Text mining for processing conditions of solid-state battery electrolytes

Abstract The search for safer next-generation lithium ion batteries has motivated development of solid-state electrolytes (SSEs), owing to their wide electrochemical potential window, high ionic conductivity (10-3 to 10-4 S·cm-1) and good chemical stability with a wide range of high charge capacity electrode materials. Still, optimization of the processing conditions of SSEs without sacrificing the performance of the complete cell assembly remains challenging. Insights extracted from scientific literature can accelerate the optimization of processing protocols of SSEs, yet digesting the information scattered over thousands of journal articles is tedious and time consuming. In this work, we demonstrate the role of text mining to automatically compile materials synthesis parameters across tens of thousands of scholarly publications using machine learning and natural language processing techniques that glean information into the processing of sulfide and oxide-based Li SSEs. We also gain insight on low temperature synthesis of highly potential oxide-based Li garnet electrolytes, notably Li7La3Zr2O12 (LLZO), which can reduce the interface complexities during integration of the SSE into cell assembly. This work demonstrates the use of text and data mining to expedite the development of all-solid-state Li metal batteries by guiding hypotheses during experimental design.