1322
Langmuir 2005, 21, 1322-1327
Controlled Polymerization of Substituted Diacetylene Self-Organized Monolayers Confined in Molecule Corrals Shawn P. Sullivan, Albert Schnieders,† Samuel K. Mbugua,‡ and Thomas P. Beebe, Jr* Department of Chemistry & Biochemistry, and University of Delaware Surface Analysis Facility, University of Delaware, Newark, Delaware 19716 Received June 8, 2004. In Final Form: August 24, 2004 We have shown that STM-tip-induced chain polymerization of 10,12-tricosadiynoic acid (TCDA) in a self-organized monolayer at the liquid-solid interface of TCDA on highly oriented pyrolytic graphite is possible. The oligomers thus produced started at the point where a voltage pulse was applied between the STM tip and the sample during a short period when the feedback condition was momentarily suspended (as it is for scanning tunneling spectroscopy). Polymerization probabilities depended upon the length of the applied voltage pulse and were generally higher for longer pulse widths in the 10-ms to 100-µs time scales, approaching unit probability for the former and decreasing quickly to a few tens of percent for the latter. The polymerization could be confined to certain nanometer-sized areas by using “molecule corrals,” and polymerization appeared to be governed by topochemical constraints. Polymerization across domain boundaries, or over molecule corral edges, was never observed in over ∼150 observations. Due to the constant supply of nonpolymerized molecules from the covering solution, a dynamic exchange between molecules on the surface and in the solution was possible. This exchange occurred on a time scale that was comparable to the image acquisition time (∼101 s), and appeared to depend weakly upon the length of the desorbing oligomer. The desorption process was probably also influenced by interactions with the STM tip.
Introduction Nanometer-scale organic surface structures have, due to their size-dependent structural, electronic, physical and chemical properties, significant potential applications, e.g., in the field of molecular electronics or nanoelectronics. In recent years, these subjects have become major international research fields due to their vast potential in many technical areas. It is widely hoped that nanoelectronics will hold the key to solving many of today’s technological limitations, e.g., in further miniaturizing integrated circuits. However, most of the properties of nanometerscale organic surface structures are not fully understood, and therefore, their characterization and controlled production is of increasing importance. A widely used substrate for structural studies of organic monolayers is highly oriented pyrolytic graphite (HOPG). HOPG is a two-dimensional solid with strong covalent bonding in each layer and only weak van der Waals interactions between the layers. HOPG is easy to cleave, its surface is atomically flat, and it is relatively inert, even in air and in contact with fluids. By heating HOPG in air, nanometer-sized etch pits are formed around preexisting (i.e., naturally occurring) or intentionally created defects in the basal plane of the upper surface layer. This technique produces pits that are uniformly circular and mostly one monolayer deep.1,2 Using ion bombardment to introduce defects into the surface can easily allow control of the pit diameter, the pit density, * Author to whom correspondence should be addressed. Phone: 302-831-1888. Fax: 302-831-6335. E-mail:
[email protected]. † Present address: ION-TOF USA, Inc., Chestnut Ridge, NY 10977. ‡ Visiting Howard Hughes Medical Institute Summer Scholar from Lincoln University, Oxford, PA. (1) Chu, X.; Schmidt, L. D. Carbon 1991, 29, 1251-1255. (2) Patrick, D. L.; Cee, V. J.; Beebe, T. P., Jr. Science 1994, 265, 231-234.
the pit depth,3,4 and even the pit position and pit pattern on the surface.5 These etch pits can be used as so-called “molecule corrals” to confine domains of self-assembled monolayers to certain regions. Our group has used this technique for the study of ordering phenomena of selfassembled molecular monolayers by STM, among other uses.2,6 Etch pits in HOPG have also been used as templates for the formation of metal and semiconductor nanostructures.7-9 In recent years, it has been shown that scanning tunneling microscopy (STM) is, besides its wide use in the characterization of organic monolayers,10,11 also well suited for manipulation of these layers.6 Recently, Okawa and Aono used the tip of an STM to initiate linear chain polymerization of 10,12-pentacosadiynoic acid in a Langmuir-Blodgett layer (LB) that had been transferred to an HOPG surface.12-15 While scanning the tip over the surface, a voltage pulse was applied to the tip at predefined positions. Starting from these points, linear conjugated (3) Zhu, Y.-J.; McBride, J. D.; Hansen, T. A.; Beebe, T. P., Jr. J. Phys. Chem. B 2001, 105, 2010-2018. (4) Zhu, Y.-J.; Hansen, T. A.; Ammermann, S.; McBride, J. D.; Beebe, T. P., Jr. J. Phys. Chem. B 2001, 105, 7632-7638. (5) Zhu, Y.-J.; Schnieders, A.; Alexander, J. D.; Beebe, T. P., Jr. Langmuir 2002, 18, 5728-5733. (6) Stevens, F.; Buehner, D.; Beebe, T. P., Jr. J. Phys. Chem. B 1997, 101, 6491-6496. (7) Ho¨vel, H.; Becker, T.; Bettac, A.; Reihl, B.; Tschudy, M.; Williams, E. J. J. Appl. Phys. 1997, 81, 154-158. (8) Ho¨vel, H.; Becker, T.; Bettac, A.; Reihl, B.; Tschudy, M.; Williams, E. J. Appl. Surf. Sci. 1997, 115, 124-127. (9) McBride, J. D.; Van Tassell, B.; Jachmann, R. C.; Beebe, T. P., Jr. J. Phys. Chem. B 2001, 105, 3972-3980. (10) Claypool, C. L.; Faglioni, F.; Goddard, W. A., III.; Gray, H. G.; Lewis, N. S.; Marcus, R. A. J. Phys. Chem. B 1997, 101, 5978-5995. (11) Giancarlo, L. C.; Flynn, G. Annu. Rev. Phys. Chem. 1998, 49, 297-336. (12) Okawa, Y.; Aono, M. Nature 2001, 409, 683-684. (13) Okawa, Y.; Aono, M. J. Chem. Phys. 2001, 115, 2317-2322. (14) Okawa, Y.; Aono, M. Top. Catal. 2002, 19, 187-192. (15) Okawa, Y.; Aono, M. Surf. Sci. 2002, 514, 41-47.
10.1021/la048586g CCC: $30.25 © 2005 American Chemical Society Published on Web 12/04/2004
Controlled Polymerization of Self-Organized Monolayers
Figure 1. Structure of 10,12-tricosadiynoic acid (TCDA).
Langmuir, Vol. 21, No. 4, 2005 1323
induced chain polymerization to self-organized monolayers of TCDA at the liquid-solid interface of a solution of TCDA on HOPG. This provided the possibility to study the dynamic exchange of molecules in the solution with single molecules and oligomers in the self-organized monolayer, an option that was not possible in the LB films of Okawa, Aono, et al. We also used molecule corrals to confine the polymerization reaction to certain areas in a first attempt at controlling the length and direction of oligomer formation. We view this work as one approach for controlling the length, orientation, and connection points of surfacebound molecular “wires”. Experimental Section
Figure 2. Scheme of the topochemical polymerization of TCDA. Epitaxy of the alkyl “tails,” which are always parallel to one another within any given domain, places the -CtC- moieties in the proper orientation and at the proper distance for chain polymerization in the monolayer.
oligomer “nanowires” were formed.16 The polymerization stopped at domain boundaries and defects in the molecular layer. In addition, De Schryver et al.17 used both UV-light and STM-tip-induced polymerization to initiate the formation of polymerized diacetylenes. Depending on the number of diacetylene functional groups in one given molecule, they were able to create not only 1-D nanostructures but also 2-D nanostructures. Their work, like Okawa and Aono’s, explored this using a LB layer transferred on HOPG, as well as a drop cast method. Substituted diacetylene compounds such as 10,12tricosadiynoic acid, TCDA, (Figure 1) or 10,12-pentacosadiynoic acid are characterized by the interior -CtC-CtC- group. These molecules can form 3-D and 2-D crystals.18,19 Polymerization can be induced by irradiation with UV light. Supplied from a solution, substituted diacetylenes form physisorbed self-ordered monolayers on HOPG after a sufficient time and under the correct conditions. These layers at the liquid-solid interface have been studied by STM before.20 Diacetylene compounds are known to undergo topochemical polymerization in the bulk,21 in two-dimensional monolayers, and along one-dimensional lines,18,22,23 as seen in Figure 2. On a surface, the molecules can polymerize with retention of the packing structure in the original monolayer lattice. The polymeric backbone formed is composed of alternating conjugated double and triple bonds. In this paper, we present our work using a technique similar to that of Okawa and Aono using a related diacetylene molecule. But rather than using an LBtransferred molecular layer, we extended the use of tip(16) Akai-Kasaya, M.; Shimizu, K.; Watanabe, Y.; Saito, A.; Aono, M.; Kuwahara, Y. Phys. Rev. Lett. 2003, 91, 255501-255504. (17) Miura, A.; De Feyter, S.; Abdel-Mottaleb, M. M. S.; Gesquiere, A.; Grim, P. C. M.; Moessner, G.; Sieffert, M.; Klapper, M.; Muellen, K.; De Schryver, F. C. Langmuir 2003, 19, 6474-6482. (18) Wegner, G. Makromol. Chem. 1972, 154, 35-48. (19) Tieke, B.; Lieser, G.; Wegner, G. J. Polym. Sci. Polym. Chem. 1979, 17, 1631-1644. (20) Rabe, J. P.; Buchholz, S.; Askadskaya, L. Synth. Met. 1993, 54, 339-349. (21) Enkelmann, V. Adv. Polym. Sci. 1984, 63, 91-136. (22) Ozaki, H.; Funaki, T.; Mazaki, Y.; Masuda, S.; Harada, Y. J. Am. Chem. Soc. 1995, 117, 5596-5597. (23) Grim, P. C. M.; De Feyter, S.; Gesquiere, A.; Vanoppen, P.; Rucker, M.; Valiyaveetil, S.; Moessner, G.; Mullen, K.; De Schryver, F. C. Angew. Chem., Int. Ed. Engl. 1997, 36, 2601-2603.
HOPG samples, grade ZYB, were generously supplied by Dr. Arthur W. Moore of Union Carbide/Advanced Ceramics. The samples were cleaved with adhesive tape in ambient conditions. When it was desired to control the pit density, immediately after cleaving, samples were bombarded with 500-eV Cs+ ions in the ultrahigh vacuum analysis chamber of a time-of-flight secondary ion mass spectrometer (TOF-SIMS IV, ION-TOF, Mu¨nster, Germany). The ion beam was rastered over a 500 µm × 500 µm area of the sample surface to ensure a uniform ion exposure. The applied ion dose density was extremely low, 5 × 108 cm-2, corresponding to the equivalent of ∼3 × 10-7 monolayer. The ion-bombarded HOPG samples were subsequently thermally oxidized by heating the samples in a Lindberg Hevi-Duty tube furnace. The samples were placed in the center of a 60-cm-long horizontal quartz tube with 2.5-cm diameter. The ends of the tube were left open, and no effort was made to cause or prevent airflow during thermal oxidation of the HOPG samples. Heating at 650 °C for 20 min resulted in pits of about 60-nm diameter. The average pit density was about 2 × 108 cm-2, in accordance with our previously reported observation that the pit yield under these conditions (HOPG with defects artificially created by 500eV Cs+ ion bombardment) is about 0.5 pits per ion impact.3 10,12-tricosadiynoic acid, TCDA, was purchased from Fluka (Buchs, Switzerland). It was dissolved in octylbenzene (also purchased from Fluka) with a typical concentration of 30 mg mL-1. A volume of approximately 2 µL of the solution was pipetted onto the pitted HOPG surface, which it immediately wet, creating a thin film. Immediately after this preparation, STM imaging was started. In this study, we used a custom-built STM that had been specially designed for high-resolution imaging of molecular layers. It has been described elsewhere.24 Typical scanning conditions used were bias voltage, U ) -1 V; tunneling current, I ) 130 pA resulting in a low tip-surface interaction (tunnel gap impedance ≈ 10 GΩ). All images were acquired in constantheight mode using tips mechanically cut from wires (80% Pt and 20% Ir, Goodfellow, Berwyn, PA). Using control software, we were able to apply voltage pulses of predefined shape between the tip and sample at predefined positions while the feedback condition was momentarily interrupted, as is commonly done for scanning tunneling spectroscopy. For these studies, an amplitude of U ) -6 V (sample negative) and a pulse width, τ, ranging from 16 ms to 200 µs were chosen. It was not our intention in the present work to systematically explore parameter space for these parameters.
Results and Discussion A. Self-Ordered Monolayers of TCDA on HOPG Prior to Polymerization. Figure 3 shows a typical highresolution constant-height STM image of a self-ordered monolayer of TCDA physisorbed onto HOPG from a solution in octylbenzene. For clarity, models of five molecules, drawn to scale, have been superimposed on the STM image. Clearly visible are rows of parallel-packed TCDA molecules forming long lamellae with a bright stripe in the middle. The bright bands in Figure 3 correspond (24) Zeglinski, D. M.; Ogletree, D. F.; Beebe, T. P., Jr.; Hwang, R. Q.; Somorjai, G. A.; Salmeron, M. B. Rev. Sci. Instr. 1990, 61, 3769-3774. (25) Information provided by Fluka, Buchs, Switzerland.
1324
Langmuir, Vol. 21, No. 4, 2005
Sullivan et al.
Figure 3. High-resolution STM image of a self-ordered monolayer of TCDA physisorbed from a solution in octylbenzene onto HOPG. A molecular model of five molecules is overlayed on the correct relative scale. See text for a detailed discussion of the unit cell dimensions and symmetry. The image contrast is formed by collecting the tunneling current at each pixel in a constant-height scanning mode (i.e., a “current image”). Field of view: 11 × 11 nm2.
to the acetylene moieties that image with a high tunneling current due to the unsaturated C-C bonds.10 As is almost always the case in the literature, the alkyl chains of the molecules are packed commensurate with the graphite lattice and play a dominant role in determining the monolayer’s structure and especially its epitaxy. The unit cell consists of one molecule and measures Rˆ ) 2.90 ( 0.12 nm (along the molecular long axis), and βˆ ) 0.475 ( 0.05 nm (corresponding to two times the repeat unit of the HOPG lattice), with an angle θ ) 80° ( 5°. Reported errors are one standard deviation based on dozens of unit cell measurements. The 2-D TCDA overlayer is described by the symmetry matrix Μ h given in eq 1,
( ) ( )(
) (
aˆ m m 5.89aˆ + 0.97aˆ Rˆ ) aˆ 1 m11m12 ) 0.58aˆ 1 + 1.11aˆ 2 βˆ 2 21 22 1 2
)
(1)
where aˆ 1 and aˆ 2 are unit cell vectors of the HOPG substrate lattice, and Rˆ and βˆ are the unit cell vectors of the TCDA monolayer. Since the hexagonal lattice of HOPG is not a primitive lattice, we chose to define the substrate unit cell vectors aˆ 1 and aˆ 2 with a length of 0.492 nm (the width of two adjacent hexagons), separated by an angle of 120°. The ordered overlayer is also 2-D chiral, with two energetically degenerate mirror-image monolayer structures, both of which were observed with approximately equal frequency. Okawa and Aono, working with transferred LB monolayer films,12-15 observed an arrangement of the lamellae in which the COOH end groups of one lamella were adjacent to those of a neighboring lamella; we did not observe a pattern like this. This is due to the water-free preparation conditions used in the present studies. The formation of the LB layer in Okawa and Aono’s experiments on a water surface probably allowed water molecules to participate in hydrogen bonding between neighboring carboxylic acid groups, while no such interactions are present in our studies. By using HOPG with molecule corrals as the substrate, we observed the formation of self-ordered monolayers of TCDA inside some corrals (Figure 4A). The layers inside corrals have row directions, in most cases, of the same orientation as rows on the surrounding terrace, although relative orientations of (120° are also possible due to the HOPG symmetry and were occasionally observed. As Figure 4B indicates, the lamellae are interrupted at the edge of the corral. The probability of observing ordered rows within a given corral seemed to depend on corral size. This effect is presently under study and would be
Figure 4. STM current image of self-ordered monolayers of TCDA on HOPG with molecule corrals. Note that in A, one of the three corrals, self-assembly had taken place at the time of image acquisition, whereas it had not taken place in the other two corrals. Note also that the row direction of lamellae in the corral is parallel to that on the surrounding terrace. Field of view: A, 150 × 150 nm2; B, 50 × 50 nm2.
consistent with previous observations for different molecules in which the rate-limiting nucleation step was proportional to corral area.2 A preferred orientation of the TCDA monolayers on HOPG was observed, within the limits of constraints imposed by epitaxy and the HOPG symmetry axes. The lamellae tended to be oriented as close as possible to perpendicular to the fast-scan direction, i.e., the molecular long-axis was parallel to the fast-scan direction. In all images presented here, the fast-scan direction is horizontal. This observation raised the question of whether the STM was “blind” to ordered monolayers in which the lamellae were oriented parallel to the fast-scan direction. The observation might also suggest that there is an interaction between the scanning tip and the molecules on the surface that affects the self-assembly process. To address this question, a series of consecutive STM images of a TCDA monolayer on HOPG was acquired (not shown here). The bias voltage was U ) -870 mV and the current set-point was I ) 170 pA, corresponding to a relatively weak interaction between tip and surface (tunnel gap impedance ≈ 5 GΩ). In the first scan, the monolayer was not completely formed. There were several domains of differently oriented layers. After one additional scan (approximately 40 s), nearly the whole scanned area was covered by a uniform layer. Even over step edges, the molecular orientation was the same. However, after abruptly changing the fast-scan direction by 90°, the layer started to realign on a 10-min time scale during continuous scanning. Eventually, several new domains could be observed. This experiment indicates that the interaction between the TCDA molecules and the HOPG surface is weak and comparable to the magnitude of the tip-molecule interaction, even at gigaohm tunnel gaps. Therefore, the molecules preferentially align with their long axes as close to the fast-scan direction as epitaxy will allow. However,
Controlled Polymerization of Self-Organized Monolayers
Langmuir, Vol. 21, No. 4, 2005 1325
Figure 6. STM current images from a series of more than 20 scans showing tip-induced oligomers starting and ending at edges of molecule corrals. In the first image, the positions at which voltage pulses (U ) -6 V, τ ) 200 µs) were applied during scans are indicated by circles.
Figure 5. STM current images showing a successful initiation of a chain polymerization of TCDA. The images were consecutively acquired with about 40 s intervals. N ) 1: original layer; during N ) 2, voltage pulses (U ) -6 V, τ ) 200 µs) were applied at the circled positions; scans N ) 4-14 were collected immediately following the voltage pulse used to initiate the polymerization of TCDA. The inset in scan N ) 11 was collected in a right-to-left direction, whereas in the larger N ) 11 image, scans were simultaneously collected in a left-to-right direction. A second shorter oligomer, parallel to, and immediately adjacent to the larger oligomer, is evident in scan N ) 11. By scan N ) 12, this shorter oligomer had desorbed, and by scan N ) 13, the larger oligomer had begun to break and desorb.
in the absence of scanning, or under extremely limited continuous scanning at very high tunnel gap impedances, there will be a statistical distribution of the three equivalent lamellae row directions, reflecting the 3-fold HOPG symmetry and its effect on molecular epitaxy. This statistical distribution in the absence of scanning was proved for other molecular monolayers in previous work from our group.2,6 B. Tip-Induced Polymerization of TCDA Monolayers on HOPG. By applying voltage pulses between the tip and surface while the feedback condition was momentarily interrupted, chain polymerization of TCDA molecules along a single lamella could be induced. The events described in Figure 5 are representative of ∼150 observations of tip-induced polymerizations. Some prior work on tip-induced voltage pulses and the defects that resulted were done “in air,” rather than under a liquid, as was the case here. The magnitude of voltage pulses used here rarely caused noticeable substrate damage. In the relatively rare instances in which damage was obvious, the damage occurred to the tip. Such damage often made imaging impossible with good resolution and contrast. Figure 5 shows several STM images acquired during initiation of a polymerization reaction of TCDA after applying voltage pulses (U ) -6 V, τ ) 200 µs) at the positions marked by circles in scan N ) 2. The scans were consecutively acquired with a 40-s time interval, and not all images in the series are shown here. Images presented here were collected in “constant-height” mode to improve image quality by reducing 1/f noise. Because of the shadowing that always results from constant-height imaging parameters, a left-to-right image tends to hide features immediately to the right of a given feature.
Likewise, a right-to-left image (as for the inset of N ) 11 image in Figure 5) tends to hide features immediately to the left of a given feature. In scan N ) 4, three oligomers can be seen along different lamellae indicated by bright lines. After scan N ) 5, only one oligomer line remained on the surface, and the disturbance created by the voltage pulse was still causing a slight degradation of image quality. The inset of scan N ) 11 is a right-to-left image (it was recorded line-byline simultaneously with the larger left-to-right image). Although it might appear in image N ) 11 as if there is a cut in the oligomer, the inset helps to eliminate this interpretation. The larger left-to-right image clearly shows a second shorter oligomer parallel to the larger one. By image N ) 12 in Figure 5, the shorter oligomer had already desorbed, leaving the longer oligomer in its original position. After scan N ) 12, the oligomer broke at several points, desorbed, and was eventually replaced by nonpolymerized TCDA molecules from the covering solution. This dynamical exchange was possible due to the quasiinfinite supply of nonpolymerized molecules in the covering solution. It occurred on a time scale with an approximate upper limit of ∼101 s, and a lower limit that we could not measure. This exchange was probably affected by interactions with the STM tip since, in all experiments presented here, continuous scanning was used. We are currently investigating the role of the covering solution in the polymerization process. For example, it was observed that an ordered monolayer was still present after rinsing the covering solution from the HOPG substrate, although limited attempts to induce polymerization on such a monolayer have been unsuccessful in the absence of the covering solution. The four images in Figure 6 are from a series (originally consisting of more than 20 images) of images of the same molecule corral. In every other scan, voltage pulses (U ) -6 V, τ ) 200 µs) were applied at the positions indicated by circles in Figure 6A. This experiment demonstrated that the production and the desorption of oligomer chains was relatively reproducible. We further found that the probability of producing an oligomer was dependent on the width of the applied pulse. Pulse widths of τ ) 200 µs and amplitude U ) -6.0 V resulted in very low polymerization probabilities (statistics too low to report with confidence), while a pulse width of τ ) 16 ms and amplitude U ) -6.0 V produced 90 oligomers for 86 pulses, corresponding to a probability of polymerization of ∼1.1 oligomers/pulse. The value greater than 1.0 was due to
1326
Langmuir, Vol. 21, No. 4, 2005
the fact that in some cases one pulse initiated polymerization of more than one oligomer, although the most likely outcome was 1 oligomer per pulse. While this manuscript was in revision, we performed additional experiments using a pulse width of τ ) 8 ms and amplitude U ) -6.0 V, resulting in 72 oligomers formed out of 124 pulses applied, corresponding to a probability of polymerization of ∼0.58 oligomer per pulse. Experiments using pulse widths ranging from milliseconds to microseconds are currently underway to further explore how the pulse width influences the polymerization event. In general, it can be reported with confidence that a voltage pulse applied for a longer period of time is more likely to result in a polymerization event. It is remarkable that no oligomer extended over the edge of the molecule corrals. This was true for polymerized structures inside corrals (eg., Figure 6A) as well as for polymerized structures on terraces that extended up to a corral edge (eg., Figure 6B-D). This provides strong evidence that the polymerization propagated via reaction between direct neighbors of TCDA molecules following the polymerization induced by a voltage pulse, as suggested by Okawa and Aono.12-15 Since the lamellae did not extend over the step edge (see also Figure 4), the required topochemistry was disrupted at such a site. Together with the ability to control the size, pattern, and position of molecule corrals,5 this provided us with the means to control the length and direction of molecular oligomers that might someday be used to connect nanostructures formed inside molecule corrals.9 It should be noted that, like almost all STM-related work, success predominantly depended on the quality of the tip. Some tips proved to work very well (yield ≈ 1), whereas with other tips, the yield was zero. We therefore restricted the statistical evaluations of the data, and probabilities reported here, to those obtained with a “good” tip. Another problem occurred especially when using highconcentration solutions of TCDA. In that case, larger chunks of material (with diameters up to 20 nm) were often deposited onto the surface at the positions of the voltage pulses. The composition of these chunks was not clear. It might be that these chunks were surfacepolymerized oligomers induced by the voltage pulse, or bulk-polymerized material initially in solution and later deposited on the surface. It is known that there is always some portion of TCDA polymerized in any given solution of TCDA, as evidenced by a colored solution.21,24 C. Surface Residence Lifetime of Oligomers. The lifetime of an oligomer formed on the surface might be directly related to the length of the oligomer. This hypothesis is based on the argument that more points of interaction (i.e., bonds) will result in a higher desorption activation energy. Therefore, one could deduce that a longer chain would require more time to successfully desorb and be exchanged with the quasi-infinite supply of monomers surrounding it. After many observations, using various scanning conditions, the length of the oligomer seemed to have only a weak influence on its surface lifetime, as seen in Figure 7. Careful analysis of the lifetime of each oligomer, based on scan speed and tunnel impedance, indicated that some other factors are affecting the apparent lifetime of the oligomer on the surface. It is believed that tip-sample interactions may be another important factor in addition to oligomer length. Using tunneling gaps ranging from 5 to 10 GΩ, it was observed that some of the longest oligomers remained on the surface for a short time (approximately 2 min), whereas some small oligomers remained for more than 10 min, contributing to the scatter observed in Figure 7. By varying
Sullivan et al.
Figure 7. Oligomer length vs surface residence lifetime for 111 oligomers polymerized on the HOPG substrate. In the text, it was concluded that the surface residence lifetime depended weakly on oligomer length. Oligomers that extended beyond the image area were excluded from counting statistics. Oligomers desorbing from the surface into the covering solution were immediately replaced by monomers that adopted the templated structure of the surrounding self-assembled monolayer.
the bias voltage at a constant current of I ) 100 pA, the average surface residence lifetimes for oligomers varied systematically. Whereas an oligomer remained on the surface for several scans at a bias voltage of U ) -1000 mV, oligomers desorbed within one scan when lowering the bias voltage to about U ) -800 mV. Additional experiments are currently being carried out in an attempt to quantify the tip-sample interactions, using a statistically significant number of observations. This could potentially allow for the variation of oligomer lifetime on the surface, for example, as a means to control a localized chemical reaction. IV. Conclusions We have shown that STM-tip-induced chain polymerization of 10,12-tricosadiynoic acid in a self-organized monolayer at the liquid-solid interface of a solution of TCDA on HOPG is possible. The resulting oligomers started at the point where a voltage pulse was applied between the STM tip and sample during a scan. Polymerization probabilities depended upon the length of the applied voltage pulse and were generally higher for longer pulse widths in the 10-ms (unit probability) to 100-µs time scales (∼10%). The polymerization could be confined to certain areas by using molecule corrals. Due to the constant supply of nonpolymerized molecules from the covering solution, a dynamic exchange of molecules on the surface and in the solution was possible. This exchange occurred on a time scale of ∼101 s, depended weakly on the length of the desorbing oligomer, and was probably influenced by interactions with the STM tip. The technique of tip-induced chain polymerization of diacetylene compounds in a monolayer on HOPG with molecule corrals can be used for engineering networks of “nanowires” on surfaces. In recent studies,5,9 our group was successful at growing Au and Si nanostructures on HOPG by using molecule corrals as templates. We hope to connect these nanostructures by nanowires of polydiacetylene. This will provide us with natural connection points to larger devices to measure electronic properties on the oligomer chains and possibly provide a means to ultimately develop new structures on the nanometer scale.
Controlled Polymerization of Self-Organized Monolayers
Acknowledgment. A.S. was supported by a research fellowship of the Deutsche Forschungsgemeinschaft (DFG) (SCHN 675/1-1). Financial support from the National Science Foundation (CHE-0415979, CHE-9814477, DMR-9724307) is gratefully acknowledged. S.P.S. was partially supported by the NSF-IGERT program at UD (DGE-0221651). S.K.M. was supported by the Howard
Langmuir, Vol. 21, No. 4, 2005 1327
Hughes Medical Institute through a summer visiting scholar program at the University of Delaware. The authors would like to thank Dr. Arthur Moore from Advanced Ceramics Corporation for his generous provision of HOPG samples. LA048586G
