Rapid Exploration of Mixture Adsorption via Adiabatic Sampling.

The efficient computational screening of large libraries of porous materials for gas storage applications relies on the ability of simulations to rapidly predict the phase diagrams of confined fluids, their adsorption properties, and the selectivity of adsorbents in separation applications. While isothermal grand-canonical simulations are commonly used to this end, their convergence rate can dramatically decrease when the thermodynamic conditions approach coexistence, giving rise to metastability and the onset of large free-energy barriers. To address this challenge, we develop an adiabatic counterpart for the simulation of mixtures, derive the acceptance rules for a grand-isochoric adiabatic Monte Carlo method, and assess its accuracy and efficiency through comparisons with isothermal grand-canonical simulations of Ar-Kr mixtures, both in the bulk and when adsorbed in an MCM-41 porous material. The results show that adiabatic simulations allow for large temperature variations, thereby resulting in the rapid sampling of the configuration space and faster convergence than the grand-canonical simulations by up to 2 orders of magnitude.