Metal d-orbital Occupation Dependent Images in the Scanning Tunneling Microscopy of Metal Phthalocyanines.

Copyright: Journal of Physical Chemistry (1996)

by
K.W. Hipps,* Xing Lu, X.D. Wang,+ and Ursula Mazur
Department of Chemistry and Materials Science Program
Washington State University
Pullman, WA 99164-4630

ABSTRACT

A strong d-orbital dependence in the scanning tunneling microscopy image of metal phthalocyanines is demonstrated. Unlike copper phthalocyanine (CuPc) wherein the central metal appears as a hole in the molecular image, the cobalt atom in CoPc is the highest point (about 0.3 nm) in the molecular image. On the other hand, the phenyl ring regions of CoPc and CuPc appear to have the same height. These data are consistent with theoretical calculations that predict a large contribution of cobalt d-orbitals near the Fermi energy. An intriguing aspect of this work is that it may be possible to chemically identify the different metal phthalocyanines simply by their appearance. This is demonstrated for the case of a mixed monolayer of CuPc and CoPc on Au(111).

*) Author to whom correspondence is to be addressed.
+) Current Address: Charles Evans and Associates, Redwood City, CA

INTRODUCTION

Metal phthalocyanines are of great technological and fundamental interest. A schematic structure of a typical metal(II) phthalocyanine is presented in Figure 1 . Recent theoretical studies of the electronic structure of metal phthalocyanines include, but are not limited to, density functional treatments [1,2] unrestricted (open shell) Hartree-Fock calculations, multi-configuration SCF, and extended Huckel MO based elastic scattering quantum chemical calculations. [3-5] The phthalocyanines are models for biologically important species such as porphyrins, hemoglobin, and chlorophyll, and they are extensively used as pigments and dyes. They can serve as the active elements in chemical sensors, and are of great interest in optoelectronic devices and solar cells. [6-8] Their catalytic properties have been of interest for some time, most recently for redox catalysis such as in fuel cell applications.[9-12] They have interesting semiconductivity, and can be used to form well behaved field effect transistors. [13-14]

An understanding of the interaction between metal phthalocyanines (MPc) and surfaces is a critical element required for optimizing their use in the many of the applications listed above. Two elements of interest are the nature of the bonding between the MPc and the support, and the structure of the MPc-support entity. In principal, both of these issues can be addressed by obtaining sub-molecular resolution images of MPc adsorbed on the substrate of interest. Until now, however, only limited information about the chemical nature of the adsorbed MPc was extracted by STM studies.

STM images of individually distinguishable copper phthalocyanine (CuPc) molecules have been presented by Gimzewski and co-workers, Moeller, Lippel et al., Ludwig and associates, Fritz and co-workers, Kanai et al., and Petracek. [15-21] In addition, images of the free acid (H2Pc) co-adsorbed on graphite with a liquid crystalline carrier have also been published by Freund and associates. [22] While there have been reports of PbPc studied by STM, no molecular images have been observed. [23,24] To our knowledge, no other phthalocyanine systems have been studied by STM. In all of these cases, the predominant features of the molecular image could be understood based on the organic material alone. Theoretical calculations of the STM image of CuPc predict that there should be an apparent hole in the center of the molecule [5,17] and this is what is normally observed. [17,19] Even for H2Pc, there is an apparent hole in the center of the molecule [22] The explanation for these 'holes', it seemed to us, was that both the occupied and unoccupied orbitals localized on Cu lay 1 eV or more from the Fermi energy, while the MPc ligand LUMO lay close to the Fermi energy.[25,26] We reasoned that MPc systems having a greater metal d-orbital participation in molecular orbitals near the Fermi surface should show profoundly different STM images. Alternatively, dramatic changes in the apparent molecular shape might also occur in systems where interactions between the metal d-orbitals and a metallic substrate were significant. In the latter case, the metal surface density of states might 'shine through' giving enhanced height to the central metal.

As a test of this concept, a study of various metal phthalocyanines was initiated. The first system chosen was cobalt phthalocyanine (CoPc), wherein cobalt has a d7 configuration.. Figure 2 depicts the occupied (heavy lines) and unoccupied levels close to the Fermi energy as determined by the UHF-X* calculation of Reynolds and Figgis. [3] (The levels have been shifted by about 1.5 eV to account for the polarization energy in the solid state.) It is clear from this open shell calculation that there should be considerable d-orbital participation in both the HOMO and LUMO orbitals, making CoPc an excellent candidate for observing direct d-orbital charge density. In this letter we will demonstrate that the sub-molecular resolution STM image of CoPc on Au(111) is dramatically different from that of CuPc on the same substrate.

EXPERIMENTAL

In a single continuous operation, 0.4 nm of CoPc, CuPc, or a mixture of both were deposited onto a 100 nm thick layer of Au(111) epitaxially grown on mica. This was accomplished in a liquid nitrogen trapped diffusion pumped bell jar system having a base pressure of about 2x10-7 torr. In detail, the mica was heated to 550 C in vacuum for a period of 2 hours and then allowed to cool to 375 C. The Au layer was deposited at this temperature at a rate of 7.8 nm/minute. The substrate was then allowed to cool to 180 C and the MPc was vapor deposited at a rate of 1.5 nm/minute from an ME-1 source (R. D. Mathis). The sample was then allowed to cool in vacuum. The metal phthalocyanines were purified by multiple sublimation before use.

The completed sample was removed from the preparation chamber, mounted on a sample carrier, immediately inserted through a load lock into the UHV chamber housing a McAllister STM. Both W and Pt/Ir tips were used after electron bombardment in UHV. The STM is controlled by a Digital Instruments Nanoscope III. All the data reported here was obtained using a Pt/Ir tip, but similar results have been obtained using W tips. Typical experimental parameters were a sample bias of -0.100 V, a current of 1.0 nA, and a 4 Hz scan rate. Current voltage, i(v), curves were taken using the standard Nanoscope software.

RESULTS AND DISCUSSION

While a large number of pure CoPc/Au(111) and CuPc/Au(111) images were recorded, these results can accurately and compactly be portrayed using the mixed complex film results. Figure 3 presents a surface plot of a small section of the MPc coated Au(111) surface. This image was Fourier filtered. Notice first that the four-leaf clover like shape of the MPc molecules is clearly apparent. Note also that some of the molecules have an intense bright (high) area in the center while others have a pronounced dark region (hole). Based on our experience with chemically pure samples, the bright centered molecules are CoPc while those with an apparent hole in the center are, as expected, CuPc. Thus, chemical analysis of these mixed films can be performed in a trivial way. It is also interesting to note that larger area images show the CuPc distributed in small apparently randomly sized clusters within the CoPc, suggesting weak differences in the lateral forces between the MPc species.

Figure 4 is a combined top-view and cross-sectional plot of a different area of the same film. This image was not Fourier filtered. The heavy line in the image served as the cut for the cross section. Note that (from top to bottom) it passes through a phenyl group of CoPc, the central Co, and the opposite phenyl group. It later crosses a phenyl group of a CuPc molecule, the central Cu, and the opposite phenyl group of the CuPc. The positions of the central cobalt and copper are shown by triangular markers in the sectional plot. From Figure 4 , the phenyl groups have an apparent height of about 0.1 nm above the gold surface, independent of the nature of the central metal. The copper appears to have near zero height while the cobalt towers over the surface with a 0.3 nm total elevation.

It is clear from these results that our conjecture regarding the role of d-orbitals in the STM image must have some validity. On the other hand, there are at least three separate mechanisms (consistent with our conjecture) that could lead to the observed differences in height. If the tunneling current is primarily LUMO mediated, then the bright central region would be due to the 3dz2 orbital of cobalt. If the HOMO is the critical participant, then the dxy, dxz, and dyz orbitals provide the observed contrast. Alternatively, it may be that the unpaired electron on cobalt forms a partial bond with the underlying gold, as has been suggested for CoPc on platinum electrodes.10 Thus, enhanced electron density in the dz2 orbital could also lead to the apparent height of the central atom. The combination of higher resolution STM images (the dxy, dxz, and dyz orbitals should all have a small hole in their center) and current-voltage, i(v), data (unoccupied orbitals enhance the current when the sample is biased positively) should resolve this issue. To date, we have been frustrated in obtaining good i(v) data by the difficulty in keeping the tip clean. Some i(v) curves show a marked current increase at about +0.5 volts bias while others show a similar change at -0.5 volts. We interpret this to mean that we sometimes have MPc molecules on the tip. We are currently working to solve this problem and also to obtain higher resolution images of a range of metal phthalocyanines.

ACKNOWLEDGMENT

We thank the American Chemical Society Petroleum Research Fund and the National Science Foundation for their assistance in the form of grants PRF-25763-AC6, DMR-9201767 and DMR-9205197. We also thank Dr. Matthew Antonik for his assistance.

FIGURE CAPTIONS

Figure 1. . Molecular structure of a typical metal(II) phthalocyanine.

Figure 2. Spin-orbital energies of cobalt(II) phthalocyanine near the Fermi energy. Heavy lines depict occupied levels while the thinner ones represent virtual (empty) orbitals. This diagram is based on data taken from reference 3.

Figure 3. Surface plot of a mixed CoPc and CuPc monolayer on Au(111). The gray scale extends over a range of 0.5 nm. The image was Fourier filtered.

Figure 4. Top View and cross sectional plot of a mixed CoPc and CuPc monolayer on Au(111). The gray scale extends over a range of 0.5 nm. This data was not Fourier filtered.

  1. ) Rosa, A.; Baerends, E. J. Inorg. Chem. 1994, 33, 584.
  2. ) Rosa, A.; Baerends, E. J. Inorg. Chem. 1992, 31, 4717.
  3. ) Reynolds, Philip A.; Figgis, Brian N. Inorg. Chem. 1991, 30, 2294.
  4. ) Carniato, S.; Dufour, G.; Rochet, F.; Roulet, H.; Chaquin, P.; Giessner-Prettre, C. J. Electron Spectrosc. Relat. Phenom. 1994, 67, 189.
  5. ) Sautet, P.; Joachim, C.; Bocquet, M. L.; Salmeron, M. Ann. Chim. (Paris) 1992, 17, 217
  6. ) Collins, R.A.; Mohammed, K.A. J. Phys. D. ,1988, 21, 142.
  7. ) Flynn, B.W.; Owen, A.E.; Mayor, J. J. Phys. C. ,1977, 10, 4051.
  8. ) Ghosh, A.H.; Morel, D.L.; Feng, T.; Shaw, R.F.; Rowe, C.A. J. Appl. Phys. 1974, 1, 20.
  9. ) "Katalyse an Phthalocyaninen," Kropf, H.; Steinbach, F. Ed.; Georg Theme Verlag, Stuttgart (1973).
  10. ) Mho, Sun-il; Ortiz, Bertha; Park, Su-Moon; Ingersoll, David; Doddapaneni, Narayan J. Electrochem. Soc. 1995, 142, 1436.
  11. ) Zagal, J.H. Coordin. Chem. Rev. 1992, 119, 89.
  12. ) "Phthalocyanine Research and Applications," Ed: Thomas, A.L.; CRC Press, Boca Raton FL (1990).
  13. ) Madru, R.; Guillaud, G.; Sadoun, M.; Maitrot, M.; Andre, J.J.; Simon, J.; Even, R. Chem. Phys. Lett. 1988, 145, 343.
  14. ) Maitrot, M.; Guillaud, G.; Boudjema, B.; Andre, J.J.; Strzelecka, H.; Simon, J.; Even, R. Chem. Phys. Lett. 1987, 133, 59.
  15. ) Gimzewski, J. K.; Stoll, E.; Schlittler, R. R. Surf. Sci. 1987, 181, 267.
  16. ) Moeller, R,; Coenen, R.; Esrlinger, A.; Kaslouski, B. J. Vac. Sci. Technol. A 1990, 8, 659.
  17. ) Lippel, P. H.; Wilson, R. J.; Miller, M. D.; Woell, C.; Chiang, S. Phys. Rev. Lett. (1989), 62, 171.
  18. ) Ludwig, C.; Strohmaier, R.; Petersen, J.; Gompf, B.; Eisenmenger, W. J. Vac. Sci. Technol. B., 1994, 12, 1963
  19. ) Fritz, Torsten; Hara, Masahiko; Knoll, Wolfgang; Sasabe, Hiroyuki Mol. Cryst. Liq. Cryst. Sci. Technol. Sect. A 1994, 252-253, 561.
  20. ) Kanai, M.; Kawai, T.; Motai, K.; Wang, X. D.; Hashizume, T.; Sakura,T. Surf. Sci. (1995), 329, L619.
  21. ) Petracek, V. Int. J. Electron. 1995, 78, 267.
  22. ) Freund, J.; Probst, O.; Grafstroem, S.; Dey, S.; Kowalski, J.; Neumann, R.; Woertge, M.; zu Putlitz, G. J. Vac. Sci. Technol. B, 1994, 12, 1914.
  23. ) Pester, O.; Mrwa, A.; Hietschold, M. Phys. Status Solidi A 1992, 131, 19.
  24. ) Hamann, C.; Laiho, R.; Mrwa, A. Phys. Status Solidi A (1989), 116(2), 729
  25. ) Mazur, Ursula; Hipps, K. W. J. Phys. Chem. 1995, 99, 6684.
  26. ) Mazur, Ursula; Hipps, K. W. J. Phys. Chem. (1994), 98(33), 8169