Communication—Electrochemical Stability of Pt/Pd(111) Model Core-Shell Structure in 80 ◦ C Perchloric Acid

We investigated the electrochemical stability of a Pt/Pd(111) model core–shell structure for oxygen reduction reaction catalysts in 0.1 M HClO 4 at 80 ◦ C by performing potential cycles (PCs) between 0.6 and x V vs. RHE ( x = 0.8–1.0; x V-PCs). Pristine Pt/Pd(111) shows an ORR activity four times higher than that of Pt(111), which considerably decreased after 5000 cycles of 0.9-and 1.0 V-PCs and remained nearly unchanged for the 0.8-and 0.85 V-PCs samples. These results suggest that the atomic-scale structural changes at the Pt/Pd interface that depend on PC windows determine the durability of Pd–Pt core–shell structures. © The Author(s) 2017. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0571709jes]

Pd-Pt core-shell nanoparticles (NPs), where Pt atoms are located only at the near surface regions, have been extensively studied as cathode catalysts in polymer electrolyte fuel cells (PEFCs) for the oxygen reduction reaction (ORR). [1][2][3][4] These NPs have the advantage of reducing Pt loading and enhancing the ORR activity. A major problem of core-shell ORR catalysts is durability in practical operating conditions of PEFCs, e.g., potential regions from 0.6 to 1.0 V, strongly acidic conditions, and temperatures of around 80 • C. However, to date, most durability evaluations of the catalysts are conducted with a half-cell setup in solutions at room temperature instead of 80 • C. To investigate the degradation mechanisms of actual core-shell-type ORR catalysts, it is essential to evaluate the electrochemical (EC) stabilities of core-shell model catalysts at 80 • C. We have previously discussed the relation between the Pt/Pd interface structure and ORR activity using Pt monolayers grown on a Pd(111) surface that were prepared using molecular beam epitaxy (MBE) in ultra-high vacuum (UHV). 5,6 In this communication, we focus on the EC stabilities of Pt/Pd(111) core-shell model structures in 0.1 M HClO 4 at 80 • C while performing potential cycles (PCs) with various potential windows.

Experimental
Sample fabrication procedures, surface structural analysis, and electrochemical measurements are presented elsewhere. 6 Briefly, after surface cleaning of Pd(111) in UHV, 4-monolayer (ML)-thick Pt was deposited on the Pd(111) substrate via electron beam evaporation at a substrate temperature of 300 • C. Hereinafter, the sample is referred to as Pt/Pd(111). As for the EC measurements, cyclic voltammetry (CV) curves were recorded in N 2 -purged 0.1 M HClO 4 . Subsequently, linear sweep voltammetry (LSV) measurements were performed in an O 2 -saturated solution under various disk rotation rates. The ORR activities were evaluated from the kinetic current densities j k estimated using Koutecky-Levich plots. The EC stabilities were investigated by applying square-wave PCs between 0.6 and x V (x = 0.8, 0.85, 0.9, and 1.0 V; denoted hereafter xV-PCs) for 3 s at each potential to an applied potential in the O 2 -saturated electrolyte at 80 • C.

Results and Discussion
UHV scanning tunneling microscopy (STM), reflection highenergy electron diffraction (RHEED), and low-energy ion-scattering spectroscopy (LEIS) were performed for the as-prepared Pt/Pd(111) model catalyst surface. Figures 1a and 1b show the experimental results. The RHEED pattern (inset in (a)) shows sharp streaks, with slightly larger (by ca. 0.4 %) than those expected for a Pt(111) lattice, indicating that Pt layers epitaxially grow on the Pd(111) substrate surface. The UHV-STM image (a) shows island-like structures having * Electrochemical Society Member. z E-mail: n-todoroki@material.tohoku.ac.jp hexagonal shape and approximately 20-nm-wide terraces. The LEIS spectrum for the corresponding surface ( Figure 1b) exhibits a sharp and single band stemming only from the Pt atoms located at the topmost surface of Pt/Pd(111). We use this Pt/Pd(111) as a Pt shell-layer model for actual Pd-Pt core-shell catalysts. Figure 1c shows a CV curve of Pt/Pd(111) (red); a CV curve of clean Pt(111) (dotted) is added for reference. In the OH-related region (0.6-1.0 V; enlarged in inset), Pt/Pd(111) shows a symmetrical current-potential wave that is similar to the so-called butterfly peak at 0.8 V for clean Pt(111). Compared to clean Pt(111), the redox peak of the OH adsorption reaction on Pt/Pd(111) shifts to higher potential by approximately 30 mV and the total charge for the OH adsorption and desorption reactions significantly decreases. Figure 1d shows the estimated specific ORR activities j k of the samples at 0.9 V, where Pt/Pd(111) exhibits an activity approximately four times higher than that of clean Pt(111). Figure 2 shows the LSV curves of Pt/Pd(111) recorded in solution with a disk rotation rate of 1600 rpm before and after performing 1000 cycles of 1.0V-PCs at RT and 80 • C. The curve of the 1000-cycled sample at RT (blue) is nearly identical to that of the initial one (red). In contrast, the half-wave potential of the 1000-cycled sample at 80 • C (green) notably shifts to lower potential values and is close to that of Pt(111) (dashed black line). Therefore, the solution temperature when performing PCs drastically affects the electrochemical stability of the Pt/Pd(111) sample. We then investigated the effects of the upper limit potential (ULP; x V) of PCs on the EC stabilities at 80 • C. Figure 3 shows the changes in the j k values of Pt/Pd(111) estimated at 0.9 V with different ULPs (xV-PCs) at 80 • C. The activity of the sample exposed to the 1.0V-PCs (black) steeply decreases with increasing number of PCs, eventually becoming less active than clean Pt(111) even at 1000 cycles. Potential cycle numbers when the j k values of the Pt/Pd(111) become less than that of Pt(111) are 5000 cycles at RT 6 and 1000 cycles at 80 • C, respectively. Accordingly, the activitydeterioration-acceleration-factor by temperature increment from RT to 80 • C can be estimated to ca. 5. However, the deactivation of the sample is effectively suppressed by lowering the ULPs. In particular, the activities of the 0.8V-(green) and 0.85V-PCs (red) after 5000 cycle loadings were 120 % and 90 %, respectively, compared to the corresponding initial activity. Figure 4a shows the CV curves recorded after xV-PCs at 80 • C. It is clearly evident that both the hydrogen and oxygen species-related regions grow with an increase in the ULPs (x V) of the PCs. The result shows that the PCs with higher ULPs tend to cause more roughening of the Pt/Pd(111) model catalyst surface.
To further investigate the influence of the PC-induced surface structural changes on the ORR activity, UHV-STM and LEIS were performed for the samples in solution after 5000 cycles at 80 • C. Judging from the corresponding UHV-STM images (Figure 4b), although the surface subjected to the 0.8V-PCs showed nanosized surface roughening with heights below 0.7 nm, the 0.9V-PCs resulted in agglomerated particle-like structures that were 20-nm wide and 3nm high. These results indicate that the ULPs of the square-wave PCs, i.e., surface oxidation, including oxygen or OH-related species adsorptions, strongly influence the EC stability of the Pt/Pd(111) sample. The deconvoluted LEIS spectra show that the signals are due to surface Pd atoms, irrespective of the ULPs, indicating that the topmost surfaces comprise Pt/Pd alloys instead of pure Pt. Furthermore, the surface Pd/Pt atomic ratio of the 0.9V-PCs sample is higher than that of the 0.8V-PCs sample by approximately 30 %. Therefore, the topmost surface Pd/Pt atomic ratios are strongly correlated with the ORR deactivation process during the PCs loadings, which is accompanied by an increase in the Pd/Pt atomic ratios. Nevertheless, the observed structural changes (STM and LEIS) are consistent with the increase in the charge of the CVs (Figure 4a).
To the best of our knowledge, very few studies related to the effect of the electrolyte temperature on EC stabilities of Pt-based ORR catalysts have been reported. Ohma et al. investigated the influence of potential ranges on commercial Pt/C catalysts in an accelerated durability test (ADT) and demonstrated that the electrochemical surface areas (ECSA) significantly decreased after 15k cycles of the ADT with ULPs over 0.90 V. 7 Xing et al. reported that increasing solution temperatures cause lower shifts of the onset potentials for the oxidation reaction of the polycrystalline Pt surface, resulting in larger dissolution of the Pt atoms. 8 The degradation behaviors of Pt/Pd(111) analyzed in this study seem to be more prominent than the aforementioned results of pure Pt catalysts. Accordingly, because Pd exhibits much lower activity than Pt, the activity degradations of the Pt/Pd(111) sample should be governed by atomic-level surface structural changes induced by the electrochemical redox reactions and following interdiffusion of the Pt and Pd atoms located near the surface regions. 9 Two possible origins of structural (STM) and compositional (LEIS) changes of Pt/Pd(111) near surface region can be considered. One is a decrease in the effective thickness of the Pt(111)-surface layers induced by the aggregation of surface Pt atoms during the PC loadings. Actually, Conway proposed a place-exchange reaction of oxygen species and that surface Pt atoms could cause atomic-level surface roughing of the Pt electrode. 10 Another is oxygen adsorptioninduced surface segregation of the substrate Pd atoms. It has been reported that more oxophilic metals tend to segregate to the topmost surface of Pt shell layers when oxygen adsorbs on Pt/M bimetallic alloy surfaces. 11 Hence, surface roughening of Pt shell layers as well as surface segregation of Pd atoms during the PC loadings are correlated with the deactivation of Pt/Pd(111). Once the substrate Pd atoms segregate to the topmost surface, they would immediately dissolve into the electrolyte, thereby leading to extreme activity decrease for ULPs above 0.9 V. Nonetheless, although further atomic-scale investigations are required for clarifying atomic-level degradation mechanisms of Pt/Pd(111) core-shell model structures, Pt and Pd near-surface behaviors at a given ULP determine the EC stability, i.e., the durability of the actual Pd-Pt core-shell ORR catalysts.

Summary
We investigated the EC stabilities of 4 ML-thick Pt deposited on Pd(111) model core-shell structures during PCs with various potential windows in 0.1 M HClO 4 at 80 • C. The initial ORR activity was estimated to be approximately four times greater than that of clean Pt(111). While the 1.0V-PCs Pt/Pd(111) sample was remarkably deactivated, lowering the ULPs of the PCs suppressed the activity deterioration. Specifically, 0.85-and 0.8V-PCs Pt/Pd(111) samples showed activities of approximately 90 % and 120 %, respectively, for the pristine sample activity, even after 5000 cycles. These results suggest that changes in the Pt/Pd interfacial structures during PCs determine the ORR durability of Pd-Pt core-shell catalysts.