In Situ Scanning Tunneling Microscopy of Molecular Assemblies of

Mar 9, 2004 - Molecules of copper(II) and cobalt(II) 5,10,15,20-tetraphenyl-21H,23H-porphine (CuTPP and CoTPP) and cobalt(II) phthalocyanine (CoPc) ar...
22 downloads 12 Views 1MB Size
Langmuir 2004, 20, 3159-3165

3159

In Situ Scanning Tunneling Microscopy of Molecular Assemblies of Cobalt(II)- and Copper(II)-Coordinated Tetraphenyl Porphine and Phthalocyanine on Au(100) Soichiro Yoshimoto,† Akinori Tada,† Koji Suto,† Shueh-Lin Yau,‡ and Kingo Itaya*,†,‡ Department of Applied Chemistry, Graduate School of Engineering, Tohoku University, Aoba-yama 04, Sendai 980-8579, Japan, and Core Research Evolutional Science and Technology organized by Japan Science and Technology Agency (CREST-JST), Kawaguchi Center Building, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan Received October 20, 2003. In Final Form: January 16, 2004 Molecules of copper(II) and cobalt(II) 5,10,15,20-tetraphenyl-21H,23H-porphine (CuTPP and CoTPP) and cobalt(II) phthalocyanine (CoPc) are spontaneously adsorbed onto reconstructed Au(100) substrate from a benzene solution containing each individual complex. In situ scanning tunneling microscopy (STM) was used to examine the real-space arrangement and the internal molecular structure of each of the individual molecules in 0.1 M HClO4 under potential control. The adsorption of CuTPP and CoTPP produced the same highly ordered square array with an intermolecular spacing of 1.44 nm on a reconstructed Au(100) surface. These molecular superlattices and the underlying reconstructed Au(100) predominated between 0 and 0.9 V, but lifting of the reconstructed Au(100) surface and elimination of the ordered adlayers occurred at more positive potentials. Molecular resolution STM revealed propeller-shaped admolecule with its center imaged as a protrusion for Co(II) and a depression for Cu(II). In contrast, the spontaneous adsorption of CoPc molecules resulted in a rapid phase transition from the reconstructed Au(100) surface to the (1 × 1) phase, coupled with the production of locally ordered, square-shaped arrays with an intermolecular distance of 1.65 nm. This molecular adlayer and the Au(100)-(1 × 1) remained unchanged when the potential was modulated between 0 and 1.0 V. These results indicate that the subtle variation in the molecular structure of adsorbate influenced not only its spatial arrangement but also the structure of the underlying Au(100) substrate.

Introduction It is now recognized that the precise control of molecular assembly on a substrate is one of the key foundations to harness nanomolecular devices. Consequently, there has been an upsurge of interest recently to prepare and characterize the molecular architecture on an atomically flat substrate such as gold and highly oriented pyrolytic graphite (HOPG). There are many reports describing the preparation of ordered molecular assemblies from vapor phase or solution phase. For example, porphyrins and phthalocyanines have found applications in such diversified fields as biology,1 photosynthesis,1 electrocatalysis,2,3 and molecular devices.4 More specifically, copper(II) phthalocyanine (CuPc) is involved in research on lightemitting diodes (LED)5 and field effect transistors (FET).4,5 In addition, the use of metallophthalocyanines (MPcs) for catalyzing the electroreduction of dioxygen represents another well-known research subject needed for the development of an efficient cathode for fuel cells.2,3,6-10 * To whom correspondence should be addressed. Phone/Fax: +81-22-214-5380. E-mail: [email protected]. † Tohoku University. ‡ Core Research Evolutional Science and Technology. (1) Electron Transfer in Chemistry; Balzani, V., Ed.; WILEY-VCH: New York, 2001; Vol. 3. (2) Yeager, E. Electrochim. Acta 1984, 29, 1527. (3) Collman, J. P.; Wagenknecht, P. S.; Hutchison, J. E. Angew. Chem., Int. Ed. Engl. 1994, 33, 1537 and references therein. (4) Molecular Electronics; Jortner, J., Ratner, M., Eds.; IUPAC: Oxford, 1997. (5) Guillaud, G.; Simon, J.; Germain, J. P. Coord. Chem. Rev. 1998, 178-180, 1433. (6) Jasinski, R. J. Electrochem. Soc. 1965, 112, 526. (7) Alt, H.; Binder, H.; Sandstede, G. J. Catal. 1973, 28, 8.

Intensive effort in the past decade has made substantial progress in the research on molecular assemblies on the surfaces of Au,11-21 Ag,11,22 and Cu11,23,24 in a vacuum (UHV) and solution phases. The adsorption of 5,10,15,20-tetrakis(3,5-di-tertiarybutylphenyl)porphine copper(II) (CuTBPP) on Cu(100), Au(110), and Ag(110) surfaces has been examined to elucidate the role of substrate.11,12 The adsorption of 5,10,15,20-tetrakis(3,5(8) Savy, M.; Andro, P.; Bernard, C.; Magner, G. Electrochim. Acta 1973, 18, 191. (9) Mho, S.-i.; Ortiz, B.; Park, S.-M.; Ingersoll, D.; Doddapaneni, N. J. Electrochem. Soc. 1995, 142, 1436. (10) Ca´rdenas-Jiro´n, G. I.; Gulppi, M. A.; Caro, C. A.; del Rı´o, R.; Pa´ez, M.; Zagal, J. H. Electrochim. Acta 2001, 46, 3227 and references therein. (11) Jung, T. A.; Schlittler, R. R.; Gimzewski, J. K.; Tang, H.; Joachim, C. Science 1996, 271, 181. (12) Jung, T. A.; Schlittler, R. R.; Gimzewski, J. K. Nature 1997, 386, 696. (13) Yokoyama, T.; Yokoyama, S.; Kamikado, T.; Mashiko, S. J. Chem. Phys. 2001, 115, 3814. (14) Yokoyama, T.; Yokoyama, S.; Kamikado, T.; Okuno, Y.; Mashiko, S. Nature 2001, 413, 619. (15) Scudiero, L.; Barlow, D. E.; Hipps, K. W. J. Phys. Chem. B 2000, 104, 11899. (16) Scudiero, L.; Barlow, D. E.; Mazur, U.; Hipps, K. W. J. Am. Chem. Soc. 2001, 123, 4073. (17) Lu, X.; Hipps, K. W.; Wang, X. D.; Mazur, U. J. Am. Chem. Soc. 1996, 118, 7197. (18) Hipps, K. W.; Lu, X.; Wang, X. D.; Mazur, U. J. Phys. Chem. 1996, 100, 11207. (19) Lu, X.; Hipps, K. W. J. Phys. Chem. B 1997, 101, 5391. (20) Barlow, D. E.; Hipps, K. W. J. Phys. Chem. B 2000, 104, 5993. (21) Chizhov, I.; Scoles, G.; Kahn, A. Langmuir 2000, 16, 4358. (22) Lackinger, M.; Hietschold, M. Surf. Sci. 2002, 520, L619. (23) Lippel, P. H.; Wilson, R. J.; Miller, M. D.; Wo¨ll, C.; Chiang, S. Phys. Rev. Lett. 1989, 62, 171. (24) Yanagi, H.; Mukai, H.; Ikuta, K.; Shibutani, T.; Kamikado, T.; Yokoyama, S.; Mashiko, S. Nano Lett. 2002, 2, 601.

10.1021/la0359474 CCC: $27.50 © 2004 American Chemical Society Published on Web 03/09/2004

3160

Langmuir, Vol. 20, No. 8, 2004

di-tertiarybutylphenyl)porphine platinum (PtTBPP) on a Cu(100) substrate is reported to shed insight into the mechanism of molecular organization.24 Yokoyama et al. found that CN-substituted TBPP molecule is adsorbed on Au(111) to produce one-dimensional supermolecular assemblies such as monomers, trimers, tetramers, and extended wire-like structures.14 Hipps and co-workers investigated various metal-coordinated tetraphenyl-21H,23H-porphine (MTPP) such as NiTPP, CuTPP, and CoTPP on the reconstructed Au(111) surface.15,16 Among these macrocyclic molecules, only the Co(II) metal center exhibited a pronounced protrusion in the STM image. This difference is presumably due to the presence of a halffilled dz2 orbital at the Co(II) site, which acts as the electron-tunneling mediator between the tip and the gold substrate. Furthermore, molecular resolution STM has revealed the internal molecular structure of MPc adlayers, allowing detailed characterization of their adsorption configurations.17-24 Furthermore, STM has been extensively used to probe the molecular assembly of various metal-coordinated Pc’s such as CuPc,17-19 CoPc,17,18 NiPc,19 FePc,19 and VOPc20 on Au(111). The metal centers appear to be unimportant in the organization of the admolecules as they are arranged similarly on the Au(111) substrate. These metal centers yield uneven corrugations, presumably due to the difference in their electronic structures.15-20 The above highlighted studies were conducted in a vacuum environment, showing that the vapor-phase deposition is useful to produce ordered overlayers of macrocyclic molecules.10-23 On the other hand, solutionphase dosing can be an inexpensive and convenient alternative for preparing highly ordered molecular adlattices on gold and HOPG substrates. An aqueous dosing solution works well if the admolecule is water soluble. However, it is obviously difficult to use for most organic molecules in aqueous solutions. To continue to develop a technique for producing molecular assemblies from a solution phase, we explored the use of organic solvents such as benzene and ethanol as the medium for the deposition of organic compounds. Preliminary results have already been reported.25-27 For example, water-insoluble tetraphenyl-21H,23H-porphine cobalt(II) (CoTPP) and copper(II) (CuTPP) formed highly ordered adlayers on Au(111) from their benzene solutions, as indicated by molecular resolution STM.25,27 The as-prepared adlayers of CoTPP and CuTPP possess the same structure as those found in UHV.15,16 On the other hand, it was noted that CoPc and CuPc molecules adapt different spatial structures on reconstructed Au(111). To further explore the role of substrate in the formation of molecular assemblies, we examined CoTPP, CuTPP, and CoPc on the Au(100) surface. Chart 1 illustrates their molecular structures. Meanwhile, the reconstruction of Au(100) surface, a popular subject in the study of electrified interfaces, is also of our present interest.28-30 Experimental Section Compounds of 5,10,15,20-tetraphenyl-21H,23H-porphine cobalt(II) (CoTPP) and copper(II) (CuTPP) and cobalt(II) phtha(25) Yoshimoto, S.; Tada, A.; Suto, K.; Narita, R.; Itaya, K. Langmuir 2003, 19, 672. (26) Yoshimoto, S.; Narita, R.; Wakisaka, M.; Itaya, K. J. Electroanal. Chem. 2002, 532, 331. (27) Yoshimoto, S.; Tada, A.; Suto, K.; Itaya, K. J. Phys. Chem. B 2003, 107, 5836. (28) Magnussen, O. M.; Hotlos, J.; Behm, R. J.; Batina, N.; Kolb, D. M. Surf. Sci. 1993, 296, 310. (29) Gao, X.; Edens, G. J.; Hamelin, A.; Weaver, M. J. Surf. Sci. 1993, 296, 333. (30) Kolb, D. M. Surf. Sci. 2002, 500, 722.

Yoshimoto et al. Chart 1. Molecular Structure of CoTPP (or CuTPP) and CoPc

locyanine (CoPc) were purchased from Aldrich and used without further purification. Benzene was obtained from Kanto Chemical Co. (spectroscopy grade). The supporting electrolyte for voltammetric and STM measurements was 0.1 M HClO4 prepared by mixing ultrapure perchloric acid obtained from (Cica-Merck) and Milli-Q water (resistivity g 18.2 MΩ cm). The Au(100) single-crystal electrode was prepared by the Clavilier method, and the size of the crystal was roughly 2 mm in diameter.31 As reported previously, the surface structure of Au(100) is critically dependent on details in the preparation process, and we employed the procedure we established previously to make a reconstructed or ideal (1 × 1) surface.28-30 Specifically, the reconstructed Au(100) surface was prepared by annealing in a hydrogen flame for several seconds, followed by cooling in air for 3-5 min.28 This Au(100) electrode was immersed in a benzene dosing solution containing a saturated amount of CoPc or 10 µM CoTPP or CuTPP. (The former was only slightly soluble in benzene, and its concentration was lower than 10 µM.) In situ STM imaging (vide infra) showed that an immersion time of 5 min or longer enabled the formation of a full monolayer of CoPc. In contrast, CoTPP and CuTPP readily dissolved in benzene, and a dipping time as short as 10 s was sufficient to produce a full monolayer of these molecules on Au(100).25 The Au(100) electrode coated with one of the molecules of CoTPP, CuTPP, and CoPc was then rinsed with ultrapure water before it was transferred into the electrochemical or STM cell to perform subsequent experiments. The elapsed time was about 10 min before the STM imaging was actually started. Cyclic voltammetry was carried out at 20 °C with a potentiostat (HOKUTO HAB-151, Tokyo). The Au(100) electrode configured by the hanging meniscus method was placed in a threecompartment electrochemical cell blanketed with N2. The reference and counter electrodes were a reversible hydrogen electrode (RHE) and Pt wire, respectively. The electrochemical STM measurement was performed with a Nanoscope E (Digital Instruments, Santa Barbara, CA). The tip was made of a tungsten wire etched in 1 M KOH, and a thin layer of nail polish was applied to minimize residual faradic currents. All STM images were recorded in the constant-current mode. A Pt wire was used as a quasi-reference electrode in the STM experiments, but all potentials reported are referred to the reversible hydrogen electrode (RHE).

Results and Discussion Voltammetry. Parts a, b, and c of Figure 1 show cyclic voltammograms (CVs) obtained with an Au(100) electrode modified with CoTPP, CuTPP, and CoPc in 0.1 M HClO4 at a scan rate of 50 mV s-1, respectively. The dotted traces in Figure 1 represent the CV for a bare Au(100). It is mostly featureless between 0 and 1.0 V, except for the anodic peak at 0.85 V. This feature stems from the structural transition from the reconstructed to the (1 × 1) phase, but the process is so slow that the CV appears to be irreversible at the scan rate used. These CVs are identical to those reported for a well-ordered Au(100) (31) Clavilier, J.; Faure, R.; Guinet, G.; Durand, R. J. Electroanal. Chem. 1980, 107, 205.

Co(II)- and Cu(II)-Coordinated TPP and Pc

Figure 1. Typical cyclic voltammograms of bare Au(100) (dashed line) and CoTPP (a), CuTPP (b), and CoPc (c) adsorbed (solid line) Au(100) electrodes in pure 0.1 M HClO4. The potential scan rate was 50 mV s-1.

electrode.32,33 Because the molecular adlayers were deposited in benzene solutions, the possibility of interference by benzene adsorption was scrutinized by running voltammograms of a Au(100) electrode after soaking in benzene for 10-60 s. The CV (not shown here) was essentially the same as that of a bare Au(100), indicating that benzene molecule was not adsorbed on Au(100). This finding agrees with the results obtained in UHV. The CV profiles obtained with an Au(100) electrode modified with CoTPP (a) and CoPc (c) were mostly featureless between 0 and 1.0 V. The lack of any feature of these profiles was somewhat unexpected, because these molecules contain an electroactive metal center, Co(II); so they were expected to give a peak near 0.22 V in the negative-going scan, as observed previously at Au(111).25,27 This phenomenon appears to indicate that the Co3+/2+ reduction reaction was kinetically slow at Au(100). However, the reason behind this effect of electrode orientation on the rate of electron transfer is not clear. Another peculiar feature of these profiles is the increase in the reduction current at the negative end, which should result from the evolution of hydrogen and the possible desorption of the admolecules. However, it is not clear why CoPc resulted in the most pronounced increase of current among these molecules. Furthermore, the peak at 0.85 V, associated with the lifting of reconstructed Au(100) surface, is either substantially reduced (Figure 1a and b) or totally disappeared (Figure 1c). These results imply that the Au(100) surface after being modified with these molecules remained in either the reconstructed (5 × 20) or the unreconstructed (1 × 1) state within the potential range between 0 and 1.0 (32) (a) Hamelin, A.; Weaver, M. J. J. Electroanal. Chem. 1987, 223, 171. (b) Hamelin, A. J. Electroanal. Chem. 1996, 407, 1. (33) Kolb, D. M.; Schneider, J. Electrochim. Acta 1986, 31, 929.

Langmuir, Vol. 20, No. 8, 2004 3161

V. Results from in situ STM imaging (vide infra) support this view, and it was the reconstructed (5 × 20) structure that predominated within this potential regime. A similar phenomenon was also noted in the case of Au(111), where the adsorption of a CoTPP molecular adlayer favors the reconstructed (x3 × 22) over the (1 × 1). Ultimately, the molecular adlayer might block the adsorption of anions, frequently a necessary condition for lifting the reconstruction. The adsorption of organics on gold electrode has been extensively examined, where the adsorbate exerts a strong impact on the stability of the atomic structure of the Au(100) substrate.34,35 It is concluded that the relative adsorption energy of an organic molecule on the reconstructed and ideal (1 × 1) surfaces dominates the structure of a gold electrode. If an organic adsorbate binds more strongly to the reconstructed phase, the lifting of reconstruction would shift positively, as compared to that of a bare gold electrode. In situ STM. Well-Defined Au(100) Surface. We first conducted STM imaging to ensure that the Au(100) substrate was structured as expected. The STM results are sufficient to show that the Au(100) surface was initially unreconstructed but changed into the reconstructed phase at more negative potentials.28-30 This potential-induced phase transition of Au(100) reflects the effects of charge density on the bonding energy between surface gold atoms and on the adsorption of anions.28-30 It is emphasized that one can differentiate between the two atomic structures of the Au(100) substrate by a simple examination of the surface morphology. The reconstructed Au(100) surface would be flat with the prevalent strand features with 0.07 nm corrugation. In contrast, the (1 × 1) surface contains a high density of mesas resulting from the 24% difference in the density of surface gold atoms between these two phases. An even higher resolution STM scan was able to reveal atomic details of the surface structure. However, we will not describe further details because they have already been reported in the literature.28-30 CoTPP Adlayer. Figure 2 shows STM images of a CoTPP adlayer on Au(100) prepared by immersing into a benzene dosing solution for 10 s. The potential of the gold electrode was held at 0.75 V, which is slightly more negative than the open-circuit potential, and the imaging parameters, such as feedback current and bias voltage, are typically 5 nA and -0.3 V, respectively. Figure 2a is an STM topography scan intended to reveal the gross surface morphology of the as-prepared Au(100) sample. Within the 250 × 250 nm2 area two uniform terraces separated by a step line marked by a white arrow in the diagonal direction of the image are clearly seen with only a few minor defects of dots and cracks. This homogeneous appearance of the surface clearly indicates that a reconstructed Au(100) surface existed under the monolayer of CoTPP. A faint modulation of intensity was found on the terrace. The higher resolution STM scan shown in Figure 2b indicates the formation of a highly ordered molecular array spanning 75 × 75 nm2 with only a few vacancy defects. The bright dots forming the periodical square array are ascribed to the CoTPP molecules. Occasionally, this structure was found to be near on a reconstructed Au(100) domain, for example, as observed in the upper portion of the STM image in Figure 2c. Although the underlying Au(100) was not imaged, the island-free surface morphology indicates that it was in the form of a reconstructed phase. Figure 2c further allows extrapola(34) Ho¨lzle, M. H.; Kolb, D. M. Ber. Bunsen-Ges. Phys. Chem. 1994, 98, 330. (35) Skoluda, P. Electrochem. Commun. 2003, 5, 142.

3162

Langmuir, Vol. 20, No. 8, 2004

Yoshimoto et al.

Figure 2. Large-scale (250 × 250 nm2) (a), middle-scale (75 × 75 nm2) (b), and small-scale (20 × 20 nm2) (c) STM images of the CoTPP adlayer on the Au(100)-(hex) surface in 0.1 M HClO4 acquired at 0.75 V versus RHE. The tip potentials and the tunneling currents were, respectively, (a and b) 0.45 V and 5.0 nA and (c) 0.35 V and 3.5 nA.

Figure 3. High-resolution (9 × 9 nm2) STM image (a) and structural model (b) of the CoTPP adlayer superimposed on the STM image of atomically resolved Au(100)-(hex) surface in 0.1 M HClO4, acquired at 0.75 V versus RHE. The tip potential and the tunneling current were 0.45 V and 3.5 nA, respectively.

tion of the gross registry of CoTPP admolecule with respect to the corrugated strands, if the reconstruction pattern is extended into the molecular arrays. It appears that the CoTPP molecules are located on the peaks of the corrugated strands. A further close-up view of the CoTPP array on the reconstructed Au(100) is shown in Figure 3a, where each individual CoTPP molecule is seen as a propeller with a bright dot at the center, which are ascribed, respectively,

to the peripheral porphyrin molecule and the Co(II) cation. This configuration resembles that of CoPc and CoTPP admolecules on Au(111) in UHV17,18 and in solution.25,27 The intermolecular spacing is measured to be 1.44 ( 0.05 nm. Figure 3b reveals a proposed model of this molecular square array superimposed on the image of a reconstructed Au(100) substrate. This assignment is in line with the result shown in Figure 3b. On the other hand, selective imaging with different STM parameters may be one approach for elucidating the registry of an admolecule, as demonstrated previously.36 Modulation of potential between 0 and 0.85 V did not affect the structure of this molecular adlayer. However, at potentials positive of 0.95 V, the reconstructed Au(100) was lifted, resulting in disordering of the adlayer. It is noteworthy that the lifting of reconstruction occurred at a potential more positive than that in 0.1 M HClO4, suggesting that the CoTPP overlayer expanded the stable range of potential for the reconstructed phase. This phenomenon unambiguously observed by the in situ STM is also in agreement with the voltammetric result, in which the phase transition feature at 0.85 V was largely suppressed (Figure 1a). The effect of organic adsorption on the structure of Au(100) has been extensively examined. Ultimately, the relative adsorption energy of organic adsorbate on the two phases of Au(100) determines the atomic structure of Au(100). For example, coumarin, cyclohexanone and γ-butyrolactone are shown to bind more strongly to the hex phase so that the lifting of reconstructed surface occurs at more positive potentials.35 As a corollary, CoTPP could be more strongly held at the reconstructed Au(100) surface, rendering the phase transition more difficult to proceed. Alternatively, these organic adlayers might inhibit the adsorption of anions, such as perchlorate in the present case. In fact, a similar phenomenon was noted for the adsorption of porphyrin on Au(111), where the (x3 × 22) reconstruction is lifted at 0.8 V, which is 0.2 V more positive than the potential of bare Au(111).25 On the other hand, the reason for the preference toward the structure of Au(100) is not clear. CuTPP Adlayer. We then studied the adsorption of a CuTPP adlayer on the reconstructed Au(100) to unveil the role of the metal cation in the molecular adsorption. The same procedure was used to prepare a CuTPP adlayer, and the in situ STM result (Figures 4 and 5) showed that this molecule also formed a long-range ordered molecular array. Although these images were obtained at 0.8 V in 0.1 M HClO4, the same structure predominated also at (36) Kunitake, M.; Batina, N.; Itaya, K. Langmuir 1995, 11, 2337.

Co(II)- and Cu(II)-Coordinated TPP and Pc

Langmuir, Vol. 20, No. 8, 2004 3163

Figure 4. STM image (20 × 20 nm2) of the CuTPP adlayer on reconstructed Au(100)-(hex) in 0.1 M HClO4 acquired at 0.8 V versus RHE. The tip potential and the tunneling current were 0.35 V and 15.0 nA, respectively.

Figure 5. High-resolution STM images (8 × 8 ) of CuTPP adlayer on reconstructed Au(100)-(hex) in 0.1 M HClO4 acquired at 0.8 V versus RHE. Images a and b correspond to those of regions I and II in Figure 4, respectively. The tip potential and the tunneling current were 0.35 V and 15 nA, respectively. nm2

other potentials, as long as it is not more positive than 0.95 V. Further close-up views of the molecules in regions I and II in Figure 4 are shown in Figure 5a and b, respectively. The internal molecular structure is clearly seen. The four spots located at the four corners of molecular squares are assigned to the four phenyl rings lying roughly perpendicularly to the porphyrin ring. The unit length of the as-marked square lattice is 1.45 nm, the same as that of CoTPP. Each CuTPP molecule was also found to be of a propeller shape, indicating the parallel adsorption configuration of CuTPP molecule. It is likely that the most efficient molecular packing on the surface is obtained when all molecules are rotated with their C2 axis intersecting the substrate atomic rows at an angle of about 30° (clockwise for Figure 5a or anticlockwise for Figure 5b). These STM images are almost the same as those observed on Au(111), presented in our recent paper.25 The adlattices of CoTPP and CuTPP are practially indistinguishable, except in their molecular resolution images. More specifically, the center of a CuTPP molecule or the location of the Cu2+ cation appears as a depression, whereas the cobalt ion in CoTPP is apparently protruded. This difference in three-dimensional configuration between these two molecules was previously noted also on Au(111).14,15 This result is explained by an electronic effect arising primarily from the difference in the occupation of the d orbitals.14-19 The electronic configuration of CoPc is dxz2, dyz2, dxy2, dz21, whereas that of CuPc is dxz2, dyz2, dxy2,

Figure 6. Large-scale (50 × 50 nm2) STM images of CuTPP adlayer on Au(100)-(1 × 1) surface in 0.1 M HClO4 acquired at 0.95 V (a) and at 0.75 V stepped from 0.95 V (b) versus RHE, respectively. The tip potential and the tunneling current were 0.35 V and 10 nA, respectively. The set of two arrows indicate the close-packed directions of the Au(100) substrate.

Figure 7. Large-scale (50 × 50 nm2) STM image of the CoPc adlayer on Au(100)-(1 × 1) surface in 0.1 M HClO4 acquired at 0.75 V versus RHE. The tip potential and the tunneling current were 0.43 V and 15 nA, respectively. The two arrows indicate close-packed directions of the Au(100) substrate.

dz22. The half-filled dz2 orbital in CoTPP is believed to mediate electron tunneling, which leads to the protruded appearance in the STM image.16 This electron-mediating channel apparently does not exist for CuTPP, giving the depressed appearance in the STM image. The structure of CuTPP adlayer was also very stable on the reconstructed Au(100) surface; no structural change

3164

Langmuir, Vol. 20, No. 8, 2004

Yoshimoto et al.

Figure 8. STM image (25 × 25 nm2) (a) of the CoPc adlayer on Au(100)-(1 × 1) surface in 0.1 M HClO4 acquired at 0.75 V versus RHE and the corresponding structural model (b). The tip potential and the tunneling current were 0.43 V and 15 nA, respectively.

within the adlayer and the gold substrate was observed in the potential range between -0.1 and 0.85 V. However, applying a potential equal to or more positive than 0.9 V destroyed the ordered molecular overlayer and formed many monatomically high mesas on the terraces, as revealed by the STM image in Figure 6a. The emerging mesas are indicative of the lifting of reconstruction of Au(100) to the (1 × 1) phase. This phase transition was reversible with respect to potential, as locally ordered arrays of CuTPP appeared again at a more negative potential (0.75 V), as seen in the STM image of Figure 6b. These two images were not acquired at the same location on the Au(100), and therefore they have rather different morphologies. Also, it is believed that the admolecules were not desorbed at 0.9 V, and they simply became disordered. This is indicated by the fact that the ordered molecular overlayer was restored during the prolonged imaging at 0.75 V. On the other hand, it is not clear why CuTPP did not form an ordered 2D structure on the (1 × 1) phase.37 CoPc Adlayer. To explore the effect of molecular structure on the spatial arrangement and the atomic structure of the supporting gold substrate, we investigated the adsorption of CoPc on Au(100). The Au(100) electrode was made to be reconstructed, and the dosing of CoPc was done in a benzene solution, as described earlier. Figure 7 shows a large-scale STM image of a CoPc adlayer on Au(100) acquired at 0.8 V in 0.1 M HClO4. The CoPc admolecules imaged as bright spots were found on the terraces and islands. The size and shape of the island are evidently poorly defined, but they all exhibit the same height (∆z ) 0.23 nm). These areas of the protruded mesas add up to account for ca. 24% of the surrounding substrate, which roughly matches the excess in atomic density of the hex phase with respect to the (1 × 1) phase. In other words, the CoPc adlayer, prepared by using the same procedure as was used to produce CoTPP and CuTPP adlayers, was formed on (1 × 1), suggesting that the adsorption of CoPc simultaneously caused the lifting of reconstruction. Intriguingly, a similar phenomenon was also noted when CoPc was adsorbed on Au(111)(x3 × 22), as reported in our recent paper.27 This result is in marked contrast with the results obtained with CoTPP and CuTPP, which formed long-range ordered adlattices on reconstructed Au(100). (37) Yoshimoto, S.; Tada, A.; K.; Itaya, K. Manuscript in preparation.

A modulation of potential was applied to the Au(100) electrode to investigate the stability of the CoPc adlayer and the Au(100)-(1 × 1) phase. Remarkably, the CoPc adlayer was so stable that no change occurred between 0 and 1.0 V. In other words, the Au(100)-(1 × 1) surface, which is normally reconstructed into the hexagonal phase at potentials negative of 0.45 V in 0.1 M HClO4, remained stable at potentials as negative as 0 V in the presence of a CoPc adlayer. To the best of our knowledge, such a high stability of the reconstructed phase of Au(100) has not been observed thus far in an electrochemical environment. This result implies that CoPc molecule binds much more strongly on the reconstructed Au(100) surface than on the unreconstructed phase. The evenly distributed mesas on the terrace made it difficult to form a well-ordered molecular array and also to achieve high-quality molecular resolution STM. Occasionally, CoPc molecule was imaged as a propellershaped object, as seen in some portions (e.g., the area circled by a dotted line in Figure 8a). Interpretation of the STM result is straightforward. The center protrusion and the four weaker spots at the four corners of squares are readily associated with the Co ions and phenyl moieties of the Pc molecules, respectively. It is stressed, however, that details in the sample preparation procedure can influence the spatial molecular arrangements of CoPc. For example, CoPc adlayer was found to be square-like on reconstructed Au(111) at an immersion time less than 5 min, but a nearly close-packed hexagonal phase was produced when the dosing time was increased to 10 min.27 The adsorption of CoPc on reconstructed Au(100) was attempted by reducing the dosing time to less than 3 min, but the results showed that the (1 × 1) phase always prevailed even with a dosing time as short as 1 min. This result indicates that the adsorption of CoPc monolayer increased the mobility of gold atoms. The intermolecular distance between nearest neighbor CoPc molecules in this adlayer was about 1.65 nm, which is greater than the adlattice constant of CoTPP or CuTPP on reconstructed Au(100) as described above. The intermolecular spacings between CoPc molecules are 2.05 ( 0.05 and 2.61 ( 0.07 nm, respectively, in the [011h ] and the [01 h1 h ] directions, which correspond to 7 and 9 times the Au lattice constants. This local structure is thus assigned as c(7 × 9)rect, whose real-space structural model is shown in Figure 8b. This model features a high symmetry of the Pc molecules with respect to the Au(100) substrate, where

Co(II)- and Cu(II)-Coordinated TPP and Pc

all phenyl moieties can occupy 2-fold bridge sites. This assignment is supported by the two possible rotational domains imaged in Figure 8a. The fact that the same structure was also observed for CuPc on Au(100)-(1 × 1) suggests that the real-space structures of Pc complexes are largely determined by the framework of the organic portion of the molecule rather than by the center cations. Furthermore, in view of the known fact that vapor deposition of CuPc on reconstructed Au(100) surface in UHV produces a (5 × 5) structure,38 it is clear that details in the sample preparation procedure can control the amounts and structures of these adsorbate molecules. Conclusions Immersing a Au(100) electrode in a benzene solution containing CoTPP, CuTPP, or CoPc molecules produces a highly ordered adlayer of CoTPP or CuTPP on the reconstructed Au(100) surface and a local ordering of CoPc on the unreconstructed Au(100)-(1 × 1) surface. Molecular resolution STM has revealed the internal molecular structure of each individual admolecule in 0.1 M HClO4 under potential control. CoTPP and CuTPP adapt the same spatial arrangement on the reconstructed Au(100) surface with both molecules located on the peak of corrugated strands of gold atoms. The spacing of 1.44 nm between two neighboring molecular rows equals to the distance between two neighboring strands of reconstructed Au(38) Auerhammer, J. M.; Knupfer, M.; Peisert, H.; Fink, J. Surf. Sci. 2002, 506, 333.

Langmuir, Vol. 20, No. 8, 2004 3165

(100). The metal centers of Co(II) and Cu(II) in the CoTPP and CuTPP admolecules appear as a protrusion and depression, respectively, in the molecular resolution STM. The highly ordered CoTPP arrays stabilize reconstructed Au(100) at potentials up to 0.85 V, whereas those of CuTPP sustain the structure even at a potential as positive as 0.9 V. The adsorption of CoPc molecule results in a rapid phase transition from the hexagonal phase to the (1 × 1) phase of the Au(100) substrate, leading to the formation of an only locally ordered adlayer. Molecular structure apparently governs the interaction between the admolecule and the gold substrates. This study further illustrates the feasibility and readiness of using the solutionphase dosing to produce a well-ordered molecular adlayer, which may be important in the development of molecular electronics. Acknowledgment. This work was supported in part by the Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Agency (JST) and by the Ministry of Education, Culture, Sports, Science and Technology, a Grant-in-Aid for the Center of Excellence (COE) Project, Giant Molecules and Complex Systems, 2003. The authors acknowledge Dr. Y. Okinaka for his assistance in writing this manuscript and Dr. J. Inukai of Tohoku University for his useful discussion. LA0359474