Skip to main content
ARS Home » Research » Publications at this Location » Publication #166452

Title: LIQUID CO2 EXTRACTION OF HEXANES FROM SOYBEAN OIL

Author
item Taylor, Scott
item Eller, Fred

Submitted to: Intl Symposium on Supercritical Fluid Chromatography and Extraction
Publication Type: Abstract Only
Publication Acceptance Date: 5/12/2004
Publication Date: 8/4/2004
Citation: Taylor, S.L., Eller, F.J. 2004. Liquid co2 extraction of hexanes from soybean oil. Intl Symposium on Supercritical Fluid Chromatography and Extraction.

Interpretive Summary:

Technical Abstract: Soybean oil is currently extracted on an industrial scale by hexane and the largest operating cost in this process is the separation of hexane from the extracted oil. The removal of hexane from the oil is critical because of hexane's demonstrated toxic effects and hexane is listed as a hazardous air pollutant by the U.S. Environmental Protection Agency. Recently, it has been demonstrated that SC-CO2 (1740 psi, 40-60 deg C) could be used to remove hexane from a soybean oil (SBO) and hexane mixture using a countercurrent packed tower. Residual hexane levels as low as 20 ppm were obtained. With conventional methods, the concentration of hexane in SBO after the first stripper is typically 1000 ppm, with a minimum level of 500 ppm being attainable. The study also indicated residual hexane decreased as the extraction temperature decreased, and thus residual hexane was inversely proportional to CO2 density. Because CO2 density is significantly higher at 25 deg C than 40 deg C, it seems that residual hexane levels could be reduced further using liquid CO2 L-CO2). Hexane has a high solubility in L-CO2 while triglyceride (e.g., SBO) solubility in L-CO2 is very low. In the present study, L-CO2 was used to remove n-hexane and isohexane from solvent/soybean oil mixtures. A fractionation column was constructed that included a precooling section and four separate zones and had a total height of 253.2 cm. It was packed with a protruded stainless-steel packing, which provided a 94% void volume. The solvent/SBO mixtures were introduced into the column above the first zone by a liquid metering pump. L-CO2, at 25 deg C and 1350 psi, was pumped through the solvent/SBO mixture at a flow rate of 3-4 L/min (expanded gas). After passing through the solvent/SBO mixture, the CO2 stream was directed through the injector of a supercritical fluid chromatograph (SFC). At timed intervals, the SFC program was initiated and a sample was injected onto a capillary column and analyzed with a flame ionization detector. Relative peak areas of the hexanes were compared to determine the residual hexane levels. The extractions proceeded for set volumes of expanded CO2 to determine the amount required to effectively remove the hexanes from the mixtures. The SFC program was optimized to analyze the hexane portion of the sample. This technique provided real time on-line monitoring of the extraction process. The SFC data were compared to the residual hexane that remained in the extracted SBO. ISO Method 9832:2002 was used to determine the residual hexane levels. n-Hexane and isohexane constituted 25 wt% of their respective feed stocks. Previous research had shown that the passage of 300 L of CO2 through a 25 wt% solution of hexane in soybean oil yielded 38 ppm of residual hexane in the raffinate soybean oil. This is well below the value of 500 to 1000 ppm obtained by the first stage stripper during the conventional extraction methodology. The current research yielded average values of residual solvent in the soybean oil of 3.84 and 3.28 ppm for n-hexane and isohexane, respectively. These data were determined by headspace GC analysis. SFC analyses yielded average values of 0.05% and 0.04% for n-hexane and isohexane, respectively, in the CO2 stream. Calibration curves for n-hexane and isohexane in CO2 were determined and gave R2 values of 0.972 and 0.988, respectively. Correlating the SFC data of solvent in the CO2 stream to the residual hexane in SBO, it can be seen that solvent removal is complete when the solvent/CO2 percentage is below 0.1%. This is applicable for both n-hexane and isohexane in soybean oil mixtures. This SFC analysis technique can be used for real time monitoring of solvent removal during a continuous feed process. Liquid CO2 readily removes n-hexane and isohexane from soybean oil/solvent mixtures. Both are reduced from 250