Seminar on Securing Lithium for an Energy Sustainable Future: Lithium Extraction from Salt-Lakes and Battery Leachates with Membranes

Dr Foo Zi Hao

Massachusetts Institute of Technology

Owing to its high electrochemical activity and heat capacity, lithium is the central component of modern-day batteries and is a resource of increasing strategic importance for most economies. In spite of its abundance in continental and geothermal salt-lakes, however, conventional evaporative technology for lithium mining exacerbates freshwater scarcity and wetland destruction, and suffers from protracted production cycles. Simultaneously, the burgeoning volume of lithium-ion battery waste necessitates urgent development in battery recycling technologies that bypass the traditional limitations of absorbents and ion exchange resins. Here, we unravel the fundamentals of ion transport across nanofiltration (NF) and electrodialysis (ED) membranes, and quantify the techno-economic feasibility of membrane processes for direct lithium extraction from salt-lakes and battery leachates. We first construct non-equilibrium molecular dynamics (NEMD) models to visualize the potential field within the active layers and elucidate the mechanism of counter-ion selectivity. Thereafter, as informed by our NEMD simulations, a highly crosslinked polyethyleneimine surface layer is fabricated onto conventional polyamide NF and polysulfonate cation exchange membranes, to enhance counter-ion selectivity with the Gibbs-Donnan effect. Subsequently, we quantify the selectivity enhancements of the composite NF and ED membranes with bench-scale pressure- and electrochemically-driven cells, leveraging >3000 original concentration measurements with salt-lake brines and battery leachates that span four feed salinities, three pH levels, and five current densities. The measurements are used to calibrate semi-empirical phenomenological models for NF & ED, and are employed to elucidate the dependence of the ion selectivity and thermodynamic efficiency on intrinsic membrane parameters, feed composition, salinity and pH level. Lastly, we train a physics-informed neural network (PINN) with our empirical dataset to project system-scale energy consumption and process cost for lithium recovery from salt-lakes and battery leachates.