ABSTRACT
Epoxidation of high-linolenic perilla oil was carried out in the presence of solid acidic ion-exchange resin at varying reaction temperatures for 8 h. A pseudo two-phase kinetic model that captures the differences in reactivity of double bonds at various positions in the fatty acid of a triglyceride molecule during both epoxy formation and cleavage was developed. The proposed model is based on the Langmuir-Hinshelwood-Hougen-Watson (L-H-H-W) postulates and considers the adsorption of formic acid on the catalyst as the rate-determining step. To estimate the kinetic rate constants of various reactions, genetic algorithm was used to fit experimentally obtained iodine and epoxy values of epoxidized perilla oil. A re-parametrized form of Arrhenius equation was used in the proposed model to facilitate the precise estimation of parameters with least computational effort. The obtainment of the least error between experimentally determined and theoretically predicted iodine and epoxy values indicates the robustness of the proposed model.
