Stable Solid Electrolyte Interphase Formation with Optimal Charging Protocol

Stable solid electrolyte interphase (SEI) layers are key to the high performance of lithium-ion batteries, impacting metrics such as coulombic efficiency and cycle life. The instability of the SEI on the surface has severely impeded its practical applications due to poor cycling stability and safety concerns. A non-uniform SEI layer can cause delamination or promote dendrite growth on the surface of the material. Dendrite growth is accompanied by parasitic responses and high electrolyte consumption. Dendritic structures can penetrate the separator and eventually reach the electrode side, compromising battery safety. Therefore, it is important to prevent local concentration of lithium and form a uniform SEI. The SEI varies as a function of electrode material, electrolyte and additives, temperature, potential, and formation protocol. In this study, the charging method and conditions for stable SEI formation are identified using design of experiment method. To achieve high coulombic efficiency, variables such as charging and discharging time, charging and discharging current, and rest time were optimized. Because experiments are time-consuming and expensive, formal experimental design methods optimize measurements of the design space to obtain the best model with the fewest observations. In classical experimental design, the processes of modeling and optimization are distinct; however, contemporary approaches that are based on models have the capability to modify the response surface for more effective sampling and incorporate optimization directly within the modeling process. In this study, charging conditions are identified using machine learning based design of experiments. As a result, optimal charging protocol conditions with high coulombic efficiency are found with fewer experiments compared to classical experimental design methods.