Alkali halide force fields: Search for versatility

•Constructing semi-empirical force fields for atomic-molecular particles.•Molecular dynamics simulation.•Alkali halide crystals melting temperature.•Gaussian charges and exponential repulsion. In previous papers we were looking for a versatile still simple form of alkali halide force fields. Since the original polarizable version was devised mainly for aqueous solutions, properties of alkali halides as crystals and melts got less attention (P.T Kiss and A. Baranyai, J.Chem.Phys.,141,114501 (2014)). In that paper the non-electrostatic part of alkali halide force fields was formed as a combination of individual repulsion wings, exponentials with two parameters. To obtain correct melting temperature, we left out dipolar polarization as a negligible player in a symmetric environment, and massaged the repulsive potentials without moving considerably away from the original form. However, the results were not satisfactory (A. Baranyai, J. Mol. Liquids, 297, 111,762 (2020)). Then we created a different set preserving only the original Gaussian charges with exponential repulsion. Since we wanted to be as accurate as possible, we could obtain potentials for pairs (A. Baranyai, J. Mol. Liquids, 343, 117,575 (2021)). In the present paper we try to re-establish the pairwise combination of repulsions fulfilling the requirement of accurate energy, density and melting temperature. We applied a simpler combination rule for an exponential function, square root both of the pre-exponentials and the arguments. The repulsion of cations was fixed for all combinations. The size difference of cations did not allow covering the entire set with identical anion parameter values. We could create two separate groups were the anion repulsion was also constant. The sodium–potassium and the rubidium-cesium pairs used the same anion parameters. Exceptions are fluorides, we had to use individual anion parameters for K+, Rb+, and Cs+. For lithium the fluoride used the same parameters as in the case of sodium. For other anions of Li+, however, they had to use different anion parameters from that of Na+. While fluoride deficiency can be attributed to the inadequacy of the form of functions applied, problems of the lithium halides, namely, the nonadditive character of their radii for large anions, is physical. We describe the fitting process and present the results.