An Experimentally Driven Automated Machine Learned lnter-Atomic Potential for a Refractory Oxide
This work addresses the need for efficient potential development in materials science, particularly for high-temperature applications, but it is incremental as it applies known methods to a specific material.
The researchers tackled the challenge of developing an inter-atomic potential for HfO2, a refractory oxide, by combining experimental data with an automated active-learning loop, resulting in a multi-phase potential that covers temperatures from room temperature to ~2900°C with reduced development time and human effort.
Understanding the structure and properties of refractory oxides are critical for high temperature applications. In this work, a combined experimental and simulation approach uses an automated closed loop via an active-learner, which is initialized by X-ray and neutron diffraction measurements, and sequentially improves a machine-learning model until the experimentally predetermined phase space is covered. A multi-phase potential is generated for a canonical example of the archetypal refractory oxide, HfO2, by drawing a minimum number of training configurations from room temperature to the liquid state at ~2900oC. The method significantly reduces model development time and human effort.