Dynamic Monte Carlo simulations were used to examine binary self-diffusion on heterogeneous cubic lattices and in ZSM-5 (MFI), which included both strong and weak adsorption sites due to the presence of Al atoms in addition to Si. It was observed that the diffusion behavior in such zeolites strongly depended upon the pore network topology, the fraction of strong adsorption sites and the differing relative adsorption strengths of the diffusing species on different types of sites. The Maxwell-Stefan approach previously proposed to predict diffusivities in silicalite-1, the all-Si version of MFI, was re-examined and generalized so as to include the effect of strong adsorption sites. It was concluded that this approach was qualitatively and quantitatively inaccurate when correlation effects were large; as a result of zeolite heterogeneity. An alternative theory based upon an effective medium approximation and other mean-field arguments was more successful for the case where differences in relative adsorption strengths of the species were very high. This theory could be further refined to include intermolecular correlations.

Dynamic Monte-Carlo Simulations of Binary Self-Diffusion in Zeolites: Effect of Strong Adsorption Sites. V.Iyengar, M.O.Coppens: Chemical Engineering Science, 2004, 59[22-23], 4747-53