Durability of Polymer Electrolyte Membrane Fuel Cells Operated at Subfreezing Temperatures

Abstract

The structure, composition, and interfaces of membrane electrode assemblies (MEA) and gas-diffusion layers (GDLs) have a significant effect on the performance of single-proton-exchange-membrane (PEM) fuel cells operated isothermally at subfreezing temperatures. During isothermal constant-current operation at subfreezing temperatures, water forming at the cathode initially hydrates the membrane, then forms ice in the catalyst layer and/or GDL. This ice formation results in a gradual decay in voltage. High-frequency resistance initially decreases due to an increase in membrane water content and then increases over time as the contact resistance increases. The water/ice holding capacity of a fuel cell decreases with decreasing subfreezing temperature (−10 • C vs. −20 • C vs. −30 • C) and increasing current density (0.02 A cm −2 vs. 0.04 A cm −2 ). Ice formation monitored using in-situ highresolution neutron radiography indicated that the ice was concentrated near the cathode catalyst layer at low operating temperatures (≈−20 • C) and high current densities (0.04 A cm −2 ). Significant ice formation was also observed in the GDLs at higher subfreezing temperatures (≈−10 • C) and lower current densities (0.02 A cm −2 ). These results are in good agreement with the long-term durability observations that show more severe degradation at lower temperatures (−20 • C and −30 • C).

Freezing Site of Super-Cooled Water and Failure Mechanism of Cold Start of PEFC

“…The fitting results shown in Figure 2d reveal that the ohmic resistance and R ct remains nearly constant, while R mt increases 20 times rapidly in the latter stage. This observation is consistent with Mukundan et al 15,16 The dynamic PEIS (H 2 /N 2 ) results during cold start are shown in Figure 3. The catalyst layer behaves as a blocking porous electrode at the voltage higher than 0.1 V and a porous electrode with faradaic reactions at the OCV.…”

“…The last piece of CV was measured after ice formation, which is evidenced by a sharp potential drop. The five CV curves almost coincide with each other and the ECSA values vary within 4%, indicating that ice formation does not reduce the TPI, which is consistent with Mukundan et al 15,16 but contradicts with Ge et al 14 This discrepancy may be ascribed to the fact that different samples were used in these experiments. Ge et al 14 employed a homemade CCM incorporating an 18 μm Gore-Select membrane while Mukundan et al 15,16 employed a Gore PRIMEA 57 series MEA.…”

“…In this regard, traditional electrochemical characterization methods, such as cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS), are, currently at least, indispensable. [11][12][13][14][15][16] However, CV and EIS results during cold start are often open to dispute. Mukundan et al reported that the electrochemically active surface area (ECSA) obtained from the CV remains constant during cold start, 15,16 while Ge et al observed a decreasing trend.…”

“…[11][12][13][14][15][16] However, CV and EIS results during cold start are often open to dispute. Mukundan et al reported that the electrochemically active surface area (ECSA) obtained from the CV remains constant during cold start, 15,16 while Ge et al observed a decreasing trend. 14 Discrepancy also exist in the charge-transfer resistance (R ct ) and the mass-transfer resistance (R mt ) fitted from EIS data.…”

“…14 Discrepancy also exist in the charge-transfer resistance (R ct ) and the mass-transfer resistance (R mt ) fitted from EIS data. Oszcipok et al 13 attributed the start-up failure to the sharp increase of R ct , while Mukundan et al 15,16 considered the sharp increase of R mt as the culprit. In terms of modeling activities, Mao et al modelled the freezing process by considering the effect of ice formation on both ECSA and gas diffusion coefficient, concluding that accumulation of ice in pores decreases the voltage.…”