Viewing the water challenge of PEM fuel cells from a biological perspective

At high current density, proton-exchange membrane (PEM) fuel cells tend to produce too much water, flooding the stack and impeding performance. Koizumi et al. draw inspirations from biological systems and offer a new perspective on this problem.

During biological cellular respiration, energy-carrying molecules store chemical energy obtained from glucose, and generate water as a byproduct when releasing the energy. This process takes place as an open system in a non-equilibrium and steady state.

Using time-resolved small-angle neutron scattering (SANS), the researchers observed the process of what they see as the equivalent to a biological respiration mode in a PEM fuel cell. In this process, the cell voltage, detected by a segmented electrode, varies along a gas-flow channel from the upper to bottom stream and oscillates in time. As the current density increases, the cell voltage at different positions begins to synchronize.

To further investigate water’s role in this respiration process, the researchers also introduced a contrast variation for SANS to study a polymer electrolyte film humidified with water in a single fuel cell, but switched the fuel gas from hydrogen to deuterium.

Their results showed that the scattering intensity oscillates with a time interval of roughly 100 seconds — as if the system is “breathing,” likely caused by interparticle interference of water clusters, which originate from the water microdomains in the polymer electrolyte.

“Usually, researchers observe only the cell voltage,” said author Satoshi Koizumi. “But we added a new SANS element to show that this respiration behavior is a nonlinear, non-equilibrium phenomenon in an open system, where flooding plays a feedback role to decelerate fuel transportation and the chemical reaction of water generation.”

Source: “Heterogeneous cell performance of polymer electrolyte fuel cell at High Current Operation: Respiration mode as non-equilibrium phenomenon,” by Satoshi Koizumi, Satoru Ueda, Putra Annada, and Yasuyuki Tsutsumi, AIP Advances (2019). The article can be accessed at https://doi.org/10.1063/1.5099498 .