The formation of lithium-ion batteries is a crucial process step in battery production. Although it covers only the first few charging and discharging cycles, it impacts both the long-term degradation as well as the performance characteristics of the final cell. Commonly, formation times are on the order of multiple days due to relatively low C-rates. Although this prevents safety-critical lithium plating, research has shown that neither the slowest nor the fastest formation result into the best possible cell characteristics. [1]

To allow for a knowledge-based design of the cell formation process, detailed cell diagnostics including the characterization of the formed solid electrolyte interphase (SEI) are indispensable. However, the preparative effort for most experimental methods for SEI characterization like SEM or XPS is significant and will also require cell disassembly. An in-operando cell diagnosis can be realized with physicochemical modeling based on non-destructive dynamic electrochemical measurements. However, the dynamics of the SEI are commonly either modeled in a simplified way or the models are not designed for the simulation of various measurement types. To overcome these limitations, we extended the classic battery model from Doyle et al. [2] with a detailed SEI modeling. This finally allows to describe C-Rate and EIS data with the same parameter set (see Fig. 1a,1b), providing detailed insights into performance-limiting processes and their changes along cell aging. [3]

On the experimental side, we performed a broad formation study at different temperatures with different currents and current profiles, using small-scale three-electrode test cells. Fig. 1c) shows the discharge capacity for different formation procedures. Clearly, the performance significantly depends on the chosen formation conditions. The model-based cell diagnosis helps to shed light onto this interrelation. Surprisingly, we found that the bulk and interfacial properties of the SEI are not the root cause for the substantial differences in the cell’s fast charge/discharge capability. In fact, the effective transport properties in the anode electrolyte phase are driving the performance differences. Furthermore, the cathode reaction kinetics are affected by the chosen cell formation protocol. Ultimately, our experimental formation study in combination with the model-based cell diagnosis highlights that the cell formation process is not only about a stable SEI but also about minimizing the impact of reaction products from the SEI formation on the bulk electrolyte phase.

References:

[1] H. Mao et al. (2018) J Power Sources 402, 107-115

[2] M. Doyle et al. (1993) J Electrochem Soc 140 (6), 1526-1533

[3] D. Witt et al. (2022) Batteries Supercaps, 10.1002/batt.202200067