pubs.acs.org/Langmuir © 2010 American Chemical Society
Pt-Ru/CeO2/Carbon Nanotube Nanocomposites: An Efficient Electrocatalyst for Direct Methanol Fuel Cells Zhenyu Sun, Xiang Wang, Zhimin Liu,* Hongye Zhang, Ping Yu, and Lanqun Mao* Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, P. R. China Received March 16, 2010. Revised Manuscript Received May 6, 2010 Pt-Ru/CeO2/multiwalled carbon nanotube (MWNT) electrocatalysts were prepared using a rapid sonicationfacilitated deposition method and were characterized by X-ray diffraction (XRD), X-ray photoemission spectroscopy (XPS), transmission electron microscopy (TEM), energy-dispersive spectroscopy (EDS), and voltammetry. Morphological characterization by TEM revealed that CeO2 nanoparticles (NPs) were in intimate contact with Pt-Ru NPs, and both were highly dispersed on the exteriors of nanotubes with a small size and a very narrow size distribution. Compared with the Pt-Ru/MWNT and Pt/MWNT electrocatalysts, the as-prepared Pt-Ru/CeO2/MWNT exhibited a significantly improved electrochemically active surface area (ECSA) and a remarkably enhanced activity toward methanol oxidation. The effects of the Pt-Ru loading and the Pt-to-Ru molar ratio on the electrocatalytic activity of Pt-Ru/ CeO2/MWNT for methanol oxidation were investigated. We found that a maximum activity toward methanol oxidation reached at the 10 wt % of Pt-Ru loading and 1:1 of Pt-to-Ru ratio. Moreover, the role of CeO2 in the catalysts for the enhancement of methanol oxidation was discussed in terms of both bifunctional mechanism and electronic effects.
1. Introduction Direct methanol fuel cells (DMFCs) show a great potential as power sources for vehicles and portable electronic devices due to their high power density, low operating temperature, and the ease in handling using a liquid fuel.1 Metallic platinum (Pt) has always been considered as an effective anode electrocatalyst for methanol oxidation.2 However, Pt is susceptible to CO poisoning during the oxidation of methanol, resulting in a substantial decrease in its activity. Moreover, a high loading of this expensive noble metal is usually required in the anode, making the cost of DMFCs very high. As a consequence, extensive efforts have been directed toward improving the CO tolerance and reducing Pt loading while maintaining high mass current density. To this end, binary or ternary Pt-based metallic/metal oxide catalysts including PtRu,3 PtPd,4 PtRuIr,5 and PtRuTiO26 have been developed, among which bimetallic PtRu, in particular, has been shown to behave as a robust catalyst for methanol oxidation.7 The role of the second metal is widely accepted as a promoter for the oxidation of CO strongly adsorbed on Pt by oxygen-containing species formed on the neighboring second metal particles, for example, Ru (Ru-OHad).8 Besides this bifunctional mechanism, an electronic or ligand effect may be relevant as well; i.e., the
electronic nature of the surface atoms is modified, and thus the binding strength of Pt-CO is weakened.9 Compared with smooth Pt and PtRu bulks, carbon-supported Pt and PtRu nanoparticles (NPs) show higher current efficiency for methanol oxidation due to their larger surface areas whereby the soluble intermediates formed are more readily adsorbed and oxidized to CO2 before they are transported away from the electrode surface by the continuous flow of electrolyte.3 In recent years, carbon nanotubes (CNTs) have attracted special attention as catalyst supports in fuel cell applications owing to their unique properties, such as large surface areas, high chemical resistivity, and superior mechanical strength. In addition, the hairlike structure of CNTs allows for entanglements among nanotubes, enabling better electron conductivity than traditional carbon systems.10 From this context, numerous studies have reported higher electrocatalytic activities of Pt-based NPs supported on CNTs as compared with those deposited on high-surface-area carbon.5,10-18 However, most of the established protocols to prepare carbon nanotube-based electrocatalysts involve covalent functionalization of CNTs,10-12,15,18 leading to the severe disruption of nanotubes’ intrinsic electronic properties and thus increasing application expenditures. Other strategies consist of an adsorption of small organic coupling molecules,19 polyelectrolytes,13,14 polymers,16
*Corresponding authors. E-mail:
[email protected] (Z.L.); lqmao@iccas. ac.cn (L.M.).
(10) Hsin, Y. L.; Hwang, K. C.; Yeh, C. J. Am. Chem. Soc. 2007, 129, 9999– 10010. (11) Lin, Y. H.; Cui, X. L.; Yen, C.; Wai, C. M. J. Phys. Chem. B 2005, 109, 14410–14415. (12) Chetty, R.; Xia, W.; Kundu, S.; Bron, M.; Reinecke, T.; Schuhmann, W.; Muhler, M. Langmuir 2009, 25, 3853–3860. (13) Okamoto, M.; Fujigaya, T.; Nakashima, N. Small 2009, 5, 735–740. (14) Wu, B. H.; Hu, D.; Kuang, Y. J.; Liu, B.; Zhang, X. H.; Chen, J. H. Angew. Chem., Int. Ed. 2009, 48, 4751–4754. (15) Xing, Y. C. J. Phys. Chem. B 2004, 108, 19255–19259. (16) Wang, S. Y.; Jiang, S. P.; White, T. J.; Guo, J.; Wang, X. J. Phys. Chem. C 2009, 113, 18935–18945. (17) Huang, T.; Jiang, R. R.; Liu, J. L.; Zhuang, J. H.; Cai, W. B.; Yu, A. S. Electrochim. Acta 2009, 54, 4436–4440. (18) Zhou, C. M.; Wang, H. J.; Peng, F.; Liang, J. H.; Yu, H.; Yang, J. Langmuir 2009, 25, 7711–7717. (19) Yang, D. Q.; Hennequin, B.; Sacher, E. Chem. Mater. 2006, 18, 5033–5038.
(1) Scott, K.; Shukla, A. K. In Modern Aspects of Electrochemistry; White, R. E., Ed.; Springer: Berlin, 2007; Vol. 40, p 127. (2) Wilson, M. S.; Gottesfeld, S. J. Appl. Electrochem. 1992, 22, 1–7. (3) Wang, H. S.; Alden, L. R.; DiSalvo, F. J.; Abru~na, H. D. Langmuir 2009, 25, 7725–7735. (4) Wang, H.; Xu, C. W.; Cheng, F. L.; Zhang, M.; Wang, S. Y.; Jiang, S. P. Electrochem. Commun. 2008, 10, 1575–1578. (5) Liao, S. J.; Holmes, K.; Tsaprailis, H.; Birss, V. I. J. Am. Chem. Soc. 2006, 128, 3504–3505. (6) Tian, J. A.; Sun, G. Q.; Jiang, L. H.; Yan, S. Y.; Mao, Q.; Xin, Q. Electrochem. Commun. 2007, 9, 563–568. (7) Petrii, O. A. J. Solid State Electrochem. 2008, 12, 609–642. (8) Du, B. C.; Tong, Y. Y. J. Phys. Chem. B 2005, 109, 17775–17780. (9) Tong, Y. Y.; Kim, H. S.; Babu, P. K.; Waszczuk, P.; Wieckowski, A.; Oldfield, E. J. Am. Chem. Soc. 2002, 124, 468–473.
Langmuir 2010, 26(14), 12383–12389
Published on Web 05/20/2010
DOI: 10.1021/la101060s
12383
Article
or surfactants20 to enable strong affinity of metal NPs on nanotubes. Nevertheless, such noncovalent modification introduces an issue of concern that the molecules absorbed on the surfaces of NPs may adversely affect their catalytic efficiencies. Moreover, electrodeposition of metal particles on nanotubes is often tedious and struggles to achieve quantitative control of the metal coverage.21 In the case of impregnation method, it is experimentally difficult to control the NP size and attain high NP loading by the colloid scheme. The supercritical-fluid-depositon method always requires high operation temperature and pressure, inhibiting its practical extension although it affords to coat nanotubes with a variety of metals and metal oxides.11,22 Unfortunately, previous protocols available for synthesizing CNT hybrid catalysts suffer from either the sparse deposition of metal particles or the formation of large-sized aggregates on nanotube surfaces. Another important consideration is the enhancement of catalytic activity for methanol oxidation coupling with the decrease of catalyst cost which still presents an ongoing challenge in the development of practical DMFCs. Herein we developed a kind of novel and efficient electrocatalyst, Pt-Ru/CeO2/MWNT, for methanol oxidation. As one part of significance in this work, we presented a facile, rapid, and straightforward process to attach ultrafine CeO2 and subsequently Pt-Ru NPs onto the external surfaces of MWNTs with the aid of sonication. In this process no stabilizers (i.e., organic species, surfactants) for NPs and harsh pretreatments for MWNTs are required, which makes the synthetic procedures very simple. The chemical, morphological, and crystallographic properties of the obtained catalysts were studied in detail by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), and energy-dispersive spectroscopy (EDS) performed in the scanning TEM mode. The catalytic activity of Pt-Ru/CeO2/MWNT toward methanol electrooxidation was investigated, and comparisons were made with Pt-Ru/MWNT and Pt/MWNT catalysts. The effects of the Pt-Ru loading as well as the Pt-to-Ru molar ratio in the catalysts were studied in an attempt to maximize the electrocatalytic activity for methanol oxidation. The role of CeO2 in the catalysts for the oxidation of methanol also was discussed.
2. Experimental Section 2.1. Materials. All chemicals used in this work were of analytical grade and used as supplied. High-purity (95%) multiwalled carbon nanotubes (MWNTs) were purchased from Shenzhen Nanotech Port Co., Ltd., and used without further treatment. MWNTs were prepared by the catalytic decomposition of CH4 using La2NiO4 as a catalyst precursor. To remove the catalyst, CNTs were purified via being dispersed in 1 M HNO3 solution for 6 h, followed by filtering and washing with distilled water several times. IR analysis for the treated CNTs showed the absence of detectable peaks assigning to carboxylic or hydroxyl groups, suggesting that the pretreatment with dilute nitric acid had little influence on the intrinsic structure of CNTs. TEM characterization reveals that the outer diameters and lengths of MWNTs were 40-60 nm and 1-12 μm, respectively. 2.2. Catalyst Synthesis. In a typical experiment to prepare PtRu/CeO2/MWNT, 5 mg of MWNTs was initially dispersed in 20 mL of a Ce(NO3)3 3 6H2O ethanol solution by 2 min tip sonication (Vibra Cell CVX, 500 W, 20 kHz, 20% amplitude), and KOH (KOH/Ce(NO3)3 3 6H2O molar ratio is 3:1) dissolved in a ethanol aqueous solution (90% v/v) was dropped slowly into the
Sun et al.
Figure 1. XRD pattern for the Pt-Ru/CeO2/MWNT catalyst with the Pt-Ru loading at 20 wt %.
dispersion within 2 min under tip sonication. Subsequently, 1 mL of the precursor (H2PtCl6 3 6H2O þ RuCl3 3 3H2O) dissolved in ethanol at a certain concentration was dropped into the above suspension, followed by the dropwise addition of 1 mL of a NaBH4 ethanol solution (its concentration was 4 times higher than that of the precursor) in a similar manner under tip sonication. All synthetic processes were conducted at 308 K. The obtained mixture was ultracentrifuged, and the collected precipitate was first washed repeatedly with absolute ethanol and distilled water and then vacuum-dried at 60 °C for 6 h. The loading of CeO2 onto MWNTs in all cases was kept constant at 25 wt % of MWNT. The loadings of Pt-Ru on CeO2/MWNTs were tuned in the range from 5 to 40 wt % by changing the concentration of the feeding precursor. For comparison, Pt-Ru/ MWNT and Pt/MWNT were also synthesized via similar procedures without the addition of Ce(NO3)3 3 6H2O. 2.3. Characterization. XRD was operated on a D/MAXRC diffractometer at 30 kV and 100 mA with Cu KR radiation. XPS measurements were carried out on an ESCAL Lab 220i-XL specrometer at a pressure of ∼3 10-9 mbar (1 mbar = 100 Pa) using Al KR as the excitation source (hν = 1486.6 eV) and operated at 15 kV and 20 mA. TEM observations were performed on a transmission electron microscope (JEOL JEM-2010) at 200 kV. High-resolution TEM measurements were made on a transmission electron microscope (Philips Tecnai F30 FEG) operated at 300 kV equipped with an energy-dispersive X-ray spectrometer (EDS) and a Gatan annular dark-field (ADF) detector. 2.4. Electrochemical Study. Electrochemical measurements were carried out with a computer-controlled electrochemical analyzer (CHI 600A CH Instruments) in a conventional twocompartment cell. A glassy carbon (GC) (3 mm in diameter) was used as working electrode, a Pt wire as counter electrode, and an Ag/AgCl electrode (KCl-saturated) as reference electrode. Catalyst powders dispersed in ethanol were subjected to bath sonication to obtain homogeneous colloidal dispersions, and 3 μL of each kind of dispersion was pipetted onto GC electrode. Before the surface modification, the GC electrodes were polished with 0.3 and 0.05 μm alunima slurries, washed with distilled water and acetone, followed by 1 min ultrasonic agitation in distilled water, and then dried by N2 blowing. The coated GC electrode was dried by a lamp irradiation. The electrodes were scanned for five cycles, and the last cycle was used for comparing the electrocatalytic activity of the catalysts prepared in this study. Chronoamperometry was conducted in 0.5 M H2SO4 and 1 M methanol solution at a polarized potential of 0.5 V over a time scale of 1000 s. All electrochemical measurements were carried out at ambient temperature. In all cases, the Pt loading on electrode was quantitatively identical in an attempt to minimize the ensemble effect owing to different Pt coverage.8
3. Results and Discussion (20) Lee, C. L.; Ju, Y. C.; Chou, P. T.; Huang, Y. C.; Kuo, L. C.; Oung, J. C. Electrochem. Commun. 2005, 7, 453–458. (21) Wu, G.; Xu, B. Q. J. Power Sources 2007, 174, 148–158. (22) Sun, Z. Y.; Liu, Z. M.; Han, B. X.; Wang, Y.; Du, J. M.; Xie, Z. L.; Han, G. J. Adv. Mater. 2005, 17, 928–932.
12384 DOI: 10.1021/la101060s
3.1. X-ray Diffraction. Figure 1 shows the XRD profile of the Pt-Ru/CeO2/MWNT catalyst with Pt-Ru loading at 20 wt %. The strongest diffraction peak at 26.5° and peaks with weak intensity at 43.16°, 44.6°, 54.2°, and 77.7° were attributed to Langmuir 2010, 26(14), 12383–12389
Sun et al.
Article
Figure 2. Wide survey XPS spectrum (a) and core-level XPS patterns for Ce 3d (b), Pt 4f (c), and Ru 3d (d) regions of the Pt-Ru/CeO2/ MWNT catalyst with the Pt-Ru loading at 20 wt %.
the (002), (100), (101), (004), and (110) reflections of the graphitic structure of MWNTs, respectively.22 The peaks at 28.5°, 33.1°, 47.5°, 59.1°, 76.7°, and 79.1° were assigned to (111), (200), (220), (311), (222), (331), and (420) planes of a face-centered cubic (fcc) fluorite structure for CeO2 (JCPDS file 13-1002), indicating the formation of ceria in the catalyst. Recent contributions from Eichhorn’s group23,24 have offered a comprehensive structural evaluation of Pt-Ru bimetallic systems, which is very useful to recognize the diffraction peaks regarding the Pt-Ru crystal structure in our catalysts. The diffraction peaks around 40° (38°-42°) were free from the diffraction influence of the MWNT support and ceria NPs and hence could provide accurate information with respect to Pt-Ru phases. In such scenarios, a Pt (111) diffraction peak at 40.53° could be identified from the XRD pattern, as shown in Figure 1, the position of which shifted to higher 2θ, as compared with that for bulk Pt (39.76°, JCPDS file 04-0802). The Pt lattice parameter gives a 3.852 A˚ fcc lattice constant, which was slightly compressed with that of pure Pt at 3.923 A˚. These anomalies in the diffraction data for Pt in the catalyst may be attributable to incomplete lattice formation and strains associated with the two-dimensional structure, which have been elegantly described for both PtRu alloys and Ru@Pt core-shell systems.25 However, a diffraction peak at 38.8° was also discerned, which may arise from the crystalline hexagonal close-packed Ru in contrast with only the Pt fcc peaks in XRD spectra for PtRu alloys. These considerations suggest that Pt and Ru in the catalyst did not exist mainly as PtRu alloys but as other bimetallic architectures, such as core-shell or linked structures. The broadened peaks in the XRD pattern may be indicative of high dispersion of Pt-Ru and CeO2 NPs with small sizes in the catalyst. 3.2. X-ray Photoelectron Spectroscopy. The nature of surface species in the Pt-Ru/CeO2/MWNT catalyst was investigated (23) Alayoglu, S.; Nilekar, A. U.; Mavrikakis, M.; Eichhorn, B. Nature 2008, 7, 333–338. (24) Alayoglu, S.; Zavalij, P.; Eichhorn, B.; Wang, Q.; Frenkel, A. I.; Chupas, P. ACS Nano 2009, 3, 3127–3137. (25) Schlapka, A.; Lischka, M.; Gross, A.; Kasberger, U.; Jakob, P. Phys. Rev. Lett. 2003, 91, 0161011.
Langmuir 2010, 26(14), 12383–12389
by XPS. The wide survey profile for the catalyst with Pt-Ru loading at 20 wt % shows pronounced Pt, Ce, and O peaks in addition to the overlapped C 1s þ Ru 3d regions (Figure 2a). The C signal originated from the MWNT support and backgound. No trace of other heteroelements including Cl and N was detectable, ruling out the presence of unreacted precursors or formation of byproducts in the catalyst. Figure 2b shows the Ce 3d core-level XPS profile of the sample, which consists of two series of V and U peaks corresponding to 3d5/2 and 3d3/2 states, respectively, in line with previous report on Ce(IV).26 The peaks of V and V0 were attributed to a mixing configuration of 3d94f2 (O 2p4) and 3d94f1 (O 2p5) Ce4þ states and V00 to the 3d94f0 (O 2p6) Ce4þ state. The same assignment could be applied to the U structures, which correspond to the Ce 3d3/2 level.26 This manifests the major valence of Ce (IV) in the catalyst, consistent with the XRD result. The ill-defined valley between V and V0 was assigned to 3d94f1 (O 2p6) Ce3þ final state, indicating the presence of small amounts of Ce2O3 in the sample. The Pt 4f core level XPS pattern (Figure 2c) shows asymmetric peaks with a doublet, i.e., Pt 4f7/2 and 4f 5/2 centered at 72.1 and 75.4 eV, respectively. The main valence of Pt in the catalyst should be ascribed to Pt(0) because the Pt species were produced by the reduction via excessive NaBH4 during the synthetic process. Nevertheless, a remarkable shift to higher binding energy (BE) occurred, as compared with those of Pt foil (Pt 4f7/2 = 70.7 eV) and monometallic Pt nanoparticles (6.1 nm) (Pt 4f7/2 = 71.39 eV). This shift is probably due to the strong interaction between Pt and MWNT support or Pt and CeO2.27 Other relevance regarding this shift may be associated with the presence of Ru@Pt bimetals in the catalyst23 since a similar shift was also found for Pt in Ru@Pt bimetallic system. The other weaker doublet in the Pt 4f core level XPS pattern (Figure 2c) could be assigned to small amounts of Pt oxides, which may result from the slight oxidation of Pt upon exposure of the products to ambient air. Figure 2d displays the deconvoluted C 1s þ Ru 3d core level XPS pattern. The peak located at 285.2 eV may arise from C-OH present in pristine MWNTs. The doublet of Ru 3d5/2 (26) Ji, P. F.; Zhang, J. L.; Chen, F.; Anpo, M. J. Phys. Chem. C 2008, 112, 17809–17813. (27) Holmgren, A.; Duprez, D.; Andersson, B. J. Catal. 1999, 182, 441–448.
DOI: 10.1021/la101060s
12385
Article
Sun et al.
Figure 3. TEM images: (a) Pt-Ru/CeO2/MWNT at 10 wt % of Pt-Ru loading, the inset is the histogram of Pt-Ru NPs deposited on CeO2/ MWNT; (b) Pt-Ru/CeO2/MWNT at 10 wt % of Pt-Ru loading when mechanical stirring was applied instead of tip sonication under otherwise identical conditions; (c) HRTEM image of the sample shown in (a); (d) Pt-Ru/CeO2/MWNT at 40 wt % of Pt-Ru loading; (e) Pt/ MWNT at 20 wt % of Pt loading; (f) Pt-Ru/MWNT at 20 wt % of Pt-Ru loading. The insets in images e and f correspond to the EDS analyses of the composites.
and 3d3/2 centered at 281.9 and 286 eV, respectively, is indicative of Ru (0) in the catalyst.22 The other doublet of Ru 3d with higher BE reveals another chemically different Ru entity of Ru oxide as the result of Ru oxidation in contact with air. Note that both BEs of Ru(0) and Ru oxide in the catalyst shift to higher values relative to bulk Ru(0) and Ru oxide. Taking into account the fact that the BE of Ru shifts to lower values in Ru@Pt bimetals,23 this shift toward higher BEs is probably attributed to the strong interplay between Ru and CeO2 or Ru and MWNT support. 3.3. Transmission Electron Microscopy. Figure 3a shows a representative TEM image of the Pt-Ru/CeO2/MWNT at a 10 wt % Pt-Ru loading. It can be clearly seen that the exterior surfaces of nanotubes were decorated with well-dispersed fine nanoparticles. More importantly, careful TEM observations for different parts of the sample revealed that all nanoparticles deposited on the nanotube support, and all nanotubes were distributed with nanoparticles. By contrast, quite a few large NP aggregates with nonuniform sizes free from the nanotube support occurred, as can be confirmed by TEM observations, as shown in Figure 3b, when mechanical stirring was applied instead of tip sonication under otherwise identical conditions. Thus, it may be concluded that the good dispersion of NPs on nanotubes was resulted from the sonication utility. Put specifically, first, sonication enabled exfoliation of nanotube aggregates into individuals and simultaneously introduced defects on the nanotube surfaces. Second, it greatly accelerated the mass transfer and diffusion of the reactants and products, which promoted better wetting of nanotube surfaces with metal precursors. Third, sonication may prevent the obtained NPs from aggregating. As Ce(NO3)3 3 6H2O was mixed with KOH, Ce(OH)3 was produced and further converted into ceria, which favorably nucleated on 12386 DOI: 10.1021/la101060s
the defects of nanotubes. Likewise, sonication also facilitated the uniform dispersion of the produced metals (i.e., Pt-Ru) and their heterogeneous nucleation on the defective sites of nanotube surfaces. In addition, the metal NPs preferred to be in intimate contact with ceria probably due to their strong affinity.27 HRTEM measurements offer evidence that the nanoparticles were composed of CeO2 and Pt-Ru crystallites. Figure 3c illustrates the high crystallinity of nanoparticles with an interplanar spacing of 0.31 and 0.22 nm, corresponding to the {111} facets of CeO228 and Pt(0),29 respectively. A clear layer separation of 0.34 nm is attributed to the MWNT. Furthermore, careful HRTEM observations (as shown in Figure 3c) reveal that most Pt-Ru NPs are in close contact with ceria NPs whereas bare core-shell structures of (Pt-Ru)/CeO2, if any, were found. The histogram of Pt-Ru particle size (diameter) distribution (shown in the inset of Figure 3a) was extracted by directly measuring 110 particles with an interplanar spacing of ca. 0.22 nm in the HRTEM view, from which the mean diameter of Pt-Ru NPs was estimated to be ∼2.4 nm at the 10 wt % of Pt-Ru loading. Given the constant ceria content while increasing the Pt-Ru loading, the sizes of Pt-Ru NPs increased accordingly, and the surfaces of CeO2/MWNT were even densely wrapped in Pt-Ru NP layers at higher Pt-Ru loadings (i.e., 40 wt %), as shown in Figure 3d. Figures 3e,f show typical TEM images for the reference catalysts of Pt/MWNT and Pt-Ru/MWNT obtained in this work. It also can be observed that metal NPs with average sizes less than 3.5 nm were highly dispersed on the surfaces of MWNTs in both cases. (28) Sun, Z. Y.; Zhang, H. Y.; An, G. M.; Yang, G. Y.; Liu, Z. M. J. Mater. Chem. 2010, 20, 1947–1952. (29) Chen, J.; Herricks, Y. T.; Geissler, M.; Xia, Y. N. J. Am. Chem. Soc. 2004, 126, 10854–10855.
Langmuir 2010, 26(14), 12383–12389
Sun et al.
Figure 4. (a) An annular dark field image taken in the scanning TEM (STEM) mode for Pt-Ru/CeO2/MWNT at 10 wt % of Pt-Ru loading. The red line indicates the regions scanned for the EDS data shown in (b). The top line in (b) shows the counts for Pt, while the bottom line shows the counts for Ru.
3.4. Scanning TEM and Energy-Dispersive Spectroscopy. Figure 4a shows an annular dark field (ADF) image of the catalyst obtained in the scanning TEM (STEM) mode. The brightly white spots corresponded to CeO2 or Pt-Ru NPs deposited on MWNTs, and the average particle size was estimated to be around 3.0 nm. Elemental profiles of Pt and Ru taken across the scanned distance denoted by the red line are shown in Figure 4b. The EDS counts are proportional to the atomic concentration of Pt and Ru and reflect variations in the Pt and Ru concentrations over the scanned region. Note that the EDS profiles showed clear signs of segregation as illustrated in the areas between dot lines 1 and 2 and lines 3 and 4. To minimize the experimental error, we made such measurements for different regions, and similar results were achieved in all cases. This indicates that the major phase of Pt-Ru in the catalyst is not alloy, but other architectures, such as core-shell or linked structures, in accordance with the XRD result presented above. 3.5. Electrocatalytic Performance for Methanol Oxidation. The electrocatalytic activity of the Pt-Ru/CeO2/MWNT catalyst was investigated by cyclic voltammetry (CV) and chronoamperometry for methanol oxidation. For comparison, the electrocatalytic activities of Pt-Ru/MWNT and Pt/MWNT, with the identical metal loading (i.e., 20 wt %), were also examined. Shown in Figure 5A are a set of cyclic voltammograms for Pt-Ru/CeO2/MWNT, Pt-Ru/MWNT, and Pt/MWNT recorded in 0.5 M H2SO4 without methanol. The cathodic and anodic peaks appearing between -0.13 and 0.2 V were ascribed to the adsorption/desorption of hydrogen on the surfaces of Pt-Ru or Pt NPs correlating with Pt (110), Pt (111), and Pt (100) sites.30 Note that, in contrast to the PtRu alloy NPs whose typical CV was lack of Pt oxide reduction peak,31 both Pt-Ru/CeO2/ MWNT and Pt-Ru/MWNT samples showed well-defined reduction peaks of Pt oxides. This demonstration further confirms that few PtRu alloy NPs, if any, were formed in the as-prepared catalysts. From Figure 5A, it can be seen that Pt-Ru/CeO2/ MWNT showed remarkably higher specific current magnitude than both of Pt-Ru/MWNT and Pt/MWNT in the hydrogen region (-0.13 to 0.2 V) after normalization by Pt mass. Moreover, we estimated the electrochemically active surface area (ECSA) of our catalysts by ECSA = QH/(210 ( μC cm-2) Pt loading on electrode), which varied as Pt-Ru/CeO2/MWNT (122.4 m2 g-1 of Pt) > Pt-Ru/MWNT (77.4 m2 g -1 of Pt) > Pt/MWNT (36.6 m2 g-1 of Pt), where QH is the charge exchanged during the electroadsorption of hydrogen on Pt (-0.13 to 0.2 V vs (30) Kosaka, M.; Kuroshima, S.; Kobayashi, K.; Sekino, S.; Ichihashi, T.; Nakamura, S.; Yoshitake, T.; Kubo, Y. J. Phys. Chem. C 2009, 113, 8660–8667. (31) Du, B. C.; Rabb, S. A.; Zangmeister, C.; Tong, Y. Y. Phys. Chem. Chem. Phys. 2009, 11, 8231–8239.
Langmuir 2010, 26(14), 12383–12389
Article
Ag/AgCl).13,30 It is worthwhile to point out that the ECSA values of Pt-Ru/MWNT (77.4 m2 g-1) and Pt/MWNT (36.6 m2 g-1) prepared in this work were comparable to those reported for Pt-Ru/MWNT (53.5 m2 g-1 of Pt) and Pt/MWNT(47.1 m2 g-1) in the literature.14 However, the ECSA value of the as-synthesized Pt-Ru/CeO2/MWNT was higher than those of these two catalysts and also higher than those of typical nanostructured Pt films (38 m2 g-1), mesoporous Pt powders (60 m2 g-1), and Pt/C (14-87 m2 g-1).32 The reason for this phenomenon is unclear at present. TEM measurements showed that the size of Pt-Ru NPs in the Pt-Ru/CeO2/MWNT was approximately identical to that in the Pt-Ru/MWNT, implying that the presence of CeO2 had marginal effect on the dispersion quality of metal NPs in the composites. As such, we speculate that the higher ECSA value for Pt-Ru/CeO2/MWNT compared to the Pt-Ru/MWNT may result from the synergistic effect between Pt and the ionic conductor of CeO2, which may promote the hydrogen spillover rate of Pt-H and hence increases the dissociation of hydrogen adsorption. Similar hydrogen spillover occurrence was observed by Sung et al. using electrochromism for a Pt-WOx system.33 Methanol electrooxidation over Pt-Ru/CeO2/MWNT, Pt-Ru/ MWNT, and Pt/MWNT catalysts were studied by cyclic voltammetry measurements performed in 0.5 M H2SO4 and 1 M methanol solution at a scan rate of 0.02 V s-1, and the corresponding CV curves are shown in Figure 5B. All the curves showed voltammetric characteristics as typical electrooxidation of methanol on Pt-based catalysts, each of which included a methanol oxidation peak during the forward scan at around 0.6-0.8 V and another anodic peak during the reverse scan associated with the removal of incompletely oxidized carbonaceous species (such as CO) formed in the forward scan. Comparing the peak current in the forward scan after normalization in terms of Pt mass allows us to evaluate their catalytic activity for methanol oxidation. As illustrated in Figure 5B, the electrocatalytic performance of Pt-Ru/CeO2/ MWNT indeed outweighed those of both Pt-Ru/MWNT and Pt/MWNT for methanol oxidation. Additionally, Pt-Ru/MWNT showed higher electrocatalytic performance than Pt/MWNT as expected, probably mainly due to a synergistic bifunctional mechanism well-articulated for a Pt-Ru surface8 as well as a possible additional electronic effect.9 Alternatively, the ratio of the forward anodic peak current (If) to the backward anodic peak current (Ib) is another important factor for evaluating the tolerance of catalysts to the accumulation of carbonaceous species. Higher If/Ib suggests more efficiency in removal of poisoning species from the surface of catalysts.34 We found that both of Pt-Ru/CeO2/MWNT and Pt-Ru/MWNT showed higher If/Ib (≈2.31) than that of Pt/ MWNT (≈1.63). In such a scenario, we are not surprised to find that the reverse peak potential for carbonaceous species oxidation on both of Pt-Ru/CeO2/MWNT and Pt-Ru/MWNT was ∼0.49 V, in contrast with 0.55 V on Pt/MWNT. This negative shift was also the indication of more readily removing the intermediates from the catalysts. Chronoamperometric measurements offer a straightforward comparison of the electrocatalytic activity for methanol oxidation among the three catalysts. As clearly illustrated in Figure 5C, Pt-Ru/CeO2/MWNT showed the highest oxidative current, further indicating its superior electrochemical performance for methanol oxidation compared to Pt-Ru/MWNT and Pt/MWNT catalysts. This enhancement can be ascribed to the presence of CeO2 phase in Pt-Ru/CeO2/MWNT as the metal (32) Liu, Y.; Chen, J.; Zhang, W. M.; Ma, Z. F.; Swiegers, G. F.; Too, C. O.; Wallace, G. G. Chem. Mater. 2008, 20, 2603–2605. (33) Park, K. W.; Ahn, K. S.; Nah, Y. C.; Choi, J. H.; Sung, Y. E. J. Phys. Chem. B 2003, 107, 4352–4355. (34) Gu, Y. J.; Wong, W. T. Langmuir 2006, 22, 11447–11452.
DOI: 10.1021/la101060s
12387
Article
Sun et al.
Figure 5. (A) Cyclic voltammograms for electrodes prepared using Pt-Ru/CeO2/MWNT (solid line), Pt-Ru/MWNT (dotted line), and Pt/
MWNT (dashed line) in 0.5 M H2SO4 at a scan rate of 0.3 V s-1. The molar ratio of Pt and Ru in both Pt-Ru/CeO2/MWNT and Pt-Ru/ MWNT catalysts is 1:1. (B) Cyclic voltammograms for the three electrodes in 0.5 M H2SO4 þ 1 M methanol at a scan rate of 0.02 V s-1. (C) Chronoamperometry curves for the three electrodes. The currents were normalized in terms of Pt mass.
Figure 6. (A) Cyclic voltammograms for electrodes prepared using Pt-Ru/CeO2/MWNT with Pt-Ru loadings at 5 wt % (dotted line), 10 wt
% (solid line), 30 wt % (dashed line), and 40 wt % (dash-dotted line) in 0.5 M H2SO4 at a scan rate of 0.3 V s-1. The inset corresponds to the reduction potential of Pt oxides as a function of Pt-Ru loading where the reduction potential of Pt oxides on Pt/MWNT is also included as denoted by the dotted line. The molar ratio of Pt and Ru is 1:1 for the four catalysts. (B) Cyclic voltammograms for the four electrodes in 0.5 M H2SO4 þ 1 M methanol at a scan rate of 0.02 V s-1. (C) Chronoamperometry curves for the four electrodes. The currents were normalized in terms of Pt mass.
loading and NP size in the catalysts are reasonably the same, and thus metal size and loading effects can be excluded. Figure 6A shows the CV curves of Pt-Ru/CeO2/MWNT with metal loadings varying from 5 to 40 wt % in 0.5 M H2SO4. Note that the Pt-Ru/CeO2/MWNT at a 10 wt % of Pt-Ru loading showed the highest electrochemical surface area of 134.8 m2 g-1 of Pt. This value was comparable to those of PtRu/PIL/MWNT (91.2 m2 g-1 of Pt),14 PtRuSnOx/MWNT (63.2 m2 g-1 of Pt),17 and PtRu/MnO2/MWNT (112 m2 g-1 of Pt)18 reported in recent papers. Pt utilization efficiency (ηPt) is an important gauge to describe the catalyst performance, which is generally calculated by dividing ECSA by the geometrical specific surface area (Sgeom) of Pt NPs. Sgeom can be determined by Sgeom = 6/(Fd ), where F is the density of Pt (21.09 g cm-3) and d is the mean diameter of Pt NPs.13 Taking the average size of Pt-Ru NPs as dPt, the Sgeom of Pt-Ru/ CeO2/MWNT (10 wt % of Pt loading) was calculated to 118.5 m2 g-1, and thus ηPt was as high as 114%. This unexpected occurrence resulted from the overestimation of Pt NP size and hence underestimation of Sgeom by assuming that the size of Pt NPs was identical to that of Pt-Ru NPs. In principle, the size of Pt is less than that of Pt-Ru bimetallic (not alloy) NPs. Despite the uncertainty of ηPt, the Pt utilization efficiency might be reasonably improved taking into account the remarkable enhancement of the ECSA for Pt-Ru/CeO2/ MWNT in contrast to Pt/MWNT. Shown in Figures 5B and 6B are the CVs recorded for Pt-Ru/CeO2/MWNT with different metal loadings in 0.5 M H2SO4 and 1 M methanol. Despite almost no pronounced variation in If/Ib, the oxidation peak current of methanol oxidation differed sharply with respect to the Pt-Ru loadings, which followed the trend as 10 > 20 > 30 > 5 > 40 wt %, agreeing well with that of their 12388 DOI: 10.1021/la101060s
corresponding ECSA values. This trend also has been confirmed by chronoamperometric measurements, as shown in Figures 5C and 6C. In light of these findings, we can conclude that an optimal Pt-Ru loading for the as-prepared Pt-Ru/ CeO2/MWNT catalysts regarding methanol oxidation exists probably at 10 wt % as opposed to the monotonic change of electrocatalytic activity of Pt/MWNT with metal loading reported previously.15,16 The exact reason for this phenomenon remains elusive. As revealed by TEM observations, the sizes of Pt-Ru increased with the increment of Pt-Ru loading on CeO2/MWNT. In addition, superimposition and aggregation of NPs occurred at high metal loadings. Both of these aspects unequivocably lead to the reduction in the surface areas of NPs available and consequently to the decrease in their catalytic activity. However, the increase in metal loading simultaneously results in the increase of NP interconnectivity and hence the enhancement of catalytic activity. The catalytic activity of Pt-Ru/CeO2/MWNT for methanol oxidation was maximized at 10 wt % of Pt-Ru loading as a result of the balance between these two competitive processes. We also made a preliminary study of the Ru content effect on the methanol electrocatalytic oxidation. Shown in Figure 7 are the CVs for electrodes prepared using Pt-Ru/CeO2/MWNT at a 10 wt % of Pt-Ru loading with Pt-to-Ru molar ratios varying from 1:3 to 2:1 in 0.5 M H2SO4 and 1 M methanol solution. The plausibly lower onset potential for Pt-to-Ru ratio of 1:2 and 2:1 as compared to 1:1 was highly probably originated from the influence of oxygen (i.e., the reduction of oxygen) remained in the solution. We note that the sample with the Pt-to-Ru ratio at 1:1 exhibited the highest methanol oxidation current, which likely indicates that the optimal Pt-to-Ru ratio was 1:1 for the Pt-Ru/ CeO2/MWNT at 10 wt % of Pt-Ru loading. This optimal value Langmuir 2010, 26(14), 12383–12389
Sun et al.
Article
CeO2, which can be further accelerated in Pt/CeO2 catalyst, resulting in more oxygen vacancies on the oxide surface:27 CeO2 T CeO2 - x þ x=2O2
Figure 7. Cyclic voltammograms for electrodes prepared using Pt-Ru/CeO2/MWNT at a 10 wt % of Pt-Ru loading with Ptto-Ru molar ratios varying from 1:3 to 2:1 in 0.5 M H2SO4 and 1 M methanol. The scanning rate of the applied voltage is 0.02 V s-1. The currents were normalized in terms of Pt mass.
was in great contrast to the most recognization that methanol oxidation was generally favored on Pt-rich surfaces with less neighboring Ru surface atoms.35-39 An explanation for this may be partly due to the presence of ceria in the catalyst, which changed the optimal surface composition of Pt-to-Ru ratio for methanol oxidation. It is well recognized that methanol oxidation over Pt catalysts proceeds through complex parallel and consecutive reaction pathways.3 Nevertheless, the mechanistic role of cocatalysts is still a subject of ongoing discussion. In the case of PtRu bimetallic catalysts for methanol electrooxidation, the oxidative removal of COad is promoted probably due to a bifunctional mechanism as well as electronic effects. Recent studies have shown that an alloy formation is not a prerequisite for obtaining high CO tolerance while the Pt and Ru atoms are necessarily to be in close proximity probably to satisfy an ensemble requirement for methanol oxidation.40 The active state of Ru is another subject of hot debate. Ru(0) is mostly considered as the active component, while recent contributions suggest the significant role of oxides of Ru41 and hydrous ruthenium oxides.12 More specifically, Chetty et al.12 proposed that hydrous ruthenium oxide expressed as RuOxHy is a mixed proton and electron conducting species, which can adsorb large quantities of OH species and thus facilitate the oxidative removal of COad. We note that there were some ruthenium oxide in the as-prepared Pt-Ru/CeO2/MWNT and Pt-Ru/MWNT catalysts corroborated by XPS measurements, which may partly contribute to the higher catalytic activity through a synergistic effect compared to Pt/MWNT. The significance of CeO2 in the catalyst toward methanol oxidation may be explained from both an electronic effect and a bifunctional mechanism point of view. It is known that cerium ion can switch between Ce4þ and Ce3þ in (35) Gasteiger, H. A.; Markovic, N.; Ross, P. N.; Cairns, E. J. J. Phys. Chem. 1993, 97, 12020–12029. (36) Iwasita, T.; Hoster, H.; John-Anacker, A.; Lin, W. F.; Vielstich, W. Langmuir 2000, 16, 522–529. (37) Cuesta, A. J. Am. Chem. Soc. 2006, 128, 13332–13333. (38) Lin, M. L.; Lo, M. Y.; Mou, C. Y. J. Phys. Chem. C 2009, 113, 16158– 16168. (39) Waszczuk, P.; Solla-Gullon, J.; Kim, H. S.; Tong, Y. Y.; Montiel, V.; Aldaz, A.; Wieckowski, A. J. Catal. 2001, 203, 1–6. (40) Maillard, F.; Gloaguen, F.; Leger, J. M. J. Appl. Electrochem. 2003, 33, 1–8. (41) Gu, Y. J.; Wong, W. T. J. Electrochem. Soc. 2006, 153, A1714.
Langmuir 2010, 26(14), 12383–12389
ð0exe0:5Þ
ð1Þ
This may provide active sites for CO species adsorption and hence favorably leaving more Pt sites for further electrocatalytic reaction. On the other hand, the binding strength of Pt-CO may be weakened due to the electron transfer from Pt to CeO2. On the basis of the bifunctional mechanism, CeO2 may promote the dissociation of coordinated water resembling Ru forming more OH species to oxidize COad and release more Pt active sites. In this regard, some possible reaction routes for the methanol electroxidation on Pt-Ru/CeO2/MWNT can be expressed as follows: Pt þ CeO2 - x þ H2 O f CeO2 - x - ðOHÞad þ Pt- Had Pt þ Ru þ H2 O f Ru- ðOHÞad þ Pt- Had
ð2Þ ð3Þ
2Pt- ðCOÞad þ CeO2 - x - ðOHÞad þ Ru- ðOHÞad f 2CO2 þ 2Hþ þ 2Pt þ CeO2 - x þ Ru þ 2e -
ð4Þ
Note that the role of CeO2 in the catalyst with respect to methanol electroxidation has not been well established. In addition, the optimal CeO2 loading for maximizing the electrocatalytic activity of Pt-Ru/CeO2/MWNT is unknown at present. Both of these two aspects warrant further investigation.
4. Conclusions To conclude, we reported an efficient Pt-Ru/CeO2/MWNT catalyst for DMFCs. Pt-Ru/CeO2/MWNT catalysts were readily synthesized by means of a very simple and rapid sonication process. Both TEM and STEM observations manifested that CeO2 and Pt-Ru NPs were in intimate contact and uniformly deposited on the outer surfaces of nanotubes with very narrow size distributions. XRD and EDS analyses confirmed that the major structure of Pt-Ru was other bimetallic architectures rather than alloy. Electrochemical measurements showed that the presence of CeO2 significantly contributed to the enhancement of activity toward methanol oxidation compared to Pt-Ru/ MWNT and Pt/MWNT catalysts. Further studies demonstrated that the optimal value of Pt-Ru loading and Pt-to-Ru ratio occurred at 10 wt % and 1:1, respectively, for a maximum activity toward methanol oxidation over the Pt-Ru/CeO2/MWNT catalyst. Considering that ceria is much cheaper than Pt coupling with the fact that Ru is currently 7 times less expensive than Pt, we should point out the as-synthesized Pt-Ru/CeO2/MWNT has great promising applications in DMFCs. Acknowledgment. This work is financially supported by the National Natural Science Foundation of China (No. 20903105) and the Chinese Academy of Sciences (KJCX2.YW.H16).
DOI: 10.1021/la101060s
12389
