Evaluation of "Intrinsic" Oxygen Reduction Reaction Selectivity on Carbon-Supported Hafnium Oxynitride Catalysts

Non- platinum-group-metal (PGM) catalysts have been extensively developed for use in cathodes of many energy-conversion devices such as fuel cells and metal-air batteries. Some of them showed oxygen reduction reaction (ORR) activities higher than that of platinum in alkaline media;¹ however, none of them showed such high activity in acidic environments. Therefore, most researchers developing non-PGM ORR catalysts for use in polymer electrolyte fuel cell (PEFC) cathodes, which operate in highly acidic environments, have focused on improving their activity. Selectivity for ORR, i.e., the number of electrons transferred per unit oxygen molecule, n, is another important factor that should be considered for durability. When ORR proceeded via the so-called 2-electron pathway wherein an oxygen molecule is first reduced to hydrogen peroxide (O₂ + 2H⁺ + 2e^– → H₂O₂), the H₂O₂ intermediate degraded the proton-conductive perfluorosulfonate-ionomer (PFSI) in catalyst layers and PFSI membranes.² Thus, non-PGM catalysts on which ORR proceeded via a 4-electron pathway wherein an oxygen molecule was directly reduced to water (O₂ + 4H⁺ + 4e^– → H₂O) was necessary for the long-term operation of inexpensive PEFCs.

In this study, the amount of H₂O₂ formed on carbon-supported hafnium oxynitride (HfO_xN_y-C) catalyst³ during ORR was successfully suppressed without a decrease in activity by simply changing the synthesis conditions. A decrease in the NH₃ treatment time from 50 to 6 h and NH₃-flow rate from 200 to 100 sccm at 1223 K resulted in a decrease in H₂O₂formation by a factor of 3 at 0.6 V versus a standard hydrogen electrode as shown in Figure 1.

The “apparent, not intrinsic” selectivity as well as activity depended on the mass fraction of PFSI, Nafion in catalyst layers, χ_N as shown in Figure 2. The n value was almost constant when χ_N was increased up to 0.4, and significantly decreased when χ_N was increased further to 0.5. The Koutecky-Levich plots of the catalyst layers for two different χ_N are shown in the inset of Figure 2 (b). The intercept of the y–axis at ω → ∞ was almost zero for catalyst layers with χ_N = 0.3, indicating that the ORR process was not controlled kinetically nor the other factors at the low disk potential, E_d of 0.1 V versus SHE, as expected. However, the intercept for catalyst layers with χ_N = 0.5 was much larger than zero under identical conditions, suggesting the appearance of Nafion film on the catalyst surface due to the large χ_N. Such Nafion film should be a barrier to both O₂ and H₂O₂ transport through catalyst layers, which affect n. These results indicated that both the activity and selectivity of non-PGM catalysts should be evaluated after the optimization of χ_Nfor the accuracy.

References

(1)           H. T. Chung et al., Nature Commun., 4, 1922 (2013).

(2)           S. Hommura et al., J. Electrochem. Soc., 155, A29 (2008).

(3)           M. Chisaka et al., J. Phys. Chem. C, 115, 20610 (2011).

Acknowledgment

This work was partially supported by Grant-in-Aid for Young Scientists (B), 23760185, from the Japanese Ministry of Education, Culture, Sports, Science and Technology, the Adaptable and Seamless Technology Transfer Program through target-driven R&D, AS242Z00224L, from the Japan Science and Technology Agency, and a research grant from the Kao Foundation for Arts and Sciences, Japan.