Integrating Macrostate Probability Distributions with Swing Adsorption Modeling for Binary/Ternary Gas Separation

TL;DR

This paper presents a new modeling framework that combines macrostate probability distributions with process optimization to accurately and efficiently predict multicomponent gas adsorption equilibria, aiding material discovery.

Contribution

It introduces a novel integration of macrostate probability distributions from Monte Carlo simulations with cyclic process modeling for improved gas separation predictions.

Findings

High-fidelity mixture adsorption predictions without repeated simulations.

Efficient evaluation of binary and ternary gas separations.

Establishment of MPD-based modeling as a general predictive method.

Abstract

Accurate and efficient prediction of multicomponent adsorption equilibria across pressures, temperatures, and compositions remain a central challenge for designing energy-efficient adsorption-based separation processes. Traditional approaches, including model fitting and ideal adsorbed solution theory (IAST), often fail to balance accuracy, computational efficiency, and transferability under process-relevant conditions. Here, we introduce a material-to-process modeling framework that integrates macrostate probability distributions (MPDs) from flat-histogram Monte Carlo simulations with rigorous cyclic process optimization. MPDs directly capture the joint occupancy distributions of adsorbates, producing reweightable landscape that enables high-fidelity mixture adsorption equilibria without repeated simulations or model assumptions. We show that coupling this statistical mechanical foundation with process modeling delivers accurate and computationally efficient evaluations for binary and ternary gas mixture separations. This integration establishes MPD-based modeling as a generalized method for predictive multicomponent adsorption equilibria, accelerating the discovery and design of adsorbent materials for carbon capture and other separation challenges.