Machine learning interatomic potential can infer electrical response

TL;DR

This paper demonstrates that machine learning interatomic potentials can be extended to predict electrical responses such as polarization and Born effective charges directly from energy and force data, enabling scalable modeling of electric-field effects.

Contribution

It introduces a method to extract electrical response tensors from MLIPs using the Latent Ewald Summation framework without additional training on charges or polarization.

Findings

Accurately predicts infrared spectra of water under electric fields

Models ionic conductivities in superionic ice

Simulates phase transitions and hysteresis in ferroelectric PbTiO3

Abstract

Modeling the response of material and chemical systems to electric fields remains a longstanding challenge. Machine learning interatomic potentials (MLIPs) offer an efficient and scalable alternative to quantum mechanical methods but do not by themselves incorporate electrical response. Here, we show that polarization and Born effective charge (BEC) tensors can be directly extracted from long-range MLIPs within the Latent Ewald Summation (LES) framework, solely by learning from energy and force data. Using this approach, we predict the infrared spectra of bulk water under zero or finite external electric fields, ionic conductivities of high-pressure superionic ice, and the phase transition and hysteresis in ferroelectric PbTiO$_3$ perovskite. This work thus extends the capability of MLIPs to predict electrical response--without training on charges or polarization or BECs--and enables accurate modeling of electric-field-driven processes in diverse systems at scale.