3.1. Effects of Cell Temperature
After a predetermined number of cycles, a cyclic voltammetry measurement for electrochemical surface area (ECA) was conducted to determine the degradation status.
Figure 2 shows the effect of the cell temperature on the relationship between the normalized ECA ratio and the number of potential cycles in the load cycle durability test. The influence of cell temperature on the rate of ECA decrease was not so large, and was increased a little when the cell temperature was raised.
In order to calculate the amount of carbon support corrosion during the load cycle durability test, the concentration of CO
2 in the nitrogen gas exhausted from the cathode outlet was measured with a non-dispersive infrared analyzer.
Figure 3 shows the effect of the cell temperature on the relationship between the normalized carbon corrosion ratio and the number of potential cycles in the load cycle durability test. The influence of cell temperature on the rate of carbon decrease was not so large, and was increased a little when the cell temperature was raised. From
Figure 2 and
Figure 3, both ECA and carbon ratio decreased by number of potential cycles, but the decline rate of ECA was larger than that of carbon. The ECA decrease rate in the load cycle test is mainly influenced by the decrease of the Pt surface area due to increased particle size rather than the carbon support corrosion [
4,
5,
6]. The influence of cell temperature on the Pt agglomeration rate was considered not to be large under the test conditions in this study.
The durability of the Pt/C catalyst is dependent on the durability of both the carbon support and platinum, and both the degradation of platinum and carbon corrosion decrease the ECA of the Pt/C catalyst.
Figure 4 shows the relationship between the normalized ECA ratio and the normalized carbon support corrosion ratio under each cell temperature condition. The relationship was similar under each cell temperature condition, as shown in
Figure 4. Therefore, it was thought that an almost identical Pt/C degradation phenomenon occurred even if the cell temperature was different in the load cycle durability test.
The influence of cell temperature on the decline in power generation performance in the load cycle durability test was investigated by I-V measurements as diagnostics of the degradation status.
Figure 5 shows the relationship between the cell voltage at 1 A cm
−2 and the ECA in the load cycle durability test. No influence of cell temperature on the relationship was observed in the range of these examination conditions. From the viewpoint of the power generation performance, we confirmed that the same agglomeration phenomenon occurred even if the cell temperature was different in the load cycle durability test.
3.2. Effects of Humidity
Next, we investigated the influence of the humidity of the gas supplied to the cell on the ECA and carbon corrosion in the load cycle durability test.
Figure 6 shows the effect of the humidity of the gas supplied to the cell on the relationship between the normalized ECA ratio and the number of potential cycles in the load cycle durability test. Relative humidity influenced the rate of ECA decrease, which was increased when the humidity was raised.
In order to calculate the amount of carbon support corrosion during the load cycle durability test, the concentration of CO
2 in the nitrogen gas exhausted from the cathode outlet was measured with a non-dispersive infrared analyzer.
Figure 7 shows the effect of the humidity of the gas supplied to the cell on the relationship between the normalized carbon corrosion ratio and the number of potential cycles in the load cycle durability test. The carbon corrosion ratio was not influenced by the relative humidity under the test conditions in this study. It was thought that the higher the humidity, the faster the Pt agglomeration rate. The Pt dissolution rate is thought to be increased due to the large amount of water, but more detailed investigation is necessary.
Figure 8 shows the relationship between the normalized ECA ratio and the normalized carbon support corrosion ratio under each humidity condition. The relationship was similar under each humidity condition, as shown in
Figure 8. Therefore, it was thought that an almost identical Pt/C degradation phenomenon occurred even if the humidity supplied to the cell was different in the load cycle durability test.
The influence of humidity supplied to the cell on the decline in power generation performance in the load cycle durability test was investigated by I-V measurements as diagnostics of the degradation status.
Figure 9 shows the relationship between the cell voltage at 1 A cm
−2 and the ECA in the load cycle durability test. No influence of humidity on the relationship was observed in the range of these examination conditions. From the viewpoint of the power generation performance, we confirmed that the same agglomeration phenomenon occurred even if the humidity was different in the load cycle durability test.
Figure 10 shows the effect of humidity on the relationship between the normalized cell voltage at a current density of 1 A cm
−2 and the number of potential cycles. Relative humidity influenced the rate of cell voltage decrease, which was increased when the humidity was raised. The reason why the rate of cell voltage decrease increased in a smaller number of potential cycles was thought to be that the rate of ECA decrease was higher under more humid conditions.
In order to investigate in detail the difference in performance degradation for each humidity condition, the amount of performance degradation was divided by each overvoltage [
7]. During the power generation test, the cell resistance was measured with an AC resistance meter (10 kHz). The product of the resistance and the current was used as the resistance overvoltage. From the measured cell voltage, the IR-free cell voltage was calculated by the correction of the resistance polarization. The activation overvoltage was then defined as the difference between the theoretical cell voltage and the corresponding voltages of the linear extrapolation of the Tafel region (low current density region) of the IR-free polarization curve. The diffusion overvoltage was the voltage obtained by subtracting the activation overvoltage and the resistance overvoltage from the cell voltage.
Figure 11 shows the changes in the relationship between overvoltage and ECA in the load cycle durability test. ECA decreases according to the number of potential cycles. A remarkable increase was not seen in resistance overvoltage or diffusion overvoltage when ECA decreased. Regardless of the humidity, it was considered that there was no deterioration in the performance of the electrolyte membrane or ionomer due to the potential cycle test and no change in the pore structure of the catalyst layer. The activation overvoltage increased with a decline of ECA. It was believed that the agglomeration of Pt particles reduced the reaction area required for power generation. No influence of humidity was seen in the increase in activated overvoltage. These results show that the performance degradation mechanisms caused by the load cycle test were almost the same regardless of humidity conditions.
Therefore, it was confirmed that the decline of ECA with the number of potential cycles was accelerated at high humidity.