Global warming caused by excessive emission of greenhouse gases and consumption of fossil fuels is calling for rapid action. A crucial strategy to counteract its potentially harmful effects is the implementation of sustainable technologies which do not rely on fossil fuels. In this context, polymer electrolyte fuel cells (PEFC) are particularly promising examples. Their potential is huge, notably in the areas of transportation and portable as well as stationary applications.
Within PEFC, gas is transported to the membrane electrode assembly (MEA) via channels embedded in its bipolar plates (conductive components which connect cathodes and anodes in neighbouring cells in a fuel cell stack). These channels, referred to as "flow fields", are designed to have serpentine-like forms (with single, double or multi-channel designs) or alternatively run parallel to each other.
Water is the only chemical product resulting from the reactions taking place within PEFC. The removal of water from the PEFC needs to be carefully managed, since too much water can partly block the flow field channel, thereby causing flooding and reducing the cell’s performance. Too little water, and the cell dries out, which reduces the proton mobility and can also lead to a reduced performance. Understanding how water is distributed within the channels is therefore crucial for optimising the design of future generations of fuel cells.
Previous studies, both experimental and theoretical, investigated the water management and channel flooding in PEFCs. A systematic four-dimensional comparison of the influence of different channel arrangements is, however, lacking to date. This knowledge gap motivated a study conducted by a team of researchers from Great Britain, Germany and the ILL.
"The unique sensitivity of neutrons to hydrogen was crucial to this experiment. To carry out these measurements, we successfully exploited the high-speed neutron tomography capabilities of NeXT", says Lukas Helfen, one of the responsibles of NeXT. The researchers were able to look ‘inside’ the cells in operando, which brought their experiment very close to real-life conditions. Lukas' colleague Alessandro Tengattini adds: "Our experimental setup allowed us to rotate them continuously, ensuring that we could reconstruct their tomography images from any angle without any gaps in the data acquisition." Tomography was chosen to unambiguously separate the in-plane distributions at the anode and cathode sides. The challenge was in order to attain a relatively high spatial resolution and sufficient image contrast at sub-minute scan durations (36 s).
The researchers visualised the evolution of the water volume contained within the flow fields as a function of varying current density. "While all three designs showed full hydration of the MEA, the parallel (PAR) channel alignment appeared less able to efficiently remove water from the gas-conducting channels as opposed to the single and double serpentine designs", explains Jennifer Johnstone-Hack, the first author of the study. Indeed, the serpentines filled up more slowly and featured lower water volumes at the cathode while maintaining water levels sufficient to hydrate the cell, reflecting a better water management capability over the parallel design.
In agreement with theoretical predictions, water was found predominantly in the outer bends of the single serpentine channel, where flow velocity is lower. The single channel design was also able to efficiently purge large amounts of water (referred to as "slugs)". In contrast, because of the four channels in the poorly-performing parallel design, the flow velocity across each channel was lower, which resulted in significant accumulation of water in multiple channels that could not be efficiently removed.
The sophisticated experimental setup and the unique spatio-temporal resolution available at NeXT also allowed the team to look into the formation and the evolution of individual water droplets within the flow fields. Across all channel designs, the number of droplets decreased over time due to their coalescence into larger droplets. However, important differences were observed with respect to the droplet morphology: within the single serpentine design, droplets were either of the "slug" morphology or wetting the channel surfaces as thin films; in the case of the PAR design, droplet evolution began with a plug-like shape and eventually coalesced into a large droplet. Interestingly, in the double serpentine design, one channel was found to be filled with water "plugs", while the second one contained many isolated droplets.
Overall, the serpentine design had clear advantages with respect to water management. These results demonstrate the power of neutron imaging for direct visualisation of water movement within confined spaces and its immediate relevance for industrial applications.
"This is the first study directly comparing the effect of different flow field designs on water management and cell performance, and quantifying water droplet characteristics in operating PEFCs in four dimensions ", says Jennifer Johnstone-Hack. "We hope that our results will inspire the further development of optimised flow field designs to realise future generations of sustainable, high-performance fuel cells."
Reference: Hack J, Ziesche RF, Fransson M, Suter T, Helfen L, Couture C, Kardjilov N, Tengattini A, Shearing P, Brett D. Understanding water dynamics in operating fuel cells by operando neutron tomography: investigation of different flow field designs. Journal of Physics: Energy. 2024 Apr 12;6(2):025021.
ILL contacts: Lukas Helfen, Alessandro Tengattini
ILL instrument: NeXT