Robust Edge Design Facilitated By 4D X-Ray Computed Tomography Visualization of Membrane Degradation in Fuel Cells

Abstract

Membrane electrode assembly (MEA) edges are sensitive regions that could strongly influence the durability of polymer electrolyte membrane fuel cells. Membrane failure in poorly designed edges can promote gas crossover and often lead to premature cell failures as well as reduced durability; hence, shortened fuel cell lifetimes. Currently, there is limited knowledge addressing the mechanisms of edge failure in the literature and this gap is explored in this work. Two different MEA edge designs for our small scale fixture and research scale MEA (<1 cm2), where edge effects are more pronounced, were implemented to study their robustness during a combined chemical/mechanical membrane degradation accelerated stress test (AST)[1]. Four-dimensional in-situ visualization[2], enabled by X-ray computed tomography, was performed to understand and thus mitigate the design issues responsible for edge failures. Interfacial interaction between the adhesive-containing polyimide (PI) gasket layer and the catalyst coated membrane (CCM) was identified as a key contributor to premature edge failures, which introduced significant voltage decay and fluctuations due to permanent membrane deformation subjected to the hygrothermally severe AST conditions. CCM slipping was observed at the edges likely due to adhesive melting, which led to excessive CCM cracks. This issue was mitigated in a subsequent design by using a non-adhesive polytetrafluoroethylene (PTFE) layer at the CCM interface along with changes in gasket coverage area, which led to: (i) delayed onset of edge failure; (ii) nearly five times reduction in edge crack size; (iii) elimination of membrane tearing; and (iv) minimal impact of edge failures on cell performance. This mitigation enabled a robust MEA edge wherein the performance-impacting failure was shifted from the edges to the active area regions where operational factors are expected to play a role in eventual degradation of performance. The active area is broadly composed of uncompressed channel regions and compressed land regions. Membrane cracks, observed exclusively in channel regions, were the predominant cause of performance loss by opening paths for internal gas crossover. All membrane cracks were driven by membrane buckling into macro pores in gas diffusion layer (GDL) through surface pores on microporous layer (MPL). Cracks were initiated from the catalyst layer (CL) surface, and gradually penetrated the entire membrane thickness. In-situ diagnostics data showed significant gas crossover through membrane cracks, which induced dramatic open circuit voltage loss. Formation of membrane buckling was highly dependent on preliminary non-uniformities in CL and GDL such as cracks and macro pores that cause uneven stress distribution. In land regions, membrane creep into macro GDL pores was observed instead of buckling, but creep did not lead to membrane crack. However, both creep and buckling were induced by similar mechanisms. The accumulation of ionomer and catalyst at creep spots led to local membrane thinning and loss in effective platinum surface area. Although mechanical stress was the main contributor to membrane failure, it is believed that chemical stress accelerated the degradation process since the number of AST cycles needed to achieve ultimate MEA failure was significantly reduced compared to our previously reported pure mechanical degradation tests[2]. Overall, a robust MEA edge design was the most significant outcome from this work. In addition, new knowledge was gained on membrane degradation under combined chemical/mechanical AST, which could contribute to enhanced fuel cell durability. Keywords: fuel cell; membrane durability; edge design; mechanical degradation; chemical degradation; X-ray computed tomography Acknowledgement Funding for this research was provided by the Natural Sciences and Engineering Research Council of Canada, Canada Foundation for Innovation, British Columbia Knowledge Development Fund, and Ballard Power Systems through an Automotive Partnership Canada grant. This research was undertaken, in part, thanks to funding from the Canada Research Chairs program. Reference [1] D. Ramani, et al., J. Electrochem. Soc. 165 (2018) F3200.[2] Y. Singh, et al., Journal of Power Sources. 412 (2019) 224–237. Figure 1