Langmuir 2007, 23, 9495-9500
9495
Electrochemical Scanning Tunneling Microscopic Observation of the Preoxidation Process of CO on Pt(111) Electrode Surface Changhoon Jung, Jandee Kim, and Choong Kyun Rhee* Department of Chemistry, Chungnam National UniVersity, Daejeon, 305-764, South Korea ReceiVed April 16, 2007. In Final Form: June 14, 2007 Presented are sequential images of CO on Pt(111), observed with electrochemical scanning tunneling microscopy, during its electrochemical preoxidation process. In the course of the well-known phase transition from the (2 × 2)-3CO-R structure to the (x19 × x19)R23.4°-13CO structure, various structures were observed: (2 × 2)-3CO-β (Chem. Comm. 2006, 2191-2193), (1 × 1)-CO, and (x13 × x13)R46.1°-9CO. Based on an analysis of the populations of the structures averaged over imaging time and imaged location at the preoxidation potential range (0-0.25 V vs Ag/AgCl), the structures of CO domains changed sequentially in the order of (2 × 2)-3CO-R, (2 × 2)-3CO-β, (1 × 1)-CO, (x13 × x13)R46.1°-9CO, and (x19 × x19)R23.4°-13CO as the potential shifted from 0 to 0.25 V. Such a sequential structural change demonstrates that the structures of (2 × 2)-3CO-β, (1 × 1)-CO, and (x13 × x13)R46.1°-9CO are transient ones during the preoxidation of CO on Pt(111). Discussed are the transient structures in terms of various aspects, such as the absence of CO in solution and the origin of compressed structures.
1. Introduction In understanding the catalytic oxidation of hydrocarbon molecules on various platinum metal surfaces in gaseous and electrochemical environments, the oxidation of CO has been focused on extensively for several decades.1-3 The intensive highlight on CO oxidation stems from recognition of CO as a model compound mimicking the catalytic poisons formed during the oxidation of hydrocarbons. The electrochemical oxidation of CO on Pt surfaces, in particular, has been investigated extensively with an eye toward fuel cell technology.1-9 Various aspects of adsorbed CO molecules and their oxidation have been revealed, because of the relative ease of access, with in situ and ex situ methods like vibrational spectroscopies,8,10-17 ultrahigh vacuum techniques,10,11,18,19 scan* Corresponding author. E-mail:
[email protected]. Fax: 82-42-8218896. Tel: 82-42-821-5483. (1) Chrzanowski, W.; Wieckowski, A. Interfacial Electrochemistry; Marcel Dekker, Inc: New York, 1999; pp 937-954. (2) Markovic, N. M.; Ross, P. N. Surf. Sci. Rep. 2002, 45, 117-229. (3) Iwasita, T. Handbook of Fuel Cells; John Wiley & Sons Ltd.: New York, 2003; Vol. 2, pp 603-624. (4) Lin, W. F.; Iwasita, T.; Vielstich, W. J. Phys. Chem. B 1999, 103, 32503257. (5) Lin, W. F.; Zei, M. S.; Eiswitrh, M.; Ertl, G.; Iwasita, T.; Vielstich, W. J. Phys. Chem. B 1999, 103, 6968-6977. (6) Maillard, F.; Gloaguen, F.; Hahn, F.; Leger, J.-M. Fuel Cells 2002, 2, 143-152. (7) Spendelow, J. S.; Lu, G. Q.; Kenis, P. J. A.; Wieckowski, A. J. Electroanal. Chem. 2004, 568, 215-224. (8) Yajima, T.; Uchida, H.; Watanabe, M. J. Phys. Chem. B 2004, 108, 26542659. (9) Maillard, F.; Lu, G.-Q.; Wieckowski, A.; Stimming, U. J. Phys. Chem. B 2005, 109, 16230-16243. (10) Chang, S.-C.; Weaver, M. J. Surf. Sci. 1990, 238, 142-162. (11) Kizhakevariam, N.; Jiang, X.; Weaver, M. J. J. Chem. Phys. 1994, 100, 6750-6764. (12) Villegas, I.; Weaver, M. J. J. Chem. Phys. 1994, 101, 1648-1660. (13) Yoshimi, K.; Song, M.-B.; Ito, M. Surf. Sci. 1996, 368, 389-395. (14) Rodes, A.; Gomez, R.; Feliu, J. M.; Weaver, M. J. Langmuir 2000, 16, 811-816. (15) Batista, E. A.; Iwasita, T.; Vielstich, W. J. Phys. Chem. B 2004, 108, 14216-14222. (16) Lopez-Cudero, A.; Cuesta, A.; Gutierrez, C. J. Electroanal. Chem. 2005, 579, 1-12. (17) Lagutchev, A.; Lu, G. Q.; Takeshita, T.; Dlott, D. D.; Wieckowski, A. J. Chem. Phys. 2006, 125, 154705-154714. (18) Wasberg, M.; Palaikis, L.; Wallen, S.; Kamrath, M.; Wieckowski, A. J. Electroanal. Chem. 1988, 256, 51-63.
ning tunneling microscopy (STM),12,20,21 surface X-ray scattering,22,23 surface nuclear magnetic resonance,24-26 and electrochemical means.5-7,15,16,28 Despite such numerous endeavors, the details of CO oxidation (e.g., reaction sites, relative rates of charge transfer, and surface diffusion of CO) are not clearly understood yet. In 1994, for example, the adlayer of CO on Pt(111) was reported to transform from a (2 × 2)-3CO structure to a (x19 × x19)R23.4°-13CO structure during preoxidation (0.0 - 0.3 V vs SCE).12 After the advent of the first structural information at the molecular level, however, it is true that there was not much of a significant advance in understanding CO oxidation from a structural point of view. In this work, we present sequential electrochemical STM images of CO observed during preoxidation on the Pt(111) electrode surface. To our knowledge, the reported images, including our previous work,21 are the first sequential images at the molecular level, which is surely a break-through in understanding the CO oxidation processes. The key element of this work is the phase transitions of CO adlayers, including several new adstructures,21 during the preoxidation around reaction sites on a Pt(111) electrode surface. 2. Experimental Section The Pt(111) single crystals electrode used in this work was made by the bead method. After being annealed with a flame and quenched with hydrogen-saturated water, the single crystal bead, protected (19) Zurawski, D.; Wasberg, M.; Wieckowski, A. J. Phys. Chem. 1990, 94, 2076-2082. (20) Song, M.-B.; Yoshimi, K.; Ito, M. Chem. Phys. Lett. 1996, 263, 585590. (21) Jung, C.; Ku, B.; Kim, J.; Rhee, C. K. Chem. Commun. 2006, 21912193. (22) Markovic, N. M.; Grgur, B. N.; Lucas, C. A.; Ross, P. N. Surf. Sci. 1997, 384, L805-L814. (23) Lucas, C. A.; Markovic, N. M.; Ross, P. N. Surf. Sci. 1999, 425, L381L386. (24) Yahnke, M. S.; Rush, B. M.; Reimer, J. A.; Cairns, E. J. J. Am. Chem. Soc. 1996, 118, 12250-12251. (25) Tong, Y.; Rice, C.; Wieckowski, A.; Oldfield, E. J. Am. Chem. Soc. 2000, 122, 1123-1129. (26) Tong, Y.; Kim, H. S.; Babu, P. K.; Waszczuk, P.; Wieckowski, A.; Oldfield, E. J. Am. Chem. Soc. 2002, 124, 468-473. (27) Lebedeva, N. P.; Koper, M. T. M.; Herrero, E.; Feliu, J. M.; Santen, R. A. v. J. Electroanal. Chem. 2000, 487, 37-44. (28) Bergelin, M.; Herrero, E.; Feliu, J. M.; Wasberg, M. J. Electroanal. Chem. 1999, 467, 74-84.
10.1021/la701103u CCC: $37.00 © 2007 American Chemical Society Published on Web 07/25/2007
9496 Langmuir, Vol. 23, No. 18, 2007
Jung et al.
Figure 1. Typical stripping voltammogram of CO on Pt(111) in a 0.05 M H2SO4 solution (solid line). The dashed line is a voltammogram of Pt(111) in a 0.05 M H2SO4 solution. The inset presents a enlarged voltammogram in the preoxidation region. Scan rate: 50 mV/s. with a water drop, was incorporated into a homemade EC-STM cell for an STM instrument (Nanoscope III, Digital Instrument, U.S.A.) and was immediately brought into contact with a 0.05 M H2SO4 solution (Merck, Suprapur) saturated with CO (99.99%, Air Products and Chemicals, U.S.A.) at 0.05 V for 10 min. Without any loss of potential control, the CO-containing solution was replaced with a CO-free 0.05 M H2SO4 solution by iteration of adding several aliquots of the CO-free solution into the cell and removing a similar amount of the CO-containing solution, to an extent to ensure that there was no dissolved CO in the cell. Then, a W tip, electrochemically polished and coated with molten polyethylene, was attached to one of the (111) facets. It took normally less than 10 min to observe the first STM image after the beginning of the replacement. For voltammetric experiments, on the other hand, a (111) facet of a bead crystal, polished to a mirror-like finish, was used in a configuration of meniscus. The electrode potential was controlled against a Ag/AgCl reference electrode and reported as measured.
3. Results 3.1. Electrochemistry of CO Preoxidation. Figure 1 presents typical voltammograms of Pt(111) before (dashed line) and after (solid line) adsorption of CO in a 0.05 M H2SO4 solution, respectively. The preoxidation takes place from 0.05 to 0.28 V with the current maximum at 0.20 V, whereas the main oxidation starts at 0.35 V. The oxidative charge under the curve of preoxidation, after a proper background correction, was 34 µC/ cm2. This specific charge is equivalent to the coverage of 0.07, defined as a ratio of the number of adsorbed CO molecules to the number of surface Pt atoms. The charge of a complete oxidation of CO reflects that the total CO coverage was 0.75. 3.2. Various Adstructures of CO. In Figure 2 are shown sequential electrochemical STM images for the preoxidation of CO adsorbed on Pt(111) observed in a CO-free 0.05 M H2SO4 solution. After taking a few images at one potential roughly for 5 min, the electrode potential was stepped to the next higher potential for another 5 min of imaging. The molecular level image in Figure 2a represents a well-known (2 × 2)-3CO structure of 0.75 coverage.12,21 As the electrode potential shifts to 0.10 V, the adstructure of CO changes to another (2 × 2)-3CO superlattice of 0.75 coverage (Figure 2b), recently found and assigned to (2 × 2)-3CO-β (abbreviated as (2 × 2)-β).21 The former CO structure in Figure 2a is named as (2 × 2)-3CO-R (abbreviated as (2 × 2)-R). With a further anodic potential shift to 0.15 V, the CO
Figure 2. Sequential electrochemical STM molecular level images (10 nm × 10 nm) of CO adlayers on Pt(111) in 0.05 M H2SO4 solution at (a) 0.05, (b) 0.10, (c) 0.15, (d) 0.20, and (e) 0.25 V.
adlayer of (2 × 2)-β converts to a domain of a new hexagonal structure whose spot period is 0.29 ( 0.01 nm as in Figure 2c. Because the particular spot period is comparable to one Pt-Pt distance of Pt(111) (0.277 nm, designated with a), this specific CO domain is assigned to a (1 × 1)-CO structure with 1.0 coverage (see the Discussion section). It should be underlined that the details of the (1 × 1) structure of CO differs slightly from those of the (2 × 2)-4CO structure with an identical CO coverage observed by us.21 A further anodic shift to 0.20 V modifies the (1 × 1) structure to a CO domain whose spots are not uniform but distinctive as shown in Figure 2d. Despite the uneven contrast, the distances between the hexagonally arrayed spots are 0.34 ( 0.02 nm, indicating that the coverage of the CO domain in Figure 2d becomes lower than 1.0. In addition, it is quite remarkable that the angle of the densest molecular row in Figure 2d (line 2) with respect to the [01h1] direction (line 1) is 14° in the clockwise direction. Based on the observed spot period and rotation angle, the superlattice in Figure 2d is assigned to the (x13 × x13)R46.1°-9CO structure (abbreviated as (x13 × x13)) whose coverage is 0.692 (see below). Upon observation of the (x13 × x13)-9CO structure, the well-known Moire´ pattern of (x19 × x19)R23.4°-13CO (abbreviated as (x19 × x19)) with 0.682 coverage (Figure 2e) follows at 0.25 V. Notably, the densest molecular row in (x19 × x19) (line 3) rotates counterclockwise by 9° against the [01h1] direction (line 1). Figure 3 shows proposed schematics of the molecular CO arrangements as observed in Figure 2. In the (2 × 2)-R structure as illustrated in Figure 3a, there is one CO molecule at one atop site and two CO molecules at threefold hollow sites in a unit cell.12,21 During the anodic shift, the CO molecules in the hollow sites in the (2 × 2)-R structure may move to the adjacent atop
Preoxidation of CO on Pt(111)
Figure 3. Schematics of the superlattices of CO adsorbed on Pt(111): (a) (2 × 2)-3CO-R, (b) (2 × 2)-3CO-β, (c) (x13 × x13)R46.1°-9CO, and (d) (x19 × x19)R23.4°-13CO and superimposion of (e) pseudo-(1 × 1)-CO structure (blue balls) and (x13 × x13) structure (red balls) and (f) (x13 × x13) (blue balls) structure and (x19 × x19) structure (red balls). The hexagonal grids represent the Pt(111) surface, and the corners of the triangles stand for Pt atoms. The contrasts in the schematics of panels a-d reflect the observed contrasts in STM images. The pseudo-(1 × 1)-CO structure (blue balls in panel e) is a structure which differs from the conventional (1 × 1) structure only in the size of unit cell (0.34 nm).
sites to form the (2 × 2)-β structure without any change in the CO coverage (Figure 3b).21 A further potential increase induces a structural change from the (2 × 2)-β structure to the (1 × 1)-CO lattice of uniform contrast (not shown in Figure 3), probably by compressing the molecular rows of 2a periodicity in the (2 × 2)-β structure along the parallel direction (e.g., along the [01h1] direction in Figure 3b). The (1 × 1)-CO lattice may transform further to the (2 × 2)-4CO structure, as detailed in ref 21, without changing the CO coverage. On the other hand, the appearance of the (x13 × x13) structure (Figure 3c), concomitant with the decrease in CO coverage to 0.692, would be described as a convolution of two processes: one is a simple expansion of the base lattice vectors of the (1 × 1)-CO structure from 0.277 to 0.34 nm (“pseudo-(1 × 1)-CO structure”) as the CO coverage decreases to 0.692 and the other is a clockwise rotation of the pseudo-(1 × 1)-CO structure by 13.4°. In Figure 3e, the pseudo(1 × 1)-CO structure (blue balls) and the (x13 × x13) structure (red balls) are superimposed. Surprisingly, a repeating unit, composing of two concentric 6-membered and 12-membered hexagons around a central CO molecule (dark ball), is recognizable. The central CO molecule is obviously the molecule belonging to the different arrays but overlapping exactly. Thus, the transformation to the (x13 × x13) structure could be achieved by a simple clockwise rotation of the CO circles of the pseudo-(1 × 1)-CO structure in each repeating unit. Another example of such a local rotational repeating unit is the transformation of the (x13 × x13) structure to the (x19 × x19) structure (Figure 3d). Here again, an overlapped schematic of the two structures is shown in Figure 3f. Obviously, a repeating unit of 6-membered hexagon is clearly discernible. A simple counterclockwise rotation
Langmuir, Vol. 23, No. 18, 2007 9497
Figure 4. Sequential electrochemical STM images (20 nm × 20 nm) of a location on CO-covered Pt(111) during the preoxidation: (a) 0.05, (b) 0.15, (c) 0.17, (d) 0.20, (e) 0.25, and (f) 0.25 V.
of the circles in the (x13 × x13) structure by 9° leads to the structure of (x19 × x19), accompanied with a slight lattice vector expansion. The recognition of the repeating units, as exemplified in Figure 3, panels e and f, absolutely makes it possible to discard the rotation of an entire CO domain causing a huge movement of the CO molecules around its perimeter. 3.3. Structural Evolution during Potential Excursion. The structural change of CO on Pt(111) during the preoxidation is obviously related to the CO coverage decrease from 0.750 to 0.685. The sequential images in Figure 4 demonstrate the phase transitions during the preoxidation on Pt(111) terraces. At the onset of the preoxidation, the entire surface is covered with the domains of (2 × 2)-R (Figure 4a). A potential increase induces dramatic changes as shown in Figure 4b (enlarged in Figure S1 of the Supporting Information): (1) the right upper terrace is covered with the adstructure of (2 × 2)-β, (2) the area, designated with an oval, is covered with (1 × 1)-CO, (3) the region, enclosed with dashed line, is a mixed domain of (2 × 2)-β and (1 × 1)-CO, and (4) the rectangle of solid line is still a domain of (2 × 2)-R. A further potential increase (Figure 4c, enlarged in Figure S2 of the Supporting Information), however, resumes a significant amount of the CO domain to the (2 × 2)-R one as shown in the oval drawn with a solid line with a simultaneous transition from the (2 × 2)-β structure to the (1 × 1)-CO superlattice in the region confined with the oval drawn with a dashed line. This particular reversibility between the structures of (2 × 2)-3CO and (1 × 1)-CO implies that the CO molecules are quite mobile to form even molecular rows with the period of a. Indeed, a mixed domain of the (2 × 2)-β structure and (1 × 1)-CO (i.e., the rectangular area in Figure 4c), is recognizable at a location different from that in Figure 4b. At 0.2 V where the preoxidation current is maximized, the observed surface takes a totally different
9498 Langmuir, Vol. 23, No. 18, 2007
Figure 5. Plot of potential vs relative preoxidation charge (Q/Qo) after holding for 10 and 30 min. Q and Qo are preoxidation charges obtained in a CO-free 0.05 M H2SO4 solution with potential holding and without potential holding, respectively.
shape (Figure 4d): the right part of the imaged area is covered with the (x19 × x19), whereas the left part is filled with spots of (x13 × x13), featured with apparent irregularity in contrast. At the end of the preoxidation, a domain of (x19 × x19) is emerging among the (x13 × x13) domain at the right part of Figure 4e, and the appearance of the (x19 × x19) is prevailing in Figure 4f as the imaging time elapses at 0.25 V. 3.4. Does the Preoxidation Proceed during STM Imaging? The stability of the adsorbed CO on Pt(111) during STM imaging was examined. It would be logical to cast a question on the stability of the adsorbed CO, especially in the potential region of oxidation, because during STM imaging, the oxidation may proceed to alter the adstructures. For the specific purpose, two methods were employed: measuring coulometrically the amount of CO after holding the electrode potential at a potential in the preoxidation region for a certain amount of time, and imaging continuously a particular location at a potential in the preoxidation region for a certain amount of time. Figure 5 represents a plot of relative preoxidation charge (Q/ Qo) along with holding potential and time. Here, Q and Qo are preoxidation charges obtained in a CO-free 0.05 M H2SO4 solution after holding at a potential and without potential holding, respectively (see Figure S3 for the actual stripping voltammograms). At 0.05 V, the preoxidation charge does not decrease at all even after holding for 30 min. Clearly and remarkably demonstrated is that, at 0.10 and 0.15 V, significant amounts of preoxidation charge remain after potential holding for 30 min, indicating incomplete preoxidation. At 0.20 V, the preoxidation does not finish within 10 min, but the reaction completes within 30 min. At higher potential, however, the preoxidation comes to the end within 10 min. These results indicate that during STM imaging, alteration of the CO adstructures can take place depending on the holding potential. Time-dependent EC-STM images are shown in Figure 6. On the top side of the panel, the potentials where the images were taken are designated. On the left side of the panel, the times after STM tip engagement are labeled, also. The images observed at 0.05 V are filled with the (2 × 2)-R structure regardless of observation time, indicating definitely that the particular structure is stable at 0.05 V. On the other hand, the adstructure of CO at potentials higher than 0.05 V evolutes as the time flows. Upon shifting the potential from 0.05 to 0.20 V, for example, narrow domains of (2 × 2)-β (designated with ovals in the top frame) appear within 5 min. In the next frame, domains of (2 × 2)-β shrink and new domains of (1 × 1)-CO emerge, as indicated with the ovals and solid polygons, respectively. Wider domains
Jung et al.
of (1 × 1) are notable in the subsequent frame. After 10 min more (the last frame), the surface becomes filled with the structure of (x19 × x19) but not fully developed yet. Because the preoxidation proceeds continuously at 0.20 V, a change from the (2 × 2)-R structure to the (x19 × x19) one via the (2 × 2)-β and (1 × 1)-CO adlattices is rather straightforward to understand. Below 0.20 V within the STM experimental time scale (25 min), indeed, the (x19 × x19) structure was not observed. Within 10 min after shifting to 0.25 V, however, the (x19 × x19) structure became immediately discernible as in the second frame among the rightmost column. The structural evolution in Figure 6 is coherent with the charge variation in Figure 5, at least qualitatively. Therefore, it could be concluded that the extent of the CO preoxidation is limited by potential and time. 3.5. Potential-Dependent Populations of Various CO Adstructures. Figure 7 presents the probabilities to observe the adstructures of CO on Pt(111) at the potentials of interest. The probabilities were roughly estimated with a 5% increment from more than 30 replica experiments performed within 10 min time scale. Because of the local probing nature of STM and because of the progress in preoxidation under the employed experimental conditions, a statistical treatment of the images taken in different locations and times (i.e., averaging over time and replica experiment) is reasonable to the generalized probabilities or populations of CO structures at each potential. At lower potentials than 0.05 V, the (2 × 2)-R structure is unique. As the potential become higher than 0.05 V, the population of the adstructures of (2 × 2)-β and (1 × 1)-CO increases gradually and sequentially at the expense of the (2 × 2)-R domain. At 0.20 V, the whole surface become a mosaic of domains related to (1 × 1)-CO and (x13 × x13) (featured with spots of nonuniform contrast and irregular spot distances ranging from 0.28 to 0.34 nm) and growing domains of (x19 × x19). At 0.25 V, the transformation to the (x19 × x19) structure is almost completed. The presented results confirm that, during the preoxidation of CO on Pt(111), the (2 × 2)-R structure of 0.75 coverage transforms to the (x19 × x19) structure of 0.685 coverage via the intermediate structures of (2 × 2)-β, (1 × 1)-CO, and (x13 × x13). The particular phase transition from the (2 × 2)-R structure to the (2 × 2)-β one is closely related to the reaction sites on the Pt surfaces (i.e., steps and CO-adsorption defects).14,16,27 As shown in Figure 8a, small domains of (2 × 2)-β, grown perpendicularly to a step, were commonly observed in (2 × 2)-R domains. Furthermore, (2 × 2)-β domains in conjunction with adsorbate defects, as shown in Figure 8b, were repetitively recognized as well. The close relation between the active sites and the transit structures indicates that the (2 × 2)-β structure, later transforming to the (1 × 1)-CO one, may be a precursor toward the preoxidation (see below).
4. Discussions 4.1. Does the Absence of CO in Solution Cause the Phase Transition from (2 × 2)-R to (x19 × x19)? The presence of CO in solution has been known to affect the adstructures of CO on Pt(111), especially at low potential. Specifically, Feliu and co-workers14 have reported that, in the presence of CO in solution, spectral features of adsorbed CO indicated the registry of CO was (2 × 2)-R below 0.5 V vs RHE (roughly 0.2 V in our scale), whereas in the absence of CO, the spectral features changed to reflect a mixed layer of (2 × 2)-R and (x19 × x19) even below 0.3 V vs RHE (0 V in our scale). In this work, inconsistent with the previous findings is the observation of the (2 × 2)-R structure even in the absence of CO in solution at 0.05 V as a unique structure.
Preoxidation of CO on Pt(111)
Langmuir, Vol. 23, No. 18, 2007 9499
Figure 6. Time-dependent STM images of CO on Pt(111) at various potentials. The areas enclosed with the dotted lines and solid lines are the domains of the (2 × 2)-3CO-β and (1 × 1)-4CO structures, respectively. Images size: 20 nm × 20 nm.
Figure 8. Typical electrochemical STM images (10 nm × 10 nm) around reaction sites of CO oxidation on Pt(111) at 0.13 V: small domains of (2 × 2)-β adjacent to (a) a step of Pt(111) substrate and (b) adsorbate defects in a (2 × 2)-R domain. Figure 7. Probabilities to observe the adstructures of CO on Pt(111) as a function of potential. The probabilities are averaged values over imaging times and replica experiments.
The particular discrepancy may come from different experimental procedures to remove dissolved CO in solution after contact with Pt(111) electrodes. There are two ways to remove the dissolved CO in solution: purging it with an inert gas and replacing it with a CO-free solution. Both methods were performed in this work and are compared here. After purging the COcontaining solution with N2 for 5 min even at 0 V, most of the preoxidation charge disappeared, revealing that a significant amount of adsorbed CO desorbed somehow. When the CO-
containing solution was carefully replaced with a copious amount of CO-free solution at 0 V, on the other hand, the preoxidation charge did not change. Actually, a CO-free 0.05 M H2SO4 solution was fed continuously into the electrochemical cell at least for 5 min with simultaneous drainage of the CO-containing solution to keep the solution level enough to secure potential control, but not to flood the cell. After completion of such replacement, the electrode potential was shifted to a potential in the preoxidation region, the preoxidation charge was dependent on the holding potential and time, as shown in Figure 5 (see Figure S3). A cyclic voltammogram following a complete CO removal up to 0.55 V, furthermore, indicated that there was no dissolved CO in solution after replacement. Therefore, it is clear that the absence
9500 Langmuir, Vol. 23, No. 18, 2007
Jung et al.
of CO in solution does not induce the phase transition from (2 × 2)-R to (x19 × x19). Regarding the presence of CO in solution, two EC-STM observations are worth mentioning here. After observation of the (x19 × x19) structure at 0.25 V, the electrode potential was shifted to 0.05 V. When the dissolved CO was present in the solution, the (2 × 2)-R structure was recovered immediately. By contrast, when the dissolved CO was absent, the (x19 × x19) structure was persistent. Thus, it is safe to state that the phase transition of interest is due to preoxidation of CO. 4.2. Origin of the (1 × 1) Structure. The (1 × 1) structure observed during preoxidation of CO on Pt(111) is certainly a highly compressed structure if CO is the adsorbate. Thus, it is worth discussing further the origin of the (1 × 1) structure. Conceivable origins of the (1 × 1) structure are bisulfate, bare Pt(111), and CO. Saravanan and co-workers29 have reported, based on a dynamic Monte Carlo simulation, that the preoxidation voltammetric wave became apparent in the case of a competitive adsorption of bisulfate anion and OH to reaction sites under a condition of slow CO diffusion. It is well-known, on the other hand, that bisulfate forms a (x3 × x7) structure on Pt(111).30,31 Because a domain of spot periods of x3a (0.48 nm) and x7a (0.73 nm) was not observed in this work, the (1 × 1) structure of interest is certainly not related to anion adsorption. Due to a possibility not to observe adsorbed anions under the employed imaging condition, however, involvement of anions during the preoxidation could not be excluded. Another possible origin of the (1 × 1) structure would be a bare Pt(111) after removal of CO during the preoxidation. Considering that the population of the (1 × 1) structure is roughly 70% at 0.20 V in Figure 7, the coverage of CO left at the specific potential is easy to estimate logically to be 0.23 with a simple calculation (0.75-0.75 × 0.7). In reality, however, the CO coverage after the preoxidation is 0.685, rejecting an assumption that the domain of the (1 × 1) structure is a domain of bare Pt(111). Then, the last choice accountable for the (1 × 1) registry is adsorbed CO. 4.3. Roles of the Intermediate Adstructures. A naturally emerging question concerns the existence of the intermediate structures. In particular, the structures of (2 × 2)-β and (1 × 1)-CO are certainly less stable than the (2 × 2)-R adlattice, because of the highly repulsive lateral force in the molecular row with
the period of a. The existence of locally dense CO rows, such as solitons, has been demonstrated during an increase of CO coverage on Pt(111) in an ultrahigh vacuum environment.32 The repulsive force between the CO molecules in the dense rows may play a critical role in the preoxidation (e.g., pushing the CO molecules in the molecular rows with the period of a to reaction sites (like bullets in a magazine) to facilitate the preoxidation). As the coverage of CO in a terrace decreases, the compact CO domains, most likely (1 × 1)-CO domains, transit sequentially to the less dense domains of (x13 × x13) and (x19 × x19) by a convolution of intermolecular distance expansion and rotation. Thus, the repulsive structures of (2 × 2)-β and (1 × 1)-CO may be stairways from the (2 × 2)-R adlattice up to a reaction complex, and the (x13 × x13) superlattice may be a cascade path down to the (x19 × x19) structure.
(29) Saravanan, C.; Koper, M. T. M.; Markovic, N. M.; Head-Gordon, M.; Ross, P. N. Phys. Chem. Chem. Phys. 2002, 4, 2660. (30) Funtikov, A. M.; Linke, U.; Stimming, U.; Vogel, R. Surf. Sci. 1995, 324, L343. (31) Itaya, K. Prog. Surf. Sci. 1998, 58, 121.
LA701103U
5. Summary In summary, we have presented structural changes of adsorbed CO, observed with electrochemical STM, during the phase transition from the (2 × 2)-R adlayer to the (x19 × x19) adlattice. The specific transition was clearly demonstrated to be due to preoxidation on the Pt(111) electrode surface. In the phase transition, the (2 × 2)-R adlayer was transformed to compressed adstructures of (2 × 2)-β, (2 × 2)-4CO, and (1 × 1)-CO, which may be precursors to a reaction complex. As the coverage of CO decreased, the adlayer of CO transitioned to the (x19 × x19) adlattice via the (x13 × x13) one. A theoretical work on the reaction path as revealed in this work is needed to understand this in terms of reaction kinetics. Acknowledgment. The authors appreciate the permission of Central Research Facility, Chungnam National University, Korea, to use the STM instrument. This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD; R05-2004-000-10247-0). Dedicated to Professor Ho-In Lee, Seoul National University, Korea, on occasion of his 60th birthday. Supporting Information Available: Enlarged versions of Figure 4b (Figure S1) and Figure 4c (Figure S2). Actual stripping voltammograms for Figure 5 (Figure S3). This material is available free of charge via the Internet at http://pubs.acs.org.
(32) Tushaus, M.; Berndt, W.; Conrad, H.; Bradshaw, A. M.; Persson, B. Appl. Phys. A 1990, 51, 91-98.
