High Performance Ag Rich Pd-Ag Bimetallic Electrocatalyst for Ethylene Glycol Oxidation in Alkaline Media

Efﬁcientelectrocatalysts atdecreasingcostforethyleneglycol(EG)oxidationreactionarehighlydemandedforgeneratingelectricity out of direct EG fuel cells and producing hydrogen from EG electroreforming cells. In this work, carbon-supported Pd-Ag bimetallic catalysts (Pd x Ag y /C) with varying atomic ratios (x: y = 2: 1, 1: 1, 1: 2, 1: 3 and 1: 4) are synthesized by adding a Pd(II) precursor solution in a preformed Ag colloid in the presence of ascorbic acid followed by mixing with carbon black, and then are screened for EG electrooxidation in alkaline media. ICP-AES, TEM, XRD and XPS are used to characterize the obtained Pd x Ag y /C catalysts. The resulting Pd x Ag y nanoparticles exhibit similarly mean sizes of ca. 9.5 nm and an essentially alloy structure with a Pd surface enrichment rather than a typical Ag core-Pd shell structure. Electrocatalytic evaluation reveals that the as-synthesized Pd x Ag y /C catalysts display a “volcano” proﬁle in terms of the Pd mass activity with increasing y: x value, and the Pd 1 Ag 3 /C yields the most enhanced and durable activity among all the catalysts examined. Both electronic and bifunctional effects may account for this enhancement based on the existing guideline for bimetallic catalysts

Among various alcohols fed for low temperature fuel cells, ethylene glycol (EG) receives increasing interest, owing to its low toxicity and permeability, high boiling point (473 K), high theoretical energy density (7.56 kWh L −1 ) as well as renewability from biomass-based resources. 1,2 EG could also be used as a promising sacrificial molecule at the anode of a polymer electrolyte membrane-based electrolyzer targeting for hydrogen production at the cathode with a much lower voltage input as compared to a conventional water splitting electrolyzer. 3 Both the two types of applications involve the EG oxidation reaction (EGOR) at the anode catalysts in alkaline media. In fact, EGOR in alkaline media deserves a greater attention, given recent advancement of anion exchange membrane technology and relative facile EGOR kinetics due to easy formation OH ad species, less corrosive attack on catalysts and supports, minimal alcohol permeation rate, and use of Pt-free catalysts. 4,5 Along this line, Pd, with a lower cost and a larger reserve but a comparable activity, is a promising catalytic metal for EGOR in alkaline media to substitute Pt. [6][7][8] Nevertheless, EGOR on monometallic Pd surfaces still suffer from limited electrocatalytic activity and durability, mainly because of the accumulation of carbonaceous species including CO and HO-CH x as a result of the C-C bond cleavage. 9 On one hand, the CO and HO-CH x intermediates are essential for the C1 pathway to form the desired CO 2 and its derivative products; on the other hand, severe accumulation of these species if not removed at a sufficient rate blocks the active sites for EGOR. 9,10 Thus, new catalysts need developing, in order to address the above concern. To this end, a number of bimetallic [11][12][13][14][15] and trimetallic [16][17][18][19][20] Pd-based catalysts or Pd-based catalysts supported on transition metal oxidecarbon composites have been reported, 21 showing improved performance. Among them, bimetallic Pd catalysts deserve special attention, given their simplicity in controlled synthesis and characterization as well as in data explanation.
According to the catalyst design guideline based on the d-band theory proposed by Nørskov et al. and other scholars, 22,23 the d-band center of Pd may shift upward and downward for the bimetallic catz E-mail: liqiaoxia@shiep.edu.cn; wbcai@fudan.edu.cn alysts in either alloy or core-shell structure, corresponding to the enhanced and weakened adsorption of surface species, thus may change the surface reactivity toward targeted reactions. In addition to the electronic effect, the bifunctional effect in the bimetallic Pd catalysts may also affect the surface reactivity. Therefore, synthesizing and screening bimetallic Pd catalysts provides an effective way to enhance the activity and stability of Pd toward many electrocatalytic reactions.
Pd-Ag bimetallic catalysts have been extensively investigated for electrocatalytic oxidation of alcohols in alkaline media, including but not limited to methanol oxidation reaction (MOR), 24,25 ethanol oxidation reaction (EOR), [26][27][28][29][30][31][32] glycerol oxidation reaction (GOR), 33 and propanol. 34 As for the electrocatalysis of EGOR, two tactics are mainly adopted in the synthesis of Pd-Ag bimetallic catalysts: one is the morphological control and the other is compositional control. Du et al. 35 and Zhu et al. 36 synthesized unsupported Pd-Ag nanoflowers (NFs) by using CTAB and PVP respectively as the surfactant with ascorbic acid as the reductant for the coreduction of Pd(II) and Ag(I), Pd 1 Ag 1 and Pd 1 Ag 2 alloy NFs were reported respectively to exhibit the optimal performance toward EGOR. In practice, carbon supported Pdbased catalysts are more popular given the better utilization of precious metals. Along this line, Lei et al. 37 reported carbon supported heterostructured Pd-Ag catalyst through two-step ethylene glycol synthesis at raising temperatures, showing also an enhanced activity for EGOR, but no compositional dependence was reported. Li et al. 38 prepared Pd, Pd 1 Ag 1 and Pd 1 Ag 3 catalysts on carboxylic acid functionalized carbon nanotubes (CNTs) by co-reduction of Pd(II) and Ag(I) with NaBH 4 , suggesting that the electrocatalytic activity toward EGOR was in the order of Pd 1 Ag 1 /CNTs > Pd/CNTs > Pd 1 Ag 3 /CNTs. In other words, there is no consensus regarding the optimal composition for Pd-Ag bimetallic catalysts toward EGOR. Whether and what Ag-rich Pd-Ag bimetallic catalyst exhibits the highest Pd mass activity toward EGOR remains unclear, calling for a comprehensive investigation into bimetallic Pd-Ag catalysts obtained by new controlled synthesis.
In the present work, a series of carbon supported Pd x Ag y (with x: y values varied from 2: 1 to 1: 4) bimetallic catalysts with same metal loadings and close particle sizes were obtained by a unique seeded-growth in aqueous phase using the preformed Ag colloidal nanoparticles as seeds followed by unique restructuring of Ag and Pd. Attention was paid in the synthesis to avoid the use of hardly removed surfactants. The resulting Pd x Ag y /C catalysts show essentially an alloy nature with surface Pd enrichment, the x: y dependency on electrocatalytic activity and stability for EGOR was examined carefully, revealing that the Pd 1 Ag 3 /C catalyst was the best one for the desired reaction.

Experimental
Reagents.-PdCl 2 (A.R.), Pd(NO 3 ) 2 ·2H 2 O (A.R.) and Vulcan XC-72 carbon black were obtained from Aldrich and Cabot, respectively. NaBH 4 (A.R., purity ≥96%), NaCl(A.R., purity ≥96%) and AgNO 3 (A.R., purity ≥99.8%) were purchased from Sinopharm Chemical Reagent Company. All the aqueous solutions were prepared with 18. Catalyst synthesis.-The nominal metal loading for all catalysts was 20 wt%. All the Pd x Ag y bimetallic catalysts were prepared by using the seeded-growth method, in which an Ag colloidal solution was synthesized and then mixed with a PdCl 4 2− solution, followed by addition of ascorbic acid, and finally loaded on carbon black. Specifically, the procedures for preparing the most promising (vide infra) Pd 1 Ag 3 /C catalyst is briefly described as follows: 130 mg of sodium citrate and 17.7 mg of AgNO 3 were mixed in 50 mL of H 2 O. Under vigorously stirring, 15 mL of freshly prepared ice-cold 0.01 M NaBH 4 aqueous solution was added dropwise into the above solution through a constant-flow pump at 0.5 mL · min −1 , the reduction temperature was maintained for 3 h by using an ice-water bath. The resultant solution was kept stirring at 30 • C for 4 h, followed by mixing with 3.5 mL of 0.01 M Na 2 PdCl 4 . Then 5 mL of freshly prepared 0.1 M ascorbic acid aqueous solution was then added dropwise to the solution through a constant-flow pump at 0.5 mL · min −1 . Such a mixture was vigorously stirred for 1 h, and then 45 mg of Vulcan XC-72 carbon pretreated with HNO 3 was added to the solution. The powder in the solution was stirred continuously at 30 • C for 48 h, then filtered and rinsed with ultrapure Milli-Q water repeatedly. The Pd 1 Ag 3 /C catalyst was finally vacuum-dried at 50 • C overnight. For preparing the other Pd x Ag y /C bimetallic catalysts, all the other procedures are largely same except the amounts of the Pd(II) and Ag(I) precursors and carbon black were adjusted according to the atomic Pd/Ag ratio.
For the synthesis of the Ag/C, Ag colloidal nanoparticles was directly loaded on carbon black support without adding Pd(II) precursor and ascorbic acid in the Ag colloid. The Pd/C was prepared with NaBH 4 as the reducing agent. Briefly, 3.15 mL of 50 mM Na 2 PdCl 4 , 250 mg of sodium citrate and 67 mg of Vulcan XC-72 carbon were mixed in 50 mL of H 2 O. The mixture was ultrasonicated for 20 min to disperse carbon powder. Then, 15 mL of a freshly prepared ice-cold 0.1 M NaBH 4 aqueous solution was added dropwise into the aqueous mixture under vigorous stirring through a constant-flow pump at 0.5 mL · min −1 in an ice-water bath. The suspension was kept stirring for 3 h in the ice-water bath and another 12 h at 30 • C. After that, the powder in the solution was filtered and rinsed with ultrapure Milli-Q water repeatedly, and vacuum-dried at 70 • C for overnight.
For comparison, another Pd 1 Ag 3 /C (20 wt%) catalyst was also prepared by using the coreduction of Pd(NO 3 ) 2 and AgNO 3 with NaBH 4 as the reductant, according to the report of Reference 39 (noted as Pd 1 Ag 3 /C-NaBH 4 ). 39 Materials characterization.-The compositions of as-synthesized catalysts were determined on inductively coupled plasma-atomic emission spectrometry (ICP-AES, Hitachi P-4010). The structures of these catalysts were examined by X-ray diffraction (XRD, D8 Advance X-ray Diffractometer) with the Cu Kα radiation at the 2θ angle from 30 • to 90 • . The morphology and size distribution of the catalysts were characterized by transmission electron microscopy (TEM, JEOL JEM-2010 microscope). The binding energies of core level electrons of Pd and Ag were analyzed by X-ray photoelectron spectroscopy (XPS, Perkin-Elmer PHI-5000C ECSA) with the Mg Kα radiation (1253.6 eV), and the C 1s peak at 284.6 eV was used as the reference for calibration.
Electrochemical measurements.-The working electrode was a glassy carbon electrode (GCE, 3 mm in diameter) coated with a catalyst layer. The catalyst ink was prepared by mixing 1 ml of ethanol and 120 μL of 5 wt% Nafion solution and 2 mg of a catalyst and sonicating for 20 min, and 5.6 μL of this ink was transferred onto a freshly polished GCE and dried naturally to form a catalyst layer. A platinum foil and a saturated calomel electrode (SCE) served as counter and reference electrodes, respectively. The electrochemical measurements were run with a CHI 660B electrochemistry workstation.
The cyclic voltammetry (CV) was run on a GCE-supported catalyst layer in a N 2 -saturated 1 M NaOH solution at 50 mV s −1 within the potential range from −1.0 V to 0.1 V. The electrochemical surface area (ECSA) was calculated by integrating the charge associated with the peak area of CO desorption region through anodic CO stripping experiment, first by bubbling CO (>99.9% purity) for 20 min, and then purging the dissolved CO by bubbling N 2 for 40 min while holding the potential at −0.9 V. At last, cyclic voltammogram for the CO stripping was recorded from −1.0 V to 0.1 V at 50 mV s −1 . The corresponding ECSA of a catalyst was evaluated by using Equation 1: [ 1 ] in which Q is the electro-oxidation charge of the CO ad-layer, m is the total mass of Pd loaded on GCE, and Q co is approximately 0.42 mC cm −2 for the oxidation of a monolayer of CO on smooth Pd surface. Cyclic voltammograms of EGOR on different catalysts were measured in 1 M ethylene glycol + 1 M NaOH solution between −1.0 V and 0.1 V at 50 mV s −1 while chronoamperometric curves were recorded in the same solution at −0.3 V for 3600 s. All electrochemical measurements were run at 18 ± 1 • C.

Results and Discussion
TEM and HRTEM.-The metal weight percentages and the Pd: Ag atomic ratios for the as-synthesized Pd-based catalysts are close to the desired values within measurement errors according to the ICP-AES assay, as listed in Table I. Figs. 1a-1f are TEM images for Pd 2 Ag 1 /C, Pd 1 Ag 1 /C, Pd 1 Ag 2 /C, Pd 1 Ag 3 /C, Pd 1 Ag 4 /C and Pd/C, respectively, showing that metal nanoparticles are well dispersed on carbon black with approximately spherical shape of similar mean sizes of ca. 9.5 nm, facilitating the comparison of electrochemical behaviors of different Pd x Ag y /C catalysts. The EDS mapping patterns for Pd 1 Ag 3 /C in Figs. 1g-1j in a larger scale suggest that Pd and Ag elements are distributed rather uniformly from particle to particle.
The structure and composition of an arbitrarily selected metallic nanoparticle from the Pd 1 Ag 3 /C catalyst were further characterized by HRTEM as well as EDS elemental analysis. The HRTEM image of the Pd 1 Ag 3 nanoparticle showing in Fig. 2a reveals a d-  a (111) plane of an face-centered-cubic (fcc) structured crystal, this value is slightly larger than that for Pd(111), or 2.24 Å, but smaller than that for Ag(111), or 2.36 Å. At a first glance, our preparation procedures might lead to a distinct Ag core-Pd shell structure, however, this may not be true. On one hand, due to different equilibrium redox potentials, i.e, ϕ 0 Ag + /Ag (0.79 V vs SHE) and ϕ 0 Pd 2+ /Pd (0.95 V vs SHE), galvanic displacement reaction occurred to some extent when Ag seeds met with Pd(II) species, in which Ag + ions diffused out-wards and Pd atoms deposited inwards. As ascorbic acid was dropwise added, both the resulting Ag + ions and the remaining Pd(II) species were co-reduced gradually. 27 This, together with the inter-diffusion of two metals, helps to mix Ag and Pd atoms, to form an alloy structure with surface Pd enriched. On the other hand, the mismatch of lattice constants between Pd (3.89 Å) and Ag (4.09 Å) is as high as 5.14%, resulting in the non-coherent interface with a large density of defects, which may further prevented Pd conformal overgrowth on the  preformed Ag seeds to form distinct Ag core-Pd shell structure. 40,41 The galvanic displacement and the lattice mismatch may also lead to reconstruction and formation of smaller Pd-Ag bimetallic nanoparticles. 27 The evidence of an alloy structure can be found later in the XRD characterization, as neither isolated Ag nor Pd phase can be clearly identified. In fact, Figs. 2b-2d indicate that the Pd and Ag elements had almost consistent mapping across the nanoparticle, a replot of the Pd/Ag elemental signal ratio in Fig. 2f along the arrow direction in Fig. 2e further suggests relative higher Pd concentration in the outer layers of the Pd 1 Ag 3 alloy nanoparticles. In other words, the Pd 1 Ag 3 nanoparticles assume an alloy structure with surface Pd enrichment, in agreement with the following XPS analysis.
XRD characterization.-The XRD patterns for Pd/C and Pd x Ag y /C catalysts are shown in Fig. 3, where the four features correspond to the fcc Pd crystalline structure. Sharp and intense XRD peaks for the Ag/C suggest larger Ag nanoparticles. In fact, the formation of Ag colloidal particles with a mean size of ca. 20 nm was found using the similar synthetic procedures. 42 The relative broader peaks with Pd x Ag y /C are consistent with smaller particle sizes, or ca. 9.5 nm, suggesting the formation of the Pd-Ag nanoalloy from Ag colloidal seeds is probably accompanied with restricted aggregation or geometric restructuring of Ag nanoparticles. If it were a typical Ag-core@Pd-shell structure, one would expect rather significant XRD features for the monometallic Ag core, especially for Ag-rich bimetallic catalysts. As mentioned in the above, the preferred alloy structure formation may arise from significant lattice mismatch between Ag and Pd and in-out inter-diffusion. The characteristic diffraction peaks of the Pd x Ag y /C catalysts shift toward smaller angles from those of Pd/C with increasing y: x value, owing to a larger Ag lattice constant. 28,30 Additionally, as arrowhead-marked in Fig. 3, for the Pd 1 Ag 2 /C, Pd 1 Ag 3 /C, Pd 1 Ag 4 /C and Ag/C catalysts a weaker shoulder peak around ca. 36 • possibly due to Ag 2 O resulted from the aerial oxidation of exposed surface Ag atoms. 37 XPS analysis.-X-ray photoelectron spectroscopy (XPS) was further conducted to determine the surface composition and the oxidation state of the two metals of the Pd/C, Pd 2 Ag 1 /C and Pd 1 Ag 3 /C catalysts (Fig. 4). Shown in Fig. 4a, the binding energy of Pd 3d 5/2 electrons is positively shifted by 0.35 eV for Pd 1 Ag 3 /C as compared to that for Pd/C. In contrast, Fig. 4b shows that the binding energy of Ag 3d 5/2 electrons is negatively shifted by 0.50 eV as compared to that for monometallic Ag. Apart from Ag(0) species, Ag(I) also accounts for a significant composition of the surface layers determined by ex situ XPS, in line with the above XRD result. The shifts of core-level binding energies for Ag(0) and Pd(0), consistent with the result reported by Lam et al., 43 indicate a partial electron transfer between these two metals. Furthermore, based on the XPS peak areas and corresponding calibration factors, the Pd: Ag molar ratio estimated is ca. 2.1: 1 for the outer layers of the Pd 1 Ag 3 nanoparticles, significantly higher than the bulk average value (ca. 1: 2.8) as determined by ICP-AES. This, together with the above STEM EDS line scan elemental analysis result, confirms a surface Pd-enriched nanoalloy structure.
Normally, for a metal of different valences (like in metal oxides), the gain (or loss) of partial electron would cause a decrease (or increase) of binding energies of its core level electrons in XPS spectrum. However, this may not apply in the case of bimetallic alloys. In fact, M. Watanabe et al. 44 45 Also notably reported is that the direction of the core-level shift is commensurate with that of the valence band shift. 44 Accordingly, the XPS result suggests the d-band center of Pd in the as-synthesized Pd 2 Ag 1 /C or Pd 1 Ag 3 /C is downshifted and that of Ag upshifted. This suggestion is to some  extent different from a theoretical prediction. According to the DFT model calculation on random Pd-Ag alloy, the d-band center of Pd is upshifted slightly by 0.14 eV while that of Ag is upshifted moderately by 0.59 eV. 22,23 We attributes tentatively the difference to unique structure and composition of the as-synthesized Pd 1 Ag 3 /C. It should be pointed out that using the lattice expansion or shrinkage as the only criterion for the upshift or downshift of the d-band center is not sufficient, 39,46 the electron transfer in between should also take effect. In fact, both the theoretical prediction and the present XPS result indicate that the Ag d-band center is upshifted moderately despite its lattice shrinks in Pd-Ag alloys, suggesting the electron transfer effect is more significant in determining the Ag d-band center in the case of Pd-Ag alloys. Similarly, for B-doped Pd, the Pd lattice expands but its d-band center is downshifted. 47,48 The modified electronic properties of Pd and Ag favor the electrocatalytic EGOR. Fig.  5 are the cyclic voltammograms (CVs) for the Pd x Ag y /C and Pd/C catalysts in a N 2 -saturated 1 M NaOH solution, with currents being normalized to the Pd masses loaded (Fig. 5a) and the geometric electrode areas (Fig. 5b), respectively. Since the sum of Pd and Ag mass loadings is same for the electrochemical measurement, Fig. 5a facilitates the comparison of Pd-mass activities of different catalysts. Fig. 5b presents the apparent current densities observed on the same glassy carbon electrode respectively coated with different catalysts. The Pd/C and Pd x Ag y /C catalysts produced a typical voltammetric profile in the alkaline media with corresponding cathodic peaks appearing at ca. −0.4 V being attributable to the reduction of PdO to Pd. The PdO to Pd peak areas in Fig. 5b reflected the exposed surface Pd sites of the tested catalyst layers of a given total metal mass. Specifically, the Pd/C catalyst layer yields the largest surface active area, consistent with its higher Pd loading and much smaller Pd particle sizes. Also it was noted that the Pd x Ag y /C catalysts did not exhibit significant H adsorption/desorption peaks, especially for the Ag-rich ones, again in accordance with the modified electronic property of Pd. It was mentioned that the decreased H diffusivity in Pd-Ag bimetallic alloys may also minimize the H adsorption/desorption feature. 28,49 The electrocatalytic activities of the Pd x Ag y /C catalysts toward EGOR were examined by cyclic voltammetry. Fig. 6a shows the typical Pd-mass normalized CVs for the Pd x Ag y /C catalysts in 1 M NaOH solution containing 1 M EG at 50 mV s −1 and Fig. 6b the corresponding mass and specific activities, using the peak currents in the positive-going potential (forward) scan to evaluate the activities while the electrochemical surface area (ECSA) of each catalyst was obtained by anodic CO stripping measurement and listed in Table  I. It can be seen that Pd 1 Ag 3 /C exhibited both the highest Pd-mass activity and specific activity, i.e., 7.93 A mg −1 Pd and 7.82 mA cm −2 , which are 5.54 and 3.15 times higher than those of Pd/C, i.e., 1.43 A mg −1 Pd and 2.31 mA cm −2 , respectively. In addition, the onset oxidation potential on the Pd 1 Ag 3 /C is −0.69 V, that is, the most negative one among those detected for all the tested catalysts. Additionally, the ratio between the peak current in the forward scan (I f ) and that in the backward scan (I b ) has been often used to roughly indicate the tolerance of a catalyst surface to poisoning species: the higher the ratio, the higher tolerance. 30 As Table II shows, the I f /I b ratio for a Pd x Ag y /C catalyst was higher than that for Pd/C, and the ratio for the Pd 1 Ag 3 /C is among the highest, indicating its potentially good anti-poisoning property toward EGOR. Further durability verification of the catalysts toward EGOR at constant potential will be compared using chronoamperometry curves (vide infra). The parameters in Table II demonstrate a "volcano-like" dependence of electrocatalytic activity on bimetallic composition x: y, and the Pd 1 Ag 3 /C catalyst exhibits the most durable oxidation current. As discussed in the above section, the electron transfer occurs from Ag to Pd. It is known that the C-C bond cleavage for ethanol or ethylene glycol is favored on Pt or Pd electrode at negative potentials. 20 Along this line, electron-enriched Pd surface may also favor the C-C bond cleavage during EGOR. Moreover, the downshift of Pd d-band center in the as-synthesized Pd x Ag y /C catalyst may weaken the adsorption of possible poisoning intermediates such as CO and HO-CH x on Pd sites from the C-C bond cleavage, 9 in favor of leaving more Pd active sites. On the other hand, although Ag is largely inactive for EGOR as seen in the inset of Fig. 6a, the moderate d-band upshift of Ag favors the adsorption of OH ad , or the enhancement of the oxophilicity of Ag. In other words, Ag provides more OH ad at lower potentials to promote the EGOR process including the removal of the poisoning species on adjacent Pd sites, given that OH ad is the essential reactant pair. 9,39,50 Synergically, the Pd x Ag y /C catalysts demonstrated significantly upgraded electrocatalytic activities, featuring with a negative shifted onset oxidation potential and an increased oxidation peak current. Nevertheless, too high Ag content may reduce significantly the active Pd sites on the surface, which may account for the start of decrease in electrocatalytic activity when the Pd: Ag atomic ratio changes from 1: 3 to 1: 4. In this regard, the role of Ag in Pd-Ag alloy catalyst toward EGOR is similar to that of Ru in Pt-Ru alloy catalyst toward methanol oxidation reaction, in which both electronic effect and bifunctional effect synergically account for.

Electrocatalytic properties of Pd x Ag y /C catalysts.-Shown in
To highlight the advantage of the present seeded-growth synthesis, carbon supported Pd 1 Ag 3 catalyst was also prepared by the traditional NaBH 4 co-reduction method and tested for EGOR in otherwise same conditions. The TEM image in Fig. 7a shows that serious agglomeration of metallic nanoparticles existed in the Pd 1 Ag 3 /C-NaBH 4 catalyst, and Fig. 7b indicates the corresponding CV peak current toward EGOR on Pd 1 Ag 3 /C-NaBH 4 is only 3.92 A mg −1 Pd , much lower than that on Pd 1 Ag 3 /C, and the peak potential of the former is also more positive than that of the latter.
In order to further evaluate the durable electrocatalytic activity of the Pd x Ag y /C catalysts toward EGOR in the same solution, the chronoamperometry was applied at −0.3 V for 3600 s. Fig. 8 indicates that Pd 1 Ag 3 /C kept the highest oxidation current during the whole process, for example, at 3600 s the mass-activity was 2.79 A mg −1 Pd for this catalyst, that is 3.3, 2.6, 1.3, 3.6 and 8 times as high as that observed on Pd 2 Ag 1 /C, Pd 1 Ag 1 /C, Pd 1 Ag 2 /C, Pd 1 Ag 4 /C and Pd/C, respectively. Notably, the Pd 1 Ag 4 /C catalyst deactivated to a larger extent than expected from the above CV measurement. Based on all the above characterizations and measurements, we can conclude that the as-synthesized Pd 1 Ag 3 /C is the optimal bimetallic catalysts under examined for EGOR in alkaline media.

Conclusions
Carbon supported Pd x Ag y bimetallic catalysts with varying compositions have been synthesized through seeded-growth method using Ag colloid as seeds, resulting in Pd x Ag y nanoparticle sizes of ca. 9.5 nm with an essentially alloy structure with surface Pd enrichment, as characterized by variety of methods. Electrocatalytic evaluation reveals that the as-synthesized Pd x Ag y /C catalysts display a "volcanolike" relationship for the Pd-mass activity as a function of bimetallic composition, and the Pd 1 Ag 3 /C catalyst yields the most enhanced and durable EGOR oxidation current among all the catalysts examined. The slightly downshifted d-band center of Pd and the bifunctional effect of more oxophilic Ag may account for this enhancement based on the existing guideline for bimetallic catalysts. The present work provides a unique synthetic tactic for obtaining carbon supported Pd-Ag bimetallic nanoalloy catalysts with potential application in efficient electrooxidation of ethylene glycol, thus promoting the development