Applying the Maxwell-Stefan Diffusion Framework to Multicomponent Battery Electrolytes

The Doyle-Fuller-Newman model is widely applied for the simulation of Li-ion batteries (LIBs), and considers a binary electrolyte (one anion, one cation, one solvent) using Maxwell-Stefan diffusion. However, commercial electrolytes composed of LiPF6 dissolved in a mixture of cyclic and linear organic carbonates, are not accurately captured within this description. Here, using Raman spectroscopy of an optically-accessible electrochemical cell, we obtain concentration profiles for each electrolyte species, confirming they deviate significantly from those expected from a single-solvent assumption. Starting from parameters obtained from classical molecular dynamics, we thus develop an iterative fitting approach to parametrize multicomponent LIB electrolytes based on such concentration profiles. Importantly, each electrolyte component can be accurately simulated, to better understand their contributions to battery performance and degradation. The robust methodology developed for the parameterization of multicomponent LIB electrolytes is expected to aid the development and optimization of future LIBs.