Abstract
Although it is of scientific and practical importance, the state-of-the-art of predicting the thermal expansion of oxides over broad temperature and composition ranges by physics-based atomistic simulations is currently limited to qualitative agreements. We present an emerging machine learning (ML) approach to accurately predict the thermal expansion of cubic oxides with a dataset consisting of experimentally measured lattice parameters while using the metal cation polyhedron and temperature as descriptors. High-fidelity ML models that can accurately predict temperature- and composition-dependent lattice parameters of cubic oxides with isotropic thermal expansions have been successfully trained. The ML-predicted thermal expansions of oxides not included in the training dataset have shown good agreement with available experiments. The limitations of the current approach and challenges to go beyond cubic oxides with isotropic thermal expansion are also briefly discussed.
Original language | English |
---|---|
Article number | 111034 |
Journal | Computational Materials Science |
Volume | 210 |
DOIs | |
State | Published - Jul 2022 |
Funding
This research was sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U. S. Department of Energy. This research used resources of the Compute and Data Environment for Science (CADES) at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725. The authors thank Chris Layton for his support for using CADES.
Funders | Funder number |
---|---|
CADES | |
Data Environment for Science | |
U.S. Department of Energy | DE-AC05-00OR22725 |
Office of Science | |
Oak Ridge National Laboratory |
Keywords
- Lattice Parameters
- Machine learning
- Oxides
- Polyhedron
- Thermal expansion