Pt-Encapsulated Pd−Co Nanoalloy Electrocatalysts for Oxygen

Considerable effort is being made worldwide with respect to the .... (30% Gaussian) method after background subtraction using Shirley's method...
1 downloads 0 Views 5MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Pt-Encapsulated Pd-Co Nanoalloy Electrocatalysts for Oxygen Reduction Reaction in Fuel Cells A. Sarkar, A. Vadivel Murugan, and A. Manthiram* Electrochemical Energy Laboratory, Materials Science and Engineering Program, The University of Texas at Austin, Austin, Texas 78712 Received July 27, 2009. Revised Manuscript Received September 20, 2009 Pt-encapsulated PdxCo100-x nanoalloy electrocatalysts supported on carbon have been synthesized by a rapid microwave-assisted solvothermal (MW-ST) method within 15 min at as low as 300 °C. Subsequently, the samples have been heat treated at 900 °C in a reducing gas atmosphere to obtain Pt-Pd-Co nanoalloys. X-ray diffraction (XRD) analysis of the as-synthesized and 900 °C heat-treated samples reveals interesting changes in phase compositions and degree of alloying with Co and Pt contents and heat treatment. Transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) data of the as-synthesized samples confirm Pt enrichment on the surface of the Pd-Co nanoparticles. Rotating disk electrode (RDE) and single cell proton exchange membrane fuel cell measurements reveal that the as-synthesized Pt-encapsulated Pd80Co20 (i.e., 75 wt % Pd80Co20 þ 25 wt % Pt) with 20 wt % total metal loading on carbon or 5 wt % Pt exhibit higher catalytic activity for the oxygen reduction reaction (ORR) compared to Pt with 20 wt % Pt loading on carbon. Significant changes in the catalytic activity for ORR occur on heat treatment at 900 °C as a result of changes in the phase composition and increase in particle size. This study demonstrates that the encapsulation of Pd-Co alloys with Pt offers a significant enhancement in activity for ORR per unit mass of Pt, offering a significant cost savings.

1. Introduction Fuel cells are modular electrochemical devices that convert the chemical energy of fuels directly into electrical energy, offering high efficiency with minimal pollutants, and are attractive for transportation and stationary applications. Considerable effort is being made worldwide with respect to the commercialization of both proton-exchange membrane fuel cells (PEMFC) and direct methanol fuel cells (DMFC) because of their low temperature of operation.1-6 However, commercialization is hampered by the sluggish oxygen reduction reaction (ORR) and methanol oxidation reaction (MOR) along with the high cost of Pt and Pt-based alloy catalysts.1-4 The alloying of Pt with other elements such as V, Cr, Fe, Co, Ni, and Cu has been shown to improve the kinetics of ORR while reducing the cost.5-11 Among the various Pt-based alloy catalysts investigated, Pt-Co alloys have received much attention because they exhibit almost 2 times higher mass-based activity compared to Pt.6-9 The enhanced activity has been attributed to lattice contraction that facilitates the adsorption of molecular oxygen, the formation of a Pt skin due to the dissolution of Co atoms in acidic solutions, and changes in the *Corresponding author. Fax: 512-471-7681. E-mail: [email protected]. edu. (1) Kinoshita, K. Electrochemical Oxygen Technology; Wiley: New York, 1992. (2) Ralph, T. R.; Hogarth, M. P. Platinum Met. Rev. 2002, 46, 3. (3) Arico, A. S.; Srinivasan, S.; Antonucci, V. Fuel Cells 2001, 1, 133. (4) Steele, B. C. H.; Heinzel, A. Nature 2001, 414, 345. (5) Mukerjee, S.; Srinivasan, S. J. Electroanal. Chem. 1993, 357, 201. (6) Mukerjee, S.; Srinivasan, S.; Soriaga, M.; McBreen, J. J. Electrochem. Soc. 1995, 142, 1409. (7) Toda, T.; Igarashi, H.; Uchida, H.; Watanabe, M. J. Electrochem. Soc. 1999, 146, 3750. (8) Antolini, E.; Passos, R. R.; Ticianelli, E. A. Electrochim. Acta 2002, 48, 263. (9) Xiong, L.; Manthiram, A. Electrochim. Acta 2004, 49, 4163. (10) Yano, H.; Kataoka, M.; Yamashita, H.; Uchida, H.; Watanabe, M. Langmuir 2007, 23, 6438. (11) Koh, S.; Leisch, J.; Toney, M. F.; Strasser, P. J. Phys. Chem. C 2007, 111, 3744. (12) Markovic, N. M.; Schmidt, T. A.; Stamenkovic, V.; Ross, P. N. Fuel Cells 2001, 1, 105.

2894 DOI: 10.1021/la902756j

d-band occupancy.12-22 Similarly, alloying Pd with other elements such as Fe, Co, Mo, and W has also been found to increase the activity for ORR significantly.23-26 Additionally, it has been experimentally verified that a Pt monolayer on Pd and Pd-based alloys exhibit catalytic activity several times higher than that of Pt alone.27 The underpotential deposition of Cu on Pd-Co alloys and subsequent displacement by Pt4þ has been shown to increase the ORR activity substantially.27-33 Thus, the growth of Pt overlayers on (or encapsulation by Pt of) nonplatinum noble (13) Markovic, N. M. In Handbook of Fuel Cells: Fundamentals, Technology and Application; Vielstich, W., Lamm, A., Gasteiger H. A., Eds.; John Wiley: Hoboken, NJ, 2003; Vol. 2. (14) Noerskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard, T.; Jonsson, H. J. Phys. Chem. B 2004, 108, 17886. (15) Chen, S.; Ferreira, P. J.; Sheng, W. C.; Yabuuchi, N.; Allard, L. F.; ShaoHorn, Y. J. Am. Chem. Soc. 2008, 42, 13818. (16) Stamenkovic, V.; Mun, B.; Mayrhofer, Karl J. J.; Ross, P. N.; Markovic, N. M.; Rossmeisl, J.; Greeley, J.; Noerskov, J. K. Angew. Chem., Int. Ed. 2006, 45, 2897. (17) Koh, S.; Toney, M. F.; Strasser, P. Electrochim. Acta 2007, 52, 2765. (18) Adzic, R. R. In Electrocatalysis; Lipkowski, J., Ross, P. N., Eds.; VCH: New York, 1998; Vol. 5. (19) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayerhofer, Karl J. J.; Lucas, C. A.; Wang, G.; Ross, P. N.; Markovic, N. M. Nat. Mater. 2007, 6, 241. (20) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (21) Hammer, B.; Noerskov, J. K. Adv. Catal. 2000, 45, 71. (22) Balbuena, P. B.; Altomare, D.; Vadlamani, N.; Bingi, S.; Agapito, L. A.; Seminario, J. M. J. Phys. Chem. A 2004, 108, 6378. (23) Fernandez, J. L.; Raghuveer, V.; Manthiram, A.; Bard, A. J. J. Am. Chem. Soc. 2005, 127, 13100. (24) Sarkar, A.; Vadivel Murugan, A.; Manthiram, A. J. Phys. Chem. C 2008, 112, 12037. (25) Sarkar, A.; Vadivel Murugan, A.; Manthiram, A. J. Mater. Chem. 2009, 19, 159. (26) Shao, M. H.; Sasaki, K.; Adzic, R. R. J. Am. Chem. Soc. 2006, 128, 3526. (27) Zhang, J.; Vukmirovic, M. B.; Sasaki, K.; Nilekar, A. U.; Mavrikakis, M.; Adzic, R. R. J. Am. Chem. Soc. 2005, 127, 12480. (28) Zhang, J.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, N. M.; Adzic, R. R. Angew. Chem., Int. Ed. 2005, 44, 2132. (29) Zhang, J.; Mo, Y.; Vukmirovic, M. B.; Klie, R.; Sasaki, K.; Adzic, R. R. J. Phys. Chem. B 2004, 108, 10955. (30) Zhang, J.; Lima, F. H. B.; Shao, M. H.; Sasaki, K.; Wang, J. X.; Hanson, J.; Adzic, R. J. Phys. Chem. B 2005, 109, 22701.

Published on Web 10/16/2009

Langmuir 2010, 26(4), 2894–2903

Sarkar et al.

Article

metals or alloys is appealing in realizing high Pt-mass-based activity at a reduced cost. We present here a systematic investigation of a series of Ptencapsulated Pd-Co as well as alloyed Pt-Pd-Co electrocatalysts. The electrocatalysts are synthesized by a novel microwaveassisted solvothermal method (MW-ST) within a short reaction time at a temperature as low as 300 °C. The MW-ST approach involves the uniform heating of polar solvents by absorbing microwave energy and subsequent transfer of the heat selectively to the reactants, which reduces thermal gradients inside the reaction vessel and increases the reaction kinetics. The MW-ST method provides a uniform nucleation environment and offers highly crystalline monodisperse multimetallic nanoparticles. Recently, Lai et al.34 have synthesized Pt-Co binary alloys by the microwave irradiation of aqueous solutions containing Pt and Co precursors, and a significant enhancement in activity for ORR has been observed for compositions such as Pt50Co50 as a result of a higher degree of alloying between Pt and Co. Our MW-ST method presented here is anticipated to offer an even better degree of alloying because of the simultaneous application of self-generated pressure under the solvothermal condition along with temperature and microwave irradiation. To ascertain the effect of alloying on ORR, the samples are also subsequently heat treated at 900 °C in a 10% H2 þ 90% Ar mixture. The samples are characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), cyclic voltammetry (CV), CO (carbon monoxide) stripping measurements, hydrodynamic polarization measurements with rotating disk electrodes (RDE), and single-cell PEMFC measurements.

2. Experimental Section Carbon-supported Pt-encapsulated Pd-Co electrocatalysts (100 mg) with a 20 wt % metal loading on carbon were synthesized by a two-step MW-ST process as described below. Required amounts of (NH4)2PdCl4 (Alfa Aesar) and CoCl2.6H2O (Alfa Aesar) were dissolved in 25 mL of tetraethylene glycol (TEG) (Alfa Aesar) under constant stirring to give PdxCo100 - x with 0 e x e 100. After complete dissolution of the salts, 1.5 mL of 1 M NaOH was added along with the required amount of Vulcan XC72R carbon black (Cabot Corp.) with stirring. The mixture was then transferred to a quartz vessel, sealed, and placed on a turntable for uniform heating in an Anton Paar microwave synthesis system (Anton Parr Synthos-3000) equipped with a wireless pressure and temperature sensor. At an operating frequency of 2.45 GHz and power of 600 W, the sample temperature was ramped to and kept at 300 °C for 15 min while the pressure was raised to 40 bar. The resulting precipitate was repeatedly washed with acetone and deionized water to remove traces of TEG and other impurities, and the powder was dried in an air oven. The carbon-supported PdxCo100 - x thus obtained was then added to 25 mL of a TEG solution containing the required amount of H2PtCl6 3 6H2O (Strem Chemicals Inc.), followed by the addition of 1.5 mL of 1 M NaOH to realize (100 - y) wt % PdxCo100 - x þ y wt % Pt (y = 25, 50, and 75) on Vulcan carbon; the total PdxCo100 - x þ Pt metal loading was 20 wt % on 80 wt % carbon, so the y = 25, 50, and 75 samples represent, respectively, (31) Shao, M. H.; Huang, T.; Liu, P.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Adzic, R. R. Langmuir 2006, 22, 10409. (32) Adzic, R. R.; Zhang, J; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.; Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Top. Catal. 2007, 46, 249. (33) Lima, F. H. B.; Zhang, Z.; Shao, M. H.; Sasaki, M.; Vukrimovic, M. B.; Tiacianelli, E. A.; Adzic, R. R. J. Solid State Electrochem. 2008, 12, 399. (34) Lai, F.-J.; Sarma, L. K.; Chou, H.-L.; Liu, D.-G.; Hsieh, C.-A.; Lee, J.-F.; Hwang, B.-J. J. Phys. Chem. C 2009, 113, 12674.

Langmuir 2010, 26(4), 2894–2903

5, 10, and 15 wt % Pt loadings on carbon. The mixture was again subjected to the MW-ST process in the Anton Paar microwave synthesis system as described above, followed by washing the precipitate with acetone and deionized water and drying in an air oven to obtain the Pt-encapsulated Pd-Co electrocatalysts supported on carbon. For comparison, 100 mg of 20 wt % Pt and 100 mg of 20 wt % Pd80Co20 on 80 wt % carbon were also synthesized in a single step by the MW-ST process. Finally, all of the carbon-supported (100 - y) wt % PdxCo100 - x þ y wt % Pt samples were heat treated at 900 °C in a 10% H2 þ 90% Ar gas mixture for 2 h with a heating and cooling rate of 5 °C/min. The samples were all characterized by XRD with Cu KR radiation. All of the XRD patterns were fit with a mixture of Gaussian and Lorentzian profiles (50% Gaussian) after background subtraction using the Jade MDI software. For most of the samples, the lattice parameters were evaluated using the first four peaks at 2θ = 30-85°. The Pt/Pd/Co ratios in the synthesized samples were assessed by averaging the ratios obtained at four different spots in energy-dispersive spectroscopic (EDS) analysis with a JEOL-JSM5610 SEM having an Oxford instruments EDS attachment. Morphological and particle distribution studies were carried out with a JEOL 2010F high-resolution transmission electron microscope (TEM) operating at 200 keV. XPS studies were conducted to assess the surface compositions of the samples with a Kratos Analytical spectrometer using a monochromatic Al KR X-ray source. All of the XPS profiles were fit by the Gaussian-Lorentzian (30% Gaussian) method after background subtraction using Shirley’s method. CV characterizations were carried out with a standard singlecompartment three-electrode cell having a Pt mesh counter electrode, a glassy carbon (5 mm diameter) working electrode, and a double junction Ag/AgCl reference electrode employing an Autolab PGSTAT302N potentiostat (Eco Chemie B.V., Netherlands). All potentials are, however, reported against a normal hydrogen electrode (NHE). In a typical experiment, 10 mg of the carbon-supported catalyst was ultrasonicated in 5 mL of deionized water and 5 mL of 0.15 wt % Nafion solution (diluted from a 5 wt % Nafion solution obtained from Electrochem Inc. by adding an appropriate amount of ethanol) until a dark, homogeneous dispersion was formed. Twenty microliters of the aliquot was drop cast onto the glassy carbon electrode (5 mm in diameter, Pine Instruments) to give an effective carbon-supported catalyst loading of 20.37 μg metal/cm2. The CV experiments were conducted in N2-purged 0.5 M H2SO4 (solutions prepared from Fisher Scientific high-purity Optima grade 18 M H2SO4) at a scan rate of 50 mV/s between 0.0 and 1.1 V (vs NHE) under ambient conditions. The stable voltammograms obtained with the CV experiments are reported here. Rotating disk electrode (RDE) experiments were conducted with a glassy carbon disk electrode (5 mm diameter) mounted onto an interchangeable RDE holder (Pine Instruments) in O2saturated 0.5 M H2SO4. Before each experiment, the glassy carbon electrode was polished to a mirrorlike finish with 0.05 μm alumina (Buehler). The catalyst-coated working electrode was prepared in the same way as mentioned above. Before the hydrodynamic polarization curves were recorded, adventitious organic impurities and thermally adsorbed oxygen were cleaned off of the catalyst surface by cycling 50 times between 0.0 and 1.1 V (vs NHE) at a scan rate of 50 mV/s. The potential was scanned from 1.0 to 0.0 V (vs NHE) at 5 mV/s.35,36 For CO-striping experiments, the catalysts-coated electrode was prepared and cleaned of impurities in the same way as mentioned earlier. CO was then adsorbed on the catalyst surface by the continuous bubbling of CO at 0.05 V vs NHE for 0.5 h in 0.5 M H2SO4. Thereafter, the solution was purged with N2 for (35) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Markovic, N. M.; Ross, P. N. J. Phys. Chem. B 2002, 106, 4181. (36) Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J. J. Electroanal. Chem. 2001, 495, 135.

DOI: 10.1021/la902756j

2895

Article 0.5 h while maintaining the potential at 0.05 V to remove the dissolved CO. Finally, the potential was scanned from 0.05 to 1.1 V vs NHE at 50 mV/s. To determine the electrochemically active surface area, the charge corresponding to CO oxidation was evaluated and a conversion factor of 420 μC/cm2 was assumed. For single-cell fuel cell tests, the catalyst ink prepared by ultrasonicating the required amount of the carbon-supported catalyst, isopropyl alcohol, water, and 40 wt % Nafion was sprayed on top of a commercial gas diffusion layer (BASF) (2.5 cm  2.5 cm), followed by drying in air at 90 °C. The catalyst loading was kept at 0.4 mg/cm2 both for the anode (Alfa Aesar HiSpec 3000) and the cathode in PEMFC. After the required amount of the catalyst was sprayed, 50 mg of 5 wt % Nafion solution was sprayed additionally for better adhesion between the membrane and the catalyst-coated gas diffusion layer. The membrane electrode assemblies (MEA) for PEMFC were made by sandwiching a Nafion 112 membrane (Electrochem Inc.) between the anode and cathode by hot pressing at 140 °C for 2 min at 2500 psi (172 bar) in a Carver temperature-controlled hot press. Fuel cell testing was carried out at 60 °C with hydrogen and oxygen pressures of, respectively, 10 psig (0.689 bar) and 12 psig (0.827 bar), and the humidifier temperature was same as the cell temperature. The gas flow rates were initially 0.2 L/min for H2 and 0.4 L/min for O2, but they varied with load as, respectively, 0.022 and 0.055 L/minA. The fuel cell (5 cm2 active area) was run for at least 6 h before the polarization curves were recorded.

3. Results and Discussion 3.1. Synthesis. Tetra-ethylene glycol (TEG), having a higher boiling point (328 °C) and higher viscosity (50.21 cP) than ethylene glycol, served both as a reaction medium and as a reducing agent during the MW-ST synthesis process.37 However, initial XRD analysis of the precipitate formed before the MW-ST process indicated the formation of Co(OH)2 as a result of the addition of NaOH to the CoCl2.6H2O solution in TEG. Although the subsequent microwave radiation of the mixture of [PdCl4]-2 and the Co(OH)2 nanoprecipitate causes a rapid rise in temperature and pressure, resulting in the formation of Pd-Co nanoparticles with reducing agent TEG, it is possible that the Co(OH)2 precipitate acts as a seed for the nucleation and growth of the Pd-Co nanoparticles. Similarly, the Pd-Co nanoparticles act as a seed during the microwave irradiation of carbon-supported Pd-Co þ [PtCl6]-2 in presence of TEG to form the Pt-encapsulated Pd-Co nanoparticles. Also, because the reduction potential of Pt (þ1.2 V vs NHE) is higher than those of both Pd (þ0.92 V) and Co (-0.28 V), the galvanic displacement of Pd and Co by Pt4þ is possible during the MW-ST process. 3.2. Structural and Compositional Characterizations. 3.2.1. X-ray Diffraction (XRD) Analysis. Figure 1 compares the XRD patterns of the as-synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt on 80 wt % carbon samples prepared by the two-step MW-ST method at 300 °C and after heat treatment at 900 °C in a flowing 10% H2/90% Ar mixture for 2 h. As seen in Figure 1a, the as-synthesized 75 wt % Co100 þ 25 wt % Pt exhibit reflections corresponding to mainly cobalt oxide (Co3O4). No reflections corresponding to metallic Pt or Pt-Co alloy are seen, possibly because of the low concentration of Pt in the carbonsupported electrocatalyst or because of the small particle size. Interestingly, the incorporation of a small amount of Pd as in the case of 75 wt % Pd20Co80 þ 25 wt % Pt results in reflections corresponding to a metallic fcc phase along with very weak reflections corresponding to Co3O4. Additionally, the significant differences in the XRD patterns of the 75 wt % Co100 þ 25 wt % (37) Kidney, A. J.; Parish, W. R. Fundamentals of Natural Gas Processing; Taylor & Francis: Boca Raton, FL, 2006.

2896 DOI: 10.1021/la902756j

Sarkar et al.

Pt and 75 wt % Pd20Co80 þ 25 wt % Pt samples suggest that the presence of Pd catalyzes the reduction of cobalt ions and facilitates the formation of Pd-Co alloys. More importantly, the peaks are highly asymmetric to lower 2θ values for samples with increasing Pd (Pd/Co > 4:6), and the peak positions shift to lower angles with increasing Pd content compared to those of both Pt and Pd, indicating an increase in the degree of alloying between Pd and Co. The highly asymmetric nature of the peaks suggests the presence of two or more phases with similar structures and close lattice parameters. Indeed, a deconvolution of the first three peaks located around 2θ = 40, 45, and 65°, which correspond to the (111), (200), and (220) planes of the fcc structure, clearly indicates the presence of a primary fcc phase with significantly smaller lattice parameter than for Pd (aPd = 3.8902 A˚), which corresponds to a Pd-Co alloy and a secondary fcc phase with a lattice parameter close to that of Pt (aPt = 3.923 A˚). In addition, the peak positions of the primary phase as well as the secondary phase shift gradually to lower 2θ values with increasing Pd content, indicating an expansion of the lattice due to the substitution of smaller Co atoms for larger Pd and Pt atoms. Table 1 gives the lattice parameter values of the various samples. We also point out that in the case of 75 wt % Pd80Co20 þ 25 wt % Pt, the lattice parameter of the primary phase (3.844 A˚) closely matches that of 100 wt % Pd80Co20 (3.861 A˚) synthesized under the same conditions in a single step. The small discrepancy in the lattice parameter could be due to further alloying in the second synthesis step of 75 wt % Pd80Co20 þ 25 wt % Pt. Additionally, the lattice parameter of the secondary phase (3.924 A˚) closely matches that of Pt (3.923 A˚). Although it is possible that the addition of Pt in the second synthesis step results in a primary phase with Pd-rich Pd-Co-Pt alloy and a secondary phase with Pt-rich Pt-Pd-Co alloy, it could not be resolved from the lattice parameter values because of the broad nature of the peaks and the inherent complexity of a trimetallic system. However, in the case of 75 wt % Pd100 þ 25 wt % Pt, the asymmetric nature is much less pronounced and the primary phase has a lattice parameter that is larger than that of Pd, suggesting alloying between Pt and Pd. Increasing the Pt content while keeping the Pd/Co ratio constant at 4:1 on going from 75 wt % Pd80Co20 þ 25 wt % Pt to 50 wt % Pd80Co20 þ 50 wt % Pt results in a noticeable decrease in the asymmetry and a decrease in the lattice parameter of the Ptrich secondary phase, suggesting a slight increase in the degree of alloying between Pt and Pd . However, the lattice parameter of the primary phase (Pd-rich Pd-Co alloy) remains constant at 3.844 A˚. On increasing the Pt content further as in the case of 25 wt % Pd80Co20 þ 75 wt % Pt, the two phases could not be resolved and only one lattice parameter could be obtained. Overall, a small decrease in the lattice parameter of the secondary phase and no increase in the lattice parameter of the primary phase (Table 1) on increasing the Pt content while keeping the Pd/Co ratio constant at 4:1 strongly suggests the possibility of Pt existing both as separate nanoparticles formed by the nucleation and growth of Pt alone and as encapsulating Pt on the Pd-Co nanoparticles formed by the nucleation and growth of Pt on the surface of Pd-Co nanoparticles during the second synthesis step. Indeed, Adzic et al.30,31 have reported strong asymmetry in the XRD profiles observed for Au-Ni and Pd-Co nanoparticles after heat treatment as a result of surface segregation and the formation of Au- and Pd-rich overlayers, respectively, for Au-Ni and Pd-Co nanoparticles. Similarly, we believe that the strong asymmetry observed in the XRD patterns of most of our samples in Figure 1 is partially due to the growth of Pt overlayers on the Pd-Co nanoparticles. However, from the asymmetry of the XRD peaks Langmuir 2010, 26(4), 2894–2903

Sarkar et al.

Article

Figure 1. XRD patterns of the (a) as-synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt (y = 25, 50, and 75) samples obtained by the two-

step MW-ST method and (b) 900 °C heat-treated samples and (c) variations of the lattice parameters with heat-treatment temperature. The solid and dotted lines refer, respectively, to reflections corresponding to the (111) reflections of Pd and Pt. The reflections marked with b, O, and f refer, respectively, to Co3O4, Co, and the ordered PtCo alloy. The % values in the legend refer to wt %. All of the samples have a total metal (PdxCo100 - x þ Pt) loading of 20 wt % on 80 wt % carbon, and the y = 25, 50, and 75 samples represent, respectively, 5, 10, and 15 wt % Pt loadings on carbon.

alone, complete encapsulation of Pt on Pd-Co alloy nanoparticles cannot be established. The XRD patterns of the (100 - y) wt % PdxCo100 - x þ y wt % Pt samples after heat treatment at 900 °C are presented in Figure 1b. The primary phase in all of the samples belonging to the series 75 wt % PdxCo100 - x þ 25 wt % Pt after heat treatment at 900 °C has fcc lattice parameter values that are smaller than those for Pd as evident from a large shift in the peak positions toward higher angles due to an increase in the degree of alloying with Co at higher temperatures. However, the 75 wt % Co100 þ 25 wt % Pt sample consists of metallic fcc Co as the primary phase along with ordered PtCo as the secondary phase. In particular, the 75 wt % Pd20Co80 þ 25 wt % Pt sample after heat treatment at 900 °C consists of two distinct fcc phases with lattice parameters of 3.811 and 3.682 A˚ and some impurity phases, which could not be identified. Clearly, the primary phase consists of a Pd-rich Pt-Pd-Co alloy, and the secondary phase consists of a Co-rich Pt-Pd-Co alloy. Increasing the Pd content further causes a noticeable increase in the lattice parameter values of the secondary phase (Co-rich alloy) and the primary phase (Pd-rich alloy) as a result of the substitution of smaller Co atoms for larger Pd Langmuir 2010, 26(4), 2894–2903

atoms. Indeed, a single phase with a lattice parameter value of 3.817 A˚ was observed for the 75 wt % Pd60Co40 þ 25 wt % Pt sample. Increasing the Pd content still further causes an expansion of the primary phase, as can be noticed in the XRD pattern of 75 wt % Pd80Co20 þ 25 wt % Pt. Moreover, the 75 wt % Pd100 þ 25 wt % Pt sample has a lattice parameter of 3.894 A˚, which is close to the expected lattice parameter value of 3.895 A˚ based on Vegard’s law for a solid solution between 75 wt % Pd100 and 25 wt % Pt, indicating a full degree of alloying. A systematic increase in the concentration of Pt while keeping the Pd/Co ratio constant at 4:1 results in an expansion of the lattice as a result of the insertion of larger Pt atoms, as can be noticed on going from 75 wt % Pd80Co20 þ 25 wt % Pt to 50 wt % Pd80Co20 þ 50 wt % Pt to 25 wt % Pd80Co20 þ 75 wt % Pt (Figure 1b). Interestingly, the XRD pattern of 50 wt % Pd80Co20 þ 50 wt % Pt reveals the presence of two fcc phases with lattice parameters of 3.870 and 3.840 A˚ indicating the presence of a miscibility gap in the ternary phase diagram of Pt-Pd-Co. To investigate further the phase behavior of the Pt-Pd-Co system with increasing amounts of Pt while keeping the Pd/Co ratio constant at 4:1, the 75 wt % Pd80Co20 þ 25 wt % Pt, 50 wt % DOI: 10.1021/la902756j

2897

Article

Sarkar et al.

Table 1. Crystallographic Data of the As-Synthesized and 900 °C Heat-Treated (100 - y) wt % PdxCo100 - x þ y wt % Pt samples heat-treatment structure temperature type (°C) (space group)

lattice parameter (A˚)

75% Pd20Co80 þ 25% Pt 75% Pd40Co60 þ 25% Pt 75% Pd60Co40 þ25% Pt

as synthesized as synthesized as synthesized

75% Pd80Co20 þ 25% Pt

as synthesized

75% Pd100 þ 25% Pt 50% Pd80Co20 þ 50% Pt

as synthesized as synthesized

25% Pd80Co20 þ 75% Pt 100% Pd80Co20 75% Pd20Co80 þ 25% Pt

as synthesized as synthesized 900

75% Pd40Co60 þ 25% Pt

900

75% Pd60Co40 þ 25% Pt 75% Pd80Co20 þ 25% Pt 75% Pd100 þ 25% Pt 50% Pd80Co20 þ 50% Pt

900 900 900 900

25% Pd80Co20 þ 75% Pt

900

3.866 3.872 3.907 3.820 3.924 3.844 3.899 3.918 3.844 3.897 3.861 3.811 3.682 3.822 3.788 3.817 3.849 3.894 3.870 3.840 3.873

samplea

a

Pd(225) Pd(225) Pt(225) Pd(225) Pt(225) Pd(225) Pd(225) Pt(225) Pd(225) Pt(225) Pd(225) Pd(225) Co(225) Pd(225) Co(225) Pd(225) Pd(225) Pd(225) Pt(225) Pd(225) Pt(225)

All of the % values refer to wt %.

Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples were each heat treated at 500 and 700 °C and the variations of the lattice parameters with heat-treatment temperature are shown in Figure 1c. The data reveal that both the 75 wt % Pd80Co20 þ 25 wt % Pt and 25 wt % Pd80Co20 þ 75 wt % Pt samples form a single-phase solid solution with complete alloying after heat treatment at 700 °C. However, the 50 wt % Pd80Co20 þ 50 wt % Pt sample shows two phases at all temperatures. Interestingly, the lattice parameters of the two phases of the 50 wt % Pd80Co20 þ 50 wt % Pt (3.840 and 3.870 A˚) closely match the lattice parameter values of 75 wt % Pd80Co20 þ 25 wt % Pt (3.849 A˚) and 25 wt % Pd80Co20 þ 75 wt % Pt (3.873 A˚). Therefore, it can be suggested that the 75 wt % Pd80Co20 þ 25 wt % and 25 wt % Pd80Co20 þ 75 wt % Pt samples are close to the phase boundaries in the Pt-Pd-Co system and Pt gets redistributed into the Pt-rich phase and the Pt-poor phase. 3.2.2. TEM Analysis. The particle size and distribution were characterized by TEM and are presented in Figure 2a,b. As seen, the as-synthesized 75 wt % Pd80Co20 þ 25 wt % Pt has a mean particle diameter of 11.2 ( 3.6 nm with a narrow particle size distribution without much agglomeration of the nanoparticles. Moreover, the high-resolution TEM as shown in the inset of Figure 2a demonstrates the crystalline nature of the nanoparticles. However, the Pt nanoparticles synthesized by one-step MWST have a slightly larger mean particle size of 13.9 ( 3.9 nm. Although the 75 wt % Pd80Co20 þ 25 wt % Pt sample has been synthesized by a two-step method compared to Pt, the unimodal nature of the particle size distribution indicates either a rapid growth of individual Pt nanoparticles compared to bimetallic PdCo nanoparticles during the second synthesis step or the nucleation and growth of Pt on the surfaces of already-formed Pd-Co nanoparticles. The similarity of the mean particle sizes suggests that both the above particle growth mechanisms might be occurring simultaneously. Figure 2b shows the TEM image and the particle size distribution of the 75 wt % Pd100 þ 25 wt % Pt sample after heat treatment at 900 °C. It indicates a considerable increase in the mean particle diameter accompanied by a 2898 DOI: 10.1021/la902756j

Figure 2. TEM images of the (a) as-synthesized and (b) 900 °C heat-treated 75 wt % Pd80Co20 þ 25 wt % Pt samples. The insets show the particle size distributions and HR-TEM images.

broadening of the particle size distribution and particle agglomeration after heat treatment at 900 °C. 3.2.3. X-ray Photoelectron Spectroscopy (XPS) Analysis. Figure 3a-c shows the Pt 4f, Pd 3d, and Co 2p core-level X-ray photoelectron spectroscopic (XPS) profiles of 75 wt % Co100 þ 25 wt % Pt, 75 wt % Pd80Co20 þ 25 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt. A decovolution of the Pt 4f profiles as shown in Figure 3a clearly indicates a lower-bonding-energy component corresponding to Pt0 and higher-binding-energy components corresponding to surface oxides of Pt such as PtO (Pt2þ) and PtO2 (Pt4þ). Similarly, the samples containing Pd reveal the presence of metallic Pd along with oxides on the surface. Interestingly, the XPS signals corresponding to Co or Co-oxides are weak in the case of the 75 wt % Pd80Co20 þ 25 wt % Langmuir 2010, 26(4), 2894–2903

Sarkar et al.

Article

Figure 3. Core-level XPS profiles (dots) of the (a) Pt 4f, (b) Pd 3d, and (c) Co 2p regions and the corresponding fitting results (dotted line) of the 75 wt % Co100 þ 25 wt % Pt, 75 wt % Pd80Co20 þ 25 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples obtained by the MW-ST method. The red and blue lines in panels a and b refer to metallic and oxide peaks, respectively.

Pt sample, and no detectable Co signals are observed in the 25 wt % Pd80Co20 þ 75 wt % Pt sample. However, the Co signals could be clearly observed in 75 wt % Co100 þ 25 wt % Pt as seen in Figure 3c. These observations suggest that at both high concentrations of Pt and Pd, the surface is predominantly Pt- and Pd-rich. The absence of Co on the samples with high Pd and Pt contents also suggests that the addition of NaOH in the first synthesis step results in the formation of nanoparticles of cobalt hydroxide that act as nucleation centers for the precipitation of Pd and subsequently the Pd-Co particles act as seeds for the nucleation of Pt. Table 2 gives the surface concentrations as well as the compositions determined from SEM-EDS analysis for all of the samples. A comparison of the relative surface concentrations of Pt, Pd, and Co reveals a significantly higher surface concentration of Pt, particularly with samples containing higher Pd content. This can be attributed to either a large number of smaller Pt particles or the Langmuir 2010, 26(4), 2894–2903

formation of Pt surface layers on the already formed Pd-Co nanoparticles during the second synthesis step. Moreover, the XRD data, which shows presence of a secondary phase having a lattice parameter close to that of Pt, and the TEM data, which shows the absence of bimodal particle size distribution, strongly support the existence of Pt-encapsulated Pd-Co particles. We also emphasize that XPS is not purely a surface technique, with the photoelectrons originating from layers that are several atoms thick. 3.3. Electrochemical Characterization. Figure 4 presents the cyclic voltammograms (CV) of selected as-synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt samples prepared by the two-step MW-ST method at 300 °C and after heat treatment at 900 °C in a flowing 10% H2 þ 90% Ar mixture for 2 h. As seen in Figure 4a, the absorption of hydrogen in palladium causes significant interference in the CV profiles of 75 wt % Pd100 þ 25 wt % Pt. DOI: 10.1021/la902756j

2899

Article

Figure 4. Cyclic voltammogramms of selected (a) as-synthesized and (b) 900 °C heat treated (100 - y) wt % PdxCo100 - x þ y wt % Pt samples. The data were collected in N2-saturated 0.5 M H2SO4 at room temperature at a scan rate of 50 mV/s. The % values in the legend refer to wt %.

However, the absorption is significantly less in samples containing Co possibly because of the alloying of Pd with Co. Additionally, the nearly identical CV profiles observed for the 75 wt % Pd80Co20 þ 25 wt % Pt, 50 wt % Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples suggest that the electrochemically active surface area does not increase with increasing Pt content. Consequently, it may be concluded that the increase in Pt content does not significantly increase the number of individual nanoparticles but results in thicker Pt layers on the Pd-Co nanoparticles. Moreover, after heat treatment at 900 °C, significantly less hydrogen absorption is seen because of the alloying of Pd with both Co and Pt. In addition, the CV profile of 75 wt % Pd100 þ 25 wt % Pt after heat treatment at 900 °C has a significantly higher surface oxide reduction peak compared to those of the other samples. As in the case of as-prepared samples, the CV profiles of 75 wt % Pd80Co20 þ 25 wt % Pt, 50 wt % Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt are found to be similar. However, no specific trend is observed from the CVs of the as-synthesized samples or samples heated to 900 °C. Because of the uncertainty in the accurate determination (38) Jerkiewicz, G. Prog. Surf. Sci. 1998, 57, 137. (39) Wang, D.; Lee, K. Y.; Luo, S.; Flangan, T. B. J. Alloys Compd. 1997, 252, 209. (40) Grden, M.; Piacik, A.; Koczorowski, Z.; Czerwinski, A. J. Electroanal. Chem. 2002, 532, 35.

2900 DOI: 10.1021/la902756j

Sarkar et al.

of the adsorption characteristics of hydrogen on Pd,38 Pd alloys,39,40 and Pd/Pt core/shell nanoparticles,41 the electrochemical surface area, which assumes monolayer coverage of hydrogen, could not be determined. The rotating disk electrode (RDE) data given in Figure 5a compares the oxygen reduction reaction (ORR) activity of the (100 - y) wt % PdxCo100 - x þ y wt % Pt series of electrocatalysts synthesized by the MW-ST method at 300 °C along with commercial (Alfa Aesar HiSpec 3000) 20 wt % Pt supported on carbon. As seen, a clear limiting current is obtained for all of the samples. Additionally, as seen in Figure 5a, the characteristic polarization curves shift toward higher potentials with increasing Pd content in the electrocatalyst, indicating an increase in activity with Pd. Indeed, the mass-transfer-corrected Tafel plots as shown in Figure 5b and the variation of the kinetic current density at 0.8 V versus NHE with composition as shown in the inset of Figure 5b clearly demonstrate a monotonic increase in the activity of ORR with increasing Pd content up to a Pd/Co ratio of 4:1 and a decrease in activity thereafter. The 75 wt % Pd80Co20 þ 25 wt % Pt sample that has a total metal loading of 20 wt % or a Pt loading of 5 wt % on carbon shows the maximum activity for ORR. Moreover, the activity of the 75 wt % Pd80Co20 þ 25 wt % Pt sample is significantly higher (∼10-fold) than that of the Pd80Co20 sample synthesized by the same method. This clearly indicates that the ORR activity is influenced by Pt encapsulation on the Pd80Co20 nanoalloys. A comparison of the activities for ORR (kinetic current density at 0.8 V vs NHE) of the 75 wt % Pd80Co20 þ 25 wt % Pt sample, the Pt sample synthesized by the single-step MW-ST method, and commercial Pt on the basis of per unit mass of metal is presented in Figure 5c. As seen, the activity for ORR of the 75 wt % Pd80Co20 þ 25 wt % Pt sample is significantly higher than that of the 100% Pt sample synthesized by the single-step MW-ST method. Although the mass-based activity of the 75 wt % Pd80Co20 þ 25 wt % Pt sample is lower than that of commercial Pt (Alfa Aesar HiSpec 3000) because of the larger particle size, the activity per unit mass of Pt is considerably higher. Additionally, increasing the Pt content while keeping the Pd/Co ratio at 4:1 did not increase the activity for ORR, as can be seen in Figure 5d. Although the TEM data do not show the core-shell morphology, the XRD and XPS data strongly suggest the presence of Pt-encapsulated Pd-Co nanoparticles along with Pt nanoparticles. Thus, an increase in Pt concentration possibly results in a thicker Pt layer on Pd80Co20 nanoparticles, accompanied by a small increase in particle size. This explains the almost identical CV profiles and similar activities for ORR of the 25 wt % Pd80Co20 þ 75 wt % Pt, 50 wt % Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples. Although one would expect the voltammetric behavior to resemble that of Pt on increasing the Pt content, we did not observe such features possibly because of the degree of alloying between Pd and Pt, as can be seen from the lattice parameter values of the Pt phase (i.e., Pt(225) in Table 1) in the 75 wt % Pd80Co20 þ 25 wt % Pt and 50 wt % Pd80Co20 þ 50 wt % Pt samples. The CO stripping curves for selected catalysts are shown in Figure 6a. For commercial Pt, the electrochemically active surface area evaluated from the charges corresponding to CO oxidation was found to be 67.9 m2/g, which is in good agreement with the electrochemical surface area evaluated from the hydrogen desorption region of the CV in N2-purged H2SO4 (76.0 m2/g). The absence of hydrogen desorption peaks in the hydrogen region of the CO stripping curve (0.0 V to 0.4 V vs NHE) indicates nearly (41) Kobayashi, H.; Yamauchi, M.; Kitagawa, H.; Kubota, Y.; Kato, K.; Takata, M. J. Am. Chem. Soc. 2008, 130, 1818.

Langmuir 2010, 26(4), 2894–2903

Sarkar et al.

Article

Figure 5. (a) Hydrodynamic polarization curves of the as-synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt samples along with commercial Pt in O2-saturated 0.5 M H2SO4 at room temperature with a rotation rate of 1600 rpm, (b) mass-transfer-corrected Tafel plots of the as-synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt samples along with as-synthesized Pt, as-synthesized Pd80Co20, and commercial Pt, with the inset showing the variation of kinetic current density at 0.8 V vs NHE with Pd content, (c) mass specific current density at 0.8 V vs NHE on the basis of per unit mass, per unit mass of Pt, and per unit mass of noble metal, and (d) hydrodynamic polarization curves of the as-synthesized 75 wt % Pd80Co20 þ 25 wt % Pt, 50 wt % Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples. The % values in the legend refer to wt %.

complete coverage of CO on the surface of Pt nanoparticles. However, the electrochemically active surface area of the 75 wt % Pd80Co20 þ 25 wt % Pt sample has been found to be 11.0 m2/g. The much decreased electrochemical surface area is due to the significantly larger particle sizes as observed in TEM. Additionally, the peak potential due to CO oxidation has shifted in the positive direction compared to that of commercial Pt. As-synthesized Pt still shows a smaller surface area of 8.5 m2/g consistent with the TEM data, which indicates a larger particle size for the assynthesized Pt nanoparticles compared to that for the 75 wt % Pd80Co20 þ 25 wt % Pt nanoparticles. Moreover, the CO stripping curves of commercial Pt and as-synthesized Pt are nearly identical in shape (Figure 6). Interestingly, the peak potential due to CO oxidation in the case of as-synthesized Pd80Co20 has shifted in the positive direction by almost 0.2 V. We would also like to emphasize that the CO oxidation profile of the 75 wt % Pd80Co20 þ 25 wt % Pt sample is significantly different from that of as-synthesized Pt and as-synthesized Pd80Co20. Furthermore, the current signal of the 75 wt % Pd80Co20 þ 25 wt % Pt sample is not a summation of the current signals corresponding to as-synthesized Pt and assynthesized Pd80Co20 samples, indicating that our sample is not a mechanical mixture of separate Pd80Co20 and Pt nanoparticles. The surface area specific activities obtained by dividing the mass activity (current density per unit mass at 0.80 V vs NHE) by Langmuir 2010, 26(4), 2894–2903

the electrochemical surface area per unit mass (evaluated from the CO striping charge) of the commercial Pt, 75 wt % Pd80Co20 þ 25 wt % Pt, and as-synthesized Pt samples are compared in Figure 6b. The surface area specific current density of commercial Pt has been found to be 124 μA/cm2, which is in good agreement with the literature value.34,35 Significantly, Pt and 75 wt % Pd80Co20 þ 25 wt % Pt samples synthesized by the MW-ST method show considerably higher activity per unit electrochemically active surface area. Moreover, the surface area specific activities of both the as-synthesized Pt and 75 wt % Pd80Co20 þ 25 wt % Pt samples are nearly the same, which demonstrates the effect of encapsulation of Pt on the Pd80Co20 alloy nanoparticles. The polarization curves for ORR of the samples after heat treatment at 900 °C are presented in Figure 7. As seen, the electrocatalytic activities for ORR of most of the samples decrease considerably after heat treatment at 900 °C compared to that of the as-synthesized samples because of the increase in particle size. The increased activity for ORR of the 75 wt % Co100 þ 25 wt % Pt sample after heat treatment at 900 °C compared to that of the assynthesized sample is possibly due to the formation of an ordered PtCo phase as revealed by the XRD data. The addition of Pd to Co increases the activity as seen in the 75 wt % Pd20Co80 þ 25 wt % Pt sample compared to that in the 75 wt % Co100 þ 25 wt % Pt sample. However, increasing the Pd content further first decreases DOI: 10.1021/la902756j

2901

Article

Sarkar et al.

Figure 7. (a) Hydrodynamic polarization curves of the (100 - y)

Figure 6. (a) CO stripping curves of commercial Pt, 75 wt % Pd80Co20 þ 25 wt % Pt, as-synthesized Pt, and as-synthesized Pd80Co20 samples in 0.5 M H2SO4 and (b) surface area specific current density of commercial Pt, 75 wt % Pd80Co20 þ 25 wt % Pt, and as-synthesized Pt samples at 0.85 V vs NHE. (The electrochemical surface area was evaluated from the CO stripping charge.) The % values in the legend refer to wt %.

the activity until a Pd/Co ratio of 3:2 is attained and then increases the activity until the Co-free Pt-Pd alloy is attained. Overall, the maximum activity for ORR is observed with the 75 wt % Pd100 þ 25 wt % Pt sample. A comparison of the ORR activities of the 75 wt % Co100 þ 25 wt % Pt and 75 wt % Pd20Co80 þ 25 wt % Pt samples illustrates the synergistic effect of the addition of a small amount of Pd to the Pt-Co alloy system. Additionally, similar to the results obtained for the as-synthesized samples, the activity for ORR does not change appreciably on increasing the Pt content while keeping the Pd/Co ratio constant at 4:1. However, unlike the as-synthesized samples, the origin of similar activities can be traced to the phase structure of the samples. Because the 50 wt % Pd80Co20 þ 50 wt % Pt sample consists of two phases having compositions similar to 75 wt % Pd80Co20 þ 25 wt % Pt and 25 wt % Pd80Co20 þ 75 wt % Pt, the activity for ORR is probably a mass-based average of the activities of both phases. Figure 8 compares the performance in a single-cell protonexchange-membrane fuel cell (PEMFC) at 60 °C for commercial Pt, as-synthesized 75% Pd80Co20 þ 25% Pt, and Pt synthesized by the MW-ST method. Although the performance of 75% Pd80Co20 þ 25% Pt is lower than that of commercial Pt, it is higher than that of the Pt sample synthesized by the same method and having a 2902 DOI: 10.1021/la902756j

wt % PdxCo100 - x þ y wt % Pt samples after heat treatment at 900 °C, with the inset showing the variation of kinetic current density at 0.8 V vs NHE with Pd content and (b) hydrodynamic polarization curves of the 75 wt % Pd80Co20 þ 25 wt % Pt, 50 wt % Pd80Co20 þ 50 wt % Pt, and 25 wt % Pd80Co20 þ 75 wt % Pt samples after heat treatment at 900 °C. The % values in the legend refer to wt %.

Figure 8. Comparison of the ORR activities at 60 °C of the commercial Pt, as-synthesized Pt, and as-synthesized 75 wt % Pd80Co20 þ 25 wt % Pt samples in a single-cell proton-exchangemembrane fuel cell. The % values in the legend refer to wt %.

similar particle size. The single-cell data are consistent with the RDE data presented earlier. The electrochemical data thus Langmuir 2010, 26(4), 2894–2903

Sarkar et al.

Article

Table 2. Nominal Compositions, Bulk Compositions, and Surface Compositions Determined from the XPS Profiles of the As-Synthesized (100 - y) wt % PdxCo100 - x þ y wt % Pt Samples nominal atomic composition

samplea 75% Co100 þ 25% Pt 75% Pd20Co80 þ 25% Pt 75% Pd40Co60 þ 25% Pt 75% Pd60Co40 þ 25% Pt 75% Pd80Co20 þ 25% Pt 75% Pd100 þ 25% Pt 50% Pd80Co20 þ 50% Pt 25% Pd20Co80 þ 75% Pt a All of the % values refer to wt %.

Pt9.1Co90.9 Pt10.5Pd17.9Co71.6 Pt11.8Pd35.3Co52.9 Pt13.0Pd52.2Co34.8 Pt14.2Pd68.6Co17.2 Pt15.4Pd84.6 Pt33.2Pd53.4Co13.4 Pt59.8Pd32.2Co8.0

demonstrate that encapsulating the PdxCo100 - x nanoalloys within Pt offers an attractive approach to lowering the Pt loading and overall catalyst cost, whereas the Pd-Co core may have an important influence in enhancing the activity of the outer Pt layer.

4. Conclusions A rapid two-step MW-ST synthesis process taking advantage of both the microwave and solvothermal effects has been pursued to obtain carbon-supported Pt-encapsulated Pd-Co nanoparticles with high crystallinity and controlled particle size. A detailed analysis by XRD, TEM, and XPS of the as-synthesized samples confirms the presence of Pt overlayers on the surface of Pd-Co nanoparticles. RDE experiments show that the 75 wt % Pd80Co20 þ 25 wt % Pt sample with a Pt loading of 5 wt % exhibits higher catalytic activity for ORR compared to either Pd80Co20 or Pt synthesized by the same method (all having the same total metal

Langmuir 2010, 26(4), 2894–2903

bulk composition determined by SEM-EDS

surface composition determined by XPS

Pt3.7Co96.3 Pt12.5Pd21.7Co65.8 Pt8.5Pd38.0Co53.5 Pt10.0Pd60.3Co29.7 Pt12.2Pd76.2Co11.6 Pt13.7Pd86.3 Pt37.3Pd52.3Co10.4 Pt60.4Pd30.5Co9.1

Pt8.7Co91.3 Pt20.2Pd16.6Co63.2 Pt11.6Pd31.0Co57.4 Pt23.4Pd50.0Co26.6 Pt30.2Pd59.0Co10.8 Pt30.4Pd69.6 Pt58.7Pd41.3 Pt64.5Pd35.5

loading of 20 wt %), demonstrating the influence of the Pd-Co core on the outermost Pt layer in enhancing the catalytic activity. Furthermore, the 75 wt % Pd80Co20 þ 25 wt % Pt sample shows almost 3 times higher activity per unit electrochemical surface area compared to commercial Pt. Single-cell PEMFC data confirms the RDE data and affirms the enhancement in catalytic activity per unit mass of Pt on encapsulating Pd-Co by Pt. The activity for ORR, however, decreases on heat treating at 900 °C as a result of the changes in phase composition and particle growth. This study demonstrates that the encapsulation of lower-cost metals or alloys with Pt by novel synthesis approaches such as the MW-ST method presented here represents an attractive strategy for lowering the Pt loading and catalyst cost without sacrificing catalytic activity. Acknowledgment. Financial support by the National Science Foundation (grant CBET-0651929) is gratefully acknowledged.

DOI: 10.1021/la902756j

2903