A General Nonbonded\nForce Field Based on Accurate\nQuantum Mechanics Calculations for Elements H–La and Hf–Rn

Noncovalent interactions (NCI) play a central role in\nnumerous\nphysical, chemical, and biological phenomena. An accurate description\nof NCI is the key to success for any theoretical study in such areas.\nAlthough quantum mechanics (QM) methods such as dispersion-corrected\ndensity functional theory are sufficiently accurate, their applications\nare practical only for <300 atoms and <100 ps of simulation\ntime. Thus, empirical force fields (FF) have generally been the only\nchoice for systems with thousands to millions of atoms and for nanoseconds\nand longer. We want to develop a FF that can be applied to applications\nof thousands to millions of atoms with an accuracy comparable to QM\nmethods. As the first step, we develop here a new general nonbonded\npotential (GNB) based on a novel functional form with four adjustable\nparameters for each element. We report here parameters for elements\nH–La, Hf–Rn (excluding lanthanides and actinides) by\nfitting the interaction energy of molecular complexes to QM calculations\nusing the accurate Head-Gordon ωB97M-V density functional. We\nperformed extensive testing of GNB for organic molecules, organometallic\nmolecules, and metal organic–framework (MOF) systems. The mean\nabsolute errors of GNB are 0.37 kcal/mol for the dispersion and mixed\ngroups of the S66 × 8 benchmark set, 0.35 kcal/mol for CO<sub>2</sub> adsorption on MOF materials, and 4.53 kcal/mol for the XTMC43\nbenchmark. GNB outperforms existing FF and in many cases has accuracy\ncomparable to that of QM methods such as PBE-D3. GNB can potentially\nreplace the nonbonded part of existing FFs in a wide range of applications.