Development of an Oxygen Mass Transport Coefficient Measurement and Separation Method for Proton Exchange Membrane Fuel Cells

Abstract

The cost and durability of proton exchange membrane fuel cells (PEMFCs) still limit their adoption as sustainable power systems [1]. Operation at high current densities, which decreases the inventory of costly materials [2], is still considered an option [3] in addition to concurrent reductions in Pt catalyst loadings. However, these approaches intensify mass transfer losses especially at the cathode because the oxidant stream is more dilute (21 % O2 in air) than the fuel (100 % H2) and the heavier O2 molecule moves more slowly. The improvement of characterization methods able to measure and separate oxygen mass transfer coefficients is desirable for more accurate identification and localization of resistances in membrane/electrode assemblies (MEAs). Average limiting currents iave are first measured with a dilute oxidant stream for different flow rates. Subsequently, data are fitted to a mathematical model [4] to obtain the overall mass transfer coefficient k: iave =ie (1–exp(–nFkpr /RTief)) (1) with ie the inlet reactant flow rate equivalent current density, n the number of electrons exchanged in the electrochemical reaction, F the Faraday constant, pr the dry inlet reactant stream pressure, R the ideal gas constant, T the temperature, and f the inert gas to reactant fraction in the dry inlet reactant stream. The oxidant diluent is also varied (He, N2, CO2) to separate the molecular diffusion mass transfer coefficient km by extrapolating to the origin the linear correlation between the overall mass transfer resistance and the diluent molecular mass M [4]: k –1=km (M)–1+k e+K (c)–1=km (M)–1+ke (c)–1+kK –1 (2) The ionomer permeability mass transfer coefficient contribution ke was separated from 1/k e+K by taking advantage of the constant Knudsen diffusion resistance (equation 2 above and equation 7 in [5]). For that last step, the linear correlation between the lumped ionomer permeability and Knudsen diffusion mass transfer resistance 1/k e+K and the oxygen concentration c (1 to 7 %) was extrapolated to the origin, which yielded the Knudsen diffusion mass transfer coefficient kK . This approach to isolate the Knudsen contribution is novel. Other techniques are used, including temperature variations [5]. Experiments were conducted with a 50 cm2 active area cell and General Motors MEAs with a commercially relevant cathode catalyst loading of 0.05 to 0.15 mg Pt cm–2 (Pt catalyst supported on a high surface area carbon). All MEA tests were duplicated and completed for several temperatures (30 to 80 °C) and relative humidities (50 to 100 %). Four diagnostic methods were employed to supplement the analysis of mass transfer data. The use of O2, 21 % O2 in He, and air enabled the derivation of overpotentials from polarization curves [6]. The high frequency cell resistance was measured with a milliohmmeter. The catalyst area was extracted from the hydrogen adsorption region of cyclic voltammograms [7]. The overall cell water balance was also calculated to assess the presence of liquid water [8]. The smallest mass transfer resistance was assigned to the ionomer permeability, which supported the need for repetitive measurements and statistics for an accurate quantification. The slope of the linear correlation between the ionomer permeability and Knudsen diffusion mass transfer resistance and the oxygen concentration was dependent on the water balance. A positive slope occurred under sub-saturated streams whereas a negative slope was correlated with the wettest operating conditions, which will require a method modification to extract the Knudsen diffusion resistance. A plot of the ionomer and molecular diffusion mass transfer resistance versus the total mass transfer overpotential showed two regimes that are differentiated by the absence/presence of liquid water (see figure). Catalyst layers, only several microns thick, have a larger mass transfer resistance and overpotential under flooding conditions. [1] Y. Wang, D. F. R. Diaz, K. S. Chen, Z. Wang, X. C. Adroher, Mater. Today, 32 (2020) 178. [2] A. Kongkanand, M. F. Mathias, J. Phys. Chem. Lett., 7 (2016) 1127. [3] N. Ramaswamy, W. Gu, J. M. Ziegelbauer, S. Kumaraguru, J. Electrochem. Soc., 167 (2020) article 064515. [4] T. Reshetenko, J. St-Pierre, J. Electrochem. Soc., 161 (2014) F1089. [5] N. Nonoyama, S. Okazaki, A. Z. Weber, Y. Ikogi, T. Yoshida, J. Electrochem. Soc., 158 (2011) B416. [6] J. St-Pierre, M. Angelo, K. Bethune, J. Huizingh, T. Reshetenko, M. Virji, Y. Zhai, Modern Fuel Cell Testing Laboratory, in Springer Handbook of Electrochemical Energy, Part D, Chapter 19, Edited by C. Breitkopf, K. Swider-Lyons, Springer, 2017, p. 611. [7] R. N. Carter, S. S. Kocha, F. T. Wagner, M. Fay, H. A. Gasteiger, ECS Trans., 11(1) (2007) 403. [8] J. St-Pierre, N. Jia, M. van der Geest, A. Atbi, H. R. Haas, United States patent 7,132,179, November 7, 2006. Figure 1