For polymer electrolyte fuel cells, Pt-based alloy catalysts have been widely studied. We have developed the nanocapsule method^1,2 for the synthesis of PtCo alloy nanoparticles (n-PtCo) with well-controlled size and composition. On n-PtCo subsequently treated in hydrogen gas at high temperature (n-PtCo_H2-HT), additional Pt-skin layers were formed (Pt_xAL-PtCo).³

Figure 1 shows the area-specific activities j_k for the oxygen reduction reaction (ORR) at carbon-supported n-PtCo_H2-HT and Pt_xAL-PtCo (n-PtCo_H2-HT/C and Pt_xAL-PtCo/C, respectively) and a commercial carbon-supported Pt catalyst (c-Pt/C) coated with Nafion in O₂-saturated 0.1 M HClO₄ solution at 65 ^oC and 0.85 V vs. RHE measured by a multi-channel flow double-electrode cell technique.² Potential-step cycles were applied to the catalyst electrodes as the accelerated durability tests (ADTs) simulating load-change cycles of fuel cell vehicles⁴ in N₂-saturated solution with no solution flow. Both the initial j_k and the durability were the highest at Pt_2AL-PtCo/C.³

Figure 2 shows changes in the electrochemically active areas (ECAs) of n-PtCo_H2-HT/C and Pt_xAL-PtCo/C during the ADTs in a constant solution flow at 65 °C so that the re-deposition of Pt became negligible. The ECA values of the catalysts decreased as the number of potential-step cycles increased, and the losses of the ECA of n-PtCo/C and Pt_xAL-PtCo/C after 5000 cycles were 42% and 26%, respectively. Therefore, the ECA loss was much mitigated by the existence of Pt skins.

Figure 3 shows the in-situ X-ray diffraction (XRD) pattern of the (111) peak of Pt-Co alloy of n-PtCo/C and Pt_xAL-PtCo/C measured during the ADT in a solution flow. For n-PtCo/C, the diffraction peak at ca. 27.6° was close to that for Pt (27.5°), suggesting that Co easily dissolved from n-PtCo into the hot acid electrolyte solution. The peak was small and broad even before the ADT, probably because of a low order of atoms in crystallite. In the XRD pattern of Pt_xAL-PtCo/C, a significantly large, distinct peak was observed before the potential cycles; the ordering of atoms in Pt_xAL-PtCo was very high. The peak was almost unchanged after the potential-step cycles, indicating that the crystallite structures of Pt_xAL-PtCo remained the same. The parameters of the Pt-Co alloy catalysts from the data obtained by the electrochemical measurements, transmission electron microscopy, and in-situ XRD are listed in Table 1. For n-PtCo/C, the crystallite size was approximately 2.3 nm during 5000 cycles. The number of crystallites per area and the metal retention ratio gradually decreased from 14 × 10³ to 10 × 10³ mm^-2 and from 100% to 50%, respectively. Not only the dissolution but the migration/coalescence also played a role. For Pt_xAL-PtCo/C, the crystallite size was approximately 2.9 nm during 5000 cycles. No change was indicated in the shape and the number of the crystallites per area, as well as the loss of the metals. Interestingly, the ECA of Pt_xAL-PtCo/C gradually decreased during 5000 cycles by 26%. Therefore, it is assumed that the Pt_xAL-PtCo crystallites were aggregated into a larger particle consisting of multi-crystallites, while keeping the shape of each crystallite unchanged. The loss of ECA is explained by the loss of surface areas by a contact between crystallites. As shown in Fig. 1, the j_k value at Pt_xAL-PtCo/C remained the same level during the potential-step cycles. The coalescence of alloy crystallites without any dissolution of Pt and Co clearly explains this constant value of j_k. Analyzing an EXAFS of Pt in Pt_xAL-PtCo/C, the Debye-Waller factor for Pt–Pt in Pt_xAL-PtCo/C was even smaller than that of Pt nanoparticles with a similar size, revealing that the crystallinity of the outer Pt skin was extremely high.⁶ Thus, the existence of a rigid Pt skin realized the extensively high durability of Pt_xAL-PtCo/C. Further structural analyses of the Pt-Co catalyst at the atomic level are now in progress.

Acknowledgements

This work was carried out under the “SPer-FC” project, NEDO, Japan. The synchrotron radiation experiments were performed at BL14B2, BL16B2, BL19B2, and BL46XU of SPring-8 (No. 2015B3388, 2016A1760, 2016A1805, 2016A5390, 2016B1550, and 2016B1856, 2016B5390).

References

1. Yano, H.; Kataoka, M.; Yamashita, H.; Uchida, H.; Watanabe M. Langmuir, 2007, 23, 6438.
2. Yano, H.; Song, J. M.; Uchida, H.; Watanabe, M. Phys. Chem. C, 2008, 112, 8372.
3. Watanabe, M.; Yano, H.; Tryk, A.; Uchida, H. J. Electrochem. Soc., 2016, 163, F455.
4. http://fccj.jp/pdf/23_01_kt.pdf in Japanese.
5. Watanabe, M.; Tryk, D. A.; Wakisaka, M.; Yano, H; Uchida, H. Acta, 2012, 84, 187.
6. Shi, G.; Yano, H; Tryk, D. A.; Matsumoto, M.; Tanida, H.; Arao, M.; Imai, H; Inukai, J.; Iiyama A.; Uchida H. Sci. Technol., 2017, 7, 6124.