The Application of Deep Learning on Room Temperature Conductivity of Li₁₀GeP₂S₁₂ Type Solid State Electrolyte

Abstract Room temperature ionic conductivity of the solid-state electrolyte is essencial element commercialization lithium-ion batteries. First-principles method a good tool as it simulates atomic level dynamics, and provides accurate insides ion moving velocity, path, microscopic environment. However, first-principles time computational resources-consuming. It usually applied to systems with hundreds atoms simulation less than 100 picoseconds. The classical molecular dynamics applies containing thousands can reach scale nanoseconds. based on experimental force-field parameters only be limited experimentally-proofed parameters. used in this paper combination dynamics. Deep-learning Deepmd-kit train data targeted system into tailor-made model. verified model for specific will then simulation. With precise efficiency, movement at room achieved. results Li 10 GeP 2 S 12 SiP showed agreement experiments.