The University of British Columbia

Published purpose

CO2 electrolysis is a promising energy storage technology that uses renewable electricity to convert CO2 into valuable fuels and chemicals. Commercializing CO2 electrolyzers requires the development and optimization of new materials, such as membranes and electrocatalysts. Today, these materials are typically tested in isolation under conditions that are much different than that of an actual electrolyzer (wherein catalysts on gas diffusion electrodes (GDEs) are interfaced with membranes in a membrane electrode assembly (MEA)). As a result of this non-representative testing methodology, seemingly promising material candidates may fail to exhibit high performance once integrated into an electrolyzer. MEAs can be tested in benchtop reactors, however existing benchtop approaches for screening materials are inefficient and labour-intensive. The Recipient proposes to circumvent these challenges by designing and building a robotics-based materials acceleration platform (MAP) for MEA development. This MAP will robotically fabricate and characterize GDEs before transferring them to a semi-automated, high-throughput electrochemical testing station where the GDEs will be integrated into MEAs and tested under conditions representative of an operating electrolyzer. To further increase platform productivity, machine learning algorithms will be employed to automatically analyze experimental data and design highly-informative follow-up experiments. This combination of experimental automation and artificial intelligence will result in an autonomous platform for accelerating the development of new, scaleable, and industrially-relevant materials for CO2 applications.