As part of the PREDICTOR project, my work focuses on the development of an automated, high-throughput electrochemical characterization platform for redox flow batteries (RFBs). The core of this effort is the implementation of potentio-dynamic techniques, such as cyclic voltammetry, linear sweep voltammetry, electrochemical impedance spectroscopy, etc., in an automated workflow to rapidly evaluate redox-active species and electrolyte formulations. This platform enables the systematic and efficient screening of large chemical libraries, significantly accelerating the experimental evaluation phase that traditionally relies on labor-intensive, manual testing.

A key innovation in this approach is the integration of Bayesian Optimization into the experimental workflow. This machine learning technique is used to intelligently guide the selection of experimental conditions, allowing the system to iteratively explore the vast chemical and parameter space with maximum efficiency. By using real-time feedback from electrochemical measurements, the platform continuously refines its understanding of promising candidate materials, enhancing discovery while minimizing resource consumption.

My work contributes to building this closed-loop system, where automated experimentation, data analysis, and optimization are seamlessly connected. This enables not only faster material characterization but also data-driven decision-making that supports the development of predictive models. This work helps realize a self-driving laboratory tailored for electrochemical energy storage research, aligned with PREDICTOR’s mission to accelerate materials discovery and reduce development time for next-generation, scalable, and sustainable energy storage systems.