pubs.acs.org/Langmuir © 2009 American Chemical Society
Oligothiophene Template Effects on Packings and Orientations of C60 Molecules on Ag(111) Surface Jin Wen and Jing Ma* Institute of Theoretical and Computational Chemistry, Key Laboratory of Mesoscopic Chemistry of MOE, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, P. R. China Received October 12, 2009. Revised Manuscript Received December 4, 2009 The packing conformations of sexithiophene (6T) and the orientations of the C60 molecules on top of the preadsorbed 6T monolayer on Ag(111) surface have been investigated by the molecular dynamics simulations (on the basis of molecular mechanics) in conjunction with quantum mechanics calculations of the relative strength of intermolecular and interfacial interactions. It is demonstrated that the flat-lying oligothiophene (nT, n=4 and 6) monolayer is formed on the Ag(111) surface, and the arrangement of 6T molecules is more ordered than that in 4T film. It is also shown that the underlying 6T stripes make C60 molecules aggregate in chainlike arrays on the 6T covered Ag(111) surface, showing significant template effects on the directed self-assembly of C60 cages. For the absorbed C60 molecule on the 6T prepatterned Ag(111) surface, four typical orientations, hexagon, pentagon, 6:6 bond, and 5:6 bond, are found to appear with populations of 26.3%, 2.7%, 37.5%, and 18.8%, respectively. When the deposition order is changed, the 6T stripes are shown to tilt with corrugation on the underlying C60 carpet, revealing the important role of the deposition order in modulation of the ordered supramolecular nanostructures.
1. Introduction The bottom-up fabrication of highly ordered self-organized nanostructures is a key issue in the realization of highly efficient electronic devices.1-9 It has been demonstrated that the arrangement of fullerenes on bare substrate or well-ordered nanotemplates was controlled by temperature, surface coverage, structures of substrates, and the deposition order.10-13 On various semiconducting and metal surfaces (which have isotropic lattice) C60 molecules were *Corresponding author. E-mail:
[email protected]. (1) Halik, M.; Klauk, H.; Zschieschang, U.; Schmid, G.; Dehm, C.; Sch€utz, M.; Maisch, S.; Effenberger, F.; Brunnbauer, M.; Stellacci, F. Nature 2004, 431, 963–966. (2) Halik, M.; Klauk, H.; Zschieschang, U.; Schmid, G.; Ponomarenko, S.; Kirchmeyer, S.; Weber, W. Adv. Mater. 2003, 15, 917–922. (3) Witte, G.; W€oll, C. J. Mater. Res. 2004, 19, 1889–1916. (4) Thayer, G. E.; Sadowski, J. T.; zu Heringdorf, F. M.; Sakurai, T.; Tromp, R. M. Phys. Rev. Lett. 2005, 95, 256106. (5) Li, L.; Tang, Q.; Li, H.; Hu, W. J. Phys. Chem. B 2008, 112, 10405–10410. (6) Kowarik, S.; Gerlach, A.; Sellner, S.; Schreiber, F.; Cavalcanti, L.; Konovalov, O. Phys. Rev. Lett. 2006, 96, 125504. (7) Sirringhaus, H.; Brown, P. J.; Friend, R. H.; Nielsen, M. M.; Bechgaard, K.; Langeveld-Voss, B. M. W.; Spiering, A. J. H.; Janssen, R. A. J.; Meijer, E. W.; Herwig, P.; de Leeuw, D. M. Nature 1999, 401, 685–688. (8) Briseno, A. L.; Mannsfeld, S. C. B.; Ling, M. M.; Liu, S.; Tseng, R. J.; Reese, C.; Roberts, M. E.; Yang, Y.; Wudl, F.; Bao, Z. Nature 2006, 444, 913–917. (9) Schreiner, P. R.; Fokin, A. A.; Pascal, R. A., Jr.; de Meijere, A. Org. Lett. 2006, 8, 3635–3638. (10) Theobald, J. A.; Oxtoby, N. S.; Phillips, M. A.; Champness, N. R.; Beton, P. H. Nature 2003, 424, 1029–1031. (11) Bonifazi, D.; Kiebele, A.; St€ohr, M.; Cheng, F.; Jung, T.; Diederich, F.; Spillmann, H. Adv. Funct. Mater. 2007, 17, 1051–1062. (12) Bonifazi, D.; Spillmann, H.; Kiebele, A.; de Wild, M.; Seiler, P.; Cheng, F.; G€untherodt, H. J.; Jung, T.; Diederich, F. Angew. Chem., Int. Ed. 2004, 43, 4759–4763. (13) Zhang, H. L.; Chen, W.; Huang, H.; Chen, L.; Wee, A. T. S. J. Am. Chem. Soc. 2008, 130, 2720–2721. (14) Sakurai, T.; Wang, X. D.; Hashizume, T.; Yurov, V.; Shinohara, H.; Pickering, H. W. Appl. Surf. Sci. 1995, 87/88, 405–413. (15) Schull, G.; Berndt, R. Phys. Rev. Lett. 2007, 99, 226105. (16) Zhang, X.; Yin, F.; Palmer, R. E.; Guo, Q. Surf. Sci. 2008, 602, 885–892. (17) Altman, E. I.; Colton, R. J. Phys. Rev. B 1993, 48, 18244–18249. (18) Hou, J. G.; Yang, J.; Wang, H.; Li, Q.; Zeng, C.; Lin, H.; Wang, B.; Chen, D. M.; Zhu, Q. Phys. Rev. Lett. 1999, 83, 3001–3004. (19) Chen, W.; Zhang, H.; Huang, H.; Chen, L.; Wee, A. T. S. ACS Nano 2008, 2, 693–698.
Langmuir 2010, 26(8), 5595–5602
√ √ observed to form the hexagonally close-packed (2 3 2 3)R30 layer,14-19 while various organic templates made C60 molecules line up in directed supramolecular arrays.10,20-23 For examples, scanning tunneling microscope (STM) experiments displayed that codeposition of C60 molecules on top of sexithiophene (6T), sexiphenyl, and porphyrin monolayers led to the long-range ordered arrangements on the Ag(111) surface.11,19,24-27 These well-controlled films are conceived to have great potential in constructing the low-cost electronic devices such as organic solar cells and field-effect transistors. Over the past years, theoretical calculations also contributed a lot to the understanding of interplay between the C60 and various substrates.18,28-36 In spite of the surge of experimental data and theoretical models, the template effects of (20) Leigh, D. F.; N€orenberg, C.; Cattaneo, D.; Owen, J. H. G.; Porfyrakis, K.; Li Bassi, A.; Ardavan, A.; Briggs, G. A. D. Surf. Sci. 2007, 601, 2750–2755. (21) Li, M.; Deng, K.; Lei, S. B.; Yang, Y. L.; Wang, T. S.; Shen, Y. T.; Wang, C. R.; Zeng, Q. D.; Wang, C. Angew. Chem., Int. Ed. 2008, 47, 6717–6721. (22) Nakanishi, T.; Takahashi, H.; Michinobu, T.; Takeuchi, M.; Teranishi, T.; Ariga, K. Colloids Surf., A 2008, 321, 99–105. (23) Yoshimoto, S.; Honda, Y.; Ito, O.; Itaya, K. J. Am. Chem. Soc. 2008, 130, 1085–1092. (24) Chen, L.; Chen, W.; Huang, H.; Zhang, H. L.; Yuhara, J.; Wee, A. T. S. Adv. Mater. 2008, 20, 484–488. (25) Chen, W.; Zhang, H. L.; Huang, H.; Chen, L.; Wee, A. T. S. Appl. Phys. Lett. 2008, 92, 193301. (26) Huang, H.; Chen, W.; Chen, L.; Zhang, H. L.; Wang, X. S.; Bao, S. N.; Wee, A. T. S. Appl. Phys. Lett. 2008, 92, 023105. (27) Zhang, H. L.; Chen, W.; Chen, L.; Huang, H.; Wang, X. S.; Yuhara, J.; Wee, A. T. S. Small 2007, 3, 2015–2018. (28) Lee, J. Y.; Kang, M. H. Surf. Sci. 2008, 602, 1408–1412. (29) Lu, X.; Grobis, M.; Khoo, K. H.; Louie, S. G.; Crommie, M. F. Phys. Rev. Lett. 2003, 90, 096802. (30) Wang, H.; Zeng, C.; Wang, B.; Hou, J. G.; Li, Q.; Yang, J. Phys. Rev. B 2001, 63, 085417. (31) Hou, J. G.; Yang, J.; Wang, H.; Li, Q.; Zeng, C.; Yuan, L.; Wang, B.; Chen, D. M.; Zhu, Q. Nature 2001, 409, 304–305. (32) Yuan, L. F.; Yang, J.; Wang, H.; Zeng, C.; Li, Q.; Wang, B.; Hou, J. G.; Zhu, Q.; Chen, D. M. J. Am. Chem. Soc. 2003, 125, 169–172. (33) Jalkanen, J. P.; Zerbetto, F. J. Phys. Chem. B 2006, 110, 5595–5601. (34) Ogawa, A.; Tachibana, M.; Kondo, M.; Yoshizawa, K.; Fujimoto, H.; Hoffmann, R. J. Phys. Chem. B 2003, 107, 12672–12679. (35) Larsson, J. A.; Elliott, S. D.; Greer, J. C.; Repp, J.; Meyer, G.; Allenspach, R. Phys. Rev. B 2008, 77, 115434. (36) Wang, L. L.; Cheng, H. P. Phys. Rev. B 2004, 69, 165417.
Published on Web 12/16/2009
DOI: 10.1021/la903869g
5595
Article
Wen and Ma
Figure 1. Schematic sketch of the studied molecular thin films on Ag(111) surface. Each inset shows the MD snapshot (at 1 ns), in which the upmost layer is highlighted and the Ag(111) surface is removed for clarity.
organic/metal substrate on the packings and orientations of the upmost C60 molecules remain illusive. In this work, we attempt to investigate packings of the C60 and/ or 6T monolayers on Ag(111) surface, as shown in Figure 1, in a multiscale that covers the molecular and atomic levels down to the electronic structures. In order to explore the relationship between the intermolecular (6T-6T, C60-6T, and C60-C60) or interfacial (Ag-6T and Ag-C60) interactions and various packing arrangements, we combine the sophisticated quantum mechanism (QM) calculations with computationally economic molecular mechanism (MM) models. The hybrid QM/MM method is an approximated scheme to reduce the computational cost of complex systems (e.g., solutions, interfaces, materials, and enzymes) that cannot be completely treated by conventional QM methods. In our QM/MM calculations, QM is used to describe the close contact regions in C60-6T, C60-C60, and C60/Ag(111) systems, while the remaining parts that are far away from the adsorption or binding sites are treated by MM. Furthermore, within the framework of MM the variation in packing structures of nT (n=4, 6) and codeposition of C60 and 6T nanostripes on Ag(111) surface with different coverage and deposition order are demonstrated by molecular dynamics (MD) simulations. Drawn from the MD snapshots, some typical deposition configurations of C60 molecules are further subjected to density functional theory (DFT) calculations to get the STM images. It can be seen that the 6T templates are highly ordered, flat-lying on Ag(111) surface, and most of C60 molecules reside with one fused bond between hexagons (6:6 bond) faced to 6T nanostripes. When the 6T molecules are adsorbed on top of C60 monolayer, some bend and twist 6T stripes are formed with obvious corrugation and anisotropy. These results may enrich our knowledge of microscopic structures of the self-assembled thin films toward nanoelectronics applications.
2. Computational Details 2.1. MD Simulations. The performance of molecular mechanics mainly depends on the selected force field. In this work, we adopted the polymer consistent force field (PCFF),37-40 which had been extensively tested in our previous works on the packing (37) Marder, S. R.; Perry, J. W.; Tiemann, B. G.; Gorman, C. B.; Gilmour, S.; Biddle, S. L.; Bourhill, G. J. Am. Chem. Soc. 1993, 115, 2524–2526. (38) Sun, H. Macromolecules 1995, 28, 701–712. (39) Sun, H.; Mumby, S. J.; Maple, J. R.; Hagler, A. T. J. Phys. Chem. 1995, 99, 5873–5882. (40) Hwang, M. J.; Stockfisch, T. P.; Hagler, A. T. J. Am. Chem. Soc. 1994, 116, 2515–2525. (41) Zhang, G.; Pei, Y.; Ma, J.; Yin, K.; Chen, C.-L. J. Phys. Chem. B 2004, 108, 6988–6995. (42) Zhang, G.; Ma, J.; Wen, J. J. Phys. Chem. B 2007, 111, 11670–11679.
5596 DOI: 10.1021/la903869g
structures of nTs in amorphous phase,41,42 on Ag(111),43 and even in solutions.44,45 In order to further validate the applicability of PCFF in describing the intermolecular interactions between nT (n= 4, 6) dimers as well as the interfacial interactions (Ag-nT and Ag-C60), we compare the PCFF binding energies with QM and QM/MM results in Figures S1 and S2 of the Supporting Information. The qualitative agreements between the PCFF and QM results are depicted in these curves, though PCFF slightly underestimates the interlayer interaction between nTs (Figure S1) and overestimates the interaction between C60 and Ag cluster (Figure S2). The 4T/Ag(111) and 6T/Ag(111) systems are represented by a periodic slab of 35.0 34.7 44.7 A˚ and 55.0 34.7 44.7 A˚, respectively, with the three layers of Ag atoms fixed during the simulations. It is demonstrated that a reliable simulation of the codeposition of C60 and 6T molecules on Ag(111) surface requires a larger slab model of 110.1 69.3 44.7 A˚. MD simulations were performed in the canonical (NVT) ensemble with constant moles (N), volume (V), and temperature (T). The Discover module in Material Studios package46 with the selection of PCFF is employed to run MD simulations. In order to reproduce the experimental condition, the temperature is set to 298 or 360 K, respectively, by using an Andersen thermostat.47 The equations of motion are integrated by the velocity Verlet method48 with the time step of 1 fs. The 1 ns MD trajectories are collected after the equilibrium stage at every 50 fs. 2.2. QM Calculations. The STM images of some typical orientations of C60 molecule deposited on the substrate are (43) Wen, J.; Ma, J. J. Theor. Comput. Chem. 2009, 8, 677–690. (44) Meng, S.; Ma, J.; Jiang, Y. J. Phys. Chem. B 2007, 111, 4128–4136. (45) Meng, S.; Ma, J. J. Phys. Chem. B 2008, 112, 4313–4322. (46) Materials Studio, version 4.0, Accelrys Inc., San Diego, 2006. (47) Andersen, H. C. J. Chem. Phys. 1980, 72, 2384–2393. (48) Allen, M. P.; Tildesley, D. J. Computational Simulation of Liquids; Oxford University Press: New York, 1987. (49) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865– 3868. (50) Tersoff, J.; Hamann, D. R. Phys. Rev. Lett. 1983, 50, 1998–2001. (51) Vanderbilt, D. Phys. Rev. B 1990, 41, 7892–7895. (52) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2004.
Langmuir 2010, 26(8), 5595–5602
Wen and Ma
Article
Figure 2. Snapshots of (a) 4T and (b) 6T monolayers on Ag(111) surfaces are given in top and side views, respectively. The unit cell parameters, cell-edge length a, b, the angle γ between the axes, and adsorption height, d, are given in the insets. The experimental data in parentheses are taken from ref 58 (for 4T/Ag(111)) and ref 27 (for 6T/Ag(111)).
simulated by DFT as implemented in Material Studios packages.46 It is performed using the Perdew, Burke, Enzerholf (PBE) functional for exchange correlation49 by the TersoffHamann scheme.50 A slab model of 10.0 8.7 24.7 A˚ with two layers of Ag atoms is taken to reproduce the experimental STM image. Ultrasoft pseudopotentials for C and Ag atoms and a plane-wave basis set with a 330 eV cutoff energy are employed.51 The first Brillouin zone is approximated on a (2 3 1) k-point mesh. By using the popular Gaussian 03 package,52 QM/MM and ONIOM calculations are carried out to understand the interplay between the interfacial (Ag-nT and Ag-C60) and the intermolecular (nT-nT, nT-C60, and C60-C60) interactions. Similar to QM/MM, the ONIOM method applies different theoretical levels to treat different parts of a complicated system. In this work, we choose the ONIOM (MP2:HF) scheme, in which the sophisticated second-order Moeller-Plesset perturbation theory (MP2) is employed for the binding sites and simple Hartree-Fock (HF) is used for the remaining parts. Here, the distance between Ag atoms is fixed at 2.89 A˚, and the surface reconstruction is ignored. C60 and thiophene molecules are optimized by DFT calculations with B3LYP functional, respectively. LANL2DZ basis sets are used for Ag, and 6-31G(d) basis sets are employed for H, C, and S atoms. Basis set superposition error (BSSE) is remedied through the usage of the counterpoise (CP) correction.
3. Results and Discussion 3.1. Oligothiophenes on Ag(111). The packing structures of nT thin films on Ag(111) surface are the consequence of subtle balance between interfacial and intermolecular interactions. The weak physisorption and little site preference of thiophene adsorbed on Ag(111) system have been demonstrated by both DFT and MP2 calculations.43 It is well recognized that DFT suffers from the limitation in treating the weak interactions between the adsorbate and substrates.35,53 Although performance of DFT has (53) Haran, M.; Engstrom, J. R.; Clancy, P. J. Am. Chem. Soc. 2006, 128, 836–847. (54) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput. 2006, 2, 364–382. (55) Thonhauser, T.; Cooper, V. R.; Li, S.; Puzder, A.; Hyldgaard, P.; Langreth, D. C. Phys. Rev. B 2007, 76, 125112.
Langmuir 2010, 26(8), 5595–5602
been improved by taking the van der Waals interaction into account,33,54,55 MP2 calculations are still desired in many cases.43,56,57 The binding energy between nT (n = 4, 6) pairs is calculated by both QM (MP2/6-31G*) and MM (PCFF) calculations, with the results shown in Figure S1. The intralayer interaction is surveyed in x and y directions, respectively, whereas the interaction between nT layers is investigated by the binding energy as a function of the intermolecular distance in the z direction. It can be seen that the energy curves in x and y directions are rather flat without the appearance of evident minima. The interaction between nT chains in the xy plane is much weaker than the interlayer interaction in the z direction. It means the π-π stacking interaction in two different nT layers is much stronger than the side-by-side interaction in the same molecular layer. It is also conceivable that at low coverage oligothiophene molecules may be quite free in the nT monolayer, resulting in some disorder phase or domains. Moreover, the interaction between 6T layers is obviously stronger than that between 4T-4T pair, implying the more ordered packing in 6T monolayer than 4T layer. Monolayer Morphology: 4T vs 6T. On Ag(111), both 4T and 6T molecules are shown to grow into a well-ordered molecular layer with their molecular plane parallel to the surface. Figure 2 shows the snapshots along with the cell-edge length a, b, and the angle, γ, of the space unit cell. The experimental data27,58 are also given in parentheses for comparison. The unit cell parameters are evaluated as the average values over a large area on surface and the 1 ns duration. The values of cell length a and b are determined along the end-to-end vectors of the nearest neighboring nTs, and the angle γ is the cross-angle between a and b vectors. For both 4T/Ag(111) and 6T/Ag(111) systems, a very good agreement between the simulation (4T: a=17.4 A˚, b=6.0 A˚, γ=88.0; 6T: a= 27.6 A˚, b=6.0 A˚, γ=129) and STM experiment (4T:58 a=17.3 A˚, b = 6.7 A˚, γ = 83.5; 6T:27 a = 27.1 A˚, b = 6.3 A˚) is achieved, implying not only the applicability of PCFF in treating nTs but also a possibility of error cancellations in theoretical modeling. (56) Magnko, L.; Schweizer, M.; Rauhut, G.; Sch€utz, M.; Stoll, H.; Werner, H. J. Phys. Chem. Chem. Phys. 2002, 4, 1006–1013. (57) Ikeda, A.; Kameno, Y.; Nakao, Y.; Sato, H.; Sakaki, S. J. Organoment. Chem. 2007, 692, 299–306. (58) Gebauer, W.; B€assler, M.; Soukopp, A.; V€aterlein, C.; Fink, R.; Sokolowski, M.; Umbach, E. Synth. Met. 1996, 83, 227–230.
DOI: 10.1021/la903869g
5597
Article
Wen and Ma
Figure 4. Order parameter Sij changes with the coverage of nT. The cross-angle, δ, is defined as the angle between two long axes of nT molecules shown in the lower inset. The snapshots of 6T/ Ag(111) and 4T/Ag(111) with different coverage are also shown in the top insets.
Figure 3. Distribution of orientation angles, ω, defined as the long axes in nT (n = 4, 6) relative to the [110] direction: (a) 4T/Ag(111) and (b) 6T/Ag(111) with different coverage in the process of growth. The domains I, II, and III correspond to the range of 30-50, 70-90, and 100-120, respectively.
The highly ordered 6T monolayer is formed via side-by-side packing with a periodicity of 6.0 A˚ perpendicular to the chain direction, and the intermolecular distance along the molecular long axis is around 27.6 A˚. The average adsorption height and the torsion angle of the nT (n=4, 6) on the Ag(111) surface are further drawn from MD trajectory (Figure S3). Most of the fivemembered rings in oligothiophenes take the trans-conformation and lie flat above the Ag(111) surface, as reflected by the peak at around 173-178 in the distribution of torsion angles in Figure S3a. It is interesting to notice that 6T and 4T monolayers have identical average adsorption heights on Ag(111) surface, seemingly veiling the difference between these two systems. Taking a closer look at the height distribution curves in Figure S3b, however, we find that the 6T/Ag(111) system has a higher peak at around 3.2 A˚ than the 4T monolayer, indicating that 4T molecules pack less orderly than those 6T segments on the Ag(111) surface. Growth from Monolayer to Bilayer. In order to better understand the arrangement of oligothiophene on surface, we simulate the growth process of the nT molecules with the increase of nT coverage on the Ag(111) surface. The geometrical change in the local nT domains can be represented by variations in orientation angle, ω, with respect to the [110] direction of the Ag(111) surface (shown in Figure 3). The orientation angle of 4T distributes in a much broader range than that of 6T, indicating 4T molecules can move more freely on the Ag(111) surface. With the increase of coverage, both 4T and 6T adlayers show a noticeable increase in the population of domain II with ω=70-90. The enlargement of the size of the simulation model does not change the conclusion (see Figure S4). It is also interesting to detect the global change of morphology in the growth process of nT layers by using the order parameter, Sij 3 1 Sij ¼ Æcos2 δij æ 2 2 5598 DOI: 10.1021/la903869g
ð1Þ
where the cross-angle, δij, is the angle between the long axes of two nT chains (shown in the inset of Figure 4). A basic trend can be seen from Figure 4 that Sij of 6T/Ag(111) increases with the increase of coverage. The transformation from less ordered phase to long-range-ordered phase in the growth process of nT layers is clearly shown in the selected snapshots (Figure 4). It can be also found that the 6T layers maintain the ordered packing structure with an increasing film thickness. In addition, the order parameter of 4T is smaller than that of 6T at the same coverage, in accordance with the weaker 4T-4T interaction than 6T-6T binding. Therefore, 6T is suggested to be a promising template for developing self-assemblies prepatterned Ag(111) surface. 3.2. C60 on Ag(111) Surface. Before the MD study on the packing structures of the C60 monolayer on Ag(111) surface, the local electronic structures have been addressed by QM calculations. A series of Agn cluster models with increasing sizes, n=10, 16, 22, and 28, are used to investigate the substrate-adsorbate interaction. MP2 calculations are rather computationally demanding for complex systems such as C60-Ag(111); the hybrid QM/MM approach57,59-62 is thus applied. Here, the MP2 method is used in the QM part for the close contact region, including both the Agn cluster and a hexagon of C60 that faced the surface, and the universal force field (UFF) is chosen in the MM part for the rest of the C60 cage that is far away from the surface. Table S1 lists the binding energy of C60 molecule deposited on different adsorption sites (top, bridge, hcp-hollow, and fcchollow) by using both QM/MM and MM calculations. Although QM/MM can treat larger systems than QM methods, the C60-Ag28 complex is really an obstacle for routine QM/MM calculations. Thus, we resort to MM calculations for the purpose of making comparisons between the Ag22 and Ag28 surface models. One can see that the difference in binding energy upon the replacement of Ag22 by the larger Ag28 models is less than 1.0 kcal/mol. The Ag22 cluster model is hence adopted in the following calculations on binding energies of C60 molecule adsorbed on the Ag(111) surface, with an emphasis on the (59) Fomina, L.; Reyes, A.; Guadarrama, P.; Fomine, S. Int. J. Quantum Chem. 2004, 97, 679–687. (60) Kita, Y.; Wako, K.; Okada, I.; Tachikawa, M. J. Theor. Comput. Chem. 2005, 4, 49–58. (61) Fomina, L.; Reyes, A.; Fomine, S. Int. J. Quantum Chem. 2002, 89, 477–483. (62) Froese, R. D. J.; Morokuma, K. Chem. Phys. Lett. 1999, 305, 419–424.
Langmuir 2010, 26(8), 5595–5602
Wen and Ma
Article
Figure 5. Simulated STM images of C60 monolayer with different molecular orientations adsorbed on Ag(111) surface: (a) hexagon; (b) pentagon; (c) 6:6 bond; and (d) 5:6 bond under Vs = -2.0 V (above) and Vs = þ2.0 V (below).
different orientations. There is no significant site preference for adsorption of C60 on Ag(111), similar to what was reported for 6T/Ag(111) systems.43 Our QM/MM and MM results agree well with the DFT adsorption energies of C60 deposited on Ag(111) surface with various molecular orientations.36 The packing structure of C60 monolayer is governed by not only the substrate-molecule (C60-Ag) but also the intermolecular (C60-C60) interactions. The binding energies between two C60 molecules are thus calculated by the QM/MM (MP2:UFF) method in both the pentagon-pentagon (with the pentagon rings stacking face-to-face) and 6:6 bond-6:6 bond (with the 6:6 bonds faced to each other) models. As shown in Figure S5, the QM part is highlighted with the ball-stick style and the MM region is shown in the line pattern. Because of the high symmetry of fullerene, the C60-C60 interaction depends strongly on their separation, d, but weakly on their relative orientation. In fact, the pentagon-pentagon and 6:6 bond-6:6 bond configurations of C60 dimers give nearly identical curves of binding energies varied with d. We find that the calculated C60-C60 binding energies match the Lennard-Jones 12-6 potential functions very well (see the dash-dotted lines in Figure S5), revealing the essential van der Waals interaction among C60 molecules. Furthermore, binding energy curves for various orientations obtained by MM (PCFF) are also close to the QM/MM results. Hence, the MM (PCFF) is further applied to the MD simulations of C60 monolayer on Ag(111) surface. Within the √ PCFF √ framework, our MD simulations give an ordered (2 3 2 3)R30 packings of C60 monolayer on the Ag(111) surface, as shown in Figure 5. The simulated C60 arrays accord with those deduced from the STM images of C60 on the Ag(111) surface.17,19 The average C60-C60 distance (between the centers of adjacent cages) in our simulated C60/Ag(111) system is 9.9 A˚ (Figure 5), in good agreement with the experimental observations (∼10.0 A˚)17,19,24,27 on hcp close-packed C60 monolayers. It is also interesting to test the influence of Ag(111) surface on the C60-C60 distance through the comparison between the C60/Ag(111) and C60 dimeric systems. As shown in Figure S5, the equilibrium distance between the centers of the facing pentagon rings in a C60 dimer is around 3.5 A˚. So the center-to-center C60-C60 distance is 10.0 A˚ when the radius of a C60 cage is taken into account. In fact, the intermolecular distances between C60 cages with and without Ag(111) template are very close to each other, indicating very weak template effect of Ag(111) surface on packings of C60. Langmuir 2010, 26(8), 5595–5602
There are also many theoretical simulations of the electron density near the Fermi level of the C60/metal systems to reproduce the experimental STM images.18,28-30,35,63 Subsequently, on the basis of MD snapshots of C60/Ag(111), we select four typical conformations of C60 on Ag(111), (a) hexagon, (b) pentagon, (c) 6:6 bond, and (d) 5:6 bond, as outlined in the insets of Figure 5. Among them, the hexagon and 6:6 bond orientations appear in abundance with the percentages of 26.3% and 37.5%, respectively, in good agreement with the experiments (hexagon: 32.2%; 6:6 bond: 38.4%).64 In addition, the orientation of 5:6 bond faced to the surface has a non-negligible population of around 18.8%. DFT (PBE) calculations are then carried out on these four orientations with the simulated STM images under both negative and positive bias of sample shown in Figure 5. Comparison of the STM images of C60 with the same orientation but under different bias is made between Figure 5 (with the bias of (2.0 V) and Figure S6 ((1.0 V). It can be found that the strength of bias has little influence on the STM images provided that voltage is not very high. The three-leaf shape and a symmetric dumbbell pattern are associated with the bonds between pentagon and hexagon rings of C60 cage at positive bias in Figure 5a (S6a) and Figure 5c (S6c), respectively, corresponding to the 3-fold and 2-fold symmetric feature that demonstrated in experiments.19 A ringlike shape appears for the pentagon ring of C60 cage at positive bias in Figure 5b (S6b). The stripelike internal features of C60 are shown at the negative bias, related to the cage structure of the molecule.31 The highest occupied molecular orbital (HOMO) of C60 molecule is centered on the 6:6 bond (a double bond between hexagon rings), while the lowest unoccupied molecular orbital (LUMO) is mainly involved with the 5:6 bond (a bond between hexagon and pentagon rings). Recalling the similar STM pictures for the depositions of C60 on Au(111), Cu(111), and Si (111)-(7 7) surfaces,14-19 the interfacial interactions between C60 and (111) terraces are deduced to bear a quite similar nature. 3.3. C60-Thiophene on Ag(111). The supramolecular assemblies composed of 6T and C60 on Ag(111) surface brought new possibilities in constructions of metal-organic nanoarchitecture.24,25,27 A detailed insight into the patterning of the complex binary molecular layers on surface requires the information on the relative strength of C60-Ag(111), 6T-Ag(111), C60-C60, 6T-6T, and C60-6T interactions. Therefore, we carry out a (63) Maruyama, Y.; Ohno, K.; Kawazoe, Y. Phys. Rev. B 1995, 52, 2070–2075. (64) Chen, W.; Wee, A. T. S., personal communications.
DOI: 10.1021/la903869g
5599
Article
Figure 6. Snapshots of 0.5 ML C60 on (a) 1.0 ML and (b) 2.0 ML 6T/Ag(111) with the unit cell parameters shown in the insets. The data in parentheses are measured from experimental STM images of ref 24.
systematic study on the interactions among the models of Ag22, thiophene, and C60, with results collected in Figures S2 and S7-S9. The insensitivity in orientations of adsorbate on surface is displayed for both T/Ag(111) and C60/Ag(111) systems. The binding energy between C60 and thiophene molecules is also evaluated with the QM/MM (MP2:UFF) method. The thiophene and partial Cn fragment (n=12-36) of C60 are taken into the QM part. When the QM core of C60 is large enough (n g 20), binding energy will not change much with the increase of model size. The C60 with C32 fragment is hence chosen in higher QM level of hybrid ONIOM (MP2:HF) scheme to calculate the binding energies of C60-T dimer with different packing orientations. The binding energies are further corrected by basis set superposition error (BSSE). The relative tilting angle, j, of the dimer is defined as the vertical orientation of C60 to the direction of sulfur atom in thiophene. The minimum at j = 90 (in Figure S8b) indicates that thiophene molecule favors the planar conformation lying on top of C60. The change in binding energy with different orientation angle, β, is also shown in Figure S8c when the thiophene molecule is flat-lying on C60 molecule. The flatness of the potential curves with the increase in β may originate from the high symmetry of the C60 cage. It is also interesting to investigate the effect of curvature of C60 on the intermolecular interaction (Figure S9). As expected, there is little effect from the lateral distortion of C60 (in x or y directions) but dramatic influence from the stretching of C60 along the longitude (z direction) on the C60-T intermolecular interactions. 3.4. C60 Molecular Chanis on 6T Prepatterned Ag(111). The template effects of 6T monolayer on the upmost C60 selfassembly can be seen from the changes in packing structure with respect to the molecular arrangement of C60/Ag(111). It is found 5600 DOI: 10.1021/la903869g
Wen and Ma
Figure 7. Distributions of four orientations, hexagon, pentagon, 6:6 bond, and 5:6 bond, for C60 deposited on (a) Ag(111) surface and (b) 6T monolayer preadsorbed Ag(111) surface.
by both PCFF and QM/MM calculations that the interfacial interaction between C60 and Ag surface is stronger than that of thiophene and Ag surface and intermolecular interactions. From Figure 6a, one can find that when ca. 0.5 monolayer (ML) C60 deposited on top of 6T monolayer, there are two parallel chains with 46 tilt angle relative to the [112] direction of the Ag(111) surface. The C60 chains are aligned in alternating distances of 19.0 and 21.5 A˚, respectively, between the two neighboring C60 arrays. The interchain distances are in good agreement with the experimental data (18.4 and 22.8 A˚).24 As illustrated in the inset of Figure 6a, the template effect of 6T is exhibited from the matching of the interchain distance of 21.5 A˚ between C60 arrays with the head-to-end length of 22.3 A˚ for a 6T molecule. It should be mentioned that the sufficiently large surface model is desired to reproduce the experiments. The enlargement of slab’s size from 55.0 34.7 44.7 A˚ to 110.1 69.3 44.7 A˚ significantly reduces the deviations from experimental results. From MD simulations, we notice that it is also possible for the C60 molecules to directly deposit on Ag surface, squeezing the bottom 6T molecule to the nearby adsorption site due to the stronger interaction between Ag surface and C60 than the surface-6T interaction, which agrees with experimental observations.64 Deposition of C60 molecules on the bilayer of 6T on Ag(111) surface further √ lead to√some cluster domains, exhibiting a partial hexagonal (2 3 2 3)R30 organized structure, as shown in Figure 6b. The growth directions of C60 wires in two different domains are in a cross angle of about 60. Going to the details, the unit cell parameters are also shown in the inset of Figure 6b. C60 monolayer is close-packed and rotated by 5 with respect to the underlying Ag(111) lattice. Again, the distance between the centers of two neighboring C60 molecules is about 9.9 A˚. The effect of substrate on the C60 layer is negligible when the hybrid adlayer contains ca. 2.0 ML of 6T preadsorbed on Ag(111) Langmuir 2010, 26(8), 5595–5602
Wen and Ma
Article
Figure 8. Distributions of (a) plane angle, σ, and (b) orientation angle, ω, of thiophene ring in the C60-on-6T/Ag(111) and 6T-on-C60/ Ag(111) systems. The coverage of the upmost C60 or 6T is changed on (c) 1 ML 6T/Ag(111) and (d) 1 ML C60 /Ag(111) substrates.
surface. It can be seen that the molecular packing of the fullerene adlayers is mainly directed by the orientations and anisotropy of 6T layers that sandwiched between the C60 monolayer and Ag(111) substrate. It is also shown in Figure S10 that the temperature has little influence on the C60 interchain distance. The curves in distributions at 298 and 360 K both have similar peaks in interchain distance around 10.0, 17.0, and 20.0 A˚. The insensitivity to the simulated temperature may partially come from the temperature-independent parameters in the traditional force fields. In analogy to the C60/Ag(111) system, there are also four major orientations of C60 facing the 6T carpet on Ag(111). The populations of four orientations of 1.0 and 0.5 ML fullerene adsorbed on Ag(111) and 6T/Ag(111) systems, respectively, are compared in Figure 7. On Ag(111) surface, most adsorption configurations prefer to facing the substrate by hexagon and 6:6 bond of C60. However, when C60 is added on the predeposited 6T monolayer on Ag(111) surface, nearly 40% C60 cages are adsorbed with their 6:6 bond faced to 6T template. The other three orientations have similar populations in the range of around 8-22%. Moreover, the modification of Ag(111) surface by 6T monolayer renders the hexagon orientation reduces remarkably from 26.3% to 12.2%, while the orientation with pentagon facing the substrate increases by more than 10%. In addition, the populations of the four orientations change little with the variations in C60 and 6T coverages, as shown in Figure S11. 3.5. Effect of Deposition Order. As a comparison with the C60-on-6T/Ag(111) system, 6T molecules are deposited on top of C60 monolayer, from which the influence of deposition order can be shown. Obviously, the 6T-on-C60 layer is less ordered than the C60-on-6T/Ag(111). In order to survey the extent of twist and tilt of 6T segment on C60, the plane angle, σ, is calculated in terms of the angle between normal vector υi of the ith thiophene ring and the [111] direction of Ag(111) lattice. In Figure 8a there is a peak near σ=10 in the distribution curve of the C60-on-6T/Ag(111) system, while an evident tilting of 6T backbone with σ=20 is observed for the 6T-on-C60/Ag(111) system. It reflects that 6T backbones partially bend on the waved C60 carpet, which is quite different from the flat-lying 6T template on Ag(111) surface. As the coverage increases, some 6T molecules even tend to stand perpendicularly on C60 monolayer. The C60 substrates also affect the orientation of 6T segments. The long axes of most of 6T molecules tilt Langmuir 2010, 26(8), 5595–5602
about 80 against the direction of [110] on the Ag(111) surface, but the orientation angle increases to 100 when 6T deposited on C60 substrate (Figure 8b). For the C60-on-6T/Ag(111) system, we also find in Figure 8c that 6T chains become less ordered with the increasing orientation angle as the coverage increases. In contrast, on the C60 premodified Ag(111) surface, the increase of 6T coverage gives rise to more ordered packing in the upmost layer (Figure 8d).
4. Conclusions In this study, the packing structures of self-assembly nT (n=4, 6) on Ag(111) surface and subsequent codeposition of C60 and 6T on substrate have been investigated by MD simulations in conjunction with quantum mechanics (QM) calculations of intermolecular and interfacial interactions. The oligothiophene (nT, n=4, 6) chains turn to ordered packing as the coverage of the monolayer increases. The highly ordered nT monolayers are flatlying on Ag(111) surface with the long axes deviating from [110] direction by about 70-90. The predeposition of Ag(111) surface by 6T nanostripes √ √makes C60 molecules change from the close-packed (2 3 2 3)R30 structure on isotopic Ag(111) surface to the chainlike arrays with alternating interchain spacing. Most of C60 molecules deposit with 6:6 bond facing the 6T template. The orientation of self-organized hybrid bilayer is governed by both C60 and 6T coverages. When we change the deposition order of 6T and C60 coadsorption system, the 6T segments are found to curve on the underlying C60 carpet. Useful information is provided for both the simulation of complex systems and rational design of the electronic devices. Acknowledgment. The authors sincerely thank three referees for their good suggestions. This work was supported by the National Natural Science Foundation of China (Grant No. 20825312) and the Fok Ying Tong Education Foundation (Grant No. 111013). Supporting Information Available: Table S1 gives binding energies of C60-Agn with different binding sites and increasing cluster size, n=10-28; Figures S1 and S2 compare the binding energies of nT-nT (n=4, 6), T-Ag22, and C60-Ag22 complex systems among various theoretical levels, such as QM(MP2/6-31G*), MM(PCFF), and QM/MM(MP2: DOI: 10.1021/la903869g
5601
Article
UFF); different packing parameters of nT (n = 4, 6) on Ag(111) surface obtained from MD simulations is depicted in Figures S3 and S4, varied with different coverage and different surface models, respectively; Figure S5 shows the change in binding energy as a function of distance between C60 dimer; the simulated STM images of C60 on Ag(111) surface are shown in Figure S6 under the bias of (1.0 V; Figures S7, S8, and S9 give the binding energies of T-C60,
5602 DOI: 10.1021/la903869g
Wen and Ma
changed with the different models, packing styles, and size of C60 cage, respectively; Figure S10 displays the distributions of the center-to-center distances between C60 molecules at temperature of 298 and 360 K; distributions of the four orientations of C60 molecules on Ag(111) surface are shown in Figure S11, varied with different coverage ratio of 6T to C60 molecules. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2010, 26(8), 5595–5602