Adsorption of Formaldehyde on Pt (111) and Pt (100) Electrodes

On both Pt electrodes, HCHO oxidation commenced preferentially at step sites at the onset potential of this reaction, but it occurred uniformly at the...
0 downloads 0 Views 866KB Size
4964

Langmuir 2005, 21, 4964-4970

Adsorption of Formaldehyde on Pt(111) and Pt(100) Electrodes: Cyclic Voltammetry and Scanning Tunneling Microscopy Chen-Fu Mai,† Chia-Haw Shue, Yaw-Chia Yang, Liang-Yueh Ou Yang,†,‡ Shueh-Lin Yau,*,§ and Kingo Itaya*,‡,§ Department of Chemistry, National Central University, Chungli, Taiwan 320, Department of Applied Chemistry, Graduate School of Engineering, Tohoku University, Aoba-yama 04, Sendai 980-8579, Japan, and CREST, JST, 4-1-8 Kawaguchi, Saitama 332-0012, Japan Received October 28, 2004. In Final Form: February 26, 2005 The adsorption of formaldehyde (HCHO) on Pt(111) and Pt(100) electrodes was examined by cyclic voltammetry (CV) and in situ scanning tunneling microscopy (STM) in 0.1 M HClO4. The extent of HCHO adsorption at both Pt electrodes was evaluated by comparing the CVs, particularly for the hydrogen adsorption and desorption between 0.05 and 0.4 V, obtained in 0.1 M HClO4 with and without HCHO. The adsorption of HCHO on these Pt electrodes was significant only when [HCHO] g 10 mM. Adsorbed organic intermediate species acted as poisons, blocking Pt surfaces and causing delays in the oxidation of HCHO. Compared to Pt(111), Pt(100) was more prone to poisoning, as indicated by a 200 mV positive shift of the onset of HCHO oxidation. However, Pt(100) exhibited an activity 3 times higher than that of Pt(111), as indicated by the difference in peak current density of HCHO oxidation. Molecular resolution STM revealed highly ordered structures of Pt(111)-(x7 × x7)R19.1° and Pt(100)-(x2 × x2) in the potential region between 0.1 and 0.3 V. Voltammetric measurements further showed that the organic poisons produced by HCHO adsorption behaved differently from the intentionally dosed CO admolecules, which supports the assumption for the formation of HCO or COH adspecies, rather than CO, as the poison. On both Pt electrodes, HCHO oxidation commenced preferentially at step sites at the onset potential of this reaction, but it occurred uniformly at the peak potentials.

Introduction The simple molecular structure of formaldehyde (HCHO) has prompted its use as a model to gain insight into the processes of the electrooxidation of small organic molecules at Pt electrodes, a subject of long-term interest in the development of fuel cell technology.1 The dual mechanism, originally proposed to account for the oxidation of methanol at Pt electrodes, was also applied to explain the results of HCHO oxidation at Pt electrodes.2 More specifically, HCHO is oxidized to CO2 either directly or indirectly through the formation of an intermediate of formyl (HCO), COH, or carbon monoxide.1 The intermediate acts as a poison, as it blocks the active sites on Pt surfaces.1 The chemical nature of the poisons derived from HCHO has been examined, but not all studies lead to consistent conclusions, mainly because different techniques were employed in experiments performed under different conditions.2-5 Possible intermediates generated in HCHO oxidation include formyl (HCO),3 COH,4,5 and CO,6,7 but the idea of poisoning by CO has been prevalent.1 * To whom correspondence should be addressed. Telephone: 81-22-2174177. Fax: 81-22-2174177. E-mail: yausl@ atom.che.tohoku.ac.jp (S.-L.Y.). † National Central University. ‡ Tohuku University. § JST. (1) Sun, S. G. Studying Electrocatalytic Oxidation of Small Oragnic Molecules with In Situ Infrared Spectroscopy. In Electrocatalysis; Lipkowski, J., Ross, P. N., Eds.; Wiley-VCH: New York, 1988; p 243. (2) Parsons, R.; VanderNoot, T. J. Electroanal. Chem. 1988, 257, 9. (3) Kamath, V. N.; Lal, H. J. Electroanal. Chem. 1968, 19, 137. (4) Bagotzky, V. S.; Vassiliev, Yu. B. Electrochim Acta 1966, 11, 1439. (5) Kazarinov, V. E.; Vassiliev, Yu. B.; Andreev, V. N.; Kuliev, S. A. J. Electroanal. Chem. 1981, 123, 345.

Most studies performed prior to 1980 employed polycrystalline Pt electrodes with ill-defined atomic structure, which makes it difficult to gain insight into the structural effect in electrocatalysis. In the early 1980s a simple method was devised for preparing Pt single-crystal electrodes with controllable surface defects of kinks and steps. This method has enabled the examination of electrocatalysis on the atomic level.8 Quite often, singlecrystal Pt electrodes with more open atomic structures such as Pt(110) exhibit a higher electrocatalytic activity, but they are more prone to poisons than electrodes with a close-packed atomic structure such as Pt(111).1,8 Research in the past decade focused attention on the kinetics of methanol oxidation, whereas a few studies dealt with formaldehyde at Pt single-crystal electrodes.9-11 Sun et al. showed that the oxidation of HCHO on Pt(111) produces CO species bonded at atop and 2-fold bridge sites, whereas the adsorption from CO-saturated perchloric acid yielded only terminal CO.9 Olivi et al. reported IR results on the adsorption of HCHO on Pt(100) and Pt(110) electrodes and identified linear and multibonded CO on Pt(100), but only terminal CO on Pt(110).10 More recently, the second harmonic generation (SHG) technique was used to examine Pt(111) in HCHO-containing HClO4 under gal(6) Sideswaran, P.; Lal, H. J. Electroanal. Chem. 1972, 123, 143. (7) Nishimura, K.; Ohnishi, R.; Kunimastu, K.; Enyo, M. J. Electroanal. Chem. 1989, 258, 219. (8) Clavilier, J.; Rodes, A.; El Achi, K.; Zamakhchari, M. A. J. Chim. Phys. 1991, 88, 1291. (9) Sun, S.-G.; Lu, G.-Q.; Tian, Z.-W. J. Electroanal. Chem. 1995, 393, 97. (10) Olivi, P.; Bulhoes, L. O. S.; Leager, J.-M.; Hahn, F.; Beden, B.; Lamy, C. J. Electrochim Acta 1996, 41, 927. (11) Mishina, E.; Karantonis, A.; Yu, Q. K.; Nakabayashi, S. J. Phys. Chem. 2002, 106, 10199.

10.1021/la047342t CCC: $30.25 © 2005 American Chemical Society Published on Web 04/22/2005

Adsorption of HCHO on Pt Electrodes

Langmuir, Vol. 21, No. 11, 2005 4965

vanostatic and potentiodynamic conditions,11 revealing the phenomena of potential oscillation and an orderdisorder phase transition of adspecies. However, this SHG study did not identify the real-space arrangement of adspecies. Herein, we report results of a study of the adsorption of HCOH on Pt(111) and Pt(100) electrodes performed by using in situ scanning tunneling microscopy (STM) and cyclic voltammetry, which revealed the formation of two ordered adlattices, (x7 × x7)R19.1° and (x2 × x2), on Pt(111) and Pt(100), respectively. The adsorbates on both electrodes caused delays in HCHO oxidation in a manner different from an intentionally dosed CO adlayer. Our results indicate that poisons other than CO are formed when HCHO was adsorbed on Pt electrodes in the potential range between 0.1 and 0.3 V in 0.1 M HClO4. Experimental Section The procedure for preparing and treating Pt single-crystal bead electrodes followed the well-known annealing and quenching method.8 Kibler et al. examined how the pretreatment process affects the surface morphology and atomic structures of the Pt(111) and Pt(100) electrodes.12 It is shown that the environment in which the Pt electrodes are quenched determines the outcome, particularly for Pt(100) because its surface can reconstruct itself into a complicated hex structure. According to that report, the reduction environment containing hydrogen or CO gas is beneficial for the formation of an unreconstructed (1 × 1) structure.12 If Pt(100) is quenched in hydrogen stream, the surface always contains mesas whose dimensions are determined by the pressure of hydrogen. In addition, the annealing and quenching procedure always produces a monolayer thick thermal oxide on the Pt electrode. To expose a Pt(100) surface for studying HCHO adsorption, this oxide layer had to be removed by applying a potential of 0.05 V in 0.1 M HClO4. Ultrapure perchloric acid and formaldehyde were purchased from Merck (Darmstadt, Germany) and Aldrich (Saint Louis, MO), respectively. Note that commercially available formaldehyde solution contains about 10% (by volume) methanol, which is added intentionally to stabilize the HCHO. Formaldehyde is known to hydrate to produce methylene glycol, CH2(OH)2, with an equilibrium constant of about 2300 in favor of the hydrated form.13 The STM used was a Nanoscope-E (Santa Barbara, CA), and the tip was made of tungsten (diameter 0.3 mm) prepared by electrochemical etching in 2 M KOH. Details of STM imaging procedure were described previously.14,15 We always used the constant-current mode in conducting STM imaging experiments. A reversible hydrogen electrode (RHE) was used as the reference electrode in the electrochemical and STM experiments, and all potentials in this paper refer to the RHE.

Results and Discussion Cyclic Voltammetry (CV). Pt(111) in HCHO-Containing Perchloric Acid. The effect of HCHO concentration on its adsorption on Pt(111) was first investigated by using voltammetry. The initial contact between a Pt electrode and HCHO was always made under potential control at 0.05 V, where HCHO is least likely to be oxidized. Figure 1a and 1b show CVs recorded in the first positive potential sweep from 0.05 to 1.0 V at the potential scan rate of 50 mV/s at Pt(111) electrodes in 0.1 M HClO4 containing 1 and 10 mM HCHO, respectively. The dotted trace in Figure 1a is a steady-state voltammogram of Pt(111) in 0.1 M HClO4. The positive potential scan in 1 mM HCHO yielded (12) Kibler, L. A.; Cuesta, A.; Kleinert, M.; Kolb, D. M. J. Electroanal. Chem. 2000, 484, 73. (13) Doherty, A. P.; Christensen, P. A.; Hamnett, A.; Scott, K. J. Electroanal. Chem. 1995, 386, 39. (14) Itaya, K. Prog. Surf. Sci. 1998, 58, 121. (15) Zang, Z. H.; Wu, J. L.; Yau, S. L. J. Phys. Chem. B 1999, 103, 9624.

Figure 1. Cyclic voltammograms of Pt(111) in 0.1 M HClO4 containing (a) 1 mM HCHO, (b) 10 mM HCHO, and (c) 1 mM CH3OH. The CVs in (d) and (e) were obtained with CO-modified Pt(111) in 0.1 M HClO4 without and with 10 mM HCHO. The scan rate was 50 mV/s, except (b) was recorded at 10 mV/s.

peak A1 at 0.45 V, presumably due to the oxidation of methylene glycol, the hydrated form of HCHO. Another peak, A2, at 0.7 V superimposed on the butterfly feature is known to be associated with the redox process Pt + H2O f PtOH + H+ + e-, occurring only at a long range ordered Pt(111) electrode surface.16 The origin of A2 is somewhat more complicated. It could be due to the oxidation of CO produced in A1, but it appeared at a potential 200 mV more negative than that of surface-bound CO molecules, denoted CO(s). It is unlikely to be the oxidation of HCHO, because [HCHO], roughly 0.41 µM in 1 mM HCHO, was too low to produce a prominent peak like A2. Instead, it is ascribed to the oxidation of methanol, the stabilizer added in commercial HCHO. Indeed, the i-E profile of Pt(111) in 0.1 M HClO4 containing 1 mM CH3OH (Figure 1c) contains a pronounced peak at 0.7 V. The adsorption of HCHO was not important, as it hardly affected the hydrogen characteristics between 0.05 and 0.35 V in 1 mM HCHO (Figure 1a). Raising [HCHO] by 10 times produced rather different CV profiles. The positive potential scan shown in Figure 1b contains no peak at 0.45 V, which is observed in 1 mM HCHO (Figure 1a). Instead, the oxidation of methylene glycol produced a broad oxidation wave spanning from 0.5 to 0.9 V, and the hydrogen characteristic between 0.05 and 0.35 V was eliminated. Continuous potential sweeps between 0.05 and 1.0 V caused no change in the morphology of the CV profile, although the effect of concentration polarization gradually reduced the current density. These (16) Wagner, F. T.; Ross, P. N. Surf. Sci. 1985, 160, 305.

4966

Langmuir, Vol. 21, No. 11, 2005

Mai et al.

Figure 2. Cyclic voltammograms of Pt(111) in 0.1 M HClO4 containing 0.1, 1, 10, and 100 mM HCHO. The scan rate was 50 mV/s.

results indicate that HCHO was adsorbed when [HCHO] g 10 mM. Pt(111) regained its activity at potentials positive of 0.4 V, but the activity suffered again from the formation of an OH adlayer at E > 0.7 V, as the current density dropped to a minimum at the end of the butterfly feature, and no current flowed in the negative potential scan until the OH adlayer was reduced at 0.9 V. These voltammetric results are mostly consistent with those reported by Olivi et al.10 Because CO could be the possible cause of the passivation, we examined the electrochemistry of CO(s) and its effect on the electooxidation of HCHO. Shown in Figure 1d is a positive potential sweep from 0.05 to 1.0 V at a Pt(111) electrode intentionally dosed with a monolayer of CO. This CV shows a broad and weak oxidation peak at 0.6 V, a sharp current spike at 0.7 V, and a shoulder at 0.8 V. These features are ascribed, respectively, to the phase transition of the CO adlayer from (2 × 2) to (x19 × x19)R23.4°, the oxidation of CO(s) to CO2, and the formation of an OH adlayer.15 It was noted that the peak potential of 0.72 V hardly varied with the coverage of CO(s). Next, we compared the CV shown in Figure 1b with that obtained at a Pt(111) electrode intentionally dosed with a monolayer of CO in 0.1 M HClO4 containing 10 mM HCHO. Experimentally, the Pt(111) electrode was dosed with CO at 0.1 V, followed by purging with N2 to remove the dissolved CO and adding HCHO to make its final concentration equal to 10 mM. The obtained CV at a scan rate of 50 mV/s is shown in Figure 1e. It is seen that the positive sweep yielded a featureless region between 0.05 and 0.9 V, followed by a current spike at 0.92 V, resulting from the oxidation of CO(s) to CO2. The removal of CO(s) released Pt sites, which became available for the oxidation of methylene glycol, producing a broad peak at 0.65 V in the negative potential scan. These results illustrate that the inhibitive effect of CO(s) and that of HCHO-derived adspecies on the electrocatalytic activity of Pt(111) are evidently different from each other, indicating the formation of species other than CO from the decomposition of HCHO. It is also intriguing to note that CO(s) was oxidized at a potential 200 mV more positive than in pure 0.1 M HClO4. We are unable to explain this result at present. Further voltammetric experiments with the potential window set between 0.05 and 0.4 V in 0.1 M HClO4 containing 0.001-0.1 M HCHO were conducted to elucidate the mechanism of adsorption of HCHO on Pt(111). The resultant CVs (Figure 2) obtained in 0.1 and 1 mM HCHO were nearly identical with that of a bare Pt(111). Adsorption of HCHO, as revealed by the change in hydrogen characteristics, became apparent when [HCHO] g 10 mM. Note that Olivi et al. concluded that HCHO, rather than CH2(OH)2, is the precursor for CO(s) on Pt electrodes.10 This effect of HCHO concentration on the extent of adsorption can be explained by the ratio of 1:106 for [HCHO]/[H+] in 0.1 M HClO4 containing 1 mM HCHO.

Figure 3. Cyclic voltammograms of Pt(100) in 0.1 M HClO4 containing (a) 1 mM HCHO, (b) 1 mM CH3OH, and (c) 10 mM HCHO. The CVs in (d) and (e) were obtained with CO-modified Pt(100) in 0.1 M HClO4 without and with 10 mM HCHO, respectively. The scan rate was 50 mV/s, except (c) was recorded at 10 mV/s.

The slight increase of current at potentials positive of 0.3 V in 0.1 M HCHO could be due to the commencement of methylene glycol oxidation. Finally, the adsorption of HCHO barely affected the evolution of hydrogen, as revealed by the precipitous increase of current at 0.05 V. As for CO(s), the hydrogen evolution reaction would be blocked at a potential as negative as -0.2 V. These CV results again disfavor the formation of CO(s) on Pt(111) even in 0.1 M HCHO. However, the situation might be different for Pt electrodes with different crystallographic orientations. Pt(100) in HCHO-Containing Perchloric Acid. Similar voltammetric experiments were performed with a Pt(100) electrode to explore the structural effect on the oxidation of HCHO and CO. The resultant CVs are summarized in Figure 3 with the dotted line in Figure 3a representing the background CV of a bare Pt(100) electrode in 0.1 M HClO4. The CV in the solid line of Figure 3a obtained in the presence of 1 mM HCHO reveals two peaks at 0.38 (A1) and 0.58 V (A2) in the positive potential scan from 0.05 to 0.9 V. As in the case of Pt(111), these features are ascribed to the oxidation of methylene glycol and methanol. Shown in Figure 3b is the CV of Pt(100) in 0.1 M HClO4 containing 1 mM methanol, where a prominent oxidation peak at 0.7 V is clearly seen. This corresponds to the A2 peak in Figure 3a. Raising [HCHO] to 10 mM or higher dramatically changed the i-E profile, showing a featureless region

Adsorption of HCHO on Pt Electrodes

Langmuir, Vol. 21, No. 11, 2005 4967

Figure 4. Constant-current in situ STM images of Pt(111) obtained at 0.1 V in 0.1 M HClO4 containing 10 mM HCHO. The imaging conditions were 100 mV in bias voltage and 5 nA in set-point current. The rhombus shown in (c) indicates a unit cell of (x7 × x7)R19.1°.

between 0.05 and 0.67 V, followed by a sharp current spike at 0.72 V in the positive potential scan from 0.05 to 1.1 V (Figure 3c). Because the sharp peak appeared at nearly the same potential as that for the oxidation of CO(s) (Figure 3d), it is probable that CO(s) is produced from the adsorption of HCHO. We investigated this possibility by running voltammetry with a CO-modified Pt(100) electrode in 0.1 M HClO4 + 10 mM HCHO. The resultant CV shown in Figure 3e indicates that the Pt(100) electrode was completely inactive toward HCHO oxidation until the poisons were removed at 0.85 V, which is 0.13 V more positive than the potential for the peaks in Figure 3b,e. This result suggests that the adsorption of HCHO on Pt(100) from 0.1 M HClO4 + 10 mM HCHO at 0.1 V did not produce CO(s), as was also found at Pt(111). The structure of Pt electrodes apparently dominated their interaction with HCHO and hence the extent of poisoning, as indicated by the onset potential for the features associated with oxidation reaction in the CVs (Figures 1b and 3c). Furthermore, these CVs reveal that Pt(111) and Pt(100) produced different current densities (0.8 and 1.7 mA/cm2, respectively) of the feature associated with the oxidation of methylene glycol. These results indicate that Pt(100) is more prone to poisoning but is more active than Pt(111), as noted by others.1,10 This structural effect could stem from the strength of adsorption of organic poisons and the ease with which it produces oxygen species needed in the oxidation reactions (vide infra). In Situ Scanning Tunneling Microscopy (STM). Ordered Structures of Adsorbate on Pt(111) in HCHOContaining Perchloric Acid. The typical surface state of an as-prepared Pt(111) electrode at 0.1 V in 0.1 M HClO4, as revealed by in situ scanning tunneling microscopy (STM), contained atomically flat terraces measuring 5001000 Å delineated by monatomic steps (∆z ) 2.5 Å). The hexagonal atomic lattice of Pt(111) with a nearest-neighbor spacing of 2.8 Å was imaged also, as already reported.14,15 STM atomic resolution of the Pt(111) substrate allowed precise determination of the structures formed by HCHO adsorption. As was established by CV, HCHO concentration of at least 10 mM was needed to enable its adsorption at 0.1 V. There was a 10 min delay before the ordered adlattices of HCHO became visible, possibly because the adsorption of HCHO on Pt(111) was an activation process, involving the cleavage of a C-H bond in the HCHO molecule. The barrier for this reaction can be substantial, because of the strength of the C-H bond (bond energy).17 (17) Morrison, R. T.; Boyd, N. R. Organic Chemistry; Prentice-Hall: Englewood Cliffs, NJ, 1999.

The topographic STM scan shown in Figure 4a illustrates typical surface morphology of the Pt(111) electrode in 0.1 M perchloric acid containing 10 mM HCHO. In addition to the features of extended terraces, STM imaging revealed the presence of blotches protruding ca. 2.5 ( 0.3 Å on terraces with nearly identical size and shape. It is likely that they are images of identical species produced by chemical or electrochemical processes involving HCHO. For example, they could be paraformaldehyde, HO(CH2O)nOH, produced by the polymerization of HCHO, which is favored at low temperatures and low pH values.13 Attempts to resolve atomic features on these blotches were unsuccessful. The STM scans in Figure 4b,c are closeup views of the long-range ordered adlattice on terraces and the internal molecular arrangement within this adlayer. (The image in Figure 4c was filtered by using a twodimensional Fourier transform technique to remove noise.) The unit cell of this ordered array, i.e., the rhombus in Figure 4c, was determined to be (x7 × x7)R19.1° (hereafter referred to as x7), as the two unit vectors are 7.4 Å in length and rotated 19° from the close-packed atomic rows of Pt(111). In addition to the prominent protrusions at the lattice points of this adlayer (or the four corners of the rhombus), two weaker features appearing at least 0.4 Å lower in elevation were discerned within the cell. Although adsorbed CO molecules also form a x7 structure on Pt(111),18 the present x7 structure is more likely due to COH or OCH, rather than CO.1,10 Ordered Adlattices on Pt(100) in 0.1 M HClO4 Containing 10 mM HCHO. The STM image of a bare Pt(100) electrode recorded at 0.1 V prior to the addition of HCHO revealed that the protruding islands (∆z ) 2.5 Å) are due to aggregates of Pt atoms.12,19-22 The small clusters with an edge length of 10 Å or with an area less than 100 Å2 were poorly defined in shape, but the larger ones were clearly square or rectangular. They were monatomic in height and aligned with their edges parallel to the closepacked atomic directions of Pt(100), which is possibly driven by the formation of (111) oriented microfacets at the borders between the islands and the underlying Pt(100) substrate. As reported by Kibler et al., the density of protruding islands varied with the pressure of H2 in the cooling process of Pt(100).12 Reducing the number of islands improves substantially the quality of STM resolution. (18) Villegas, I.; Weaver, M. J. J. Chem. Phys. 1994, 101, 1648. (19) Villegas, I.; Weaver, M. J. J. Phys. Chem. 1991, 95, 7559. (20) Villegas, I.; Weaver, M. J. J. Electroanal. Chem. 1994, 373, 245. (21) Chang, S. C.; Weaver, M. J. J. Phys. Chem. 1991, 95, 5391. (22) Shue, C. H.; Yau, S. L. J. Phys. Chem. B 2001, 105, 5489.

4968

Langmuir, Vol. 21, No. 11, 2005

Mai et al.

Figure 5. Constant-current in situ STM images of Pt(100) obtained at 0.1 V in 0.1 M HClO4 containing 10 mM HCHO. The imaging conditions were 100-200 mV in bias voltage and 5-10 nA in feedback current. The square in (c) represents the unit cell of Pt(100)-(x2 × x2).

Figure 6. Time-dependent in situ STM images acquired at Pt(111) after its potential was stepped from 0.1 to -0.1 V in 0.1 M HClO4 containing 10 mM HCHO. All images were acquired consecutively at intervals of 1 min with image (a) obtained 3 min after the potential step. Scan sizes are all 2500 × 2500 Å. The (x7 × x7)R19.1° structures are highlighted.

With the potential held at 0.1 V, HCHO was added to give a final concentration of 10 mM in the STM cell. Highresolution STM scans (Figure 5b,c) reveal the formation of an ordered square array with the nearest-neighbor spacing of 4 Å, which corresponds to x2 times the Pt-Pt interatomic spacing. Meanwhile, the close-packed rows of protrusions are aligned in the [002] (or x2) direction. These results lead to a (x2 × x2)R45° (hereafter referred to as x2) structure with all protrusions in the STM image (Figure 5c) exhibiting an identical corrugation height. This structure appears often on Pt(100), and it has been found for adsorbates of iodine, sulfur, CN, and CO.18-21 It is likely that this structure is due to the adsorption of COH or HCO at sites with an identical symmetry on Pt(100). Despite the fact that 4-fold hollow sites are generally favored, it is difficult to determine exact registries of the adspecies from the STM image alone. For CO adsorbate, IR results indicate that its adsorption takes place predominantly on atop sites.20 Potential Effect on the Stability of Pt(111) -(x7 × x7)R19.1°. The potential of Pt(111) was altered to examine

the stability of the x7 adlattice. First, the potential was stepped from 0.1 to -0.1 V to see if the x7 structure sustains the potential-induced hydrogenation reaction. This experiment was expected to provide indirectly a chemical identity for the x7 structure. Being strongly adsorbed on Pt(111), CO(s), if it is indeed produced from HCHO, should remain stable after the potential is stepped from 0.1 to -0.1 V. Figure 6 displays time-sequenced STM images obtained after the potential was stepped from 0.1 to -0.1 V. Hydrogen evolved at -0.1 V caused some instability in the imaging, but fortunately it did not jeopardize this experiment. Typically, the STM was operated at 200 mV in bias voltage and 5 nA in feedback current. These images contain distinct blotches, ascribed to paraformaldehyde, HO(CH2O)nOH, produced by polymerization of HCHO. In addition, in situ STM revealed local x7 arrays as parallel lines (within the highlighted areas in the images) on terraces. The STM image in Figure 6a, acquired 3 min after the potential was held at -0.1 V, shows that the x7 structure covered roughly 50% of the Pt(111) surface,

Adsorption of HCHO on Pt Electrodes

Figure 7. Potential-dependent STM images of Pt(111) acquired in 0.1 M HClO4 containing 10 mM HCHO at 0.1 V (a), 0.4 V for 3 min (b), 0.4 V for 6 min (c), and 0.1 V 2 min after the potential step (d). Scan sizes 840 × 840 Å for (a) and 570 × 570 Å for (b), (c), and (d).

whereas the remainder (the dimmer areas) is believed to contain hydrogen. The x7 domains existed only on terraces: the edges and their vicinities were free of x7 structure. These results indicate that hydrogen adatoms were adsorbed preferentially near step ledges, displacing adspecies derived from HCHO. This displacement process proceeded swiftly to reduce the domain size of the x7 structure, which vanished completely after the potential was held at -0.1 V for 6 min. The rate of this process was potential-dependent. Applying a potential of -0.2 V reduced the duration of the displacement process to approximately 2 min. However, stepping the potential from -0.1 V back to 0.1 V restored the x7 ordered arrays, revealing the reversibility of this event. We propose that the organic adspecies, possibly formyl (-CHO) or COH, was hydrogenated to HCHO and desorbed at -0.1 V. This negative potential step from 0.1 and -0.1 V hardly affected characteristics of the unidentified blotches. The stability of the x7 structure at more positive potentials was also examined. Shown in Figure 7 are potential-dependent STM images of Pt(111), obtained after the potential was stepped from 0.1 to 0.4 V and back to 0.1 V. The image in Figure 7a, which was acquired at 0.1 V, shows a prominent x7 structure on terraces T1 and T2, and in trench H. The former contained the identical x7 structure, while the structure in H was a mirror image of those in T1 and T2. At potentials between 0 and 0.3 V, no change was observed in the x7 structure. However, once the potential was made more positive than 0.4 V, the x7 structure near the upper end of steps disappeared selectively. This is illustrated by the STM image in Figure 7b obtained 2 min after the potential was held at 0.4 V. Nearly all x7 structures disappeared after the potential was held at 0.4 V for 4 min. It seems that the adspecies was oxidized to CO2 or some unknown species, which dissolved into the electrolyte. Electrochemical potential controlled the rate of this phase transition, as stepping the potential from 0.1 to 0.5 V immediately transformed the x7 structure into a disordered adlayer. This potential-

Langmuir, Vol. 21, No. 11, 2005 4969

induced phase transition was reversible: stepping the potential from 0.5 V back to 0.1 V restored the highly ordered x7 structure within 2 min, as shown in Figure 7d. The effect of potential on the stability of the organic adlayer on Pt(100) was also examined by using in situ STM. Similar desorption and oxidation characteristics were observed at negative and positive potentials, respectively. The difference between Pt(111) and Pt(100) lies in the potential where they became active and the anodic current began to flow. The activation potentials were 0.4 and 0.7 V for Pt(111) and Pt(100), respectively. These two electrodes also exhibited different surface morphologies after disappearance of the x7 and x2 structures. The surface of Pt(111) appeared mostly disordered at E g 0.4 V, whereas the ordered atomic lattice of Pt(100) was observed after a brief period of fuzziness in the STM imaging. This contrast between Pt(111) and Pt(100) implies that the kinetics of HCHO oxidation are different at these two Pt electrodes. In particular, the adspecies was oxidized to CO2 at a faster rate on Pt(100), leaving no poison on the substrate, whereas they went through the formation of an intermediate species before being oxidized to CO2 on Pt(111). The relationship between catalytic activity and atomic structure of substrates in a vacuum is known to be determined by complicated interplays among geometric (spatial arrangements of atoms), thermodynamic (strength of adsorption), and electronic (tendency for bond formation and cleavage) factors. The situation at liquid-solid interfaces is even more complicated, because the omnipresent components such as water and ionic species frequently compete with interested reactants for active sites. In the context of HCHO/Pt(111) and Pt(100), it is possible that the thermodynamics for the adsorption of HCHO and water molecules could be of prime importance. More specifically, the CHO or COH intermediates could be more strongly bonded to Pt(100) than to Pt(111), resulting in the more pronounced poisoning, or the 200 mV positive shift in potential for the oxidation of methylene glycol at Pt(100). It is thought that the 4-fold coordination available on Pt(100) may create a more favorable binding environment for OCH or COH species than the 3-fold binding on Pt(111). The higher efficiency for HCHO oxidation at Pt(100) could be associated with its higher tendency to bind and dissociate water molecules than that on Pt(111). Finally, the structures of the adspecies, presumably CHO or OCH, will be discussed in relation to the ordered adlattices; i.e., x7 on Pt(111) and x2 on Pt(100), are discussed. Desai et al. conducted a theoretical study on the dehydration of methanol on Pt(111) in the gas phase, producing a formyl adsorbed intermediate with η2,η1-C,O configuration.23 Their theoretical study yielded the adsorption energy of -237 kJ/mol for formyl as compared to -49 and -168 kJ/mol for formaldehyde and CO on Pt(111), respectively. The formyl moiety may well be dehydrogenated into CO and H at the gas-solid interface of Pt(111), producing CO adspecies as the poison, as Desai et al. concluded. This pathway is considered to be unfavorable on the basis that the Pt electrodes were already covered with H adatoms. It is also noteworthy that an IR study of acetaldehyate on Pt(111) and Pt(100) also indicates the generation of strongly adsorbed species other than CO, such as η1-acetal or η1-acetaldehyde.24 The (23) Desai, S. K.; Neurock, M.; Kourtakis, K. J. Phys. Chem. B 2002, 106, 2559. (24) Rodriguez, J. L.; Pastor, E.; Xia, X. H.; Iwasita, T. Langmuir 2000, 16, 5479.

4970

Langmuir, Vol. 21, No. 11, 2005

Mai et al.

x7 unit cell, which is explained by their 2-fold bridge coordination, as described by Villegas et al.18 Finally, STM experiments with Pt(111) in 0.1 M HClO4 containing 10 mM methanol were performed to scrutinize the possible link between methanol adsorption and the ordered adlattices seen by STM. However, there was no evidence for the presence of the x7 structure on Pt(111). Thus, it is concluded that the 10% methanol included in formaldehyde was not responsible for the ordered structures observed on Pt electrodes in this study.

Figure 8. (a) In situ STM image of Pt(111) obtained at 0.1 V in 0.1 M HClO4 containing 10 mM HCHO. (b) In situ STM image of CO adlattice on Pt(111) in 0.1 M HClO4. Data taken from ref 18. The unit cells of (x7 × x7)R19.1° are indicated by the rhombus in the images.

STM image of each HCO moiety appearing as a single spot implies that it has a C-bonded, vertically oriented configuration on Pt(100). The lack of information on critical data such as coverage and spectra makes it difficult to offer a sensible interpretation of the x7 structure on Pt(111). One can only speculate that the CHO intermediate on Pt(111) is oriented in a manner similar to that on Pt(100), and that the difference in corrugation height among the spots in the STM molecular image (Figure 4c) resulted from the difference in their registries, with the brighter spots occupying atop sites and the dimmer spots residing on fcc and hcp sites. Identical x7 structures have been observed for iodine adatoms and iodobenzene admolecules on Pt(111), and the corresponding models have been described repeatedly.25,26 Meanwhile, we draw a comparison of the x7 structure observed in this study (Figure 8a) with that reported by Village et al. (Figure 8b).15 These two x7 structures are similar in appearances, but close examination reveals that the weak spots within the unit cell are located differently. They are situated on the edges of the (25) Schardt, B. C.; Rinaldi, F.; Yau, S. L. Science 1989, 243, 1050. (26) Chen, J. H.; Yau, S. L.; Chang, S. C. J. Phys. Chem. B 2002, 106, 9079.

Conclusions The adsorption and oxidation of HCHO on Pt(111) and Pt(100) electrodes were examined with voltammetry and in situ scanning tunneling microscopy (STM) under potential control. The adsorption of HCHO or the electrode poison was not important at [HCHO] lower than 10 mM. According to the results obtained in this study, CO was not the poisonous species on both Pt electrodes. The adspecies is thought to be C-bonded formyl (CHO) standing upright on both Pt electrodes. Molecular resolution STM shows that the adspecies are well ordered in Pt(111)-(x7 × x7)R19.1° and Pt(100)-(x2 × x2) with coverages of 0.43 and 0.5, respectively. The adspecies are adsorbed at atop and 3-fold sites on Pt(111), but they occupy only 4-fold hollow sites on Pt(100). Time- and potential-dependent STM imaging shows that the electrooxidation of HCHO at 0.4 V did not produce poisons, or it was oxidized completely to CO2. In addition, in situ STM suggests the possibility of poisoning by the polymerization of HCHO to paraformaldehyde, HO(CH2O)nOH, which is adsorbed on Pt even at -0.1 V. The oxidation of HCHO at Pt singlecrystal electrodes as a function of potential and time will be investigated further by using in situ STM. Acknowledgment. This work is supported by the National Science Council of the Republic of China under Contract No. NSC 93-2119-M-008-002. This work was partially supported by the Ministry of Education, Culture, Sport, Science and Technology, a Grant-in-Aid for the COE Project, Giant Molecules and Complex Systems, 2004. The authors thank Dr. Y. Okinaka for his help in writing this paper. LA047342T