al., 2007; Torres et al., 2009). In this case, esterification and methanolysis of the DD produced a mixture containing tocopherols, phytosterol esters and FAME, with the process goal of the SC-CO2 process being the elimination of FAME to concentrate tocopherols and sterol esters in the raffinate.

Lee et al. (1991) studied the feasibility of tocopherols concentration from SODD by SC-CO2 at different temperatures and pressures. It was observed that by increasing the CO2 pressure, the SODD solubility also increases for all the studied temperature (45°C, 55°C and 70°C) and that esterified SODD has four to six times higher solubility in SC-CO2 than the sterol-removed SODD. The results showed that SODD initially containing about 13-14% tocopherols may require a countercurrent multistage column to be efficiently concentrated.

King and Dunford (2002) described a solid fluid fractionation method to recover sterol-enriched triglyceride fractions from vegetable oil DD (rice bran and soybean oil DD) using a pilot scale high pressure packed column.

It was possible to obtain oil fractions containing 20-31% sterols and 30-38% TAG, respectively. The method consists of two extraction steps, one carried out at 14 MPa and 45°C and the second extraction was performed at 20 MPa and 80°C. The described method does not leave any solvent or chemical residues in the final product, nor generates additional waste streams requiring subsequent disposal. However, another purification step should be applied in order to obtain a high-purity sterol fraction.

Bondioli et al. (1993) described a process to recover squalene from OODD after transformation of the FFA into TAG in order to increase the separation efficiency. OODD was converted into FFA by saponification and splitting. The mixture was further dried and esterified with glycerol in the presence of an acid catalyst into the corresponding TAG, the latter ones being extracted with SC-CO2. The process was carried out on a pilot-plant scale with a column operating in the countercurrent mode. Using this process, squalene was recovered in high purity and yields of about 90%.

Nagesha et al. (2003) described a process where SC-CO2 extraction of chemically modified SODD was studied at three levels of pressure (180-300 bar) and temperature (40-60°C) to optimize the conditions for enrichment of tocopherols in the raffinate. The modification process includes esterification, saponification, acid hydrolysis and cold crystallization to remove sterols, and again esterification of the FFA obtained from acid hydrolysis of the triglycerides (Fig. 22.8). After modification, SODD containing about 90% of FAME showed improved solubility in SC-CO2 and a better extraction rate. Since FAMEs are more volatile, they were extracted preferentially over tocopherols and other high molecular weight compounds. The extraction at higher pressures and temperatures resulted in a better yield of FAME along with tocopherols and this in turn decreased the degree of enrichment of tocopherols in the raffinate. However, a specific level of pressure and temperature of the extraction caused the increase in the solubility of FAME due to their volatility and results in the enhanced enrichment of tocopherols in the raffinate. It was observed that the enrichment of tocopherols (36%) to ten times the original concentration of the feed (4%) occurred at an extraction pressure of 180 bar and a temperature of 60°C.

The recovery of tocopherols and sterols from sunflower oil deodorizer distillates (SfODD) using countercurrent supercritical carbon dioxide extraction (CC-SCCO2) has been studied by Vazquez et al. (2006). The chemical transformation of the SfODD composition significantly enhances the concentration of minor lipids in the raffinate product. This pretreatment resulted in a mixture (ethylated SODD) which mainly consists of tocopherols, sterols and fatty acid ethyl ester (FAEE). Additionally, the reaction step produced a solid phase, mainly consisting of sterols, which was isolated from the liquid product.

After two consecutive extractions with hexane, the sterol purity in the new solid phase increased up to ca. 88%, which corresponds to 18% of recovery of the total sterols present in the original SODD. A similar procedure was accomplished replacing hexane by ethanol. In this case, the purity of the sterols obtained was similar, although the recovery was reduced to ca. 10%. This low value of recovery indicates a higher solubility of the sterol solid phase in ethanol compared to that in hexane.

The main drawback of the CC-SFE process described in the present study is related to the high amount of unidentified compounds present in the original SODD (20%). Around 50% of this unidentified material corresponds to non­volatile compounds which preferably accumulated in the raffinate.

CC-SCCO2 extractions of the ethylated and original SODD resulted in a 3.7- fold increase in the tocopherol + phytosterol concentration (ca. 80% recovery)

with the ethylated material, while only a 1.3-fold increase was obtained with the original SODD.

Additionally, during the formation of FAEE, partial crystallization of free sterols occurs, and around 20% of the sterols present in the original SODD can be recovered with high purity (88%) in the solid phase.

Liu et al. (2006) studied the vapor-liquid phase equilibrium data for SC-CO2 and methyl esterified DD (ME-DD) at 40°C and in the pressure range of 9.7-16.2 MPa in order to determine the feasibility of SC-CO2 to concentrate natural tocopherols from SODD. The results showed that the separation factor between tocopherols and FAME was from 2.5 to 3.8 at 40°C and 9.7-16.2 MPa, which is fundamental and necessary for future process designs. For this purpose a modification process of DD was applied that includes esterification, cold crystallization for removing sterols and alcoholysis. The FFAs obtained from TAGs by alcoholysis were further esterified to FAMEs. The detailed procedure is shown in Fig. 22.9. Through such pretreatment, the obtained methyl esters (ME-DD) contained 52% FAME and 8% of tocopherols and other compounds. After the reactions, most of the sterols were easily removed because of their low solubilities in FAMEs below 4°C.

22.9 Oil deodorizer distillates (DD) modification process (from Liu et al., 2006).

Fang et al. (2007) described a process where SC-CO2 fractionation was employed to concentrate tocopherols from ME-DD. The initial pressure, feed location, temperature gradient and ratio of CO2 to ME-DD were optimized for separating FAMEs. For the following tocopherol concentration step, a final pressure of 20 MPa resulted in the greatest average tocopherol content (>50%) and tocopherol recovery (about 80%).

ME-DD was prepared from DD through the pretreatment process that include two steps of esterification and methanolysis, which converted FFA and glycerides into FAMEs. The two reactions were conducted with the catalysts of sulfuric acid and sodium methoxide, respectively. After each reaction, the mixture was washed until neutral. Finally, the mixture was stored in a refrigerator for 12 h. As a result most of the sterols were crystallized and removed by filtering under reduced pressure. A fractionation column was required for the ME-DD separation. Low pressure (the initial pressure) was used in combination with a temperature gradient along the column to separate the FAMEs. Then, the pressure was increased to separate the tocopherols from other impurities.

Torres et al. (2009) reported a two-step enzymatic reaction to obtain phytosterol esters, where the SODD was initially modified by the addition of oleic acid in order to decrease the DD melting point. After esterification steps, the product obtained comprised mainly FAEEs, tocopherols and phytosterol esters, together with minor amounts of squalene, FFAs, free sterols and triacylglycerols. The FAEEs were eliminated by SC-CO2 and the phytosterol esters and tocopherols were concentrated in the raffinate. The separation between the last two compounds was carried out in an isothermal countercurrent column (without reflux), with pressures ranging from 200 bar to 280 bar, temperatures of 45-55°C and solvent- to-feed ratios from 15 kg/kg to 35 kg/kg. Using these extraction conditions, the fatty acid esters were completely extracted. The phytosterol esters were concentrated in the raffinate up to 82% with a yield of 72%.