Machine learning Landau free energy potentials

TL;DR

This paper introduces a machine-learning method to construct Landau-like free energy models for ferroelectric materials, demonstrating accurate predictions of PbTiO$_{3}$'s behavior using minimal experimental data and simple polynomial models.

Contribution

The authors develop a novel machine-learning framework that identifies optimal polynomial free energy models from experimental-like data, capturing complex couplings without extensive training sets.

Findings

A simple polynomial model accurately describes PbTiO$_{3}$'s free energy surface.

The model captures elastic strain effects on ferroelectric properties.

Physically motivated models can predict material behavior with limited data.

Abstract

We show how to construct Landau-like free energy potentials using a machine-learning approach. For concreteness, we focus on perovskite oxide PbTiO$_{3}$. We work with a training set obtained from Monte Carlo simulations based on an atomistic ''second-principles'' potential for PbTiO$_{3}$. We rely exclusively on data that would be experimentally accessible -- i.e., temperature-dependent polarization and strain, both with and without external electric fields and stresses applied --, to explore scenarios where the training set could be obtained from laboratory measurements. We introduce a scheme that allows us to identify optimal polynomial models of the temperature-dependent free energy surface, mapped as a function of the homogeneous electric polarization and homogeneous strain. Our results for PbTiO$_{3}$ show that a very simple polynomial -- where only two parameters depend linearly on temperature -- is sufficient to yield a correct description of the material's behavior. Remarkably, the obtained models also capture the subtle couplings by which elastic strain controls key features of ferroelectricity in PbTiO$_{3}$ -- i.e., the symmetry of the polar phase and the discontinuous character of the transition --, despite the fact that no effort was made to include such information in the training set. We emphasize the distinctive aspects of our methodology (which relies on an original form of validation step) by comparing it with the usual machine-learning approach for model construction. Our results illustrate how physically motivated models can have remarkable predictive power, even if they are derived from a limited amount of data. We argue that such ''third-principles'' models can be the basis for predictive macroscopic or mesoscopic simulations of ferroelectrics and other materials undergoing non-reconstructive structural transitions.