Modeling gas transport in organic molecular sieve membranes

Abstract

Organic molecular sieves are promising materials for gas separation membranes due to their versatile geometric and chemical tunability. Understanding molecular transport mechanisms within these materials is crucial for identifying optimal candidates from extensive libraries. Advances in computational technology and data science have introduced molecular simulation and machine learning (ML) as powerful tools to deliver atomistic insights and enable high-throughput screening. This review emphasizes modeling gas transport through the adsorption-diffusion framework, which better represents intrinsic molecular behavior in confined environments. We address its derivation, implementation, validation, and comparison with the solution-diffusion model. Computational methods, including molecular dynamics, Monte Carlo, density functional theory, and ML, are discussed regarding their application in determining membrane properties. Challenges and limitations in aligning computational predictions with experimental data are critically analyzed. This review underscores the importance of methodological consistency, transparency, and data availability for leveraging opportunities in artificial intelligence and advancing membrane separation technologies.