Molecular Calipers Control Atomic Separation at a Metal Surface

Sep 9, 2011 - Molecular Calipers Control Atomic Separation at a Metal Surface. Lydie Leung,. †. Tingbin Lim,. †. John C. Polanyi,*. ,† and Werne...
0 downloads 0 Views 3MB Size
Download PDF
LETTER pubs.acs.org/NanoLett

Molecular Calipers Control Atomic Separation at a Metal Surface Lydie Leung,† Tingbin Lim,† John C. Polanyi,*,† and Werner A. Hofer‡ †

Lash Miller Chemical Laboratories, Department of Chemistry and Institute of Optical Science, University of Toronto, 80 St. George Street, Ontario, M5S 3H6, Canada ‡ Surface Science Research Centre, The University of Liverpool, Liverpool, L69 3BX, U.K.

bS Supporting Information ABSTRACT: If a molecule controls the length of some other moiety, it can be termed a “molecular caliper”. Here we image individual molecular calipers of this type by scanning tunneling microscopy. These consist of linear polymers of p-diiodobenzene, (pDIB)n, of varying length, 0.7 2.9 nm, physisorbed on Cu(110) at 4.6 K. Through electron-induced reaction these chemically imprint their terminal I-atoms on the copper, 0.7 nm further apart than their initial separations. The physisorbed monomer or polymer, therefore, constitutes a molecular-caliper with variable terminal I..I separation. The localized nature of the I-atom reaction at the copper surface relative to the parent molecule, constitutes a novel finding reported here. It ensures that the separation of the I-atoms in the physisorbed molecular caliper correlates with their subsequent separation when chemisorbed at the surface. KEYWORDS: STM, electron-induced reaction, halogenation, metal surface

S

ingle molecules have been proposed as rulers for measuring molecular length, particularly for biomolecules.1 3 If the ruler were actively engaged in controlling the molecular length of some other specie, it could be termed a “molecular caliper”. This designation has been used to describe the function of yeast enzymes (“Elop”) thought to control molecular chain-length in certain fatty acids.4 The term “caliper” describes a measuring instrument, or, as here, a marking device “for scribing lines at specified distances”.31 The calipers demonstrated here are comprised of linear polymers of p-diiodobenzene, (pDIB)n, of varying length, physisorbed on cooled Cu(110). As a consequence of electron-induced reaction these chemically imprint their terminal I-atoms at controlled separations on the copper surface. The physisorbed monomer or linear polymer of variable length, therefore, constitutes the molecular caliper. Structurally a caliper must not distort prior to reaction; functionally the chemisorbed reaction product formed at the surface must be localized in the vicinity of its former position in the caliper. Neither condition is assured since physisorbed reagent can isomerize, and reaction products migrate. These conditions for functioning of a molecular caliper are, however, shown to be met in the present instance. Structural rigidity of the reagent caliper was ensured by the stability of benzene rings π-bonded flat on metal surfaces at low temperatures.5 9 The localized nature of the I-atom reaction at the metal surface, constitutes a significant finding, anticipated by Maksymovych and Yates who showed a preponderance of retention of conformation in the electron-induced dissociation of CH3SSCH3 to give CH3S pairs at Au(111).10 Localized atomic reaction (LAR) has been reported extensively in the halogenation of silicon.11,12 Localized reaction ensures that the separation of the I-atoms in the molecular caliper correlates with their subsequent separation at the surface. r 2011 American Chemical Society

Earlier scanning tunneling microscopy (STM) studies of surface halogenation by dihalobenzenes gave the pattern of photoinduced11 and thermal13 15 reaction at a semiconductor surface, Si(111)-7  7. The three principal findings were that the halogen atom recoiled along the C X bond direction, reaction was localized to the vicinity of the original position of the X-atom in the adsorbed reagent, and chemisorbed pairs of X were found at a separation exceeding by a few angstroms the previous separation in the physisorbed reagent. A simple model16 gave values of 0.3 0.4 nm for the increase in pair separation from the adsorbate to the final chemisorbed state. The thermal dissociation of pDIB on Cu(111) has been studied by McCarty and Weiss17 who found pairs of I-atoms 0.4 nm apart, markedly less than the initial separation of 0.72 nm in the undissociated molecule. The I-atoms exhibited mobility on room temperature copper, which could therefore account for the observed short pair separation. The reaction dynamics reported here, involving an increase in I..I separation in reaction, resemble those reported above for dihalobenzenes on silicon. In the present instance, however, the reaction was electron-induced at a metallic (rather than a semiconductor) surface, and the terminal halogen atom separation in the physisorbed reagent was systematically varied over the range 0.69 2.89 nm by collinear self-assembly. The end-to-end I..I separation in the reagent will be shown to correlate closely with the subsequent I-atom separation at the surface. The reagent acted, therefore, as a molecular caliper used to mark the surface with chemisorbed atoms at a set distance apart. Received: May 19, 2011 Revised: September 2, 2011 Published: September 09, 2011 4113

dx.doi.org/10.1021/nl2023788 | Nano Lett. 2011, 11, 4113–4117

Nano Letters

LETTER

Figure 1. STM images obtained before and after electron-induced reaction. (A) From a monomer (Vs = +1.0 V, I = 0.5 nA), (B) from a staggered dimer (Vs = +1.0 V, I = 0.6 nA), (C) from a linear dimer (Vs = +1.0 V, I = 0.6 nA) and (D) from a linear trimer (Vs = +1.0 V, I = 0.6 nA). Panel 1 gives the STM images of the initial state (I.S.). Panel 2 shows the simulated image overlaid by the computed adsorption geometry (top view). Sites on the copper substrate are indicated in panel 2 of (A): long-bridge (L), short-bridge (S) and 4-fold hollow (H). Panels 3 show side views of the calculated adsorption geometries. Panels 4 show the F.S. imaged by STM. The number of cases is indicated as N in each panel. From the maxima in height profiles, we obtained the positions of the I-atoms (black spots) at either side of the organic residue. In panels 4, the distances in nanometers are the F.S. separations between the I-atoms. On average the I-atoms were displaced 0.36 nm from their initial positions. The diagrammatic calipers indicate the terminal I..I distances in nm in both the I.S. and F.S. 4114

dx.doi.org/10.1021/nl2023788 |Nano Lett. 2011, 11, 4113–4117

Nano Letters The caliper action was obtained by electron-induced reaction of the terminal I-atoms of aromatic di-iodobenzene molecules having four different end-to-end I..I spacings. The (pDIB)n molecules were composed of a sequence of benzene rings that lay flat on a cooled metal surface due to pi-bonding (see below). In order to circumvent the need to distill large molecules into vacuum, we assembled them in situ on the copper surface following Yates and co-workers who reported 1D self-assembly of monoiodobenzene dimers on Cu(110).18 Experiments were performed in a low-temperature scanning tunneling microscope cooled to 4.6 K. Dissociation of the stable monomer, para-diiodobenzene (pDIB), was electron-induced by a tip-pulse or scanning (see Supporting Information for experimental details). Both modes of electron-induced reaction gave the same mean separation of I-atom product for pDIB monomer. Only the scanning mode was employed for the dimers and trimer. Electroninduced dissociation of the C I bond was previously reported with a similar positive surface-bias for pDIB19 and for monoiodobenzene (MIB)18,20 adsorbed on a copper surface. In the case of MIB the I-atom product formed along the C I bond direction, but recoil distances were not reported. The pDIB monomer was observed in the present study (Figure 1A, initial state, (I.S.), panel 1) as an asymmetric bright protrusion with its I..I axis along the [001] direction. The positions of the two I-atoms within the pDIB image spaced 0.69 nm apart (panel 1 of Figure 1A) were computed by densityfunctional theory (DFT) calculation of the pDIB adsorbate on Cu(110). The computed adsorbate geometry is shown in Figure 1A, panels 2 and 3, with a simulated STM image shown in panel 2. The computed alignment of the I..I axis was along [001], which is in agreement with experiment. The most stable calculated geometry was with the molecule lying flat on the copper surface with the phenyl ring located over one short-bridge site, and the two I-atoms bent toward the substrate over shortbridge sites at either side (Figure 1A, panels 2, 3). In addition to monomers, staggered dimers, linear dimers, and linear trimers were observed (initial states, panel 1, of Figure 1B D). The linear dimers and trimers were composed of evenly spaced physisorbed pDIB molecules, aligned along the [001] direction. (These differ from the lines of phenylene previously found to arise from room temperature reaction of pDIB at Cu(111)21). Computed adsorption geometries, are shown in Figure 1B D (panels 2, top view, and panels 3, side view)22 25 (see Supporting Information for details). The computed heats of adsorption were 0.16, 1.03, 1.20, and 3.06 eV for the monomer, staggered dimer, linear dimer, and trimer, respectively. The increase in the adsorption energies resulted from the interaction between the polymer units. In an earlier report of polymerization of MIB dimers on Cu(110), it was surmised that the I-atoms were directed at one another.18 Head-to-head halogen-atom configurations in alkyl halides have also been proposed by Flynn and co-workers on graphite.26 Ab initio calculations reported here confirm the headto-head structure for the linear dimer and also trimer. The few “staggered dimers” observed can be presumed, based on their low yield, to have weaker total binding. Calculations by DFT for the staggered dimer gave the simulated image and adsorption geometry shown in Figure 1B, panel 2 and panel 3). For the staggered dimer the central I-atoms can be seen to be no longer head-to-head; the H-atoms of the benzene ring provide weak I.. H C hydrogen-bonding. A comparable staggered arrangement

LETTER

Figure 2. (A) Distribution of the end I-atom separations after the electron-induced reaction (bin size: 1.8 Å corresponding to half a unit cell). The designations M, SD, LD and T correspond to monomer, staggered dimer, linear dimer, and trimer. (B) Correlation between reagent I..I separation (I.S.) and product I..I separation (F.S.). The I..I separation for the physisorbed adsorbate increases linearly, by a constant 0.72 nm, in forming the chemisorbed I..I imprints. Error bars represent the standard deviation of the mean (falling sometimes within the points).

was observed by M€uller et al. for halohexanes on graphite26 and by Lee et al. for 2,6-dimethylpyridine dimers on Cu(110).27 Electron-induced reaction of the terminal I-atoms yielded the four final state images (F.S.) in panel 4 of Figure 1A D. For the monomer, reaction was studied over a range of positive surface biases. In the region of 1.3 1.5 V only single C I bond breaking was observed for the monomer. For biases above 1.6 V, two C I bonds of the monomer reacted with the copper surface. Bond breaking will be shown below to be a single-electron process. In some cases, the breaking of the two C I bonds of the monomer could be resolved in time. For both dimers and for the trimer, the successive bond-breaking events were invariably resolved in time. The products of the electron-induced reaction (Figure 1A, panel 4) were two round protrusions (height: 0.06 ( 0.02 nm) identified as I-atoms28 on either side of a smaller protrusion (0.02 ( 0.02 nm) ascribed to the phenyl residue.29 The I..Iimprints for all four cases were collinear with the principal axis of the polymer, consistent with a high degree of localization of the products. 4115

dx.doi.org/10.1021/nl2023788 |Nano Lett. 2011, 11, 4113–4117

Nano Letters

LETTER

Figure 4. Electron transfer from the linear dimer to Cu(110). (A) The electron density loss (isocontour 0.004 e /Å3), shown in red, is mainly located at the two pDIB molecules. The loss, in addition to that from the benzene rings, forms a torus in the plane perpendicular to the Cu surface, wrapped around the I-atoms. (B) The electron density gain (isocontour +0.004 e /Å3) is displayed in blue. There is electron density gain at the Cu-atoms underneath the I-atoms and the Cu-atoms underneath the benzene rings.

Figure 3. Electron-induced dissociation of a C I bond in the monomer at a bias of +1.4 V. (A) Current versus time plot showing one downward break, corresponding to one C I bond-breaking event. (B) Kinetics of monomer dissociation, where n (1.3 ( 0.1) is the slope of the least-squares fit to the reaction rate as a function of tunneling current (log log coordinates). Error bars give the standard deviation. (C) The calculated pDOS over p-diiodobenzene on Cu(110). The total DOS of the molecule is indicated by the solid curve. The pDOS p-orbital component of the I-atom is shown by short dashes, and both the C-atom and their adjacent H-atom pDOS by long dashes.

For the dimer and trimer reagents the bias voltage was kept in the neighborhood of the threshold of 1.3 V, so that the scanning STM tip induced bond-breaking sequentially at the terminal I-atoms. For linear dimer and trimer, a single “central” I-atom (never two), responsible for holding the polymer together also reacted in about one-third of the scans. The reduced reactivity of the central I-atoms of the linear dimer and trimer was ascribed to their caging in approximately head-to-head configuration. In the case of the staggered dimer, the caging of the central I-atoms, which were no longer opposed, was less effective; reaction of only the terminal I-atoms (Figure 1B, F.S. panel 4) was achieved by reducing the scanning voltage to 1.15 V in midscan.

The I..I distances shown in Figure 1B D for the products of pDIB polymer reaction were for cases in which the scan caused exclusively reaction of the terminal I-atoms. The observed displacement of each terminal I-atom from its initial position in the monomer, staggered dimer, linear dimer and trimer to its final position at the surface was a mean of 0.36 nm. Along the C I direction of motion the three populated sites at the copper surface were a 4-fold hollow (H), a short-bridge (S), and a further 4-fold hollow (H). Figure 2A shows a histogram of the measured separations between the chemisorbed I-atoms, and Figure 2B shows the linear relationship between physisorbed reagent terminal I..I separation, and the chemisorbed product I..I separation. The increase in separation on reaction, as stated above was a constant 0.72 nm. The close correlation between initial and final atomic separation constitutes molecular-caliper action dependent on the localization of the I-atom reaction. Figure 3A gives a typical current versus time curve recorded with the STM tip over a monomeric pDIB at a bias of 1.4 V. The break in the curve corresponds to the time at which a single C I bond broke, forming I Cu at the surface. The yield at threshold was 9  10 11 per electron rising to ∼10 9 above threshold. Figure 3B gives the dependence of reaction rate on current at a constant bias of 1.4 V; the slope of the line was found to be 1.3 ( 0.1, indicative of predominantly single-electron reaction for the breaking of a single C I bond. The projected density of states (pDOS) over the adsorbed monomer, shown in Figure 3C, evidence an energy state 1.3 1.4 eV above the Fermi-level, corresponding to the observed threshold for electron-induced reaction (see Supporting Information for computational details). In a subsequent paper on monomeric pDIB, we shall detail the motions of the iodine atoms and the organic residue that lead to the observed I..I separation of 1.44 nm. The source of the internal binding in the linear dimer with approximately opposed central I-atoms, is indicated in Figure 4. 4116

dx.doi.org/10.1021/nl2023788 |Nano Lett. 2011, 11, 4113–4117

Nano Letters The figure shows contours of equal charge transfer,30 namely 0.004 e /Å3 (see Supporting Information for details). The computed geometry has the central I-atoms displaced slightly from one another in the vertical plane. The figure shows electron transfer from the I-atoms to the underlying pair of copper atoms that form four “short-bridges” directly beneath (see also Figure 1C, panel 2). The charge-transfer was 0.1 electron for each of the three I-atoms closely neighboring the surface. In sum, we have found for the electron-induced reaction at Cu(110) in four linear assemblies of p-diiodobenzenes, that the overall dynamics of terminal I-atom reaction involved linear extension of each C I bond by a mean distance of 0.36 nm in going from physisorbed reagent to chemisorbed product. Reaction even at this smooth metal surface was, therefore, highly localized. The prevalence of localized reaction led to a linear relationship between the initial separation of the terminal I-atoms in the physisorbed reagent and their subsequent separation as chemisorbed imprints on the metal surface. The four physisorbed reagent (pDIB)n molecules functioned as molecular calipers controlling the separations between the chemisorbed I-atom products. Such localized reaction may provide a useful means to fingerprint the structure of complex molecules adsorbed at metal surfaces.

’ ASSOCIATED CONTENT

bS

Supporting Information. Experimental and computational methods. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: 416-978-3580. Fax: 416-978-7580. E-mail: jpolanyi@ chem.utoronto.ca.

LETTER

(10) Maksymovych, P.; Yates, J. T., Jr. J. Am. Chem. Soc. 2006, 128, 10642. (11) Lu, P. H.; Polanyi, J. C.; Rogers, D. J. Chem. Phys. 1999, 111, 9905. (12) Harikumar, K. R.; McNab, I. R.; Polanyi, J. C.; Zabet-Khosousi, A.; Hofer, W. A. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 950. (13) Dobrin, S.; Harikumar, K. R.; Polanyi, J. C. Surf. Sci. 2004, 561, 11. (14) Dobrin, S.; Harikumar, K. R.; Matta, C. F.; Polanyi, J. C. Surf. Sci. 2005, 580, 39. (15) McNab, I. R.; Polanyi, J. C. Chem. Rev. 2006, 106, 4321. (16) Matta, C. F.; Polanyi, J. C. Philos. Trans. R. Soc. London, Ser. A 2004, 362, 1185. (17) McCarty, G. S.; Weiss, P. S. J. Phys. Chem. B 2002, 106, 8005. (18) Dougherty, D. B.; Lee, J.; Yates, J. T., Jr. J. Phys. Chem. B 2006, 110, 20077. (19) Hla, S.-W.; Meyer, G.; Rieder, K.-H. Chem. Phys. Chem. 2001, 2, 361. (20) Hla, S.-W.; Bartels, L.; Meyer, G.; Rieder, K.-H. Phys. Rev. Lett. 2000, 85, 2777. (21) McCarty, G. S.; Weiss, P. S. J. Am. Chem. Soc. 2004, 126, 16772. (22) Kresse, G.; Furthm€uller, J. Phys. Rev. B 1996, 54, 11169. (23) Grimme, S. J. Comput. Chem. 2004, 25, 1463. (24) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. (25) Palotas, K.; W. A. Hofer, W. A. J. Phys.: Condens. Matter 2005, 17, 2705. (26) M€uller, T.; Werblowsky, T. L.; Florio, G. M.; Berne, B. J.; Flynn, G. W. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5315. (27) Lee, J.; Dougherty, D. B.; Yates, J. T., Jr. Surf. Sci. 2007, 601, L91. (28) Lipton-Duffin, J. A.; Ivasenko, O.; Perepichka, D. F.; Rosei, F. Small 2009, 5, 592. (29) Simic-Milosevic, V.; Bocquet, M.-L.; Morgenstern, K. Surf. Sci. 2009, 603, 2479. (30) Hofer, W. A.; Fisher, A. J.; Wolkow, R. A.; P. Gr€utter, P. Phys. Rev. Lett. 2001, 87, 236104. (31) Caliper. Encyclopedia Britannica online; Encyclopedia Britannica: Chicago, IL, 2011; http://www.britannica.com/EBchecked/topic/ 89719/caliper. Accessed July 11, 2011.

’ ACKNOWLEDGMENT We thank Dr. I. R. McNab for valuable discussions. J.C.P. thanks the Natural Sciences and Engineering Research Council of Canada (NSERC), Photonics Research Ontario (PRO), Ontario Centre of Excellence (OCE), and the Xerox Research Centre Canada (XRCC) for their support of this work. W.A.H. thanks the Royal Society of London for support. J.C.P. and W.A.H. thank the Canadian Institute for Advanced Research (CIFAR) for support. ’ REFERENCES (1) Weiss, S. Science 1999, 283, 1676. (2) Yildiz, A.; Forkey, J. N.; McKinney, S. A.; Ha, T.; Goldman, Y. E.; Selvin, P. R. Science 2003, 300, 2061. (3) Sonnichsen, C.; Reinhard, B. M.; Liphardt, J.; Alivisatos, A. P. Nat. Biotechnol. 2005, 23, 741. (4) Denic, V.; Weissman, J. S. Cell 2007, 130, 663. (5) Stranick, S. J.; Kamna, M. M.; Weiss, P. S. Science 1994, 266, 99. (6) Stranick, S. J.; Kamna, M. M.; Weiss, P. S. Nanotechnology 1996, 7, 443. (7) Lauhon, L. J.; Ho, W. J. Phys. Chem. A 2000, 104, 2463. (8) Komeda, T.; Kim, Y.; Fujita, Y.; Sainoo, Y.; Kawai, M. J. Chem. Phys. 2004, 120, 5347. (9) Zotti, L. A.; Teobaldi, G.; Palotas, K.; Ji, W.; Gao, H.-J.; Hofer, W. A. J. Comput. Chem. 2008, 29, 1589. 4117

dx.doi.org/10.1021/nl2023788 |Nano Lett. 2011, 11, 4113–4117