Li-ion batteries are expected to perform in a wide range of operating conditions. Nevertheless, extreme operating temperature, voltage or current can result in significant safety issues including Li- and Cu-plating, gas formation, venting and even thermal runaway. Normally, BMSs ensure that the battery operates within the boundaries of these critical parameters. However, internal cell defects usually cannot be detected in time and hence cannot be avoided, representing unpredictable safety risks. Well-established test procedures, design to proof battery quality, emulate absurd worst-case scenarios and rarely reflect real circumstances leading to safety issues. Therefore, we aim to develop more realistic test methods for characterization of internal cell defects in Li-ion cells and their effect on cell safety-behavior as well as propagation in the battery module.
To this end, we developed two test procedures based on two different light-overcharge approaches using lab-scale pouch-cells, deliberately generating Li-plating as realistic trigger mechanism. Both methods were validated on large-scale, commercial pouch- and cylindrical cells including different cell chemistries using a specially designed testing setup to perform the test under safe conditions. Cell capacity and internal resistance were determined before and after testing to assess cell degradation. Thermal sensors as well as thermal imaging were employed to record the surface cell-temperature during testing. Cell degradation at component-level was addressed via optical post-mortem analysis.
Both test procedures lead to the successful generation of Li-plating in a reproducible and reliable manner as confirmed by optical post-mortem analysis of the cells. Dendrites were shown to have grown on the anode surface and through the separator in all tests. Our most promising test method induces massive Li-plating, resulting in a dramatic increase of cell surface-temperature accompanied by a cell voltage drop. After the test, we find a significant increase in internal cell resistance and a capacity loss of up to 85%. This behavior is characteristic for all investigated cells. Despite the substantial cell degradation, most high-quality cells do not yield any external damage or critical safety issues, contrary to the forced thermal runaway expected under unrealistic cell abuse testing, whereas low-quality cells exhibit external damage, gas formation, venting and even thermal runaway. Accordingly, cells of different quality show characteristic behaviors under the developed test method, allowing cell classification and risk assessment.