(13d) Spatiotemporal Decoupling of Cerium-Mediated Water Electrolysis for Grid Energy Storage and Hydrogen Generation

The implementation of electrolysis systems for hydrogen production has continued to grow as the paradigm shift towards renewable energy and fuels progresses. However, when intermittent renewable energy sources power conventional polymer electrolyte membrane (PEM) electrolysis systems, the performance and safety of PEM electrolyzers degrade due to gas crossover. In order to make electrochemical hydrogen production safer when using intermittent renewable energy, decoupled electrolysis systems have been introduced to temporally and spatially separate the evolution of hydrogen and oxygen. This makes the electrolysis system safer, but it sacrifices efficiency compared to conventional electrolysis systems. In this presentation, I will discuss a cerium-mediated decoupled electrolysis system that can both produce hydrogen and store energy in the redox couple. Cerium was chosen due to its large redox potential of 1.6V, which is higher than water oxidation. This means that in one cell (charging cell), an energy input allows cerium to be oxidized while hydrogen is reduced, and in the other cell (discharging cell), energy can be generated by reducing cerium and oxidizing water. By combining experimental results and grid modeling, we identify optimal grid-integration strategies for the decoupled electrolysis system and provide guidelines for its dynamic operation with variable renewable electricity sources. Furthermore, since the oxidation of cerium and water are competing in the charging cell, Faradaic efficiency (FE) for the cerium reaction is less than 100%, which limits total efficiency of the system. To improve the FE of cerium oxidation, dynamic potential pulsing was employed to maintain a high concentration of cerium at the electrode surface improving the FE. The four variables that affect the behavior of the pulses include the oxidation pulse voltage, the rest voltage, the oxidation pulse duration, and the rest duration. Since the effect of these four factors on the FE and cerium conversion rate were unknown, we employed a Bayesian optimization (BO) to approach to identify the optimum combination of the four variables. The BO method was combined with electrochemical models of the reaction to accelerate the search. Finally, the optimal results will be put into context in terms of their effect on the performance and feasibility of the proposed cerium-mediated decoupled water electrolysis and energy storage system.