Hydrogen Diffusion in Magnesium Using Machine Learning Potentials: a comparative study

TL;DR

This study demonstrates that machine learning potentials, especially when fine-tuned, can accurately predict hydrogen diffusion in magnesium, matching experimental results while significantly reducing computational costs compared to traditional methods.

Contribution

It provides a comprehensive comparison of ML interatomic potentials and shows how fine-tuning improves accuracy for hydrogen diffusion in magnesium.

Findings

ML potentials match experimental diffusion coefficients

Fine-tuning enhances model accuracy for defective materials

ML reduces computational effort compared to traditional methods

Abstract

Understanding and accurately predicting hydrogen diffusion in materials is challenging due to the complex interactions between hydrogen defects and the crystal lattice. These interactions span large length and time scales, making them difficult to address with standard ab initio techniques. This work addresses this challenge by employing accelerated machine learning (ML) molecular dynamics simulations through active learning. We conduct a comparative study of different ML-based interatomic potential schemes, including VASP, MACE, and CHGNet, utilizing various training strategies such as on-the-fly learning, pre-trained universal models, and fine-tuning.By considering different temperatures and concentration regimes, we obtain hydrogen diffusion coefficients and activation energy values which align remarkably well with experimental results, underlining the efficacy and accuracy of ML-assisted methodologies in the context of diffusive dynamics. Particularly, our procedure significantly reduces the computational effort associated with traditional transition state calculations or ad-hoc designed interatomic potentials. The results highlight the limitations of pre-trained universal solutions for defective materials and how they can be improved by fine-tuning. Specifically, fine-tuning the models on a database produced during on-the-fly training of VASP ML force-field allows the retrieving of DFT-level accuracy at a fraction of the computational cost.