The structure modification and activity improvement of Pd–Co/C electrocatalysts by the addition of Au for the oxygen reduction reaction

Abstract

In this study, Pd₇₅Co_25−xAu_x/C ternary catalysts with varying x content are synthesized by the deposition–precipitation approach with hydrogen reduction at 390 K for the oxygen reduction reaction (ORR). The roles of Au in the modification of structures, surface species and electrochemical properties of PdCo/C catalysts are investigated. X-ray diffraction results reveal that although the low reduction temperature does not benefit the Co alloying with Pd, Pd–Au alloys are preferentially formed. Moreover, it confirms that the incorporation of Au into a Pd–Co system contributes to the generation of inhomogeneous alloy structure. Fine structural details determined by X-ray absorption spectroscopy indicate that Au addition improves the heteroatomic intermixing extent of alloy nanocatalysts, especially for Pd₇₅Co₁₀Au₁₅/C (Au₁₅) catalysts. Surface characterization by temperature programmed reduction suggests that a Pd-rich surface gradually changes to Pd, Au and alloy mixed surfaces when the Au content is larger than 15 at%. Regarding the electrochemical results, Au₁₅ displays the superior ORR performance among all samples due to the improved heteroatomic intermixing extent, large electrochemical surface area and multiple coexisting surface species. Furthermore, it also displays a better stability than Pd/C and Pd₇₅Co₂₅/C catalysts after accelerated durability tests.

---

Introduction

To date, fuel cells have been appealing devices owing to their high chemical-to-electrical energy conversion efficiency and high power density properties.¹ Synthesis of structure and surface controllable nanocatalysts with the aim of manipulating their electrochemical and catalytic characteristics is extensively highlighted in the research area of proton exchange membrane fuel cells (PEMFCs).^2,3 Recently, Pt nanoparticles (NPs) dispersed on a C matrix with high surface areas have been extensively employed as the cathode catalysts for the oxygen reduction reaction (ORR) in the present PEMFCs owing to their high performance in an acid environment.⁴ Although studies regarding the development of Pt or Pt-based catalysts have been extensively investigated to date, some researchers have put their endeavors on the non-Pt or Pd-based bimetallic systems, showing high potential as promising candidates to replace Pt.^5,6 Through the alloying effect with the second transition metals such as Fe, Co, or Ni, the ORR activity of Pd-based catalysts can be significantly enhanced because of the change of Pd–Pd inter-atomic distance, modification of the electronic structures, formation of more favorable active sites for O₂ absorption and alteration of the surface species.^7–10

Fernandez et al.¹¹ have proposed a main concept for choosing high performance Pd-based alloy catalysts on the basis of the thermodynamic model. As they have stated, the Pd–Co alloy NPs with a Pd/Co atomic ratio of 3/1 can display the best performance. In addition, Suo et al.⁹ have also proposed the conceptual consideration for designing high performance Pd–Co alloy NPs and have indicated that the bimetallic catalysts with Pd shell–alloy core structure are highly desired for exhibiting the extraordinary ORR activity. Apart from the bimetallic systems, some researchers also emphasize on the activity enhancement toward ORR by introducing the third elements, such as Au or Mo into a Pd–Co system for synthesizing ternary catalysts.^12,13 By way of Au addition, the oxophilicity and stabilization of catalysts can be accordingly modified.¹⁴ For example, in the Pt–Au system, the unfilled d-orbital vacancy of Pt is decreased by incorporating Au and thus the adsorption energy of O₂ on the alloy surface is remarkably modified, further enhancing the ORR activity.^15,16 On the other hand, Zhang et al.¹⁴ have reported that the modification of Pt by Au clusters can contribute to an increase in Pt oxidation potential, thereby promoting the long-term durability during the ORR measurement.

In order to obtain highly effective alloy catalysts, it is particularly important to thoroughly understand the variations in structures and ORR activities caused by alloying with the second or third elements in the catalysts.¹⁷ For the Pt-based alloy system, on the basis of density functional theory (DFT), the alteration in electronic structure driven by the introduction of Fe into the Pt lattice is delicately investigated.¹⁸ Also, in terms of the Pd-based alloy system, a volcano-type relationship between the ORR activity and d-band center energy of Pd over underlying Pd alloys is elaborately unravelled via similar computational calculation.¹⁹ In contrast to theoretical calculation, the fundamental understanding between the catalytic activity toward ORR and structures of Pt–Co NPs by X-ray absorption spectroscopy (XAS) has been reported.²⁰ Basically, the oxidation states and electronic structures, or fine structural details including the short-range ordering arrangement, the bond pair distance between heteroatoms, the coordination numbers and types inside alloy NPs can be revealed by combining X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) characterizations.

Besides the atomic ratio and alloying effect of the catalysts, the surface species involved in the catalysis reaction needs to be considered.^21,22 For example, with regard to the widely used PtRu/C catalysts at the anode of direct methanol fuel cells (DMFCs), it has been suggested that the surface majorly comprising of Pt species exhibits the superior performance toward methanol decomposition rather than crystalline RuO₂ ones.²³ In terms of the cathode catalysts of PEMFCs, the manipulation of surface and structural properties by means of ceria addition, post annealing process or electrochemical leaching of transition metals improves their ORR activity.^24–26 Clearly, the surface compositions of alloy NPs are a deciding factor in determining the catalytic activity directly.^24–27 In dealing with surface characterization, various techniques such as X-ray photoelectron spectroscopy (XPS), Fourier transform infrared (FTIR) spectroscopy, scanning transmission electron microscopy with nano-probe energy dispersive spectroscopy (STEM), anodic CO stripping and cyclic voltammetry (CV) in alkaline medium are believed to be useful to provide the insight into the changes in surface compositions.^28–30 Additionally, temperature programmed reduction (TPR) with its surface characterization capability is able to qualitatively offer not only the oxidation states but also the surface information of alloy NPs.^31–34 According to our previous study, based on the position of reduction peak temperature (T_r) in TPR profiles, the evolution of surface compositions of alloy catalysts can be extensively investigated.^31–34

In light of the previous literature, our main impetus is to explore the Pd–Co alloy catalysts exhibiting a comparable ORR performance to Pt/C. Meanwhile, it is also of tremendous significance to thoroughly comprehend the structure–surface dependent electrochemical behaviour of the prepared alloy catalysts. Herein, attempts are made to synthesize the Pd–Co based ternary catalysts by alloying Au to form Pd₇₅Co_25−xAu_x/C (x = 0–25 at%) catalysts for the ORR through the deposition–precipitation (DP) method along with a relatively low reduction temperature (390 K) to retain the small size and specific surface species of the catalysts.^32,33 For structural characterizations, besides X-ray diffraction (XRD), the XAS technique is employed for acquiring evolutions of detailed local structures within alloy NPs. To the best of our knowledge, despite that the fine structural details of alloy NPs or single crystal substrates with monolayers of Pt or Pd shells have been elucidated, such a technique has rarely been employed to resolve the structural variation in PdCoAu/C ternary systems.^35,36 Besides, the TPR method is applied to provide the information regarding the near surface region of various alloy NPs.

Experimental

Preparation of catalysts

The 20 wt% carbon-supported Pd₇₅Co_25−xAu_x (x = 0, 5, 15, and 25 at%) alloy catalysts were prepared through a conventional DP route. Three involved metal precursors, palladium nitrate (Pd(NO₃)₂), cobalt nitrate (Co(NO₃)₂·6H₂O) and gold chloride (HAuCl₄), in the desired stoichiometry were completely dissolved in DI water and co-deposited onto the commercial carbon black (Vulcan XC-72R) at 340 K. During the deposition process, 1.0 M of NaOH aqueous solution was used to adjust the pH value to near 9. Afterwards, the solution was stirred for 24 h at room temperature, filtered extensively with numerous DI water and dried at 320 K for 24 h. The dried catalysts were subjected to the reduction treatment in H₂/N₂ (10/90 vol%) gas at 390 K for 1 h and they were stored as the pristine catalysts. The Pd₇₅Co_25−xAu_x/C alloy catalysts with varying amounts of Au (x) are hereafter designated as Au_x for simplicity. Similarly, carbon supported 20 wt% Pd (Pd/C), Au (Au/C) and Co (Co/C) were also prepared by the same method for comparison.

Physical analyses

XRD characterization: the crystal phases and crystallite sizes of various alloy catalysts were characterized by a Shimadzu XRD using CuK_α radiation operated at 40 kV and 25 mA. The XRD profile was obtained at a scan rate of 0.12 and 0.024° s⁻¹ for the 2θ values ranging from 20° to 70° and 35° to 50°, respectively.

High resolution transmission electron microscopy (HR-TEM) characterization: the morphologies of various alloy catalysts were observed by using a high resolution transmission electron microscope (JEOL-2100) equipped with a LaB₆ electron gun and operated at 160 kV. Briefly, the powder alloy catalysts were ultrasonically suspended in 2-propanol. Afterwards, the suspension was then immediately dropped and dried on carbon supported on 200 mesh/inch copper grids.

TPR characterization: in each TPR analysis, a sample of approximately 20 mg was inserted into a U-shape quartz tube and pre-oxidized in air at room temperature (300 K) for 1 h. Subsequently, the pre-oxidized catalysts (alloy or reference samples) were reduced by a flow of 20% H₂ in N₂ at a flow rate of 30 mL min⁻¹ upon increasing the temperature from 100 to 800 K at a heating rate of 7 K min⁻¹. Once the reduction process was started, the rate of hydrogen consumption presented in the TPR profile was measured automatically by a thermal conductivity detector (TCD). Moreover, silica gel and molecular sieve absorbents were utilized for the purpose of water removal before the flowing gas reached the detector.

XAS analyses

The typical XAS spectra of various alloy NPs were obtained in fluorescence mode at the BL01C1 beamline at the National Synchrotron Radiation Research Center (NSRRC). The incident beam was monochromated using a double crystal monochromator equipped with a Si(111) crystal. A Si monochromator was employed to adequately select the energy with a resolution ΔE/E better than 1 × 10⁻⁴ at the Pd K-edge (24 350 eV), the Au L_III-edge (11 919 eV) and the Co K-edge (7709 eV). In general, all alloy NPs were dispersed uniformly on the tape and prepared as thin pellets with an appropriate absorption thickness (μx = 1.0, where μ is the absorption edge and x is the thickness of the sample) so as to attain the proper edge jump step at the absorption edge region. In order to acquire acceptable quality spectra, each XAS measurement was repeated at least twice and averaged for successive comparison. Moreover, the ionization chamber filled with different mixing gases such as Ar, N₂, He or Kr was used to detect the intensities of the incident beam (I_o), the fluorescence beam (I_f) and the beam finally transmitted by the reference foil (I_r). For the EXAFS analysis, the backgrounds of the pre-edge and the post-edge were subtracted and normalized with respect to the edge jump step from the XAS spectra (χ(E)). The normalized χ(E) spectra were transformed from energy to k-space and further weighted by k³ to distinguish the contributions of back scattering interferences from different coordination shells. Subsequently, the extracted k³-weighted spectra in k-space ranging from 3.3 to 11.5, 3.3 to 12.8 and 3.3 to 12.8 Å⁻¹ for the Pd K-edge, Co K-edge and Au L_III-edge were Fourier transformed into r-space, respectively. The phase correction was set on all spectra in the r-space. Finally, the filtered EXAFS data of the Pd K-edge were analyzed by a nonlinear least-squares curve fitting method in the r-space ranging from 1.7 to 2.8 Å⁻¹ depending on the bond to be fitted. The reference phase and amplitude for the Pd–Pd coordination were initially acquired from the Pd foil. Normally, the backscattered amplitude and phase shift functions for specific atom pairs were theoretically estimated by means of utilizing the FEFF7 code.³⁷ In addition, the reduction amplitude (S²₀) value for Pd was fixed at 0.83 in order to determine various structural parameters for each bond pair.

Electrochemical measurements

Cyclic voltammetry (CV): the CV test was assessed by using a glassy carbon (GC, area = 0.196 cm²) as a working electrode in a conventional electrochemical cell with a three-electrode configuration. A Pt wire and a saturated calomel electrode (SCE) separated from the working electrode by using a salt bridge were used as counter and reference electrodes, respectively. All potentials throughout this study were quoted with respect to the normal hydrogen electrode (NHE). Generally, the preparation of catalyst inks for the electrochemical measurement is as follows. At the initial step, 5 mg of catalysts was ultrasonically suspended in 1 mL of 2-propanol and 50 μL of 5 wt% Nafion solution (Aldrich) for about 60 min to obtain a well-dispersed catalyst ink, and then 20 μL of the slurry was spread on the surface of a GC electrode. After the fabrication of the electrode, it was then immersed into the N₂-saturated 0.1 M HClO₄ and the potential was scanned for 20 cycles from 0–1.2 V (vs. NHE) at a scan rate of 30 mV s⁻¹ for the pretreatment procedure. Afterwards, the CV experiment was carried out in the same electrolyte at a scan rate of 20 mV s⁻¹. Moreover, the electrochemical surface area (ECSA) is hereafter estimated by integrating charges associated with the adsorption of H.^38,39

Linear sweep voltammetry (LSV): the equipment for measuring the electrocatalytic activity toward ORR of catalysts and the preparation approach of catalyst inks was the same as that for the CV measurement. In brief, 20 μL of the catalyst slurry was transferred on the surface of the GC electrode and subsequently dried at room temperature. The estimated metal loading deposited on the GC electrode was around 19 μg_metal. For the examination of ORR activity on each alloy catalyst, the electrochemical test was carried out by way of a rotating disk electrode (RDE) at a rotation rate of 1600 rpm in 0.1 M of HClO₄ aqueous solution saturated with highly purified O₂. Prior to each test, the electrode was cycled several times between 0 and 1.2 V (vs. NHE) in order to produce clean surfaces and activate the electrocatalysts. During each measurement, a moderate O₂ gas flow was kept above the electrolyte to maintain the total concentration of O₂ molecules in the electrolyte. The scan rate of each LSV was set at 5 mV s⁻¹ with a potential of 1.0 to 0 V (vs. NHE) and all of the ORR polarization profiles were plotted in the negative sweep direction. Furthermore, the same experiment at various rotational speeds was conducted to identify the electron transfer numbers during the course of ORR.⁴⁰ The comparison of ORR activity for all alloy catalysts within the mixed kinetic-diffusion region (current density and mass activity at E = 0.75 V) was suggested and employed by some literature studies.^40,41

Accelerated durability test (ADT): in order to investigate the stability of Pd/C and modified samples, a repetitive potential cycling experiment of ADT was performed.^42,43 Basically, the fabrication of the electrode for ADT was identical to that mentioned in CV and LSV experiments. For the testing procedure of ADT, it was conducted by continuous potential cycling between 0.6 and 1.4 V (vs. NHE) at a scan rate of 50 mV s⁻¹ with periodic measurements of ECSA and ORR activity after every 50 scan. In addition, 0.1 M HClO₄ solution was used as a supporting electrolyte throughout this durability experiment. It is believed that alloy NPs or carbon supports were collapsed or leached out under such a strict acidic environment with a high cycling potential and therefore the electrochemical stability as well as variations of ORR performance of diverse catalysts could be examined.

Results and discussion

The demonstrative XRD profiles of Pd/C and various alloy catalysts synthesized by the DP route are depicted in Fig. 1. In Fig. 1a, the diffraction peak observed in all alloy catalysts at 2θ near 25° corresponds to C supports. Typically, when compared to the Pd patterns (JCPDS 46-1043), reflections of Au₀ (Pd₇₅Co₂₅/C) catalysts found at 40.5°, 47.0° and 69.0° slightly shift toward higher angles, implying that the incorporation of Co atoms into the Pd lattice results in the formation of the Pd–Co alloy. Similarly, in the case of Au₂₅ (Pd₇₅Au₂₅/C) catalysts, the formation of the Pd–Au phase can also be apparently detected based on the shift of the diffraction peaks relative to those of Pd. According to Vegard's law, the Co and Au atomic fractions (X_Co and X_Au) of Au₀ and Au₂₅ can be estimated as tabulated in Table 1.⁴⁴ As shown in Table 1, the X_Au is higher than X_Co (52.6% vs. 8.9%), indicating that Pd–Au alloys can be formed at the low reduction temperature of 390 K. On the other hand, the low X_Co for Au₀ is ascribed to the insufficient reduction temperature for Co precursors since the temperature for complete reduction of Co is at least 620 K.⁴⁵ Despite Au₀ has the low X_Co, our previous studies suggest that a high reduction temperature results in the serious agglomeration of Pd–Co alloy NPs together with Co surface enrichment and the concomitant decline in ORR activity.^32,33 On the other hand, for the XRD patterns of Au₅ (Pd₇₅Co₂₀Au₅/C) and Au₁₅ (Pd₇₅Co₁₀Au₁₅/C) catalysts, their major diffraction (111) peaks lie between those of the Au₀ and Au₂₅ ones, suggesting that these two samples may be concurrently comprised of Pd–Co and Pd–Au alloys. This can be evidenced by their (111) asymmetric peaks, confirming that Au₅ and Au₁₅ samples have inhomogeneous alloy structures. Therefore, the atomic fractions of Au₅ and Au₁₅ cannot be easily calculated owing to the coexistence of Pd–Au and Pd–Co alloys. Table 1 compares the lattice constants of various Au_x samples. It is noted that the value increases with increasing Au content, signifying that Au atoms are indeed alloyed with Pd and affect the atomic fraction of Au_x to different extents.

Fig. 1 XRD patterns for Pd/C and various Au_x (x = 0–25) alloy catalysts. The XRD profiles were obtained at a scan rate of (a) 0.12° s⁻¹ and (b) 0.024° s⁻¹. The solid and dashed vertical lines represent the (111) diffraction peaks for the monometallic Pd and Au, respectively.

Table 1 Structural parameters of various alloy catalysts determined from XRD

Samples	Phases	2θ (°)	Lattice constant (Å)	X Co or XAu^a (%)	L alloy ^b (nm)
a The Co or Au atomic fractions are calculated by Vegard's law. b The crystallite size of alloy NPs is estimated by Scherrer's equation. c The atomic fractions cannot be calculated owing to the coexistence of Au and Co atoms in the Pd lattice. d The crystallite size cannot be determined due to the asymmetric peak shape of (111) diffraction.
Au0	C, PdCo	40.48	3.86	8.9	3.9
Au5	C, PdCo, PdAu	40.14	3.89	—^c	—^d
Au15	C, PdCo, PdAu	39.13	3.98	—^c	—^d
Au25	C, PdAu	39.01	3.99	52.6	4.7

---

In order to obtain more adequate structural information, the XRD measurement with slow scan rate for 2θ ranging from 35° to 50° is performed and the typical diffraction patterns of all presented catalysts are shown in Fig. 1b. Even though the structures of Au₅ and Au₁₅ composed of two mixed phases, Pd–Co and Pd–Au alloys, are observed, the delicate structural details must be further examined by the technology of XAS. The average crystallite size (L_alloy) of the Au₀ and Au₂₅ catalysts is evaluated by Scherrer's formula described in the following eqn (1).⁴⁶

L = K × λ/(FWHM × cos θ) (1)

where λ = 0.154 nm is the wavelength, K = 0.89 is the Scherrer constant, θ is the Bragg angle and FWHM is the full width at half maximum. The calculated L_alloy of Au₀ and Au₂₅ is nearly 3.9 and 4.7 nm as listed in Table 1, respectively. However, for Au₅ and Au₁₅ samples, the L_alloy is not susceptible to be calculated due to the asymmetric and broad (111) diffraction peak. This phenomenon can correspond to the formation of the inhomogeneous alloy structure and the generation of stacking faults or microstrains within alloy NPs, leading to the difficulty in calculating the L_alloy.⁴⁶

Although XRD results indicate that the incorporation of Au into Pd–Co systems gives rise to the formation of inhomogeneous alloys for Au₅ and Au₁₅, the effect of Au addition on the fine structures of various catalysts can be explored precisely by XAS. The XANES spectra of the Pd K-edge for various alloy catalysts and Pd foil are displayed in Fig. 2a. For Au₀, the presence of the strongest multiple-excitation amplitude (peak B) (i.e. the outgoing photoelectrons bump into heavy elements at the neighboring shells) within the near-edge region suggests that it has the most Pd metallic-like character among all presented catalysts. In other words, Au₀ may have the lowest degree of heteroatomic intermixing as compared with other samples. This is consistent with the XRD results in which Au_o has a low atomic fraction (X_Co) of about 8.9%. On the other hand, a pronounced peak broadening (across the region from arrows A to B) together with a slight increase in white line (WL) intensity found for Au₅, Au₁₀, and Au₂₅ samples may be caused mainly by the subtle oxidation of Pd and partially by the hybridization between Pd, Co and Au atoms within NPs.⁴⁷

Fig. 2 (a) XANES and (b) FT-EXAFS spectra of the Pd foil and various alloy catalysts for the Pd K-edge. The intensity of EXAFS spectra for the Pd foil is multiplied by 0.5 for clear comparison.

The full atomic images of Pd and Co K-edges for various alloy catalysts are sketched on the basis of the XAS derived structural information. The corresponding Fourier transformed radial structure function (RSF) of the Pd K-edge is shown in Fig. 2b. Obviously, for Au₀, the shoulder at near 2.4 Å close to the main radial peak (at 2.8 Å) can be assigned to the contribution of interference from the Pd–Co bond pair in the 1st coordination shell. In the case of Au promoted nanocatalysts (Au_x, x = 5, 15 and 25), a perceivable shoulder along with a slight peak shift is found as a result of the local structure expansion (or structure distortion) and the intercalation of Au atoms inside the framework of the Pd–Co lattice. Furthermore, it is worth noting that their 1st radial peak features are sharply damped relative to that of Au₀, suggesting the significant disruption of Pd–Co lattice structures by the insertion of atoms with strong homo-metallic affinity (for example, Pt or Au atoms). In other words, it reveals that the out-of-phase interferences between Pd–Pd, Pd–Co and Pd–Au oscillations are caused by the formation of Pd–Co, Pd–Au, or even Au clusters in bulk structure. The local structural parameters around Pd atoms of all catalysts are quantitatively determined by using the fitting model. Basically, the established model for fitting the Pd K-edge of Au₀ has two bond length parameters because Co suffers from serious oxidation and is hardly alloyed into the Pd lattice (this will be discussed hereafter). Conversely, owing to the good intermixing between Pd and Au for Au_x (x = 5–25), the bulk structure can be recognized as a perfect solid solution. Thus, only one bond length parameter is taken into account in various Au modified alloy catalysts except Au₀. The best fitting results and the corresponding profiles are summarized in Table 2 and provided in Fig. S1 (ESI†), respectively. For Au₀, the coordination numbers of Pd–Pd (N_Pd–Pd) and Pd–Co (N_Pd–Co) bond pairs are determined to be 9.1 and 1.1, respectively. Compared with the Au₀, the N_Pd–Pd of various samples substantially decreases after the Au addition. Besides, the high N_Pd–Pd structure parameter belonging to Au₀ is indicative of its metallic-like core structure (low intermixing degree) which is consistent with the XANES results. On the other side, the calculated heteroatomic intermixing extent (χ%, i.e. (N_Pd–Co + N_Pd–Au)/N_total) shown in Table 2 of all catalysts indicates that the degree of intermixing between Pd, Co and/or Au inside alloy NPs can be well-improved (especially for Au₁₅) after the incorporation of Au.²⁰ Meanwhile, for Au₅ and Au₁₅, their N_Pd–Au is found to be larger than that of N_Pd–Co, implying that the bulk structure of alloy NPs is comprised of Pd–Au mainly and Pd–Co partially.

Table 2 EXAFS fitting parameters of the Pd K-edge for the Pd foil and various alloy catalysts

Samples	Shell	N ^a	Heteroatomic intermixing extent^b [χ%]	R ^c [Å]	σ ^2 (×10^−3)^d [Å^2]	ΔE0^e [eV]	R factor^f
a N: coordination number. b χ: the heteroatomic intermixing extent is determined using (NPd–Co + NPd–Au)/Ntotal. c R: bond distance. d σ ^2: Debye–Waller factor. e ΔE0: inner potential correction. f R factor: residual error.
Pd foil	Pd–Pd	12 (±0.03)	0.0	2.74 (±0.01)	5.0 (±0.19)	2.9	<0.01
Au0	Pd–Pd	9.1 (±0.5)	10.8	2.72 (±0.01)	8.5 (±1.7)	−5.3	0.01
	Pd–Co	1.1 (±0.3)		2.63 (±0.02)
Au5	Pd–Pd	5.9 (±0.3)	23.3	2.72 (±0.01)	7.9 (±0.71)	−8.0	0.02
	Pd–Co	0.4 (±0.1)
	Pd–Au	1.4 (±0.2)
Au15	Pd–Pd	5.8 (±0.3)	24.7	2.72 (±0.02)	8.3 (±1.3)	−4.5	0.02
	Pd–Co	0.3 (±0.1)
	Pd-Au	1.6 (±0.2)
Au25	Pd–Pd	6.2 (±0.4)	19.5	2.73 (±0.01)	8.3 (±1.3)	−16.4	0.01
	Pd–Au	1.5 (±0.4)

---

In the case of Co K-edge XANES spectra plotted in Fig. 3a, the relatively high WL intensity found in Au₀ and Au₅ implies that most of the Co atoms exist in oxidized or unalloyed states. In terms of Au₁₅, the presence of a broad plateau (arrow C) illustrates the partial metallic character of Co atoms and a good intermixing and/or homogeneity in this sample, consistent with the χ value (24.7%) shown in Table 2. Besides, in the typical Co K-edge RSF shown in Fig. 3b, the radial peaks of D (at ∼1.85 Å) and E (at ∼2.81 Å) related to the interferences of photoelectrons by Co–O and Co–Co bond pairs, respectively, in a tetragonal phased CoO₂ domain are obviously perceived.^48,49 Therefore, the weak radial peak (Co–O bonding) noted for Au₁₅ accounts for the more metallic-like Co characteristic, which results in the well alloyed bulk structure. Conversely, for Au₀, the intense radial peak indirectly reveals its poor intermixing between each heteroatoms among various alloy catalysts (χ = 10.8%).

Fig. 3 (a) XANES and (b) FT-EXAFS spectra of the Co foil and various alloy catalysts for the Co K-edge. The intensity of EXAFS spectra for the Co foil is multiplied by 0.5 for clear comparison.

The representative Au L_III-edge XANES and RSF spectra are presented in Fig. S2a and b (ESI†), respectively. As shown in Fig. S2a (ESI†), peaks labeled E, F and G are characteristic of the pure Au with face center cubic structure.⁵⁰ Notably, a slight increase in WL intensity implies the mild oxidation of Au. The results of Au L_III-edge RSF shown in Fig. S2b (ESI†) suggest that the metallic peak (across 2.6 to 2.9 Å) of various Au_x (x = 5, 15 and 25) grows with increasing Au content. Such information reveals the preferential formation of Au clusters inside the bulk structures, as consistently probed by the results of Pd and Co K-edge XAS, consequently resulting in the distortion of local structure around Pd and Co atoms.

Fig. 4 shows the morphologies and size distribution histograms of various alloy catalysts determined by HR-TEM. It is clearly observed that the alloy NPs are well-dispersed on the C supports for all samples. Besides, the measured mean particle diameters (D) of various samples are tabulated in Table 3 by counting more than 200 randomly chosen particles in TEM images. Note that the D_alloy for Au₀, Au₅, Au₁₅ and Au₂₅ catalysts is around 4.9, 5.1, 5.2 and 5.4 nm, respectively. Obviously, the increase in D with increasing Au content may be attributed to the lattice expansion originating from the inclusion of Au atoms into the Pd lattice. It is worthwhile to mention that the slight inconsistencies between crystallite size (L) and particle size (D) for Au₀ as well as Au₂₅ catalysts are presumably ascribed to that the L has a mass average value which is 2/3 of the sphere equivalent diameter obtained from TEM by particle numbers.⁵¹ Even though the differences in mean D between all alloy NPs are quietly small, Au₁₅ displays a comparative narrow size distribution in the region of 4.5 to 5.9 nm compared to the other samples.

Fig. 4 TEM morphologies and the corresponding histograms of (a, b) Au₀, (c, d) Au₅, (e, f) Au₁₅ and (g, h) Au₂₅ alloy catalysts. Insets show high magnification of each sample.

Table 3 TEM and TPR results of various alloy catalysts

Samples	TEM	TPR
	D alloy [nm]	T r ^a [K]	Surface species
a Main reduction peak temperature (Tr) of different surface oxide species. b Weak peak.
Pd/C	—	282	PdOx
Co/C	—	475, 738	CoOOH, Co3O4^b
Au/C	—	125	AuOx
Au0	4.9 ± 0.9	296	PdOx
Au5	5.1 ± 0.8	301	PdOx
Au15	5.2 ± 0.7	170, 244, 289	AuOx, AOx, PdOx^b
Au25	5.4 ± 1.0	162, 235, 290	AuOx, AOx, PdOx

---

In the viewpoint of surface characterization, the TPR technique is employed as a diagnostic tool for probing the species on the topmost surface region of alloy NPs. Fig. 5 compares a series of typical TPR traces of various alloy catalysts along with Pd/C. Additionally, two sets of TPR traces corresponding to reference Au/C and Co/C catalysts are depicted in Fig. S3 (ESI†). Noticeably, the H₂ uptake signal related to the reduction of some oxygen-based functional groups from C supports can be detected at temperatures higher than 900 K.⁵² These peaks are beyond the range of our experimental measurements and hence the effect of such functional groups can be neglected rationally. Note that the major TPR profile of Pd/C displays two distinguishable peaks (T_rs) located at approximately 282 and 383 K which are dominantly assigned to the H₂ consumption and desorption of Pd, respectively.²⁶ Besides, in Fig. S3 (ESI†), the main T_r of surface oxides for Au/C is at 125 K. The low T_r demonstrates that the absorption of O₂ onto the surface of Au cannot easily occur relative to that of Pd or Co, indicative of an inferior oxophilicity property of Au.^10,53 For Co/C, two observable T_rs lying at 450–500 and 750 K are assigned to the reduction of surface CoOOH and Co₃O₄ species, respectively.^32,33,54 The evolution of surface species and major T_rs on various catalysts is tabulated in Table 3. For Au₀, no obvious H₂ consumption peak belonging to Co oxide species (CoOOH or Co₃O₄) is observed; instead, the reduction peak of PdO_x species is detected significantly, suggesting that its surface is Pd rich. This result confirms that reduction temperature at 390 K may not form Pd–Co alloys completely but can prevent Co from surface segregation and sintering.^32,33 In terms of Au₅, the surface is also composed of PdO_x species mainly. Because of high reduction potential for Au ions in comparison to Pd and Co ones, some portions of Au precipitate firstly during the deposition and precipitation process.⁵⁵ As a result, for a sample with low Au content (Au₅), most of the Au ions precipitate initially and the correspondingly Pd-rich surface is perceived. Moreover, when compared with Pd/C, the positive movement of H₂ uptake peak and peak broadening are found for Au₀ and Au₅, respectively, suggesting that alloying of Co or Au may subtly affect the reduction mechanism of surface PdO_x.^10,26 In contrast, with respect to Au₁₅ and Au₂₅, introduction of an appropriate amount of Au into the Pd–Co system leads to the formation of multiple coexisting surface species such as PdO_x, AO_x (where A means the Pd–Au alloy species) or AuO_x. From the detailed investigations on the structural, morphological and surface information of various alloy nanocatalysts, it is vital for us to correlate those imperative findings with their electrochemical properties.

Fig. 5 TPR spectra of Pd/C and various Au_x (x = 0–25) alloy catalysts.

In order to comprehend the impact of Au promotion on variations in ECSAs of various alloy catalysts, the CV test is conducted. Fig. 6 presents CV profiles in which a well-defined hydrogen adsorption/desorption (H_ads/des) region at the potential near 0–0.3 V is noted for all catalysts. Based on the integration of charges within the H_ads/des domain, the ECSAs of all nanocatalysts along with Pd/C can be roughly estimated as listed in Table 4. It is observed that the ECSA of Au₀ catalysts is very close to that of Pd/C. On the other hand, it is worthy of note that the ECSA increases prominently with increasing Au content. Although the inclusion of Au into Pd–Co alloys increases the average D slightly (see TEM results shown in Fig. 4), they still have comparatively high ECSA values among all samples. This may be explained by that the insertion of Au into the Pd–Co system decreases the total N_Pd–Pd number as well as promotes the heteroatomic intermixing extent (χ) and thus the promotion of effective ECSAs is observed accordingly. In Table 2, the coordination number N_Pd–Pd of Au₀ is larger than those of Au-promoted catalysts (Au₀, Au₁₅ and Au₂₅), implying that most of the Pd atoms enriched in the core structure since atoms covering the shell may have fewer neighbors.⁵⁶ In this concern, despite that Au₀ has a small D, its ECSA is lower than those of Au-modified catalysts. Apart from this, the onset potential of oxide reduction in the negative-going direction for all modified catalysts slightly moves to more positive potential relative to Pd/C. This finding is attributed to that the chemisorptions of oxygenated species such as OH_ads on Pd sites are inhibited due to the alternation of electronic structure of Pd from Co or Au alloying.^57,58 Thus, the improvement of ORR kinetics for various promoted catalysts can be expected reasonably.

Fig. 6 CV profiles of Pd/C and various Au_x (x = 0–25) alloy catalysts recorded in the N₂ saturated 0.1 M HClO₄ solution at a scan rate of 20 mV s⁻¹.

Table 4 CV and LSV results for various alloy catalysts

Samples	ECSA^a (m^2 g^−1Pd)	I 075
		Current density [mA cm^−2]	Mass activity^b [mA mg^−1Pd]
a ECSA: electrochemical surface area calculated by measuring the charges associated with the H2 absorption and desorption region in the CV profile. b Mass activity: current at E = 0.75 V normalized to the Pd mass of alloy catalysts.
Pd	57.6	0.9	9.3
Au0	62.3	1.4	16.9
Au5	79.8	1.6	20.7
Au15	116.2	2.6	39.2
Au25	120.3	2.1	35.8

---

Fig. 7 depicts a characteristic set of ORR polarization curves for Pd/C, and various alloy catalysts. The typical features associated with a visible diffusion-controlled region starting from 0.5 to 0.65 V and a mixed kinetic-diffusion control region going from 0.65 to 0.8 V can be apparently observed. From the LSV profile, the slight discrepancies in diffusion limiting current densities among all alloy catalysts are probably caused by several factors such as surface roughness, heat treatment temperatures, preparation means and alloy compositions which influence the flux density during the course of ORR measurement.⁵⁹ In Table 4, the ORR activities at I₀₇₅ (current density at E = 0.75 V in the mixed kinetic-diffusion control region) for all presented nanocatalysts are compared. The ORR activity enhancement is in the order of Au₁₅ > Au₂₅ > Au₅ > Au₀ > Pd/C. Furthermore, compared with Pd/C, the positive shift of onset potential for Au₁₅ is indicative of its lower overpotential for the dissociation of O₂. On the other hand, the Pd mass activities of various catalysts determined by the I₀₇₅ (current density at E = 0.75 V normalized to the Pd mass of each nanocatalyst) values are listed in Table 4. It is found that the Pd mass activity of Au₁₅ is about 4.2 and 2.3-fold higher than that of Pd/C and Au₀ (Pd₇₅Co₂₅/C).

Fig. 7 ORR activity curves for Pd/C and various Au_x (x = 0–25) alloy catalysts obtained in O₂-saturated 0.1 M HClO₄ with a rotation speed of 1600 rpm. All polarization profiles are plotted in the negative sweep direction at a scan rate of 5 mV s⁻¹.

In order to evaluate the kinetic parameter of Au₀ and Au₁₅ alloy catalysts during the course of ORR, the polarization curves recorded at different rotational speeds in O₂-saturated 0.1 M HClO₄ are provided in Fig. 8. In addition, two insets show their corresponding Koutecky–Levich plot drawn against the inverse current (I⁻¹) as a function of the inverse square root of the ration rate (ω^−1/2).⁶⁰ As is extensively known the O₂ catalyzed reduction process on Pt or Pd-based alloy catalysts involves four or two electron-transfer pathways in the acid environment. In the four electron pathway, the O₂ molecule is directly reduced into water (H₂O) with a standard potential of 1.23 V. While in the two electron pathway, the formation of an undesirable by-product of hydrogen peroxide (H₂O₂) with a standard potential of 0.69 V is observed, which leads to a decrease in the net reaction potential. Thus, based on the Koutecky–Levich equation (eqn (2)) and the measured slopes, the apparent number of electrons transferred per oxygen molecule (n) can be adequately calculated.

I⁻¹ = I_k⁻¹ + I_d⁻¹ = I_k⁻¹ + (βω^1/2)⁻¹ (2)

where I_k is the kinetic current, I_d is the diffusion current and β is equal to 0.62nFAD^2/3C_v^−1/6. To estimate the Levich constant, some fundamental constants such as a geometric area of the electrode of 0.196 cm⁻², a bulk O₂ solubility of 1.18 × 10⁻⁶ mol cm⁻³, a diffusion coefficient of 1.9 × 10⁻⁵ cm² s⁻¹ and a kinematic viscosity of 8.93 × 10⁻³ cm² s⁻¹ in 0.1 M HClO₄ solution have been used.⁶⁰ It is evident that both Au₀ and Au₁₅ alloy catalysts display a well-defined diffusion-controlled region at various rotational speeds. The straight lines shown in the inset plots suggest the characteristic of first order dependence of O₂ kinetics. Furthermore, their calculated values of n are approximately 3.8 and 3.7, implying that both of them have a similar ORR pathway (i.e. four electron pathway) and a similar rate-determining step during the course of ORR.

Fig. 8 ORR polarization curves for (a) Au₀ and (b) Au₁₅ alloy catalysts recorded in O₂-saturated 0.1 M HClO₄ at various rotation speeds. All polarization profiles are plotted in the negative sweep direction at a scan rate of 5 mV s⁻¹. Insets show their corresponding Koutecky–Levich plot obtained at a potential of 0.4 V.

The stability characteristics of Pd/C, Au₀ and Au₁₅ catalysts are obtained by the ADT method where these catalysts are subjected to repetitive potential cycling between 0.6 and 1.4 V at a scan rate of 50 mV s⁻¹ in 0.1 M HClO₄. Under these severe conditions, the redox reactions of surface Pd atoms occur quickly and consequently cause the serious dissolution of Pd or Co or Au elements.⁶¹ Moreover, polarizing the catalysts in an acid medium at a high positive potential region may also result in the corrosion of C supports, thereby leading to the loss of active area obviously. Fig. 9(a) presents a series of CV curves for Pd/C, Au₀ and Au₁₅ catalysts recorded after every 50 potential cycling scan. Clearly, significant changes in the features of CV curves (peak movement and area reduction in the cathodic sweep direction or area alternation within the H_ads/des domain) for all catalysts can be found, indicating that alloy NPs are gradually leached out during ADT. Recently, several mechanisms have been proposed to explicit the loss of ECSA for active catalysts. During a long-term measurement, Pt (or Pd) atoms supported on C supports are dissolved into electrolytes and then re-deposited onto the surface of large particles, leading to a decrease in ECSA.^43,62 On the other hand, another critical issue needed to be taken into account is the dramatic corrosion of C supports while the sweeping potential reaches as high as near 1.5 V. It has been suggested that the electrochemical corrosion of such high-surface-area supports may cause the intensive deconstruction and/or agglomeration of anchored metal NPs (Pt or Pd).⁴³Fig. 9(b) shows the enlarged CV polarization curves recorded after 200 cycling scans on various catalysts. As can be seen the H_ads/des peak of all samples can hardly be observed, confirming that most of the active Pd NPs are detached from supports and dissolved into an acid electrolyte. Besides, for Au₁₅, the presence of a split reduction peak accompanied with a small peak located at 1.2 V (reduction of surface Au oxides) may corroborate the collapse of PdCoAu alloy structure.⁶³ Note that the substantial change in electrical double layer area is ascribed to the deteriorated stability of C supports.⁶⁴ In other words, the surface or chemical states of C supports are changed significantly after ADT.

Fig. 9 (a) CVs and (b) enlarged CV plots after 200 potential cycling of Pd/C, Au₀ and Au₁₅ catalysts recorded in the N₂ saturated 0.1 M HClO₄ solution at a scan rate of 20 mV s⁻¹.

Fig. 10 compares the variations in ECSA and ORR activity during the ADT process for Pd/C, Au₀ and Au₁₅ catalysts. Considering the changes in ECSA, even though the ECSA degradation rate of Au₁₅ appears to be faster than that of Pd/C and Au₀ at the initial stage which may be due to the destruction of the alloy structure, its active area is always higher than those of others. Furthermore, a similar phenomenon can be perceived while examining the catalytic activity toward ORR at 0.75 V. Obviously, the ORR activity of Au₁₅ is undoubtedly better than those of both Pd/C and Au₀ samples at each stage, indicating that Au₁₅ is electrochemically stable. These observations are consistent with the ECSA results, and also signify that the addition of Au into Pd–Co not only increases both ECSA and ORR performances but also has a positive effect on the stabilization of alloy catalysts.

Fig. 10 Variations in ECSA and ORR activity during ADT measurements as a function of cycle numbers for Pd/C, Au₀ and Au₁₅ catalysts.

Through the systematic comprehension on structures, surface compositions and electrochemical behaviors, it can be summarized that the ORR performance of PdCo/C can be significantly improved through the optimized Au promotion (Au₁₅) where it possesses the well-promoted heteroatomic intermixing extent (χ), the narrow particle size distribution, the promoted ECSA, and the concomitant multiple coexisting species (Pd, Au and alloy) dominated on the outermost surface. In addition, a pertinent amount of Au incorporation (Au₁₅) can improve the durable character relative to Pd/C and unmodified catalysts (Au₀).

Conclusions

In this study, the development and modification of Pd–Co nanocatalysts through the Au addition based on the commonly used DP route are studied for the enhancement of ORR activity. The relationship between structures, surface species and electrochemical properties of the ternary Pd₇₅Co_25−xAu_x/C (x = 0–25 at%) catalysts is systematically investigated by various physical and electrochemical characterizations. From XRD characterization, it is observed that the atomic fraction of Pd₇₅Au₂₅/C (Au₂₅) is much higher than that of Pd₇₅Co₂₅/C (Au₀), indicating that Au is easier to alloy with Pd than Co. Besides, the addition of Au gives rise to the formation of inhomogeneous alloy structure. The structure of Au₅ and Au₁₅ is composed majorly of Pd–Au and partially of Pd–Co alloys as predicted by XRD. Detailed structural information disclosed by XAS shows similar tendency for all presented catalysts, meaning that the heteroatomic intermixing extent can be modified by the Au addition. Morphologies determined by TEM show that all of the alloy NPs are dispersed uniformly on C supports with the size of about 5 nm and Au₁₅ has the relatively narrow size distribution among all nanocatalysts. On the other side, according to the standpoint of surface characterization measured by TPR, the Pd-rich surface gradually changes to Pd, Au and alloy mixed surface species when Au content is larger than 15 at%. The ORR performance of PdCo/C is optimized with addition of 15 at% of Au. Therefore, for the preparation of high activity and stabilization catalysts toward ORR, it is highly suggested that the incorporation of 15 at% Au into Pd–Co system can not only enhance the heteroatomic intermixing degree and narrow the alloy NPs distribution but also alter the surface species (Pd rich to Pd, Au and alloy mixed species), thereby leading to the improvement of ORR performance.

Acknowledgements

We acknowledge the financial support from the National Science Council of Taiwan under Contract NSC-100-2221-E-008-051 and facility support from the National Synchrotron Radiation Research Center (NSRRC).

References

1. H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal., B, 2005, 56, 9.
2. J. Luo, L. Wang, D. Mott, P. N. Njoki, Y. Lin, T. He, Z. Xu, B. N. Wanjana, I. I. S. Lim and C. J. Zhong, Adv. Mater., 2008, 20, 4342.
3. Z. Peng and H. Yang, Nano Today, 2009, 4, 143.
4. V. R. Stamenkovic, B. Fowler, B. S. Mun, G. Wang, P. N. Ross, C. A. Lucas and N. M. Markovic, Science, 2007, 315, 493.
5. F. Kadirgan, A. M. Kannan, T. Atilan, S. Beyhan, S. S. Ozenler and S. Suzer, Int. J. Hydrogen Energy, 2009, 34, 9450.
6. E. Antolini, T. Lopes and E. R. Gonzalez, J. Alloys Compd., 2008, 461, 253.
7. K. Lee, O. Savadogo, A. Ishihara, S. Mitsushima, N. Kamiya and K. I. Ota, J. Electrochem. Soc., 2006, 153, A20.
8. L. Zhang, K. Lee and J. Zhang, Electrochim. Acta, 2007, 52, 7964.
9. Y. Suo, L. Zhuang and J. Lu, Angew. Chem., Int. Ed., 2007, 46, 2862.
10. Y. C. Wei, C. W. Liu and K. W. Wang, Chem. Commun., 2011, 47, 11927.
11. J. L. Fernandez, D. A. Walsh and A. J. Bard, J. Am. Chem. Soc., 2005, 127, 357.
12. V. Raghuveer, P. J. Ferreira and A. Manthiram, Electrochem. Commun., 2006, 8, 807.
13. C. V. Rao and B. Viswanathan, Electrochim. Acta, 2010, 55, 3002.
14. J. Zhang, K. Sasaki, E. Sutter and R. R. Adzic, Science, 2007, 315, 220.
15. C. W. Liu, Y. C. Wei and K. W. Wang, J. Colloid Interface Sci., 2009, 336, 654.
16. B. N. Wanjala, J. Luo, B. Fang, D. Mott and C. J. Zhong, J. Mater. Chem., 2011, 21, 4012.
17. S. Axnanda, K. D. Cummins, D. W. Goodman and M. P. Soriaga, ChemPhysChem, 2010, 11, 1468.
18. F. J. Lai, H. L. Chou, L. S. Sarma, D. Y. Wang, Y. C. Lin, J. F. Lee, B. J. Hwang and C. C. Chen, Nanoscale, 2010, 2, 573.
19. M. H. Shao, T. Huang, P. Liu, J. Zhang, K. Sasaki, M. B. Vukmirovic and R. R. Adzic, Langmuir, 2006, 22, 10409.
20. B. J. Hwang, S. M. S. Kumar, C. H. Chen, Monalisa, M. Y. Cheng, D. G. Liu and J. F. Lee, J. Phys. Chem. C, 2007, 111, 15267.
21. H. Nitani, T. Nakagawa, H. Daimon, Y. Kurobe, T. Ono, Y. Honda, A. Koizumi, S. Seino and T. A. Yamamoto, Appl. Catal., A, 2007, 326, 194.
22. S. Chen, P. J. Ferreira, W. Sheng, N. Yabuuchi, L. F. Allard and Y. S. Horn, J. Am. Chem. Soc., 2008, 130, 13818.
23. Y. C. Wei, C. W. Liu and K. W. Wang, ChemPhysChem, 2009, 10, 1230.
24. V. Stamenkovic, T. J. Schmidt, P. N. Ross and N. M. Markovic, J. Electroanal. Chem., 2003, 554, 191.
25. C. W. Liu, Y. C. Wei and K. W. Wang, Chem. Commun., 2010, 46, 2483.
26. Y. C. Wei, C. W. Liu, H. W. Lee, S. R. Chung, S. L. Lee, T. S. Chan, J. F. Lee and K. W. Wang, Int. J. Hydrogen Energy, 2011, 36, 3789.
27. K. J. J. Mayrhofer, K. Hartl, V. Juhart and M. Arenz, J. Am. Chem. Soc., 2009, 131, 16348.
28. P. H. C. Camargo, M. Rycenga, L. Au and Y. Xia, Angew. Chem., Int. Ed., 2009, 48, 2180.
29. R. Alvarez-Puebla, R. Contreras-Caceres, I. Pastoriza-Santos, J. Perez-Juste and L. M. Liz-Marzan, Angew. Chem., Int. Ed., 2009, 48, 138.
30. F. J. Lai, W. N. Su, L. S. Sarma, D. G. Liu, C. A. Hsieh, J. F. Lee and B. J. Hwang, Chem.–Eur. J., 2010, 16, 4602.
31. W. D. Kang, Y. C. Wei, C. W. Liu and K. W. Wang, Electrochem. Commun., 2011, 13, 162.
32. Y. C. Wei, C. W. Liu, Y. W. Chang, C. M. Lia, P. Y. Lim, L. D. Tsai and K. W. Wang, Int. J. Hydrogen Energy, 2010, 35, 1864.
33. Y. C. Wei, C. W. Liu and K. W. Wang, ChemPhysChem, 2010, 11, 3078.
34. Y. C. Wei, C. W. Liu, W. D. Kang, C. M. Lai, L. D. Tsai and K. W. Wang, J. Electroanal. Chem., 2011, 660, 64.
35. T. Y. Chen, T. L. Lin, T. J. M. Luo, Y. Chioi and J. F. Lee, ChemPhysChem, 2010, 11, 2383.
36. R. R. Adzic, J. Zhang, K. Sasaki, M. B. Vukmirovic, M. Shao, J. X. Wang, A. U. Nilekar, M. Mavrikakis, J. A. Valerio and F. Uribe, Top. Catal., 2007, 46, 249.
37. S. I. Zabinsky, J. J. Rehr, A. L. Ankudinov, R. C. Albers and M. J. Eller, Phys. Rev. B: Condens. Matter, 1995, 52, 2995.
38. W. Du, Q. Wang, C. A. LaScala, L. Zhang, D. Su, A. I. Frenkel, V. K. Mathur and X. Teng, J. Mater. Chem., 2011, 21, 8887.
39. H. Schulenburg, J. Durst, E. Muller, A. Wokaun and G. G. Scherer, J. Electroanal. Chem., 2010, 642, 52.
40. U. A. Paulus, T. J. Schmidt, H. A. Gasteiger and R. J. Behm, J. Electroanal. Chem., 2001, 495, 134.
41. C. W. Liu, Y. C. Wei and K. W. Wang, J. Phys. Chem. C, 2011, 115, 8702.
42. W. Zhu, J. P. Zheng, R. Liang, B. Wang, C. Zhang, G. Au and E. J. Plichta, J. Electrochem. Soc., 2009, 156, B1099.
43. S. Y. Huang, P. Ganesan and B. N. Popov, Appl. Catal., B, 2010, 96, 224.
44. G. Zhang, Y. Wang, X. Wang, Y. Chen, Y. Zhou, Y. Tang, L. Lu, J. Bao and T. Lu, Appl. Catal., B, 2011, 102, 614.
45. S. Srinivasan, R. J. De-angelis, P. J. Reucroft, A. G. Dhere and J. Bentley, J. Catal., 1989, 116, 144.
46. I. L. Leontyev, D. Y. Chernyshov, V. E. Guterman, E. V. Pakhomova and A. V. Guterman, Appl. Catal., A, 2009, 357, 1.
47. D. C. Koningsberger and R. Prins, X-ray absorption: Principles, applications, techniques of EXAFS, SEXAFS and XANES, Wiley, New York, 1988, p. 53.
48. ICSD #88722, J. M. Tarascon, G. Vaughan, Y. Chabre, L. Seguin, M. Anne, P. Strobel and G. Amatucci, J. Solid State Chem., 1999, 147, 410.
49. ICSD #28505, J. G. Mullen, Phys. Rev., 1968, 168, 220.
50. M. P. Casaletto, A. Longo, A. M. Venezia, A. Martorana and A. Prestianni, Appl. Catal., A, 2006, 302, 309.
51. Y. Zhang, Q. Huang, Z. Zou, J. Yang, W. Vogel and H. Yang, J. Phys. Chem. C, 2010, 114, 6860.
52. L. Jiang, G. Sun, Z. Zhou, W. Zhou and Q. Xin, Catal. Today, 2004, 93, 665.
53. K. W. Wang and C. T. Yeh, J. Colloid Interface Sci., 2008, 325, 203.
54. R. J. Wu, J. G. Wu, T. K. Tsai and C. T. Yeh, Sens. Actuators, B, 2006, 120, 104.
55. A. J. Bard, R. Parsons and J. Jordan, Standard potentials in aqueous solutions, Marcel Dekker, New York, 1985, p. 287.
56. B. J. Hwang, L. S. Sarma, J. M. Chen, C. H. Chen, S. C. Shih, G. R. Wang, D. G. Liu, J. F. Lee and M. T. Tang, J. Am. Chem. Soc., 2005, 127, 11140.
57. S. Mukerjee, S. Srinivasan, M. P. Soriaga and J. McBreen, J. Electrochem. Soc., 1995, 142, 1409.
58. V. S. Murthi, R. C. Urian and S. Mukerjee, J. Phys. Chem. B, 2004, 108, 11011.
59. R. Kant and S. K. Rangarajan, J. Electroanal. Chem., 1994, 368, 1.
60. U. A. Paulus, A. Wokaun and G. G. Scherer, J. Phys. Chem. B, 2002, 106, 4181.
61. F. Taufany, C. J. Pan, H. L. Chou, J. Rick, Y. S. Chen, D. G. Liu, J. F. Lee, M. T. Tang and B. J. Hwang, Chem.–Eur. J., 2011, 17, 10724.
62. Z. W. Chen, M. Waje, W. Z. Li and Y. S. Yan, Angew. Chem., Int. Ed., 2007, 46, 4060.
63. E. Irissou, F. Laplante, S. Garbarino, M. Chaker and D. Guay, J. Phys. Chem. C, 2010, 114, 2192.
64. J. M. Lee, S. B. Han, J. Y. Kim, Y. W. Lee, A. R. Ko, B. Roh, I. Hwang and K. W. Park, Carbon, 2010, 48, 2290.

---

Footnote

† Electronic supplementary information (ESI) available: EXAFS fitting results of the Pd K-edge (Fig. S1), XANES and EXAFS spectra of the Au L_III-edge (Fig. S2), and TPR profiles of Au/C and Co/C (Fig. S3). See DOI: 10.1039/c2cy20136d

---