Accelerated computation of lattice thermal conductivity using neural network interatomic potentials

Journal Article · 13 May 2022 · Computational Materials Science

DOI: https://doi.org/10.1016/j.commatsci.2022.111472 · OSTI ID:1892139

Choi, Jeong Min ^[1]; Lee, Kyeongpung ^[1]; Kim, Sangtae ^[1]; Moon, Minseok ^[1]; Jeong, Wonseok ^[2]; Han, Seungwu ^[1]

---

1. Seoul National University (Korea, Republic of)
2. Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)

We report with the development of the density functional theory (DFT) and ever-increasing computational capacity, an accurate prediction of lattice thermal conductivity based on the Boltzmann transport theory becomes computationally feasible, contributing to a fundamental understanding of thermal conductivity as well as a choice of the optimal materials for specific applications. However, steep costs in evaluating third-order force constants limit the theoretical investigation to crystals with high symmetry and few atoms in the unit cell. Currently, machine learning potentials are garnering attention as a computationally efficient high-fidelity model of DFT, and several studies have demonstrated that the lattice thermal conductivity could be computed accurately via the machine learning potentials. However, test materials were mostly crystals with high symmetries, and the applicability of machine learning potentials to a wide range of materials has yet to be demonstrated. Furthermore, establishing a standard training set that provides consistent accuracy and computational efficiencies across a wide range of materials would be useful. To address these issues, herein we compute lattice thermal conductivities at 300 K using neural network interatomic potentials. As test materials, we select 25 materials with diverse symmetries and a wide range of lattice thermal conductivities between 10^-1 and 10³ Wm^-1K^-1. Among various choices of training sets, we find that molecular dynamics trajectories at 50–700 K consistently provide results at par with DFT for the test materials. In contrast to pure DFT approaches, the computational cost in the present approach is uniform over the test materials, yielding a speed gain of 2–10 folds. When a smaller reduced training set is used, the relative efficiency increases by up to ~50 folds without sacrificing accuracy significantly. The current work will broaden the application scope of machine learning potentials by establishing a robust framework for accurately computing lattice thermal conductivity with machine learning potentials.

• Research Organization: Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
• Sponsoring Organization: USDOE National Nuclear Security Administration (NNSA); Korea Institute of Ceramic Engineering and Technology (KICET); National Research Foundation of Korea (NRF); Ministry of Science and ICT
• Grant/Contract Number: AC52-07NA27344
• OSTI ID: 1892139
• Report Number(s): LLNL-JRNL-831802; 1048754
• Journal Information: Computational Materials Science, Journal Name: Computational Materials Science Journal Issue: N/A Vol. 211; ISSN 0927-0256
• Publisher: ElsevierCopyright Statement
• Country of Publication: United States
• Language: English