Lithium ion batteries (LIBs) currently dominate the market for electric vehicles (EVs), due to their high energy density, high power density, and long lifespan, combined with sweeping cost reductions that have played out over the last decade. A LIB is a relatively complex system with many opportunities for deeper understanding and efficiency through data analytics. Of the various materials in a LIB, the cathode materials are the most expensive and are constantly being changed in new battery formulations. Accordingly, machine learning approaches are needed to accelerate the recycling of LIBs as new chemistries and
cell designs drive changes in supply and materials choices. Even though high purity and performance cathode materials may be recovered, they may not “close-the-loop” for a circular economy unless data-science techniques can be leveraged to accelerate the process to target leading-edge cell designs. In this project, we will use machine learning to design methods to upcycle the old cathode materials to the current cathode materials. Specially, we will focus on converting low Nickel LiNi1/3Mn1/3Co1/3O2 (NMC111) to high Nickel LiNi0.6Mn0.2Co0.2O2 (NMC622) initially. The knowledge gained from the project can be applied for other chemistries based on an analysis of this case.
As shown in Figure E, NMC111 particles will be separated, coated with a Nickel compound (for example nano-size Ni2O3), and then converted to primary particles NMC622 by high temperature sintering. During this process, different dopants can be added to improve the performance. The project will lead to high-quality battery materials from lower quality waste. Machine learning methods will be explored to help determine which input parameters (e.g., temperature, concentrations, etc.) aids in determining final product quality, and in predicting which experimental conditions will yield optimal upcycling of battery materials.
Researchers: Dr. Wang, Mat. Science & Data Science faculty, TBD.
As shown in Figure E, NMC111 particles will be separated, coated with a Nickel compound (for example nano-size Ni2O3), and then converted to primary particles NMC622 by high temperature sintering. During this process, different dopants can be added to improve the performance. The project will lead to high-quality battery materials from lower quality waste. Machine learning methods will be explored to help determine which input parameters (e.g., temperature, concentrations, etc.) aids in determining final product quality, and in predicting which experimental conditions will yield optimal upcycling of battery materials.
Researchers: Dr. Wang, Mat. Science & Data Science faculty, TBD.