NANO LETTERS

A Self-Directed Growth Process for Creating Covalently Bonded Molecular Assemblies on the HSi(100)-31 Surface
Xiao Tong, Gino A. DiLabio, and Robert A. Wolkow*,,
National Institute for Nanotechnology, National Research Council of Canada, W6-010 Sixth Floor ECERF, UniVersity of Alberta, 9107-116th Street, Edmonton, Alberta T6G 2V4 Canada, and Department of Physics, 412 AVadh Bhatia Physics Lab, UniVersity of Alberta, Edmonton, AB, Canada T6G 2J1
Received February 6, 2004; Revised Manuscript Received March 16, 2004

2004 Vol. 4, No. 5 979-983

ABSTRACT
A chain reaction initiated at a dangling bond on a H-terminated Si(100)-31 surface leads to the creation of contiguous, linear multimolecular assemblies. In contrast to a similar growth process observed on the HSi(100)-21 surface, the linear structures grow in the cross-row direction, rather than parallel to dimer rows. This process is enabled by both an uncommonly high rate of H atom diffusion, specifically in the cross-row direction, and a low barrier to H atom abstraction from dihydride sites. These results demonstrate that anisotropy inherent to the substrate can be imposed upon molecular assemblies formed via this self-directed growth process.

Introduction. Nanoscale patterning1-7 of surfaces has received considerable attention for its potential application in, for example, sensors,8 light emitters,9 and future molecular devices.10,11 The hybrid properties of small molecular assemblies on silicon have particular appeal as a basis for sensors and as a platform for a variety of molecular scale devices. Silicon is made more attractive by the wealth of compatible lithographic processes that can be adapted to simplify the challenges associated with placing and addressing such small entities.10 Intimate knowledge of surface chemical processes allows us to go beyond arduous, entirely scanned probe-based patterning procedures, relying instead on natural chemical tendencies to self-direct the building of stable, covalently bonded, hybrid organic-silicon structures.12,13 The self-directed process has a number of attractive features. Because the growth initiation points are defined, not random, absolute position over growth is achieved. Numerous initiation points can be defined, creating the potential for parallel (that is, simultaneous) fabrication. Furthermore, because the substrate provides a templating function, the structures formed have a predefined structure. The extent of growth can be controlled by the turning of a valve that feeds the reacting molecules into the vacuum
* Corresponding author. E-mail: bob.wolkow@nrc.ca. National Institute for Nanotechnology, National Research Council Canada. Department of Physics, University of Alberta. 10.1021/nl049796g CCC: $27.50 Published on Web 04/09/2004 2004 American Chemical Society

chamber. The resulting structures offer opportunities to directly study molecule-molecule interactions underlying organic electronics and also point to routes to defining nanoscale hybrid silicon-organic entities of unique properties. To date, it has been shown that styrene,12 vinyl ferrocene,13 cyclopropyl methyl ketone,14 and long-chain alkenes15 can undergo a radical mediated chain reaction on H-terminated silicon to create stable, ordered, covalently bonded molecular assemblies. Forthcoming studies will describe other variants on the self-directed growth process, describing extended control over growth dynamics and/or of resultant properties. In this work, we describe styrene molecular line growth on the 31 H-terminated Si(100) surface. In contrast to growth on the 21 H-Si(100) surface, which follows the dimer row direction, growth on the 31 surface proceeds at a right angle to the dimer row direction. This result suggests that a judicious combination of substrate specification and molecular design may lead to our goal of providing a means for arbitrary growth direction control for organic nanostructure formation on silicon surfaces. Experimental Section. The experiments were carried out at room temperature in an ultrahigh vacuum (UHV) chamber with a base pressure below 2 10-10 Torr. An As-doped Si(100) crystal (0.005 Ohm-cm) was cleaned by repeated flashing to 1200 C and H-terminated at 330 C by exposure

Figure 1. Sequence of STM images (Vs ) -2.5 V, I ) 0.06 nA) showing the growth of a styrene line on the H-Si(100) 31 surface. A clean surface with a DB on a dimer row site which diffuses in a direction perpendicular to the rows, as shown in (a) to (c). (d) A styrene molecule attaches at the DB site, abstracting a H atom from the neighboring monomer site. A diffusion event moves the DB to a position on the next dimer row. In frames (e) and (f), line growth continues. In (g), a deviation in the growth process occurs as the DB hops to the next dimer row. A styrene attaches but, as shown in (h), abstracts in the reverse direction, creating a new DB in the void between the assembled molecules. A final styrene attaches, shown in (i), and remains as a radical.

to atomic hydrogen from a tungsten filament. Single dangling bonds (DBs) were generated by hydrogen desorption with the STM tip by employing a sample bias of +3.5 V and currents below 40 nA.16 Typical imaging conditions were a sample bias of Vsb) -2.5 V and tunneling current It ) 55 pA. Styrene was degassed by performing several freezepump-thaw cycles before controlled introduction into the UHV chamber via a variable leak valve. Calculations, using the Gaussian-03 program package,17 were also carried out on small models of silicon clusters in order to obtain estimates of the bond energies and barrier heights associated with line growth. Results and Discussion. Figure 1 shows a sequence of STM images of an H-terminated Si(100) 31 surface. Alternating rows of monohydride (H-capped Si dimers) and dihydride (SiH2, monomer) units appear as relatively light and dark rows, respectively.18 To understand the H-terminated 31 Si(100) surface it is helpful to first consider the
980

native reconstruction of the 21 Si(100) surface. There, Si atoms exist in pairs (dimers), giving each atom three bonds and one dangling bond. When atomic H is added to create a 21 monohydride structure, all dangling bonds are capped and no silicon-silicon bonds are broken. As the Figure 1 schematic shows, the 31 surface contains both H-capped dimers, just as on the 21 surface, but also SiH2 dihydride units. The substrate temperature during H exposure determines the phase which forms; at room temperature the 31 H-terminated surface forms while the 21 structure results at approximately 300 C. The 31 structure does not exist in a H-free state. The indicated bright spot in Figure 1a-c corresponds to a missing H atom, that is, a Si dangling bond. DBs are found exclusively on the dimer rows. DBs are frequently observed to migrate, exclusively perpendicular to the row direction from one dimer row to another, without being imaged on an intermediate dihydride site. DBs also move back and forth
Nano Lett., Vol. 4, No. 5, 2004

on a single dimer. Counting row-to-row jumps, and assuming a preexponential of 1013, the activation energy for interrow diffusion is found to be 0.9 eV. Motion along dimer rows was not observed. Thermal DB motion has been reported on the H-Si(100) 21 surface at elevated temperatures. The activation energies for intrarow (dimer-to-dimer on the same dimer row) and intradimer (on the same dimer) H diffusion on the 21 surface were measured to be 1.75 ( 0.10 eV and 1.01 ( 0.05 eV, respectively.19 Interrow diffusion does not occur. DB motion has been induced to occur on the 21 surface at room temperature when scanning at conditions more harsh (larger current and voltage) than applied here.20 Likely, some tip-induced DB diffusion has occurred in this study. In cases where tip-induced motion is dominant, truncated images of adsorbates are observed, appearing as the adsorbate (or void) jumps at a moment when the tip is quite near it. As we typically do not see truncated images, the observed diffusion appears to be largely thermally induced. The salient point with respect to this study is that row-to-row DB motion is far more facile on the 31 H-Si(100) surface than on the 21 surface. It will emerge in the discussion to follow that a multimolecular growth process, dependent upon interrow DB diffusion, occurs in the absence of STM scanning, supporting the view that DB motion can occur via thermal activation at room temperature. We emphasize that facile dangling bond diffusion across rows on the H-Si(100) 31 surface is opposite to the preferred diffusion direction on the 21 surface. Our STM images of the 31 surface, and those reported previously,21,22 show that virtually all DBs appear on the dimer rows. Recent DFT calculations considered various distributions of H atoms (different from those in this work) on a model 31 surface and found that H atom occupation of dimer sites was disfavored.23 Our own calculations confirm that the Si-H bond on the dimer is 0.1 eV weaker than that of the monomer, which is consistent with the dangling bond residing on the dimer sites. The sequence of images 1c-i reveals the steps involved in the self-directed growth process on the H-Si(100) 31 surface. In (c), a styrene-free surface with an isolated DB positioned on a dimer row is shown. In (d), adsorption of one styrene molecule has occurred at the preexisting DB. Also apparent in (d) is a new DB, created when the attached molecule, initially bound as a radical, abstracts an H from an adjacent surface site. In frames (e) and (f), the process is seen to repeat. In (g), the DB has migrated to a more distant dimer site, causing a disruption in the subsequent growth process. The DB can then diffuse back-and-forth on the dimer. Evidently, steric interactions inhibit a second styrene from binding to the dimer Si closest to the dihydride. Because the DB in frame (h) has diffused to a position free of steric hindrance, growth can proceed in the opposite direction. In (h), a styrene molecule has attached to the DB shown in (g) and has then reversed the growth direction by abstracting a H atom from the site nearer the preexisting segment of molecules. In the final frame, (i), one last styrene adsorbs. It appears to remain in the radical state, seemingly unable
Nano Lett., Vol. 4, No. 5, 2004

Figure 2. Proposed mechanism for styrene chain reaction on H-terminated Si(100) 31 surface. (a) A styrene molecule approaches a 31 surface with a preexisting DB on a dimer row. (b) Initial reaction with the DB leads to formation of a radical center at the C adjacent to the point of attachment. (c) Abstraction of a H atom from adjacent monomer site, followed in (d) by H atom migration resulting in a DB at the adjacent dimer row. In (e) the DB is shown after moving to the more distant side of the dimer. At times, attachment to monomer sites occurs, resulting in more closely spaced molecules. After such attachments, growth proceeds in the cross-row direction as indicated here.

to abstract an H atom from a nearby site. Typically, longer lines are observed with growth stopping when defects are encountered. It is therefore possible that the repeated scanning that was required to obtain the sequences in Figure 1 led to a tip-induced DB in 1(g). Additional styrene exposure did not lead to further growth of the structure in 1(g). The perceived greater height of the last adsorbate added (Figure 1g) is consistent with that species having a radical center. A radical has associated with it a greater local state density and therefore a more prominent appearance in STM. This observation, together with the images presented here showing resolved adsorbed molecules and surface dangling bonds, substantiates our earlier claims12,24 that the growth process we observe is not polymer growth, but a multimolecular radical-mediated process creating a line of molecules, each attached to the surface, and none attached to another molecule.12 The growth stoppage described here also lends some support to our earlier claim related to the growth of irregular styrene assemblies on the H-Si(111) surface.24 That regular surface lacks the distinct in-plane anisotropy of the Si(100) surface, and thus meandering growth on H-Si(111) is observed. When a molecule becomes surrounded by other molecules and cannot achieve H abstraction, growth stops. A proposed mechanism, consistent with most of our observations, is presented in Figure 2. In frame (a), a styrene molecule approaches an H-terminated Si(100) 31 surface with a preexisting DB localized on the dimer row. In (b), the styrene molecule reacts via its vinyl group to form a C-Si bond, creating a carbon-centered radical. In a departure
981

that the determining structural qualities that lead to alongrow or across-row growth are highly local to the reaction.27 Conclusion. Dangling bond diffusion perpendicular to the rows direction is extraordinarily facile on the 31 surface. This, combined with the low barrier to H-atom abstraction by the carbon-centered radical from the dihydride, leads to cross-row growth of molecular lines via a self-directed mechanism. This result is closely related to the mechanism observed previously on the 2 1 surface on which, because of the lack of the dihydride features, line growth proceeds along the dimer row direction. The results suggest that the nature of the silicon substrate is an exploitable feature, which can be used as a means of controlling the direction of grown of organic nanostructures on silicon surfaces. Acknowledgment. We are indebted to Douglas Moffatt for helpful discussions and assistance with instrumentation. We thank WestGrid for access to computational facilities and iCORE for financial assistance. We thank K. U. Ingold for insightful comments. References
(1) Eigler, D. M.; Schweizer, E. K. Nature, 1990, 344, 524-526. (2) Avouris, Ph.; Walkup, R. E.; Rossi, A. R.; Akpati, H. C.; Nordlander, P. Surf. Sci., 1996, 363, 368-377. (3) Shen, T. C.; Wang, C.; Tucker, J. R. Phys. ReV. Lett. 1997, 78, 12711274. (4) Abeln, G. C.; Lee, S. Y.; Lyding, J. W.; Thompson, D. S.; Moore, J. S. Appl. Phys. Lett. 1997, 70, 2747-2749. (5) Xia, Y. N.; Whitesides, G. M.; Annu. ReV. Mater. Sci. 1998, 28, 153-184. (6) Hong, S.; Mirkin, C. A. Science 2000, 288, 1808-1811. (7) Kruse, P.; Wolkow, R. A. Appl. Phys. Lett. 2002, 81, 4422-4424. (8) Bollani, M.; Piagge, R.; Charai, A. Narducci, D. Appl. Surf. Sci. 2001, 175, 379-385. (9) Thirstrup, C.; Sakurai, M.; Stokbro, K.; Aono, M. Phys. ReV. Lett. 1999, 82, 1241-1244. (10) Wolkow, R. A. Annu. ReV. Phys. Chem. 1999, 50, 413-441. (11) We are currently studying a prototype device that will allow electrical monitoring of changes in molecular adsorption on silicon, to be published. (12) Lopinski, G. P.; Wayner, D. D. M.; Wolkow, R. A. Nature 2000, 406, 48-51. (13) Kruse, P.; Johnson, E. R.; DiLabio, G. A.; Wolkow, R. A. Nano Lett. 2000, 2, 807-810. (14) Tong, X.; DiLabio, G. A.; Clarkin, O. J.; Wolkow, R. A. Nano Lett. 2004, 4, 357-360. (15) DiLabio, G. A.; Piva, P. G.; Kruse, P.; Wolkow, R. A. Phys. ReV. Lett., submitted. (16) Shen, T.-C.; Wang, C.; Abelin, G. C.; Tucker, J. R.; Lyding, J. W.; Avouris, Ph.; Walkup, R. E. Science 1995, 268, 1590-1592. (17) 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.; 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 B.04, Gaussian, Inc.: Pittsburgh, PA, 2003. (18) Boland, J. J. Phys. ReV. Lett. 1990, 65, 3325-3328. Nano Lett., Vol. 4, No. 5, 2004

Figure 3. STM image (Vs ) -2.0 V, I ) 0.1 nA) of styrene lines on an H-terminated Si(100) 31 surface which before exposure had a dilute concentration of single DBs. Examples of typical lines (S) and more rare wider lines (D) are shown, as well as a line following the dimer row direction (A) where a small 21 region existed.

from growth exhibited on the 21 H-Si(100) surface, an abstraction event takes place at a dihydride site, as shown in (c). Our experimental evidence for this step is indirectafter a molecule adds, a DB appears on the next dimer row. Direct abstraction from the next row dimer is beyond reach. Our calculations indicate that H-atom abstraction from the dihydride site is quite facile. The barrier for this process is ca. 0.55 eV,25 approximately 0.3 eV lower than H-abstraction from a neighboring dimer row on the H-Si(100) surface. We surmize that abstraction from the well aligned dihydride initially occurs, followed by replacement by a hydrogen atom from the adjacent dimer site. Results from modeling and our observations of row-to-row DB motion, imaged in Figures 1a-c, support this step. The surface shown in Figure 3 has 21 and 31 domains intermixed. The juxtaposition of the two phases offers an opportunity to see, in one image, evidence of growth along rows in 21 regions and the new mechanism, which proceeds across rows in the 31 domains. In addition to providing direct confirmation that a new growth mechanism specific to the 31 domain does exist, this image also demonstrates that the substrate provides the directing function within the self-directed growth mechanism. These observations allow global or long-range effects to be ruled out as determining which mode of growth results. It is evident
982

(19) Hill, E.; Freelon, B.; Ganz, E. Phys. ReV. B 1999, 60, 15896-15900. (20) Quaade, U. J.; Stokbro, K.; Thirstrup, C.; Grey, F. Surf. Sci. 1998, 415, L1037-L1045. (21) Thirstrup, C.; Sakurai, M.; Nakayama, T.; Aono, M. Surf. Sci. 1998, 411, 203-214. (22) Sakurai, M.; Thirstrup, C.; Nakayama, T.; Aono, M. Surf. Sci. 1997, 386, 154-160. (23) Shen, T. C.; Steckel, J. A.; Jordan, K. D. Surf. Sci. 2000, 446, 211218. (24) Cicero, R. L.; Chidsey, C. E. D.; Lopinski, G. P.; Wayner, D. D. M.; Wolkow, R. A.; Langmuir 2002, 18, 305-307. (25) A two dimer, two monomer cluster of 34 Si atoms was used as a model for the 31 surface for the barrier height calculations. The truncated Si-Si (bulk) bonds were capped by hydrogen atoms and

constrained. The distance between the carbon-centered radical (from cluster-bound styrene) and the labile hydrogen atom was increment by 0.05 and held fixed while the cluster was geometry optimized at the B3LYP26/6-31G* level of theory. The transition state (TS) structure at this level of theory occurs at a C-H distance of 1.55 . The barrier height is computed to be the B3LYP/6-31G* difference between the TS and reactant complex energies. (26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (27) Figure 3 also shows a wider linear structure labeled D. Such structures appear to be related to phase boundaries between 31 domains.

NL049796G

Nano Lett., Vol. 4, No. 5, 2004

983