Demand for electric vehicles and distributed energy storage will require many terawatt- hours of new battery capacity over the coming years. Demand for lithium is growing at a rate of over 20% per year due to rapidly expanding lithium-ion battery manufacturing, and higher lithium content per unit energy will be needed for next-generation technologies such as all-solid-state batteries. Unlocking the vast lithium resources found in claystone deposits in the Western US promises to satisfy domestic lithium demand for the rest of the century, but limited understanding of the lithium distribution and phase assemblage of claystone resources is currently the key barrier to responsible extraction from these deposits at costs that are competitive with imported lithium chemicals. Available commercial platforms for drill core analysis are insensitive to lithium and other light elements such as hydrogen, carbon, and oxygen, and do not integrate disparate datasets into self-consistent thermodynamic model that can inform separations strategies. We address this urgent need by incorporating correlative hyperspectral spectroscopic mapping into a high- throughput and field-deployable resource characterization platform powered by deep learning and a quantum chemical database. Successful deployment of this characterization and analysis platform will provide real-time information about the economic potential and environmental impacts of critical element extraction and allow for the discovery of new separations strategies on-the-fly.
