Constructing a Robotic Platform for the Autonomous Design of High-Entropy Electrolytes

Abstract

The use of machine learning and robotics in laboratory experiments has seen a rapid increase in recent years, driven by the need for more accurate and efficient experimental results.Despite strong synergies when combined with machine learning models, autonomous robotics hardware for electrochemical experiments has not received the same spotlight. In this work, we introduce a customizable robotic platform designed for the electrochemical characterization of liquid battery electrolytes. The platform is applicable for synthesizing and characterizing a large number of samples with varying compositions and can be used in connection with machine learning algorithms for performance optimization.We discuss the challenges associated with working with a robotic platform compared to other solutions, demonstrate design considerations on test cells for <1 ppm O2 atmosphere, and how to design a heating/cooling system. A brief overview of different fabrication methods and tools for developing the robotic setup is offered. We present a hardware control solution that allows for the future addition of hardware like pumps and sensors and show how to integrate these components while keeping a tight code design. Finally, we demonstrate the importance of accurately monitoring the uncertainties for the included components and their implications on the training accuracy of the machine learning model.Overall, our work presents a promising approach to electrochemical synthesis and characterization that has the potential to improve the efficiency and accuracy of experimental results.