We demonstrate that classical density functional theory (DFT) based on the PC-SAFT equation of state is a fast, accurate, and predictive model to predict multicomponent adsorption in porous materials, which is an essential step toward the design of next-generation adsorbents for relevant applications. Using GPU acceleration, adsorption isotherms and adsorption enthalpies can be obtained in a matter of seconds, which is several orders of magnitude faster than grand canonical Monte Carlo (GCMC) simulations. Using metal-organic frameworks as adsorbents and non- or weakly polar molecules as adsorbates, we validate our approach by performing GCMC simulations for binary, ternary, and quaternary mixtures with practically relevant applications, such as noble gas separations (Kr/Xe, Ar/Kr/Xe), direct dry air capture (CO2/N2), hydrogen enrichment (CH4/H2, CH4/H2/N2) and adsorbed natural gas (CH4/C3H8, CH4/C2H6/C3H8, CH4/C2H6/C3H8/N2). Classical DFT reproduces loadings and adsorption enthalpies of the mixtures in close agreement with results from GCMC simulations. Thus, classical DFT expands our toolbox for studying multicomponent adsorption.
Thiele et al. (Thu,) studied this question.