NANO LETTERS
Ring-Opening Radical Clock Reactions for Hybrid Organic−Silicon Surface Nanostructures: A New Self-Directed Growth Mechanism and Kinetic Insights
2004 Vol. 4, No. 2 357-360
Xiao Tong,† Gino A. DiLabio,† Owen J. Clarkin,†,‡ and Robert A. Wolkow*,†,§ National Institute for Nanotechnology, National Research Council of Canada, W6-010, 9107 116th Street, Edmonton, Alberta, Canada T6G 2V4, Department of Chemistry, Carleton UniVersity, Ottawa, Ontario, Canada K1S 5B6, Department of Physics, UniVersity of Alberta, Edmonton, Alberta, Canada T6G 2J1 Received November 13, 2003; Revised Manuscript Received December 8, 2003
ABSTRACT Cyclopropyl methyl ketone molecules react at single dangling bonds on an H-terminated Si(100) surface to form a metastable radical species attached via the oxygen atom. While the adsorbate has the capacity to abstract an H atom from an adjacent surface site, in analogy with the previously studied reaction of styrene, it appears not to do so. Rather, the cyclopropyl ring opens, thereby shifting the radical to a position distant from the point of attachment. H-abstraction then occurs from one of a variety of surface sites within range of the radical. The surface dangling bond created by abstraction allows the process to repeat, leading to the creation of a contiguous string of molecules attached to the surface. The assembly process follows a meandering path unlike in the case of styrene where H-abstraction is restricted to adjacent surface sites, resulting in straight multimolecular assemblies. This result hints at an opportunity to gain control over growth direction through judicious addition of constraints within the ring-opened radical adsorbates. Also, because the unimolecular rates of such ring opening reactions have been well characterized (“radical clocks”) we anticipate that in the future, the kinetics of the self-directed growth process may be determined.
Introduction. We anticipate that hybrid organic-silicon nanostructures will underpin a new generation of devices that go beyond the traditional scope of semiconductor devices to include functions such as molecular sensors.1 Control over the formation of ordered structures on silicon surfaces requires that we develop a collection of techniques that allow us to control the position and conformation of individual molecules and specific physiochemical properties of the nanostructures. To this end, our group has demonstrated the remarkable ability of styrene2 and vinylferrocene3 molecules to undergo a chain reaction on an H-terminated Si(100) surface. These chain reactions result in the “self-directed” growth of lines along the silicon dimer rows. The properties of these lines can, in principle, be tuned by using ring substituents or different metal cations. The chain reaction mechanism, originally proposed by Linford et al. in the context of monolayer film growth on H-Si(111),4 can be described as follows. First, a molecule containing a double bond, typically a CdC group, attaches * Corresponding author. E-mail:
[email protected]. † NINT-NRC ‡ Carleton University. § University of Alberta. 10.1021/nl035021g CCC: $27.50 Published on Web 12/30/2003
© 2004 American Chemical Society
to the surface-silicon dangling bond. The π-bond of the molecule breaks and a σ-bond forms between one C atom of the molecule and the Si atom of the surface. This process leaves an unpaired electron on the molecule at the C atom that is R to the C attached to the surface. Second, the molecular radical is quenched by a hydrogen atom abstraction from a neighboring surface silicon, which results in a new dangling silicon bond at that adjacent site. H-atom abstraction is believed to be slow, although we have only been able to make crude estimates of the rate for this process.2 Last, a molecule attaches to the new dangling bond and the steps are repeated. On the H-Si(100)-2×1 surface, abstraction preferentially (but not always2) occurs from the next silicon dimer row, which results in line growth in a direction parallel to the dimer rows. On the H-Si(111)-1×1 surface, all surface sites are equivalent and growth proceeds via a random walk process yielding compact islands of molecules.5 The underlying surface structure is thus one means of controlling the nature of the nanostructures created by the chain reaction mechanism. In principle, the types of nanostructures formed on silicon surfaces can also be adjusted by changing the nature of the organic molecule to which the surface is exposed.
From our styrene and vinylferrocene studies, we found that the CdC double bond of the vinyl group restricts the physical distance that separates chemisorbed molecules because the “reach” of the carbon-centered radical for the H-atom abstraction step is limited by the geometry associated with the Si-C-C moiety. One interesting way of extending the radical reach is to use a molecule that will undergo a post-attachment rearrangement such that the unpaired electron is on an atom far from the site of chemisorption. One example of this type of system is the “radical clock” reaction, typified by the ring opening reaction 1 of cyclopropylcarbinyl to form the 3-butenyl radical.
The rate for reaction 1 is well known.6 By exposing Si dangling bonds to a molecular species that is capable of generating a moiety similar to cyclopropylcarbinyl after chemisorption, we expect to introduce a ring opening step to the line growth process (vide infra). In this work, we present a preliminary study of the dynamics of nanostructure formation, which is influenced by a ring opening reaction by chemisorbed species on a silicon surface. We use density functional theory (DFT) calculations and scanning tunneling microscopy (STM) under ultrahigh vaccuum (UHV) conditions to investigate the reaction of cyclopropyl methyl ketone (CPMK, chosen because it is readily available in high purity and contains a cyclopropyl moiety) with isolated single dangling bonds on an otherwise H-terminated Si(100) surface. The goal of this work is to understand how the dynamics of line growth is altered by the ring opening reaction and thus help us gain an understanding of how radical clocks may be used to measure nanostructure growth rates. Results and Discussion. Theoretical Calculations. Groundwork DFT calculations7 indicate that CPMK is an ideal candidate for reaction with Si dangling bonds. Our calculations show that CPMK adds to the empty valence of a 15silicon-atom cluster and releases 0.53 eV of energy (A to C through B, Figure 1). The chemisorbed CPMK has a carboncentered radical site that can be quenched by abstraction of a hydrogen atom from a neighboring silicon dimer site (transition state represented in Figure 1D) to form a silyloxylcyclopropylethane (Figure 1E). This process is also computed to be exothermic (by an additional 0.53 eV) and has a classical barrier height (Ebarr) of ca. 0.6 eV, 0.35 eV lower than the analogous barriers in the styrene and vinylferrocene systems.3 These barriers are large because in order to reach the H-atom transfer transition state (TS), the SiSi-O and Si-Si-H bond angles must open substantially in order to allow the short-tethered Si-O-C• moiety (RO-C ) 1.37 Å) to reach the labile hydrogen on the edge of the next dimer row. The Si-H bond must also stretch to the transition state bond length of ca. 1.7 Å. An alternative and more likely pathway for quenching the carbon-centered radical (vide infra) is for the chemisorbed 358
Figure 1. Proposed mechanism for the reaction of CPMK with a dangling Si bond on the 2×1 H-Si(100) surface.
species to first undergo ring opening to form a siloxylbut3-enyl species. Measurements performed in organic solvent have shown that the R-trimethylsilyloxycyclopropylmethyl radical (an excellent model for our system) undergoes ring opening at a rate of 2.4 × 107 s-1 at 298 K.13 Thus, ring opening occurs on a time scale of ca. 40 ns. Our calculations indicate ring opening to be slightly exothermic (by < 0.05 eV), in agreement with experimental data on related species.14 The resulting terminal carbon-centered radical is then quenched by a subsequent hydrogen atom abstraction (releasing an additional ca. 0.5 eV). The steps involved in CPMK reaction with a silicon dangling bond are shown in Figure 1. The interesting consequence associated with the ring opening is the long reach of the butenyl-type radical. In principle, this radical can abstract a hydrogen atom from any site within its ≈4.7 Å long reach because there is effectively free rotation about the anchoring Si-O bond (site x in Figure 2). Three sites (a, d, and f in Figure 2) lie within this range, as measured by distances between surface hydrogens. Additional calculations15 were performed to predict the classical barrier heights (Ebarr) associated with abstracting a hydrogen atom from nearby sites. These data are summarized in Figure 2. Abstraction from site d, the site from which styrene2 and vinylferrocene3 normally abstract, has the lowest barrier, with Ebarr ) 0.1 eV. The barrier associated with abstraction from site f is also very low, with Ebarr ) 0.2 eV. It is interesting to point out that sites a and g have the same Ebarr ) 0.3 eV despite the large difference in their distance to x (3.5 vs 5.8 Å H-H distances, respectively). All of these low barrier heights are the result of the ability of butenyl radical to adopt conformations that are ideal for abstractions from the Nano Lett., Vol. 4, No. 2, 2004
Figure 2. Schematic of the top view of a 2×1 H-Si(100) surface. Filled and open circles represent the first and second layer of silicon atoms, respectively. Hydrogen atoms are not shown. Calculated classical barrier heights associated with hydrogen atom abstraction by the butenyl-type radical (chemisorbed to site x) from surface sites (in eV) are: a ) 0.3, b ) 0.4, c ) 2.2, d ) 0.1, e ) 1.7, f ) 0.2, g ) 0.3, h ) 2.4. Barrier heights associated with other sites are much higher that 2.4 eV.
different sites. Ebarr ) 0.4 eV for site b and other sites shown in Figure 2 have Ebarr > 1.6 eV. Quantum mechanical tunneling is expected to reduce the effective barrier heights associated with all sites. The consequences of the long reach of the butenyl-type radical is a substantial reduction of the barriers associated with H-atom abstraction (compared to the barriers in the cases of styrene and vinylferrocene) and an increase in the number of sites from which abstraction can occur. Nanostructure growth due to CPMK is therefore expected to occur readily and in a number of directions. Scanning Tunneling Microscopy.17 Figure 3 contains a sequence of STM images corresponding to increasing exposure of the 2×1 H-Si(100) surface to CPMK. Figure 3a shows a clean H-Si(100) surface with a low concentration of isolated silicon dangling bonds. The dangling bonds appear as bright protrusions, ca. 0.5 Å high. The dark features in the Figure are likely defects that are due to single and multiple Si atom vacancies. The dangling bonds appear to react with CPMK upon exposure of the surface to 300 L of the gas (Figure 3b). Exposure of the surface to 900 L of CPMK (Figure 3c) results in the formation of more extensive structures, as indicated by larger, light colored areas surrounding the dark spots which represent the positions of the original dangling bonds. The structures that are formed are evidently larger than a single chemisorbed CPMK molecule. Growth does not occur in one direction but appears to proceed both along a dimer row and across dimer rows. This is clear evidence that the cyclopropyl ring of CPMK does indeed open and that the butenyl-type radical is capable of abstracting an H-atom from any one of a number of surface sites. Nano Lett., Vol. 4, No. 2, 2004
Figure 3. Sequence of STM images (170 Å × 170Å -2.0 V, 0.1 nA) corresponding to increasing exposure of the surface to CPMK: (a) clean, H-terminated 2×1 Si(100) surface with isolated dangling bonds (the white features); (b) after 300 L exposure; and (c) after 900 L exposure. The black dots in (b) and (c) indicate the positions of the initial dangling bonds and show that these sites serve to nucleate the growth of CPMK islands.
Growth is self-limiting because continued exposure of the surface to CPMK does not result in an increase in the size of the formed nanostructures. Termination of the chain reaction can occur if the growth turns back on to itself. This type of termination was also observed to occur for the random-walk type chain reaction of styrene on the H-Si(111) surface.5 The possibility that the features are physisorbed CPMK monomers or polymers is rejected on the basis that further nanostructure growth does not occur with additional exposure of the surface to CPMK. Additional evidence is provided by the constancy of the nanostructures under repeated scanning with the STM tip. Conclusions. We have studied the ring opening and chainreaction-driven CPMK growth of nanostructures on the 2×1 H-Si(100) surface. DFT calculations indicate that CPMK adds to the dangling bond of a surface silicon in a reaction that is exothermic by 0.53 eV. Ring opening is predicted to occur readily, and the resulting butenyl-type radical can abstract a hydrogen atom from any one of a number of sites. The chain-reaction continues through the addition of another CPMK molecule to the newly created surface radical. STM experiments verify these dynamics by showing that the growth occurs both down and across dimer rows. The possibility that ring opening does not occur cannot be completely dismissed based on the growth observed in the STM images. However, our theoretical calculations and a previously determined experimental rate for ring opening strongly support the ring opening mechanism. Furthermore, our previous experimental work with various molecules having no capacity to ring-open has not revealed the extraordinary row-crossing growth described in this study, 359
pointing strongly to the long-reaching abstraction process that results from ring opening. Future experiments will focus on manipulating the radical stability of the chemisorbed radical clock precursor through the use of substituents in order to alter the rate of ring opening. By doing so, we hope to be able to measure the absolute rates of line growth.19 Such substitutions, together with the addition of shape constraints upon the ring opened radical, may lead to control over the direction of growth.
(9) (10) (11) (12) (13)
Acknowledgment. We are particularly indebted to K. U. Ingold and Danial D. M. Wayner for many insightful discussions and Douglas Moffatt for assistance with instrumentation. We also thank iCORE for support.
(14) (15)
References (1) Wolkow, R. A. Annu. ReV. Phys. Chem. 1999, 50, 413-41. (2) Lopinski, G. P.; Wayner, D. D. M.; Wolkow, R. A. Nature 2000, 406, 48-51. (3) Kruse, P.; Johnson, E. R.; DiLabio G. A.; Wolkow, R. A. Nano Lett. 2002, 2, 807-810. (4) Linford, M. R.; Fenter, P.; Eisenberger, P. M.; Chidsey, C. E. D. J. Am. Chem. Soc. 1995, 117, 3145-3155. (5) Cicero, R. L.; Chidsey, C. E. D.; Lopinski, G. P.; Wayner, D. D. M.; Wolkow, R. A. Langmuir 2002 18, 305-307. (6) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317-323. (7) All calculations in this work were performed using the Gaussian 98 suite of programs.8 The silicon cluster used in the calculations contained 15 Si atoms and two dimer rows. Silicon dangling bonds were capped with hydrogen atoms that were constrained to maintain their positions during geometry optimization calculations. Geometries were optimized at the B3LYP9,10/6-31G(d) level of theory. Singlepoint energies calculated for the optimized structures using B3P868,11/ 6-311G(2d,2p), which we have shown performs well for bond dissociation enthalpies.12 (8) Gaussian 98, Revision A.9, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgonery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D.
360
(16) (17)
(18) (19)
K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; 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.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A.; Gaussian, Inc.: Pittsburgh PA, 1998. Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. Perdew, J. P. Phys. ReV. B 1986, 33, 8822-8824. Johnson, E. R.; Clarkin, O. J.; DiLabio, G. A. J. Phys. Chem. A 2003, 107, 9953-9963. Nonhebel, D. C.; Suckling, C. J.; Walton, J. C. Tetrahedron Lett. 1982, 23, 4477-4480. Beckwith, A. L. J.; Bowry, V. W. J. Am. Chem. Soc. 1994, 116, 2710-2716. Barrier height calculations were performed on a cluster consisting of 55 silicon atoms, which included six dimer rows. The cluster was capped and constrained as described.7 Hydrogen atom abstraction transition state (TS) geometries were computed using the AM1 method.16 Since AM1 predicts Si-O bond lengths and Si-O-C bond angles that are in considerable error with respect to higher levels of theory (e.g., B3LYP/6-31G(d)), we used the butenyl (C5H8) radical for the calculations. The reach of this radical is close to that of the CPMK radical species. Single-point energies were calculated for the AM1 TS geometry using B3LYP/6-31G(d). Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 3902-3909. The STM experiments were carried out at room temperature in a ultrahigh vacuum 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 then H-terminated at 330 °C by exposure to atomic hydrogen created at a tungsten filament. Single dangling bonds were generated by hydrogen desorption with the STM tip by employing a sample bias of +3.5 V and currents below 40 nA.18 Typical imaging conditions were a sample bias of Vsb) -2.0 V with tunneling current It ) 40 pA. CPMK was degassed by performing several freeze-pump-thaw cycles. CPMK was introduced to the UHV chamber via a variable valve. Shen, T.-C.; Wang, C.; Abelin, G. C.; Tucker, J. R.; Lyding, J. W.; Avouris, Ph.; Walkup, R. E. Science 1995, 268, 1590-1592. We recognize that the rate of H-abstraction by the closed-ring species establishes a lower limit in the rate of ring opening for molecules used to measure rates of nanostructure formation.
NL035021G
Nano Lett., Vol. 4, No. 2, 2004
