Syntheses of Sub-30 nm Au@Pd Concave Nanocubes and Pt-on

0 downloads 0 Views 1MB Size Report
atoms in the steps and kinks, are generally more active than those of low-index planes ... small (ca. 21 nm) Au@Pd NPs with concave cubic structures enclosed with ...... three metals (Pd, Pt, and Au) was taken into account, this. Pt. ∧. (Au@Pd) ...
Article pubs.acs.org/JPCC

Syntheses of Sub-30 nm Au@Pd Concave Nanocubes and Pt-on(Au@Pd) Trimetallic Nanostructures as Highly Efficient Catalysts for Ethanol Oxidation Gui-Rong Zhang, Jie Wu, and Bo-Qing Xu* Innovative Catalysis Program, Key Lab of Organic Optoelectronics & Molecular Engineering, Department of Chemistry, Tsinghua University, Beijing 100084, China S Supporting Information *

ABSTRACT: Designed manipulation of the morphology of metallic nanostructures containing Pd and/or Pt represents a challenge in the search of highly efficient precious metal catalysts. We provide herein the first selective synthesis of narrowly sized small (21 nm) Au@Pd concave nanocubes enclosed with high-index Pd facets by deposition of Pd atoms on truncated-octahedral Au seeds. Moreover, further controlled deposition of Pt onto these Au@Pd concave nanocubes produces sub-30 nm trimetallic Pt-on-(Au@Pd) nanostructures having jagged Pt-rich surfaces. The as-prepared Au@Pd and Pt-on-(Au@Pd) nanostructures are found highly active and fairly stable when employed as anode catalysts for ethanol electrooxidation, their activity normalized to the mass of Pd or Pd plus Pt being 7−9 times higher than conventional Pd black catalyst. These data may echo the importance of innovative small multimetallic nanostructures for highly efficient catalysts that depend critically on use of precious metals, for applications in energy, environmental, and chemical technologies. ethanol oxidation7,9 and cathode oxygen reduction reactions.8 For precious metal catalysts, high MSAs are essential in view of practical application as the metals are priced by mass, NOT by their surface area. To this end, an urgent demand in chemistry could be to develop new synthetic methods for producing highindex faceted particles in smaller sizes with either well-defined or irregular structures. We disclose in this study the first synthesis of narrowly sized small (ca. 21 nm) Au@Pd NPs with concave cubic structures enclosed with high-index Pd facets (e.g., {730}, {520}, and {610}). On having Pt deposited onto these Au@Pd concave nanocubes, we are able to further construct sub-30 nm trimetallic Pt-on-(Au@Pd) nanostructures (denoted as Pt∧(Au@Pd)) featuring a triple-layered structure with Au in the core, Pd in the inner-layer, and Pt-rich clusters in the outermost jagged-surface layers. When employed as anodic catalysts for ethanol oxidation reaction (EOR), the Au@Pd concave nanocubes and Pt∧(Au@Pd) nanostructures are found to show significantly improved catalytic activity relative to those conventional Pd samples. These data demonstrate for the first time that the catalytically highly active high-index facets (or planes) or unique jagged surfaces can be selectively assembled on sub-30 nm bi- and/or multimetallic particles in specially designed layered nanostructures.

1. INTRODUCTION In recent years, significant attention has been directed toward the innovation of Pt and/or Pd containing nanostructures due to their distinctive properties and application potentials related to thermal- and electrocatalysis.1−6 However, the high cost and supply constraint of the precious Pt and Pd have greatly restricted their large-scale applications.3,6 Development of synthetic methods that enable a designed manipulation of the size, morphology, and composition of Pt and/or Pd containing nanostructures to greatly improve the catalytic efficiency of Pt and/or Pd component(s) has remained a challenge. For most catalytic reactions, high-index facets (or planes), which are associated with high density of low-coordinated atoms in the steps and kinks, are generally more active than those of low-index planes composed of more closely packed surface atoms. Unfortunately, precious metal nanoparticles (NPs) enclosed with high-index facets were obtained so far only in well-defined particle shapes with very large sizes; for instances, Pt tetrahexahedra enclosed by {730} facets were obtained in sizes larger than 80 nm,7 and high-index faceted Pt,8,9 Au,10,11 Pd,4 and bimetallic core−shell Au@Pd12−14 particles were produced in average sizes of 40−130 nm. These high-index faceted particles are too large for surface catalysis because they could only expose a very small fraction (less than 3%) of the precious metal atoms for catalytic reactions; their mass-specific activity (MSA, activity based on a unit mass of the active metal) would be even lower than those of conventional Pt/C or Pd/C catalysts, as observed for anode formic acid and © 2012 American Chemical Society

Received: May 10, 2012 Revised: July 19, 2012 Published: August 2, 2012 20839

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

2. EXPERIMENTAL SECTION 2.1. Materials. Hydrogen tetrachloroaurate(III) trihydrate (HAuCl4·3H2O, Acros, ACS reagent), palladium(II) chloride (PdCl2, SCRC, 99.5%), potassium tetrachloroplatinate(II) (K2PtCl4, Sigma-Aldrich, 99.9%), cetyltrimethylammonium bromide (CTAB, Sigma-Aldrich, ≥99%), sodium borohydride (NaBH4, Sigma-Aldrich, 99%), ascorbic acid (Sigma-Aldrich, 99%), Pd/C (J&K Chemical, Pd-5 wt %), and Pd black (Alfa Aesar, 99.9% metal basis) were used as received. Deionized water was used in all of the preparations. Aqueous H2PdCl4 solution (10 mM) was prepared by completely dissolving 44.5 mg PdCl2 (0.25 mmol) in aqueous HCl(20 mM, 25 mL) in a water bath thermostat at 60 °C. 2.2. Preparation of Au@Pd Concave Nanocubes. The preparation of Au@Pd concave nanocubes was based on a seed mediated growth method using truncated-octahedral Au NPs as seeds, synthesized by modifying a literature method.15 A typical synthesis procedure for the Au seeds is as follows: 0.5mL of 5 mM HAuCl4 solution was mixed with 7.5 mL of 100 mM cetyltrimethylammonium bromide (CTAB) solution. Subsequently, a freshly prepared, ice-cooled 1 mL of 5 mM NaBH4 solution was quickly injected into the above mixture to make the primary Au seeds. A growth solution was prepared by adding 50 μL of 5 mM HAuCl4 solution and 4 mL of 15 mM ascorbic acid into 0.8 mL of 100 mM CTAB solution in a vial. Then 5 μL of the primary seed solution was added to the growth solution and thoroughly mixed. The mixture was left undisturbed overnight at room temperature. The color of the solution gradually changed from colorless to pink, indicating the formation of Au NPs. Au@Pd concave nanocubes were then prepared by seedmediated growth of Pd on the truncated-octahedral Au NPs. Briefly, 50 μL of 0.5 mM ascorbic acid and 100 μL of 10 mM H2PdCl4 solutions were added to 2 mL of the as-prepared truncated-octahedral Au hydrosols in turn. The mixture was gently shaken and left undisturbed overnight, a color change of the solution from a pink to a dark brown featuring the formation of the Au@Pd particles. 2.3. Preparation of Pt∧(Au@Pd) Nanostructures. The Pt∧(Au@Pd) trimetallic nanostructures were prepared by depositing Pt on the above as-prepared Au@Pd concave nanocubes through ascorbic acid reduction of PtCl42− in an aqueous solution. Specifically, 50 μL of 0.5 mM ascorbic acid and 500 μL of 2 mM K2PtCl4 solution were added to the asprepared 1 mL of Au@Pd hydrosols in turn. After a gentle shaking, the mixture was left undisturbed overnight at room temperature. 2.4. Preparation of Pd NPs. To prepare the reference Pd NPs, 125 μL of 10 mM H2PdCl4 solution was mixed with 3.75 mL of 100 mM CTAB solution, and then an ice-cooled 0.5 mL of 5 mM NaBH4 solution was injected into the mixture to obtain Pd seeds. Next, 50 μL of 0.5 mM ascorbic acid and 100 μL of 10 mM H2PdCl4 solution were added to 2 mL of Pd seed solution to make Pd NPs. The resultant products (Au@Pd, Pt∧(Au@Pd), and Pd NPs) were collected by centrifugation at 6000 rpm for 5 min. The precipitates were washed twice by redispersing in deionized water and centrifugation. 2.5. Structural Analysis. Transmission electron microscopy (TEM) images were captured using a JEOL JEM-2010 microscope operated at 120 kV. High-resolution TEM (HRTEM), high-angle annular dark-field scanning transmission

electron microscopy (HAADF-STEM), and energy dispersive spectroscopic (EDS) elemental mapping measurements were performed on FEI Tecnai G2 F20 microscope equipped with an EDAX X-ray detector. The samples were prepared by placing a drop of the colloidal solution or catalyst powder dispersion in deionized water on a Formvar/carbon film coated Cu grid (3 mm, 300 mesh), followed by drying under ambient conditions. UV−vis spectroscopy was recorded on a Unico UV-2102PC spectrometer operated at a resolution of 0.5 nm. The colloidal solution samples were filled in a quartz cell of 1 cm light-path length, and the light absorption spectra were given in reference to deionized water. X-ray photoelectron spectroscopy (XPS) studies were carried out on a PHI 5300 ESCA1610 SAM instrument using Mg Kα radiation (1253.6 eV). The binding energies (BE) were calibrated using the adventitious C 1s line at 284.8 eV. The actual metal loadings on the working electrode for all the catalysts were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES, IRIS Intrepid II XSP, ThermoFisher). To measure the composition of the Au@Pd, Pt∧(Au@Pd), and Pd NPs by ICP-AES analysis, we first enriched the as-prepared colloidal solution containing the metal NPs by high-speed centrifugation at 6000 rpm. Then, to 200 μL of the enriched colloidal solution was added a freshly prepared aqua regia to make a mixed solution of 1.0 mL. This mixed solution was ultrasonicated for 30 min to ensure complete dissolution of the precious metals and was further diluted to 10.0 mL with deionized water before it was subjected to ICP-AES analysis. In this way, the nominal concentration of each precious metal in the test solution was kept in the range of 0.5−100 μg mL−1, which would ensure accurate determination of each of the precious metals in the solution. 2.6. Electrochemical Study. Electrochemical measurements were performed using a glassy carbon rotating disk electrode (GC-RDE, Pine Instrumentation) controlled by a potentiostat/galvanostat model 263A (PAR). A saturated calomel electrode (SCE) and Pt wire with a diameter of 0.5 mm were employed as the reference and counter electrodes, respectively. The potentials reported here are given with respect to SCE. The GC-RDE was polished to a mirror finish using 0.5 and 0.05 μm γ-alumina powders (CH Instruments, Inc.) before use. To prepare the working electrode, the GCRDE was coated with the washed Au@Pd, Pt∧(Au@Pd), or Pd NPs at similar metal loadings (ca. 4 μg) and then air-dried. To enhance the samples’ attachment with the GC-RDE, 10 μL of 0.05 wt % Nafion solution was applied to cover the catalyst layer. When the catalyst was Pd black or Pd/C, a catalyst ink was prepared first by sonicating a suspension of the individual catalyst (0.5 mg for Pd black, 5.0 mg for Pd/C) in isopropanol (1.0 mL), and 10 μL of the suspension was then applied onto the GC-RDE. After solvent evaporation, the catalyst layer was covered by 10 μL of 0.05 wt % Nafion solution before electrocatalytic measurements. Cyclic voltammetry (CV) measurements were carried out at 298 K in 0.1 M KOH in the potential range from −0.85 to 0.40 V (scan rate: 50 mV s−1). For ethanol electrooxidation measurements, the electrolyte was N2-saturated 0.1 M KOH solution containing 0.5 M ethanol. An activation of the electrode catalyst was achieved by repeatedly performing CV scanning from −0.85 to 0.40 V with a scan rate of 50 mV s−1. The repetition was done for up to several tens of CV cycles until the electrode catalyst became stabilized or a stable voltammograms were obtained. The activation treatment was thought essential to minimize the influence of the protector molecules that were used during the 20840

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

Figure 1. Schematic illustration of the heterogeneous seed-mediated syntheses of Au@Pd concave nanocubes and Pt∧(Au@Pd) nanostructures. The numbers above the particle images show the particle size distribution statistics obtained from TEM measurements.

Figure 2. (a−c) TEM and (d−f) HRTEM images of (a, d) truncated-octahedral Au seeds, (b, e) Au@Pd concave nanocubes, and (c, f) Pt∧(Au@ Pd) nanostructures. The insets in (d) and (e) show the geometric models for the Au and Au@Pd particles, respectively. The inset in (f) shows the FFT pattern from the region in the square mark. (g−i) Atomic models for high-index facets showing (g) (730), (h) (520), and (i) (610) planes projected from the [001] direction.

syntheses of Au@Pd, Pt∧(Au@Pd), and Pd NPs. All electrochemical experiments were carried out at room temperature and under ambient pressure.

method in an aqueous solution employing HAuCl4 as the gold precursor, ascorbic acid as the reducing agent, and CTAB as the stabilizing agent. Figure 2a shows the representative TEM images of the as-prepared Au NPs. HRTEM images of the Au NPs and their fringe orientations indicate that they were in single-crystalline structure; the majority of the particles appeared in truncated-octahedral shape (Figure 2d). Further statistical analysis of the particle shape showed that 90% of the Au NPs assumed the truncated-octahedral shape; their average particle size was 9.6 ± 0.7 nm (Figure S1, Supporting Information). Au@Pd concave nanocubes were then prepared by the heterogeneous seed-mediated growth of a Pd layer from the

3. RESULTS AND DISCUSSION 3.1. Preparation and Characterization of Au@Pd Concave Nanocubes and Pt∧(Au@Pd) Nanostructures. A multicomponent heterogeneous seed-mediated growth route was employed to prepare the Au@Pd and Pt∧(Au@Pd) samples, as illustrated schematically in Figure 1. In the first step, truncated-octahedral Au NPs, which were used later as the structure directing cores for the formation of Au@Pd concave nanocubes, were prepared following a seed-mediated growth 20841

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

approach, was reported by Huang et al.17 Using a Pd-specific Suzuki cross-coupling reaction, Huang et al. found that their Au@Pd concave nanocubes were ca. 3 times more active than Au@Pd octahedra (121 nm) in terms of turnover frequency (TOF)17 although they made no comparison to conventional Pd black catalyst. The small sizes of the present Au@Pd particles (ca. 21 nm, Figure 2b,e) could be benefited from our use of small truncated-octahedral Au seeds (9.6 ± 0.7 nm) and would then offer an advantage in catalytic applications. In any case, the very similar morphology of the Au@Pd concave nanocubes obtained independently in this and Huang et al.’s work demonstrates that the use of truncated-octahedral seed particles (like Au colloids) in seed-mediated growth could be exploited as a general approach for selective synthesis of bimetallic concave nanocubes. UV−vis absorption spectroscopy was used to show the response to Pd deposition of the Au surface plasmon resonance (SPR) signal during the formation of Au@Pd NPs (Figure S3). The amount of Pd atoms deposited on the Au NP cores could be readily controlled during the particle growth by varying the added volume of H2PdCl4 solution. The characteristic Au SPR signal centered at ca. 510 nm gradually weakened as the added H2PdCl4 solution was increased from 5 to 10 μL. When the added H2PdCl4 solution was further increased to 20 μL or more (i.e., Pd/Au ratio ≥2.0), the SPR signal became hardly visible. As the intensity of Au SPR signal could be related directly to the exposed Au surface area and Pd does not produce any SPR signal, these results might suggest that the surface of the Au NPs became gradually covered with small Pd entities and then encapsulated with a Pd shell when the number of deposited Pd atoms amounted to more than twice that of Au atoms in the preformed Au NP seeds. This discussion, however, provides just a rough description of the early stages during the formation of the Au@Pd NPs shown in Figure 2b. To gain insight into the morphological evolution of the Au@ Pd concave nanocubes, we prepared Au@Pd NPs in the standard procedure except variation in the amount of H2PdCl4. Figure S4 shows the TEM images of the products prepared with 5, 10, 50, and 100 μL of H2PdCl4 solution, respectively. When the volume of H2PdCl4 solution was increased from 5 to 50 μL, the product Au@Pd NPs changed from truncated octahedra into cuboctahedra then into truncated cubes. On further increasing to 100 μL, most of Au@Pd NPs were formed in the cubic shape with a concave structure, thus showing a morphological transformation from truncated octahedra to concave cubes. The formation of various shapes of metal NPs is likely the outcome of the interplay between the faceting tendency of the stabilizing agent and the growth kinetics.20 Here in this work CTAB was employed as the only stabilizer in all the syntheses. It is well documented that CTAB could preferentially adsorb on Pd {100} facets and then lower their growth rate.15 Epitaxial growth of Pd would selectively occur on the truncated-octahedral Au seeds along the ⟨111⟩ and ⟨110⟩ directions, as evidenced by the morphological evolution of Au@Pd NPs from truncated octahedra to truncated cubes (Figure S4a−c). During the subsequent growth, Pd atoms would preferentially add to the corners and edges (corresponding to the ⟨111⟩ and ⟨110⟩ directions, respectively), thus leading to a selective formation of the unique concave cubic morphology (Figure S4d). These observations also indicate that a sufficiently high atomic Pd/Au ratio is a necessity for the selective formation of the concave cubic shape.

ascorbic acid reduction of aqueous H2PdCl4 in the presence of truncated-octahedral Au seeds (Figure 1). The successful growth of a Pd layer on the Au seeds led to the product Au@Pd NPs; the majority of these Au@Pd NPs were found in the cubic shape (Figure 2b) and thus can be referred to as Au@ Pd nanocubes. Extended analysis of the TEM images (Figure S1) demonstrates that the Au@Pd nanocubes were obtained in a high morphology selectivity (>85%) with a narrow size distribution at 21.0 ± 1.6 nm. Figure S2 shows some TEM images of higher magnifications for the Au@Pd nanocubes, which reveal that the product Au@Pd nanocubes were formed with concave rather than flat surfaces. To better describe the structure of the concave surfaces, it would be straightforward to image with HRTEM the high-index facets of the Au@Pd concave nanocubes. However, we failed to obtain an edge-on HRTEM image of the concave facets by using up to date the most powerful aberration-corrected HRTEM (FEI-Titan 80300) because the edges of these concave nanocubes were not concaved, which hindered the concave facets from being directly imaged by HRTEM. This same problem was reported earlier by Mirkin et al.,11 who were not successful in obtaining a HRTEM image for the concave facets of their synthesized large Au concave nanocubes (40−300 nm) unless a large concave cube (270 nm) was cut using a focused ion beam. A second approach to obtain the Miller indices of the concave facets is to use the projection angles along a selected crystallographic axis,16 which has been employed in a number of recent publications.8,17,18 The Au@Pd concave nanocubes (ca. 21 nm) in this present study were too small to be cut for obtaining the edge-on HRTEM image; we used instead the second approach to index the surface facets of these Au@Pd concave nanocubes. Figure 2e shows the HRTEM of an Au@Pd concave nanocube viewed along the [001] zone axis. The angles between the facets of Au@Pd concave nanocube and the {100} facets of an ideal nanocube were 9°, 22°, and 23°, which are in agreement with {610}, {520}, and {730} facets, respectively. Although TEM technique can be used to identify the morphology and surface structure, it can only detect a limited number of NPs. In order to determine whether most of these Au@Pd nanocubes were in concave structures, we compared the atomic Pd/Au ratios for the as-prepared Au@Pd concave nanocube sample and a similarly sized model Au@Pd NP having an ideal cubic shape (see Supporting Information for details). The theoretical atomic Pd/Au ratio would be 21.6 for the ideally modeled Au@Pd cube, which appears much higher than the measured value 6.8 (ICP-AES measurement) for the as-prepared Au@Pd nanocube sample. It should be noted that the nominal Pd/Au ratio in the synthesis from the initial precursor sources was H2PdCl4/HAuCl4 = 10.0; the lower value of Pd/Au in the final sample (6.8) could probably arise from a selective etching of Pd atoms by dissolved O2 (air) in aqueous solution.19 Anyway, the number of Pd atoms in the asprepared Au@Pd particles is reduced to ca. 30% in their regularly shaped cube counterparts. Therefore, it can be concluded that most of the as-prepared Au@Pd nanocubes were not produced in the regularly cubic shape but with concave faces. The selective formation of the unique concave structure in the present synthesis would mean an important technology advance to significantly enhance the exposure and utilization efficiency of the Pd atoms for surface catalysis. During the preparation of this paper, a new publication disclosing a synthesis of much larger (75−150 nm) high-index faceted Au@Pd concave nanocubes, also with a seed-mediated 20842

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

On having Pt deposited onto the as-prepared Au@Pd concave nanocubes, we further obtained trimetallic Pt∧(Au@ Pd) nanostructures showing the TEM images of Figure 2c. These nanostructures appeared strikingly uniform in shape but had unusually jagged surfaces. The Pt∧(Au@Pd) nanostructures (82% in the cubic and the others in irregular shapes) were also narrowly sized, showing a size distribution in the range of 27.8 ± 2.1 nm (Figure S1). A representative HRTEM image of the jagged surfaces is shown in Figure 2f. Though it was quite difficult to distinguish Pt from Pd due to their quite similar lattice constants (0.392 vs 0.389 nm), HRTEM in combination with fast Fourier-transform (FFT) analyses revealed that the near-surface regions of the Pt∧(Au@Pd) nanostructures assumed the single-crystalline structures, indicating an epitaxial growth of Pt on the underlying Pd layer (Figure 2f). The inset of Figure 2f clearly indicates that the edges of the cubic Pt∧(Au@Pd) structures shared an identical FFT pattern featuring the orientations of single crystal Pd, thus providing a piece of evidence for the direct nucleation of Pt on the Pd shell rather than a random aggregation of Pt nuclei.3 Figure 3 shows the HAADF-STEM image and STEM-EDS elemental mapping results for the trimetallic Pt∧(Au@Pd)

Figure 4. XPS spectra of Au@Pd concave nanocubes, Pt∧(Au@Pd) nanostructures, and Pd black. The inset shows the Pt 4f XPS signal for Pt∧(Au@Pd), and the dotted vertical lines mark the binding energies for conventional Pd, Au, and Pt metals.

The surface composition and electronic property of the Au@ Pd and Pt∧(Au@Pd) samples were probed by XPS. Figure 4 shows the XPS signals for Pd 3d and Au 4f, in which the dotted vertical lines mark the binding energies (BEs) for conventional Pd and Au metals. The reference Pd black showed two Pd 3d5/2 signals featuring metallic Pd at 335.4 eV and PdO at 336.7 eV.21 The Pd 3d5/2 signals for Au@Pd and Pt∧(Au@Pd) were found ca. 0.9 eV lower (334.5 eV) than that for metallic Pd in the reference Pd black, signifying no PdO but metallic Pd only. This phenomenon indicates that the Pd surfaces in the bimetallic Au@Pd and trimetallic Pt∧(Au@Pd) nanostructures were not oxidized during exposure to air. However, we found surprisingly that Au in the Au@Pd and Pt∧(Au@Pd) samples still showed some visibility in the XPS measurements though both our TEM and UV−vis measurements would suggest that Au NPs in the two samples were completely covered with the Pd and PtPd overlayers, respectively. This “visibility” of the Au cores would be taken as a further evidence for the unique concave nanostructures of the samples, which demonstrates that at least some regions of the Pd and PdPt overlayers were not thick enough to shield their underlying Au cores from being reached by the X-rays during XPS measurement. The Au 4f signals for both Au@Pd and Pt∧(Au@Pd) were found ca. 0.7 eV lower than that for conventional Au metal. The lower XPS BEs of Pd 3d and Au 4f for the Au@Pd and Pt∧(Au@Pd) samples would arise from a Au−Pd interaction (alloying) at the Au/Pd interfaces,22 as observed in Au−Pd alloys of different dimensions.23−25 These data would suggest that the truncatedoctahedral Au NPs in the Au@Pd and Pt∧(Au@Pd) samples not only acted as structure directing cores but also modified the electronic structure of the Pd shell, probably by Au−Pd alloying at the interfaces. The inset of Figure 4 shows the Pt 4f XPS signal for the Pt∧(Au@Pd) nanostructures. The Pt 4f BEs were found to be the same as for conventional Pt metal (70.9 and 74.0 eV). According to the calibrated intensity of the XPS peaks, we obtained a surface atomic Pt/Pd/Au ratio of 0:68:1 for Au@Pd and 15:102:1 for Pt∧(Au@Pd). These ratios clearly confirm an

Figure 3. (a) HAADF-STEM image of the trimetallic Pt∧(Au@Pd) nanostructures. (b) STEM-EDS elemental mapping images for a representative Pt∧(Au@Pd) particle. (c) An adduct image of the Au (b-2), Pd (b-3), and Pt (b-4) elemental mappings.

sample. Every particle in the HAADF-STEM image (Figure 3a) appears to be composed of a bright core particle and a less bright shell with rough surfaces. This bright core particle must be Au NP since Au would show a stronger atomic number contrast (Z-contrast) than Pd in the shell layers. Figure 3b shows the HAADF-STEM image of a representative Pt∧(Au@ Pd) particle and its corresponding elemental mapping images. Figure 3c is an adduct image of the Au (Figure 3b-2), Pd (Figure 3b-3), and Pt (Figure 3b-4) mappings, which clearly confirms that the trimetallic triple-layered nanostructures of the Pt∧(Au@Pd) particles, with an Au NP in the core, Pd in the inner-layer and Pt-rich clusters in the outermost jagged-surface layers. It is understood that these elemental mapping data cannot discriminate Pt from any PtPd alloy if the Pt∧(Au@Pd) particles could involve a kind of PtPd alloy. However, in consideration of the fact that our XPS data (Figure 4) to be discussed in the next couple of paragraphs uncovered no alloying between Pt and Pd in the Pt∧(Au@Pd) nanostructures, we could safely say that the outermost surfaces of our trimetallic nanocrystals were dominated by pure Pt. 20843

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

enrichment of Pd at the outer surface of Au@Pd and Pt∧(Au@ Pd) nanostructures, which agrees with the layered structures derived from the TEM data (Figure 2). However, the inconsistency in the Pd/Au ratios of Pt∧(Au@Pd) (102:1) and Au@Pd (68:1) points to a difference in surface structures of the two samples. As XPS is a surface-sensitive technique with a probing depth of less than 2 nm for precious metal elements, the significantly higher Pd/Au ratio of Pt∧(Au@Pd) (102:1) would mean a heavier “shielding” of the Au cores by their outside overall “PtPd shell” layers. In fact, the PtPd shell layers in the Pt∧(Au@Pd) nanostructures did appear substantially thicker than the Pd shell of the Au@Pd concave cubes (Figure S5). 3.2. Catalytic Properties of Au@Pd Concave Nanocubes and Pt∧(Au@Pd) Nanostructures for Ethanol Electrooxidation. We report here the potentials of our small Au@Pd concave nanocubes and their derived trimetallic Pt∧(Au@Pd) nanostructures as anode electrocatalysts for ethanol oxidation reaction in alkaline electrolyte. In the work of Huang et al., a superiority for thermal catalysis of their large Au@Pd concave nanocubes (ca. 79 nm) to octahedral Au@Pd NPs (ca. 121 nm) was demonstrated in a Suzuki cross-coupling reaction.17 Figure 5a compares the cyclic voltammetry (CV) curves in 0.1 M KOH (with no ethanol) of the present Au@Pd and Pt∧(Au@Pd) catalysts with those for Pd NPs, conventional

Pd/C, and Pd black. The redox chemistry specific to the metallic Pd was characterized by a strong cathodic peak at around −0.35 V and a broad anodic peak at above −0.25 V.26,27 The broad anodic peak in the potential range between −0.85 and −0.60 V featured the hydrogen desorption associated with the Pt surface at the Pt∧(Au@Pd) nanostructures.28 The electrochemically active surface area (EAS)29,30 data, extracted according to the method described in ref 29, are compared in Figure 5b. The Pd/C catalyst showed the highest Pd EAS (48.8 m2 g−1Pd) while the Pd black exhibited in contrast the smallest EAS (2.7 m2 g−1Pd) mainly due to extensive agglomeration (Figure S6). It is of interest to note that the trimetallic Pt∧(Au@Pd) possessed a substantially higher Pd EAS (20.8 m2 g−1Pd) than the bimetallic Au@Pd (12.9 m2 g−1Pd), though it would be more straightforward to imagine that the deposition of Pt on the preformed Au@Pd particles would lead to a loss of Pd EAS. This phenomenon can be partially rationalized by assuming occurence of a galvanic replacement reaction between PtCl42− and the Pd shell of Au@Pd NP during the synthesis of Pt∧(Au@Pd), which would roughen the Pd shell and make more Pd atoms be exposed. Figure 6a shows the anodic-scan (forward scan) CV curves for EOR on Au@Pd, Pt∧(Au@Pd), and the reference Pd

Figure 6. (a) Anodic-scan CV curves and (b) comparison of IA and MSA data for ethanol electrooxidation on the indicated samples.

catalysts (complete CV curves are shown in Figure S7). The current densities were normalized against EAS of the catalysts to show the intrinsic activity (IA) data of the metallic surfaces. For a better understanding of the observed activity difference, the IA and MAS data at −0.1 V were further compared in Figure 6b. The IA number for Au@Pd (29 A m−2Pd) appeared 6−7 times that for Pd/C (4 A m−2Pd) and Pd NPs (5 A m−2Pd) and ca. 1.5 times that for Pd black (19 A m−2Pd) and Pt∧(Au@ Pd) (16 A m−2Pd+Pt), demonstrating a faster EOR kinetics on the Pd surface of the Au@Pd concave nanocubes. It is believed that low-coordinated atoms in the steps and/or kinks of the high-index metal surfaces would act as active sites for the adsorption and activation of reactant molecules,7,16 as demonstrated 30 years ago by Somorjai et al.,31 who observed

Figure 5. (a) CV profiles and (b) EAS data of Au@Pd concave nanocubes, Pt∧(Au@Pd) nanostructures, conventional Pd/C, Pd NPs, and Pd black. 20844

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

Chronoamperometric (CA) experiments were carried out at −0.1 V to investigate the electrochemical stabilities of each catalyst (Figure 7). All catalysts featured a pronounced current

that the benzene formation rate from the dehydrogenation of cyclohexane on Pt(679) surface was 7 times higher than on Pt(111) surface. There was also evidence that dissociative chemisorption of methanol to form COad took place preferentially on the step sites of single crystal Pt surfaces such as Pt(554), Pt(553), and Pt(335).32,33 Moreover, Pt and Pd NPs having stepped surfaces produced faster reaction kinetics in the electrochemical oxidation of carbon monoxide, methanol, ethanol, and formic acid.4,9,34−36 Therefore, the high IA number for EOR of the present Au@Pd catalyst (Figure 6b) would mean a high-density of low-coordinated Pd atoms at the catalyst surface, which is consistent with the concave surface structure of the Au@Pd catalyst exposing high-index Pd facets (Figure 2). Besides, the enhanced catalytic performance of Au@Pd concave nanocubes may also be due to the modified surface electronic structure of Pd shells by interfacial alloying with the underlying Au cores. According to the XPS results (Figure 4), the Pd 3d5/2 signal for Au@Pd (334.5 eV) was ca. 0.9 eV lower than that for metallic Pd (335.4 eV), indicating that the Pd surface in the Au@Pd concave nanocubes is in electron-rich environments. The increased electron density of Pd would be beneficial to ethanol activation due to an enhanced electron back-donation from Pd to the π* orbital of the surface intermediates,37,38 which would also account for the higher intrinsic catalytic activity of Pd in Au@Pd concave nanocubes than the reference Pd black catalyst. The lower IA of the trimetallic triple-layered Pt∧(Au@Pd) catalyst (Figure 6b) could be due to (i) the deposition of Pt resulted in passivation of some active Pd sites and/or significant collapse of the high-index facets of the preformed supporting Au@Pd and (ii) Pt is known intrinsically less active for EOR than Pd in alkaline media.26,39 Surprisingly, this trimetallic triple-layered Pt∧(Au@Pd) catalyst yet exhibited the highest activity by MSA (0.46 A mg−1Pd+Pt) (Figure 6b), which is 1.2 times that of Au@Pd (0.38 A mg−1Pd) and 2−10 times higher than the other Pd catalysts. Even if the total mass of all the three metals (Pd, Pt, and Au) was taken into account, this Pt∧(Au@Pd) catalyst would still show an overall MSAPd+Pt+Au of 0.38 A mg−1metal, which is 7.6 and 1.7 times higher than the MSAPd data of the conventional Pd black (0.05 A mg−1Pd) and Pd/C (0.22 A mg−1Pd) catalysts, respectively, which would lead to a Pd saving by 40−85%. Thus, the activity data for EOR demonstrated that the Au@ Pd concave cubes and their derived trimetallic triple-layered Pt∧(Au@Pd) structures are highly advantageous to conventional Pd black for the anode catalyst. The ratio of peak currents on the forward and backward scans (i.e., If/Ib) could be used as a criterion to evaluate the poisoning of the catalyst surface by carbon-containing species (identified mainly as acyl26) during EOR.40 This ratio for the Au@Pd concave cubes (If/Ib = 1.5) was significantly higher than those for Pd/C (0.8), Pd NPs (0.5), and Pd black (1.1), indicating a superior antipoison property of the Au@Pd concave cubes during the electrooxidation of ethanol. The antipoison property of the trimetallic Pt∧(Au@Pd) catalyst (If/Ib = 1.1) appeared comparable to that of the Pd black but superior to those of the reference Pd/C (0.8) and Pd NPs (0.5) catalysts. The abundant step and/or kink sites associated with the high-index facets of the Au@Pd and Pt∧(Au@Pd) nanostructures would be responsible for their superior antipoison behavior. Such step and/or kink sites would be efficient in promoting water dissociation at lower potentials and thus entitle faster removal of the poisonous species from the catalyst surface.

Figure 7. CA curves for ethanol electrooxidation on Au@Pd concave nanocubes, Pt∧(Au@Pd) nanostructures, conventional Pd/C, Pd NPs, and Pd black. All experiments were conducted in a 0.5 M aqueous solution of ethanol using 0.1 M KOH as the supporting electrolyte.

decay in the initial 50 s. The decay slowed down but was still observable at longer times. To our surprise, the Au@Pd concave nanocubes exhibited the most pronounced current decay by showing ca. 80% current loss in the first 400 s, though this Au@Pd catalyst actually possessed the best antipoison performance in the CV test. This result may arise from the loss of Pd EAS or irreversible structure rearrangement of the intrinsically more active high-index Pd facets during the CA test. The Pd EAS data of Au@Pd before (12.9 m2 g−1Pd) and after (13.3 m2 g−1Pd) the CA measurement were found to be very similar, but a damage to the concave-structure feature of the Au@Pd nanocubes after the CA test was evidenced in the TEM image (Figure S8). Therefore, the significant current degradation observed during the CA measurement was a reflection of the damage to the intrinsically more active highindex Pd facets. In contrast, the trimetallic Pt∧(Au@Pd) catalyst presented a much less significant current decay. The steady-state mass activity of Pt∧(Au@Pd) catalyst (0.056 A mg−1Pd+Pt) remained 6 times that of Au@Pd (0.009 A mg−1Pd+Pt), and dramatically 10−60 times higher than those of the reference Pd catalysts, demonstrating the great advantage of Pt∧(Au@Pd) catalyst for precious metal-saving catalytic electrodes. Long-term repeated CV measurements of EOR were also conducted to compare the electrocatalytic stabilities of the Au@Pd, Pt∧(Au@Pd), and Pd-black catalysts in 0.1 M KOH with 0.5 M ethanol at 298 K. The CV measurements were repeated for 300 cycles. Figure 8 shows the current variations at −0.2 V on each catalyst. The reference Pd black showed a continued current decay, especially in the first 200 cycles during which the current suffered from a loss of ca. 40%. The Au@Pd concave nanocubes demonstrated a higher catalytic stability though it also showed a current decay in the first 200 cycles, but the decay was less significant (ca. 8%) and became not detectable afterward. The current decay could arise from an accumulation of poisonous carbon-containing species on the catalyst surface. In contrast, the trimetallic Pt∧(Au@Pd) nanostructures showed not only no sign for any current decay during the repeated CV measurements but even a slight increase in the current after 100 cycles. This distinctively high 20845

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C



Article

ASSOCIATED CONTENT

S Supporting Information *

Extended TEM characterization; UV−vis spectra; complete cyclic voltammetry curves of ethanol electrooxidation; determination of the concave structure of Au@Pd nanocubes. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Tel +86 10 6279 2122; Fax +86 10 6277 1149; e-mail bqxu@ mail.tsinghua.edu.cn. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors thank Ms. Ling Hu (Department of Chemical Engineering, Tsinghua University) and Ms. Lina Zhang (Tsinghua-Foxconn Nanotechnology Research Centre) for their kind help in the HRTEM and HAADF-STEM measurements. This work was financially supported by NSF (21033004 and 20921001) of China.

Figure 8. Change of current at −0.2 V in repeated CV measurements of ethanol electrooxidation. The inset shows the magnified curve for Pd black catalyst. All experiments were conducted in a 0.5 M aqueous solution of ethanol using 0.1 M KOH as the supporting electrolyte.

catalysis stability of the trimetallic Pt∧(Au@Pd) catalyst may be attributed to its unique triple-layered structure with jagged Ptrich surfaces, which could function to stabilize the high-index faceted Pd layers/fabrics with interfacial Pd−Pt interaction in or near the outer surfaces. The above CA and repeated CV measurements both demonstrate that the trimetallic Pt∧(Au@Pd) nanostructures are distinctly more stable than the Au@Pd nanocubes and Pd black catalyst in EOR, which could have important implications in practical application. However, the bimetallic Au@Pd concave nanocubes showed quite different decaying performance during the CA and repeated CV measurements of ethanol electrooxidation (Figures 7 and 8). These differences would hint that the deterioration in electrocatalysis of the concave Au@Pd nanocubes is sensitive to the mode of the electrochemical measurement, an important aspect that certainly demands for further investigations in the future.



REFERENCES

(1) Chen, J. Y.; Lim, B.; Lee, E. P.; Xia, Y. N. Nano Today 2009, 4, 81. (2) Somorjai, G. A. Chem. Rev. 1996, 96, 1223. (3) Lim, B.; Jiang, M.; Camargo, P. H. C.; Cho, E. C.; Tao, J.; Lu, X.; Zhu, Y.; Xia, Y. Science 2009, 324, 1302. (4) Jin, M.; Zhang, H.; Xie, Z.; Xia, Y. Angew. Chem., Int. Ed. 2011, 50, 7850. (5) Habas, S. E.; Lee, H.; Radmilovic, V.; Somorjai, G. A.; Yang, P. Nat. Mater. 2007, 6, 692. (6) Steele, B. C. H.; Heinzel, A. Nature 2001, 414, 345. (7) Tian, N.; Zhou, Z. Y.; Sun, S. G.; Ding, Y.; Wang, Z. L. Science 2007, 316, 732. (8) Yu, T.; Kim, D. Y.; Zhang, H.; Xia, Y. Angew. Chem., Int. Ed. 2011, 50, 2773. (9) Huang, X. Q.; Zhao, Z. P.; Fan, J. M.; Tan, Y. M.; Zheng, N. F. J. Am. Chem. Soc. 2011, 133, 4718. (10) Ming, T.; Feng, W.; Tang, Q.; Wang, F.; Sun, L.; Wang, J.; Yan, C. J. Am. Chem. Soc. 2009, 131, 16350. (11) Zhang, J. A.; Langille, M. R.; Personick, M. L.; Zhang, K.; Li, S. Y.; Mirkin, C. A. J. Am. Chem. Soc. 2010, 132, 14012. (12) Yu, Y.; Zhang, Q. B.; Liu, B.; Lee, J. Y. J. Am. Chem. Soc. 2010, 132, 18258. (13) Lu, C. L.; Prasad, K. S.; Wu, H. L.; Ho, J.; Huang, M. H. J. Am. Chem. Soc. 2010, 132, 14546. (14) Wang, F.; Li, C. H.; Sun, L. D.; Wu, H. S.; Ming, T. A.; Wang, J. F.; Yu, J. C.; Yan, C. H. J. Am. Chem. Soc. 2011, 133, 1106. (15) Fan, F.; Liu, D.; Wu, Y.; Duan, S.; Xie, Z.; Jiang, Z.; Tian, Z. J. Am. Chem. Soc. 2008, 130, 6949. (16) Tian, N.; Zhou, Z. Y.; Sun, S. G. J. Phys. Chem. C 2008, 112, 19801. (17) Yang, C.; Chanda, K.; Lin, P.; Wang, Y.; Liao, C.; Huang, M. H. J. Am. Chem. Soc. 2011, 133, 19993. (18) Yin, A. X.; Min, X. Q.; Zhu, W.; Liu, W. C.; Zhang, Y. W.; Yan, C. H. Chem.Eur. J. 2012, 18, 777. (19) Xia, Y.; Xiong, Y. J.; Lim, B.; Skrabalak, S. E. Angew. Chem., Int. Ed. 2009, 48, 60. (20) Yu, Y. T.; Xu, B. Q. Appl. Organomet. Chem. 2006, 20, 638. (21) Brun, M.; Berthet, A.; Bertolini, J. C. J. Electron Spectrosc. 1999, 104, 55. (22) Ding, Y.; Fan, F. R.; Tian, Z. Q.; Wang, Z. L. J. Am. Chem. Soc. 2010, 132, 12480.

4. CONCLUSIONS A facile approach to selectively synthesize narrowly sized small Au@Pd concave nanocubes enclosed by high-index facets was developed and extended to construct trimetallic triple-layered Pt∧(Au@Pd) nanostructures at sub-30 nm. The Au@Pd concave nanocubes and Pt∧(Au@Pd) nanostructures exhibited significantly improved catalytic activity for anodic EOR in terms of both IA and MSA relative to the conventional Pd NPs, Pd/ C, and Pd black catalysts. Although the particle sizes of the asprepared Au@Pd and Pt∧(Au@Pd) particles still seem a bit too large (20−30 nm), their concave shape and unique triplelayered structure already led to efficiently enhanced Pd utilization efficiency in electrocatalysis. The trimetallic Pt∧(Au@Pd) nanostructures also showed much higher catalytic stability in long-term EOR tests. The exposed high-index Pd facets in the Au@Pd concave nanocubes and unique triplelayered structure in the Pt∧(Au@Pd) NPs would further stimulate interest to explore their applications in both thermaland electrocatalysis. It is anticipated that these data could have important impact on the design and synthesis of other highperformance nanostructured bi- and/or multimetallic catalysts featuring high-index facets. 20846

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847

The Journal of Physical Chemistry C

Article

(23) Liu, F.; Wechsler, D.; Zhang, P. Chem. Phys. Lett. 2008, 461, 254. (24) Nascebte, P.; Decastro, S.; Landers, R.; Kleiman, G. G. Phys. Rev. B 1991, 43, 4659. (25) Yi, C. W.; Luo, K.; Wei, T.; Goodman, D. W. J. Phys. Chem. B 2005, 109, 18535. (26) Liang, Z. X.; Zhao, T. S.; Xu, J. B.; Zhu, L. D. Electrochim. Acta 2009, 54, 2203. (27) Woods, R. Electroanalytical Chemistry: A Series of Advances; Marcel Dekker: New York, 1974. (28) Yu, E. H.; Scott, K.; Reeve, R. W. J. Electroanal. Chem. 2003, 547, 17. (29) Electrochemically active surface area (EAS) of Pd was determined from the surface charge associated with the reduction of Pd oxide on the CV curves in 0.1 M KOH, assuming that the charge required for reduction of a monolayer of oxygen on the Pd surface was 406 μC cm−2.27 The EAS of Pt in Pt∧(Au@Pd) nanostructure was obtained by measuring the charges associated with the hydrogen desorption signals on the CV curves in 0.1 M KOH by adopting an assumption of 210 μC cm−2 for the desorption of a hydrogen monolayer. (30) (a) Zhao, D.; Xu, B. Q. Phys. Chem. Chem. Phys. 2006, 8, 5106. (b) Zhang, G. R.; Zhao, D.; Feng, Y. Y.; Zhang, B. S.; Su, D. S.; Liu, G.; Xu, B. Q. ACS Nano 2012, 6, 2226. (31) Somorjai, G. A.; Blakely, D. W. Nature 1975, 258, 580. (32) Shin, J.; Korzeniewski, C. J. Phys. Chem. 1995, 99, 3419. (33) Housmans, T.; Koper, M. J. Phys. Chem. B 2003, 107, 8557. (34) Lee, S. W.; Chen, S.; Sheng, W.; Yabuuchi, N.; Kim, Y.; Mitani, T.; Vescovo, E.; Shao-Horn, Y. J. Am. Chem. Soc. 2009, 131, 15669. (35) Zhang, J.; Zhang, L.; Xie, S.; Kuang, Q.; Han, X.; Xie, Z.; Zheng, L. Chem.Eur. J. 2011, 17, 9915. (36) Nakamura, M.; Hanioka, Y.; Ouchida, W.; Yamada, M.; Hoshi, N. ChemPhysChem 2009, 10, 2719. (37) Tedsree, K.; Li, T.; Jones, S.; Chan, C. W. A.; Yu, K. M. K.; Bagot, P. A. J.; Marquis, E. A.; Smith, G. D. W.; Tsang, S. C. E. Nat. Nanotechnol. 2011, 6, 302. (38) Zhang, G. R.; Xu, B. Q. Nanoscale 2010, 2, 2798. (39) Shen, P. K.; Xu, C. Electrochem. Commun. 2006, 8, 184. (40) Zhu, L. D.; Zhao, T. S.; Xu, J. B.; Liang, Z. X. J. Power Sources 2009, 187, 80.

20847

dx.doi.org/10.1021/jp304570c | J. Phys. Chem. C 2012, 116, 20839−20847