5-1. Introduction
To maintain high-performance with the minimum amount of Pt loading at the PEFC cathode, the understanding of the degradation mechanism of Pt nanoparticles highly dispersed on carbon (Pt/C) during operation is very important. It has been reported that the degradation of Pt/C cathode catalysts under steady-state or potential cycling conditions was ascribed to the significant loss of electrochemical active surface area (ECSA) of Pt nanoparticles, due to a detachment from the support, to agglomeration, and/or to Ostwald ripening (dissolution of small particles followed by redeposition on larger particles) especially at low pH,1 high potential,2 high oxygen concentration,3 and high temperature.1,4
For the start-stop cycles or fuel starvation in fuel cell vehicles (FCVs), the cathode potential was found to raise as high as 1.5 V (in the worst case) by a reverse-current mechanism.5,6 The corrosion reaction of the carbon support occurs at such a high potential,
C + 2H2O → CO2 + 4 H++ 4e- E° = +0.207 V vs. SHE
The corrosion reaction rate of carbon is generally slow, but Pt plays an important role in accelerating the corrosion rate. 3,7-10 Thus, Pt particles are easily agglomerated and/or detached from the surface. 11-17 The accelerated degradation testing (ADT) of the carbon support of Pt/C has been performed using standard potential protocols recommended by the Fuel Cell Commercialization Conference of Japan (FCCJ); E = 0.9 V ↔ 1.3 V vs. RHE, holding 30 s at each E, 1 min for one cycle (protocol 2007),18 or E = 1.0 V ↔ 1.5 V vs. RHE, with the sweep rate of 0.5 V s−1, 2 s for one cycle (protocol 2011).19 However, it has been found by scanning electron microscopy (SEM) that not only the carbon support but also Pt particles were degraded after such ADTs.20 It has also been found that graphitized carbon black (GCB) exhibits higher resistance against the corrosion than conventional carbon black (CB).
-72-
It is indeed necessary to analyze the degradation of Pt nanoparticles and carbon support separately by an in-situ quantitative technique during the ADT. This certainly provides new insights for understanding the degradation mechanism of the cathode catalysts.
In this chapter, I examine the degradation of Pt/CB and Pt/GCB by using a new EQCM system, which has long-term stability and with the potential-time sequence for the ADT recommended by the FCCJ.
5-2. Experimental
5-2-1. EQCM cell and electrode preparation
A new EQCM cell was designed to be immersed in a temperature-controlled water bath. The calibration of the new EQCM cell was performed by a galvanostatic electrodeposition of Ag on a smooth Au working electrode in He-purged 0.1 M HClO4 containing 40 mM AgClO4 at 50 oC, based on our EQCM studies.21,22 The mass sensitivity thus calibrated was 4.5 ng cm−2 Hz−1, which agrees well with the calculated one based on Sauerbrey’s equation.23
Besides conventional Pt/CB (TEC10E50E, 47 wt%-Pt, Tanaka Kikinzoku Kogyo), a commercial Pt catalyst supported on graphitized carbon black (Pt/GCB, TEC10EA50E, 46 wt%-Pt, Tanaka Kikinzoku Kogyo) was employed. The working electrode consisted of a very thin layer of the catalyst uniformly dispersed on Au electrode (on quartz crystal) etched by 0.1 M HCl at a constant loading of carbon support 11 µg-C cm-2，corresponding to two layers of carbon black.
5-2-2. Protocol of the ADT of Pt/CB and Pt/GCB
The protocol of the accelerated degradation test (ADT) of the Pt/CB catalyst consisted of four steps. In the first step, the electrode surface was stabilized by repeated cycling (50 cycles) of the potential between 0.05 and 1.0 V vs. RHE in 0.01 M HClO4 at 60 oC. In step 2, the cyclic voltammogram was taken at 100 mV s−1 between 0.05 and 1.0 V in 0.01 M HClO4 at 50oC. In step 3, the ADT was performed by the potential sweep protocol (E = 1.0 V ↔ 1.5 V vs. RHE, with the sweep rate of 0.5 V s−1, 2 s for one cycle) simulating start-stop cycles of FCVs (FCCJ protocol 2011) at 50 oC, as shown in Figure 5-1.19 After that, the cyclic voltammogram was measured by using the step 2 condition again (step 4).
The amount of Pt dissolved in the electrolyte solution was analyzed by
-73-
inductively coupled plasma–mass spectrometry (ICP-MS, Agilent Technologies 7500 series).
2 s/cycle Open circuit
Hold at
initial potential (1.0 V)
Figure 5-1. The potential sweep protocol simulating start-stop of FCVs proposed by the FCCJ (2011).
-74- 5-3. Results and discussion
5-3-1. Degradation of Pt/GCB and Pt/CB by potential cycles
Figure 5-2 shows the CVs at the Pt/GCB and Pt/CB electrodes in He-purged 0.01 M HClO4 solution before and after the ADT for (A) 1500 and (B) 2000 cycles. The values of ECSA at Pt/GCB and Pt/CB are shown in Table 5-1. The ECSA at both electrodes hardly changed after the ADT. However, the current in the double layer region (0.4-0.6 V) increased as a result of the ADT, especially at Pt/CB. This is certainly ascribed to the formation of various oxygen-containing functional groups on the surface of the carbon.24,25 Figure 5-3 shows the current and simultaneous changes in mass ∆m at Pt/GCB and Pt/CB electrodes during the ADT.The ∆m at Pt/GCB was found to be negligibly small during 1500 cycles, indicating a high durability of Pt/GCB at high potentials. In contrast, the mass at Pt/CB decreased continuously with increasing number of potential cycles N. Because the amount of Pt dissolved in the solution, analyzed by ICP-MS, was found to be less than 5 % of the Pt used, the ∆m observed at Pt/CB was mainly ascribed to the degradation of the CB. Since the anodic oxidation charge during the ADT was larger than that of the cathodic reduction, the ∆m observed at Pt/CB during the ADT was mainly caused by the electrooxidation of CB.
0 0.2 0.4 0.6 0.8 1.0
-0.2 -0.1 0 0.1 0.2
Current, I / mA
Potential, E / V vs. RHE Pt/GCB
(A)
0 0.2 0.4 0.6 0.8 1.0
-0.2 -0.1 0 0.1 0.2
Current, I / mA
Potential, E / V vs. RHE Pt/CB
(B)
Figure 5-2 Cyclic voltammograms at Pt/GCB (A) and Pt/CB (B) electrodes observed at 100 mV s-1 in He-purged 0.01 M HClO4 solution at 50 oC: (A) at Pt/GCB before (-) and after (-) the ADT for 1500 cycles; and (B) at Pt/CB before (-) and after (-) the ADT for 2000 cycles.
-75-
Table 5-1 Values of ECSA at both Pt/GCB and Pt/CB before and after the ADT.
a ECSA calculated from electrical change for the hydrogen desorption wave ∆QH in each CV.
ECSAa / cm2 pristine after ADT Pt/GCB 0.76 0.77
Pt/CB 0.87 0.90
Figure 5-3 Current and ∆m of Pt/GCB (A) and Pt/CB (B) electrodes in He-purged 0.01 M HClO4 solution at 50 oC: (A) at Pt/GCB during the ADT for 1500 cycles; and (B) at Pt/CB during the ADT for 2000 cycles.
-1.0 -0.5 0 0.5 1.0
Current, I / mA
Number of potential cycles, N
0 500 1000 1500
-20 -10 0 10 20
Mass change, ΔΔΔΔm /µµµµg cm-2
-1.0 -0.5 0 0.5 1.0
Current, I / mA
Number of potential cycles, N
0 1000 2000
-20 -10 0 10 20
Mass change, ΔΔΔΔm /µµµµg cm-2
(A) (B)
-76-
5-3-2. Effect of lower limit potential on degradation of Pt/CB
To estimate the change in ECSA during the ADT, the CVs between 0.05 and 1.0 V to obtain the hydrogen adsorption charge have been conducted intermittently after a certain number of cycles. Figure 5-4 shows the current and ∆m at Pt/CB electrode during the ADT for 2000 cycles without and with CV measurement. The ∆m decreased monotonically with increasing N without any CV measurement during the ADT. When the CV measurements during the ADT were performed between 0.05 and 1.0 V after 100, 300, 600, 1000, and 1500 cycles, abrupt, large decreases in ∆m were observed just after the CV measurement, in addition to a continuous decrease in ∆m during the ADT. Thus, total decrease in ∆m with CV measurements was ca. 2 times larger than that without CV.
Subsequently, theeffect of lower limit potential of CV on the degradation of Pt/CB was examined. As illustrated in Fig. 5-5, the lower limit potentials EL during CV measurement were chosen to be 0.30 V and 0.40 V, which correspond to the onset potentials of hydrogen adsorption and of the double layer region, respectively. Figure 5-6 indicates the ∆m of Pt/CB electrode during the ADT with intermittent CV measurements, with the lower limit potentials of 0.05, 0.30, and 0.40 V. With EL = 0.30 V, the decrease in ∆m just after the CV was mitigated, compared with the case of EL = 0.05 V. In contrast, with EL = 0.40 V, the decrease in ∆m just after CV was negligibly small, and the change in ∆m with N was nearly identical with that without any CV measurement. Thus, the carbon degradation was certainly accelerated when Pt/CB was rapidly reduced at the hydrogen adsorption potential region just after the oxidation at high potentials of ADT. This suggests that the lower limit potential for the PEFC cathode must be larger than 0.4 V to avoid the degradation, even for high current density operation such as in FCVs.
-77-
Figure 5-5 Potential-time sequence including ADT and regular CVs.
Figure 5-4. Current and ∆m at Pt/CB electrode with CV (-) and without CV (-) during the ADT for 2000 cycles in He-purged 0.01 M HClO4 solution at 50 oC.
2 s/cycle
1.0 V 1.5 V
P o te n ti a l
0.05 V
Lower limit potential