NANO LETTERS
Single-Molecule Dynamics in a Self-Assembled 2D Molecular Sieve
2006 Vol. 6, No. 7 1360-1363
Guillaume Schull,* Ludovic Douillard, Ce´line Fiorini-Debuisschert, and Fabrice Charra* SerVice de Physique et Chimie des Surfaces et Interfaces, Commissariat a` l’Energie Atomique, Centre de Saclay, F-91191 Gif-sur-YVette Cedex, France
Fabrice Mathevet, David Kreher, and Andre´-Jean Attias* Laboratoire de Chimie Macromole´ culaire, UniVersite´ Pierre et Marie Curie, CNRS-UMR 7610, 4 Place Jussieu, F-75252, Paris Cedex 05, France Received February 8, 2006; Revised Manuscript Received May 30, 2006
ABSTRACT A two-dimensional molecular sieve has been realized. It consists of a host matrix of molecularly engineered building blocks self-assembled at the liquid−solid interface. The simultaneous size- and shape-dependent dynamics of different guest molecules is observed in situ, in real time with submolecular resolution using a scanning tunneling microscope both at the liquid−solid interface and under vacuum. The temperaturedependent dynamics reveals that the diffusion proceeds through thermally activated channeling between single-molecule surface cavities.
The parallel manipulation of individual molecules is the ultimate goal of much of the current research in nanoscience. A first route is the confinement of molecular motion until the continuum approximation breaks down and discretemolecule effects appear, as observed in zeolite analogues1 or inside nanotubes.2 A bottom-up alternative consists of surface self-assemblies tailored at the molecular scale. Although most studies on nanostructured surfaces are still focused on their structure,3 the demonstrations of selective adsorption-desorption phenomena of guest molecules inside single-molecule pores4-7 and the observation of rotating molecules inside bearing in a submonolayer film,8,9 or within self-assembled monolayers,10 represent pioneering first steps toward functional systems. Here, we report on the realization of a two-dimensional molecular sieve. It consists of a host matrix of engineered molecules self-assembled at the liquid-solid interface. The size- and shape-dependent dynamics of different guest molecules is observed in situ in real time with submolecular resolution. Its temperature dependence reveals that the diffusion proceeds via the surface through thermally activated channeling between single-molecule surface cavities. This permits one to derive molecular-design guidelines for tuning cavity or channel selectivity, opening interesting perspectives toward molecular-level understandings of transport phenomena, for example, through biomembranes, and practical applications in various domains, from catalysis to biotechnologies. * Corresponding authors. E-mail:
[email protected];
[email protected]. 10.1021/nl060292n CCC: $33.50 Published on Web 06/08/2006
[email protected];
© 2006 American Chemical Society
The supramolecular host network was realized using a starshaped stilbenoid compound with a 1,3,5-tristyrylbenzene conjugated core and decyloxy peripheral chains in the 3 and 5 positions (1,3,5-tris[(E)-2-(3,5-didecyloxyphenyl)-ethenyl]benzene, denoted TSB35, see Figure 1A together with the Supporting Information). This class of compounds is known to self-assemble as a honeycomb structure on highly oriented pyrolytic graphite (HOPG),11 leaving empty cavities as illustrated by the model of Figure 1B. Such cavities can be anticipated as suitable for hosting some other smaller guest molecules. We selected coronene and hexabenzocoronene (HBC) as guest candidates because of their chemical closeness and varied sizes and shapes (see Figure 1A). The formation of the honeycomb cavities (∼1.3 nm in diameter) is confirmed by scanning tunneling microscopy (STM) (Figure 1C). Their anticipated hosting capability was checked by addition of an excess of guest molecules in the solution: the appearance of a 1.3 nm circular bright spot at the center of each honeycomb site confirms the trapping inside those cavities of coronene (Figure 1D) and HBC (Figure 1E). It should be noted that coronene and HBC are imaged by spots of the same sizes although coronene (1.0 nm) is significantly smaller than HBC (1.3 nm). This can be explained by fast guest-molecule diffusion inside its host cavity during the course of each pixel acquisition (80 µs), which results in an averaged image over all possible guestmolecule locations. To investigate the long-range mass transport inside the matrix, which requires guest-molecule diffusion from cavity to cavity, we have studied incompletely filled matrixes obtained with diluted substoichiometric solutions of coronene
Figure 1. Molecular structures of the host molecule TSB35 and the two guest candidates, coronene and hexabenzocoronene (HBC), are presented (A). The inset (B) shows the model for the supramolecular assembly of the TSB35 host-matrix. The cavities are bounded by barriers formed by the conjugated cores of TSB35 (yellow) and linked by channels constituted by epitaxially adsorbed alkyl moieties (red). STM image of the self-assembled monolayer of the matrix formed through self-assembly of TSB35 on HOPG (C) and after addition of coronene (D) or HBC (E). All images have been recorded in the current, i.e., constant height, mode with a scanning rate of 40 ms per line, a sample bias of -1000 mV, and a set-point current of 46 pA.
or HBC (∼10-7mol/L). These series of experiments were carried on at liquid-solid interface at different substrate temperature and under ultrahigh vacuum (UHV) at room temperature (see the Supporting Information for details). For coronene, at low temperature and at fast imaging rates (Figure 2A), we observed simultaneously filled and empty cavities. In the series of 50 images forming the full movie (Supporting Information), the spatial distribution of the corresponding spots varies between successive images. In most cases, the appearance of a spot in one cavity is simultaneous with the disappearance of another spot in its neighborhood. The excerpt of Figure 2A illustrates the various situations. At a given temperature, if the imaging rate is decreased, the ensemble of blinking spots is changed into uniformly distributed stripes parallel to the fast-scan direction (Figure 2B), consistently with occupation changes occurring in the course of cavity imaging. For each temperature T in the range of -9 to +32 °C, the histogram of residence times (Figure 2D) has been constructed from a series of STM images. For each temperature, an exponential decay is systematically obtained whose characteristic time constants increases with decreasing temperature from 40 ms (32 °C) to 5000 ms (-9 °C, see the inset of Figure 2D). Note that direct and reverse fast-scan directions produce the same stripe patterns precluding any influence of the STM tip. The above results show that coronene is able to move in and out of the cavities. Whether this motion proceeds within the surface plane through a cavity-to-cavity hopping mechanism or by matter exchange with the solution is a fundamental point to be addressed. Three arguments give support to a 2D diffusion confined within the host matrix. First, depending upon which mechanism is involved, the statistics on the number of filled cavities will obey different Nano Lett., Vol. 6, No. 7, 2006
laws. Specifically, for a liquid-solid exchange a binomial law would give a standard deviation of 1.9 for an area comprising 15 cavities with an averaged occupancy ∼0.5 (see the Supporting Information), whereas the measured value on 120 successive images is below 1.4, consistently with molecules moving in and out of the image frame. This confirms quantitatively the concomitant appearance and disappearance evoked above. Second, the observed molecular dynamics persists when working at the vacuum-solid interface, where no exchange with any surrounding solution can be invoked (see Figure 2C). As for the liquid-solid interface, these stripes are characteristic of the appearance and disappearance of guest molecules in the cavities. Under vacuum, the transport mechanism can only be interpreted as a surface diffusion process. The corresponding average resident time measured at 299 K is 610 ms as reported in the inset of Figure 2D. Third, the closeness between the measured guest-residence decay and its exponential fit supports a Markovian hopping process. The experimental data obey an Arrhenius dependence whose energy barrier amounts to 0.81 ( 0.05 eV. This is the upper limit of known diffusion barriers of similar molecules moving on atomically flat surfaces at the solidliquid interface (0.3-0.8 eV12,13) and is much lower than the desorption energy of coronene from HOPG surface (1.4 eV14). The preexponential factor of the Arrhenius law, about 5 × 1014 Hz, is consistent with a thermally activated firstorder kinetics. This preexponential factor is an intrinsic parameter directly related to the diffusion process of one given molecule in a given 2D matrix. As such, the same value can thus be used for extrapolating the diffusion barrier energy under vacuum considering only the guest residence 1361
Figure 2. (A) Excerpt of four successive images (11 × 11 nm2, sample temperature: 12 °C, scanning rate: 24 ms/line, current set-point: 11 pA, Vtip ) 1000 mV) from a sequence of ∼50 STM images. The resolution was decreased to 75 × 75 pixels2 in order to attain an imaging rate of 1.8 s per frame. It shows the hopping of guest molecules between neighboring cavities as blinking spots at the center of honeycomb cavities. The sequence above shows various examples of single-molecule motions as highlighted by the arrows. The full movie is available as Supporting Information. (B and C) Incompletely filled matrix as obtained with under-stoichiometric concentrations of coronene at the liquid-solid interface (B, sample temperature: 20 °C, scanning rate: 41 ms/line, current set-point: 21 pA) and under UHV (C, sample temperature: 26 °C, scanning rate: 160 ms per line, sample bias: -1000 mV, set-point current: 15 pA). The images, acquired at room temperature, show uniformly distributed stripes inherent to molecular hopping occurring during the imaging of each cavity. The fast-scan direction is horizontal. (D) The histogram represents the decay of the presence of a coronene molecule, which appeared at time zero in a given cavity, as derived from these stripes in a series of successive images. The exponential decay fit (τ ) 120 ms at T ) 293 K) is plotted as a dotted line. The insert shows the temperature dependence of the decay constant at the solid-liquid interface (red disk) in comparison to that measured under vacuum at room temperature (purple diamond). The solid line represents an Arrhenius law with an energy barrier of 0.81 ( 0.05 eV.
time measured at room temperature. Compared to the value measured at the liquid-solid interface, the barrier height under UHV is then found to be about 0.05 eV higher. This difference can be explained by a different environment of the guest molecule at the liquid interface during the hopping process. Finally, from the adsorption energy of alkyl chains on HOPG (∼ 0.1 eV per CH2 unit 15) we can extrapolate an adsorption energy as large as ∼6 eV for the whole TSB35 molecule, to which one should add the packing energy of their closely aligned chains.16 This justifies the role of a stable matrix played by TSB35 relative to its guests. In contrast with coronene, the distribution of adsorbed HBC molecules stays unchanged at any scanning rates and for temperatures up to 47 °C, at which the Arrhenius law yields a residence time of only 30 ms for coronene. Incidentally, we had the opportunity to image a defect consisting in a ∼50% increased spacing between two neighboring matrix molecules (see the arrows in Figure 3B). Both cavities located on either sides of this wider channel appear striped, as observed with coronene. Moreover, the two sets of stripes are anticorrelated, that is, during each line scan the HBC molecule is observed in either one cavity 1362
or the other. It remains trapped in these two cavities despite the availability of vacant neighboring sites. This shows that, when the size of the channel between host molecules increases, larger molecules can flow in. Hence, the spaces between conjugated TSB35 cores mediate the transport. The easier hopping through the densely packed epitaxially arranged alkyl moieties rather than the conjugated cores is consistent with the stabilization effect of alkane layers on adsorbed conjugated molecules.17 This explains the absence of diffusion inside fully conjugated host matrixes deprived of such channels.4-8 The network composed of cavities linked by channels act as a 2D molecular sieve, allowing only the diffusion of smaller molecules. As a synthesis of the above results, the full physical process consists of two steps: first, the size of the cavity, ∼1.3 nm, allows the adsorption of the smaller rigid molecules, such as HBC and below. Second, only the molecules larger than the channel width (∼1.1 nm) remain sequestered and isolated in the cavity. This 2D sieving property is illustrated by the STM images obtained with a mixed solution of coronene and HBC, both at under-stoichiometric dilutions (Figure 4). Both molecules are adsorbed simultaneously: HBC occupies a fixed distribution of cavities, whereas coroNano Lett., Vol. 6, No. 7, 2006
Figure 3. Incompletely filled matrix as obtained with under-stoichiometric concentrations of HBC (A, sample temperature: 22 °C, scanning rate: 82 ms/line, set-point: 15 pA, horizontal fast scan direction): The image, acquired at room temperature, shows a discontinuous occupation of the cavities. (B) A defect that consists of an increased distance of 50% between two neighboring matrix molecules (as outlined by the black arrow) shows the motion between the two adjacent sites of a HBC molecule. (C, sample temperature: 22 °C, scanning rate: 82 ms/line, set-point: 15 pA).
lecular manipulation and a much faster operation. The fabrication of the 2D sieve, based on self-assembly, offers many degrees of freedom for tuning molecular dynamic properties. Supporting Information Available: Detailed experimental procedures and movie made from a series of ∼50 successive images showing the hopping of guest molecules between neighboring cavities as blinking spots at the center of honeycomb cavities. Images were acquired at 12 °C with an imaging rate of 1.8 s per frame (11 × 11 nm2, scanning rate: 24 ms/line, set-point: 11 pA). This material is available free of charge via the Internet at http://pubs.acs.org. References Figure 4. Room-temperature STM image obtained with a mixed coronene and HBC under-stoichiometric solution (20 × 20 nm2, sample temperature: 23 °C, scanning rate: 160 ms/line, current set-point: 8 pA). Coronene appears as striped cavities, whereas HBC appears as a solid one. The fast-scan direction is horizontal.
nene produces the characteristic stripes in all other cavities. This shows the ability to confer, simultaneously, highly different diffusion characteristics to otherwise chemically analogous molecules. Modifying sizes of the cavities and of the channels through the choice of alkyl chain lengths provides a way to improve the control on the dynamics of each specific molecule. Analogously, the functionalizations of alkyl chain ends or the chemical modifications of the conjugated core18 might be achieved for further tuning the chemical selectivity and the height of the energy barriers. In conclusion, our results show that a purposely designed self-assembled matrix is able to selectively control the trapping and surface diffusion of individual molecules. Such a nanoengineered array can thus be viewed as a two-dimensional molecular sieve. Compared with its 3D analogues, such as zeolites, the surface operation offers significant advances: The operation is monitored in-situ, at the singlemolecule scale using scanning-probe microscopy. Our findings point toward fundamentally new effects controlling molecular sieving. The whole active region is in contact with the environment, which offers additional mechanisms of moNano Lett., Vol. 6, No. 7, 2006
(1) Ferey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F.; Dutour, J.; Surble, S.; Margiolaki, I. Science 2005, 309, 2040. (2) Majumder, M.; Chopra, N.; Andrews, R.; Hinds, B. J. Nature 2005, 438, 44. (3) Barth, J. V.; Costantini, G.; Kern, K. Nature 2005, 437, 671. (4) Theobald, J. A.; Oxtoby, N. S.; Phillips, M. A.; Champness, N. R.; Beton, P. H. Nature 2003, 424, 1029. (5) Griessl, S. J. H.; Lackinger, M.; Jamitsky, F.; Markert, T.; Hietschold, M.; Heckl, W. M. Langmuir 2004, 20, 9403. (6) Stepanow, S.; Lingenfelder, M.; Dmitriev, A.; Spillmann, H.; Delvigne, E.; Lin, N.; Deng, X. B.; Cai, C. Z.; Barth, J. V.; Kern, K. Nat. Mater. 2004, 3, 229. (7) Stepanow, S.; Lin, N.; Barth, J. V.; Kern, K. Chem. Commun. 2006, 2153, 2153. (8) Li, Z.; Han, B.; Wandlowski, Th. Langmuir 2005, 21, 6915. (9) Gimzewski, J. K.; Joachim, C.; Schlittler, R. R.; Langlais, V.; Tang, H.; Johannsen, I. Science 1998, 281, 531. (10) Charra, F.; Cousty, J. Phys. ReV. Lett. 1998, 80, 1682. (11) Xu, S.; Zeng, Q.; Lu, J.; Wang, C.; Wan, L.; Bai, C. L. Surf. Sci. 2003, 538, L451. (12) Schunack, M.; Linderoth, T. R.; Rosei, F.; Laegsgaard, E.; Stensgaard, I.; Besenbacher, F. Phys. ReV. Lett. 2002, 88, 156102. (13) Lee, K.; Yu, J. Surf. Sci. 2005, 589, 8. (14) Zacharia, R.; Ulbricht, H.; Hertel, T. Phys ReV. B 2004, 69, 155406. (15) Muller, T.; Flynn, G. W.; Mathauser, A. T.; Teplyakov A. Langmuir 2003, 19, 2812. (16) Yin, S. X.; Wang, C.; Qiu, X. H.; Xu, B.; Bai, C. L. Surf. Interface Anal. 2001, 32, 248. (17) Xu, B.; Yin, S.; Wang, C.; Qiu, X.; Zeng, Q.; Bai, C. J. Phys. Chem. B. 2000, 104, 10502. (18) Attias, A. J.; Cavalli, C.; Donnio, B.; Guillon, D.; Hapiot, P.; Maltheˆte, J. Chem. Mater. 2002, 14, 375.
NL060292N 1363
