Analysis of Oxygen Transport Resistances in the Catalyst Layers with Different Carbon Supports in PEFC

Abstract

In polymer electrolyte fuel cells (PEFCs), the catalyst layer (CL) of the cathode needs a large amount of Pt because of the slow oxygen reduction reaction. Since electron, proton, and oxygen are necessary for the cathode reaction, achieving the optimum structure of electrode CL and the efficient transport of the reactants is significantly effective to reduce the usage of Pt catalyst. In this study, the effects of carbon structure on oxygen transport were evaluated using CLs fabricated under several conditions. We compared the performances using fabricated CLs with Ketjen black and Vulcan. Furthermore, the oxygen transport resistance in the CLs was evaluated. Pore size distribution (PSD) of the CL was measured by nitrogen physisorption method, and the CL structural parameters were estimated using the estimation model (1). Using the limiting current density method, we evaluate various oxygen transport resistances in the CL and discuss the contributions of these resistances to the cell performance. Figure 1 shows the estimated schematics of Pt-supported Ketjen black and Vulcan with ionomer. To investigate the structure of the fabricated CLs, we measured the PSDs by nitrogen physisorption method, and the results are shown in Figure 2. In the case of using Ketjen black, there is a small peak around 3-4 nm pore diameter, whereas in the case of Vulcan this peak does not exist. The small peak is considered to correspond to the pores in the carbon particles with a diameter of ~ 10 nm, and the difference in the PSDs for the self-produced CL is caused by the characteristics of the Vulcan with few pores in the particle. From these results, it is expected that most of the platinum particles are located on the outer surface of the Vulcan, and oxygen can easily diffuse to the platinum particles through the ionomer. Figure 3 shows the characteristics of the oxygen transport resistance of different thick CLs. By multiplying the measured oxygen transport resistance in the CL by the thickness of the CL, it is possible to evaluate only the change in the diffusion resistance in the CL pores, eliminating the influence of the oxygen transport resistance on the platinum surface and the dissolution resistance into the ionomer (1). From the gradients of the lines in Fig. 3, the effective diffusion coefficients can be estimated: DCL,eff O2=7.2×10-7 [m2/s] for the CL with Ketjen black and DO2 CL,eff =4.3×10-7 [m2/s] for that with Vulcan. The smaller effective diffusion coefficient with Vulcan is caused by the lower porosity. Figure 4 shows the characteristics of the oxygen transport resistance of the CLs having different platinum surface areas. The vertical axis represents the oxygen transport resistance in the CL, and the horizontal axis represents the inverse of the platinum surface area. Here, Pt-supported carbons with three different Pt densities were used, and the carbon amounts were set similar. Based on the gradients of the lines, the value of the oxygen transfer rate constant kPt at the platinum surface are estimated to be kPt=3.7×10-3 [m/s] for the CL with Ketjen black and kPt=9.3×10-3 [m/s] for that with Vulcan. Figures 5 and 6 show the results calculating the oxygen transport resistance of the fabricated CLs using the parameters estimated from the experiment. The oxygen dissolution rate kdiss=7.0×10-3 [m/s], estimated in the previous study (1), was applied here. With the CL with Ketjen black, the resistance at the platinum surface, R Pt , is a dominant factor, whereas the contribution of the resistance at the platinum surface with Vulcan reduces. This is considered to be because platinum particles are arranged outside the carbon particles. However, the diffusion resistance in the pores with Vulcan increases due to the lower porosity. This suggest that it is effectiveto increase the porosity for further reduction in the oxygen transport resistance in the thick CL with Vulcan. Reference Tabe, et al., ECS Trans., 80(8), 205 (2017). Figure 1