Lithium Localization by Anions in Argyrodite Solid Electrolytes from Machine‐Learning‐based Simulations

Abstract The introduction of density functional theory (DFT) has improved the study of material properties. This has enabled significant breakthroughs in solid electrolytes, which have emerged as promising candidates for next‐generation energy storage systems. However, DFT faces limitations due to the extremely high computational costs required for large atomic cells and long simulation times. In the current study, AI‐based simulations using neural network potentials (NNPs) are introduced to extend the capabilities of DFT to explore the effect of anions on lithium diffusion in Li argyrodite (Li 6 PS 5 X, X = Cl and Br). The investigation categorizes lithium frameworks into two distinct cages, demonstrating that sulfur ions in these cage centers bind the surrounding lithium ions. From the results, a strategy is proposed to enhance lithium ion conductivity by minimizing the occupation of sulfur ions in cage centers. This research provides a benchmark for evaluating lithium ionic conductivity based on anion configuration in cage centers and advances the understanding of ionic transport in Li argyrodite, informing potential improvements in energy‐storage technologies.