Multi-scale modeling strategy for separation of alkane mixtures using zeolites

Abstract

Mixtures of alkanes can be separated by selective sorption and diffusion across zeolite membranes. We develop a multi-scale strategy for estimating the permeation fluxes of the components across the membrane relying almost entirely on simulation techniques. This strategy consists of the following steps: (1) Calculating the sorption isotherms of the pure components and the mixtures using Configurational-Bias Monte Carlo (CBMC) techniques, (2) Determination of the Maxwell-Stefan diffusivities from Molecular Dynamics or Transition State Theories, (3) Using Kinetic Monte Carlo (KMC) simulation techniques for studying and verifying the mixture diffusion rules following the Maxwell-Stefan theory, (4) Solving the transient equations of continuity of mass for each species with the Maxwell-Stefan equations describing intra-crystalline diffusion to obtain the transient permeation fluxes. The applicability of our multi-scale modeling strategy is illustrated by means of a specific example for separation of methane (C1) and n-butane (nC4) by permeation across a MFI (silicalite-1) membrane.