Abnormal Temperature Dependence of Photoemission Intensity

David Wisbey, Danqin Feng, Marshall T. Bremer, Camelia N. Borca Anthony N. Caruso, Carter M. Silvernail, John Belot, Elio Vescovo, Laurent Ranno, and ...
0 downloads 0 Views 138KB Size
1095

2006, 110, 1095-1098 Published on Web 12/30/2005

Abnormal Temperature Dependence of Photoemission Intensity Mediated by Thermally Driven Reorientation of a Monomolecular Film D.-Q. Feng,† D. Wisbey,† Y. Tai,§ Ya. B. Losovyj,‡ M. Zharnikov,*,§ and P. A. Dowben*,† Department of Physics and Astronomy and the Center for Materials Research and Analysis, UniVersity of NebraskasLincoln, Lincoln, Nebraska 68588-0111, Center for AdVanced Microstructures and DeVices, Louisiana State UniVersity, 6980 Jefferson Highway, Baton Rouge, Louisiana 70806, and Angewandte Physikalische Chemie, UniVersita¨t Heidelberg, Im Neuenheimer Feld 253, D-69120 Heidelberg, Germany ReceiVed: NoVember 17, 2005; In Final Form: December 15, 2005

The photoemission intensities of organic molecular layers generally obey the Debye-Waller temperature dependence but not always. With the example of a monomolecular film formed from [1,1′;4′,1′′-terphenyl]4,4′′-dimethanethiol, we show that pronounced deviations from Debye-Waller temperature behavior are possible and are likely caused by temperature-dependent changes in molecular orientation.

Although organic adsorbates and thin films are generally regarded as “soft” materials, the effective Debye temperature, indicative of the dynamic motion of the lattice normal to the surface, can be very high, e.g., in the multilayer film formed from [1,1′-biphenyl]-4,4′-dimethanethiol (BPDMT).1 The effective Debye temperature, determined from core level photoemission from the all-carbon arene rings, is comparable to that of graphite and follows the expected Debye-Waller behavior for the core level photoemission intensities with temperature. We associate this rigidity to the stiffness of the benzene rings and the ordering in the ultrathin multilayer molecular thin film. Temperature-assisted conductivity and molecular configurational changes can both occur in molecular systems and need not always be directly correlated.2,3 In addition, electronphonon coupling in molecular systems must of necessity consider local point group symmetry effects.4 Setting aside these known complications, failure to obey an expected temperature dependence of the photoemission intensities, consistent with the Debye-Waller scattering, can provide a very strong indication of molecular conformational changes. An interesting candidate to prove this hypothesis is the self-assembled monolayer (SAM) of [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT), which exhibits structural and conformational changes upon metal evaporation and temperature variation.5,6 Note that the orientation of the TPDMT molecules in the respective SAM is very different from that of the BPDMT moieties in the multilayer films,7,8 even though the structures of the BPDMT and TPDMT molecules are quite similar. [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT) films were prepared by immersion of evaporated Au(111) substrates in solution, as described elsewhere.5,9-12 Deposition of [1,1′biphenyl]-4,4′-dimethanethiol was accomplished by a solution method, as described by Eck et al.13 on the gold substrates, as * Corresponding authors. P.D.: tel, 402-472-9838; fax, 402-472-2879; e-mail, [email protected]. M.Z.: tel, 49-6221-544921; e-mail, [email protected]. † University of NebraskasLincoln. ‡ Louisiana State University. § Universita ¨ t Heidelberg.

10.1021/jp0566616 CCC: $33.50

carried out in prior studies.1,7,8 It is, in principle, possible to fabricate a BPDMT SAM, similar to the TPDMT case,9 but, in our case, due to specific parameters of the preparation procedure, multilayer BPDMT films were fabricated. We combined photoemission and inverse photoemission studies to characterize the molecular orbital placement of both occupied and unoccupied orbitals of the adsorbed molecules on Au(111) surfaces. All inverse photoemission (IPES) spectra were obtained with the electron gun at normal incidence and the detector positioned at 45° off the surface normal, as described elsewhere.2,3,6-8,14 The high-resolution ultraviolet photoemission spectroscopy (UPS) was accomplished using a helium lamp at hν ) 21.2 eV (He I) and a Scienta 200 hemispherical electron analyzer with a combined resolution better than 10 meV.4 The light polarization dependent ultraviolet photoemission spectroscopies were carried out using synchrotron light, dispersed by a 3 m toroidal grating monochromator, at CAMD, as described in detail elsewhere.2,3,6-8,15 The measurements were performed under ultrahigh vacuum conditions (P ) 3 × 10-10 Torr) employing a hemispherical electron energy analyzer with an angular acceptance of (1°. All angles (both light incidence angles as well as photoelectron emission angles) reported herein are with respect to the substrate surface normal, designated as the z-axis throughout this work. Binding energies are referenced with respect to the Fermi edge of gold or tantalum, and all photoelectrons were collected normal to the substrate surface and at normal incidence for inverse photoemission (k| ) 0 or Γ h ) to preserve the highest possible local point group symmetry. Temperature-dependent X-ray photoemission spectroscopy (XPS) studies were taken at 60° incidence angle with the photoelectrons collected along the film surface normal, as described elsewhere.1 This photoemission geometry provided the greatest electron momentum transfer along the surface normal. Theoretical calculations of the ground-state molecular orbital electronic structure of [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT) were undertaken by PM316,17 with the HyperChem © 2006 American Chemical Society

1096 J. Phys. Chem. B, Vol. 110, No. 3, 2006

Figure 1. Combined UPS (left) and IPES (right) spectra (top) of TPDMT/Au, along with the calculated ground-state molecular orbital energies for a single TPDMT molecule (bottom line). The UPS and IPES signals are related to the contributions from the occupied and unoccupied molecular orbitals, respectively. The experimental HOMOLUMO gap and the calculated molecular orbitals of TPDMT are indicated (see text). The photoemission spectrum was taken with unpolarized He I (21.2 eV), with the photoelectrons collected normal to the surface, and the inverse photoemission spectrum was taken at normal incidence. The inset is a schematic of [1,1′;4′,1′′-terphenyl]4,4′′-dimethanethiol (TPDMT).

package, as has been completed successfully elsewhere with molecular adsorbates.2,3,6,14,18 Geometry optimization of the molecular system was performed by obtaining lowest unrestricted Hartree-Fock (UHF) energy states and their possible molecular symmetries. A calculated density of molecular orbital states was obtained by applying equal Gaussian envelopes of 1 eV width to each molecular orbital at the appropriate binding energy (to account for the solid-state broadening in photoemission), and then summing intensities (as has been done elsewhere3,6). This calculated density of states is then combined together with a rigid energy shift applied to the calculated electronic structure, of a value typically about 4.7 eV. As seen in Figure 1, where the UPS and IPES spectra of the TPDMT film are presented, along with the calculated electronic structure, there is good qualitative agreement between theory and experiment, although we note that the calculations do not account for photoemission and inverse photoemission matrix element effects. The experimental gap between the highest occupied and lowest unoccupied molecular orbitals (the HOMO-LUMO gap) is 6.8 ( 0.2 eV for TPDMT, which is in good agreement with the theoretical estimate of the HOMO-LUMO gap of 7.6 eV. In the ground state of TPDMT on Au, the energy distance from the LUMO to EF is 3.2 eV, which suggests that this monomolecular film could act as a dielectric barrier layer in tunnel magnetoresistive devices. We can assign the various photoemission and inverse photoemission features, by comparing to the occupied and unoccupied molecular orbitals derived from semiempirical ground-state calculation, as shown in Figure 1. The features at binding energies (E - EF) relative to the Fermi level of -3.6 ( 0.1 eV, -5.0 ( 0.2 eV, -6.4 ( 0.1 eV, -7.9 ( 0.2 eV, -8.7 ( 0.1 eV, and -9.9 ( 0.2 eV, in the photoemission and inverse photoemission spectra (Figure 1), can be assigned to groups of molecular orbitals. Exploiting the known orientation of [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT) in self-

Letters

Figure 2. Comparison of the light polarization photoemission spectra of TPDMT taken with p polarized light (black lines) and s + p polarized light (red lines) as a function of temperature. In (a), the energy region with the temperature-reversible polarized light effects are enhanced compared to the broader energy scale in (b). The proposed model for the adlayer structure of TPDMT at 295 K, in accordance with ref 9, is shown in the inset. The spectra were taken at photon energy of 80 eV.

assembled monolayer (SAM) films, where the molecules adopt a dense and regular packing with an upright orientation of the terphenyl backbone somewhat (about 19.3°) canted off the surface normal and bound to the substrate via the thiolate linkage,3-5,9-12 the pronounced photoemission feature at -3.6 ( 0.1 eV (Figure 2) can be attributed to largely occupied molecular orbitals of largely sulfur px,y. Such a feature could exhibit some light polarization dependence in photoemission, if the TPDMT molecules adopt a strongly upright orientation. Similarly, the photoemission feature at -6.4 ( 0.1 eV is dominated by the arene (benzene) occupied π molecular orbitals with a contribution from the methyl σ bonds and sulfur pz, and is thus less likely to exhibit light polarization photoemission regardless of the molecular orientation. Occupied molecular orbitals containing strong benzene π and σ bond weight contribute to the photoemission feature at -8.7 ( 0.1 eV (Figure 2) that can also exhibit some light polarization dependent photoemission effects: an enhancement of this feature in s-polarized light is expected for an “upright” molecular orientation. Because of the off-normal canting (19.3°) of the TPDMT molecules with the long molecular axes slightly tilted away from the substrate at room temperature,9 the light polarization dependent photoemission effects, such as they are, are expected and observed to be quite small. This does not mean that such effects are totally absent, as seen in Figure 2 (where the photoemission spectra of the TPDMT film taken with linearly polarized p polarized light (black) and s + p polarized light (red) at different temperatures are presented). Small temperaturedependent light polarization effects in the photoemission spectra of TPDMT can be observed, as shown in Figure 2a. The slight enhancement of the photoemission feature at -3.6 ( 0.1 eV (E - EF) obtained with a large light incidence angle of 70° (p polarized light) relative to a smaller incidence angle of 45° (s + p polarized light) indicate a preferential orientation of the TPDMT molecules in the film with the long molecular axis placed “upright”, and this is consistent with the near edge X-ray

Letters

J. Phys. Chem. B, Vol. 110, No. 3, 2006 1097

absorption fine structure spectroscopy results.9 This photoemission dichroism effect is too small to claim a molecular orientation of the long molecular axis in an orientation close (within 10°) to the surface normal. This small enhancement of the -3.6 ( 0.1 eV feature in p polarized light is not preserved at all temperatures (e.g., the spectra at 235 K in Figure 2a). Temperature-dependent changes result in the feature at 8.8 ( 0.1 eV becoming intense for s + p polarized light at room temperature but enhanced in p polarized light at lower temperatures. Taken together, these subtle changes in the light polarization dependent photoemission suggest changes in the relative orientation of aromatic rings and/or the twist of aromatic rings with temperature in the TPDMT molecules. Note that the observed temperature-dependent changes are fully reversible. There are other less subtle indications of molecular configuration changes with temperature. It is of interest to examine whether these supposed orientational and conformational changes can affect the standard temperature dependence of the photoemission intensity: the experimental consequences of the effective film Debye temperature. In a general case, the intensity of an emitted or scattered electron beam exponentially decays with increasing temperature1,19-26

I ) I0 exp(-2W)

(1)

where W is the Debye-Waller factor, which is given by

2W ) |∆k|2 〈uo〉2

(2)

where ∆k is the wave vector transfer and 〈uo〉2 is the meansquare displacement of the atoms. Within the Debye model of thermal vibrations, in the case of isotropic vibrations, W is described as

2W )

3p2T(∆k)2 mkBθD2

(3)

where T is the temperature of the sample (in Kelvin), m is the mass of the scattering center, and again θD is the Debye temperature which is the signature of the dynamic motion of vibrational modes normal to the surface. In photoemission measurements, the momentum transfer is the momentum of the emitted photoelectron,1,20-23,26 and the mass of the scattering center is the mass of the specific element which is the origin of the emitted photoelectron.1,26 After background (Ibg) subtraction and normalization to the peak integral intensity at the lowest temperature I0, we have plotted in Figure 3 the integral C 1s and S 2p XPS intensities for both a self-assembled monolayer of [1,1′;4′,1′′-terphenyl]4,4′′-dimethanethiol (TPDMT) and a multilayer of [1,1′biphenyl]-4,4′-dimethanethiol (BPDMT), both on Au(111). As shown in Figure 3, both C 1s and S 2p photoemission intensities of the BPDMT multilayer do exhibit the expected temperature dependence, in accordance with eq 1. The effective Debye temperature derived from the C 1s data is quite high,1 which means that the mean-square displacement of the carbon atoms in the biphenyl moieties of the BPDMT molecule is anomalously small. The respective effective Debye temperature is smaller for the sulfur, abstracted from the S 2p core level, than is the case for the BPDMT carbon atoms. We associate the high Debye temperature of the carbon to the rigidity of BPDMT and the stiffness of the benzene rings, as well as the ordering in the molecular thin film.1 The Debye temperature of sulfur is much

Figure 3. Logarithm of the C 1s (a) and S 2p (b) total photoemission intensities for the [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT) (red triangles), and biphenyldimethyldithiol (BPDMT) (black dots and circles indicating two different representative sets of data) as a function of temperature. The effective Debye temperatures for the multilayer BPDMT films were determined in accordance with eq 3: 1397 ( 110 K from C 1s core level intensity and 456 ( 50 K from the S 2p signal.

lower, as a result of the pendant character of the sulfur end groups on the BPDMT molecule in a multilayer molecular film. In contrast to BPDMT, the C 1s and S 2p core level photoemission intensities for the TPDMT film do not follow the expected temperature dependence as given by eq 1. The observed behavior can only mean that either the effective thickness of the TPDMT film or one of the parameters in eq 3 vary with the temperature. Since no temperature-induced decomposition of the TPDMT film occurs at the given, rather moderate, temperatures, and most of the parameters in eq 3 are related to the photoemission experiment and were kept fixed during the measurements, it should be a parameter related to the sample, of which we have only one, viz., the Debye temperature. Obviously, the effective Debye temperature is not constant (and perhaps not even well-defined), but increases much more strongly than T0.5 with increasing temperature to compensate the linear dependence of W on T as given by eq 3. This is somewhat surprising in that [1,1′;4′,1′′-terphenyl]4,4′′-dimethanethiol (TPDMT) is only one phenyl group larger than biphenyldimethyldithiol (as indicated in the schematic molecular diagrams in Figure 3). The differences in the molecular film structure and molecular conformation vary with temperature for these two similar molecules. Such structural changes with temperature must be partly related to differences in the molecular orientation for these two types of molecular films. Both types of molecular films have a dense and regular packing of benzene rings. Nonetheless, in self-assembled monolayers (SAMs) of [1,1′;4′,1′′-terphenyl]-4,4′′-dimethanethiol (TPDMT), the molecular orientation9 is very different from multilayer films of biphenyldimethyldithiol (BPDMT),7,8 providing the TPDMT single molecule “thick” film several mechanisms for structural changes with pressure and temperature.5,6 The molecular orientation for the terphenyl backbone of TPDMT in the single layer SAM film is more “upright” (but again, we note, not perfectly so) while BPDMT is oriented nearly parallel to the substrate with the planes of benzene rings oriented normal to the surface.7 Taking into account the results presented in Figure 2 and the respective discussion, we can assume that the temperature-dependent variation of the Debye temperature of the TPDMT film is mediated by temperature-

1098 J. Phys. Chem. B, Vol. 110, No. 3, 2006 dependent changes in orientation and conformation of its molecular constituents. Strong deviations from the expected Debye-Waller behavior occur in the vicinity of 260-270 K, about the temperature where there are the subtle changes in the light polarization dependent photoemission. Taken together, we surmise that temperaturedependent structural changes are most likely responsible for TPDMT violating the Debye-Waller behavior. Not all changes in molecular orientation result in such significant deviations in Debye behavior as is observed here. Regioregular polyhexylthiophene undergoes a temperature-dependent reorientation of the molecular backbone,3 but the core level photoemission intensities nonetheless follow the expected Debye-Waller relation intensity relationship described above, and a Debye temperature of 645 ( 50 K can be derived from the S 2p core level intensity for regioregular polyhexylthiophene (P3HT).1 Structural changes for TPDMT must differ from molecules such as polyhexylthiophene in that, with increasing temperature, there is an increase in intermolecular interaction. Such intermolecular interactions lead to an increase in the rigidity of the molecular film, as opposed to the expected decrease in intermolecular interaction expected with a molecule like regioregular polyhexyl thiophene (P3HT).2,3 Also, a change of the molecular conformation of the TPDMT molecules, associated with a relative torsion of the individual rings around the molecular backbone, might affect the stiffness of this molecular backbone for a molecule in an densely packed molecular layer, which might also result in a change of the effective Debye temperature. Acknowledgment. This work was supported by the National Science Foundation through grant CHE-0415421 and the NSF “QSPINS” MRSEC (DMR 0213808), the DFG (JA 883/4-2), the German Federal Ministry for Education and Research BMBF (05KS4VHA/4), and the Center for Advanced Microstructures and Devices (CAMD), and the Louisiana Board of Regents. The authors would also like to acknowledge the support and assistance of M. Grunze and thank W. Eck for the synthesis of the TPDMT. References and Notes (1) Feng, D.-Q.; Rajesh, R.; Redepenning, J.; Dowben, P. A. Appl. Phys. Lett. 2005, 87, 181918.

Letters (2) Caruso, A. N.; Feng, D.-Q.; Losovyj, Ya. B.; Shulz, D.; Balaz, S.; Rosa, Luis G.; Sokolov, A.; Doudin, B.; Dowben, P. A. Phys. Status Solidi In press. (3) Feng, D.-Q.; Caruso, A. N.; Schulz, D.; Losovyj, Ya. B.; Dowben, P. A. J. Phys. Chem. B 2005, 109, 16382. (4) Rosa, L. G.; Losovyj, Y. B.; Choi, J.; Dowben, P. A. J. Phys. Chem. B 2005, 109, 7817. (5) Tai, Y.; Shaporenko, A.; Eck, W.; Grunze, M.; Zharnikov, M. Appl. Phys. Lett. 2004, 85, 6257. (6) Feng, D.-Q.; Losovyj, Ya. B.; Tai, Y.; Zharnikov, M.; Dowben, P. A. Phys. ReV. Lett. Submitted for publication. (7) Caruso, A. N.; Rajesh, R.; Gallup, G.; Redepenning, J.; Dowben, P. A. J. Phys.: Condens. Matter 2004, 16, 845. (8) Caruso, A. N.; Wang, L. G.; Jaswal, S. S.; Tsymbal, E. Y.; Dowben, P. A. Interface Sci. Accepted. (9) Tai, Y.; Shaporenko, A.; Rong, H.-T.; Buck, M.; Eck, W.; Grunze, M.; Zharnikov, M. J. Phys. Chem. B 2004, 108, 16806. (10) Azzam, W.; Wehner, B. I.; Fischer, R. A.; Terfort, A.; Wo¨ll, C. Langmuir 2002, 18, 7766. (11) Luxen, C.; Azzam, W.; Arnold, R.; Witte, G.; Terfort, A.; Wo¨ll, C. Langmuir 2001, 17, 3689. (12) Tai, Y.; Shaporenko, A.; Noda, H.; Grunze, M.; Zharnikov, M. AdV. Mater. 2005, 17, 1745. (13) Eck, W.; Golzhauser, A.; Stadler, V.; Geyer, W.; Zharnikov, M.; Grunze. M. AdV. Mater. 2000, 12, 805. (14) McIlroy, D. N.; Waldfried, C.; McAvoy, T.; Choi, J.; Dowben, P. A.; Heskett, D. Chem. Phys. Lett. 1997, 264, 168. (15) Dowben, P. A.; LaGraffe, D.; Onellion, M. J. Phys.: Condens. Matter 1989, 1, 6571. (16) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209. (17) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 221. (18) Chambers, D. K.; Karanam, S.; Qi, D.; Selmic, S.; Losovyj, Y. B.; Rosa, L. G.; Dowben, P. A. Appl. Phys. A 2005, 80, 483. (19) Borca, C. N.; Choi, J.; Adenwala, S.; Ducharme, S.; Dowben, P. A.; Robertson, L.; Fridkin, V. M.; Palto, S. P.; Petukhova, N. Appl. Phys. Lett. 1999, 74, 347. (20) Shevchik, N. J. Phys. ReV. B 1977, 16, 3428. (21) Waldfried, C.; McIlroy, D. N.; Zhang, J.; Dowben, P. A.; Katrich, G. A.; Plummer, E. W. Surf. Sci. 1996, 363, 296. (22) Tonner, B. P.; Li, H.; Robrecht, M. J.; Chou, Y. C.; Onellion, M.; Erskine, J. L. Phys. ReV. B 1986, 34, 4386. (23) Williams, R. S.; Wehner, P. S.; Stohr, J.; Shirley, D. A. Phys. ReV. Lett. 1977, 39, 302. (24) van Hove, M. A.; Weinberg, W. H.; Chan, C. M. Low-Energy Electron Diffraction. Springer Ser. Surf. Sci. 1986, 6, 134. (25) Clarke, L. J. Surface Crystallography; Wiley: New York, 1985. (26) Jeong, H.-K.; Komesu, T.; Dowben, P. A.; Schultz, B. D.; Palmstrøm, C. J. Phys. Lett. A 2004, 302, 217.