Accelerated Design of Beyond-Li Aprotic Electrolytes via Structure-Property Relationships and Transfer Learning

In this project, Professor Chunmei Ban at the University of Colorado Boulder is developing new liquid electrolytes for sodium-ion and potassium-ion batteries, which are promising alternatives to today’s lithium-ion batteries since they rely on more abundant and widely available materials. By integrating machine-learning-guided materials discovery with synthesis, characterization, and electrochemical testing, this research could accelerate the identification and optimization of high-performance electrolyte formulations. The project could deliver validated electrolyte chemistries, predictive design tools, and fundamental insights into the chemical processes that govern battery performance, enabling next-generation energy storage across a broad range of applications while strengthening U.S. leadership and competitiveness in energy innovation. Current sodium-ion and potassium-ion electrolytes largely inherit formulations developed for lithium-ion batteries, leaving critical knowledge gaps that limit optimization of beyond-lithium chemistries. To address these challenges, this project integrates experimental discovery with chemically informed machine-learning models trained on complementary experimental and simulation datasets. The resulting framework will uncover fundamental relationships between molecular structure, solvation behavior, and electrolyte properties, including nonlinear composition–property interactions. These insights will enable the rational design of beyond-lithium electrolytes for batteries, fuel cells, and capacitors, while establishing transferable machine-learning methodologies for predicting the properties of complex chemical mixtures from relatively small datasets. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.