Modeling Temperature- and Humidity-Dependent Kinetics of Oxygen Reduction Reaction

Abstract

Designing active catalysts for oxygen reduction reaction (ORR) is one of the major challenges for polymer electrolyte fuel cells. Density functional theory (DFT) calculations have made significant contributions on understanding ORR pathways, rate-limiting step and catalytic activity on various transition metal surfaces 1-5. However, all these theoretical studies are for reactions at room temperature in aqueous liquid electrolytes. The condition is significantly different from actual operation conditions of fuel cells [60 to 80 °C and 30 to 100 % relative humidity (RH)]. Although the DFT models reasonably explain experimental results obtained by the rotating disk electrode (RDE) at room temperature in aqueous liquid electrolytes, the reaction mechanism and catalytic activity at the operating conditions can be significantly different. Here, we examine effects of temperature and RH on the ORR by using a micro-kinetic model. The examination indicates that both the rate-limiting step and material-dependence of the catalytic activity are strongly influenced by both temperature and RH.The theoretical analyses were carried out by extending a micro-kinetic model proposed by Hansen et al5 to higher temperature and lower RH conditions. For the extension, following assumptions were adopted. Free energies for reaction intermediates (adsorbed O, OH and HO2) are independent of temperature and relative humidity (RH) and fixed at the values in the original model.An additional reaction barrier due to the solvent reorganization, which was introduced to explain the experimentally observed prefactor in the original model, is entirely attributed to be enthalpic. The assumption (1) can be partially validated by relatively small entropies of the adsorbed species comparing with gaseous species. Because the assumption (2) is not verified, we also carried out the same simulation assuming that the solvent reorganization barrier is fully entropic [we define this assumption as (3)].Figures 1 (a) and (c) show the calculated ORR current density at 0.9V vs. reversible hydrogen electrode (RHE) under 1 atm O2 under the assumption (2) as a function of the binding energy of OH adsorbate vs. that of Pt (111). Figures 1 (b) and (d) show the calculated ORR current density in the same way as Figures 1 (a) and (c) under the assumption (3). In all conditions, in the right leg of the volcano-plot, where the OH binding energy is larger than the optimal point, the reaction is limited by the formation of the reaction intermediates. On the other hand, in the left leg of the volcano, the reaction is limited by the removal of the OH adsorbate. Regardless the type of the assumption [(2) or (3)], both the optimal catalytic activity and the optimal OH binding energy decrease with the rise in the temperature from 25 to 80 °C [see Figs. 1 (a) and (b)]. The low catalytic activity at high temperature is attributed to the drop in the reversible potential of ORR at high temperature. The shift in the optimal OH binding energy is attributed to the destabilization of OH adsorbate relative to the non-adsorbed species, which hold larger entropies than the adsorbates. The drop in RH enhances the optimal catalytic activity, while the opposite shift is observed for the optimal OH binding energy [see Figs. 1 (c) and (d)]. The both trends stem from the lowering of the water activity, which promotes the OH removal from the surface. All in all, our theoretical analyses indicate that the optimal OH binding energy strongly depends on both temperature and RH, indicating that the optimal catalyst at low temperature and in aqueous liquid electrolytes may not be optimal anymore at high temperature and low RH. Although our theoretical analyses strongly rely on non-verified assumptions (1)-(3) and the DFT model adopting the ice-like water,1 which is unlikely a proper model of the electrolyte at high temperature and low RH, the observation indicates the significance of more precise understanding of temperature-, RH- and material-dependence of ORR.In the presentation, we will present experimental data on temperature- and RH-dependent ORR activity and discuss the validity of the theoretical model. References J. K. Nørskov et al., J. Phys. Chem. B, 108, 17886 (2004).J. Greeley et al., Nat. Chem., 1, 552 (2009).J. X. Wang et al., J. Phs. Chem. A, 111, 12702 (2007).R. Jinnouch et al., Phys. Chem. Chem. Phys., 13, 21070 (2011).H. A. Hansen et al., J. Phys. Chem. C, 118, 6706 (2014). Figure 1