Integrating automated electrochemistry and high-throughput characterization with machine learning to explore Si-Ge-Sn thin-film lithium battery anodes

High-performance batteries need accelerated discovery and optimization of new anode materials. Herein, we explore the Si-Ge-Sn ternary alloy system as a candidate fast-charging anode materials system by utilizing a scanning droplet cell (SDC) as an autonomous electrochemical characterization tool with the goal of subsequent upscaling. As the SDC performs experiments sequentially, an exploration of the entire ternary space is unfeasible due to time constraints. Thus, closed-loop optimization, guided by real-time data analysis and sequential learning algorithms, was utilized to direct experiments. The best material identified was scaled up to a coin cell to validate the findings from the autonomous millimeter-scale thin-film electrochemical experimentation. Explainable machine learning (ML) models incorporating data from high-throughput Raman spectroscopy and X-ray diffraction (XRD) were used to elucidate the effect of short and long-range ordering on material performance.

The publication is available via the link (free-access): https://doi.org/10.1002/aenm.202404961

The code utilized for data analysis is available here: https://github.com/saninalexey/SiGeSn

The version of the HELAO framework used for closed-loop experimentation is accessible under this link: https://github.com/helgestein/helao-pub/tree/SiGeSn