Quality Improvements for Anode Coating in Lithium-Ion Battery Cell Manufacturing : A Case Study at Northvolt Labs

Abstract: Lithium-ion batteries (LIB) represent a promising energy storage solution in the pursuit of electrification to combat climate change. In order for LIBs to be used across different industries, they have to be commercially viable. The viability for manufacturing LIBs at scale is increasing, with manufacturing costs decreasing 89% in the last ten years. However, the LIB manufacturing process is complex and can generate large amounts of scrap due to various non-conformities (NCs). Therefore, to further increase the ability to manufacture high-quality LIBs at scale, it is crucial to minimize the occurrence of NCs by understanding their root causes. This thesis examines the characteristics of one of the non-conformities occurring in the electrode coating process, namely the formation of craters on the coated surface of the anode electrode. The thesis was conducted at Northvolt Labs using a DMAIC approach to establish relationships between various process parameters and the formation of craters in two processes, coating, and its precursor process, slurry mixing. Utilizing the data models linear regression, CART regression, regularized linear regression, and a slurry experiment, process parameters and characteristics that affect crater formation were identified. Firstly, from the data models, it was distinguished that the speed of the supply pump used in transferring the slurry from the supply tank to the slot die, and the pressure in the filter pump, have the largest effect on crater formation. Further, the time that the slurry spends in storage, i.e. from a completely mixed slurry batch to it being applied in coating, affects crater formation. In this case, the longer the slurry is stored, the more craters are found. Another notable result is that refilling the coating supply tank induces crater formation. The mentioned results indicate that the various stages of slurry transfer undertaken before coating can result in advantageous conditions for craters to form. Moreover, it was discovered that changes in the loading level of the coated anode surface can indicate crater formation. The slurry experiment discovered that by contaminating the slurry with lubricant grease, NCs with similar characteristics to the crater could be generated. While not likely related to craters, this result provides valuable insights for slurry contamination. In addition to the data models and experiments, actions to facilitate future statistical analysis investigations are proposed. This thesis also proposes actions that can be undertaken to potentially mitigate the formation of craters. Suggested actions include methods to investigate the optimal storage time of the slurry before used in coating. Further, we recommend that the coating process should be monitored through the use of control charts on the loading level measurements of the coated surface. Consequently, large changes in loading level can be detected, entailing potential crater formation. We also propose adding lubricant grease as a potential risk in the PFMEA Northvolt uses for process risk evaluation. This recommendation is also complemented with suggested actions on how to handle the risks of lubricant grease contamination.