11064
Langmuir 2004, 20, 11064-11072
FTIR Study of the Ethanol Electrooxidation on Pt(100) Modified by Osmium Nanodeposits V. Pacheco Santos, V. Del Colle, R. Batista de Lima, and G. Tremiliosi-Filho* Instituto de Quı´mica de Sa˜ o Carlos-USP, CP 780, 13560-970, Sa˜ o Carlos, SP, Brazil Received January 5, 2004. In Final Form: August 24, 2004 In the present work, ethanol electrooxidation on a Pt(100) electrode modified by different coverage degrees of osmium nanoislands obtained by spontaneous depositions, was extensively studied employing in situ FTIR spectroscopy. A collection of spectra of the ethanol adsorption and oxidation processes was acquired during the first series of a positive potential step, to determine the intermediate species, as well as the main products formed. The spectroscopic results obtained were correlated with conventional electrochemical results obtained by cyclic voltammetry. It was shown that the catalytic activity of Pt(100) for ethanol oxidation increases significantly after osmium deposition and that the mechanistic pathway for this reaction depends directly on the osmium coverage degree. Thus, for low osmium coverage (θ Os up to 0.15) the formation of CO as an intermediate was favored and hence the full oxidation of adsorbed ethanol to CO2 was increased. For higher osmium coverages (θOs up to 0.33), the higher the coverage is, the more the direct ethanol oxidation to acetaldehyde and acetic acid is favored. For osmium coverage degree of 0.40, the catalytic activity of the electrode for ethanol oxidation decreased. On an almost complete osmium layer (θOs ) 0.92) obtained by electrodeposition at 50 mV vs reversible hydrogen electrode, the catalytic activity for ethanol oxidation shows a much lower value.
1. Introduction Over the past 20 years, great effort has been devoted to the study of the electrooxidation of small organic molecules. This is because such molecules can be used as fuels for the anodic reactions that occur in direct organic fuel cell (DOFCs) devices.1-3 A considerable number of spectroscopic, electrochemical, and mass analysis methods has been employed to determine the reaction intermediates and elucidate the manner in which the organic molecules adsorb on metal surfaces.4,5 Among the mentioned methods, the most used is FTIR spectroscopy, due to the fact that this technique affords valuable insight into the electrocatalytic processes studied in electrochemical surface science.6-14 In the beginning of the 1980s, this powerful tool attracted great attention from research groups as it * Corresponding author. Telephone: +55 (16) 33739933. Fax: +55 (16) 33739952. E-mail:
[email protected]. (1) Iwasita, T.; Vielstich, W. In Advances in Electrochemical Sciences and Engineering; Gersischer, H., Tobias, C. W., Eds.; VCH: Weinheim, Germany, 1990; Vol. 1. (2) Parsons, R.; Vandernoot, T. J. Electroanal. Chem. 1988, 257, 9-45. (3) Beden, B.; Lamy, C.; Le´ger, J.-M. In Modern Aspects of Electrochemistry; Bockis, J. O’. M., Conway, B. E., White, R. E., Eds.; Plenum: New York, 1992; Vol. 22, p 97. (4) Willsau, J.; Heitbaum, J. J. Electroanal. Chem. 1985, 194, 2735. (5) Beden, B.; Morin, M.-C.; Hahn, F.; Lamy, C. J. Electroanal. Chem. 1987, 229, 353-366. (6) Iwasita, T.; Vielstich, W. J. Electroanal. Chem. 1988, 257, 319324. (7) Iwasita, T.; Rasch, B.; Cattaneo, E.; Vielstich, W. Electrochim. Acta 1989, 34, 1073-1079. (8) Pe´rez, J. M.; Beden, B.; Hahn, F.; Aldaz, A.; Lamy, C. J. Electroanal. Chem. 1989, 262, 251-261. (9) Chang, S.-C.; Leung, L.-W. H.; Weaver, M. J. J. Phys. Chem. 1990, 94, 6013-6021. (10) Rasch, B.; Iwasita, T. Electrochim. Acta 1990, 35, 989-993. (11) Iwasita, T.; Pastor, E. Electrochim. Acta 1994, 39, 531-537. (12) Gootzen, J. F. E.; Visscher, W.; van Veen, J. A. R. Langmuir 1996, 12, 5076-5082. (13) Xia, X. H.; Iwasita, T.; Ge, F.; Vielstich, W. Electrochim. Acta 1996, 41, 711-718. (14) Xia, X. H.; Liess, H.-D.; Iwasita, T. J. Electroanal. Chem. 1997, 437, 233-240.
provided a way to understand the nature of the interactions between organic molecules and the electrode surface. The technique supplies a molecular-level description of the reactions and permits the correlation of the catalytic activity with the crystallographic structure of the electrode surface. Thus, it was possible to elucidate the mechanistic pathways of various surface reactions and analyze the electrooxidation kinetics for several organic molecules.15 Most ethanol electrooxidation studies have been carried out on polycrystalline platinum electrodes. This was done mainly to quantify the reaction intermediates, products formed, and/or kinetic parameters,6-8,11,12,16 but only a few studies have detailed FTIRS studies of this reaction on single-crystal surfaces.9,14,17,18 Some aspects, such as the reaction mechanism and surface reactivity, can be more precisely investigated by working with well-defined surfaces. Ethanol electrooxidation has been studied for a long time due to the fact that this simple molecule can be used as a prospective renewable power source, associated with its low toxicity when compared with other alcohols, in DOFC devices for mobile applications.19,20 However, further studies are necessary to ensure that ethanol can be considered an ideal candidate for such applications as significant surface poisoning of the catalyst is generally observed. This poisoning is caused by dissociative ethanol chemisorption, which leads to the formation of strongly adsorbed CO and other adsorbates such as dCHOH enol type species and/or CH3CHO, depending mainly on the bulk ethanol concentration.8 (15) Sun, S.-G. In Electrocatalysis; Lipkowski, J., Ross, P. N., Eds.; Wiley-VCH: Weinheim, Germany, 1998; p 243. (16) Hitmi, H.; Belgsir, E. M.; Le´ger, J.-M.; Lamy, C.; Lezna, R. O. Electrochim. Acta 1994, 39, 407-415. (17) Leung, L.-W. H.; Chang, S.-C.; Weaver, M. J. J. Electroanal. Chem. 1989, 266, 317-336. (18) Shin, J.; Tornquist, W. J.; Korzeniewski, C.; Hoaglund, C. S. Surf. Sci. 1996, 364, 122-130. (19) Wang, J.; Wasmus, S.; Savinell, R. F. J. Electrochem. Soc. 1995, 142, 4218-4224. (20) Fujiwara, N.; Friedrich, K. A.; Stimming, U. J. Electroanal. Chem. 1999, 472, 120-125.
10.1021/la040001v CCC: $27.50 © 2004 American Chemical Society Published on Web 11/11/2004
Ethanol Electrooxidation on Pt(100)
A possible path to minimize the poisoning effect is to add other metals (e.g., Ru,21-24 Os,25-28 Sn,21,29 and/or Rh 30) to the platinum surface. The well-known reason for the use of these metals is explained on the basis of the bifunctional mechanism.31 Another important issue that should be considered when choosing a catalyst is to encounter metals that are capable of cleaving the C-C bond of the ethanol molecule.32,33 Basically, a dual path mechanism has been proposed for ethanol electrooxidation. The first one is via direct oxidation of adsorbed ethanol to acetaldehyde9,14,32 and acetic acid14,34 and the other is via cleavage of the C-C bond, which produces adsorbed CHx species12 and carbon monoxide in a major proportion.9,12,14 According to the literature, a way to improve the ethanol oxidation reaction on platinum is to add Ru as a promoter. Ianniello et al.,33 working with 13C-labeled ethanol, ascertained that Ru atoms on the Pt surface stimulate the formation of CO2 originating from the CH2OH- group at low potentials. However, according to the authors, Ru seems to be a good modifier for methanol oxidation, but not for ethanol. On the other hand, Souza et al.,24 studying ethanol oxidation on Pt-Ru by in situ FTIR spectroscopy, observed that as soon as ethanol adsorbs on the metal surface it is readily oxidized to CO2, suggesting that this catalyst can easily break the C-C bond, even at low potentials. In the same way, a number of studies demonstrate that osmium also tends to improve the catalytic activity of the electrode for ethanol electrooxidation.27,28 In one of these studies, it was observed that osmium spontaneously deposited on platinum single-crystal electrodes forms nanoislands with diameters of 1-4 nm and monatomic thickness.28 Similar sub-monolayer osmium nanoislands deposited on Pt(100) will be used in the present study. It is well-known that there is a clear connection between the structure/reactivity of a single crystal and practical electrocatalysis, in which organic fuel activation takes place on metal catalysts dispersed as small single-crystal particles that are supported on high-surface-area carbon substrates.35,36 Such small single-crystal particles develop a distinct nanostructure that induces steric and electronic differences among the catalytic sites. The terraces on the supported particles have a relatively low reactivity (21) Frelink, T.; Visscher, W.; van Veen, J. A. R. A. Surf. Sci. 1995, 335, 353-360. (22) Schmidt, V. M.; Ianniello, R.; Pastor, E.; Gonzalez, S. J. Phys. Chem. 1996, 100, 17901-17908. (23) Del Colle, V.; Giz, M. J.; Tremiliosi-Filho, G. J. Braz. Chem. Soc. 2003, 14, 601-609. (24) Souza, J. P. I.; Rabelo, F. J. B.; de Moraes, I. R.; Nart, F. C. J. Electroanal. Chem. 1997, 420, 17-20. (25) Crown, A.; Moraes, I. R.; Wieckowski, A. J. Electroanal. Chem. 2001, 500, 333-343. (26) Rhee, C. K.; Wakisaka, M.; Tolmachev, Y. V.; Johnston, C. M.; Haasch, R.; Attenkofer, K.; Lu, G. Q.; You, H.; Wieckowski, A. J. Electroanal. Chem. 2003, 554, 367-378. (27) Pacheco Santos, V.; Tremiliosi-Filho, G. J. Electroanal. Chem. 2003, 554, 395-405. (28) Pacheco Santos, V.; Del Colle, V.; Bezerra, R M.; TremiliosiFilho, G. Electrochim. Acta 2004, 49, 1221-1231. (29) Morimoto, Y.; Yeager, E. B. J. Electroanal. Chem. 1998, 444, 95-100. (30) Souza, J. P. I.; Queiroz, S. L.; Bergamaski, K.; Gonzalez, E. R.; Nart, F. C. J. Phys. Chem. B 2002, 106, 9825-9830. (31) Watanabe, M.; Motoo, M. J. Electroanal. Chem. 1975, 60, 267. (32) Me´ndez, E.; Rodrı´guez, J. L.; Are´valo, M. C.; Pastor, E. Langmuir 2002, 18, 763-772. (33) Ianniello, R.; Schmidt, V. M.; Rodrı´guez, J. L.; Pastor, E. J. Electroanal. Chem. 1999, 471, 167-179. (34) Tarnowski, D. J.; Korzeniewski, C. J. Phys. Chem. B 1997, 101, 253-258. (35) Stonehart, P. J. Appl. Electrochem. 1992, 22, 995-1001. (36) Radmilovic _, V.; Gasteiger, H. A.; Ross, P. N., Jr. J. Catal. 1995, 154, 98-106.
Langmuir, Vol. 20, No. 25, 2004 11065
compared with those on edges and defects.37-39 Both types of sites, terraces and defects, can be modeled using singlecrystal platinum surfaces: Pt(111) and Pt(100) imitate the terraces, and Pt(110) simulates the metal defects. This approach has been proven feasible, as ethanol electrooxidation is extremely sensitive to the surface crystallographic orientation of platinum.14 There is also conclusive experimental evidence that ethanol oxidation increases with osmium addition and that the enhancement depends on surface geometry.27,28 These observations create a template for the proposed FTIRS studies on Pt(100), reported here as parts of a more complete investigation involving the three principal lowindex crystal orientations. Additionally, the Pt(100)/Os system is very active for ethanol oxidation, and the FTIRS certainly will provide insight into the effect of this surface structure at a molecular level. Thus, the aim of this work is to study the activity of Pt(100) modified by different osmium coverage degrees spontaneously deposited for ethanol oxidation, using in situ FTIR spectroscopy. It is also the objective of this work to identify the intermediates and products of this reaction at different potentials and to correlate the FTIR spectra with electrochemical results obtained by cyclic voltammetry. 2. Experimental Section 2.1. Chemicals. All solutions were prepared with MilliQ purified water (>18 MW). A solution of 0.1 M HClO4 (Merck p.a.) was prepared as the base electrolyte. The osmium deposits were obtained from a solution of 1.0 mM H2OsCl6 (Alfa Aesar) in 0.1 M HClO4. The reactant solution for ethanol oxidation study was 0.5 M ethanol (J. T. Baker) in 0.1 M HClO4. To deaerate the solutions, argon (99.9%) or nitrogen (99.999%) was used. 2.2. Electrodes and Cells. A Pt(100) single-crystal cylinder of 5.0 mm in diameter and 6.3 mm in height was used as the working electrode. The electrochemical experiments were performed in a conventional glass cell. The FTIRS experiments were carried out in a glass cell fitted with 60° prismatic CaF2 or ZnSe windows in the thin electrolyte layer pattern.13,14 All potentials in this study are referred to the reversible hydrogen electrode (RHE). 2.3. Instrumentation. The electrochemical measurements were performed using a Potentiostatic-Galvanostatic Autolab, model PGSTAT 30. In situ FTIRS experiments were performed with a Nicolet Nexus 670 spectrometer and an MCT detector. 2.4. Experimental Procedure. Before each osmium deposition, the correct surface orientation of the Pt(100) electrode was confirmed by cyclic voltammetry (always in a fresh HClO4 solution) using a meniscus configuration. The voltammetric profile also attests to the cleanliness of the system40 (see solid lines in Figure 5). Following this, the osmium depositions on the Pt(100) electrode were performed from a solution of 1.0 mM H2OsCl6 + 0.1 M HClO4. The osmium deposits were obtained at a sub-monolayer level by spontaneous deposition (Eoc ) 790 mV) at different deposition times between 2 and 150 s; and an almost complete osmium layer (θ Os ) 0.92) was electrodeposited at 50 mV during 300 s. After each deposition and before ethanol oxidation, cyclic voltammograms of the electrodes in the hydrogen adsorption-desorption region were recorded at 50 mV s-1 in the base electrolyte saturated with N2. The spontaneous osmium deposition times utilized in this work were related to the respective osmium coverage degrees according to the method described in previous studies, in which the θ Os values were estimated from the decrease of the charges of hydrogen adsorp(37) Somorjai, G. A. Chemistry in Two Dimensions: Surface; Cornell University Press: Ely House, London, 1981. (38) Somorjai, G. A. Surf. Sci. 1991, 242, 481-488. (39) van Hove, M. A.; Weinberg, W. H.; Chan, C.-M. Low-Energy Electron Diffraction: Experiment, Theory, and Surface Structure Determination; Springer-Verlag: Heidelberg, Germany, 1986. (40) Rodes, A.; Clavilier, J. J. Electroanal. Chem. 1992, 338, 317338.
11066
Langmuir, Vol. 20, No. 25, 2004
Pacheco Santos et al.
Figure 1. Collection of in situ FTIR spectra (256 scans at 8 cm-1 resolution) of the first series of positive potential steps for ethanol oxidation (0.5 M C2H5OH + 0.1 M HClO4) on nonmodified Pt(100), on Pt(100) with different osmium coverage degrees (obtained after spontaneous deposition times of 2-150 s), and on an almost complete osmium layer (θOs ) 0.92) obtained by electrodeposition at 50 mV. tion-desorption peaks in the voltammograms obtained in H2SO4 solution.27,28 For the oxidation studies, the electrode was immersed in a 0.5 M C2H5OH + 0.1 M HClO4 solution. During this process, the electrode was maintained polarized at 0.05 V in order to minimize strong ethanol adsorption on platinum. After this, a reference FTIR spectrum was obtained at this potential. Spectra were acquired after applying successive potential steps of 50 mV in the positive direction from 0.1 to 1.0 V vs RHE. Spectra were computed from the average of 256 interferograms, and the spectral resolution was set at 8 cm-1. Reflectance spectra were calculated as the ratio (R/R0) of the sample (R) and the reference (R0) spectra. Positive and negative growing bands represent loss and gain of species at the sampling potential, respectively. When extension of the low-wavenumber limit was necessary (below 1000 cm-1), a ZnSe window was used. After finishing each osmium deposition and respective ethanol oxidation FTIR studies, the electrode was cycled between 50 and 1650 mV at 200 mV s-1 in fresh electrolyte. This procedure favors the oxidation and dissolution of the osmium deposits, which can be confirmed through the return of the voltammetric profile to that of platinum in HClO4 solution. Due to surface disordering after removal of the osmium deposit, hydrogen flame annealing of the electrode was necessary in order to regenerate the original single-crystal surface.27,28
Pt(100) with different osmium coverage degrees shows that the bands of the main intermediates and products of the reaction (linearly bonded CO (2040 cm-1); bridged bonded CO (1850 cm-1); and CO2 (2340 cm-1)6,7,10) increased significantly after osmium surface modification (part I). In addition, other bands arose in the spectra during the positive steps due to the formation of acetaldehyde (933 cm-1)41 and acetic acid (1290 cm-1),6,42 as well as the carbonyl group band related to both of these species (1710 cm-1)6,10 (part II). 3.1. In Situ FTIRS Study. Part I. The collection of selected spectra presented in Figure 1 was acquired for bare Pt(100) and Pt(100) with different osmium coverage degrees. Each group of reflectance spectra was obtained in the range of 0.1-1.0 V by applying successive potential steps of 50 mV in the positive direction and calculated as the ratio between every sample spectrum and the reference spectrum recorded at 0.05 V. Figure 1a shows the spectra of ethanol oxidation at 0.2 V for nonmodified Pt(100) and modified Pt(100) with different osmium coverage degrees. The FTIRS results indicate that, at this potential, the early stages of ethanol oxidation occur only on the modified surfaces. The bands associated with the adsorption of bridge- and linear-bonded
3. Results and Discussion The collection of FTIR spectra for ethanol electrooxidation during the first series of positive potential steps on
(41) Davis, J. L.; Barteau, M. A. Surf. Sci. 1990, 235, 235-248. (42) Rodriguez, J. L.; Pastor, E. Xia, X. H. Iwasita, T. Langmuir 2000, 16, 5479-5486.
Ethanol Electrooxidation on Pt(100)
Figure 2. Band intensity of CObridge (1850 cm-1) from FTIR spectra for ethanol oxidation as a function of the potential (data obtained from Figure 1): on nonmodified Pt(100) and on Pt(100) modified with different osmium coverage degrees.
CO on the modified surfaces can clearly be observed in Figure 1a. The intensity of the CObridge band reaches a maximum value at 0.2 V for the lowest osmium coverage degree studied (θ Os ) 0.15). This observation will be emphasized later in the discussion of Figure 2. For higher osmium coverage degrees, the CObridge band tends to diminish until it disappears on the almost complete osmium layer. The increase of the COlinear band follows the decrease of the CObridge band in the potential range between 0.2 and 0.5 V, as can be seen in Figure 1a-e. The fact that the band at 1850 cm-1 starts to decrease at potentials ca. 300 mV lower than the potential in which the formation of CO2 begins (ca. 0.5 V) provides evidence of the conversion of CObridge to COlinear at potentials lower than 0.5 V.13,14 According to the FTIRS results, 0.5 V is the potential where the formation of CO2 begins. The amount of CO2 reaches a maximum value at ca. 0.8 V, while CO is quickly consumed from 0.5 to 0.8 V, as can be seen in Figure 1eh. The surfaces are completely free of CO at potentials of 0.9 V or higher (see Figure 1i,j). The quantity of CO2 decreases at potentials higher than 0.7 V, reaching ca. 30% of the maximum value at 1.0 V. All the amounts of CO2 formed can be affected by the gradual diffusion of this species out of the thin electrolyte layer. The intensity of the CObridge band (Ib) taken from the spectra in Figure 1 and plotted as a function of the potential for nonmodified and modified Pt(100) is presented in Figure 2. Figure 2 shows that the adsorption of bridgebonded CO starts at lower potentials on the surfaces with osmium deposits (ca. 100 mV earlier than that observed on nonmodified Pt(100)). Moreover, the quantity of this adsorbed species is higher on the surface with the lowest osmium coverage degree studied (θOs ) 0.15). For Pt-Os electrodes with osmium coverage degrees of 0.25 and 0.33, the amount of CObridge formed is practically constant, mainly at potentials between 0.35 and 0.60 V. For the surface with a higher osmium coverage degree (θOs ) 0.40), the CObridge band diminishes, and it disappears on an almost complete osmium layer (θOs ) 0.92). With the exception of nonmodified Pt(100) and Pt(100) with an osmium coverage degree of 0.40, the CObridge is totally removed from the surface before the potential reached 0.7 V. In the same way as that seen for the CObridge band, the COlinear band reaches a maximum value for the lowest osmium coverage degree (θOs ) 0.15). This can be seen in Figure 3, where the intensity of the band at 2040 cm-1 from the spectra presented in Figure 1 was plotted as a function of the potential for each Pt-Os surface composition. Although the intensity of the COlinear band is only
Langmuir, Vol. 20, No. 25, 2004 11067
Figure 3. Band intensity of COlinear (2040 cm-1) from FTIR spectra for ethanol oxidation as a function of the potential (data obtained from Figure 1): on nonmodified Pt(100) and on Pt(100) modified with different osmium coverage degrees.
Figure 4. Band intensity of CO2 (2340 cm-1) from FTIR spectra for ethanol oxidation as a function of the potential (data obtained from Figure 1): on nonmodified Pt(100) and on Pt(100) modified with different osmium coverage degrees.
slightly affected at lower osmium coverage degrees, its initial formation on all Pt-Os electrodes is clearly shifted to a potential ca. 50 mV lower than that observed for nonmodified Pt(100). With osmium coverage degrees of 0.25-0.40 the surfaces display a curious feature: the coverage of COlinear does not change significantly over the composition range studied, and the quantity of this species decreases slowly from 0.5 to 0.9 V, in comparison with other Pt-Os compositions (see the slopes of the curves in Figure 3). The adsorbed COlinear remains on the surfaces with 0.25 < θOs < 0.40 until potentials up to ca. 0.8 V are reached, which is different from the behavior observed for nonmodified Pt(100) and Pt(100) with low osmium coverages. On these last ones, the COlinear is totally removed from the surface before the potential of 0.8 V is reached. At osmium coverage degree of 0.92, the quantity of linearly bonded CO on the surface is significantly lower than those observed for coverage degrees in the range from 0 to 0.40. As can be seen in Figure 4, where the intensity of the band at 2340 cm-1 is plotted as a function of the potential for Pt(100) and for each Pt-Os system, the potential of the onset of CO2 formation (ca. 0.5 V) does not vary significantly for the different compositions studied. Although the quantity of CO2 observed at more positive potentials is only slightly affected by the ratio of each metal on the surface over a wide range of compositions, the curve for CO2 reaches a maximum value for the lowest osmium coverage degree (θ ) 0.15 at 0.75 V). It is also observed that the amount of CO2 formed falls drastically for the almost complete osmium layer, in comparison with other Pt-Os compositions. The most significant enhancement of the catalytic activity of Pt(100) for the complete oxidation of adsorbed
11068
Langmuir, Vol. 20, No. 25, 2004
Pacheco Santos et al.
Figure 6. Coverage degrees of osmium deposits on Pt(100) as a function of the spontaneous deposition times, estimated from the hydrogen adsorption-desorption peaks in the voltammograms in HClO4 solution.
Figure 5. Cyclic voltammograms in 0.1 M HClO4 of nonmodified Pt(100) and Pt(100) modified with different osmium coverage degrees, obtained by spontaneous depositions (2-150 s), and an almost complete osmium layer obtained by electrolysis at 50 mV. (s) Nonmodified Pt(100).
ethanol to CO2 occurs for the lowest osmium coverage degree (θOs ) 0.15). This can be related to the capability of this surface to cleave the C-C bond at lower potentials than those for nonmodified Pt(100). This leads to the preferential formation of CO as an intermediate adsorbed from the ethanol oxidation reaction, and hence CO2 is the main product. However, voltammetric results, discussed in the following section, indicate that ethanol electrooxidation is improved on Pt(100) as the osmium coverage degree is increased up to ca. 0.33. This suggests that this reaction can follow other mechanism pathways for higher osmium coverage degrees. 3.2. Voltammetric Study. The voltammetric profiles in 0.1 M HClO4 of nonmodified Pt(100) and Pt(100) with different osmium coverage degrees obtained after spontaneous deposition times of 2-150 s and with an almost complete osmium layer obtained by electrodeposition at 50 mV are shown in Figure 5. The curves in Figure 5 show the increase of the anodic wave attributed to the oxidation of the osmium deposits (around 0.65 V 27,28) as a function
of the osmium coverage degree growth, parallel to the diminishing of the hydrogen adsorption-desorption peaks (from ca. 0.2 to 0.5 V). In this work, the osmium coverage degrees were estimated from the diminishing of the hydrogen peaks with the increase of the deposition time, according to the method described in previous studies.27,28 The curve of the coverage degree of osmium deposits as a function of the spontaneous deposition times is shown in Figure 6. In Figure 6, it is possible to observe that, on Pt(100), the maximum values of osmium coverage obtained by spontaneous deposition from an H2OsCl6/HClO4 solution are around 0.40. These values are lower than those obtained on Pt(100) from an H2OsCl6/H2SO4 solution (θOs ≈ 0.65).27,28 However, to give additional information about the feature of an almost complete osmium layer in the catalytic activity for ethanol oxidation, the results for the coverage degree of 0.92 (obtained by electrolysis at 50 mV) were included in this work. After each osmium deposition, the voltammograms of ethanol oxidation in EtOH + HClO4 solution were recorded. The voltammetric profiles (first positive scan) of ethanol oxidation on nonmodified Pt(100) (solid line in all voltammograms), on Pt(100) modified by different coverage degrees of spontaneously deposited osmium, and on Pt(100) covered with an almost complete osmium layer obtained by electrodeposition, are presented in Figure 7. The voltammetric profile of ethanol oxidation on bare Pt(100) exhibits a small anodic wave that commences around 0.5 V and extends to ca. 0.7 V, at which point an intense peak can be observed. This peak presents the maximum at ca. 0.75 V and can be related to CO2 formation due to its proximity with that of the maximum intensity of the band at 2340 cm-1 in the FTIR spectra, parallel to the fast consumption of COlinear. In agreement with the FTIR results, the anodic wave observed in the voltammograms between 0.5 and 0.7 V can be related to the oxidation of the remaining CObridge to CO2. Additionally, the amount of COlinear on the surface does not vary significantly between ca. 0.5 and 0.65 V, despite the formation of CO2 having already commenced. As the osmium coverage degree increases, the anodic wave between ca. 0.5 and 0.7 V is enhanced (see Figure 7), parallel to the gradative diminishing of the peak around 0.8 V (attributed to the ethanol oxidation on free Pt(100) sites). The anodic wave below 0.7 V increases until reaching a maximum intensity for osmium coverage of ca. 0.33. The fact that the amount of CO2 formed remains almost unchanged over the range of osmium coverage from 0.25 to 0.40, according to FTIRS results, suggests that most of the increase in the anodic wave between 0.5 and 0.7 V,
Ethanol Electrooxidation on Pt(100)
Figure 7. Cyclic voltammograms of the first positive ethanol oxidation scan (0.5 M C2H5OH + 0.1 M HClO4) at 20 mV s-1: on nonmodified Pt(100) and Pt(100) modified with different osmium coverage degrees, obtained by spontaneous depositions (2-150 s), and on an almost complete osmium layer obtained by electrolysis at 50 mV. (s) EtOH oxidation on nonmodified Pt(100).
observed for θOs ) 0.33 by cyclic voltammetry, can be related to other ethanol oxidation reaction pathways, mainly the direct oxidation of ethanol to acetaldehyde and acetic acid. This will be discussed further in the next section. The voltammogram of ethanol electrooxidation on Pt(100) with an almost complete osmium layer (θ Os ) 0.92) shows only a small anodic wave, due to the well-known low activity of the osmium surface for the ethanol oxidation reaction. 3.3. In Situ FTIRS Study. Part II. To obtain additional information about the probable intermediates formed during the first series of positive potential steps of ethanol oxidation on the Pt-Os electrodes studied, FTIR spectra were obtained in another range of wavelengths in order to observe other bands. An important band for determining intermediates of the ethanol oxidation is at 933 cm-1,
Langmuir, Vol. 20, No. 25, 2004 11069
which is attributed to the presence of acetaldehyde on the electrode.41 The band at 1710 cm-1 can be related to the carbonyl group of both acetaldehyde and acetic acid and cannot be used to draw conclusions. In this section it was necessary to acquire a group of spectra at wavenumbers below the limit of 1000 cm-1, and this was only possible with the use of a ZnSe window. As for the FTIR spectra obtained in part I, collections of spectra were recorded as a function of the potential from 0.1 to 1.0 V for each Pt-Os surface composition obtained by spontaneous osmium deposition (θOs up to 0.40) and on an almost complete osmium layer (θOs ) 0.92) obtained by electrodeposition at 50 mV. Selected examples are shown in Figure 8. It can be clearly observed that the O-H band from the COOH group (1290 cm-1), the C-C-O band from the CHO group (933 cm-1), and the CdO band from both acetic acid and acetaldehyde (1710 cm-1) increase significantly on Pt(100) with θOs ) 0.33 and become very low on the almost complete osmium layer. In the latter case, the catalytic activity for ethanol oxidation decreases due to the diminishing of the active sites for ethanol adsorption, as discussed previously.27 The consumption of ethanol from the thin electrolyte layer, within the potential range studied, can be confirmed through the negative-going of the C-H bands in 2980 and 2900 cm-1, related to CH3 and CH2 groups of the ethanol, respectively.7,12,14 Selected FTIR spectra in this wavenumber range are shown in Figure 9, which is evidence that the higher values of the C-H bands in the negativegoing is observed for Pt(100) with osmium coverage degree of 0.33, mainly at potentials in which the formation of acetaldehyde and acetic acid takes place (above 0.5 V). This indicates that this surface is more active for ethanol oxidation than other Pt-Os compositions studied, in agreement with the FTIR spectra in Figure 8. For Pt(100) with θOs ) 0.92, these bands diminish significantly and become smaller than those observed for nonmodified Pt(100). This is also in agreement with the FTIR spectra in Figure 8, in relation to the low activity of the complete osmium layer for ethanol oxidation. A detailed study of the influence of osmium coverage on the quantities of acetaldehyde and acetic acid formed can be seen in the curves of the band intensity related to these species, as a function of the electrode potential in Figures 10 and 11. In Figure 10, the intensity of the band related to the formation of acetaldehyde (933 cm-1) was followed during the first series of positive potential steps for all of the Pt-Os surfaces studied. The FTIR spectra show that the formation of acetaldehyde commences around 0.45 V. As the osmium coverage increases, the intensity of the band at 933 cm-1 increases rapidly in the potential range of 0.5-0.75 V. At higher potentials the band intensity starts to decrease. The highest values of the band related to acetaldehyde formation between 0.5 and 0.75 V are obtained on Pt(100) with osmium coverages of ca. 0.33-0.40. This is in agreement with the electrochemical results, which detail the influence of acetaldehyde production on the increase of the anodic wave from 0.5 to 0.7 V, in the voltammetric profiles of ethanol oxidation on Pt-Os electrodes (Figure 7). Moreover, at this potential range (ca. 0.65 V), the arising of the anodic wave associated with the oxidation of the osmium deposits (see Figure 5) can also influence the direct ethanol oxidation to acetaldehyde. This suggests that the direct ethanol oxidation can be promoted by oxide species formed on osmium deposits at ca. 0.65 V, in accordance with the bifunctional mechanism.31
11070
Langmuir, Vol. 20, No. 25, 2004
Pacheco Santos et al.
Figure 8. In situ FTIR spectra (in the wavenumber range of acetaldehyde and acetic acid bands obtained with a ZnSe flat window, from 256 scans at 8 cm-1 resolution) of the first series of positive potential steps for ethanol oxidation (0.5 M C2H5OH + 0.1 M HClO4): on nonmodified Pt(100), on Pt(100) modified with spontaneously deposited osmium (θOs ) 0.33), and on an almost complete osmium layer (θOs ) 0.92).
Figure 9. In situ FTIR spectra (in the wavenumber range of C-H bands at ca. 2980-2900 cm-1, from 256 scans at 8 cm-1 resolution) of the first series of positive potential steps for ethanol oxidation (0.5 M C2H5OH + 0.1 M HClO4): on nonmodified Pt(100), on Pt(100) modified with spontaneously deposited osmium (θOs ) 0.33), and on an almost complete osmium layer (θOs ) 0.92).
According to FTIRS results presented here, the quantities of acetaldehyde observed on nonmodified Pt(100) and
on the electrode with the lowest osmium coverage degree (see Figure 10) are very low and increase only slightly
Ethanol Electrooxidation on Pt(100)
Figure 10. Band intensity of acetaldehyde (933 cm-1) from FTIR spectra for ethanol oxidation as a function of the potential (data obtained from Figure 8): on nonmodified Pt(100) and on Pt(100) with different osmium coverage degrees.
Figure 11. Band intensity of acetic acid (1290 cm-1) from FTIR spectra for ethanol oxidation as a function of the potential (data obtained from Figure 8): on nonmodified Pt(100) and on Pt(100) with different osmium coverage degrees.
until the maximum value at ca. 0.9 V is observed. This feature is also observed for the electrode with the almost complete osmium layer. The variation of the acetic acid band (1290 cm-1) as a function of the potential for each Pt-Os electrode can be seen in Figure 11. It can be observed that this band commences around 0.55 V but increases significantly only above 0.7 V. This fact suggests that most of the anodic wave between 0.5 and 0.7 V in the ethanol oxidation voltammograms on Pt-Os electrodes can be related to the formation of acetaldehyde, which is converted to acetic acid at higher potentials. Moreover, in the same way as that for the acetaldehyde band, the intensity of the acetic acid band reaches the highest values on Pt(100) with osmium coverage degrees of 0.33-0.40. For an almost complete osmium layer (θOs ) 0.92), the acetic acid band arises at lower potentials (ca. 0.4 V) than that observed for the acetaldehyde band, as can be confirmed in Figure 11. This fact suggests that this surface favors the direct oxidation of adsorbed ethanol to acetic acid. Furthermore, the absence of the CObridge band and the low intensity of COlinear, CO2, and CH3CHO bands support the assumption above. As opposed to acetaldehyde, the acetic acid band at 1290 cm-1 does not refer only to the adsorbed species, but to the cumulative amount of species formed during the positive potential steps. This includes the acetic acid diffused into the thin electrolyte layer. Due to this, the intensity of the acetic acid band in the FTIR spectra at potentials above 0.75 V (Figure 11) is much higher than that observed for acetaldehyde in 933 cm-1, shown in Figure 10. Due to the fact that different species (CObridge, COlinear, CO2, CH3CHO, and CH3COOH) are detected during ethanol oxidation, according to the FTIRS results pre-
Langmuir, Vol. 20, No. 25, 2004 11071
sented in this study, at least three oxidation steps should be involved in this reaction. Thus, based on previous works12,14 and on the present studies, a simplified mechanism shown in Figure 12 can be proposed within the potential and osmium coverage ranges considered here. Upon chemisorption, the ethanol molecule can dissociate to form adsorbed CO and CHx fragment (pathway 1). The total amount of formed CO molecules can adsorb as CObridge (given for the R coefficient) and as COlinear (the 1 - R coefficient). A fraction (φ) of the adsorbed CObridge remains unchanged at low potentials, while the rest (1 - φ) is converted to COlinear. Thus, most of the CO2 formed come from the direct oxidation of the remaining (φR) CObridge and from the oxidation of [(1 - R) + (1 - φ)R ) (1 - φR)] COlinear. The other part of the CO2 comes from the CHx partly converted to CO(ads) (the β coefficient in pathway 1). According to the mechanism, the global amount of carbon dioxide formed in the ethanol oxidation is given for (1 + β) CO2. The remained CHx(ads) is given for the (1 - β) coefficient. At low potentials and low osmium coverage degrees the pathway 1 is enhanced, suggesting that osmium promotes the C-C bond cleavage and the oxidation of CHx species to CO. Considering the low osmium coverage and the small amount of osmium oxide formed on Pt(100)/Os at this potential range, according to a previous study,28 an electronic effect may also play a role in the catalytic activity of Pt-Os surfaces for the ethanol oxidation at lower potentials. Pathway 2 involves the formation of adsorbed acetaldehyde, which is oxidized to acetic acid at higher potentials. This pathway is favored at higher osmium coverage degrees (0.33 < θOs < 0.40) and/or at higher potentials, in comparison with that observed for lower osmium coverages and lower potentials. On the other hand, the quantities of COlinear and CO2 do not diminish in this range of osmium coverages, when compared with those obtained for θOs ) 0.28. In accordance with a previous work,28 the average size of osmium nanoislands (ca. 4-8 nm2 for θOs ≈ 0.16) diminishes with the increase of osmium coverage degree to a value of ca. 1-4 nm2 for θOs ≈ 0.45. This is due to the fact that for higher osmium coverage degrees the increase in the number of smaller islands is much more significant than the increase observed in the number of larger islands. Thus, the number of active sites for ethanol oxidation on the surface with higher osmium coverage degrees is high enough to enhance significantly the catalytic activity of the electrode for the direct oxidation to acetaldehyde and acetic acid. Moreover, the small islands preferentially formed at θOs ≈ 0.45 exhibit high values of the perimeter in comparison with their area. This feature favors the organic molecule oxidation, which is well-known to occur in the island edges.27 On the almost complete osmium layer, the presence of acetic acid at lower potentials than those observed for acetaldehyde formation, as well as the absence of the CObridge band and the low intensity of COlinear, CO2, and CH3CHO bands, support the enhancement of the direct oxidation of adsorbed ethanol to acetic acid (pathway 3) on this surface. 4. Conclusions FTIR spectra of ethanol electrooxidation on Pt(100) modified by osmium nanodeposits show that the bands of the main intermediates and products of the reaction, CObridge, COlinear, and CO2, increase significantly at lower potentials for the surfaces with the lowest osmium
11072
Langmuir, Vol. 20, No. 25, 2004
Pacheco Santos et al.
Figure 12. Scheme of the simplified mechanistic pathways proposed for the ethanol electrooxidation on Pt(100)/Os.
coverage. This significant enhancement of the activity of Pt(100) for the complete oxidation of adsorbed ethanol to CO2 at low osmium coverages can be related to the capability of this new surface to cleave the C-C bond at lower potentials than those observed for nonmodified Pt(100). Analysis of the band at 933 cm-1 suggests that the direct oxidation of ethanol to acetaldehyde is most favored on the surfaces with osmium coverages of 0.33-0.40. This supports the electrochemical results, in which a significant increase in the anodic current between 0.5 and 0.7 V on the Pt-Os surface with θOs ) 0.33 seems to be related to direct ethanol oxidation to acetaldehyde and acetic acid. The fact that the global quantity of CO2 remains almost unchanged in the 0.25 < θOs < 0.40 surface composition range corroborates the arguments above. On the almost complete osmium layer, the presence of acetic acid at lower potentials than those observed for acetaldehyde corroborates with the direct oxidation of ethanol to acetic acid.
These features of the Pt-Os electrodes studied in this work demonstrate that the improvement of the reactivity toward ethanol oxidation to CO2 on the surfaces with lower osmium coverage degrees can be attributed to their ability to cleave the C-C bond at low potentials. However, with the increase of osmium coverage and potential, a significant increase in the number of islands with small dimensions, as well as the higher presence of osmium oxide, favor the direct oxidation of ethanol rather than cleavage of its C-C bond. Acknowledgment. The authors acknowledge the Brazilian funding bodies FAPESP (Processes 01/08087-0 and 02/11007-1) and CNPq for financial support. They also thank Elisete A. Batista for help in performing the FTIRS experiments and Geoffroy R. P. Malpass for grammatical revision of the text. LA040001V
