Deciphering Surface Hydrolysis Mechanism in Argyrodite via Large‐Scale Machine Learning Potential Simulations

Abstract

ABSTRACT Sulfide‐based argyrodite (Li 6 PS 5 Cl) stands out as a promising solid‐state electrolyte (SSE) due to its high Li‐ion conductivity; however, its poor moisture stability leads to toxic H 2 S release and interfacial degradation, severely limiting its practical application. Despite its practical significance, the atomistic mechanism origin of H 2 S generation at the argyrodite/water interface remains poorly understood. Herein, we perform large‐scale molecular dynamics (MD) simulations using a fine‐tuned universal machine learning potential (uMLP) trained on reaction product‐based datasets to investigate the hydrolysis reaction dynamics. Our results reveal that pure H 2 O preserves the [PS 4 ] 3− , inducing a surface hydration layer and Li–O bonds on the argyrodite surface. In contrast, under accelerated reactive hydrogen species (RHSs) conditions, H 3 O + and OH − rapidly trigger [PS 4 ] 3 − decomposition and H 2 S formation by attacking sulfur, indicating that dissociated water‐derived species, rather than intact water, initiate hydrolysis. Guided by this insight, we demonstrate that Sn substitution mitigates H 2 S generation by forming stronger Sn─S bonds based on the hard‐soft‐acid‐base (HSAB) theory. Our MD simulations confirm this substitution preferentially redirects H adsorption from P‐coordinated sulfur sites to Sn‐coordinated sulfur sites, suppressing H 2 S formation. This study provides atomistic‐level insights into the hydrolysis of argyrodite and elucidates the mechanistic origin of the stabilizing effect through Sn substitution.