Upscaling of Fast Charging Strategies from Coin to Pouch Cells
The formation of lithium-ion batteries is one of the most time-consuming processes during the cell production. Common formation methods consist of several relatively slow constant charging and discharging cycles as it is believed to guarantee good performance of the Solid Electrolyte Interphase (SEI) on the negative electrode. The SEI has a significant impact on the performance, aging and safety of battery cells. Aiming at faster production times, cheaper production costs and better properties of lithium-ion batteries, the optimization of common formation methods is necessary.
This contribution is focused on developing optimized fast charging formation proce-dures for high-energy NMC622/G pouch cells (NMC622, 112 µm, 4.75 mAh/cm² and SMG-A5, 126 µm, 5.97 mAh/cm²). Therefore, different parameters of influence as current density, temperature, external pressure and the degassing timing are varied. Based on measurements with three-electrode coin cells the maximum fast charging capability during the formation is determined by controlling the maximum current close to low positive potentials of the negative electrode to avoid lithium-plating. The experimental results show that this fast charging strategy works properly for coin cells but not for pouch cells. It is found that the gassing during the formation is the major current limitation effect for the pouch cells. However, the pouch cells can also be fast charged during the formation but only after slow precharging and early degassing at ca. 30% SOC. Based on optical post-mortem analyses it is shown that lithium-plating can be successfully prevented during the optimized 1C peak fast charging formation (3.5 h charging time) after degassing. In contrast, state of the art constant-current-constant-voltage fast charging methods lead to significant lithium-plating. Furthermore, the optimized formation procedure result to similar cell properties compared to a much slower 0.2C formation procedures with 6 h charging time.