Beneficial Interaction of Gold and Palladium in Bimetallic Catalysts for

Mar 24, 2009 - Promotion of Activity and Selectivity by Alkanethiol Monolayers for ...... alcohol oxidation over platinum catalysts in dioxane/aqueous...
0 downloads 3 Views 1MB Size
Download PDF
J. Phys. Chem. C 2009, 113, 6191–6201

6191

Beneficial Interaction of Gold and Palladium in Bimetallic Catalysts for the Selective Oxidation of Benzyl Alcohol Stefan Marx and Alfons Baiker* Department of Chemistry and Applied Biosciences, ETH Zurich, Ho¨nggerberg, HCI, CH-8093 Zu¨rich, Switzerland ReceiVed: September 20, 2008; ReVised Manuscript ReceiVed: February 25, 2009

Bi- and monometallic nanoparticles of Au and Pd with a rather narrow size distribution were deposited on polyaniline (PANI) and their structural properties as well as catalytic behavior in the aerobic oxidation of benzyl alcohol were investigated. The size of the mono- and bimetallic particles was controlled in a narrow range (2.4-3.7 nm) using a colloidal preparation route. Admixing Pd to Au resulted in a strong enhancement of selectivity to benzaldehyde reaching a maximum of 98% at full conversion at 100 °C with bimetallic particles containing Au/Pd in a ratio of 1:9. Pure Au particles were significantly more active than pure palladium particles of the same size. Chemical, structural, and electronic properties of the bimetallic catalysts were characterized using high angle annular dark field scanning transmission electron microscopy, atomic absorption spectroscopy, X-ray photoelectron spectroscopy (XPS), and X-ray absorption spectroscopy. The PANI-supported nanoparticles showed a core shell structure with an Au enriched core and a Pd rich shell. The electronic changes occurring upon admixing Pd to Au were examined with X-ray absorption near-edge specroscopy and XPS. The binding energy of core level electrons and the valence d-band occupation in the bimetallic particles were significantly altered in comparison to the monometallic particles, which together with the core shell structure is supposed to be the main reason for the observed changes in the catalytic behavior. 1. Introduction Bimetallic catalysts often exhibit advantageous properties compared to those of their pure constituent metals.1-3 Mixing metals in ratios that are allowed by the phase diagram offers the opportunity of target orientated tuning of the catalysts properties. A bimetallic system that has gained considerable attention is gold-palladium. Its excellent catalytic performance was shown for the acetoxylation of ethylene to vinyl acetate,4 solvent free oxidations of primary alcohols to aldehydes,5 glycerol oxidation,6 hydrochlorination of acetylene,7 and the direct synthesis of H2O2 from H2 and O2.8-13 Several investigations toward the particle structure have been performed. Depending on the synthesis method and post synthetic treatment, particles with formulations between a homogeneous alloy14-18 and a core-shell structure19-25 have been identified. When focusing on the intrinsic peculiarities of bimetallic nanoparticles with disregard of the support-metal interaction, two properties are appreciably altered: structural properties and electronic configuration.26-28 Both effects are used as an explanation for changes in catalytic performance. By altering the alloy composition, a change in the interatomic distance due to Vegard’s law occurs.18,29,30 Therefore, a proper distance between the adsorption sites can be adjusted, which facilitates the chemical conversion step.4 Theoretical studies reveal that upon alloying, a rougher surface with more low coordinated sites compared to the pure metals is established.16 Likewise, ensemble size effects have been discussed where the active metal is diluted and surrounded by an ensemble of an inert, second metal.1,31 Other investigations focus on the electronic perturbations that occur upon alloying two metals. Depending on the model of adsorbate binding established by Nørskov et al.32-35 the adsorbate strength depends * To whom correspondence should be addressed. E-mail: baiker@ chem.ethz.ch. Tel: +41 44 632 3153. Fax: +41 44 632 1163.

on the energy position and relative distance of the transition metal d-band to the Fermi level. The adsorbate strength becomes higher with increasing density of states (DOS) near the Fermi level. While weak substrate surface interaction leads to low surface coverage and results in low reaction rates, strong interactions lead to a poisoning of the catalyst surface. Mixing two metals in different ratios should therefore result in a volcano-type behavior of the adsorption energy when going from one pure metal to the other. To gain insight into the electronic structure of mono- and bimetallic particles, several techniques have been applied.36-43 Soft X-ray photoelectron spectroscopy (sXPS) has been applied to examine the width, shape, and energetic level of the valence band, and a clear hybrid state between the 4d Pd band and the 5d Au band could be detected.44 X-ray photoelectron spectroscopy (XPS) analysis examines not only surface composition, but it also yields information about the width and position of core level bands.22,44-46 To investigate the density of unfilled states above the Fermi level X-ray absorption fine structure (XAFS) studies have been applied to Au-Pd alloys.44,47-50 In this work, we examined the structural and catalytic properties of polymer-supported Au/Pd bimetallic nanoparticles. As polymer support polyaniline (PANI) was applied, which has recently attracted considerable attention for catalytic reactions as a catalyst51-54 and support55-60 owing to its high conducting and redox properties, nonsolubility in most organic solvents and water as well as controllable doping through acid- and basedoping. The structural and chemical properties of the colloidderived bimetallic Au/Pd nanoparticles of different composition were characterized applying atomic absorption spectroscopy (AAS), electron microscopy (high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM), high-resolution transmission electron microscopy (HRTEM)), X-ray photoelectron spectroscopy (XPS), and X-ray absorption

10.1021/jp808362m CCC: $40.75  2009 American Chemical Society Published on Web 03/24/2009

6192 J. Phys. Chem. C, Vol. 113, No. 15, 2009 SCHEME 1: Structure of Polyaniline in the Emeraldine Form (x ) 0.5)

spectroscopy (EXAFS, X-ray absorption near-edge specroscopy (XANES)). The selective oxidation of benzyl alcohol to benzaldehyde served as a test reaction for elucidation of the catalytic properties of the bimetallic Au/Pd nanoparticles. By applying a colloidal route for the synthesis of the mono- and bimetallic nanoparticles, a uniform and rather small size distribution for each composition could be achieved. Therefore a comparison of the electronic properties between the bimetallic nanoparticles of different composition was possible, and an attempt was undertaken to relate the changes observed in the catalytic behavior upon admixing Pd to Au to the changes in chemical, structural, and electronic properties. 2. Experimental Section Catalyst Preparation. Preparation of the mono- and bimetallic particles was performed in accordance to the route that was used earlier in our group for the synthesis of gold nanoparticles.61 In brief, to a basic solution of THPC [tetra(hydroxymethyl)phosphoniumchloride] in water an aqueous solution of the mixed metal precursors was added under rapid stirring. NaOH was used as base and HAuCl4 and/or Na2PdCl4 were used as metal precursors. The molar ratio of base/THPC/Σmetal was 6:1.4:1. By varying the ratio of the two metals, a set of different bimetallic catalysts has been prepared. The nominal molar ratio of gold to palladium was 1:0, 9:1, 1:1, 1:9, and 0:1. The amount of metal used was calculated to obtain a metal loading of the support of 2 wt %. After mixing the metal precursors and adding the mixture to the THPC solution a dark colloidal solution was obtained. The color shifted from dark brown for pure gold nanoparticles to black for pure palladium. After the reduction step the colloidal solution was added to a suspension of PANI (Scheme 1) in water. PANI was received from Aldrich with an average molecular weight of 65 000 g/mol. It was present in the emeraldine form, in which the number of oxidized (quinoide) and the reduced (benzoid) units is equal. The Brunauer-Emmett-Teller (BET) surface area of PANI as received was 6.9 m2/g. TG experiments revealed that the decomposition of PANI starts at about 440 °C in helium (Netzsch, STR449). By acidification of the suspension (pH 3-4) the nanoparticles adsorbed on the polymer. Filtration, washing of the residue, and drying the powder at 80 °C for 15 h under ambient conditions gave the as-prepared catalyst. Catalytic Test Reaction. The oxidation of benzyl alcohol was performed in a two-neck round-bottom flask. To the solution of benzyl alcohol in toluene an aqueous solution of NaOH, the catalyst powder, and tert-butylbenzene as internal standard were added. The suspension was kept under an atmosphere of oxygen for 3 h and the reaction temperature in a standard experiment was kept at 50 or 100 °C. For all reactions, the metal to substrate ratio varied between 1.5 and 1.7 mol % due to fluctuations in the metal loading of the catalyst. Three equivalents of the base have been used for the test reactions. After three hours of stirring, the reaction was quenched with 1 M HCl. After filtration and washing with saturated NaCl-solution and toluene, it was

Marx and Baiker possible to separate the phases and extract the aqueous phase with ether. The combined organic phases were used for determination of conversion and yield by gas chromatography (HP 6890 GC system, HP 5 column). Atomic Absorption Spectroscopy. Au and Pd loadings of PANI were determined by AAS. PANI was burned in a crucible and the remaining metals could be dissolved in aqua regia. The measurement was performed with a Varian SpectrAA 220 FS spectrometer. The matrix of the calibration standards was adjusted to the matrix of the sample. Electron Microscopy. Particle sizes and distributions were determined by HAADF-STEM and HRTEM. The STEM and TEM pictures were acquired with a Tecnai F30 microscope (FEI, (Eindhoven); field emission cathode, operated at 300 kV). STEM images were recorded with a HAADF detector, using almost exclusively incoherently scattered electrons (Rutherford scattering) to obtain images with atomic number (Z) contrast. BET. Nitrogen physisorption isotherms have been recorded on a Micromeritics ASAP 2010 instrument at 77 K. The samples have been outgassed for 4 h at 120 °C and the specific surface area was determined using the standard BET method. X-ray Photoelectron Spectroscopy. The measurements were performed on a Leybold Heraeus LHS11 MCD instrument using Mg KR (1253.6 eV) radiation or on a PHI Quantera SXM with an Al source (1486.6 eV). The samples were evacuated in a Load lock and transferred to the analysis chamber (10-9 mbar). The peaks were energy shifted to the binding energy of N 1s (399.5 eV) and C 1s (285.0 eV), respectively, to correct charging effects of the sample. The binding energies of the Au and Pd signals were determined by peak fitting. X-ray Absorption Spectroscopy. The experiments were performed at Hamburger Synchrotronstrahlungs-Labor (HASYLAB) at the Deutsches Elektronen-Synchrotron (DESY) at DORIS III (4.45 GeV, 120 mA current) at the beamline X1 (energy range, 7-100 keV) and C1 (energy range, 5-43 keV), at Forschungszentrum Karlsruhe (Helmoltz-Gemeinschaft) at the Angstroemquelle Karlsruhe (ANKA, 2.5 GeV storage ring, 200 mA current) using the XAS-beamline (energy range, 2.3-25 keV, resolution, 2 × 10-4 ∆E/E), and at the Swiss Light Source at the Paul Scherrer Institut (SLS, 2.4 GeV storage ring, 400 mA current) using the superXAS beamline (energy range, 4.5-35 keV, resolution, 2 × 10-4 ∆E /E). The monochromator was a Si (111) crystal for the measurement of the Au LIII edge (11.918 keV) and a Si (311) crystal for the measurement of the Pd K edge (24.35 keV). The samples were pressed into a pellet and measured in transmission mode. The scans were energy referenced to Au or Pd foil. For further investigations of the interaction of Pd and Au, the amount of remaining Pd2+ was further decreased by reduction with hydrogen. Therefore an in situ flow cell was filled with the as-prepared catalyst and a mixture of 5% H2 in He with a flow of 20 mL/min was purged through the cell. The cell was heated up to 240 °C (in the case of the 1:9 catalyst) and to 260 °C (in the case of the 1:1 catalyst), respectively, while XANES spectra were taken. The final temperature was kept as long as no changes in two sequent XANES spectra could be visible. After cooling to room temperature, EXAFS spectra of the samples were taken. Data analysis was performed with the use of the software package WinXAS 3.1. After the χ(k) function was extracted from the EXAFS data, Fourier transformation was performed on the k3-weighted data in the interval k ) 3.4-13.1 Å-1 for the Au spectra and k ) 3.2-13.1 Å-1 for the Pd spectra.

Polyaniline Supported Au/Pd Nanoparticles

J. Phys. Chem. C, Vol. 113, No. 15, 2009 6193

TABLE 1: Comparison of the Obtained Loadings with the Theoretical Nominal Values (First and Last Column Contain the Molar Ratios; the Two Middle Columns Contain the Wt %) Au/Pd mol % (expected) 1:0 9:1 1:1 1:9 0:1

loading wt % (expected)

loading wt % (obtained)

Au/Pd mol % (obtained)

2.00:0.0 1.89:0.11 1.30:0.70 0.34:1.66 0.0:2.00

1.50:0.0 1.43:0.11 1.20:0.63 0.32:1.5 0.0:1.55

1:0 7:1 1:1 1:9 0:1

Data analysis in the R-space was performed using Au-Au/ Pd and Pd-Pd/Au/O/P shells calculated by FEFF 7.0.62 To simulate the scattering paths, a feff input file was generated from the crystal structures of gold, palladium, PdCl2, PdO, and Pd3P. Amplitudes and phase shifts of the theoretical calculated scattering paths have been optimized by the experimentally obtained amplitudes and phase shifts of the Au and Pd reference foils. 3. Results and Discussion 3.1. Bulk and Surface Chemical Composition. The series of PANI supported Au/Pd catalysts with different molar ratios of gold to palladium were subjected to structural and chemical analyses using various bulk and surface characterization methods. The determination of the metal loading of the catalysts by means of AAS indicated that the loadings of the metals (Au and Pd) were generally lower than the expected nominal values. The total amount of the adsorbed metal strongly depended on the pH during the adsorption step. The metal nanoparticles are negatively charged due to adsorbed chloride on the surface. By regulating the polymer’s surface charge with protons (optimum: pH ) 3-4) a maximum amount of the metal nanoparticle could be adsorbed on PANI. Table 1 lists the observed differences between nominal (expected) and actual metal loadings and Au/ Pd molar ratios, respectively. The catalysts in this paper are denoted by referring the metal(s) followed by the actual molar ratio, for example, Au/Pd 1:9. Besides the characterization of the bulk composition of the particles, the surface composition was analyzed with XPS (Figure 1). The change of the Au/Pd ratio is also reflected by the surface composition. As expected, the typical doublets of the 4f and 4d core level bands of gold become smaller with increasing Pd content. Moreover, a shift of the peak position of gold with increasing Pd content was observed (Figure 1b). The maximum shift of 0.7 eV relative to the binding energy of pure gold nanoparticles could be observed for the Au/Pd 1:9 mixture (see below). The XPS analysis for the Pd revealed that the surface of the nanoparticles consists of Pd2+. The peak intensity increases with increasing Pd content. Only for the Au/ Pd 1:9 catalyst a clear contribution of Pd0 was found. In contrast to the results of the XPS analysis, EXAFS as a bulk technique revealed a Pd0 contribution even for the Au/Pd 1:1 and pure Pd catalyst (see below). An interesting feature became apparent when the composition of impurities was examined in comparison with the composition of the catalysts. The phosphorus content, left from the reduction procedure, was highest for the Au catalyst and diminished with increasing Pd content. For the chlorine, the opposite trend was observed. It was present in relatively high concentrations in the pure Pd catalyst and decreased with increasing Au content. Therefore, the impurities should be located mainly on the metal particles and only to a lesser extent on the PANI itself.

Figure 1. XPS spectra of the (bi)metallic catalysts in the Au 4d and Pd 3d region (a) and Au 4f region (b); Au (gray, dashed); Au/Pd 7:1 (gray, solid); Au/Pd 1:1 (black, solid); Au/Pd 1:9 (black, dashed); Pd (black, dotted).

Otherwise an almost constant amount of impurities that does not change with the catalyst composition should have been observed. That means that the Au rich particles might have been surrounded by a layer of a phosphorus species, which remained from the reducing step and which stabilized the particles. With increasing Pd content, the phosphorus species diminished, which could be explained by an overlayer of Pd on the Au particle. The increasing chlorine content with increasing molar fraction of Pd might be caused by Na2PdCl4 that was not completely reduced by THPC. 3.2. Structural Properties. In previous studies, it has been shown that the gold-catalyzed aerobic oxidation of benzyl alcohol is particle size dependent, reaching a maximum for pure gold particles with a diameter around 6.9 nm.63 The colloidal preparation route applied to prepare the bimetallic nanoparticles led to nanoparticles of almost equal size, a necessary prerequisite to minimize possible particle size effects on structural and catalytic properties. Mean particle sizes and size distributions were obtained by analyzing HAADF-STEM pictures. The mean diameter of the nanoparticles ranged between 2.4 and 3.7 nm (Table 2). The size distribution was rather small due to the used colloidal route, in which the reducing reagent also serves as a stabilizer for the nanoparticles. STEM images, histograms and EDX spectra for the Au/Pd 1:1 and pure Pd catalysts are shown in Figure 2. The size distribution graphs of the nanoparticles are given in Figure 2b,e, respectively. Where it was possible, EDX spectra of single particles were taken to probe the composition of individual particles. Because of the small size of the particles the intensity of the signals was very low. Despite the poor quality of the EDX spectra, it could be confirmed that even for single particles both signals for gold and palladium

6194 J. Phys. Chem. C, Vol. 113, No. 15, 2009

Marx and Baiker

TABLE 2: Statistical Evaluation of the Nanoparticles Prepared in This Work; the Values for the Mean Diameter, the Standard Deviation, the Median, and the Mode Are Listed sample Au Au/Pd 7:1 Au/Pd 1:1 Au/Pd 1:9 Pd

mean (nm) rel. sigma (%) median (nm) mode (nm) 2.4 3.7 2.6 2.4 2.6

42 31 32 37 35

2.3 3.3 2.5 2.3 2.5

1.9 3.3 2.5 2.3 2.5

are observable and the intensity of the signals reflected the change of the composition of the bimetallic particles (Figure 2c,f). Nevertheless the existence of subnanometer particles consisting of only one metal could not be excluded. To gain further insight into the nature of the deposited particles XAS studies at the Au LIII edge and the Pd K edge were performed and qualitatively compared to reference materials like foils (Au, Au-Pd, Pd) and PdO. The simulations of the FT EXAFS spectra were performed with respect to the consistency parameters established by Via et al. which they used for examining bimetallic particles.64 The Au LIII EXAFS spectra of the catalyst series (pure Au, 7:1, 1:1, 1:9) are shown in Figure 3. All catalysts exhibited a feature in the region 1.7-3.5 Å due to the first shell backscatter contributions. No further features in the region between 4.0-6.7 Å, which are observable for gold foil, were found. The amplitudes of all measured samples were smaller than the one of the gold foil due to the lower coordination number in small particles. For the catalysts Au and Au/Pd 7:1, the maxima of their peaks arise at 2.6 Å and the shape of them is almost symmetric. The picture changes significantly for the two catalysts Au/Pd 1:1 and Au/Pd 1:9 with high Pd content. The peak positions are shifted with respect to the Au-foil and an Au-Pd backscatter contribution becomes significant as indicated by the peak at 2.2 Å for the Au/Pd 1:1 or the shoulder at 2.1 Å for the Au/Pd 1:9 catalyst. The Pd EXAFS spectrum shows a somewhat more complicated picture (Figure 4). All catalysts show their main feature at around 1.8 Å. Compared to the Pd- or Au-Pd-foil the maxima are shifted significantly to shorter distances, which means that the main part of the nearest neighbors are not Pd or Au but lighter atoms. In addition, the main peak shows a shoulder at shorter distances for all samples. That implies that at least two light scatterer could be involved in the origin of these EXAFS spectra. From XPS analysis, one can conclude that these light scatterer are chlorine, oxygen, and (for Au/Pd 7:1) even phosphorus. With the exception of Au/Pd 7:1, all other samples exhibit a small but significant peak between 2.6 and 2.8 Å. We attribute this peak to Pd-M (M ) Au, Pd) scattering paths. Although the XPS analysis revealed that reduced Pd exists only in Au/Pd 1:9, with EXAFS the existence of some reduced Pd is also observable in Au/Pd 1:1 and the pure Pd sample. The qualitatively described EXAFS spectra of the Au LIII edge and the Pd K edge were analyzed in terms of number, kind, and distance of the next neighboring atoms. The results of the simulations are shown in Table 3. It should be noted that the errors of the coordination numbers calculated in this work are mostly less than 20% which is in agreement with the commonly estimated error of the coordination numbers given in literature.48 The experimental and simulated spectra for the Au/Pd 1:1 and 1:9 are shown in Figure 5 to get an impression of the quality of the fits. The pure gold catalyst has an average coordination

number (CN) of 6 and the distance is with 2.82 Å significantly shorter than in the bulk gold but slightly longer than reported for Au nanoparticles with the same coordination number supported on oxidic substrates (2.79-2.8 Å65). The decrease in distance with decreasing coordination number is explained with the increased d-d interaction in small particles due to rehybridization.66 For the Au/Pd 7:1 catalyst, a CNAu-Au of 4 could be found but no Au-Pd interaction appears. The Pd EXAFS reveals that in this sample no metallic Pd exists, but scattering paths for light atoms like chlorine, oxygen, and, as indicated by the XPS analysis, even phosphorus are present. Therefore, we assume that either nothing of the Pd was reduced or the amount of metallic Pd after reduction was reoxidized by oxygen during storage under atmospheric conditions. The picture changed when more Pd was added to the system. In the Au-Pd 1:1 and 1:9 samples the Au-Pd interaction could be observed in the Au as well as in the Pd EXAFS spectra. The CNAu-Pd increases with increasing Pd content from 1.9 to 2.7 and the CNPd-Pd grows slightly from 1.1 to 1.4. The reverse trend could be observed for the Au backscattering contribution. The CNAu-Au as well as the CNPd-Au decrease with decreasing Au content. The Pd EXAFS spectrum for the pure Pd catalyst shows a small amount of Pd-Pd interactions resulting in a low coordination number. The main contribution stems again from light scatter atoms that have been identified with XPS as chlorine and oxygen. When regarding the interatomic distances one has to consider two effects: first the above-mentioned effect of rehybridization in small particles resulting in shrinking of the distances, and second the influence of the composition of the catalyst. Despite the almost similar particle size for all compositions and the same range of total coordination numbers of Au and Pd for all systems (except for the pure Pd system), the interatomic distance decreases with increasing Pd content. Therefore, not only the size of the nanoparticle and thus the coordination number is important for the interatomic distance but also the ratio of the two metals in the alloy. Vegard’s law describes the relation between the crystal lattice constant and the molar fraction of the two partners.30 This law, originally found as an empirical rule in salt alloys, was extended to metal alloys under the assumption of a solid mixture, in which the atoms are considered as hard spheres. It predicts a linear relation between the lattice constant and the molar fraction of one metal as long as the ratio of the atomic diameter of the two metals is greater than 0.87,29 which is true for the palladium gold pair. In accordance to Vegard’s law, the Au-Au distance of 2.82 Å found in the pure Au catalyst shrinks to 2.79 Å in the Au/Pd 1:9 catalyst despite the total CN of gold being equal in both samples. The same trend could be found for the Au-Pd distances as well as for the Pd-Pd distance. The results of the electron microscopy and XPS investigations show that the composition of the nanoparticles reflects the ratio of metal used in preparation and even single particles contain both metals. EXAFS simulation revealed that at least the two catalysts Au/Pd 1:1 and Au/Pd 1:9 form partly an alloy. Nonetheless, the total CNPd-Metal is smaller than the CNAu-Metal and therefore the two components form an inhomogeneous mixed alloy with an Au rich core and a Pd rich shell. This result is reasonable although Au has the lower surface free energy (1.63 J/m2) compared to Pd (2.05 J/m2) and should therefore be enriched on the surface.67 A possible reason could be a sequential reduction of Au and Pd due to the higher reduction potential of gold.

Polyaniline Supported Au/Pd Nanoparticles

J. Phys. Chem. C, Vol. 113, No. 15, 2009 6195

Figure 2. Typical HAADF-STEM micrographs of two different colloids supported on PANI: (a) Au/Pd 1:1 nanoparticles with corresponding histogram (b) and the spectrum of EDX analysis (c); (d) picture of Pd nanoparticles with the corresponding histogram (e) and EDX spectrum (f).

3.3. Electronic Properties. The XANES spectra of the Au LIII edge are shown together with those of Au-foil and Au-Pdfoil as references in Figure 6. The white line, the first feature after the edge jump, is clearly visible for the Au-foil and also for the pure Au catalyst. The intensity of the white line diminishes with increasing Pd-content and disappears almost for Au/Pd 1:9 as it could be observed for the Au-Pd-reference. The influence of the Pd content on the white line of the gold spectra is described later in this paper. A comparison between the Au-foil and the Au-Pd-foil reveals some more differences. At first the feature at about 11934 eV is more pronounced for the Au-Pd-foil than for the Au-foil and only the Au-Pd-foil shows a maximum there. Second the maxima of the second and the third feature, which become apparent for the Au-foil at about 11946 and 11969 eV, are also visible for the Au-Pd-foil as it is typical for face-centered cubic structures. But a closer look

Figure 3. k3-weighted magnitude of Fourier transform of the Au-foil (gray, dashed); Au catalyst (gray, solid); Au/Pd 7:1 (black, dotted); Au/Pd 1:1 (black, solid); Au/Pd 1:9 (black, dashed).

6196 J. Phys. Chem. C, Vol. 113, No. 15, 2009

Marx and Baiker

Figure 4. k3-weighted magnitude of Fourier transform of the Au-Pdfoil (gray, dashed); Au/Pd 7:1 (black, dotted); Au/Pd 1:1 (black, solid); Au/Pd 1:9 (black, dashed); Pd (gray, solid); the vertical, dashed line represents the position of the maximum of the Pd EXAFS spectrum that is omitted due to clarity.

TABLE 3: Results of the EXAFS Fit Parameters of the PANI Supported Bimetallic Nanoparticles for the Au LIII and Pd K Edge; N ) Coordination Number, R ) Distance (Å), ∆σ2 ) Debey-Waller Factor (Å2), ∆E0 ) Inner Core Correction (eV) sample

edge

pair

Na

R

Au Au LIII Au-Au 5.9 2.82 Au/Pd 7:1 Au LIII Au-Au 4.3 2.83 Pd K Pd-Cl 0.8 2.25 Pd-O 0.7 1.99 Pd-P 0.6 2.31 Au/Pd 1:1 Au LIII Au-Au 6.1 2.81 Au-Pd 1.9 2.80 Pd K Pd-Pd 1.1 2.79 Pd-Au 1.5 2.80 Pd-Cl 1.2 2.27 Pd-O 1.4 2.01 Au/Pd 1:9 Au LIII Au-Au 2.5 2.79 Au-Pd 2.7 2.77 Au-Cl 0.2 2.28 Pd K Pd-Pd 1.4 2.76 Pd-Au 0.7 2.77 Pd-Cl 1.7 2.27 Pd-O 2.0 2.01 Pd Pd K Pd-Pd 0.6 2.73 Pd-Cl 1.8 2.28 Pd-O 1.7 1.98 a

∆σ2

∆E0

residual

0.0074 0.0084 0.0020 0.0020 0.0030 0.0075 0.0062 0.0062 0.0052 0.0030 0.0026 0.0060 0.0055 0.002 0.0062 0.0052 0.0019 0.0020 0.0050 0.0020 0.0026

0.73 3.01 4.13 9.44 2.55 3.65 4.96 -1.26 -0.37 -4.07 9.09 6.69 3.79 3.03 -0.37 -3.00 -5.00 12.91 -2.82 0.94 2.94

7.2 7.4 7.0 4.9 3.2

4.8 3.7

6.4

The relative error was typically in the range 8-22%.

reveals differences in the intensities and the positions of the second and third feature. We observed a less intense second feature and a more intense third feature and both peaks are shifted to higher energies for the Au-Pd-foil. The same trend could be seen in the catalyst series. An increase of the first peak could be observed with increasing Pd content. In accordance to the bimetallic foil, the second feature is less pronounced and the third is more intense for the Pd rich catalyst compared to the Au rich sample. The shift of the maxima of the peaks to higher energy increases with increasing Pd content and is maximal for the Au/Pd 1:9 catalyst with a shift of 1 eV for the second peak and 1.5 eV for the third one. The observed shift and intensity change of these patterns could be qualitatively explained by the alteration of the distances and by changes in the electronic structure upon alloying. Both effects alter the shape of the empty DOS and therefore the shape over the edge. Compared to the references the general intensities of the peaks observed in the spectra of the catalysts are weaker due to the decreasing size and therefore decreasing number of nearest neighbors.

The Pd XANES spectra are more complicated due to the manifestations of Pd in different oxidation states. In Figure 7 the catalyst spectra as well as Pd-, Au-Pd-foil, and PdO as references are compared. The Pd XANES spectra of the two metal foils exhibit a pronounced white line at 24366 eV due to the not filled d-band of palladium. The next two features at around 24389 and 24428 eV are slightly shifted to lower energies for the Au-Pd foil compared to the Pd-foil. Again the change of the interatomic distance and electronic effects upon alloying are responsible for the differences seen in the XANES pattern of the references. For the PdO the edge threshold energy is clearly shifted to higher energies and the white line is even more distinct than for the metal foils due to the higher oxidation state of Pd in PdO. Compared to the references, the catalysts show a different pattern. With increasing Pd content the white line increases and the double feature of the Au/Pd 7:1 and 1:1 catalysts turns into one broad peak. For all catalysts the edge threshold energy is shifted to higher energy compared to the metal references and the shift reaches a maximum of 6 eV for Au/Pd 1:9 with respect to the Pd-foil. White line intensity and edge position indicate an oxidic state for the Pd in the samples. The different content of reduced and oxidized Pd species in the samples leads to the observed XANES pattern when going through the row. We assume that one of the species present in the sample might be the precursor substance Na2PdCl4. This result is in accordance with XPS analysis that reveals chlorine as one of the concomitant elements. Additionally comparison of K2PdCl4 XANES spectrum found in literature68 shows a good accordance with the pattern found for the pure Pd catalyst. Taking this problem into account experiments have been performed to explore if Pd2+ has (additionally to Pd0) an influence on the electronic and the catalytic properties of the bimetallic nanoparticles. As well as for geometric considerations two effects come into play when elucidating the electronic changes: first the decrease of the particle size into dimension where the shape, the width, and the energetic level of bands are altered, and second the addition of a second metal. Both changes lead to a decrease in the number of like next neighbors. Subsequently, the influence of the size for pure Au nanoparticles is discussed and afterward the changes that become apparent by alloying with Pd. When going from the bulk to nanosized particles the number of atoms forming a lattice decreases and therefore the width of bands will decrease too. The decrease in bandwidth results in a less pronounced overlap of bands leading to a rehybridization of the s, p, and d orbitals. In the concrete case of gold, the ideal electron configuration for a single atom is 5d106s1. With the increase in the number of atoms, the overlap of bands becomes significant and leads to hybridization of the 5d, 6s, and 6p band. The electron count of 5d and 6s orbitals is depleted while the 6p orbitals are occupied upon an increase in the number of neighbors, resulting in a electron configuration of 5d10-x6sp1+x for bulk gold.66 This depletion of the d-band for bulk gold is visible as a small white line in the Au-foil XAFS spectrum (Figure 6). The Au LIII edge probes the transition of 2p electrons to the empty density of states that have mainly d character, or more precisely spoken the transition occurs from the 2p3/2 band to the 5d5/2 band. Hence, the intensity of the white line in the LIII spectrum is directly linked to the number of holes in the d-band. Therefore the white line decreases with decreasing Au-particle size as one can see in Figure 6. When going to the bimetallic systems, a further decrease of the white line intensity is visible and the question arises, if the

Polyaniline Supported Au/Pd Nanoparticles

J. Phys. Chem. C, Vol. 113, No. 15, 2009 6197

Figure 5. Experimental χ(k) and FT spectra of the Au/Pd 1:1 (first row) and 1:9 catalyst (second row). The Au LIII edge as well as the Pd K edge are shown. The black lines represent the spectra obtained from the experiment, the gray lines result from the simulation with the parameters shown in Table 3.

Figure 6. Stacked plot of the XANES spectra at the Au LIII edge for the Au-foil (black, solid); Au-Pd-foil (gray, solid); Au (black, solid); Au/Pd 7:1 (dark gray, solid); Au/Pd 1:1 (gray, solid); Au/Pd 1:9 (light gray, solid).

Figure 7. Stacked plot of the XANES spectra at the Pd K edge for the Pd-foil (light gray, solid); Au-Pd-foil (dark gray, solid); PdO (gray, dashed); Au/Pd 7:1 (black, solid); Au/Pd 1:1 (dark gray, solid); Au/Pd 1:9 (gray, solid); Pd (light gray, solid).

decrease in white line intensity is simply caused by the dilution of Au atoms in a Pd matrix or if a Au-Pd interaction due to hybridization of Pd valence band with Au valence band44 is responsible for the change. To distinguish the two possibilities two model systems were prepared. Each model system consisted of a pair of samples: a pure Au sample and an Au/Pd bimetallic

Figure 8. Comparison of the XANES region of pairs of Au and Au/ Pd bimetallic nanoparticles; the inset magnifies the white line region of the Au LIII spectra

one. To ensure that the Pd is reduced a pretreatment in hydrogen was performed and the amount of oxidized Pd species could be significantly reduced. The results of EXAFS fits of the pretreated systems are shown in Table 4. In Figure 8 the two pairs are depicted. The important point here is that the CNAu-Au of the pair Au/Pd 1:1 and pure Au (a) (gray lines in Figure 8) as well as for the pair Au/Pd 1:9 and Au (b) (black lines in Figure 8) are almost equal. If the alloying had no influence on the electronic structure, the white line intensities for each pair should be similar due to the same number of like-neighbors in both samples. The spectrum reveals the opposite; the white line intensity of the Au/Pd 1:1 compound is smaller than for the Au (a) sample and even smaller than for the Au (b) system. For the Au/Pd 1:9 sample, the white line diminishes almost completely. Additionally, the first feature after the edge becomes significantly more pronounced with increasing Pd content, which is not observed when only the number of next gold neighbors is reduced. These findings have been predicted from the group of Zhang et al. They simulated the XANES spectra of bulk gold, of an Au55 cluster, and an Au55@Pd core-shell cluster with feff 8.0 and found a particle size dependence of the white line intensity as well as a stronger dependence from the Pd content.69 These theoretical predictions could now be confirmed by our experiment. At the moment no conlusive explanation for the enhanced filling of the Au d band in the presence of Pd can be given. However, one can speculate

6198 J. Phys. Chem. C, Vol. 113, No. 15, 2009

Marx and Baiker

TABLE 4: Structural Parameters Derived from the Au LIII Edge and Pd K Edge of Au and Au/Pd Bimetallic Nanoparticles Obtained after Reduction; N ) Coordination Number, R ) distance (Å), ∆σ2 ) Debey-Waller Factor (Å2), ∆E0 ) Inner Core Correction (eV) sample

edge

pair

Na

R

∆σ2

∆E0

4.25 Au (a) Au LIII Au-Au 10.8 2.86 0.0074 Au/Pd 1:1 Au LIII Au-Au 9.0 2.84 0.0067 5.28 Au-Pd 1.5 2.81 0.0053 4.43 Pd K Pd-Pd 1.8 2.75 0.0062 4.47 Pd-Au 2.0 2.81 0.0052 7.55 Pd-Cl 1.0 2.28 0.0030 1.73 Au (b) Au LIII Au-Au 3.5 2.80 0.0074 -0.47 Au/Pd 1:9 Au LIII Au-Au 3.0 2.79 0.0050 10.00 Au-Pd 5.3 2.78 0.0055 5.72 Pd K Pd-Pd 3.8 2.76 0.0062 -3.94 Pd-Au 1.7 2.78 0.0050 4.45 Pd-Cl 0.5 2.28 0.0030 -1.59 a

residual 3.7 1.5 7.2 8.9 4.0 5.6

The relative error was typically in the range 6-20%.

that the needed electron density could be gained by a direct transfer of electrons from the Pd or by an intramolecular charge transfer from the 6s and 6p band to the 5d band. The latter is possible due to strong Pd-Au d-d interaction. This hybridization of the Pd 4d band with the Au 5d5/2 band leads to a shift of the Au 5d5/2 band away from the Fermi level 44,70 and facilitates in this manner the filling of it. An influence of Pd2+ on the white line intensity can be ruled out because the sample, that was reduced with hydrogen, exhibited before reduction only a minor decrease of the white line intensity due to minor amount of metallic Pd present before reduction. If Pd2+ would have had the same influence on the white line intensity like the reduced Pd, the intensities of the white line before and after reduction should have been similar. This behavior was not observed in the experiments. We assume therefore that Pd2+ has no strong influence on the Au LIII edge. At last the effect of the particle size on the Au white line has to be considered. The question arises whether the increased particle size upon reduction or the presence of Pd is responsible for the decrease in the white line intensity? In bigger particles, gold has more like next neighbors which should lead to an increase of the white line intensity. Considering the size argument alone one would therefore expect a more pronounced white line on the Au LIII edge. But the opposite trend is visible in the XANES spectra; the white line becomes less intense upon admixing Pd. Therefore we assume that the influence of Pd overcompensates the size effect and that Pd is responsible for the electronic change in the d band and the enhanced filling of it. In Figure 1b, the XPS spectra for gold in the 4f region are shown. In comparison to the value for the binding energy of bulk gold (84.00 eV40,71), the 4f7/2 peak of the pure Au catalyst is shifted to a higher binding energy by about 0.5 eV. These results are in accordance to several findings in literature.26,72 Two explanations for this shift have been given. Manson45 et al. deduce the shift to initial state effects. That means that the change in hybridization (and thereby the alteration of the filling of the d and s orbitals) of small particles leads to modification of the repulsive interaction between valence and core level bands resulting in a higher binding energy. Wertheim at al. suggest that the charge, which remains after the photoionization process, could not be shielded sufficient in small nanoparticles leading to an increase of the binding energy or to a loss of kinetic energy of the leaving electrons respectively.46 As soon as Pd is added to the catalysts the Au 4f7/2 peak shifts of to lower binding energy compared to the value for the

Figure 9. Conversion (black columns) and yield (gray columns) for the oxidation of benzyl alcohol to benzaldehyde at 50 °C (a) and 100 °C (b); comparison of selectivity (c) at 50 °C (full symbols) and 100 °C (open symbols)

pure Au nanoparticle. The maximum shift of 0.7 eV was found for the Au/Pd 1:9 catalyst. This result is accordance to literature.44,45,69,70,73,74 To examine the possible influence of Pd2+ on the Au binding energy a sample with Au nanoparticles was prepared and impregnated with the Pd precursor substance (ratio Au/Pd ) 1:1). The dried sample contains only Pd2+. Measurements of these samples gave the same results (84.4 eV) as for the pure Au nanoparticles. Mason et al. suggested that an overlap of the gold 5d5/2 band with the 4d band of Pd leads to strong interaction and to a repulsion of the gold core level bands lying beneath. And indeed, detailed investigations of Au/Pd alloys by several groups44,70,73 reveal that the Pd 4d states could interact strongly with the 5d5/2 band of gold. The determination of the partial spectral weight with soft XPS shows a strong mixing of the 4d band of Pd with the 5d5/2 state of Au. This hybridization results in a broadening of the gold 5d5/2 band and in a shift of

Polyaniline Supported Au/Pd Nanoparticles

J. Phys. Chem. C, Vol. 113, No. 15, 2009 6199

TABLE 5: Overview of the Conversion, Yield, and Selectivity of the Mono- and Bimetallic Nanoparticles at 50, 100, and 0 °C and Different Reaction Times catalyst (on PANI) Au Au/Pd Au/Pd Au/Pd Pd Pd(II) Au Au/Pd Au/Pd Au/Pd Pd Pd(II) Au Au Au Au

7:1 1:1 1:9

7:1 1:1 1:9

T (K)

t (h)

conversion (%)

yield (%)

selectivity (%)

TON

TOF 1/h

323 323 323 323 323 323 373 373 373 373 373 373 323 323 323 273

3 3 3 3 3 3 3 3 3 3 3 3 0.17 0.5 1 3

>99.9 91 >99.9 51 9 13 >99.9 >99.9 >99.9 >99.9 59 15 43 72 >99.9 47

40 52 75 41 6 0.5 30 71 82 98 51 1.6 12 31 45 29

40 58 75 81 69 3.6 30 71 82 98 87 11 27 43 45 61

54 54 45 24 4 5.6 49 49 43 49 28 6.3 21 35 48 30

g18 18 g15 8 1 1.9 g16 g16 g14 g16 9 2.1 127 70 g48 10

the maximum of this band away from the Fermi level. The consequence of the shift is a filling of the empty d-band of gold (reduced white line, see above) while the result of the broadening of the 5d5/2 band is a repulsion with the 5d3/2 band and lower lying levels like the 4f band (giving rise to the observed reduced binding energy in XPS). All together we could show that the electronic properties of gold significant altered by admixing of palladium. The d-hole count of gold reduces with increasing Pd content. We could show that this is not only an effect of the decreased number of like neighbors (size effect) but also dependent on the Pd concentration. The core levels of gold are shifted to lower values, which is caused by hybridization of the valence Pd and Au bands. To gain a more detailed insight into the shape, filling, and energetic position of the d band DOS of bimetallic nanoparticles, the combined application of UPS and XPS as well as XAFS studies at the LIII and LII edge of both elements and quantum chemical calculations will be necessary. 3.4. Catalytic Behavior in Aerobic Oxidation of Benzyl Alcohol. The catalytic aerobic oxidation of benzyl alcohol has been chosen as a suitable test reaction to probe the catalytic behavior of the bimetallic Au/Pd catalysts. This test reaction has been intensively studied on both metals and is therefore excellently suited to elucidate the effect of admixing Pd to Au on the catalytic properties of the bimetallic nanoparticles. A striking example of the beneficial interaction of gold and palladium is the work of Enache et al.,5 who achieved high conversion and selectivity under solvent less conditions with Au/Pd on TiO2. In the present study the beneficial interaction of gold and palladium supported on PANI was elucidated. The support-metal interaction of PANI is distinct from oxidic supports, because it cannot provide oxygen for the catalytic cycle as it is assumed to be the case for the aerobic oxidation of alcohols with Au/ CeO2.75 Instead PANI exhibits aromatic, amine, and imine functionalities, which are good anchor points for the metal nanoparticles.76 Moreover, PANI is expected to be a charge density donor and might have an influence on the electronic properties of the nanoparticles.55 The results of the catalytic tests at 50 and 100 °C, respectively are summarized in Table 5. The activity of pure, PANI-supported gold nanoparticles at 50 °C was very high and complete conversion (>99.9%) was observed, while the Pd particles showed significantly lower activity. The catalysts Au/Pd 7:1 and 1:1 showed similar activities. By raising the molar fraction of Pd above 0.5 the conversion dropped to 51% for the Au/Pd 1:9 catalyst and reached only 9% for the Pd

system (Figure 9a). The yield of benzaldehyde increased with higher Pd content up to a molar fraction of 0.5 due to an increase in selectivity. At higher Pd contents the selectivity reached a maximum of 81% while the conversion decreased, resulting in lower yield (Figure 9c). A similar behavior was observed when the reaction was performed at 100 °C (Figure 9b). Full conversion was observed for the pure Au catalyst but the selectivity to benzaldehyde was lower compared to the reaction performed at 50 °C. In contrast, the yield of the pure palladium catalyst was higher at 100 °C (51%) than at 50 °C (6%). Because of the increase in the selectivity to benzaldehyde with increasing molar fraction of Pd to a maximum of 98% at full conversion, the highest yield of benzaldehyde was reached for the Au/Pd 1:9 catalyst (Figure 9c). Presumable side products like benzoic acid could not be detected by GC. A possible explanation could be the formation of an amide by reaction of the formed benzoic acid with the amine functionalities of the PANI. On the basis of the catalytic tests we can conclude that the selectivity increases with higher Pd content up to an Au/Pd ratio of 1:9 at 100 °C as well as at 50 °C. At higher temperatures the selectivity of Pd-containing catalysts was generally higher. Recently, Hou et al. conducted the oxidation of benzyl alcohol under homogeneous conditions in the presence of solute bimetallic Au/Pd nanoparticles stabilized with PVP and found a similar trend of the catalytic behavior in dependence of the molar ratio of Pd.77 The catalyst with the highest Pd content (Au/Pd 1:3) afforded higher selectivity than pure Au catalysts or 3:1 and 1:1 alloys in the aerobic oxidation with air. However, Hou et al. received under almost the same reaction conditions a high TOF of 74 h-1 compared to 16 h-1 found in this work. Note that the TOF determined in the present work is a conservative estimate because it is given for full conversion. In the former work, the selectivity dropped from 90.5% (at 14.8% conversion) after one hour to 39.7% (at 33.6% conversion) after 24 h due to overoxidation, while in the present work the best catalyst reached 98% selectivity at almost full conversion within three hours. To elucidate the role of Pd2+, which is present in the samples due to the synthetic procedure used, a test experiment was performed with a catalyst loaded only with the Pd2+ precursor. With only Pd2+ the conversions reached 15% at 100 °C, which is much lower than the conversion observed with Pd nanoparticles at this temperature. The yield of benzaldehyde at both temperatures was negligible (Figure 9a-c). Therefore we assume that Pd is more active in the reduced form and the

6200 J. Phys. Chem. C, Vol. 113, No. 15, 2009 oxidized species only plays a minor role for the oxidation of benzyl alcohol. Besides the above-described electronic changes of the gold atoms in the nanoparticles the structure-activity correlation was examined. Therefore the as-prepared nanoparticles have been treated with hydrogen at 250 °C for 5 h. EXAFS indicated that this treatment led to reduction of Pd as well as a growth of the particles. The obtained catalysts have been tested in the same catalytic reaction as the corresponding unreduced catalysts. The tests revealed a strong decrease of the conversion. The Au/Pd 9:1 catalyst gave at 100 °C only 34% conversion and the selectivity reached 79%. For catalysts with higher Pd loading, the selectivity was 100% but at low conversions of 4 and 2% for the 1:1 and 1:9 catalysts, respectively. As a result of the treatment at elevated temperatures the particles grow and the available fraction of surface atoms decreases leading to a general decrease in activity. Moreover the further reduction of the Pd constituent might lead to a closed shell of Pd on the surface of the nanoparticle, thus preventing a participation of the gold atoms in the reaction. These findings suggest that the reaction might take place mainly at gold atoms or at the interface between gold and palladium in the bimetallic particles. This assumption is supported by the fact that the pure Pd particles are significantly less active than the ones with gold. To prevent these effects, we performed the synthesis of PANI-supported Au/Pd catalysts via the colloidal route and omitted the reduction step. The influence of PANI on the catalytic behavior is not clear yet, but it can be assumed that no oxygen could be provided by the PANI itself and that the support-metal interaction was similar for all nanoparticles of different composition. Therefore it appears that the changes in the selectivity and activity observed here are mainly due to the changes in the electronic and geometrical properties of the bimetallic particles induced by the admixing of Pd to Au. Conclusion We have synthesized a series of polymer (polyaniline) supported bimetallic-, Au-, and Pd-containing catalysts by a colloidal route and investigated their structural and electronic properties as well as their catalytic behavior in the oxidation of benzyl alcohol with oxygen. Admixing Pd to Au resulted in a continuous increase of selectivity to benzaldehyde reaching a maximum of 98% at 100 °C at full conversion with catalysts containing Au:Pd in a ratio of 1:9. The colloidal route applied for the preparation of the bimetallic nanoparticles allowed the control of the size of the particles in a relative narrow range (2.4-3.7 nm), thereby minimizing possible size effects on the structural and catalytic properties of the nanoparticles. The structural investigations suggest the presence of bimetallic particles with an Au rich core and a Pd rich shell. The differences in the electronic and geometric properties of the mono- and bimetallic particles have been examined with XPS and XAFS. The electronic configuration of Au and Pd was found to be altered upon mixing of Pd and Au. A clear shift to lower binding energies of Au core levels with increasing Pd content could be observed. The Au XANES spectra reveal a decrease in the d holes of the Au 5d valence band. The results from XPS and XANES indicate, that at least the Au 5d DOS is more filled compared to the pure bulk metal and that the energetic level is shifted toward the Fermi level. Although no direct correlation between the changes of these properties and the catalytic behavior can be drawn it seems likely that they are decisive for the observed catalytic behavior because other effects such as particle size and metal-support interaction were minimized in the series of investigated bimetallic particles.

Marx and Baiker Acknowledgment. The authors thank Dr. Antonella Rossi, Dr. Wolfgang Kleist for the XPS analyses, Dr. Frank Krumeich for the unresting production of STEM pictures, Bertram Kimmerle and Peter Haider for the support during measurements at the beamlines and the readiness for discussion, Dr. Eva Ro¨del for fruitful discussion concerning EXAFS analysis, Niels van Vegten for BET surface measurements, and Professor Jan-Dierk Grunwaldt for his support in XAS experiments. We thank ANKA (Forschungszentrum Karlsruhe), and SLS (Paul Scherrer Institut, Villingen) for beamtime and Dr. Stefan Mangold and Dr. Maarten Nachtegaal for technical support. We thank HASYLAB (DESY, Hamburg, Germany) for beamtime and Matthias Hermann and Adam Webb for their assistance. The work was supported by the European Community (Contract RII3-CT-2004-506008). References and Notes (1) Sinfelt, J. H. Acc. Chem. Res. 1977, 10, 15. (2) Besenbacher, F.; Chorkendorff, I.; Clausen, B. S.; Hammer, B.; Molenbroek, A. M.; Norskov, J. K.; Stensgaard, I. Science 1998, 279, 1913. (3) Ponec, V. Surf. Sci. 1979, 80, 352. (4) Chen, M. S.; Kumar, D.; Yi, C. W.; Goodman, D. W. Science 2005, 310, 291. (5) Enache, D. I.; Edwards, J. K.; Landon, P.; Solsona-Espriu, B.; Carley, A. F.; Herzing, A. A.; Watanabe, M.; Kiely, C. J.; Knight, D. W.; Hutchings, G. J. Science 2006, 311, 362. (6) Ketchie, W. C.; Murayama, M.; Davis, R. J. J. Catal. 2007, 250, 264. (7) Conte, M.; Carley, A. F.; Attard, G.; Herzing, A. A.; Kiely, C. J.; Hutchings, G. J. J. Catal. 2008, 257, 190. (8) Edwards, J. K.; Carley, A. F.; Herzing, A. A.; Kiely, C. J.; Hutchings, G. J. Faraday Discuss. 2008, 138, 225. (9) Edwards, J. K.; Solsona, B. E.; Landon, P.; Carley, A. F.; Herzing, A.; Kiely, C. J.; Hutchings, G. J. J. Catal. 2005, 236, 69. (10) Edwards, J. K.; Solsona, B.; Landon, P.; Carley, A. F.; Herzing, A.; Watanabe, M.; Kiely, C. J.; Hutchings, G. J. J. Mater. Chem. 2005, 15, 4595. (11) Landon, P.; Collier, P. J.; Papworth, A. J.; Kiely, C. J.; Hutchings, G. J. Chem. Commun. 2002, 2058. (12) Landon, P.; Collier, P. J.; Carley, A. F.; Chadwick, D.; Papworth, A. J.; Burrows, A.; Kiely, C. J.; Hutchings, G. J. Phys. Chem. Chem. Phys. 2003, 5, 1917. (13) Han, Y. F.; Zhong, Z. Y.; Ramesh, K.; Chen, F. X.; Chen, L. W.; White, T.; Tay, Q. L.; Yaakub, S. N.; Wang, Z. J. Phys. Chem. C 2007, 111, 8410. (14) Beck, A.; Horvath, A.; Schay, Z.; Stefler, G.; Koppany, Z.; Sajo, I.; Geszti, O.; Guczi, L. Top. Catal. 2007, 44, 115. (15) Devarajan, S.; Bera, P.; Sampath, S. J. Colloid Interface Sci. 2005, 290, 117. (16) Mejia-Rosales, S. J.; Fernandez-Navarro, C.; Perez-Tijerina, E.; Blom, D. A.; Allard, L. F.; Jose-Yacaman, M. J. Phys. Chem. C 2007, 111, 1256. (17) Soto-Verdugo, V.; Metiu, H. Surf. Sci. 2007, 601, 5332. (18) Tsen, S. C. Y.; Crozier, P. A.; Liu, J. Ultramicroscopy 2003, 98, 63. (19) Akita, T.; Hiroki, T.; Tanaka, S.; Kojima, T.; Kohyama, M.; Iwase, A.; Hori, F. Catal. Today 2008, 131, 90. (20) Chen, C. H.; Sarma, L. S.; Chen, J. M.; Shih, S. C.; Wang, G. R.; Liu, D. G.; Tang, M. T.; Lee, J. F.; Hwang, B. J. ACS Nano 2007, 1, 114. (21) Ferrer, D.; Torres-Castro, A.; Gao, X.; Sepulveda-Guzman, S.; Ortiz-Mendez, U.; Jose-Yacaman, M. Nano Lett. 2007, 7, 1701. (22) Herzing, A. A.; Carley, A. F.; Edwards, J. K.; Hutchings, G. J.; Kiely, C. J. Chem. Mater. 2008, 20, 1492. (23) Knecht, M. R.; Weir, M. G.; Frenkel, A. I.; Crooks, R. M. Chem. Mater. 2008, 20, 1019. (24) Liu, H. F.; Mao, G. P.; Meng, S. J. J. Mol. Catal. 1992, 74, 275. (25) Wang, D.; Villa, A.; Porta, F.; Prati, L.; Su, D. S. J. Phys. Chem. C 2008, 112, 8617. (26) Guczi, L. Catal. Today 2005, 101, 53. (27) Ponec, V. AdV. Catal. 1983, 32, 149. (28) Ponec, V. Appl. Catal., A: 2001, 222, 31. (29) Denton, A. R.; Ashcroft, N. W. Phys. ReV. A 1991, 43, 3161. (30) Vegard, L. Z. Phys. 1921, 5, 17. (31) Clarke, J. K. A. Chem. ReV. 1975, 75, 291. (32) Falsig, H.; Hvolbaek, B.; Kristensen, I. S.; Jiang, T.; Bligaard, T.; Christensen, C. H.; Norskov, J. K. Angew. Chem., Int. Ed. 2008, 47, 4835. (33) Hammer, B.; Nørskov, J. K. AdV. Catal. 2000, 45, 71. (34) Lopez, N.; Norskov, J. K. J. Am. Chem. Soc. 2002, 124, 11262.

Polyaniline Supported Au/Pd Nanoparticles (35) Remediakis, I. N.; Lopez, N.; Norskov, J. K. Appl. Catal., A 2005, 291, 13. (36) Bzowski, A.; Kuhn, M.; Sham, T. K.; Rodriguez, J. A.; Hrbek, J. Phys. ReV. B 1999, 59, 13379. (37) Bzowski, A.; Sham, T. K.; Watson, R. E.; Weinert, M. Phys. ReV. B 1995, 51, 9979. (38) Cho, S. J.; Kang, S. K. Catal. Today 2004, 93-95, 561. (39) Hlil, E. K.; Baudoing-Savois, R.; Moraweck, B.; Renouprez, A. J. J. Phys. Chem. 1996, 100, 3102. (40) Kuhn, M.; Rodriguez, J. A.; Hrbek, J.; Bzowski, A.; Sham, T. K. Surf. Sci. 1995, 341, L1011. (41) Pedersen, M. Ø.; Helveg, S.; Ruban, A.; Stensgaard, I.; Lægsgaard, E.; Nørskov, J. K.; Besenbacher, F. Surf. Sci. 1999, 426, 395. (42) Sham, T. K.; Bzowski, A.; Kuhn, M.; Tyson, C. C. Solid State Commun. 1991, 80, 29. (43) Tyson, C. C.; Bzowski, A.; Kristof, P.; Kuhn, M.; Sammynaiken, R.; Sham, T. K. Phys. ReV. B 1992, 45, 8924. (44) Nahm, T. U.; Jung, R.; Kim, J. U.; Park, W. G.; Oh, S. J.; Park, J. H.; Allen, J. W.; Chung, S. M. Phys. ReV. B 1998, 58, 9817. (45) Mason, M. G. Phys. ReV. B 1983, 27, 748. (46) Wertheim, G. K.; Dicenzo, S. B.; Youngquist, S. E. Phys. ReV. Lett. 1983, 51, 2310. (47) Davis, R. J.; Boudart, M. J. Phys. Chem. 1994, 98, 5471. (48) Guczi, L.; Beck, A.; Horva´th, A.; Koppa´ny, Z.; Stefler, G.; Frey, K.; Sajo, I.; Geszti, O.; Bazin, D.; Lynch, J. J. Mol. Catal. 2003, 204, 545. (49) Reifsnyder, S. N.; Lamb, H. H. J. Phys. Chem. B 1999, 103, 321. (50) Zhang, P.; Zhou, X. T.; Tang, Y. H.; Sham, T. K. Langmuir 2005, 21, 8502. (51) He, J. L.; Wu, T. B.; Zhang, Z. F.; Ding, K. L.; Han, B. X.; Xie, Y.; Jiang, T.; Liu, Z. M. Chem.sEur. J. 2007, 13, 6992. (52) Chi, K.-W.; Hwang, H. Y.; Park, J. Y.; Lee, C. W. Synth. Met. 2009, 159, 26. (53) Luo, X. L.; Killard, A. J.; Smyth, M. R. Chem.sEur. J. 2007, 13, 2138. (54) Lihua, Z.; Heqing, T.; Jing, L.; Yinghui, W.; Harima, Y.; Aiqing, Z. Polymer 2006, 47, 7361. (55) Steffan, M.; Klasovsk, F.; Arras, J.; Roth, C.; Radnik, J.; Hofmeister, H.; Claus, P. AdV. Synth. Catal. 2008, 350, 1337. (56) Ishida, T.; Kuroda, K.; Kinoshita, N.; Minagawa, W.; Haruta, M. J. Colloid Interface Sci. 2008, 323, 105.

J. Phys. Chem. C, Vol. 113, No. 15, 2009 6201 (57) Feng-Jiin, L.; Li-Ming, H.; Ten-Chin, W.; Gopalan, A. Synth. Met. 2007, 157, 651. (58) Mourato, A.; Wong, S. M.; Siegenthaler, H.; Abrantes, L. M. J. Solid State Electrochem. 2006, 10, 140. (59) Houdayer, A.; Schneider, R.; Billaud, D.; Ghanbaja, J.; Lambert, J. Synth. Met. 2005, 151, 165. (60) Majumdar, G.; Goswami, M.; Sarma, T. K.; Paul, A.; Chattopadhyay, A. Langmuir 2005, 21, 1663. (61) Duff, D. G.; Baiker, A. Langmuir 1993, 9, 2301. (62) Rehr, J. J.; Booth, C. H.; Bridges, F.; Zabinsky, S. I. Phys. ReV. B 1994, 49, 12347. (63) Haider, P.; Kimmerle, B.; Krumeich, F.; Kleist, W.; Grunwaldt, J.-D.; Baiker, A. Catal. Lett. 2008, 125, 169. (64) Via, G. H.; Drake, K. F.; Meitzner, G.; Lytle, F. W.; Sinfelt, J. H. Catal. Lett. 1990, 5, 25. (65) Miller, J. T.; Kropf, A. J.; Zha, Y.; Regalbuto, J. R.; Delannoy, L.; Louis, C.; Bus, E.; van Bokhoven, J. A. J. Catal. 2006, 240, 222. (66) van Bokhoven, J. A.; Miller, J. T. J. Phys. Chem. C 2007, 111, 9245. (67) Boscoboinik, J. A.; Plaisance, C.; Neurock, M.; Tysoe, W. T. Phys. ReV. B 2008, 77, 045422. (68) Sobczak, J. W.; Sobczak, E.; Kosinski, A.; Bilinski, A. J. Alloys Compd. 2001, 328, 132. (69) Liu, F.; Wechsler, D.; Zhang, P. Chem. Phys. Lett. 2008, 461, 254. (70) Weinberger, P.; Szunyogh, L.; Bennett, B. I. Phys. ReV. B 1993, 47, 10154. (71) Gegner, J.; Wu, H.; Berggold, K.; Sebastian, C. P.; Harmening, T.; Pöttgen, R.; Tjeng, L. H. Phys. ReV. B 2008, 77, 035103. (72) Minati, L.; Speranza, G.; Calliari, L.; Micheli, V.; Baranov, A.; Fanchenko, S. J. Phys. Chem. A 2008, 112 (34), 7856. (73) Nascente, P. A. P.; Decastro, S. G. C.; Landers, R.; Kleiman, G. G. Phys. ReV. B 1991, 43, 4659. (74) Meitzner, G.; Sinfelt, J. H. Catal. Lett. 1995, 30, 1. (75) Abad, A.; Concepcion, P.; Corma, A.; Garcia, H. Angew. Chem., Int. Ed. 2005, 44, 4066. (76) Yonezawa, T.; Toshima, N. J. Chem. Soc., Faraday Trans. 1995, 91, 4111. (77) Hou, W. B.; Dehm, N. A.; Scott, R. W. J. J. Catal. 2008, 253, 22.

JP808362M