PHAST-MBD: Implementing Many-Body Dispersion in the PHAST 2.0 Potential, Results for Noble Gases.

A recently published empirical force field (herein PHAST or PHAST 2.0) is employed in its many-body dispersion-corrected form (PHAST-MBD) to examine the effects of collective dispersion interactions. Rare gases are used as a systematic way to test increasing importance of van der Waals attractions in systems dominated by repulsion-dispersion that are a challenge to extant force fields. The effects of many-body dispersion were studied for liquid and supercritical fluid regime for the series Neon, Argon, Krypton and Xenon. The PHAST force field is a condensed phase atomistic molecular modeling potential that includes contributions from repulsion-dispersion, permanent electrostatics, and many-body polarization. Each of these pieces is physics based and seeks to mimic their constituent first-principles counterparts with as few fitting parameters as possible. Critically, it is built to reproduce accurate gas phase pair interactions. This facilitates the efficacy of mixing rules for unlike interactions while many-body effects are added via explicit polarization and dispersion models. The effectiveness of PHAST-MBD is demonstrated calculating rare gas densities as compared to experiment over a wide pressure range. Pair potentials fail systematically at high pressure and density as dispersion grows while PHAST-MBD reproduces experiment in all regimes. This is strong evidence in favor of the PHAST 2.0 paradigm of physically motivated empirical potentials that reproduce gas phase interactions and facilitate accurate mixing rules with many-body effects included explicitly. This work suggests a hybrid future approach that will be adopted in PHAST-MBD that keeps the accurate PHAST pair interactions and only includes many-body terms via the coupled dipole method (CDM); such an approach avoids the issues identified here that the CDM many body van der Waals (MBVDWs) formalism has reasonable but nonoptimal implicit mixing rules and can alter pair potentials.