Monte Carlo Simulation and Exact Statistical Mechanical Lattice Models as a Development Tool for Zeolite Multi-Component Adsorption Isotherm Derivation

Abstract

Developing a suitable separation and purification procedure for multi-component gas mixtures is a ubiquitous problem. Very little is known experimentally about multi-component adsorption isotherms, yet these are crucially important for designing separation processes. Multi-component adsorption isotherms over a wide range of composition, pressure and temperature are virtually unknown. Adsorption onto zeolites is widely used in adsorptive separation processes due to the unique characteristics of its structural pores which may be “engineered” to permit adsorption to take place on the basis of molecular shape and size. In this article, we discuss the methodological development of rapidly obtaining multi-component adsorption isotherms using the interplay between the essentially exact (in the statistical mechanical sense) lattice model and configurational bias Monte Carlo simulations. The latter technique has the known capability of being able to reproduce experimental isotherms reliably, but can be very time consuming and expensive. In contrast, the lattice model adsorption isotherms can be parameterized by comparing with those obtained by Monte Carlo simulation and thereafter used to cheaply and rapidly explore a very wide range of mixture concentrations and conditions. The interactions required in the lattice model are effective interaction-free energies. This approach has been applied to the adsorption isotherms of selective binary and ternary mixtures of small alkanes (methane, ethane) and carbon dioxide adsorbed in silicalite. The lattice model results are shown to have the potential, when suitably parameterized, to rapidly and cheaply reproduce the main features of the much more expensive Monte Carlo isotherms over a very wide range of mixture compositions, temperatures and pressures.