Chemical control of double barrier tunnelling in α,ω- dithiaalkane molecular wires

Article abstract of DOI:10.1039/b709576g

Single molecule conductance measurements on 1,4-bis-(6-thia-hexyl)-benzene derivatives reveal (i) that benzene rings serve as an effective indentation in the tunnelling barrier, and (ii) that more electron-rich benzene rings give higher conductances, cons

Full text of DOI:10.1039/b709576g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Chemical control of double barrier tunnelling in a,v-dithiaalkane
molecular wires{
Edmund Leary, Simon J. Higgins,* Harm van Zalinge, Wolfgang Haiss and Richard J. Nichols
Received (in Cambridge, UK) 26th June 2007, Accepted 7th August 2007
First published as an Advance Article on the web 13th August 2007
DOI: 10.1039/b709576g
orbital energies for the alkyl groups lie far from the gold Fermi
energy, while the frontier orbitals for the p-system of the viologen
are much closer to the Fermi energy. This is apparent upon
comparing the single molecule conductance of 6V6 (0.5 nS)16,18
with that of dodecane-1,12-dithiol (i.e. two back-to-back barriers
with no well; 0.028 nS).19 Even though 6V6 is longer than
dodecane-1,12-dithiol, it is more conductive due to the influence of
the viologen acting as a barrier indentation.
Single molecule conductance measurements on 1,4-bis-(6-thia-
hexyl)-benzene derivatives reveal (i) that benzene rings serve as
an effective indentation in the tunnelling barrier, and (ii) that
more electron-rich benzene rings give higher conductances,
consistent with hole conduction (i.e. via the benzene HOMO).
The advent of reliable techniques for measuring the electrical
properties of single molecules bridging two metallic contacts has
led to an explosion of interest in this field.1,2 There is still a lack of
understanding of basic structure–property relationships in mole-
cular electronics. Attempts to address this are often complicated by
the fact that more than one potentially significant parameter (e.g.
molecular length, conformation, contact geometry, steric and
The viologen group in 6V6 is electroactive. When measured
under potential control in aqueous electrolyte, the conductance
increased approximately six-fold as the potential was swept
negative to access the viologen cation radical redox state, giving
a value approximately the same as that for a single C6 alkane
barrier (in hexane-1,6-dithiol; 2.5 nS16,17), indicating that the
viologen radical cation acts as a particularly deep well.
electronic effect of substituents) is varied simultaneously.3–5
A
common approach in molecular electronics has been to search for
organic molecular devices capable of basic electrical functions.
Molecules designed to act as molecular wires,6,7 switches8–10 and
diodes11 have recently been examined theoretically and/or
experimentally, and devices have been produced which show
rectification and negative differential resistance (NDR) behaviour.
This points to the possibility of generating organic devices that
mimic inorganic semiconductor junctions, although technological
deployment seems some way off.
In this paper, we address the following questions. Firstly, do
smaller, non-redox-active moieties also act as ‘‘wells’’? Secondly, if
so, can they be chemically tuned by altering the electronic structure
of the well? Thirdly, is there a relationship between the frontier
orbital energies of the well and the conductance of the whole
junction? Accordingly, we have made a range of molecules 1–4 in
which the contacts and barriers are constant (HS(CH2)6–) and the
wells are varied. These are listed in Table 1 together with the
HOMO and LUMO energies determined using the SPARTAN04
implementation of DFT (B3LYP/6-31G**).
An architecture that is expected to be important in future
generations of devices is the double tunnelling barrier junction.12
These have been introduced recently into field effect transistors,13
diodes14 and bipolar transistors.15 In these inorganic devices,
metalorganic vapour phase epitaxy is used to assemble a ‘well’ or
barrier indentation, sandwiched between the two tunnelling
barriers.
Table 1 Well structures used, and their frontier orbital energies (in
eV). R 5 HS(CH2)6– in all cases
Molecule
Well structure
HOMO
LUMO
We recently determined the single molecule conductance of the
molecule 6V6 (Fig. 1).16,17 When contacted to gold electrodes via
the thiol groups, this can be regarded as the single molecule
equivalent of a double tunnelling barrier because the frontier
1
26.25
+0.06
2
26.62
20.59
3
4
25.23
25.97
+0.09
+0.12
Fig. 1 The molecule 6V6 as a double tunnelling barrier junction.
Department of Chemistry, University of Liverpool, Crown Street,
Liverpool, UK L69 7ZD. E-mail: shiggins@liv.ac.uk;
Fax: +44 (0) 151 794 3588; Tel: +44 (0) 151 794 3512
{ Electronic supplementary information (ESI) available: Syntheses and
characterisation of 1–4; experimental details. See DOI: 10.1039/b709576g
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 3939–3941 | 3939

View Article Online
Molecules 1–4 (Table 1) were synthesised (as dithioacetates) and
fully characterised by microanalytical and spectroscopic methods.{
We then employed the scanning tunnelling microscopy (STM)
based I(t) technique of Haiss20 to determine the single molecule
conductances of 1–4. A flame-annealed gold-coated glass slide is
dipped into a dilute (1024–1025 M) solution of the appropriate
dithioacetate for one minute to allow the formation of a low
coverage monolayer (our recent experience is that it is not
necessary to employ CH3C(O)Cl–MeOH deprotection21 prior to
monolayer formation). A gold STM tip is then brought to a fixed
distance above the gold surface under perfluorodecane, controlled
using the set point current. The feedback loop is switched off and
the current is monitored as a function of time. As molecules
spontaneously form and break bridges between tip and surface,
jumps in the tunnelling current are observed (for example, Figure
S4 in Supplementary Information) due to conductance through the
molecule(s). These jumps are analysed statistically. Fig. 2 shows a
typical histogram for such an experiment. For molecules 1–4, we
repeated this experiment at different tip–substrate bias potentials,
and the mean single molecule current was then plotted against the
bias potential. The slope gives the molecular conductance (Table 2).
The results were checked using a complementary method, the
I(s) technique. Here, the tip is withdrawn while maintaining a
constant x–y position, and a current–distance (I(s) where s 5
relative tip–sample distance) curve is collected. We typically
observe current–distance behaviour characteristic of the formation
of molecular wires (Figure S1, Supplementary Information) with a
plateau in the current (I(w)) due to conductance through the fully-
extended molecule in its lowest-energy conformation. As the tip is
withdrawn further, the molecule then detaches at a distance
characteristic of its length. Again, the experiment is repeated many
times, and the results are analysed statistically as for the I(t)
measurements. The results (Table 2) are valuable in two ways.
Firstly, results for the two techniques are in good agreement,
giving us added confidence in the I(t) measurements. Secondly,
they provide confirmation that we are dealing with molecular
Table 2 Conductances (standard deviations in parentheses) of 1–4
measured using I(t) and I(s) methods, and mean detachment
distances (standard deviations in parentheses) measured using I(s)
method
Molecule
I(t) (nS)
I(s) (nS)
s1/2 (nm)
1
2
3
4
0.67 (0.07)
0.48 (0.05)
0.92 (0.14)
0.77 (0.11)
0.74 (0.24)
0.40 (0.09)
0.90 (0.19)
0.69 (0.27)
1.83 (0.19)
2.35 (0.28)
1.33 (0.14)
1.38 (0.15)
events, because the mean break-off distances are reasonably
consistent with the lengths of the molecules (2 nm S–S distances for
the fully-extended, transoid conformers of 1–4 as determined by
molecular mechanics). The degree of uncertainty in the detachment
distance measurement (see the histogram in Figure S2) is larger
than in the determination of conductance because of the range of
contact geometries, and the difficulty of maintaining a constant
starting-point in the measurements.
The results in Table 2 are interesting in several respects. Firstly,
it is clear that the conductances are considerably greater than
would be expected for an alkanedithiol of the same length as 1–4.
This indicates that the aromatic units do indeed act as a ‘‘well’’ in
the tunnelling barrier for 1–4. Secondly, it is clear that the
conductance varies with substituent; electron-donating groups lead
to higher conductances. In Fig. 3, we plot the conductances against
the HOMO energy for the aromatic unit, obtained from the
Spartan04 implementation of DFT (B3LYP, 6-31G**; equilibrium
geometry calculated for the transoid, extended form of 1–4
minimised using molecular mechanics).
Fig. 3 Plot of conductances determined by I(t) method (with standard
deviations) against HOMO energy for molecules 1–4.
For molecules such as alkanedithiols in which superexchange is
the conductance mechanism, conductance G decreases exponen-
tially with the length of the molecule (G 5 Aexp(2bNN), equation
1) where A is characteristic of the metal–molecule coupling and is
therefore contact-dependent, bN is a decay constant characteristic
of the repeat unit (–CH2–), and N is the number of repeat units.
Superexchange is also expected to be the operable mechanism for
molecules 1–4, especially given that the central chemical group (the
Fig. 2 Histogram of the current jumps observed in one of the I(t)
measurements on molecule 4. The tip–substrate potential in this example
was +600 mV. The large peak, J1, corresponds to jumps involving a single
molecule, and the smaller peak J2 to jumps involving two molecules.
3940 | Chem. Commun., 2007, 3939–3941
This journal is ß The Royal Society of Chemistry 2007

View Article Online
‘‘well’’) is off-resonance.18 However, simple barrier tunnelling
(equation 1, where the decay constant b or k is proportional to
the square root of barrier height) cannot adequately describe
tunnelling across 1–4 since the ‘‘well’’ leads to a double
tunnelling barrier with the barrier height only reduced at the well
rather than across the entire molecule. Fig. 3 is plotted in linear
form since a priori the relationship between conductance and the
HOMO position is not known and in any case is not likely to
follow the simple single barrier tunnelling model implicit in
equation 1. The observation that a clearly apparent but rather
weak (linear) relationship exists between conductance and HOMO
position is noteworthy, and demonstrates that these molecular
double tunnelling barriers are not represented adequately by
equation 1, for which a reduction in barrier height across the
whole molecular wire would lead to a non-linear relationship
between conductance and frontier orbital position. Indeed, for a
selection of conjugated molecular wires in which a single frontier
orbital extends between the two sulfur terminal groups, we observe
that ln(conductance) scales with the square root of the frontier
orbital position, consistent with the barrier tunnelling model of
equation 1.22
Notes and references
{ Sonogashira cross-coupling of the appropriate 1,4-dibromo-arene with
6-chloro-hex-1-yne, H2/Pd reduction of the resulting di-alkyne, and
subsequent KSAc–NaI(cat.)–acetone treatment of the 1,4-bis(6-chloro-
hexyl)arene afforded the corresponding thioacetic acid S-{6-[4-(6-acetylsul-
fanyl-hexyl)-aryl]-hexyl} esters, 1–4; full details are in the Supplementary
Information.
1 S. M. Lindsay and M. A. Ratner, Adv. Mater., 2007, 19, 23.
2 N. J. Tao, Nat. Nanotechnol., 2006, 1, 173.
3 L. Venkataraman, J. E. Klare, C. Nuckolls, M. S. Hybertsen and
M. L. Steigerwald, Nature, 2006, 442, 904.
4 J. R. Quinn, F. W. Foss, L. Venkataraman, M. S. Hybertsen and
R. Breslow, J. Am. Chem. Soc., 2007, 129, 6714.
5 L. Venkataraman, Y. S. Park, A. C. Whalley, C. Nuckolls,
M. S. Hybertsen and M. L. Steigerwald, Nano Lett., 2007, 7, 502.
6 A. V. Danilov, S. E. Kubatkin, S. G. Kafanov, K. Flensberg and
T. Bjornholm, Nano Lett., 2006, 6, 2184.
7 G. J. Ashwell, B. Urasinska, C. S. Wang, M. R. Bryce, I. Grace and
C. J. Lambert, Chem. Commun., 2006, 4706.
8 A. M. Moore, A. A. Dameron, B. A. Mantooth, R. K. Smith,
D. J. Fuchs, J. W. Ciszek, F. Maya, Y. X. Yao, J. M. Tour and
P. S. Weiss, J. Am. Chem. Soc., 2006, 128, 1959.
9 A. S. Blum, J. G. Kushmerick, D. P. Long, C. H. Patterson, J. C. Yang,
J. C. Henderson, Y. X. Yao, J. M. Tour, R. Shashidhar and B. R. Ratna,
Nat. Mater., 2005, 4, 167.
A
detailed theoretical description of the mechanism of
10 L. H. Yu, Z. K. Keane, J. W. Ciszek, L. Cheng, M. P. Stewart,
J. M. Tour and D. Natelson, Phys. Rev. Lett., 2004, 93, 266802.
11 M. Elbing, R. Ochs, M. Koentopp, M. Fischer, C. von Hanisch,
F. Weigend, F. Evers, H. B. Weber and M. Mayor, Proc. Natl. Acad.
Sci. U. S. A., 2005, 102, 8815.
12 L. L. Chang, L. Esaki and R. Tsu, Appl. Phys. Lett., 1974, 24, 593.
13 D. P. Wang, B. R. Perkins, A. J. Yin, A. Zaslavsky, J. M. Xu,
R. Beresford and G. L. Snider, Appl. Phys. Lett., 2005, 87, 152102.
14 T. Kanazawa, R. Fujii, T. Wada, Y. Suzuki, M. Watanabe and
M. Asada, Appl. Phys. Lett., 2007, 90, 092101.
conduction is outside the scope of this paper. A very recent study
of the single molecule conductances of a series of very short
conjugated molecules (substituted 1,4-diaminobenzenes) using
an STM-based break junction technique found that electron
donating substituents resulted in higher molecular conductances,
and there was a (very approximate) correlation between the
conductance and the Hammett sp parameter, consistent with
hole transport (i.e. one could regard the molecule as being in a
‘positive transition state’ as it conducts).5 Also, recent thermo-
electric effect measurements on conjugated dithiols trapped
between gold electrodes in the presence of a temperature gradient
gave positive values for the junction Seebeck coefficients,
indicating hole conduction.23 In our molecules, the presence of
the alkyl tunnelling barriers means that the aryl substituents are far
removed from the gold contacts during measurement, so that
possible complications from steric effects, and consequent
differences in contact geometry, are much less likely. It is
interesting that although the conduction mechanism may well be
different for short, conjugated molecules, the effect of substitution
is the same.
15 W. C. Liu, W. S. Lour and Y. H. Wang, IEEE Trans. Electron Devices,
1992, 39, 2214.
16 W. Haiss, H. van Zalinge, S. J. Higgins, D. Bethell, H. Ho¨benreich,
D. J. Schiffrin and R. J. Nichols, J. Am. Chem. Soc., 2003, 125, 15294.
17 W. Haiss, H. van Zalinge, H. Ho¨benreich, D. Bethell, D. J. Schiffrin,
S. J. Higgins and R. J. Nichols, Langmuir, 2004, 20, 7694.
18 W. Haiss, T. Albrecht, H. van Zalinge, S. J. Higgins, D. Bethell,
H. Ho¨benreich, D. J. Schiffrin, R. J. Nichols, A. M. Kuznetsov,
J. Zhang, Q. Chi and J. Ulstrup, J. Phys. Chem. B, 2007, 111, 6703.
19 W. Haiss, unpublished observations, 2007.
20 W. Haiss, R. J. Nichols, H. van Zalinge, S. J. Higgins, D. Bethell and
D. J. Schiffrin, Phys. Chem. Chem. Phys., 2004, 6, 4330.
21 W. Haiss, R. J. Nichols, S. J. Higgins, D. Bethell, H. Ho¨benreich and
D. J. Schiffrin, Faraday Discuss., 2004, 125, 179.
22 W. Haiss, C. S. Wang, I. Grace, C. Lambert, M. R. Bryce, S. J. Higgins
and R. J. Nichols, manuscript in preparation, 2007.
23 P. Reddy, S. Y. Jang, R. A. Segalman and A. Majumdar, Science, 2007,
315, 1568.
We thank EPSRC (grant no. EP/C00678X/1) for funding.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 3939–3941 | 3941