Data-Driven Interatomic Potentials for Molten Salts

Data-driven generation of interatomic potentials for molecular dynamics simulations has been shown to bridge the time and length scales of atomistic simulations with the accuracy of ab-initio methods. This scale bridging leads to new insights into material properties currently unavailable to ab-initio methods, and too complex for classical potentials.

In our work, we have developed an interatomic potential for molten NaCl using Gaussian process regression. This potential has been used to perform molecular dynamics simulations and investigate the effects of temperature on the structural and dynamic properties of the melt, at near density functional theory accuracy.

In this session, we discuss the results of this investigation, including the generation of the machine learned potential. We also present an outlook on the application of these machine learning methods to the generation of potentials for more complex ionic liquids.