Dispersive Electronic States of the π-Orbitals Stacking in Single

Mar 21, 2013 - Dispersive Electronic States of the π-Orbitals Stacking in Single Molecular Lines on the Si(001)-(2×1)-H Surface. Shin-ichi Kamakuraâ...
0 downloads 3 Views 4MB Size
Download PDF
Letter pubs.acs.org/JPCL

Dispersive Electronic States of the π‑Orbitals Stacking in Single Molecular Lines on the Si(001)-(2×1)‑H Surface Shin-ichi Kamakura,†,‡ Jaehoon Jung,§ Taketoshi Minato,†,§,∇ Yousoo Kim,§ Md. Zakir Hossain,†,⊥ Hiroyuki S. Kato,*,†,# Toshiaki Munakata,# and Maki Kawai*,†,‡ †

Surface Chemistry Laboratory and §Surface and Interface Science Laboratory, RIKEN, Wako, Saitama 351-0198, Japan Department of Advanced Materials Science, University of Tokyo, Kashiwa, Chiba 277-8501, Japan ⊥ Advanced Engineering Research Team, Advanced Scientific Research Leaders Development Unit, Gunma University, Kiryu, Gunma 376-8515, Japan # Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan ‡

S Supporting Information *

ABSTRACT: One-dimensional (1D) molecular assemblies have been considered as one of the potential candidates for miniaturized electronic circuits in organic electronics. Here, we present the quantitative experimental measurements of the dispersive electronic feature of 1D benzophenone molecular assemblies on the Si(001)-(2×1)-H. The well-aligned molecular lines and their certain electronic state dispersion were observed by scanning tunneling microscopy (STM) and angleresolved ultraviolet photoemission spectroscopy (ARUPS), respectively. Density functional theory (DFT) calculations reproduced not only the experimental STM image but also the dispersive features that originated from the stacking phenyl πorbitals in the molecular assembly. We obtained the effective mass of 2.0me for the hole carrier along the dispersive electronic state, which was comparable to those of the single-crystal molecules widely used in organic electronic applications. These results ensure the one-dimensionally delocalized electronic states in the molecular lines, which is requisitely demanded for a charge-transport wire. SECTION: Physical Processes in Nanomaterials and Nanostructures

C

molecular wire because the Ph rings are perfectly aligned in an assembly with a periodicity of 3.84 Å, which makes their πorbitals stack along the line. The key requirement for such an assembly to act as a conductive channel is to interact with each other, that is, π−π interaction among the stacking Ph rings. Indeed, some theoretical calculations have predicted that such 1D molecular assemblies, for example, a π-stacking styrene assembly, on a Si(001) surface can transport electrical current.20−25 Modulation of the conduction channel through the molecular assembly and substrate is expected to be achieved through the substitution in a molecule.23−25 Rochefort et al. have quantitatively computed the typical π-band dispersion to be ∼340 meV for such 1D assemblies.22 While the recent observation and analysis of tunneling height gradations in a constant current STM image of styrene assembly have claimed nonlocalized interaction in the assembly,12,18 it is highly desirable to experimentally investigate the π−π interaction in 1D molecular assemblies on Si substrates for exploring insight into the electronic structure as a conducting channel.

onductive nanowires are indispensable for realization of the ultimately miniaturized electronic circuits on the molecular scale. Together with the recent significant progress in the developments of single molecular devices, for example, rectifiers,1,2 switch junctions,3 and transistors,4−6 a number of challenging studies on the molecular nanowires have been conducted to gain control over the size, conductivity, and also constructiveness between the single molecular devices on a substrate.7−25 As a candidate for the charge-transport nanowires, one-dimensional (1D) molecular assemblies grown through the dangling-bond-initiated chain reactions on the Hterminated Si(001)-(2×1) surface are promising because not only length and location9−19 but also direction13−17 of the molecular line growth can be controlled. Despite a number of the studies focusing on the control of growth modes, details of the electronic states relating to the conductivity of such 1D molecular assemblies have not been experimentally investigated19 because of difficulties in preparing the dense nanostructure samples and measuring. To date, a variety of individual 9−19 and interconnected11,14,15,18 molecular assemblies have been successfully fabricated using different kinds of molecules on the Hterminated Si(100) surface. The molecular assemblies consisting of a phenyl (Ph) ring hold great promise to function as a © 2013 American Chemical Society

Received: February 20, 2013 Accepted: March 21, 2013 Published: March 21, 2013 1199

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204

The Journal of Physical Chemistry Letters

Letter

In this study, therefore, the electronic state characterization of the 1D benzophenone lines on Si(001)-(2×1)-H were carried out by the combination of an angle-resolved ultraviolet photoelectron spectroscopy (ARUPS) and periodic density functional theory (DFT) calculations with morphological verification by STM. The theoretically optimized molecular assembly was corroborated in comparison with their highresolution STM images. The observed peaks in the UPS spectra were reproduced by the calculation in which the dispersive feature consists of the π-orbitals of stacking Ph rings in the assembled benzophenone. The energy shift in k-space was comparable to that of the organic materials giving the high carrier mobility in the thin film transistors. Figure 1a shows a typical STM image of benzophenonedosed Si(001)-(2×1)-H; various length but equally directed single molecular lines following the Si dimer row direction on each terrace are seen.17,19 The self-assembled molecular lines are formed through a chain reaction initiated by dangling bond sites on Si(001)-(2×1)-H, as shown in Figure 1b. The >CO double bond of benzophenone becomes a >CH−O−Si linkage by creation of an O−Si covalent bond at a dangling bond site followed by H-atom abstraction from the nearest-neighbor monohydrated dimer Si to the >C•− radical site.17,19 The H abstraction from the nearest-neighbor site, consequently, leads the molecular line to be grown at a side of the Si dimer row, as in the inset of Figure 1a, in which the center of the molecular lines is off from the center of dimer rows (red bars).17 Note that the sp2-type >C site of the carbonyl group becomes an sp3-type >CH− site after the H abstraction. Hence, the two Ph rings of the assembled molecule are no longer in a π-conjugate, and the electronic states of the two may not be equivalent. It is fact that the on-dimer side of the benzophenone assembly is slightly brighter than the off-dimer side, as shown in the highresolution STM image (inset of Figure 1a). Because the molecular periodicity in the assembly is close, that is, the distance between the Si dimers along the row (3.84 Å), the πorbitals of each Ph ring must stack individually and thus possibly interact along the assembly. Details of the benzophenone assembly on Si(001)-(2×1)-H were examined using periodic DFT calculations. The optimized geometry of the assembly was corroborated in comparison between the high-resolution STM images and the calculated one based on the Tersoff−Hamann approach.26 Figure 1c shows the simulated STM image with the top and side views of the optimized benzophenone assembly, which reproduces well the observed high-resolution STM image. Our computational results revealed that the different electronic structure between two Ph rings is attributed to the symmetry breaking of the isolated benzophenone molecule (C2 symmetry) at the adsorption; the off-dimer Ph plane becomes nearly perpendicular to the −C−O− linker, while the on-dimer Ph plane still includes the linker (see details in the Supporting Information). This symmetry breaking results from the change in hybridization of the −C−O− linker group, that is, sp2 to sp3, and the asymmetrical environment of the substrate, that is, on-dimer and off-dimer rows, with respect to the positions of Ph rings. Therefore, this geometric difference induces the splitting of πstates of two Ph rings, as shown in Figure 1d, where the first occupied π-state (∼1.1 eV below the Fermi level), that is, the highest occupied molecular orbital (HOMO), is distributed mainly at the on-dimer Ph and the next dominant π-state (∼1.4 eV), attributed to HOMO−1, shows wider distribution on both

Figure 1. View graphs of benzophenone molecular assemblies on Si(001)-(2×1)-H: (a) a typical STM image (Vs = −2.8 V, It = 0.2 nA, 90 × 62 nm2) and a zoomed image (inset: Vs = +1.0 V, It = 0.2 nA, 2.8 × 1.5 nm2); (b) schematic reaction steps; (c) top and side views of calculated benzophenone lines; and (d) calculated orbitals of the benzophenone assembly, with the HOMO at ∼1.1 eV in binding energy (left) and the HOMO−1 at ∼1.4 eV (right). In the inset of (a), the center positions of the Si dimers are marked by red bars. In (c), the upper panel shows a simulated constant current STM image (Vs = +1.5 V), with red arrows marking the Si dimer center. The unit cell in our calculation is shown by dotted lines in (c).

on- and off-dimer Ph rings with a substrate component due to level alignment between the adsorbate and substrate orbitals. To elucidate the π-state features of the benzophenone assembly, we measured UPS spectra of them. The UPS spectra are also compared with the calculated results, as shown in 1200

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204

The Journal of Physical Chemistry Letters

Letter

ratio difference might have originated from a lack of taking account of the transition dipole moments for photoemission from individual molecular orbitals highly oriented in the assemblies. The π−π interaction in the stacking Ph rings was examined by ARUPS measured from the surface normal toward the [110] direction. To ascertain carefully the π−π interaction, particularly the peak at 3.5 eV, the substrate signals were removed by subtraction of the Si(001)-(2×1)-H spectrum from that of the benzophenone-dosed sample following an intensity normalization at the binding energy of 1.2 eV (as described in the Supporting Information). Figure 2b shows a set of the ARUPS results obtained after the substrate signal subtraction at each detection angle. The shape of peaks in the ARUPS spectra for the benzophenone assembly was not simply symmetric. At the surface normal detection, the peak was located at about 3.5 eV (open triangle), and a broad shoulder appear at about 3.0 eV (filled triangle). With increasing detection angle, the shoulder decreased in intensity, and the peak was shifted gradually to a deeper binding energy. The spectral change depending on the detection angle reflects an interaction between the stacking Ph π-orbitals in the assembly. The π-orbital interaction in the assembly was evaluated by the calculation of the electronic state dispersion. Figure 3 shows

Figure 2a. The spectra (i) and (ii) were measured at the Si(001)-(2×1)-H surface before and after the benzophenone

Figure 2. UPS spectra of benzophenone molecular assemblies on Si(001)-(2×1)-H: (a) comparison between measured and calculated spectra; (b) a set of ARUPS spectra. For (a), the spectra of the Si(001)-(2×1)-H surface before (i) and after (ii) the benzophenone dose were detected at the surface normal, and the calculated spectrum (iii) was obtained by the estimated LDOS intensities of the benzophenone assemblies at the Γ̅ point in k-space, with broadening by a Gaussian function of σ = 0.5 eV and an energy shift by +2.0 eV. The signal intensity of the spectra (i) and (ii) was numerically normalized with that at 1.2 eV for subtraction of the substrate signals. For (b), the ARUPS spectra were detected at every 2.5° from the surface normal, 0°, to 30° toward the [110] direction. The indicated ARUPS spectra were obtained after the substrate signal subtraction (Supporting Information) and normalized by the peak height. The peak and shoulder are marked for a guide for the eyes. The bold spectra in (b) indicate those at the Γ̅ point and near the first Brillouin zone boundary. Figure 3. Calculated electronic state dispersion in the k-space for the benzophenone assembly on Si(001)-(2×1)-H using the 4 × 1 supercell shown in Figure 1c. The thin lines indicate energy levels of the substrate, and the dots describe the energy levels and the LDOS intensity of the benzophenone assembly by their positions and diameters, respectively.

dose, respectively. Their onsets below 2 eV in binding energy overlapped each other, and thus, the three additive intense peaks at 3.5, 6.5, and 8.6 eV on spectrum (ii) were attributed to the signals from the benzophenone assemblies, as previously reported.19 A small energy shift of the peaks observed in the current and previous studies is probably due to a Fermi level difference arising from the dopant in different wafers used. Because the electronic states are affected by a number of molecules in line assemblies at especially less than five,22 a part of the broadened peak features is attributed to the inhomogeneous molecular line length. Spectrum (iii) was obtained from the calculated local density of states (LDOS) of the benzophenone assembly on Si(001)-(2×1)-H at the Γ̅ point in k-space with broadening by the Gaussian function of σ = 0.5 eV. The calculated spectra reproduced the three intense peaks, though the energy levels and relative intensity ratio of the peaks were somewhat incomplete. The energy level shift by a few eV is due to a weakness of standard DFT exchange− correlation functionals for band gap materials.27,28 The intensity

the computed electronic states in the topmost components parallel (toward X̅ ) and perpendicular (toward M̅ ) to the molecular lines (in the k-space). Several series of dispersive features that originated from the staking Ph orbitals were obtained. In this calculation, the 4 × 1 supercell described in Figure 1c was used, and thus, the molecular lines are located on every second Si dimer row in the periodic boundary conditions. Though the separation between the molecular lines in this periodicity must be shortest because of the molecular size, only a small peak shift perpendicular to the lines, that is, little interaction between the molecular lines, was expected. In contrast to that, the energy shift by several sub-eV was estimated parallel to the lines, indicating a strong interaction in 1201

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204

The Journal of Physical Chemistry Letters

Letter

effective mass of the holes, mh, is given by ℏ2/2ta2 around the Γ̅ points in the TB approximation and are thus estimated to be mh = 2.0me, where me is the electron rest mass. This value is comparable to that of the single-crystal molecules providing the high carrier mobility in organic field effect transistors, that is, 1.26me at pentacene30 and 0.65me at rubrene.31 It is, therefore, inferred that the assembled benzophenone lines and possibly the other Ph stacking molecular lines on Si(001)-(2×1)-H have the delocalized electronic states as requisitely demanded for conductive nanowires. In order to realize the actually useful nanowires, however, electric/electronic isolation between the stacking π-orbitals and the substrate must be considered. Although some of molecular orbitals decouple from the substrate in proper conditions, for example, the HOMO π-state shown in Figure 1d, the orbital coupling to that of the substrate often occurs, which is recognized as the fragmented dispersion of the molecular orbitals in Figure 3. Recent theoretical charge-transport analyses of the molecular lines have also claimed that the coupling between the stacking π-orbitals of molecular assemblies and the silicon substrate is not negligible.23−25 One of the solutions might be the spatial separation between the stacking π-orbitals and the substrate by a wide band gap segment. However, such bulky and/or rigid molecules may hardly proceed to the radical chain reaction by steric hindrance on the active dangling bond site neighboring the first adsorbed molecules.17 The other solution would be the energy level tuning of stacking Ph rings in the assemblies, for instance, by substituting an electronegative or electropositive group for H of Ph.18,23,25 Further continuous attempts to fabricate the real conductive nanowires have been required. In summary, the electronic states of the 1D benzophenone molecular assemblies on Si(001)-(2×1)-H were studied using STM, ARUPS, and DFT calculations. The high-resolution STM images corroborated a calculated stable structure in the molecular lines. The dispersive features of its HOMOs were obtained in both ARUPS and calculation results, consistently, for the first time, to our knowledge. Under the 1D TB approximation, the dispersive HOMO state was expected to be the hole effective mass of 2.0me, which ensured the delocalized electronic states requisitely demanded for a molecular chargetransport wire.

the stacking π-orbital range. This dispersion well explains the observed peak features in Figure 2b. Note that, as shown in Figure 1d, the HOMO π-state consisting of the stacking Ph rings is isolated from any substrate electronic states, while the next dominant π-state couples to them. Therefore, the dispersion of HOMO at around 1.1 eV in Figure 3 can be mainly attributed to the π-state delocalization in the stacking Ph rings, without electronic coupling to the substrate. Figure 4 shows the k-space map converted from the ARUPS results in Figure 2b, with the plots of the computed π-state

Figure 4. Overlap of the measured and the calculated dispersions for the benzophenone assembly on Si(001)-(2×1)-H. The measured kspace map was converted from Figure 2b. The computed components along the molecular lines were plotted by black circles, with an energy shift of +2.0 eV. Under the TB approximation, the 1D dispersion model is fit to a series of the HOMO states; the fit curve is indicated as the black solid line in the fit area, and the dashed line is for its extension.

■

energy levels parallel to the molecular lines in Figure 3 following an energy shift by +2.0 eV. The dispersive band structures along the molecular lines are in good agreement with the observed spectral features; the broad shoulder of about 3.0 eV near the Γ̅ point and the gradual main peak shift to deeper binding energy with increasing k∥. Because both molecular lines parallel and perpendicular to the [110] direction, depending on terraces, are included in our ARUPS spectra, the nondispersive components perpendicular to the molecular lines should overlap and obscure the dispersive features. By the sum of these LDOS components, indeed, the gradual spectral change is well reproduced (as shown in the Supporting Information). To elucidate its properties as a molecular nanowire, fitting of a representative model to the dispersion was performed. Within a simple 1D tight binding (TB) approximation, the energy dispersion is explained as Ec + 2t cos(ak∥), where Ec, t, and a represent the energy of the band center, transfer integral, and periodicity in the molecular line, respectively.29 By fitting this equation to a series of the π-orbitals derived from the HOMO, the model reproduces the dispersion as the solid line in Figure 4 and provides the least-squares fit parameter t = 0.127 eV. The

METHODS Experimental Section. All of the experiments were performed in an ultrahigh vacuum (UHV) chamber with a base pressure better than 5 × 10−11 mbar. The UHV chamber is equipped with a set of a He discharge lamp and a hemispherical electron analyzer (Omicron EA125HR) for UPS and a variabletemperature scanning tunneling microscope (Omicron VTSTM). The silicon substrate (8 × 1.5 mm2) was cut from a ptype silicon wafer (0.5 mm thick, B-doped, 0.02 Ωcm). The surface was cleaned by prolonged annealing at ∼850 K (∼ 8 h) followed by repeated flashing up to 1400 K. The H-terminated surface was prepared by exposure to atomic H, generated by a hot W filament (∼2100 K), at the surface temperature of ∼625 K. The prepared H-terminated surface commonly contained a dilute concentration of dangling bond sites. Benzophenone (solid) purchased from Tokyo Chemicals Industries Co., Ltd. (TCI) was purified by several times of pumping the vapor above the solid phase in a glass ampule. The benzophenone molecules were dosed by opening a gate valve between the ampule and the silicon sample in a subchamber. To obtain the 1202

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204

The Journal of Physical Chemistry Letters

Letter

Education, Culture, Sports, Science and Technology of Japan (MEXT) and the Element Innovation (EI) Project granted to Gunma University.

fully grown molecular lines, the benzophenone molecules were dosed onto the freshly prepared H-terminated surface for ≥100 Langmuir (≥10−4 Torr·s) at room temperature.19 Using this sample, all of the measurements for the benzophenone line assemblies were conducted at room temperature. The STM images of the surfaces were observed in constant current mode. At the ARUPS measurements, because the UV light source and the electron analyzer have been fixed at the UHV chamber with an angle of 45°, the detection angle of the emitted electrons was changed by the rotation of the sample mount axis from the surface normal to the [110] direction. Theoretical Basis. In order to elucidate detailed electronic structures of the 1D benzophenone assembly on Si(001)(2×1)-H, periodic DFT calculations were performed using the Vienna ab initio simulation package (VASP) code32,33 with the Perdew−Burke−Ernzerhof (PBE) exchange−correlation functional.34 The core electrons were replaced by projectoraugmented wave (PAW) pseudopotentials35 and expanded in a basis set of plane waves up to a cutoff energy of 400 eV. The 4 × 1 supercell was used in order to suppress the spatial hindrance between molecular assembly lines. The slab model consisted of six Si layers, and the bottom atoms were terminated with hydrogen. The periodically replicated slabs were separated by a vacuum region of ∼12 Å. Dipole correction was applied in order to avoid interactions between periodic slab images. During ionic relaxations, the two bottom Si layers were fixed in their bulk positions. Ionic relaxations were performed until atomic forces were less than 0.01 eV/Å. A 4 × 16 × 1 Γcentered k-point grid was used for Brillouin zone sampling. Because the optimized lattice constant for bulk Si was 5.47 Å, the periodicity of the Si dimers along the rows was estimated to be 3.87 Å, which provides the center-to-center separation of the Ph rings in the uniform benzophenone assembly. It is noted that the value of 3.87 Å is close to the center-to-center separation of parallel benzene dimers optimized in the precise calculations including the dispersion force.36 Though the dispersion force was not included in our calculations, its perturbation to our results should be restrictive.

■

■

(1) Aviram, A.; Ratner, M. A. Molecular Rectifiers. Chem. Phys. Lett. 1974, 29, 277−283. (2) Metzger, R. M. Unimolecular Electrical Rectifiers. Chem. Rev. 2003, 103, 3803−3834. (3) Luo, Y.; Collier, P.; Jeppesen, J. O.; Nielsen, K. A.; Delonno, E.; Ho, G.; Perkins, J.; Tseng, H.-R.; Yamamoto, T.; Stoddart, J. F.; Heath, J. R. Two-Dimensional Molecular Electronics Circuits. ChemPhysChem 2002, 3, 519−525. (4) Park, H.; Park, J.; Lim, A. K. L.; Anderson, E. H.; Allvisatos, A. P.; McEuen, P. L. Nanomechanical Oscillations in a Single-C60 Transistor. Nature 2000, 407, 57−60. (5) Park, J.; Pasupathy, A. N.; Goldsmith, J. I.; Yaish, Y.; Petta, J. R.; Rinkoski, M.; Sethna, J. P.; Abruna, H. D.; McEuen, P. L.; Ralph, D. C. Coulomb Blockade and the Kondo Effect in Single-Atom Transistors. Nature 2002, 417, 722−725. (6) Liang, W.; Shores, M. P.; Bockrath, M.; Long, J. R.; Park, H. Kondo Resonance in a Single-Molecule Transistor. Nature 2002, 417, 725−729. (7) Okawa, Y.; Aono, M. Nanoscale Control of Chain Polymerization. Nature 2001, 409, 683−684. (8) Okawa, Y.; Mandal, S. K.; Hu, C.; Tateyama, Y.; Goedecker, S.; Tsukamoto, S.; Hasegawa, T.; Gimzewski, J. K.; Aono, M. Chemical Wiring and Soldering toward All-Molecule Electronic Circuitry. J. Am. Chem. Soc. 2011, 133, 8227−8233. (9) Lopinski, G.. P.; Wayner, D. D. M.; Wolkow, R. A. Self-Directed Growth of Molecular Nanostructures on Silicon. Nature 2000, 406, 48−51. (10) Basu, R.; Guisinger, N. P.; Greene, M. E.; Hersam, M. C. Room Temperature Nanofabrication of Atomically Registered Heteromolecular Organosilicon Nanostructures using Multistep Feedback Controlled Lithography. Appl. Phys. Lett. 2004, 85, 2619−2621. (11) Kirczenow, G.; Piva, P. G.; Wolkow, R. A. Linear Chains of Styrene and Methylstyrene Molecules and their Heterojunctions on Silicon: Theory and Experiment. Phys. Rev. B 2005, 72, 245306/1− 245306/4. (12) Piva, P. G.; DiLabio, G. A.; Pitters, J. L.; Zikovsky, J.; Rezeq, M.; Dogel, S.; Hofer, W. A.; Wolkew, R. A. Field Regulation of SingleMolecule Conductivity by a Charged Surface Atom. Nature 2005, 435, 658−661. (13) Hossain, M. Z.; Kato, H. S.; Kawai, M. Controlled Fabrication of 1D Molecular Lines Across the Dimer Rows on the Si(100)-(2×1)-H Surface through the Radical Chain Reaction. J. Am. Chem. Soc. 2005, 127, 15030−15031. (14) Hossain, M. Z.; Kato, H. S.; Kawai, M. Fabrication of Interconnected 1D Molecular Lines along and across the Dimer Rows on the Si(100)-(2×1)-H Surface through the Radical Chain Reaction. J. Phys. Chem. B 2005, 109, 23129−23133. (15) Hossain, M. Z.; Kato, H. S.; Kawai, M. Selective Chain Reaction of Acetone Leading to the Successive Growth of Mutually Perpendicular Molecular Lines on the Si(100)-(2×1)-H Surface. J. Am. Chem. Soc. 2007, 129, 12304−12309. (16) Zikovsky, J.; Dogel, S. A.; Haider, M. B.; DiLabio, G. A.; Wolkew, R. A. Self-Directed Growth of Contiguous Perpendicular Molecular Lines on H-Si(100) Surfaces. J. Phys. Chem. A 2007, 111, 12257−12259. (17) Hossain, M. Z.; Kato, H. S.; Kawai, M. Self-Directed Chain Reaction by Small Ketones with the Dangling Bond Site on the Si(100)-(2×1)-H Surface: Acetophenone, A Unique Example. J. Am. Chem. Soc. 2008, 130, 11518−11523. (18) Piva, P. G.; Wolkow, R. A.; Kirczenow, G. Nonlocal Conductance Modulation by Molecules: Scanning Tunneling Microscopy of Substituted Styrene Heterostructure on H-Terminated Si(100). Phys. Rev. Lett. 2008, 101, 106801/1−106801/4.

ASSOCIATED CONTENT

S Supporting Information *

Details of the data analysis and the comparison between the measured and computed results. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (H.S.K.); [email protected] (M.K.). Present Address ∇

T. Minato: Office of Society-Academia Collaboration for Innovation (SACI), Kyoto University, Kyoto 606-8501, Japan. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are grateful for the computational resources of the RIKEN Integrated Cluster of Clusters (RICC) supercomputer system. This study was financially supported in part by the Grants-inAid for “Nanoscale Science and Technology Research” in RIKEN. M.Z.H. acknowledges partial support by the Program to Disseminate Tenure-Track System of the Ministry of 1203

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204

The Journal of Physical Chemistry Letters

Letter

(19) Hossain, M. Z.; Kato, H. S.; Kawai, M. Valence States of OneDimensional Molecular Assembly Formed by Ketone Molecules on the Si(100)-(2×1)-H Surface. J. Phys. Chem. C 2008, 113, 10751− 10754. (20) Cho, J.-H.; Oh, D.-H.; Kleinman, L. One-Dimensional Molecular Wire on Hydrogenated Si(001). Phys. Rev. B 2002, 65, 081310(R)/1−081310(R)/4. (21) Rochefort, A.; Martel, R.; Avouris, P. Electrical Switching in πResonant 1D Intermolecular Channels. Nano Lett. 2002, 2, 877−880. (22) Rochefort, A.; Boyer, P.; Nacer, B. Resonant Tunneling Transport in Highly Organized Oligoacene Assemblies. Org. Electron. 2007, 8, 1−7. (23) Geng, W. T.; Oda, M.; Nara, J.; Kondo, H.; Ohno, T. Electron Transport in a π-Stacking Molecular Chain. J. Phys. Chem. B 2008, 112, 2795−2800. (24) Smeu, M.; Wolkow, R. A.; Guo, H. Conduction Pathway of πStacked Ethylbenzene Molecular Wires on Si(100). J. Am. Chem. Soc. 2009, 131, 11019−11026. (25) Smeu, M.; Wolkow, R. A.; Guo, H. Conduction Modulation of π-Stacked Ethylbenzene Molecular Wires on Si(100) with Substituent Groups. Theor. Chem. Acc. 2012, 131, 1085/1−1085/8. (26) Tersoff, J.; Hamann, D. R. Theory and Applications for the Scanning Tunneling Microscope. Phys. Rev. Lett. 1983, 50, 1998− 2001. (27) Perdew, J. P.; Levy, M. Physical Content of the Exact KohnSham Orbital Energies: Band Gaps and Derivative Discontinuities. Phys. Rev. Lett. 1983, 51, 1884−1887. (28) Sham, L. J.; Schlüter, M. Density-Functional Theory of the Energy Gap. Phys. Rev. Lett. 1983, 51, 1888−1891. (29) Ueno, N.; Kera, S. Electron Spectroscopy of Functional Organic Thin Films: Deep Insights into Valence Electronic Structure in Relation to Charge Transport Property. Prog. Surf. Sci. 2008, 83, 490− 557. (30) Kakuta, H.; Hirahara, T.; Matsuda, I.; Nagao, T.; Hasegawa, S.; Ueno, N.; Sakamoto, K. Electronic Structures of the Highest Occupied Molecular Orbital Bands of a Pentacene Ultrathin Film. Phys. Rev. Lett. 2007, 98, 247601/1−247601/4. (31) Machida, S.; Nakayama, Y.; Duhm, S.; Xin, Q.; Funakoshi, A.; Ogawa, N.; Kera, S.; Ueno, N.; Ishii, H. Highest-Occupied-MolecularOrbital Band Dispersion of Rubrene Single Crystals as Observed by Angle-Resolved Ultraviolet Photoelectron Spectroscopy. Phys. Rev. Lett. 2010, 104, 156401/1−156401/4. (32) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for Liquid Metals. Phys. Rev. B 1993, 47, 558−561. (33) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169−11186. (34) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (35) Kresse, G.; Joubert, D. From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method. Phys. Rev. B 1999, 59, 1758− 1775. (36) Puzder, A.; Dion, M.; Langreth, D. C. Binding Energies in Benzene Dimers: Nonlocal Density Functional Calculations. J. Chem. Phys. 2006, 124, 164105/1−164105/8.

1204

dx.doi.org/10.1021/jz400389k | J. Phys. Chem. Lett. 2013, 4, 1199−1204