Unveiling Lithium-Atom Insertion and Clustering in Superionic Solid-State Electrolytes via Mechanical Deformation

Solid-state electrolytes (SSEs) are pivotal for next-generation energy systems, offering enhanced safety and higher energy density. In particular, a fundamental challenge in garnet-type Li7La3Zr2O12 (LLZO) and lithium superionic conductor Li10GeP2S12 (LGPS) applications is understanding and controlling the mechanical behavior at interfaces and under operational stresses, which can induce microcracking, void formation, and lithium dendrite growth. To address this gap, we employ machine learning interatomic potentials (MLIPs) with near-DFT accuracy and perform large-scale MD simulations to investigate two critical aspects: (1) the mechanical deformation of bulk LLZO under uniaxial compression and tension, revealing atomic-level mechanisms of elastic and plastic deformation, crack initiation, and failure; and (2) the kinetics of prescribed Li-atom insertion at model interfaces, exploring nucleation, clustering, and interface dynamics. Specifically, we identify two distinct lithium diffusion mechanisms: lithium ions in LGPS migrate via stepwise hopping along one-dimensional channels; in contrast, LLZO exhibits cooperative crowding (or “pushing”) within a three-dimensional network, where ion migration involves collective displacements and local structural distortions. These studies provide novel atomistic insights into the mechanical stability of SSEs under stress and the interfacial processes governing Li-atom insertion and clustering, contributing to the understanding and mitigation of key failure mechanisms in solid-state batteries.