using machine-learning potential

Phase-change material Ge₂Sb₂Te₅ exhibits interesting behavior in electrical and thermal conduction during the phase transition from the amorphous, to metastable rock-salt, and to stable hexagonal structure. About 10000-fold variation of electrical conductivity is produced by the cation and vacancy rearrangement during the transition, whose atom-wise tracking is yet very limited. Crystallization simulation by ab-initio molecular dynamics is computationally too expensive, and the machine-learning potential (MLP) approach is a good alternative in this respect. The training set construction is a critical step for proper sampling of potential-energy surface and we developed a cost-effective and efficient training scheme, namely, randomized atomic-system generator (RAG) scheme [1]. Using the RAG-trained MLP, we simulated the crystallization of amorphous Ge₂Sb₂Te₅ and calculated the thermal conductivity of several intermediate structures during the crystallization simulation.

[1] Young-Jae Choi and Seung-Hoon Jhi, Efficient Training of Machine Learning Potentials by a Randomized Atomic-System Generator. J. Phys. Chem. B 124, 8704-8710 (2020).