Universal and interpretable classification of atomistic structural transitions via unsupervised graph learning

Journal Article · 31 August 2023 · Applied Physics Letters

DOI: https://doi.org/10.1063/5.0156682 · OSTI ID:2229026

Aroboto, Bamidele  ^[1]; Chen, Shaohua  ^[2]; Hsu, Tim  ^[3]; Wood, Brandon C.  ^[4]; Jiao, Yang  ^[5]; Chapman, James  ^[1]

---

1. Boston Univ., MA (United States)
2. Arizona State Univ., Tempe, AZ (United States); Nanjing University of Aeronautics and Astronautics, Nanjing (China)
3. Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States). Center for Applied Scientific Computing
4. Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
5. Arizona State Univ., Tempe, AZ (United States)

Materials processing often occurs under extreme dynamic conditions leading to a multitude of unique structural environments. These structural environments generally occur at high temperatures and/or high pressures, often under non-equilibrium conditions, which results in drastic changes in the material's structure over time. Computational techniques, such as molecular dynamics simulations, can probe the atomic regime under these extreme conditions. However, characterizing the resulting diverse atomistic structures as a material undergoes extreme changes in its structure has proved challenging due to the inherently non-linear relationship between structures as large-scale changes occur. Here, we introduce SODAS++, a universal graph neural network framework, that can accurately and intuitively quantify the atomistic structural evolution corresponding to the transition between any two arbitrary phases. We showcase SODAS++ for both solid–solid and solid–liquid transitions for systems of increasing geometric and chemical complexity, such as colloidal systems, elemental Al, rutile and amorphous TiO₂, and the non-stoichiometric ternary alloy Ag₂₆Au₅Cu₁₉. Finally, we show that SODAS++ can accurately quantify all transitions in a physically interpretable manner, showcasing the power of unsupervised graph neural network encodings for capturing the complex and non-linear pathway, a material's structure takes as it evolves.

• Research Organization: Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
• Sponsoring Organization: USDOE National Nuclear Security Administration (NNSA)
• Grant/Contract Number: AC52-07NA27344
• OSTI ID: 2229026
• Report Number(s): LLNL--JRNL-857208; 1086479
• Journal Information: Applied Physics Letters, Journal Name: Applied Physics Letters Journal Issue: 9 Vol. 123; ISSN 0003-6951
• Publisher: American Institute of Physics (AIP)Copyright Statement
• Country of Publication: United States
• Language: English