Electrochemical Modeling of Silicon in Lithium-Ion Batteries Using a Multi-Species, Multi-Reaction Framework with Atomistic Insights

Abstract

Silicon is a promising anode material for lithium-ion batteries due to its high capacity and potential for fast charging. However, its electrochemical behavior is dominated by pronounced voltage hysteresis, particle-size-dependent voltage plateaus, and relaxation processes induced by hysteresis. Conventional Doyle-Fuller-Newman models cannot capture these phenomena. Here, we present a multi-species, multi-reaction framework that explicitly considers the multiphase lithium-silicon system by assigning an independent equilibrium potential to each phase, derived from modified Nernst equations and parameterized with experimental and atomistic data. The model captures both asymmetric lithiation and delithiation pathways as well as phase-fraction evolution in silicon half-cells. Quantitative comparison yields root-mean-square errors of 5.4-36.9 mV during constant-current and pulse protocols, corresponding to a relative RMSE of 0.6-4.1% of the overall voltage window. Simulations further reveal that phase fractions continue to evolve during relaxation through thermodynamic redistribution of lithium between phases, governed by phase-specific equilibrium potentials and kinetics. This cross-dimensional approach enables a mechanistic representation of voltage hysteresis, providing a pathway toward improved state estimation, cell design, and battery management.