Unsupervised Characterization of Interfacial Reaction and Diffusion Mechanisms in ML-Driven Molecular Simulations

Recent advances in machine learning interatomic potentials have allowed access to large-scale molecular dynamics (MD) simulations, necessitating automatic and flexible analysis of their dynamics. Conventional techniques, like common neighbor analysis or polyhedral template matching, constrain analysis to crystalline structures or rely entirely on prior knowledge of the system dynamics. Therefore, we employ an unsupervised clustering algorithm to characterize phase transitions and product formation of large-scale interfacial reaction simulations. Representations of local atomic environments of MD trajectories are clustered based on their geometric descriptors. We demonstrate this technique by examining high-temperature and pressure phase transitions of silicon carbide and interfacial reactions leading to the formation of the solid-electrolyte interphase (SEI) in solid-state lithium batteries. We identify the time evolution of possible phases, including complex molecules and crystalline, amorphous, or solid state compounds, to characterize the SEI formation in solid-state batteries. Broadly, this framework provides automated data-driven insights into the dynamics of solid state reactions and phase nucleation using large-scale MD simulations.