(Invited) PEM Water Electrolysis Stack Development and Related Characterization Tools
Dmitri Bessarabov
DOI: https://doi.org/10.1149/ma2024-01341864mtgabs
2024-08-09
ECS Meeting Abstracts
Abstract:For decades, proton-exchange membrane (PEM) water electrolysis (WE) has been mainly used for oxygen generation in anaerobic environments. Over the past two decades, however, it has been increasingly used for hydrogen generation in the industrial sectors at various scales. The PEMWE technology is also considered a key in the ongoing energy transition, if the process of hydrogen generation by means of WE is linked to renewable energy sources, such as wind, solar, etc. The following key elements enable the operation of a PEM WE plant for hydrogen production: a PEM-based WE stack and the balance of plant. The related system includes, but not limited to such key modules as the water and oxygen management systems, hydrogen gas management system, water input system, safety system, power electronics and electrolysis cell stack power supply, control system, and some other. A WE stack comprises several cells connected in series with electrically conductive bipolar plats and end plates. General design principles for the stack involve reducing efficiency losses, while aiming at lower costs and increase in durability. WE stacks should be designed in such a way, that an even current distribution is maintained, the water feed is optimized, suitable compression ratios are achieved, and preferably high discharge pressure of hydrogen can be enabled. For the stack level, in simple terms, efficiency is benchmarked by the total applied voltage, which includes the Nernst potential, anode and cathode overpotentials, and ohmic overpotentials due to the membrane ionic resistance and interfacial resistance at given current density. Overpotential represents inefficiencies in a cell stack, and some of stack design efforts focus on reducing these overpotential contributions. However, a compromise has to be found between (a) costs reduction challenges (that are often associated with the reduction of PGM-based catalyst loading, attempts to replace Ti as a key material for the bipolar plates, and the less expensive SPE membrane), (b) sufficient durability, (c) performance of the stack, (d) increase in power density of a stack. This talk will focus on the WE stack, and its subcomponents review, as well as characterization methods for both, a single cell and a stack and also challenges of the large-scale stack manufacturing. The key components of the WE stack include ion-conductive solid polyelectrolyte membranes (SPE), anode and cathode catalyst layers (CL), bipolar plates and current collectors/ gas diffusion layers (porous transport layers). In order to gain fundamental understanding of the relationships between electrical loses, degradation of the stack components and strategies to increase power density of the stacks, advanced characterization and modelling tools need to be used. These include, but not limited to electrochemical impedance spectroscopy (EIS), current interrupt (CI), dynamic compression measurements with piezoelectric senor plates, current mapping, gas cross-over analysis, computational fluid dynamics (CFD) simulations modelling, various visualization tools, etc. For example, CFD could assist in understanding and improving fluid flow dynamics in a WE stack as it is a complex challenge that requires a detailed examination of the geometry, channel design, and operating conditions within the stack. As for EIS, the usage of the distribution of relaxation times (DRT) approach would potentially allow to focus on the direct analysis of the data rather focusing on the equivalent circuit development. Approximately 15 years ago the South African Government approved a national program HySA: Hydrogen South Africa, which resulted in the development of the expertise and capacity to conduct research, development, and earlier commercial activities around green hydrogen production by means of water electrolysis. These activities include the development of local IP at the components, stack, and system levels.