Oxygen Reduction Activity of Carbon-Supported Pt ... - ACS Publications

The use of such a nanocapsule as the limited reaction space should provide monodispersed alloy particles with uniform composition. Liu et al. examined...
0 downloads 0 Views 1019KB Size
Download PDF
6438

Langmuir 2007, 23, 6438-6445

Oxygen Reduction Activity of Carbon-Supported Pt-M (M ) V, Ni, Cr, Co, and Fe) Alloys Prepared by Nanocapsule Method Hiroshi Yano,† Mikihiro Kataoka,† Hisao Yamashita,† Hiroyuki Uchida,‡ and Masahiro Watanabe*,† Clean Energy Research Center, UniVersity of Yamanashi, Takeda 4, Kofu 400-8510, Japan, and Interdisciplinary Graduate School of Medicine and Engineering, UniVersity of Yamanashi, Takeda 4, Kofu 400-8511, Japan ReceiVed January 11, 2007. In Final Form: March 13, 2007 Monodispersed Pt and Pt-M (M ) V, Cr, Fe, Co, and Ni) alloy nanoparticles supported on carbon black (denoted as Pt/CB and Pt-M/CB) were prepared by the simultaneous reduction of platinum acetylacetonate and the second metal acetylacetonate within nanocapsules formed in diphenyl ether in the presence of carbon black. For the Pt/CBs, the average Pt diameters measured by scanning transmission electron microscopy (STEM) or X-ray diffraction (XRD) ranged from 2.0 to 2.5 nm, regardless of the catalyst-loading level from 10 to 55 wt % on CB. The alloy composition was found to be well-controlled to the projected value among the supported particles. The activities for the oxygen reduction reaction (ORR) at Nafion-coated catalysts in O2-saturated 0.1 M HClO4 solution were evaluated by using a channel flow electrode (CFE) cell at 30 °C. The area-specific ORR activities at Pt-M/CB were found to be 1.3 to 1.8 times higher than that at Pt/CB. The ORR activity increased in the order Pt/CB < Pt-Ni/CB < Pt-Fe/CB < Pt-Co/CB < Pt-V/CB < Pt-Cr/CB.

1. Introduction Polymer electrolyte fuel cells (PEFCs) have been intensively developed as a primary power source for electric vehicles or residences. Development of highly active electrocatalysts for the oxygen reduction reaction (ORR) is one of the most important subjects in order to achieve high efficiency. To evaluate the intrinsic catalytic activity for the ORR, planar electrodes with well-defined characteristics (surface area, surface and bulk compositions, and crystal structure) have been examined in aqueous acidic electrolyte solutions. Enhanced ORR activities at the planar electrodes of Pt alloyed with non-precious-metals such as Fe,1-4 Co,1,4-8 Ni,1,3-11 Mn,4 Cr,5-7,12,13 and V14 have been reported. Recently, we reported that the apparent rate constant for the ORR at the Pt alloy electrode was larger than * To whom correspondence should be addressed. Phone: +81-55-2208620. Fax: +81-55-254-0371. E-mail: [email protected]. † Clean Energy Research Center. ‡ Interdisciplinary Graduate School of Medicine and Engineering. (1) Toda, T.; Igarashi, H.; Uchida, H.; Watanabe, M. J. Electrochem. Soc. 1999, 146, 3750-3756. (2) Toda, T.; Igarashi, H.; Watanabe, M. J. Electroanal. Chem. 1999, 460, 258-262. (3) Toda, T.; Honma, I. Trans. Mater. Res. Soc. Jpn. 2003, 28, 215-220. (4) Mukerjee, S.; Srinivasan, S.; Soriaga, M. P.; McBreen, J. J. Electrochem. Soc. 1995, 142, 1409-1422. (5) Mukerjee, S.; Srinivasan, S. J. Electroanal. Chem. 1993, 357, 201-224. (6) Min, M.; Cho, J.; Cho, K.; Kim, H. Electrochim. Acta 2000, 45, 42114217. (7) Neergat, N.; Shukla, A. K.; Gandhi, K. S. J. Appl. Electrochem. 2001, 31, 373-378. (8) Stamenkovic, V.; Schmidt, T. J.; Markovic, N. M.; Ross, P. N., Jr. J. Phys. Chem. B 2002, 106, 11970-11979. (9) Stamenkovic, V.; Schmidt, T. J.; Ross, P. N., Jr.; Markovic, N. M. J. Electroanal. Chem. 2003, 554-555, 191-199. (10) Drillet, J.-F.; Ee, A.; Friedemann, J.; Ko¨tz, R.; Schnyder, B.; Schmidt, V. M. Electrochim. Acta 2002, 47, 1983-1988. (11) Yang, H.; Vogel, W.; Lamy, C.; Alonso-Vante, N. J. Phys. Chem. B 2004, 108, 11024-11034. (12) Paffett, M. T.; Berry, J. G.; Gottesfeld, S. J. Electrochem. Soc. 1988, 135, 1431-1436. (13) Thamizhmani, G.; Capuano, G. A. J. Electrochem. Soc. 1994, 141, 968975. (14) Antolini, E.; Passos, R. R.; Ticianelli, E. A. Electrochim. Acta 2002, 48, 263-270.

that at the Pt electrode by a factor of 4.0 (Pt54Fe46) to 2.4 (Pt63Ni37) in the temperature range from 20 to 50 °C and that H2O2 yields at the alloy electrodes were low, ranging from 0.2% to 1.0%.15 In order to obtain high mass activity of platinum in PEFCs, it is essential to disperse such Pt alloy nanoparticles on high surface area supports. Although various binary4-6,11,14,16-19,22-24 and ternary13,20-23 Pt-based alloys supported on carbon have been prepared, a large discrepancy has been seen among their ORR activities. Because the ORR activities of these supported Pt-alloy catalysts are not always consistent with those evaluated for bulk alloys, one of the reasons could be ascribed to nonuniform chemical composition among the whole alloy nanoparticles. Usually, carbon-supported Pt-based alloy catalysts were prepared by the impregnation of the second metal precursors on Pt/CB, followed by heating at above 700 °C under an inert gas or hydrogen to form alloys.6,13,22-24 However, such a heat treatment frequently made the metal particles sintered as well as nonuniform composition depending on the particle size, resulting in a decrease in the mass activity for the ORR and presumably in faster degradation of the activity with the operation time. By reducing the reaction temperature around 200 °C with the use of carbonyl complex process11 or water-oil microemulsion method,16 the sintering of the alloy particles was suppressed, but the control (15) Wakabayashi, N.; Takeichi, M.; Uchida, H.; Watanabe, M. J. Phys. Chem. B 2005, 109, 5836-5841. (16) Xiong, L.; Manthiram, A. Electrochim. Acta 2005, 50, 2323-2329. (17) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Radmilovic, V.; Markovic, N. M.; Ross, P. N. J. Phys. Chem. B 2002, 106, 4181-4191. (18) Salgado, J. R. C.; Antolini, E.; Gonzalez, E. R. J. Electrochem. Soc. 2004, 151, A2143-A2149. (19) Lima, F. H. B.; Ticianelli, E. A. Electrochim. Acta 2004, 49, 4091-4099. (20) Luo, J.; Kariuki, N.; Han, L.; Wang, L.; Zhong, C.-J.; He, T. Electrochim. Acta 2006, 51, 4821-4827. (21) Freund, A.; Lang, J.; Lehmann, T.; Starz, K. A. Catal. Today 1996, 27, 279-283. (22) Beard, B. C.; Ross, P. N. J. Electrochem. Soc. 1990, 137, 3368-3374. (23) Shukla, A. K.; Neergat, M.; Bera, P.; Jayaram, V.; Hegde, M. S. J. Electroanal. Chem. 2001, 504, 111-119. (24) Arico, A. S.; Poltarzewski, Z.; Kim, H.; Morana, A.; Giordano, N.; Antonucci, V. J. Power Sources 1995, 55, 159-166.

10.1021/la070078u CCC: $37.00 © 2007 American Chemical Society Published on Web 04/19/2007

ActiVity of Carbon-Supported Pt-M Alloys

of the alloy composition was still difficult. The other conventional methods for the synthesis of Pt-based alloy electrocatalyst are colloid methods such as sodium borohydride,18 sulfite-complex,25 and alcohol (or polyol) reduction process.26,27 However, the average particle size obtained was usually large, and the size distribution was also broad. Hence, besides the control of the alloy composition uniformly among the whole supported nanoparticles, the control of the particle size in a narrow range (monodispersed state) is very important. Recently, for the permanent magnetic materials, Sun et al.28 and Chen et al.29 succeeded in preparing monodispersed Pt-Fe based alloy nanoparticles via the simultaneous reduction of platinum (II) acethylacetonate; Pt(acac)2 and FeCl2 and/or cobalt acetylacetonate; Co(acac)3 in reversed micelles. The use of such a nanocapsule as the limited reaction space should provide monodispersed alloy particles with uniform composition. Liu et al. examined to apply this method for the anode catalysts in direct methanol fuel cells (DMFCs).30,31 Pt or Pt-Ru alloy particles prepared in nanocapsules were once precipitated by removing organic moieties consisting of the capsules, followed by an addition of carbon black powder to support them. However, we found that these naked particles without any protection by the organic moieties were partially aggregated on the support. It was very difficult to prepare highly dispersed catalysts by such a “post-supporting” method. In the present research, we have developed a modified nanocapsule method to prepare pure Pt and Pt-M (M ) V, Cr, Fe, Co, and Ni) alloy nanoparticles highly dispersed on carbon black (abbreviated as Pt/CB and Pt-M/CB, respectively) for the ORR catalysts in PEFCs. These catalysts are characterized by X-ray diffraction (XRD) and scanning transmission microscopy (STEM) with electron-probe for microanalysis (EPMA). The kinetically controlled ORR activities at these catalysts were also evaluated by a channel flow electrode (CFE) method.15,32-34 We will demonstrate a usefulness of the present preparation method resulting in monodispersed alloy particles with well-controlled composition. 2. Experimental Section 2.1. Preparation of Carbon Supported Pt and Pt-M Nanoparticle Catalysts. Figure 1 illustrates the preparation protocol of Pt and Pt-M (M ) V, Cr, Fe, Co, and Ni) particles highly dispersed on carbon black support based on a nanocapsule method. We referred to a method reported by Sun et al.,28 and modified it for the preparation of Pt-M/CB or Pt/CB electrocatalysts. First, 0.5 mmol metal acetylacetonates, M(acac)x, listed in the third row of Table 1, were dissolved in a mixed solvent of 1,2-hexadecanediol (520 mg) and diphenyl ether (25.0 mL). As the first screening test for various Pt-M alloys, the atomic ratio of M to Pt was adjusted to 50:50 in the starting solution. The reaction flask was equipped with a (25) Ravikumar, M. K.; Shukla, A. K. J. Electrochem. Soc. 1996, 143, 26012606. (26) Li, W.; Zhou, W.; Li, H.; Zhou, Z.; Zhou, B.; Sun, G.; Xin, Q. Electrochim. Acta 2004, 49, 1045-1055. (27) Spinace´, E. V.; Neto, A. O.; Vasconcelos, T. R. R.; Linardi, M. J. Power Sources 2004, 137, 17-23. (28) Sun, S.; Andersw, S.; Thomson, T.; Baglin, J. E. E.; Toney, M. F.; Hamann, H. F.; Murray, C. B.; Terris, B. D. J. Phys. Chem. B 2003, 107, 5419-5425. (29) Chen, M.; Nikles, D. E. Nano Lett. 2002, 2, 211-214. (30) Liu, Z.; Shamsuzzoha, M.; Ada, E. T.; Reichert, W. M.; Nikles, D. E. J. Power Sources 2007, 164, 472-480. (31) Liu, Z.; Ada, E. T.; Shamsuzzoha, M.; Thompson, G. B.; Nikles, D. E. Chem. Mater. 2006, 18, 4946-4951. (32) Wakabayashi, N.; Takeichi, M.; Itagaki, M.; Uchida, H.; Watanabe, M. J. Electroanal. Chem. 2005, 574, 339-346. (33) Yano, H.; Higuchi, E.; Uchida, H.; Watanabe, M. J. Phys. Chem. B 2006, 110, 16544-16549. (34) Yano, H.; Inukai, J.; Uchida, H.; Watanabe, M.; Panakkattu, K. B.; Kobayashi, T.; Chung, J. H.; Oldfield, E.; Wieckowski, A. Phys. Chem. Chem. Phys. 2006, 8, 4932-4939.

Langmuir, Vol. 23, No. 11, 2007 6439

Figure 1. Illustration of the preparation protocol for Pt/CB and Pt-M/CB catalysts. condenser, and all of the reactions were performed in N2 atmosphere with magnetic stirring. The mixture was heated at 110 °C for 20 min. Then, oleic acid (OAC; 170 µL) and oleylamine (OAM; 160 µL) were added, followed by addition of carbon black (300 mg; Ketjen Black EC, specific surface area ) 800 m2/g, Lion Co, Ltd.). The temperature was elevated to 220 °C, and kept constant for 30 min. LiBEt3H (2.0 mL) was added dropwisely into the mixture. After heating at 220 °C for 10 min, the mixture was refluxed at 270 °C for 30 min. As a reference catalyst to the Pt-Fe/CB prepared by the present method, the same combination catalyst was also prepared by an ethylene glycol reduction (EG) method following refs 26 and 27. The mixture of Pt(acac)2, Fe(acac)3, ethylene glycol (as the solvent and reducing agent) and carbon black as the support was refluxed at 200 °C for 3 h under N2 atmosphere. The mixture formed by each method was cooled to room temperature and filtered. The powder thus obtained was dried at 60 °C in a vacuum. To remove organic moieties, the powder was heat-treated at 230 °C for 4 h in a flow of N2. The catalyst powder thus prepared was examined by X-ray diffraction (XRD, Rigaku RINT2000) with Cu KR radiation (50 kV, 300 mA). The microstructure was observed by scanning transmission electron microscopy (STEM, Hitachi S-5200 and HD-2300C) with an energy dispersive X-ray analysis (EDX, EDAX Genesis). The loading amount of the catalysts on the CB was quantified from the weight loss by combustion of carbon at 600 °C in air. The alloy compositions were quantitatively analyzed by an inductively coupled plasma mass analyzer (ICP-MS, Seiko SPQ9200) after dissolving the ash remaining after the combustion in hot aqua resia. 2.2. Preparation of Pt/CB and Pt-M/CB-Dispersed Au Electrodes. To make ultrathin and uniform dispersion of Pt/CB or Pt-M/CB on a gold substrate in the CFE cell, we just employed the recipe optimized for commercial Pt/CB.33,34 The preparation conditions are summarized in Table 2. Here, for example, we briefly explain the procedure for the case of Pt (25.2 wt %)/CB in Table 2. The Pt/CB powder (1.18 mg) was ultrasonically dispersed in 2 mL of ethanol (99.5% ethanol, CPt/CB ) 0.59 g/L). The gold substrate (geometric area ) 0.04 cm2) was mirror-finished and treated by an atmospheric plasma surface treater (ST-7000, Keyence Co. Ltd., Japan, work distance ) 10 mm for 10 s) to increase hydrophilicity of the surface just before coating the catalyst. A 0.5 µL aliquot of

6440 Langmuir, Vol. 23, No. 11, 2007

Yano et al.

Table 1. Typical Properties of Synthesized Pt/CB and Pt-M/CB Catalystsa metal acetylacetonateb M(acac)x

no.

catalysts

1 2

Pt/CB Pt-V/CB

3

Pt-Cr/CB

4

Pt-Fe/CB

5

Pt-Co/CB

6

Pt-Ni/CB

initial metal loadingc (wt %) Pt-M

analyzed metal loadedd (wt %) Pt-M

analyzed compositione (at. %) Pt:M

dXRDf (nm)

dSTEMg (nm)

lattice constanth (Å)

24.5 28.9

25.2 26.4

53:47

2.2 2.5

2.3 2.5

3.93 ( 0.01 3.90 ( 0.01

28.0

28.4

49:51

2.3

2.5

3.90 ( 0.02

29.4

25.6

50:50

1.7

2.1

3.85 ( 0.02

29.6

27.0

46:54

2.0

2.0

3.85 ( 0.02

29.6

31.5

55:45

2.3

2.5

3.78 ( 0.01

Pt(acac)2 Pt(acac)2 V(acac)3 Pt(acac)2 Cr(acac)3 Pt(acac)2 Fe(acac)3 Pt(acac)2 Co(acac)3 Pt(acac)2 Ni(acac)2

a Projected composition (at. %) of Pt:M was 50:50. Analyzed values were those for the samples before the electrochemcial experiment. b Dissolved in a mixed solvent of 1,2-hexadecanediol and diphenyl ether for preparation of Pt and Pt-M. c Calculated from the amount of 0.5 mmol M(acac)x and CB (300 mg). d Pt or Pt-M weight percent in Pt/CB or Pt-M/CB catalysts estimated by weight loss using thermogravimetry (TG). e The composition of Pt-M alloy was analyzed by ICP after dissolving the ash remaining after TG. f Average crystallite size calculated from Scherrer’s equation. g Average particle size based on the STEM observation. h Average lattice constants and standard deviations calculated from (111), (200), (110), and (200) diffraction peak in Figure 5.

Table 2. Preparation Conditions of Suspensions and Resulting wPt-M and wCB by Pipetting the Suspension with 12.5 µL/cm2 on Au Substrate no.

catalysts

CPt-M/CBa (g/L)

wPt-Mb (µg/cm2)

wPtc (µg/cm2)

wCBd (µg/cm2)

1 2 3 4 5 6

Pt/CB Pt-V/CB Pt-Cr/CB Pt-Fe/CB Pt-Co/CB Pt-Ni/CB

0.59 0.60 0.61 0.59 0.60 0.64

1.85 1.97 2.18 1.89 2.03 2.53

1.85 1.60 1.71 1.47 1.50 2.03

5.50 5.50 5.50 5.50 5.50 5.50

a Amount of Pt/CB and Pt-M/CB catalysts in the suspension. Amount of Pt-M attached on the Au substrate. c Amount of Pt attached on the Au substrate. d Amount of CB attached on the Au substrate.

b

the suspension was pipetted onto the gold substrate (12.5µL/cm2), giving wPt ) 1.85 µg/cm2 and wCB ) 5.50 µg/cm2. This was dried at room temperature in a glass Petri dish containing 99.5% ethanol. The dish was covered by a lid with a small gap to evaporate the solvent droplet slowly (for about 30 min) under nearly saturated vapor pressure of ethanol. A 0.5 µL aliquot of a 0.2 wt % Nafion [diluted with ethanol and water (3:2, volume ratio)] solution was placed on top of the dried catalyst layer to yield the film thickness of 0.1 µm. This was dried at room temperature under the similar manner to that described above. Finally, the Nafion-coated catalyst layer on the Au was heat-treated at 130 °C for 30 min in air. Assuming a uniform dispersion of each catalyst in the layer, the thickness at wCB ) 5.50 µg/cm2 was calculated to be 28 nm, corresponding to almost the monolayer height of the CB particles. Because the CB particles form agglomerate as seen in STEM images of Figure 2, an excess amount of the CB should result in nonuniform stacked layer. Thus, our recipe is very important to provide a favorable condition for the mass-transportation of the reactants (O2 and H+) to the supported Pt catalyst surfaces.35 2.3. Electrochemical Measurements. Details of the experimental setup of the CFE cell and the flow circuit of the electrolyte solution are described in our previous papers.15,32,33 The CFE cell was made of Kel-F blocks and a Teflon sheet. The substrate for working electrode was mirror-finished gold plates (flow direction length 0.1 cm × width 0.4 cm, geometric area ) 0.04 cm2). A platinum mesh and reversible hydrogen electrode (RHE) were used as the counter and reference electrodes, respectively. All electrode potentials were controlled by a potentiostat (ALS 700A, BAS, Inc.) with respect to the RHE. The electrolyte solution of 0.1 M HClO4 was prepared from reagent grade chemicals (Kanto Chemical Co., Japan) and (35) Higuchi, E.; Uchida, H.; Watanabe, M. J. Electroanal. Chem. 2005, 583, 69-76.

Milli-Q water (Milli-pore Japan Co., Ltd.) and purified in advance by conventional pre-electrolysis methods.36,37 Prior to the ORR experiments, the working disk electrode surface was cleaned and electrochemically stabilized by potential cycling between 0.05 and 1.0 V at 0.5 V/s in 0.1 M HClO4 solution deaerated with N2 gas until a steady-state cyclic voltammogram (CV) was obtained. The electrochemical active area of Pt, SPt, was evaluated from an electric charge for the hydrogen desorption QH in the positivegoing potential scan from 0.05 to 1.00 V in cyclic voltammetry at the sweep rate of 0.1 V/s at 30 °C. After bubbling O2 in the same electrolyte solution for 30 min, hydrodynamic voltammograms for the ORR at the working disk electrode were recorded by sweeping the potential from 0.2 to 1.0 V at 0.5 mV/s under the mean flow rate of electrolyte (Um) from 10 to 50 cm/s. The kinetically controlled current (Ik) was determined from the hydrodynamic voltammograms in the CFE by using the following equation:38 1/I ) 1/Ik + 1/IL ) 1/{1.165 × nF[O2]w(UmD2x12/h)1/3} (1) where n is the number of electrons transferred, F is the Faraday constant, [O2] is the O2 concentrations in the bulk of electrolyte solution, w is the width of the working electrode, Um is the mean flow rate of the electrolyte solution, D is the diffusion coefficient of O2, x1 is the length of the working electrode in the electrolyte flow direction, and h is the half channel height (or half the thickness of the electrolyte flow over the electrodes). The kinetically controlled current density (jk) was calculated by dividing Ik with the SPt. All of the electrochemical experiments were performed at 30 °C.

3. Results and Discussion 3.1. Characterization of Pt/CB with Various Loading Levels. It is desirable for the catalyst preparation method to enable us to control the loading level while keeping the highly dispersed state. First, we prepared Pt/CB catalysts with various Pt loading levels to know the advantages of the nanocapsule method. By the XRD of the Pt/CB catalysts, the formation of Pt particles with face-centered cubic (fcc) structure was confirmed, i.e., the lattice constant of 3.93 ( 0.01 Å (Table 1) agrees well with the crystallographic data. Figure 2 shows typical STEM images of the Pt/CB catalysts with various Pt loadings of 10.1 (36) Watanabe, M.; Motoo, S. J. Electroanal. Chem. 1970, 60, 259-266. (37) Uchida, H.; Ikeda, N.; Watanabe, M. J. Electroanal. Chem. 1997, 424, 5-12. (38) Levich, V. G. In Physicochemical Hydrodynamics; Prentice Hall: Englewood Cliffs, NJ, 1962; p 112.

ActiVity of Carbon-Supported Pt-M Alloys

Langmuir, Vol. 23, No. 11, 2007 6441

Figure 2. STEM images (left row: low magnification, second row: high magnification) and particle size distribution histograms of Pt/CB powders with various Pt loading level. (A) 10.1 wt %, (B)17.5 wt %, (C) 25.2 wt %, (D) 33.0 wt %, (E) 42.6 wt %, (F) 55.0 wt %.

6442 Langmuir, Vol. 23, No. 11, 2007

Yano et al.

Figure 3. Changes in the average particle sizes determined by STEM, dSTEM, (4) and the crystallite sizes determined by XRD, dXRD, (b) as a function of Pt loading level on Pt/CB catalysts.

to 55.0 wt % together with the particle size distribution histograms among more than 500 particles for each sample. It was found that Pt nanoparticles were uniformly dispersed on the CB support and their size distribution was fairly narrow, i.e., 2 to 4 nm, irrespective of the loading level from 10.1 to 55.0 wt %. The average crystallite size was also calculated from Scherrer’s equation for the XRD peak assigned to Pt(220). Figure 3 shows plots of the average values of particle size based on STEM (dSTEM) or crystallite size based on XRD (dXRD) as a function of the Pt loading level. Both d values are in fairly good agreement. The Pt particle size increases only slightly with increasing the Pt loading level, and reaches almost the constant value of ca. 2.6 nm beyond 33.0 wt % loading. Such an excellent size-controlling is one of the great advantages in the present nanocapsule method. Because the metal catalyst particles were formed by the reduction of metal precursor(s) at the limited space in each nanocapsule, the particle size could be settled into such a narrow size distribution. Recently, Liu et al. examined to apply Sun’s method28 for the anode catalysts in DMFCs.30,31 Pure Pt or Pt-Ru alloy particles were prepared in nanocapsules with diphenyl ether as the solvent. Just following to the Liu’s recipe, the organic moieties adsorbed on the nanoparticles were first removed in hexane and then carbon black powder was added to support the “naked” particles without any protection by the organic moieties. However, such naked particles were partially aggregated on the support. We have checked the superiority of the present method to Liu’s one.30,31 Figure 4 shows STEM images of 20.8 wt % and 18.0 wt % Pt/CB catalysts prepared by the present method and Liu’s (post-supporting)30 one. It is clear in Figure 4B and 4C that Pt particles are aggregated when prepared by Liu’s post-supporting method. In contrast, Pt particles are quite uniformly dispersed over the CB supports for our present catalyst (Figure 4A). This is because the aggregation of Pt particles could be suppressed by a repulsive action in the organic moieties adsorbed on the surface as imagined in Figure 1. Based on the thermogravimetry, we confirmed these organic protectors (no longer in use for the electrocatalyst) to be removed by heat treatment at 230 °C for 4 h in a flow of N2. Appearance of high ORR activity by the method will be shown in the later section. Recently, we found that the apparent ORR rate constants (per real Pt active surface area) at bulk-Pt (with and without Nafioncoating) and Nafion-coated commercial Pt/CB (19.3 and 46.7 wt % Pt, d ) 2.6 to 2.7 nm) thin-film electrodes were in good agreement with each other in the operation conditions of PEFCs, i.e., 30-110 °C and ca. 0.7 to 0.8 V vs RHE.33 Therefore, we can expect that the use of highly loaded Pt/CB prepared by the present method can provide an advantage of reducing the catalyst layer thickness without losing the kinetic activity, if the catalyst layer is properly designed to achieve a good balance between the catalyst utilization and the gas diffusivity.

Figure 4. STEM images of 20.8 and 18.0 wt % Pt/CB catalysts prepared by (A) the present method and (B and C) by Liu’s postsupporting method,31 respectively. Image C is the zoomed up one of the dotted square part in image B.

Figure 5. X-ray diffraction patterns of (A) Pt-V/CB, (B) Pt-Cr/ CB, (C) Pt-Fe/CB, (D) Pt-Co/CB, (E) Pt-Ni/CB, and (F) Pt/CB powder.

3.2. Characterization of Pt-M/CB. Figure 5 shows XRD patterns of the Pt/CB, Pt-V/CB, Pt-Cr/CB, Pt-Fe/CB, PtCo/CB, and Pt-Ni/CB powders with the total metal loading of 25.2-31.5 wt % (in Table 1). The broad peak at 2θ ) ∼25° for

ActiVity of Carbon-Supported Pt-M Alloys

Langmuir, Vol. 23, No. 11, 2007 6443

Figure 6. STEM images (left row: low magnification, second row: high magnification) and particle size distribution histograms of (A) Pt-V/CB, (B) Pt-Cr/CB, (C) Pt-Fe/CB, (D) Pt-Co/CB, and (E) Pt-Ni/CB powders.

all of the samples is assigned to CB. The diffraction peaks for Pt/CB are well assigned to the fcc phase of Pt. The XRD patterns for the Pt-M/CB resemble those of Pt/CB, but the peak positions observed shift to a higher angle than the latter. For example, although the value of 2θ assigned to Pt(111) at Pt/CB was 39.8°, the corresponding values for Pt-V/CB, Pt-Cr/CB, Pt-Fe/CB, Pt-Co/CB, and Pt-Ni/CB were 40.0°, 40.0°, 40.6°, 40.7°, and 41.6°, respectively. No extra peaks assigned to V, Cr, Fe, Co, and Ni (or their oxides) as well as the Pt-M ordered alloys39 were identified. The average lattice constants of these catalysts calculated from (111), (200), (220), and (311) are summarized in Table 1. The changes in the lattice parameters by alloying with metal M were found to be quite isotropic as shown in Table S1 in the Supporting Information. Because the crystal structure of both Ni and Pt is fcc, the change in the lattice constant of the solid solution should follow the Vegard’s law (linear relationship between the composition and the lattice constant). Indeed, the (39) Watanabe, M.; Tsurumi, K.; Mizukami, T.; Nakamura, T.; Stonehart, P. J. Electrochem. Soc. 1994, 141, 2659-2668.

value for Pt-Ni/CB just accords with that expected for the alloy composition. However, the crystal structure of V, Cr, and Fe in low temperature form is body centered cubic (bcc), whereas Co exhibits a hexagonal closest packing (hcp) structure. In such Pt-M alloys, the change in the lattice constant was smaller than that expected from simple Vegard’s law.1 The change in the lattice parameters of Pt-V, Pt-Cr, Pt-Fe, and Pt-Co/CB showed qualitatively similar behavior to that in ref 1. These results indicate the formation of a corresponding solid solution with fcc structure. The crystallite sizes dXRD of Pt-V, Pt-Cr, Pt-Fe, Pt-Co, and Pt-Ni, calculated in the same manner as described above, were 2.5, 2.3, 1.7, 2.0, and 2.3 nm, respectively (see Table 1). Figure 6 shows STEM images of Pt-M/CB (M ) V, Cr, Fe, Co, and Ni) together with the particle size distribution histograms. Pt-M alloy nanoparticles were uniformly dispersed on the CB support, and their size distributions were fairly narrow similar to that of Pt/CB. The average particle sizes dSTEM of Pt-M alloys are consistent with those of dXRD,

6444 Langmuir, Vol. 23, No. 11, 2007

Yano et al. Table 3. EDX Spot Analysis of Compositions of Pt-Fe Particles Supported on CB, Prepared by Nanocapsule Method and EG Methoda composition (at. %) nanocapsule (present method)

EG (conventional method)

no.

Pt

Fe

Pt

Fe

1 2 3 4 5 6 7 8 9 10 average standard deviation

41.89 50.40 41.08 46.91 40.97 37.61 42.89 38.69 47.92 45.76 43.41 4.17

58.11 49.60 58.92 53.09 59.03 62.39 57.11 61.31 52.08 54.24 56.59 4.17

40.59 38.47 73.17 77.57 74.00 71.96 39.65 40.58 92.32 45.59 59.39 20.28

59.41 61.53 26.83 22.43 26.00 28.04 60.35 59.42 7.68 54.41 40.61 20.28

a Analyzed values were those for the samples before the electrochemcial experiment.

Figure 7. TEM images (high angle annular dark field) of Pt-Fe/ CB powders synthesized by the present nanocapsule method (A) and EG reduction method (B). The numbers shown in the figures indicate the particles performed spot-analysis with EDX.

as shown in Table 1. It was found that the analyzed values of the loading level and the average composition of all of the alloys agreed well with the starting (projected) values used for the preparation, i.e., metal loading level from 24.5 to 29.6 wt % and the composition of 50:50 at. % used as the metal precursor. Figure 7 shows the TEM images (high angle annular dark field) of Pt-Fe/C catalysts synthesized by the nanocapsule method (present method) and the EG method (conventional method). The composition of individual Pt-Fe particles was also analyzed by EDX at 10 particles randomly selected from each catalyst. Table 3 summarizes the atomic ratio of Pt:Fe at the each particle from nos. 1 to 10. For the catalysts prepared by the conventional EG method, the average composition was Pt59Fe41 with standard deviation of more than 20 at. %; that is, the alloy composition changed significantly from particle to particle. In contrast, the alloy catalyst prepared by the nanocapsule method exhibited the average composition of Pt43Fe57 with the standard deviation of only 4 atom %, which was close to the projected composition of Pt50Fe50. Thus, the composition of each alloy particle was well-controlled. Furthermore, the particle size distribution was fairly narrow compared to that of the EG method in the same scale images. 3.3. ORR Activities of Nafion-Coated Pt-M/CB. Recently, we demonstrated that non-precious-metal elements in Pt-Fe, Pt-Co, and Pt-Ni alloys were leached out in an acidic solution, but a Pt skin layer with a modified electronic structure was formed

Figure 8. Cyclic voltammograms on Nafion-coated Pt/CB and PtM/CB catalysts in 0.1 M HClO4 solution purged with N2 at 30 °C. Sweep rate ) 0.1 V/s. See Figure S1 in the Supporting Information, which includes all of the CV data in color.

on the alloy surface.1,2,40-42 Prior to the ORR measurements, all of the Nafion coated Pt-M/CB electrodes were electrochemically stabilized by repetitive potential sweeps (typically 20 times) between 0.05 and 1.00 V vs RHE in N2-purged 0.1 M HClO4 solution at 30 °C. Figure 8 shows typical CVs at Nafion-coated Pt-M/CB and Pt/CB in deaerated 0.1 M HClO4 solution at 30 °C. The other catalysts showed a similar behavior (see Figure S1 in the Supporting Information). Since the stabilized CVs at Pt-M/CB resembled that of a polycrystalline platinum or Pt/CB electrode, the electrochemically active surface area (SPt) was evaluated from the electric charge of the hydrogen desorption wave ∆QH in each CV, supposing ∆QH° ) 210 µC/cm2 for a smooth polycrystalline Pt.36,43 The SPt of all of the electrodes were calculated to be 0.10 (Pt/CB), 0.11 (Pt-V/CB), 0.12 (PtCr/CB), 0.12 (Pt-Fe/CB), 0.12 (Pt-Co/CB), and 0.11 cm2 (PtNi/CB), which correspond to the specific surface area (SA) of 136, 134, 136, 160, 153, and 116 m2/g, respectively. The value of SA thus obtained for the Pt/CB (dSTEM ) 2.3 nm) is in accord with that calculated for spherical Pt particles of 2.3 nm. Figure 9A shows the hydrodynamic voltammograms for the ORR at the Nafion-Pt-Cr/CB working electrode in O2-saturated (40) Wan, L-J.; Moriyama, T.; Ito, M.; Uchida, H.; Watanebe, M. Chem. Comm. 2002, 58-59. (41) Uchida, H.; Ozuka, H.; Watanabe, M. Electrochim. Acta 2002, 47, 36293636. (42) Wakisaka, M.; Mitsui, S.: Hirose, Y.; Kawashima, K.: Uchida, H.; Watanabe, M. J. Phys. Chem. B 2006, 110, 23489-23496. (43) Watanabe, M.; Motoo, S. J. Electroanal. Chem. 1975, 60, 275-283.

ActiVity of Carbon-Supported Pt-M Alloys

Langmuir, Vol. 23, No. 11, 2007 6445 Table 4. Kinetically Controlled Current Density (jk) at 0.80 V on the Various Electrode Catalysts Pt/ CB 0.82 jk (mA/cm2) jk(Pt-M/CB)/jk(Pt/CB) 1.0

Figure 9. Hydrodynamic voltammograms for the ORR (A) at Nafion-coated Pt-Cr/CB working electrode at a mean flow rates of 10, 15, 23, 30 and 50 cm/s, (B) at the Pt/CB and Pt-M /CB working electrode at 50 cm/s, in O2-saturated 0.1 M HClO4 solution at 30 °C. Sweep rate ) 0.5 mV/s. See Figure S2 (color version of Figure 9B) in the Supporting Information.

Pt-V/ Pt-Cr/ Pt-Fe/ Pt-Co/ Pt-Ni/ CB CB CB CB CB 1.35 1.6

1.45 1.8

1.10 1.3

1.32 1.6

1.03 1.3

per SPt (denoted as jk) at 0.80 V for Pt/CB and Pt-M/CB are shown in Table 4. The values of jk at alloys were 1.3 to 1.8 times higher than that of the Pt/CB, and increased in the order Pt/CB < Pt-Ni/CB < Pt-Fe/CB < Pt-Co/CB < Pt-V/CB < PtCr/CB. The activities at Pt-Ni/CB, Pt-Co/CB, and Pt-Fe/CB, however, seemed to be lower than those evaluated for bulk sputtered alloys.15 We consider that this is not ascribed to socalled “particle size effect”. Recently, we performed electrochemical ORR and 195Pt EC-NMR measurements for Pt/CB catalysts with different particle sizes dPt ) 1.6, 2.6, and 4.8 nm together with bulk-Pt film.34 It was found that the ORR rate constants and H2O2 yields evaluated from hydrodynamic voltammograms did not show any particle size dependency at the potential range of 0.70-0.80 V in the temperature range of 30110 °C. Therefore, the major reason for the discrepancy between the ORR activities at Pt-M/CB and the bulk-alloys might be the change in the surface state (thickness or roughness) of the Pt skin layer, because the corrosion level of the second non-preciousmetal component certainly depends on the crystallite size. This also suggests that the optimum compositions before the stabilization at the nanoparticle alloys could be changed from that at bulk ones. Because the alloy composition of the Pt-M/CB is easily controlled by the present method, we are performing further experiments together with the evaluation of the activity and H2O2 yield at elevated temperature by using a channel flow double electrode (CFDE) method.15,32-34

4. Conclusion Figure 10. I-1 vs Um-1/3 plots obtained from hydrodynamic voltammograms for the ORR at (b) 0.80, (2) 0.76, and (9) 0.70 V vs RHE on Pt-Cr/CB electrode in O2-saturated 0.1 M HClO4 solution at 30 °C.

0.1 M HClO4 solution at 30 °C under the various flow rate conditions. The ORR currents commenced at ca. 0.98 V and reach to diffusion limits around 0.5 V. The hydrodynamic voltammograms for the ORR at Pt-M/CB alloy catalysts are shown in Figure 9B [see Figure S2 (color version of Figure 9B) in the Supporting Information], in comparison with that of Pt/ CB at a mean solution flow rate of 50 cm/s. The ORR currents at the alloy catalysts commence to increase at more positive potential than that at the Pt/CB by ca. 40 mV, indicating the enhanced ORR activity at these alloys. Plots of I-1 vs Um-1/3 for the ORR on the Nafion-Pt-Cr/CB catalyst are shown in Figure 10. Linear relationships with a constant slope are obtained at all of the potentials of 0.80, 0.76, and 0.70 V. The other catalysts showed a similar behavior (Figure S3 in the Supporting Information). By extrapolating Um-1/3 to 0 (infinite mass transfer rate), the value of kinetically controlled current Ik was calculated. The kinetically controlled current density

We have succeeded in preparing Pt and Pt-M (M ) V, Cr, Fe, Co, and Ni) alloy nanoparticles highly dispersed on carbon black support by the simultaneous reduction of Pt(acac)2 and M(acac)x in organic nanocapsule. Two great advantages of the present preparation method were evidently demonstrated, i.e., good control of both the particle size (monodispersion) and uniform alloy composition, which are essential characteristics for the reaction(s) within the limited nanospace. We also found that the electrocatalytic activities for the ORR at the Pt-V/CB, Pt-Cr/CB, Pt-Fe/CB, Pt-Co/CB, and Pt-Ni/CB catalysts were higher than that of Pt/CB catalyst. The present method, of course, can be applied to prepare various kinds of alloys (binary, ternary, or more), which can contribute to explore highly active new electrocatalysts for PEFCs. Acknowledgment. This work was supported by the fund for “Leading Project” of Ministry of Education, Science, Culture, Sports and Technology of Japan. Supporting Information Available: All of the lattice constants calculated and electrochemical data of all the alloys. This material is available free of charge via the Internet at http://pubs.acs.org. LA070078U