Machine-learning-driven modelling of amorphous and polycrystalline BaZrS$_{3}$

TL;DR

This paper demonstrates how machine learning can accurately model the atomic structure of BaZrS$_{3}$ in both amorphous and polycrystalline forms, aiding in the development of photovoltaic materials.

Contribution

The study introduces a bespoke machine-learned interatomic potential for BaZrS$_{3}$, enabling detailed atomic-scale modeling of its phases and grain boundaries, advancing materials simulation techniques.

Findings

Successfully modeled amorphous BaZrS$_{3}$ structure

Quantified grain-boundary formation energies

Created realistic polycrystalline models for comparison

Abstract

The chalcogenide perovskite material BaZrS$_{3}$ is of growing interest for emerging thin-film photovoltaics. Here we show how machine-learning-driven modelling can be used to describe the material's amorphous precursor as well as polycrystalline structures with complex grain boundaries. Using a bespoke machine-learned interatomic potential (MLIP) model for BaZrS$_{3}$, we study the atomic-scale structure of the amorphous phase, quantify grain-boundary formation energies, and create realistic-scale polycrystalline structural models which can be compared to experimental data. Beyond BaZrS$_{3}$, our work exemplifies the increasingly central role of MLIPs in materials chemistry and marks a step towards realistic device-scale simulations of materials that are gaining momentum in the fields of photovoltaics and photocatalysis.