With the fast growing of the electric vehicle (EV) market and soaring production of the EV lithium-ion batteries (LiBs) in China, more and more life cycle assessment (LCA) studies has been focused on their impacts towards resources, energy and environment in recent years. As the indispensable background data, the life cycle inventories (LCIs) of the cradle-to-gate stage of the commonly used materials for the production of LiBs play an important role in the related LCA studies, since their accuracy and reliability can affect the results and conclusions greatly. However, LCIs of the LiB materials produced in China are in scarcity nowadays, which has become an obstacle for the further understanding of the eco-performances of EVs and LiBs. In this study, a general unit-process model was established firstly, and then a variety of foreground data were collected from the representative manufacturers of LiB materials in China. With a sequential calculation procedure carried out step by step, the targeted LCIs of the materials were obtained, along with the corresponding 95% confidential intervals through Monte-Carlo simulations. According to the inventories analysis, it was found that among the materials studied, the polyethylene separator made by wet drawing process, lithium hexafluorophosphate, carbon nanotubes, lithium nickel manganese cobalt oxide and nickel cobalt aluminum are with relatively higher primary energy demands and greenhouse gas emissions in their cradle-to-gate stage. Besides, among the cathode materials studied, lithium iron phosphate consumes the lowest non renewable mineral resources. In addition, through a comparison with GREET, we found that the energy consumptions and greenhouse gas emissions of the LiB materials in this study are significantly higher than that drawn from GREET. This will lead to 62–91% more greenhouse gas emissions from materials made in China than those in the United States in order to produce LiB cells of the same capacity. However, it was also found the discrepancies may be stemmed from many difference in the processes of modeling and data acquisition between this study and the supportive studies for GREET.
Read full abstract