Incorporating long-range electrostatics in neural network potentials via variational charge equilibration from shortsighted ingredients

Yusuf Shaidu,Franco Pellegrini,Emine Küçükbenli,Ruggero Lot,Stefano de Gironcoli

DOI: https://doi.org/10.1038/s41524-024-01225-6

IF: 12.256

2024-03-05

npj Computational Materials

Abstract:We present a new approach to construct machine-learned interatomic potentials including long-range electrostatic interactions based on a charge equilibration scheme. This new approach can accurately describe the potential energy surface of systems with ionic and covalent interactions as well as systems with multiple charge states. Moreover, it can either be regressed against known atomic charge decompositions or trained without charge targets, without compromising the accuracy of energy and forces. We benchmark our approach against other state-of-the-art models and prove it to have equivalent performances on a set of simple reference systems while being less computationally expensive. Finally, we demonstrate the accuracy of our approach on complex systems: solid and liquid state sodium chloride. We attain accuracy in energy and forces better than the model based on local descriptors and show that our electrostatic approach can capture the density functional theory tail of the potential energy surface of the isolated Na-Cl dimer, which the local descriptor-based model fails to describe.

materials science, multidisciplinary,chemistry, physical

What problem does this paper attempt to address?

The paper aims to address the issue of incorporating long-range electrostatic interactions in neural network potential functions. Specifically, the authors propose a new method to construct machine learning-derived interatomic potentials through a charge equilibration scheme. This method can accurately describe potential energy surfaces of systems containing ionic and covalent bonds, as well as systems with multiple charge states. Moreover, the method can be trained either with regression on known atomic charge decompositions or without relying on charge targets, without compromising the accuracy of energy and forces.

The method has been validated against other state-of-the-art models, showing comparable performance on some simple reference systems but with lower computational costs. Finally, the authors demonstrate the accuracy of the method in complex systems (such as solid and liquid sodium chloride), proving its superior precision in energy and forces compared to models based on local descriptors. Additionally, it can capture the tail of the density functional theory potential energy surface of an isolated Na-Cl dimer, which local descriptor models fail to describe.