Comparison of Electrolyte Transport Modelling in Lithium-ion Batteries: Concentrated Solution Theory Vs Generalized Nernst-Planck Model

Mathematical modelling is widely used to provide insights into lithium-ion battery operation, mainly by using Doyle-Fuller-Newman (DFN) porous electrode theory. A key aspect of thermo-electrochemical models is the description of electrolyte transport phenomena and their implications on thermal effects, which are the subject of this study. We show that the so-called generalized Poisson-Nernst-Planck approach (here re-named generalized Nernst-Planck, gNP) for electrolyte transport is equivalent to DFN concentrated solution theory only if the electrolyte thermodynamic factor obeys a specific gNP expression as a function of three electrolyte parameters. However, such an expression does not capture accurately the experimental dependence of the thermodynamic factor for concentrations lower than 0.5 mol l−1 or higher than 1.5 mol l−1 in a common LiPF6-based electrolyte, causing discrepancies between model predictions. The deviation between simulation results of the DFN and gNP models is negligible at low C-rates and ambient temperature. However, as the operative conditions get more challenging as for C-rate > 1 and/or extreme temperatures, detectable deviations are shown in terms of predicted voltage, maximum temperature, and accessible/ restored capacity. Furthermore, the electrolyte transport models predict different onsets of lithium plating upon charge, showing moderate deviations in the estimated penetration depth of plating.