Queryable Gaseous Adsorption Properties of Pure Components and Mixtures in Metal-Organic Frameworks

Abstract

Grand canonical Monte Carlo (GCMC) has been the most common simulation method for high-throughput screening of metal-organic frameworks (MOFs) in adsorption applications. However, GCMC results are inevitably limited to specific pressure and temperature conditions, limiting their applicability across the varying conditions in industrial operations. To address this, we develop a macrostate-probability-distribution-based (MPD-based) adsorption database for several key gas molecules (i.e., CO₂, N₂, CH₄, and CO) in approximately 2900 rigid MOF structures. Crucially, MPD computed from flat histogram Monte Carlo can be analytically reweighted to any pressures and temperatures, enabling instant access to adsorption uptakes under any desired conditions. By integrating the ideal adsorbed solution theory (IAST), the database also provides mixture isotherms at user-defined compositions to allow efficient assessment of multicomponent separations. With the MPD-based database, we demonstrate its capability in rapid screening and uncovering structure-performance relationships that govern separation behavior under different operational conditions. Overall, by employing MPD, we present a versatile and interactive adsorption database to facilitate the future discovery of optimal MOFs.