Evaluating adsorbed-phase activity coefficient models using a 2D-lattice model

Abstract

Despite the wide use of the real adsorbed solution theory to predict multicomponent adsorption equilibrium, the models used for the adsorbed phase activity coefficients are usually borrowed from the gas -liquid phase equilibria. In this work, the accuracy of the Wilson and NRTL models for evaluating adsorbed phase activity coefficients is tested using a 2D-lattice model. An accurate model for adsorbed-phase activity coefficients should have no problem in fitting adsorption data obtained using this simple lattice model. The results, however, show that the commonly used Wilson and NRTL models cannot describe the adsorbed phase activity coefficients for slightly non-ideal to strong non-ideal mixtures. Therefore, until new models for adsorbed phase activity coefficients are developed, we should use existing models for liquids with care. In the second part of this work, the use of Monte Carlo simulations on a segregated 2D-lattice model, for predicting adsorption of mixtures is investigated. The segregated model assumes that the competition for adsorption occurs at isolated adsorption sites, and that the molecules from each adsorption site interact with the bulk phase independently. Two binary mixtures in two adsorbent materials were used as case studies for testing the predictions of the segregated 2D-lattice model: the binary system CO 2 -N 2 in the hypothetical pure silica zeolite PCOD8200029, with isolated adsorption sites and normal preference for adsorption, and the binary system CO 2 -C 3 H 8 in pure silica mordenite (MOR), with isolated adsorption sites and inverse site preference. The segregated 2D-lattice model provides accurate predictions for the system CO 2 -N 2 in PCOD8200029 but fails in predicting the adsorption behaviour of CO 2 -C 3 H 8 in pure silica MOR. The predictions of the segregated ideal adsorbed solution theory model are superior to those of the 2D-lattice model.