Frustrated rotations in single-molecule junctions

Article abstract of DOI:10.1021/ja903731m

(Graph Presented) We compare the conductance of 1,4-bis(methylthio)benzene with that of 2,3,6,7-tetrahydrobenzo[1,2-b:4,5-b′]dithiophene and the conductance of 1,4-bis(methylseleno)benzene with that of 2,3,6,7- tetrahydrobenzo[1,2-b:4,5-b′]diselenophene a

Full text of DOI:10.1021/ja903731m

Published on Web 07/17/2009
Frustrated Rotations in Single-Molecule Junctions
Young S. Park,†,§ Jonathan R. Widawsky,‡,§ Maria Kamenetska,‡,§ Michael L. Steigerwald,†
Mark S. Hybertsen,| Colin Nuckolls,†,§ and Latha Venkataraman*,‡,§
Department of Chemistry, Department of Applied Physics and Applied Mathematics, and Center for Electron
Transport in Molecular Nanostructures, Columbia UniVersity, New York, New York, and Center for Functional
Nanomaterials, BrookhaVen National Laboratories, Upton, New York
Received May 7, 2009; E-mail: lv2117@columbia.edu
Understanding transport through single-molecule junctions is crucial
for the development of nanoscale devices.1 Electron conduction through
molecules bonded between metal electrodes depends not only on the
molecular structure but also on the metal and the chemical linking
group used to bind the molecule to the metal.2,3 When junctions are
formed using gold metal electrodes, thiol links are frequently used.4,5
However, the conductance of the single-molecule junction varies
significantly from junction to junction, making it difficult to map the
relation between molecular structure and junction conductance.6,7 This
is in contrast to measurements using amine linking groups8 for both
aliphatic and aromatic compounds9 and using methyl sulfide and
dimethyl phosphines for aliphatic chains.2 Measurements of single-
molecule junctions with thiol links result in a broad distribution of
Figure 1. (a) Schematic representation of the measurement. (b) Conduc-
conductances, attributed to different binding geometries at the Au-S
bond.6,7 In particular, the conductance of aromatic molecules is strongly
influenced by the orientation of the π system relative to the Au-S
bond.6 While a detailed study of such properties for junctions with
thiol links is not possible given the variability in bonding configura-
tions, they can be investigated for junctions with methyl sulfide linkers,
where a donor-acceptor bond is formed between an undercoordinated
Au atom on the electrode and the S lone pair. Here we compare the
conductance of 1,4-bis(methylthio)benzene (1) with that of 2,3,6,7-
tetrahydrobenzo[1,2-b:4,5-b′]dithiophene (3) and the conductance of
1,4-bis(methylseleno)benzene (2) with that of 2,3,6,7-tetrahydrobenzo
[1,2-b:4,5-b′]diselenophene (4). Because the orientation of the lone
pair is rigidly locked in 3 and 4 but not in 1 and 2 (see the structures
in Figure 1B), the results explicitly demonstrate the relationship
between the conductance and the orientation of the π system relative
to the Au-S or Au-Se donor-acceptor bond.
The syntheses of the molecules are outlined below and described
in detail in the Supporting Information (SI). To synthesize 1 or 2, 1,4-
dibromobenzene was dilithiated, and elemental sulfur or selenium was
added; the resulting intermediate was then alkylated with methyl
iodide.10 3 and 4 were synthesized by hydrogenations of the corre-
sponding benzodithiophene and benzodiselenophene, which were
performed in the presence of [Ru(triphos)(MeCN)3](BPh4)2 at 60 °C.11
Single-molecule junctions were created by repeatedly forming
and breaking Au point contacts4 with a modified STM in a solution
of the molecules in 1,2,4-trichlorobenzene (Figure 1a). For each
molecule studied, the measured conductance traces reveal steps at
molecule-dependent conductance values less than the quantum of
conductance G0 ) 2e2/h; these are due to conduction through a
molecule bonded in the gap between the two Au point contacts
(Figure S1). Figure 1b shows conductance histograms generated
(without any data selection) from 10 000 consecutively measured
traces. We see a peak in the conductance histograms for 3 and 4
tance histograms for the four molecular junctions. Measurements were
performed with a 25 mV bias on ∼1 mM solutions of the molecules in
1,2,4-trichlorobenzene. Inset: histograms on a linear scale. The structures
of the linkers 1-4 are also shown.
when the S (or Se) lone pair is oriented parallel to the π orbital of
the phenyl ring. Lorentzian fits to these peaks yield most probable
conductance values (defined as the position of the peak) of (1.20
( 0.05) × 10-2 and (1.12 ( 0.04) × 10-2 G0, respectively. For 1
and 2, we see a broad increase in histogram counts at a lower
conductance range; a Lorentzian (or Gaussian) could not be fit to
the data, indicating that there are no dominant conductance values
in these measurements (Figure 1b).
In order to understand the difference between these two
molecules, we analyzed in detail all of the individual conductance
traces used to generate the histograms shown in Figure 1b using
an automated algorithm2 (see the SI). For each conductance trace
with a step, we determined the average conductance of the step,
the slope of the step normalized to its average conductance, and
the step length (Figure S3). We found that for all four molecules
studied here, ∼80% of the traces exhibit a step. Figure 2a provides
histograms of the average step conductance showing that all of the
distributions have clear peaks: (0.7 ( 0.05) × 10-2 G0 for 1, (0.9
( 0.05) × 10-2 G0 for 2, and (1.4 ( 0.05) × 10-2 G0 for 3 and 4.
Analysis of the conductance data based on logarithmic binned
histograms also brings out peaks (see Figure S4). These histograms
contrast the full-trace histograms shown in Figure 1b and indicate
that the junction conductance could change during junction elonga-
tion for 1 and 2, when compared with 3 and 4. This is further
confirmed in Figure 2b, which shows the distribution of normalized
step slopes. Although both molecules have sloped steps, a larger
fraction of the steps for 1 and 2 are steeper than those for 3 and 4,
as can be seen from the larger full width at half-maximum of the
step slope histograms for 1 and 2 than for 3 and 4.
† Department of Chemistry, Columbia University.
There are two factors that control the orientation of the S (or
Se) lone pair relative to the π system, which in turn controls the
measured conductance of these two molecules. First, the probability
‡ Department of Applied Physics and Applied Mathematics, Columbia University.
§ Center for Electron Transport in Molecular Nanostructures, Columbia University.
| Brookhaven National Laboratories.
9
10820 J. AM. CHEM. SOC. 2009, 131, 10820–10821
10.1021/ja903731m CCC: $40.75  2009 American Chemical Society

C O M M U N I C A T I O N S
tunnel coupling on torsion is shown in Figure 3C. For 1, the effect
of misalignment to the π system of the ring is strong and relatively
symmetric. Increasing the torsion for 3 from the minimum-energy
configuration results in reduced tunnel coupling, while reducing
the torsion initially increases coupling because of buckling of the
fivefold ring (Figure S5).
The DFT calculations support a consistent physical picture. For 3,
the high energy cost for torsional angle distortion allows for only an
approximately (15° variation before the junction-formation energy
would be too small to sustain a measurement on the millisecond time
scale at room temperature. The corresponding variation in the predicted
conductance is about (25%, consistent with a well-defined peak in
the histogram in Figure 1B. On the other hand, the flat torsional energy
landscape for 1 indicates that as-formed junctions can sample a broad
range of tunnel coupling, resulting in a large variation in the measured
conductance from junction to junction. Furthermore, a specific junction
formed with 1 will thermally access a wide range of angles as well.
Therefore, although the nominal tunnel coupling calculated for the
minimum-energy geometries of 1 and 3 are similar, the thermally
averaged values are rather different. The estimated ratio of the thermally
averaged tunnel couplings (1.6) is similar to the ratio of the peak
conductance values in the step average histograms of Figure 2A.
In summary, we have established that the orientation of an Au-S
or Au-Se bond relative to the aromatic π system controls electron
transport through conjugated molecules. In cases such as those
discussed herein, the conduction pathway connects the Au electrodes
via the chalcogen p lone pairs and the aromatic π system, and
greater overlaps among these components leads to higher conduc-
tivity through the molecular junction.
Figure 2. (a) Normalized distributions of average step conductance based
on the analysis of 8752, 7415, 21687, and 9256 traces for 1, 2, 3, and 4,
respectively. Inset: Sample conductance traces for 1 and 3 showing a linear
fit to the molecular step (black). (b) Distribution of conductance step slope
normalized by average step conductance, showing that the conductance steps
are more sloped for 1 and 2 than for 3 and 4, as can be seen from the
narrower distribution for 3 and 4.
Figure 3. (a) Diagram showing the torsional angle sampled for 1 (see
Figure S5 for 3). (b) Energy as a function of Au-S-C-C torsional angle
for 1 (red) and 3 (blue) attached to Au5 clusters. (c) Square of the calculated
tunnel coupling (4t2) across 1 (red) and 3 (blue) attached to Au1 clusters as
a function of the Au-S-C-C torsional angle.
Acknowledgment. This work was supported primarily by the
NSF-NSEC (Award CHE-0641523), by NYSTAR and the Columbia
University RISE program, and in part by the DOE (DE-AC02-
98CH10886). L.V. thanks NSF (Career Award CHE-0744185) and
the ACS PRF.
of forming distinct junctions in which the angle between the Au-S
link vector and the phenyl plane (Au-S-C-C torsional angle)
varies significantly from its minimum-energy configuration in
response to other static constraints in the junction contributes
directly to the histogram width. Second, the probability of sampling
different torsional angles while a specific junction elongates and
ultimately breaks affects the measured conductance (thermal
average) and the slope of the conductance step. The observed broad
histograms for 1 and 2 could thus be explained by the low energy
cost for rotating the S (or Se) lone pair away from the π system in
the isolated molecule.
Supporting Information Available: Synthesis procedures, data
analysis, and theoretical methods. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Joachim, C.; Ratner, M. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 8800.
(2) Park, Y. S.; Whalley, A.; Kamenetska, M.; Steigerwald, M. L.; Hybertsen,
M. S.; Nuckolls, C.; Venkataraman, L. J. Am. Chem. Soc. 2007, 129, 15768.
(3) Chen, F.; Li, X. L.; Hihath, J.; Huang, Z. F.; Tao, N. J. J. Am. Chem. Soc.
2006, 128, 15874.
(4) Xu, B. Q.; Tao, N. J. J. Science 2003, 301, 1221.
(5) Reichert, J.; Ochs, R.; Beckmann, D.; Weber, H. B.; Mayor, M.; von
Lohneysen, H. Phys. ReV. Lett. 2002, 88, 176804. Reed, M. A.; Zhou, C.;
Muller, C. J.; Burgin, T. P.; Tour, J. M. Science 1997, 278, 252.
(6) Basch, H.; Cohen, R.; Ratner, M. A. Nano Lett. 2005, 5, 1668.
(7) Ulrich, J.; Esrail, D.; Pontius, W.; Venkataraman, L.; Millar, D.; Doerrer,
L. H. J. Phys. Chem. B 2006, 110, 2462. Li, X. L.; He, J.; Hihath, J.; Xu,
B. Q.; Lindsay, S. M.; Tao, N. J. J. Am. Chem. Soc. 2006, 128, 2135. Li,
C.; Pobelov, I.; Wandlowski, T.; Bagrets, A.; Arnold, A.; Evers, F. J. Am.
Chem. Soc. 2008, 130, 318. Gonzalez, M. T.; Wu, S. M.; Huber, R.; van
der Molen, S. J.; Schonenberger, C.; Calame, M. Nano Lett. 2006, 6, 2238.
(8) Li, Z.; Kosov, D. S. Phys. ReV. B 2007, 76, 035415. Quek, S. Y.;
Venkataraman, L.; Choi, H. J.; Loule, S. G.; Hybertsen, M. S.; Neaton,
J. B. Nano Lett. 2007, 7, 3477. Kiguchi, M.; Miura, S.; Takahashi, T.;
Hara, K.; Sawamura, M.; Murakoshi, K. J. Phys. Chem. C 2008, 112, 13349.
(9) Venkataraman, L.; Klare, J. E.; Tam, I. W.; Nuckolls, C.; Hybertsen, M. S.;
Steigerwald, M. L. Nano Lett. 2006, 6, 458. Venkataraman, L.; Klare, J. E.;
Nuckolls, C.; Hybertsen, M. S.; Steigerwald, M. L. Nature 2006, 442, 904.
(10) Engman, L.; Hellberg, J. S. E. J. Organomet. Chem. 1985, 296, 357.
(11) Bianchini, C.; Meli, A.; Moneti, S.; Oberhauser, W.; Vizza, F.; Herrera,
V.; Fuentes, A.; Sanchez-Delgado, R. A. J. Am. Chem. Soc. 1999, 121,
7071. Takimiya, K.; Konda, Y.; Ebata, H.; Niihara, N.; Otsubo, T. J. Org.
Chem. 2005, 70, 10569.
To explore this hypothesis more quantitatively, we performed
density functional theory (DFT) calculations12 for 1 and 3 coupled
to Au clusters to represent the contacts (see the SI). For both
molecules, the Au-S donor-acceptor bond energy ranges from
11 to 14 kcal/mol, depending on the structural details near the Au
contact atom. We constrained the natural torsion angle around the
S-C bonds in models with an inversion center (equivalent Au-S
bonds). The energy cost was evaluated for an Au5 cluster simulating
an Au adatom on the hcp hollow site of an Au(111) facet (Figure
S5). As shown in Figure 3b, the torsional energy surface for 1 is
very broad, with a barrier to full rotation of ∼1 kcal/mol. In contrast,
the energy cost to twist 3 is much higher. The tunnel coupling
through the junction was probed using Au1 clusters. The frontier
orbitals, of predominantly of Au s/S lone pair antibonding character,
are tunneling-coupled through the molecular backbone, resulting
in a symmetric and antisymmetric pair with a splitting of 2t. Similar
calculations for a variety of diamine-linked molecules have shown
that measured conductance correlates with the square of the
calculated tunnel coupling (4t2).13 The dependence of the calculated
(12) Jaguar 7.5; Schrodinger, LLC: New York, 2008. Perdew, J. P.; Burke, K.;
Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865. Wadt, W. R.; Hay, P. J.
J. Chem. Phys. 1985, 82, 284.
(13) Hybertsen, M. S.; Venkataraman, L.; Klare, J. E.; Whalley, A. C.; Steigerwald,
M. L.; Nuckolls, C. J. Phys.: Condens. Matter 2008, 20, 374115.
JA903731M
9
J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009 10821