Experimental Evaluation of PEFC Catalyst Layer Structure to Reduce Oxygen Transport Resistances

In polymer electrolyte fuel cells (PEFCs), it is required to reduce the amount of Pt used in the cathode catalyst layer (CL). With the CLs at lower Pt loading, it has been reported that the oxygen transport resistance drastically increases. Our previous research (1) evaluated the components of oxygen transport resistance for CLs fabricated with various structure and reported that the resistance at the Pt surface was dominant while the resistance of oxygen dissolution into the ionomer was negligible. This study focuses on CLs with low Pt loadings. We try to elucidate the appropriate CL structure with lower oxygen transport resistance based on our quantitative evaluation.

In this study, catalyst coated membranes were fabricated with various cathode structure. Three types of Pt-supported carbons (TEC10E20E, TEC10E30E, TEC10E50E) having different Pt carrying densities were used. The ratios of Pt to carbon (Pt carrying density) are approximately 20,30 and 50 wt% respectively. The ratios of ionomer to carbon were set with I/C = 0.8.

To evaluate the total oxygen transport resistance and its components, we applied the limiting current method (2). Here, I-V measurements were conducted at different total pressures (0.11MPa, 0.12MPa, 0.13MPa, 0.14MPa, 0.15MPa) for the cathode side. The flow rate of cathode gas (oxygen and nitrogen mixture) was 2000 sccm with different oxygen concentration (1%, 1.5%, 2%), and the flow rate of anode hydrogen was 100 sccm. The relative humidity of supplied gas was set 80% (80°C) in both sides.

Table 1 shows measured parameters of CLs fabricated with three Pt carrying densities. The sample name represents the Pt carrying densities: CL20, CL30, CL50 used TEC10E20E, TEC10E30E, and TEC10E50E. The porosities measured by nitrogen adsorption method are similar as well as the pore size distributions (not shown here), and this means that the structures of the fabricated CLs are also similar. Using the porosity and the materials used for fabricating, the thickness of the CLs was estimated in Table 1. The thickness of the fabricated CLs becomes thicker with lower Pt carrying density carbon to set the Pt loading around 0.06 mg/cm². Effective Pt surface area is calculated by multiplying the Pt loading and the electrochemical surface area (ECSA [m²/g_Pt]) estimated by CV measurements. The ECSA decreased as the Pt carrying density increased. This can be considered because the Pt particles aggregate on the carbon and the surface area becomes smaller. As a result, the effective Pt surface area of CLs fabricated with three Pt carrying densities are around 64 m²_pt/m²_CL for three CLs in Table 1.

Figure 1 shows results of the I-V characteristics with the fabricated CLs at 1% and 2% oxygen concentration. The difference is pronounced with O₂ 2%: the voltage with CL30 is higher than those with CL50 and CL20 especially at higher current densities. With O₂ 1%, CL20 performance is the lowest even with the largest effective Pt surface area as shown in Table 1.

To consider the contribution of the oxygen transport resistances to the performance differences, the oxygen transport resistance components were comparted in Figure 2. R_O2^P,dep, R_O2^P,ind are the pressure dependent and independent resistances respectively. The R_O2^P,dep is considered to be the oxygen transport resistance in the GDL and the channel. The R_O2^P,ind is considered to be the oxygen resistance in the CL, and the difference in the limiting current density in Figure 1 is caused by the R_O2^P,ind with different CLs.

The previous study by the authors (2) showed that the R_O2^P,ind can be evaluated by dividing into three components: R_O2^P,ind = R_CL = R_Pt + R_diss + R_pore. These are the oxygen transport resistance at the Pt surface, R_Pt, the dissolution resistance into ionomer, R_diss, and the oxygen diffusion resistance in the CL pores, R_pore. From the simplified evaluation formula (2), the R_Pt is inversely proportional to the effective Pt surface area. Here, the Pt surface areas of the CLs are similar, as in Table 1. The R_pore is proportional to the CL thickness, and the thickest thickness of CL20 may contribute to the largest R_O2^P,ind in Figure 2. However, the R_O2^P,ind of CL50 is also larger than that of CL30 in spite of the thinnest thickness and the similar effective Pt surface area. These indicate that the dissolution resistance, R_diss, could contribute to the R_O2^P,ind for the CLs with the lower Pt loadings. The dissolution resistance, R_diss, is inversely proportional to the CL thickness, and may contribute the larger R_O2^P,ind of CL50. Therefore, it is necessary to consider the balance of these three components of the oxygen transport resistance.

Figure 1