In cold regions, ensuring startup ability and durability of polymer electrolyte fuel cells (PEFCs) at subfreezing conditions is one of critical issues for the practical use of PEFC. At cold startup of PEFCs near 0 °C like –10 °C, the produced water remains as supercooled state before it freezes, and then the shutdown occurs with ice layer formation between the cathode catalyst layer (CL) and micro-porous layer (MPL) (1). It was estimated from our previous study (2) that water transport in the CL occurs by concentration gradient in ionomer. This study investigated the water transport in ionomer and the ice formation behavior by measurement of cold start characteristic and observation of membrane electrode assembly (MEA) by a cryo-SEM. To evaluate the water transport, we measured change in polymer membrane resistance corresponding to water content during the supercooling release. Further, to improve the cold start characteristic, we introduced a hydrophilic MPL and confirmed that the hydrophilic MPL makes the startup period longer and improves ice distribution. A single cell with an active area of 25 cm2(5 cm × 5 cm) was used in this study. After a preconditioning process to enhance the performance of a MEA, the initial conditions of the water in the cell were carefully controlled by the procedure described in Ref. 1. Then, the cell and the chamber were cooled to –10 °C, and cell operation was started and maintained at a constant current density. The anode and the cathode supplied gases were dry hydrogen and dry air. For the cryo-SEM observation, the sample of the MEA as removed from the cell after the stop of operation was immediately immersed in liquefied nitrogen, and the sample was set to a sample holder. The sample was cut by a cold knife in a vacuumed chamber, and the cut section area was coated with gold-palladium alloy (Au-Pd), and observed at –150 °C. To estimate the changes in the water content of a polymer electrolyte membrane from supercooled water to ice, we measured the change in the in-plane high frequency resistance of the polymer membrane. It is expected that when supercooled water changes to ice, the chemical potential of water decreases and the water content of the polymer membrane contacting with ice also decreases. Fig.1 shows change in resistance and temperature of the polymer electrolyte membrane. Supercooling release occurs after around 70 min at –7 °C, and the temperature slightly rises due to heat of solidification. Here, resistance increases even after the temperature becomes again –7 °C. This resistance increase suggests the decrease in the water content of the polymer electrolyte membrane due to ice formation. ΔR is the difference between the resistance contacting with ice and with supercooled water. Fig.2 shows relationship of ΔR and supercooling degree (ΔT). ΔR increases as ΔT increases, and this supports the above estimation that ice freezing from the supercooled state takes up water from the ionomer. Secondly, we introduced a hydrophilic MPL made by the gas diffusion electrode (GDE) method (3) to prevent water from freezing at the interface between the CL and MPL and to improve the cold start characteristics. In the previous study (1)(2), we observed the ice layer at the interface with a hydrophobic MPL. Fig.3 shows change in cell voltage and resistance at 0.02 A/cm2, –10 °C. This result suggests that hydrophilic MPL increases the startup period than conventional hydrophobic MPL. Fig.4 is observation results of cross-section of the cathode MEA with hydrophilic MPL after the 0.02 A/cm2shutdown. Figs.4(a) and (b) are the cryo-SEM images with low magnification and with high magnification. Fig.4(a) shows that there is no ice layer at the interface and water freezes in the MPL. Fig.4(b) shows that there is little ice in CL, and this suggests that the hydrophilic MPL enhances the migration of water through ionomer in the CL. These results also suggest that the hydrophilic MPL is effective to improve the characteristics at cold startup. Reference (1) Y. Tabe, et al. , Journal of Power Sources208 (2012), 366. (2) Y. Ishima, et al. , National Symposium on Power and Energy Systems(2015), 101. (3) Y. Aoyama, et al. , ECS Trans. 69 (17) (2015), 743. Figure 1
Read full abstract