Safety of automotive lithium-ion battery cells under abusive conditions : innovation report

The research carried out in this report focuses on the topic of safety of Li-ion battery cells, specifically for automotive applications. Electric vehicle battery safety is a challenge that must be tackled, especially with the rapid electrification of vehicles. Cell abuse testing simulates their failure process under different scenarios. This helps develop a deeper understanding of the failure process, its root cause and associated mechanisms, hence enabling the improvement of their safety. This research has experimentally investigated four abusive conditions; mechanical penetration, external short circuit, cell swelling as a result of overcharge and overcharge in an adiabatic environment. A number of potential industrial applications based on the research findings are also discussed. \n \nDuring nail penetration testing the effect of nail material and diameter were investigated. Firstly, cells were fully penetrated using 10 mm diameter nails with three different materials; copper, steel and plastic. Secondly, cells were penetrated using 10 and 3 mm diameter copper nails. It was found that there was a clear distinction between the outcome of the conducting and non-conducting nails. However, the outcome of using electrically conductive nails suffered from poor reproducibility. Post-mortem examination showed that at the point of penetration the nail dragged the copper current collector in the direction of penetration along with the separator. The hole in the positive electrode looked less circular and the aluminium current collector was not dragged as deep as the copper one. \n \nDuring external short circuit testing the effect of the short resistance and the short duration was investigated. Firstly, cells were short-circuited using a range of resistance values. Secondly, a programmable power supply to control the shorting duration was used. It was found that the degree of damage experienced by a cell during a short is not only defined by the short resistance, but also its duration. The cells were cycleable after the short circuit event and their capacity and resistance increase depended on the short circuit current magnitude and the short duration. Opening the cells after testing and studying their components using SEM showed no change in the surface morphology of the electrodes. \n \nDuring the third set of experiments, purpose-built equipment was designed and built for in-situ volume measurement. The change in cell volume during cycling, overcharge and 10 cycles after the overcharge event was monitored and measured in-situ. The effect of the degree of overcharge and the magnitude of the charging current were studied. After the overcharge event the cycling behaviour of the cells was investigated. \n \nElectrochemical Impedance Spectroscopy (EIS) and Direct Current Internal Resistance (DCIR) were used to track the change in resistance. An Equivalent Circuit Model (ECM) was built to investigate the individual components contributing to the cell’s impedance. The overcharge-induced capacity fade was analysed using incremental capacity analysis (ICA). The reversibility of cell volume after swelling was also investigated. Results show that cell swelling and the extent of damage depended on the degree of overcharge and the C-rate. Cell swelling was partially reversible and the cells were cycleable after the overcharge event. \n \nFinally, cells were overcharged in ambient and adiabatic conditions. This was carried out to study the effect of heat dissipation on the outcome of an overcharge event. Results highlighted the critical role of heat dissipation from the cell in determining the outcome of the test. The same overcharge regime under different conditions resulted in very different outcomes. Cells overcharged in ambient conditions swelled significantly, but did not vent nor catch fire, whereas, all cells overcharged under adiabatic conditions either ruptured or caught fire. The magnitude of the overcharge current in adiabatic conditions determined the failure mode. Cells overcharged using 0.13 C current ruptured after swelling significantly, but did not catch fire. Cells overcharged with 0.33 and 1.3 C currents were completely combusted.