Machine-Learned Interatomic Potential for the Mo-S-Al-O System
Abstract
We develop a machine-learned neuroevolution potential (NEP) for the Mo-S-Al-O system, targeting MoS2 heteroepitaxy on c-plane sapphire, α-Al2O3(0001). The potential is trained on 63,611 PBE+D3 structures distilled from about 1.1 million curated density-functional-theory (DFT) configurations, using an energy-difference loss function that directly targets interfacial and orientation-dependent energies. On held-out data, the energy and force errors are 29 meV/atom and 0.23 eV/mathringA. Validated against DFT on a suite of physical properties, the NEP reproduces the lattice constants within 0.3 %, the elastic constants within 17 %, the phonon dispersions with mean deviations below 2.5%, and the surface energies within 1.3 %. At the MoS2/sapphire interface, it reproduces the adhesion energy minimum within 0.8 %, the interfacial energy within 2.2 % and the most stable orientation of the monolayer. It preserves the DFT-predicted energetic ordering of edges, point defects (Spearman ρ = 0.995) and mirror-twin boundaries (ρ = 1.0), and reproduces tilt grain-boundary energies within 8 % (ρ = 0.98) and stacking-fault energy barriers within 14 %. The resulting potential enables future studies of the atomic-scale mechanisms of defect formation in two-dimensional heteroepitaxy.
Related
- A Machine-Learned Interatomic Potential Free-Energy Surface for the First Step of MIL-101(Cr) Secondary Building Unit Formation
- SoLiD26: A First Principles Solid-Liquid Interface Dataset for Machine-learned Interatomic Potentials
- From Energy-Force Weighting to Primal-Dual Optimization of Machine-Learned Interatomic Potentials
- Dielectric Response of Short and Long Range Models of Nonuniform Liquids and Implications for Machine Learned Interatomic Potentials
- Linear response for machine-learned interatomic potentials: phonons at arbitrary wave vectors from the primitive cell
BibTeX
@misc{bilgili2026machinelearned,
title = {{Machine-Learned Interatomic Potential for the Mo-S-Al-O System}},
author = {Emir Bilgili and Nicholas Taormina and Simon R. Phillpot and Youping Chen},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.02403},
eprint = {2611.02403},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.02403},
}
RIS
TY - UNPB
TI - Machine-Learned Interatomic Potential for the Mo-S-Al-O System
AU - Emir Bilgili
AU - Nicholas Taormina
AU - Simon R. Phillpot
AU - Youping Chen
PY - 2026
DA - 2026/11/02
PB - arXiv
DO - 10.48550/arxiv.2611.02403
UR - https://arxiv.org/abs/2611.02403
AB - We develop a machine-learned neuroevolution potential (NEP) for the Mo-S-Al-O system, targeting MoS_2 heteroepitaxy on c-plane sapphire, α-Al_2O_3(0001). The potential is trained on 63,611 PBE+D3 structures distilled from about 1.1 million curated density-functional-theory (DFT) configurations, using an energy-difference loss function that directly targets interfacial and orientation-dependent energies. On held-out data, the energy and force errors are 29 meV/atom and 0.23 eV/mathringA. Validated against DFT on a suite of physical properties, the NEP reproduces the lattice constants within 0.3 %, the elastic constants within 17 %, the phonon dispersions with mean deviations below 2.5%, and the surface energies within 1.3 %. At the MoS_2/sapphire interface, it reproduces the adhesion energy minimum within 0.8 %, the interfacial energy within 2.2 % and the most stable orientation of the monolayer. It preserves the DFT-predicted energetic ordering of edges, point defects (Spearman ρ = 0.995) and mirror-twin boundaries (ρ = 1.0), and reproduces tilt grain-boundary energies within 8 % (ρ = 0.98) and stacking-fault energy barriers within 14 %. The resulting potential enables future studies of the atomic-scale mechanisms of defect formation in two-dimensional heteroepitaxy.
ER -