MODELLING OF SUPERCRITICAL FLUID EXTRACTION

Abstract

Supercritical Fluid Extraction (SFE) was developed for analytical application in the mid 1980's in response to the desire to reduce the use of organic solvents in the environment. It is gentle, clean and green extraction process. SFE systems are being used to extract chemical compounds in industrial purification and recrystallization operations using supercritical fluid instead of an organic solvent because of regulatory and environmental pressures on hydrocarbon and ozone-depleting emissions. The result is an extract with little or no residual solvent, superior purity and yield, and lower operating costs compared to solvent-based systems. SFE-based processes have helped to eliminate the use of hexane and methylene chloride as solvents with supercritical carbon dioxide. With increasing scrutiny of solvent residues in pharmaceuticals, medical products, and neutraceuticals, and with stricter regulations on volatile organic chemical (VOC) and ozone depleting chemical (ODC) emissions, the use of SFE is rapidly proliferating in all industrial sectors. In the present study, the various models available for the modelling of SFE process are discussed. The model based on differential mass balances performed along the extraction bed and the local adsorption equilibrium model are presented. The first model is used when internal mass transfer is the controlling stage for the extraction process. The second model is used to analyze the dynamic behaviour of the extraction process incorporating intraparticle diffusion as well as external mass transfer. This model has also been studied to predict the effects of various pertinent model parameters like effective intraparticle diffusion coefficient and mass transfer coefficients and operating parameters like temperature, pressure, solvent flow rate and particle size on the cumulative extraction yield. The model based on differential mass balances due to Reverchon (1996) is simulated using ode45 toolbox of MATLAB 7.0 as well as FEMLAB 3.1 to predict the normalized extraction yield as a function of time. The internal diffusion coefficient is 111 used as a fitting parameter to fit the experimental data due to Reverchon (1996) and predicted extraction data shows 86% experimental data within +9% error when MATLAB environment was used for program development and within +4% to -5% error when the model is solved using FEMLAB. The comparison showed that the results prediction by FEMLAB is better than MATLAB showing a maximum deviation of 28%. The local adsorption equilibrium model due to Goto et al. (1993) is simulated using ode45 toolbox of MATLAB 7.0 to predict the cumulative extraction yield. The equilibrium adsorption coefficient is used as a fitting parameter to fit the experimental data due to Skerget et al. (2001), Kim et al. (2007), and Tonthubthimthong et al. (2004). The results obtained from the present analysis compared well with the experimental data. The present model is able to predict the experimental data within an error band of +10 % to -2%.