Lithium Plating Heterogeneity Caused by Realistic Thermal Gradients

Abstract

Fast charging is a key requirement to alleviate range anxiety and enable mainstream adoption of electric vehicles (EVs). Charging rates are limited primarily by the risk of lithium plating, a side reaction that can lead to the rapid deterioration of cell capacity and dendrite-induced thermal runaway in extreme cases. Whilst many studies have investigated lithium plating under pseudo uniform thermal conditions, much less is known about how thermal gradients impact the severity and distribution of the deposited lithium. In this work, we investigate this using a multilayer P3D physics-based model of a high energy pouch cell accounting for nonlinear diffusion. The results show that even a small thermal gradient of 2 ⁰C across the cell thickness can lead to significant plating heterogeneity. A balance between two physical effects is found to affect the relative plating rates between the cell layers: the reduced diffusion rates in the colder regions, and the increased current density in the warmer regions. It is shown that either effect can dominate depending on the average cell temperature: at moderate and higher temperatures, an increased plating rate is observed in the colder cell layers, but the trend is reversed at low temperatures. The thermal gradients also affect the overall amount of deposited lithium compared to a thermally uniform cell. At low temperatures, where the current distribution effect dominates, the average plating rate is increased. However, at moderate to high temperatures, where the diffusion limitation effect dominates, a thermal gradient can lead to a reduced average plating rate. Sensitivity studies were conducted to confirm these conclusions are likely to apply to other pouch cells. The model findings are supported through experimental validation based on both electrochemical signatures (differential voltage analysis) and post-mortem examination. The results provide insights into the effect of EV-relevant thermal gradients on lithium plating behaviour and highlight some of the limitations of 1D models in estimating the fast charging performance of real-world systems.