Unraveling the Impact of Surface Lithiation on the Initial Solid Electrolyte Interphase Formation on Graphite via Machine Learning Molecular Dynamics

Advancing modern energy storage requires a fundamental understanding of electrochemical interfaces, where the solid electrolyte interphase (SEI) critically governs efficiency, safety, and cycle life. However, the atomic-level mechanism of its initial formation remains elusive, owing to the complexity of interfacial chemistry that challenges conventional ab initio simulations. Here, we developed a reactive machine learning potential to simulate electrolyte reduction at the anode/electrolyte interface, enabling nanosecond-scale machine learning molecular dynamics to obtain the atomic structure of SEI. By systematically varying the local Li loading at edge sites corresponding to different electrochemical potentials in the anode, we decipher the reaction pathways and identify reactive species, thereby revealing the underlying mechanism of SEI formation. Low Li loading leads to insufficient reduction and facilitates the diffusion of intermediates into the electrolyte, whereas high Li loading promotes the formation of inorganic-rich interfacial clusters that suppress further solvent decomposition. This work develops a computational framework for understanding interfacial chemistry and provides atomic-level insights into potential-dependent SEI evolution on graphite, thereby informing rational interfacial engineering in high-performance batteries.