Introduction
Pt-shell/Pt-M-core (M = Fe, Co, Ni, etc.) nanoparticles have been intensively studied as oxygen reduction reaction (ORR) catalysts for polymer electrolyte fuel cells (PEFCs) [1]. However, the M elements easily dissolute into the electrolyte under PEFC’s operating conditions (strong acid and electrochemical potential fluctuations) and, thus, electrochemically-stable core materials are highly desirable to develop future core-shell type ORR catalysts. Because SiC shows high corrosion resistance to acidic solution, SiC is one of the candidates for the core materials. However, because of large lattice mismatch between Pt (0.277nm) and SiC (0.177nm), Pt/SiC interface seems to be unstable. Norimatsu and Kusunoki reported the formation of graphene layers on SiC(0001) substrate surface by high temperature annealing (ca. 1500°C) in vacuum through sublimation of the Si atoms [2]. Because lattice parameter of graphene is 0.246 nm, graphene would work as a buffer layer between Pt-shell and SiC substrate. In this study, we try to fabricate Pt/graphene/SiC(0001) model catalysts in ultra-high vacuum (UHV;~10^-8Pa) and investigate the catalyst’s ORR properties.

Experimental
The Pt/graphene/SiC model catalysts fabrication are conducted in UHV. A nitrogen-doped 4H-SiC (0001) substrate (Nippon Steel and Sumitomo Metal Corporation) was annealed at 1600°C by direct electronic heating in UHV for 1 hour to form graphene layers on the SiC(0001) substrate. Formation of the surface graphene layers was verified with low energy electron diffraction (LEED) and Raman spectroscopy. Then, 1.5nm-thick-Pt was deposited onto the graphene/SiC(0001) substrate by using electron beam evaporation method at a room temperature (RT) (as-prepared catalyst), followed by thermal annealing (annealed catalyst). The UHV-fabricated Pt/graphene/SiC(0001) model catalyst surfaces were observed by scanning tunneling microscopy (STM) in UHV and were transferred to an electrochemical apparatus set in a N₂-purged glove box without air exposure. CV and LSV measurements were conducted in N₂-purged and O₂-saturated 0.1 M HClO₄, respectively. The ORR activity was estimated based on j_k values at 0.9 V vs. RHE by using Koutecky-Levich equation.

Results and Discussion
Fig. 1(a) shows Raman spectra of the SiC(0001) substrate before and after the UHV-annealing at 1600°C. Before the UHV-annealing, two bands (1525 and 1695 cm^-1) ascribable to the Si-C lattice dominate the spectrum. In contrast, the annealed substrate shows 2D (2720 cm^-1) and G (1585 cm^-1) bands due to graphene layers, accompanied with remarkable band intensity reductions of the Si-C. The LEED pattern collected after the UHV-annealing (not shown) clearly showed sub-spots caused by graphene layers formation, exhibiting that graphene/SiC(0001) substrate can be generated after the UHV-annealing. Fig. 1(b) shows CVs of the as-prepared (blue) and 400°C-annealed (red) 1.5nm-thick-Pt deposited graphene/SiC(0001) model catalysts. For comparison, CV for the UHV-cleaned Pt(111) (black) is presented in the Fig.1(b). Although both of the Pt/graphene/SiC(0001) model catalysts show adsorption and desorption currents due to hydrogen (0.05 ~ 0.35 V) and oxygen-related-species (0.6 ~ 1.0 V), the CV shapes are clearly different from the clean Pt(111). Especially, double-layer currents for the 400°C-annealed catalyst is much larger than the as-prepared catalyst and clean Pt(111). The results suggest that the graphene/SiC(0001) surface is not completely covered by the deposited Pt for the 400°C-annealing. As for the ORR activity (Fig. 1 (c)), the as-prepared catalyst shows ca. 10 % increase in j_k values in comparison to that for the clean Pt(111). In contrast, the j_k value for 400°C-model catalyst is 60% lower. The UHV-STM image for the 400°C-annealed catalyst (Fig. 1 (d)) shows that the deposited Pt is aggregated to form hexagonal islands with ca. 20 nm in size. In contrast, the STM image of the as-prepared catalyst shows atomically-rough surface with fine islands (less than 5 nm; Fig. 1(e)). The CVs and STM results suggest that the lower ORR activity for the 400°C-annealed catalyst stems from decrease in electrochemical surface area of the deposited Pt on the graphene/SiC(0001). In contrast, the 10%-activity-increase against clean Pt(111) of the as-prepared catalyst might correlate with surface defects (e.g., (110) steps) that would enhance the ORR activity of Pt(111) epitaxial surface [3].

Acknowledgement
This study was supported by new energy and industrial technology development organization (NEDO) of Japan.

References
[1] P. Strasser et al., Nano Lett., 12 (2012) 5429.
[2] W. Norimatsu and M. Kusunoki, Chemical physical letters. 468, 52 (2009).
[3] Y. Iijima et al., J. Electroanal. Chem., 685, 79 (2012).

Fig.1 (a) Raman spectra of the SiC(0001) substrate collected before and after UHV-annealing (1600°C,1hr). CV (b) and j_k values at 0.9V (c) for the the as-prepared RT- (blue) and 400°C-annealed (red) model catalysts. For comparison, CV and j_k value for clean Pt(111) are presented by black. UHV-STM images of the 400°C- (d) and RT- (e) model catalysts.