Characterization and Model-Based Investigation of Lithium-Ion Battery Cell Formation
Felix Schomburg,Michael Anton Danzer,Fridolin Röder
DOI: https://doi.org/10.1149/ma2022-022112mtgabs
2022-11-24
ECS Meeting Abstracts
Abstract:The formation process of lithium-ion battery cells is one of the main contributors to the overall cell cost [1]. During this process step the solid electrolyte interface (SEI) is initially built. It is agreed upon that the composition and morphology of the SEI influences the safety and aging behaviour of the battery cell. However, the processes behind the initial build-up are not fully understood due to the manifold influencing factors, such as electrolyte composition, electrode structure, temperature, and formation protocol, which all impact the complex nanoscale processes at the interface [2,3,4]. High experimental effort is needed to investigate and optimize this process. Detailed multiscale models based on first principles can provide insights into the SEI formation but require many parameters that are difficult to determine and are also associated with uncertainty [5,6]. Therefore, those approaches are limited in terms of exploration and optimization of the formation process. Hence, there is a need for novel methods that enable to obtain information about the SEI growth and investigate the influential factors to effectively accelerate the development process. Here, we suggest a model-based approach that uses lumped growth models that can quickly be identified by using differential voltage data of formation experiments. Further, models are applied to simulate the impact of the formation protocol on the SEI formation with minimal experimental effort. Firstly, we present a parameter identification process for the applied growth model, which is based on a slow formation protocol and a reference measurement in a 3-electrode setup. The parameterization of the SEI model is conducted by combining coulometry and differential voltage analysis (DVA). Shifts of distinct peaks, i.e. feature, within the anode potential curve during the formation and the reference measurement are used to deduce the capacity loss due to the SEI growth. Using the SEI density and surface area of the electrode, the capacity loss is converted into an effective layer thickness, which is used to fit the parameters of the growth model. Within the SEI model different growth mechanisms can be implemented and compared. It is demonstrated that this approach can show differences in the SEI formation for a variation of electrolyte compositions, i.e. the electrolyte additives as can be seen in Figure 1. Afterwards, a simulation study for different formation protocols is presented. The identified growth model has been coupled with a pseudo-2-dimensional (P2D) model. Simulations are also compared to experimental data for various formation protocols and are shown to be in good agreement. Simulations reveal that the final effective film thickness is only slightly affected by current magnitude and mainly depends on the formation time as well as the anode potential. Furthermore, it can be seen that higher formation currents affect the homogeneity of the SEI across the electrode negatively. To sum up, the DVA analysis of the formation yields information about the SEI growth. This can be used to parametrize SEI growth models and investigate film growth mechanisms for different electrolyte compositions. Coupling growth models with P2D models enables to explore the influence of the formation protocols. To conclude, this novel approach enables to significantly accelerate the optimization of battery cell formation. In future work the approach could be refined to enable application to full cell voltage data, which would allow its use within battery production lines and could support end of line testing. Figure 1