A Density Functional Theory Study on Model Junction Devices

Feb 23, 2010 - (I-V) characteristics of these simplest model devices were calculated from density ... The Stoddart-Heath molecular switch consists of ...
0 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. C 2010, 114, 4611–4616

4611

Stoddart-Heath [2]Rotaxane Molecular Switch Made Simple: A Density Functional Theory Study on Model Junction Devices Yun Hee Jang*,† and William A. Goddard III‡ Department of Materials Science and Engineering, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea, and Materials and Process Simulation Center, California Institute of Technology, Pasadena, California 91125 ReceiVed: June 10, 2009; ReVised Manuscript ReceiVed: January 12, 2010

The essential components of the Stoddart-Heath-type [2]rotaxane molecular switch tunnel junction devices are the aromatic shuttle and stations, which are attached to other components, such as linkers, stoppers, and anchors. In this study, we explored a possibility of whether the molecular switch can be made simple by leaving only the π-stacked aromatic key components between two metal electrodes. The current-voltage (I-V) characteristics of these simplest model devices were calculated from density functional theory using the nonequilibrium matrix Green’s function formalism. When the aromatic components are in direct contact with the surfaces, the I-V characteristics depend dramatically on the orientation of the π stacks, and spatial or temporal changes in the orientation can decrease the device robustness. The robustness can be restored by introducing a buffer layer or a covalent bond (say bulky stoppers and anchors as well as titanium adhesion layers) between the π stack and the electrodes, and this is, indeed, what has been done in the actual fabrication of the working devices. We also propose an alternative strategy to build a simple switch that uses controlled orientation change as the basis for switching. 1. Introduction Molecular electronics has matured from development of new molecular components to demonstration of electronic devices composed of single molecule, self-assembled monolayer, or Langmuir-Blodgett (LB) monolayer tunnel junctions.1-14 Still, many questions about the detailed structure and function of these devices need to be answered to maximize performance and to provide guidance in developing new devices. In this paper, we examine how the tunneling depends on the packing of molecules in such a junction and use these results to propose a simple design based on the Stoddart-Heath-type [2]rotaxane/[2]catenane system for a programmable molecular switch.4-14 The Stoddart-Heath molecular switch consists of the aromatic shuttle and station compounds, which are attached to other components, such as linkers, bulky stoppers, or long anchors, and shows a reversible on/off switching which arises most likely from moving the cyclobis-(paraquat-p-phenylene) (CBPQT4+; blue in the figure below) shuttle between the tetrathiafulvalene (TTF; green) station and the 1,5-dioxynaphthalene (DNP; red) station by oxidation and reduction of the TTF.15-22 Previous molecular dynamics simulations23,24 and quantum mechanics (QM) calculation25 as well as experiments in solution and on LB films26,27 have suggested a tendency for the paraquat or phenyl ring of the CBPQT shuttle in one station to lie parallel to an adjacent station to form a π stack in a folded or tilted conformation of the [2]rotaxane monolayer. Another QM calculation28 showed that the frontier molecular orbitals (highest occupied MO and lowest unoccupied MO) come exclusively from π orbitals of these aromatic components. Thus, * To whom correspondence should be sent. Phone: +82-62-970-2323. Fax: +82-62-970-2304. E-mail: [email protected]. † Gwangju Institute of Science and Technology. ‡ California Institute of Technology.

we expect efficient electron transport via these π orbitals, which have good overlap with each other through the π stacking. Since the essential components for such π electron transport are the π-stacked aromatic (redox-active) shuttle and stations, a simple hypothetic model device having only the π-stacked components between two electrodes should also exhibit the shuttling-induced on/off switching, and this model would form a basis for developing a simplified Stoddart-Heath molecular switch. Thus, we built the simplest π-stack models for the molecular junction by removing all the nonaromatic components, such as linkers, stoppers, and anchors. We retained only the CBPQT shuttle and the TTF/DNP stations either in the green state (CBPQT@TTF; OFF) or in the red state (CBPQT@DNP; ON) and stacked them between two gold electrodes (Figures 1, 2). The current-voltage (I-V) characteristics of these model devices were calculated with the density functional theory (DFT) method combined with the nonequilibrium Green’s function (NEGF) formalism. A concern with our simple model is that the CBPQT shuttle stays adjacent to both stations whether the CBPQT shuttle is at the TTF station or at the DNP station (2 and 5 in Figure 2, for instance). The shuttling between two stations might not alter significantly the electronic structure of the π stack nor the electron transport through it, as is expected for the ring-shape

10.1021/jp905442c  2010 American Chemical Society Published on Web 02/23/2010

4612

J. Phys. Chem. C, Vol. 114, No. 10, 2010

Figure 1. π-Stacked aromatic key components of the Stoddart-Heath molecular switch in two different states. Color code: light blue (CBPQT), red (oxygen), yellow (sulfur), gray (carbon), black (hydrogen), green and purple (PF6-).

Jang and Goddard models employed in the calculations are as follows (Figure 2): (a) Parallel-parallel (denoted ParPar), in which the major π components (TTF, DNP, and the paraquat groups of CBPQT) are stacked parallel to the surface (1 and 4); (b) Perpendicularparallel (denoted PerpPar), in which the major π components (TTF, DNP, and CBPQT paraquat) are stacked perpendicular to the surface and the phenyl groups of CBPQT lie parallel to the surface (2 and 5); (c) Nonparallel (denoted NotPar), in which the hollow of CBPQT lies parallel to the surface (3 and 6). 2. Calculation Details

Figure 2. Simplest model of the Stoddart-Heath molecular switch device in the green (CBPQT@TTF; top) and red (CBPQT@DNP; bottom) states, which contains only the π components stacked between Au(111) slabs in three different orientations. Their relative energies (in kcal/mol) with respect to the most stable 2 are shown together. Color code: gold (light brown) and the same as in Figure 1 for the rest.

[2]catenane. The switching magnitude (the on/off ratio) is, indeed, significantly smaller in a catenane-based junction than in a rotaxane-based junction.6-8 We also notice that a rigid spacer has been introduced between two stations in the most recent junction device,13,22 and this should be either to avoid the π stacking or to introduce a partition between two stations, even in the π stack. On the other hand, since π interactions are highly sensitive to the relative orientations of the π orbitals, the orientation of the π stack with respect to the electrode surface might have a significant influence on the electron transport, especially in our simple model in which the π stacks have direct contact with the electrode surface without any bulky stoppers or long anchors between them. Indeed, the junctions composed of [2]catenanes and short [2]rotaxanes exhibited low performance and large current fluctuation, which was attributed to the close contact of the redox-active component to the electrodes, and the performance was significantly improved with the introduction of bulky stoppers as buffer layers between the active component and the electrodes.6-8 The question of this study is therefore the following: (1) Would the tunneling depend on the orientation even more than on the shuttle position in our simple π stack model? (2) If it is the case, could we use this orientation dependence as a basis of switching? To answer these questions, we built six model devices by positioning the π stacks (either green or red; Figure 1) between two electrodes in three different orientations (ParPar, PerpPar, or NotPar; Figure 2) and investigated how the position of the shuttle in the π stack and the orientation of the π stack affect the I-V characteristics of the device. The six

The geometries of the isolated π-stacked components in the green and red states were optimized as in our previous work25 (the B3LYP DFT functional and the 6-31G** basis set using Jaguar v5.5 software29). Each optimized structure was sandwiched between two gold electrodes in the three different orientations, as shown in Figure 2. The gold electrodes were modeled by periodic rectangular (33 × 5) unit cells23,30 of three-layer Au(111) slabs [30 Au atoms per layer with positions taken from the experimental fcc bulk structure31 (a ) 4.08 Å, Au-Au distance ) 2.884 Å)]. The size of the unit cell was chosen as 14.985 Å × 14.419 Å × 30.205 Å. The surface area of 14.985 Å × 14.419 Å () 2.16 nm2) is large enough to contain the π-stack model of the Stoddart-Heath switch which has orientation-dependent footprints of 0.9-1.9 nm2 according to our previous MD simulations.23,24 The c cell parameter (30.205 Å) allows the optimum separation of 3.2-4.2 Å between the π stacks and the Au(111) surfaces (measured from the top or bottom plane of the stacks to the Au surface), which were optimized by carrying out a series of calculations at various separations. Repeating this unit cell periodically in both directions leads to smooth stacking of the unit cell into the infinite Au(111) slabs in the top electrode and those in the bottom electrode as in the bulk. The same c parameter was used for all the model devices. We carried out DFT calculations on these periodic systems using the PBE functional of DFT32,33 with a Gaussian basis set as implemented in SeqQuest.34-36 The core electrons of each atom were replaced by norm-conserving pseudopotentials,37-40 and the outer electrons were described by a double-ζ-pluspolarization (DZP) basis set.41 The pseudopotentials and basis sets for Au were taken from closely related previous studies30,42-44 (listed in the Supporting Information of reference 30). Since similar results were produced from the calculations with DZP and single-ζ-plus-polarization (SZP) basis sets, we chose the smaller SZP basis set for Au. We used 94 × 90 × 192 points for the real space grid (∼0.16 Å/grid) and only the Γ point for the Brillouin zone sampling. Without further geometry optimization, the I-V curves for the model devices were calculated using the Landauer-Bu¨ttiker formalism and NEGF approach coupled with DFT as implemented in SeqQuest.35,44,45 (See reference 44 for the details of the method.) It should be noted that we do not calculate the wave function self-consistently with the finite applied field, so the results are most accurate for low bias.44 It should be noted that, although we chose the Au electrodes in our study to be consistent with our previous theoretical studies and to make use of the results from those studies, the actual metal electrodes used in most of the Stoddart-Heath junction devices are Si and Ti/Al.6,8,13 When transition metal electrodes, such as Au and Pt, are used with disulfide-tethered [2]rotaxanes on them, the same switching behavior as in the working devices seems to still hold,46 but the device characteristics are so sensitive to the nature of the molecule-electrode contact that

[2]Rotaxane Molecular Switch Made Simple

J. Phys. Chem. C, Vol. 114, No. 10, 2010 4613

Figure 4. Lowest unoccupied molecular orbitals (LUMOs) of the CBPQT complexes (a) with TTF and (b) with DNP.

Figure 5. Log-scale I-V curves of the model devices at each orientation. The green curves are for the green states (CBPQT@TTF; 1-3), and the red curves are for the red states (CBPQT@DNP; 4-6). Figure 3. Log-scale transmission curves of the model devices with (a) orientation ParPar (1 and 4), (b) orientation PerpPar (2 and 5), and (c) orientation NotPar (3 and 6). Vertical red lines represent the Fermi levels.

the change in the intrinsic molecular nature away from the contact tends to be masked.9 3. Results and Discussion Using the DFT/NEGF method, we calculated the electron transmission coefficient T(E) at each energy, E, for the six model devices (1-6) representing two different shuttle positions in the π stack (green and red) and three different orientations of the π stack with respect to the electrode surface (ParPar, PerpPar, and NotPar). The logarithm-scale transmission curves log10 T(E) are shown in Figure 3. Since the calculated transmission originates from the resonant tunneling through the energy levels of the molecular layer, the transmission curve is closely related to the density of states [DOS(E)] of the molecule perturbed (and, in turn, broadened or shifted) by the interaction within the molecular layer and the molecule-electrode interaction. The frontier MOs (HOMOs and LUMOs) of our π stack are exclusively located around the aromatic planes of CBPQT, TTF, and DNP.28 Thus, in our simple model with the direct molecule-electrode contact, it is expected that the π stack has more favorable interaction with the electrode in the ParPar and PerpPar orientations than in the NotPar orientation. The DOS curves and, in turn, the transmission curves would show more pronounced peak broadening due to this interaction in the ParPar and PerpPar orientations than in the NotPar orientation. This is exactly what we see in Figure 3. The sharp features observed for the NotPar orientation (Figure 3c) indicate unperturbed molecular states of the π stack. According to the partial DOS curve [PDOS(E)] projected onto each component (not shown here), the two peaks corresponding to the HOMO levels (around -5.1 and -4.9 eV) originate mostly from the TTF and DNP stations. The peak at the deeper level (HOMO-1; around -5.1 eV) among them comes from the station staying next to the CBPQT shuttle (DNP in 3 and TTF in 6). This is why this peak becomes completely broadened and shifted with the direct contact with the electrode in the ParPar orientation (1 and 4; Figure 3a), whereas the other peak (HOMO; around -4.9 eV) originating from the other station sitting inside the CBPQT shuttle remains relatively sharp without significant peak shift.

The two peaks corresponding to LUMO levels (around -4.4 and -4.1 eV) come mostly from the CBPQT shuttle with a slight mixing with TTF (but interestingly, not with DNP; see Figure 4), and thus, these levels become significantly broadened and shifted in the PerpPar orientation (2 and 5; Figure 3b) and even more so in the ParPar orientation (1 and 4; Figure 3a). The degree of the broadening and the amount of the shift of these LUMO levels due to the molecule-electrode contact is larger for the red (CBPQT@DNP) state than for the green (CBPQT@TTF) state. It seems to be because no mixing between CBPQT and DNP in these LUMO levels pushes the electron distribution out of the CBPQT shuttle in the red state (Figure 4b), whereas a significant electron distribution is found between CBPQT and TTF inside the shuttle in the green state (Figure 4a). This outward MO of the CBPQT shuttle in the red state should allow favorable interaction with electrodes in the current case as well as with other moieties, such as the other free TTF station, resulting in favorable electron transport through this interaction. From all these considerations, we expect that the electron transport should be much stronger in the ParPar and PerpPar orientations than in the NotPar orientation. A slightly stronger electron transport in the red state than in the green state is also expected. The tunneling current I(V) through the model junction device at a bias voltage V can be calculated from the integration of the transmission curve T(E) within the window of the bias voltage V around the Fermi level [EF - 0.5 V, EF + 0.5 V]. Figure 5 shows the I-V curves (in a logarithm scale) calculated for the six model devices, 1-6. Comparing the green and red curves in Figure 5, the off-toon switching due to the shuttling from the green state to the red state is not very obvious with our simple model, especially at the ParPar orientation, but the shuttle-position dependence of the I-V characteristics follows the same pattern as found in our previous study.42 That is, within the typical read bias voltage range (0.2-0.3 V), the red state becomes more conductive than the green state, reaching the maximum on/off switching ratio [I(red)/I(green); Figure 6a] of 2.6 (at the ParPar and PerpPar orientations) and 5.1 (at the NotPar orientation). At the PerpPar and NotPar orientations, which have similar packing as that proposed for [2]catenane junction devices, the red state consistently yields higher currents than the green state in a rather wide voltage range. The calculated switching ratio is close to the value (∼2) reported for the [2]catenane device.6-8

4614

J. Phys. Chem. C, Vol. 114, No. 10, 2010

Figure 6. Log-scale ratios between the tunneling currents flowing through different model junction devices within a typical read bias voltage range. (a) I(red)/I(green) at each of the three orientations; (b) I(ParPar)/I(PerpPar) and I(ParPar)/I(NotPar) in the green or red state.

The calculation also indicates that the shuttling-induced switching [I(red)/I(green); Figure 6a] can be masked by a large orientation-induced variation of the current. At each shuttle position, the T(E) curves exhibit broadened features at the ParPar and PerpPar orientations. The consequence is that significantly higher current flow is allowed through the π stack at those orientations than at the NotPar orientation (1 > 2 . 3 for the green state and 4 > 5 . 6 for the red state; Figure 5). Within the bias voltage range 0.2-0.3 V, the ratio between the currents at the orientations ParPar and PerpPar [I(1)/I(2) and I(4)/I(5); Figure 6b] can be as high as 4.4 and 5.4. This is already higher than the shuttling-induced on/off ratio. The ratio between the currents at the orientations ParPar and NotPar [I(1)/I(3) and I(4)/ I(6); Figure 6b] can reach 2.2 × 103 and 7.7 × 102 within the same voltage range, showing even stronger orientation dependence. The relative energies (in kcal/mol) of the six model devices were calculated as 0.7 (1), 0.0 (2), and 5.6 (3) for the green states, and as 16.1 (4), 16.1 (5), and 21.6 (6) for the red states (Figure 2). The higher stability of the green states compared to the metastable red states agrees with the interpretation from a number of experiments.7,16,17,20 At each state (either green or red), the low-conductance device with the orientation NotPar (3 or 6) is significantly less stable than the high-conductance devices with two other orientations because of the lack of π contact with the Au surfaces. Thus, a well-equilibrated device may not have sufficient population of this orientation to affect the overall I-V characteristics of the device. On the other hand, the devices with the orientations ParPar and PerpPar have essentially the same stabilities. Therefore, we can expect them to have similar populations unless some special precautions are taken in fabrication. For ParPar, the average conductance for small bias (0.2-0.3 V) at the green (off) and red (on) state is 148 nA [I(1)] and 298 nA [I(4)], whereas for PerpPar, we expect 48 nA [I(2)] and 80 nA [I(5)]. Thus, assuming equal contributions of the two orientations to the I-V characteristics of the device, we expect a I(red)/I(green) ratio of 1.9, whereas the variation in orientation results in even higher ratio I(ParPar)/I(PerpPar) of 3.1 [I(1)/I(2)] and 3.7 [I(3)/I(4)]. Thus, a new design to avoid this orientation effect is needed for optimum performance and robustness of such simple devices with direct molecule-electrode contacts. 4. Designs for Robustness This orientation-dependent fluctuation of on/off switching arises from the direct contact between the π components and the electrode upon reorienting the π stacks. To remove this ambiguity, one could either (a) introduce into the shuttle a substituent that would ensure exactly one orientation with respect to the electrode (thiol groups in several locations on the shuttle that would bind strongly to the electrode or an adhesion layer,

Jang and Goddard

Figure 7. Log-scale transmission curves calculated on 4 before and after adding 3 Å of more space between the π stack and each electrode.

Figure 8. A design for a new type of rotaxane-based switch device, which combines a voltage-driven programmable orientation switch with the orientation-dependent electron transport.

for example) but still allow the dumbbell to pass through the shuttle to reach the on and off position or (b) introduce extra space between the π stack and each electrode that would decrease the transmission by several orders of magnitude but would reduce its orientation-dependent variation. For instance, introduction of 3 Å of space between the π stack and each electrode in the model device 4 sharpens the transmission features (4gap in Figure 7), which now look similar to (but much smaller than) those found in the device 5 (Figure 3). Devices with the π stack positioned in different orientations (say, ParPar 4 and PerpPar 5) would show similar tunneling aspects, and the robustness would be retained, so long as there is sufficient buffer space between the π stack and the electrodes. A strategy for the approach (b) [retaining robustness by avoiding direct contact between the π stack and the electrode to reduce orientation-dependent current variation] is to introduce buffer layers by attaching the long anchors or bulky stoppers to the π components, and this is, indeed, what has been done in the actual fabrication of the working Stoddart-Heath devices.8,13 A second strategy is to build a new type of simple molecular switch device by taking advantage of this orientation-dependent current variation, especially the 3 orders of magnitude difference between the orientation ParPar (or PerpPar) and the orientation NotPar [I(1)/I(3) ∼ 2 × 103 at 0.2-0.3 V; Figure 6b]. We propose a molecular architecture that may enable a voltagedriven programmable orientation switch of the π stacks between the orientation ParPar (or PerpPar) and the less stable orientation NotPar. This could be done by attaching ionic or polar groups to a station component (Figure 8) that would reorient the shuttle reversibly in response to an applied field. At low bias voltages (V < Vc), the device with this new architecture would be at the low-resistance state (on), with the orientation ParPar (or PerpPar) favored. At high bias voltages (V > Vc), this device would switch to the high-resistance state (off) as the π stacks in the device turn into the low-conductance orientation NotPar (Figure 8). We expect an on/off ratio of ∼103 [I(1)/I(3)] for this device, and this would not be easily masked by the orientationdependent current variation at the ON state [I(1)/I(2) ∼ 3]. To realize this new architecture, we will need to solve problems such as (1) the high energy barrier for the molecular reorientation in a close-packed monolayer and (2) the formation of electric short circuits through the thin monolayer without a

[2]Rotaxane Molecular Switch Made Simple bulky stopper or a protective coat of adhesion layer. First, the voltage-driven reorientation of the shuttle-dumbbell complex might represent a rather large molecular motion, and it is not clear whether such a motion is actually possible in a closepacked monolayer, but there have been quite a few observations of electric-field-induced reorientation (or orientation change due to other external stimuli) in liquid-crystalline layers47-51 or other types of layers52-56 on surfaces, and we will also be able to make the monolayer less close-packed by introducing surface diluents [such as thiols (alkyl or aromatic)57-59 or hairy counterions (DMPA-, which has been used with the [2]catenane molecular switches)7], which can form a homogeneous (not phase-segregated) mixture with our shuttle-dumbbell complexes in the monolayer. Second, vapor deposition of the top metal electrode can result in the formation of filament-like electric short circuits due to the penetration of the metal atoms60,61 through our thin monolayer, which does not have protective bulky stoppers or adhesion layers. We first propose to use a scanning tunneling microscopy or conducting atomic force microscopy tip as the top electrode60,61 (instead of the top electrode deposition) for the proof of the concept. Then, for the actual fabrication, we propose to employ a soft contact deposition method (instead of the vapor deposition), such as a nanoimprint technology with a direct metal transfer method.61,62 We also propose to introduce single or a few layers of graphene sheets before the top electrode deposition or use a graphitic material [such as HOPG (highly oriented pyrolytic graphite), carbon nanotube, or graphene] as a top electrode to avoid the metal filament formation while keeping a good π contact between the molecule and the electrode. More detailed chemical structure for this new architecture as well as the voltagedependent molecular motion in the monolayer and the detailed device characteristics, such as the critical write bias voltage Vc and the spatial/temporal variation of the on/off ratio, will be further investigated and reported separately. 5. Summary The DFT/NEGF calculations on the simplest π-stack model inspired by the Stoddart-Heath [2]rotaxane molecular switch indicate that the electron transport through the π stack can be switched on and off by changing the position of the shuttle (by a redox control or a bias voltage), confirming the suggestion from a great deal of experimental and theoretical studies. However, the calculation shows that a change in the orientation of the π stack in direct contact with the Au(111) surfaces, which can occur during operation due to a thermal rotation or a redoxinduced conformation change, can also induce the on/off switching, hampering the robustness of such devices. Introduction of a buffer layer hindering the direct contact between the π stacks and the electrodes (such as the bulky stoppers or long tails attached to the π components) or a covalent bond between the π stack and the electrodes (such as a titanium adhesion layer), which are employed in the actual fabrication of working devices, should reduce this noiselike orientation-dependent switching and restore the device robustness. On the other hand, we propose that a new type of simple molecular switch devices can be built with a proper molecular architecture designed to introduce a programmable orientation switch of the π stack in combination with the orientation-dependent electron transport. Acknowledgment. This work was supported by PIMS (GIST, Korea), KICOS (Korea), NRF (Korea), and MARCO-FENA. In addition, the facilities used were supported by KISTI (Korea), ONR-DURIP, and ARO-DURIP.

J. Phys. Chem. C, Vol. 114, No. 10, 2010 4615 References and Notes (1) Carroll, R. L.; Gorman, C. B. Angew. Chem., Int. Ed. 2002, 41, 4378. (2) Reed, M. A.; Tour, J. M. Sci. Am. 2000, 282, 86. (3) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000, 408, 541. (4) Collier, C. P.; Wong, E. W.; Belohradsky, M.; Raymo, F. M.; Stoddart, J. F.; Kuekes, P. J.; Williams, R. S.; Heath, J. R. Science 1999, 285, 391. (5) Wong, E. W.; Collier, C. P.; Behloradsky, M.; Raymo, F. M.; Stoddart, J. F.; Heath, J. R. J. Am. Chem. Soc. 2000, 122, 5831. (6) Collier, C. P.; Mattersteig, G.; Wong, E. W.; Luo, Y.; Beverly, K.; Sampaio, J.; Raymo, F. M.; Stoddart, J. F.; Heath, J. R. Science 2000, 289, 1172. (7) Pease, A. R.; Jeppesen, J. O.; Stoddart, J. F.; Luo, Y.; Collier, C. P.; Heath, J. R. Acc. Chem. Res. 2001, 34, 433. (8) Luo, Y.; Collier, C. P.; Jeppesen, J. O.; Nielsen, K. A.; Delonno, E.; Ho, G.; Perkins, J.; Tseng, H.-R.; Yamamoto, T.; Stoddart, J. F.; Heath, J. R. ChemPhysChem 2002, 3, 519. (9) Yu, H.; Luo, Y.; Beverly, K.; Stoddart, J. F.; Tseng, H.-R.; Heath, J. R. Angew. Chem., Int. Ed. 2003, 42, 5706. (10) Heath, J. R.; Ratner, M. A. Phys. Today 2003, 56, 43. (11) Flood, A. H.; Stoddart, J. F.; Steuerman, D. W.; Heath, J. R. Science 2004, 306, 2055. (12) Beckman, R.; Beverly, K.; Boukai, A.; Bunimovich, Y.; Choi, J. W.; DeIonno, E.; Green, J.; Johnston-Halperin, E.; Luo, Y.; Sheriff, B.; Stoddart, J. F.; Heath, J. R. Faraday Discuss. 2006, 131, 9. (13) Green, J. E.; Choi, J. W.; Boukai, A.; Bunimovich, Y.; JohnstonHalperin, E.; Delonno, E.; Luo, Y.; Sheriff, B. A.; Xu, K.; Shin, Y. S.; Tseng, H.-R.; Stoddart, J. F.; Heath, J. R. Nature 2007, 445, 414. (14) Dichtel, W. R.; Heath, J. R.; Stoddart, J. F. Philos. Trans. R. Soc. A 2007, 365, 1607. (15) Balzani, V.; Credi, A.; Mattersteig, G.; Matthews, O. A.; Raymo, F. M.; Stoddart, J. F.; Venturi, M.; White, A. J. P.; Williams, D. J. J. Org. Chem. 2000, 65, 1924. (16) Tseng, H.-R.; Vignon, S. A.; Stoddart, J. F. Angew. Chem., Int. Ed. 2003, 42, 1491. (17) Kang, S.; Vignon, S. A.; Tseng, H.-R.; Stoddart, J. F. Chem.sEur. J. 2004, 10, 2555. (18) Huang, T. J.; Tseng, H.-R.; Sha, L.; Lu, W.; Brough, B.; Flood, A. H.; Yu, B.-D.; Celestre, P. C.; Chang, J. P.; Stoddart, J. F.; Ho, C.-M. Nano Lett. 2004, 4, 2065. (19) Steuerman, D. W.; Tseng, H.-R.; Peters, A. J.; Flood, A. H.; Jeppesen, J. O.; Nielsen, K. A.; Stoddart, J. F.; Heath, J. R. Angew. Chem., Int. Ed. 2004, 43, 6486. (20) Choi, J. W.; Flood, A. H.; Steuerman, D. W.; Nygaard, S.; Braunschweig, A. B.; Moonen, N. N. P.; Laursen, B. W.; Luo, Y.; DeIonno, E.; Peters, A. J.; Jeppesen, J. O.; Xu, K.; Stoddart, J. F.; Heath, J. R. Chem.sEur. J. 2006, 261. (21) DeIonno, E.; Tseng, H.-R.; Harvey, D. D.; Stoddart, J. F.; Heath, J. R. J. Phys. Chem. B 2006, 110, 7609. (22) Norgaard, K.; Laursen, B. W.; Nygaard, S.; Kjaer, K.; Tseng, H.R.; Flood, A. H.; Stoddart, J. F.; Bjornholm, T. Angew. Chem., Int. Ed. 2005, 44, 7035. (23) Jang, Y. H.; Jang, S. S.; Goddard, W. A., III J. Am. Chem. Soc. 2005, 127, 4959. (24) Jang, S. S.; Jang, Y. H.; Kim, Y.-H.; Goddard, W. A., III; Flood, A. H.; Laursen, B. W.; Tseng, H.-R.; Stoddart, J. F.; Jeppesen, J. O.; Choi, J. W.; Steuerman, D. W.; DeIonno, E.; Heath, J. R. J. Am. Chem. Soc. 2005, 127, 1563. (25) Jang, Y. H.; Goddard, W. A., III J. Phys. Chem. B 2006, 110, 7660. (26) Yamamoto, T.; Tseng, H.-R.; Stoddart, J. F.; Balzani, V.; Credi, A.; Marchioni, F.; Venturi, M. Collect. Czech. Chem. Commun. 2003, 68, 1488. (27) Lee, I. C.; Frank, C. W.; Tamamoto, T.; Tseng, H.-R.; Flood, A. H.; Stoddart, J. F.; Jeppensen, J. O. Langmuir 2004, 20, 5809. (28) Jang, Y. H.; Hwang, S.; Kim, Y.-H.; Jang, S. S.; Goddard, W. A., III J. Am. Chem. Soc. 2004, 126, 12636. (29) Jaguar 5.5 Schrodinger Inc.: Portland, OR, 2003. (30) Jang, Y. H.; Goddard, W. A., III J. Phys. Chem. C 2008, 112, 8715. (31) Maeland, A.; Flanagan, T. B. Can. J. Phys. 1964, 42, 2364. (32) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865. (33) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1997, 78, 1396. (34) Schultz, P. A. SeqQuest Electronic Structure Code; Sandia National Laboratories: Albuquerque, NM; http://dft.sandia.gov/Quest. (35) Williams, A. R.; Feibelman, P. J.; Lang, N. D. Phys. ReV. B 1982, 26, 5433. (36) Feibelman, P. J. Phys. ReV. B 1987, 35, 2626. (37) Hamann, D. R. Phys. ReV. B 1989, 40, 2980. (38) Redondo, A.; Goddard, W. A., III; McGill, T. C. Phys. ReV. B 1977, 15, 5038.

4616

J. Phys. Chem. C, Vol. 114, No. 10, 2010

(39) Troullier, N.; Martins, J. L. Phys. ReV. B 1991, 43, 1993. (40) Fuchs, M.; Scheffler, M. Comput. Phys. Commun. 1999, 119, 67. (41) Feibelman, P. J. Phys. ReV. B 1988, 38, 1849. (42) Kim, Y.-H.; Jang, S. S.; Jang, Y. H.; Goddard, W. A., III Phys. ReV. Lett. 2005, 94, 156801. (43) Kim, Y.-H.; Jang, S. S.; Goddard, W. A., III J. Chem. Phys. 2005, 122, 244703. (44) Kim, Y.-H.; Tahir-Kheli, J.; Schultz, P. A.; Goddard, W. A., III Phys. ReV. B 2006, 73, 235419. (45) Datta, S. Quantum Transport: Atom to Transistor; Cambridge University Press: Cambridge, UK, 2005. (46) Tseng, H.-R.; Wu, D.; Fang, N. X.; Zhang, X.; Stoddart, J. F. ChemPhysChem 2004, 5, 111. (47) Luk, Y.-Y.; Abbott, N. L. Science 2003, 301, 623. (48) Lim, J. K.; Kwon, O.; Kang, D. S.; Joo, S.-W. Chem. Phys. Lett. 2006, 423, 178. (49) Duzhko, V.; Singer, K. D. J. Phys. Chem. C 2007, 111, 27. (50) Piot, L.; Marie, C.; Feng, X.; Mullen, K.; Fichou, D. AdV. Mater. 2008, 20, 3854. (51) Kim, J. K.; Araoka, F.; Jeong, S. M.; Dhara, S.; Ishikawa, K.; Takezoe, H. Appl. Phys. Lett. 2009, 95, 063505. (52) Wang, D.; Wan, L.-J. J. Phys. Chem. C 2007, 111, 16109.

Jang and Goddard (53) Sagara, T.; Fukuda, M.; Nakashima, N. J. Phys. Chem. B 1998, 102, 521. (54) Ishida, T.; Koyama, E.; Tokuhisa, H.; Belaissaoui, A.; Nagawa, Y.; Nakano, M.; Mizutani, W.; Kanesato, M. Jpn. J. Appl. Phys. 2004, 43, 4561. (55) Diao, Y.-X.; Han, M.-J.; Wan, L.-J.; Itaya, K.; Uchida, T.; Miyake, H.; Yamakata, A.; Osawa, M. Langmuir 2006, 22, 3640. (56) Chen, W.; Chen, S.; Huang, H.; Qi, D. C.; Gao, X. Y.; Wee, A. T. S. Appl. Phys. Lett. 2008, 92, 063308. (57) Cooke, G.; Duclairoir, F. M. A.; Rotello, V. M.; Stoddart, J. F. Tetrahedron Lett. 2000, 41, 8163. (58) Bryce, M. R.; Cooke, G.; Duclairoir, F. M. A.; John, P.; Perepichka, D. F.; Polwart, N.; Rotello, V. M.; Stoddart, J. F.; Tseng, H.-R. J. Mater. Chem. 2003, 13, 2111. (59) Azehara, H.; Mizutani, W.; Suzuki, Y.; Ishida, T.; Nagawa, Y.; Tokumoto, H.; Hiratani, K. Langmuir 2003, 19, 2115. (60) Akkerman, H. B.; de Boer, B. J. Phys.: Condens. Matter 2008, 20, 013001. (61) Chen, F.; Hihath, J.; Huang, Z.; Li, X.; Tao, N. J. Annu. ReV. Phys. Chem. 2007, 58, 535. (62) Kim, T. W.; Lee, K.; Oh, S.-H.; Wang, G.; Kim, D.-Y.; Jung, G.Y.; Lee, T. Nanotechnology 2008, 19, 405201.

JP905442C