Supercritical CO2 extraction of pelletized oilseeds : Representation using a linear driving force model with a nonlinear sorption isotherm

Abstract

This work develops a general, predictive model for the SuperCritical (SC) CO2 extraction of pelletized oilseeds. Cumulative SC-CO2 extraction curves of small (S, dp = 2.5 mm) and large (L, dp = 4.0 mm) cylindrical evening primrose seed pellets were modeled simultaneously using a Linear Driving Force (LDF) approximation modified to consider the nonlinear oil partition between CO2 and the substrate and extrapolated using the actual solubility of oil in CO2 at extraction conditions (40 °C and 45 MPa). Model best-fitting parameter was a single interparticle porosity (εp) of pellets considering equilibrium conditions after static extraction, and mass transfer kinetics during dynamic extraction. The adaptation of general model to bisized mixtures containing 25–75% S-pellets, predicted experimental results better than a monosized version (same Sauter mean diameters) because of the negative impact on overall cumulative extraction curves of slow-extracting L-pellets. Finally, a simplified version of the general model was developed that neglected the intraparticle porosity (εp = 0) and used a microstructural factor () in the general model as the best-fitting parameter. This simplified model saved 33% of computational time at the expense of overestimating extraction rates, particularly at the beginning of the diffusion-controlled period, which diminished as εp decreased and FM increased.