Conductive molecular silicon

Article abstract of DOI:10.1021/ja211677q

Bulk silicon, the bedrock of information technology, consists of the deceptively simple electronic structure of just Si-Si σ bonds. Diamond has the same lattice structure as silicon, yet the two materials have dramatically different electronic properties. Here we report the specific synthesis and electrical characterization of a class of molecules, oligosilanes, that contain strongly interacting Si-Si σ bonds, the essential components of the bulk semiconductor. We used the scanning tunneling microscope-based break-junction technique to compare the single-molecule conductance of these oligosilanes to those of alkanes. We found that the molecular conductance decreases exponentially with increasing chain length with a decay constant β = 0.27 ± 0.01 A^-1, comparable to that of a conjugated chain of C = C π bonds. This result demonstrates the profound implications of σ conjugation for the conductivity of silicon.

Full text of DOI:10.1021/ja211677q

Communication
pubs.acs.org/JACS
Conductive Molecular Silicon
Rebekka S. Klausen,† Jonathan R. Widawsky,‡ Michael L. Steigerwald,† Latha Venkataraman,*,‡
and Colin Nuckolls*,†
†Department of Chemistry, Columbia University, New York, New York 10027, United States
‡Department of Applied Physics and Applied Mathematics, Columbia University, New York, New York 10027, United States
S
* Supporting Information
to relatively inert alkyl and aryl substituents,10 and the resultant
ABSTRACT: Bulk silicon, the bedrock of information
technology, consists of the deceptively simple electronic
structure of just Si−Si σ bonds. Diamond has the same
lattice structure as silicon, yet the two materials have
dramatically different electronic properties. Here we report
the specific synthesis and electrical characterization of a
class of molecules, oligosilanes, that contain strongly
interacting Si−Si σ bonds, the essential components of the
bulk semiconductor. We used the scanning tunneling
microscope-based break-junction technique to compare
the single-molecule conductance of these oligosilanes to
those of alkanes. We found that the molecular
conductance decreases exponentially with increasing
chain length with a decay constant β = 0.27 0.01 Å−1,
comparable to that of a conjugated chain of CC π
bonds. This result demonstrates the profound implications
of σ conjugation for the conductivity of silicon.
polymers also suffer from low yields and/or wide molecular
weight distributions.11 Herein we report the synthesis of
oligomeric silanes of specific length functionalized with anchor
groups, enabling single-molecule conductance measurements.
We carried out these conductance measurements on a set of
SiMe2 oligomers terminated with 4-(methylthio)phenyl groups
(Figure 1), which bind to undercoordinated Au to form single-
ulk silicon exclusively contains Si−Si σ bonds; indeed, the
B
material can be imagined as a series of one-dimensional
chains of Si atoms cross-linked in three dimensions. It was first
shown in the middle of the 20th century that short chains of Si
atoms (oligosilanes) have markedly different electronic proper-
ties than their structural analogues, the alkanes.1 While alkanes
do not absorb light above 190 nm, Gilman et al.2 observed that
permethyloligosilanes (e.g., hexamethyldisilane and higher
oligomers) display strong absorbance at 200−300 nm. In a
trend similar to those for linearly conjugated systems such as
oligoenes,3 the position of maximum absorbance (λmax) and the
absorption coefficient (ε) increase with the length of the Si
chain.2,4 This observation led to speculation that saturated
oligosilanes may undergo electron-transfer chemistry similar to
that of conjugated carbon-based materials. Indeed, the
pioneering work of West established that the stable delocalized
decamethylcyclopentasilane radical anion can be observed by
EPR spectroscopy5 and that linear and cyclic oligosilanes form
charge-transfer complexes with π acceptors such as tetracyano-
ethylene (TCNE).6
Figure 1. (a) Molecular structures of methyl sulfide-capped
oligosilanes Si1−5 and alkanes C2−4. (b) ORTEP of a single
molecule of silane Si2. Ellipsoids are shown at the 50% probability
level. (c) B3LYP/6-31G**-calculated lowest-energy structures of
oligosilanes Si1−5. (d) Perspective drawing of the lowest-energy
conformation of pentasilane Si5, demonstrating the flexible silicon
backbone. In (b−d), H atoms have been omitted for clarity; S is
shown in yellow, Si in blue, and C in green.
molecule junctions with linear silane backbones that are stable
under ambient conditions.12,13
We synthesized the oligomeric materials via silylation of an
aryllithium reagent with symmetric α,ω-dichlorooligosilanes
5a−e (Scheme 1A). α,ω-Dichlorooligosilanes are a known class
of molecules and are typically synthesized by chlorination of
dodecamethylcyclohexasilane.14 This method yields a complex
mixture of chlorosilanes that we found to be cumbersome.
Instead, we employed an iterative synthesis of the α,ω-
dichlorooligosilanes 5c−e in which dimethylphenylsilyllithium
(3)15 was coupled to commercially available dichlorodimethyl-
silane (5a) or dichlorotetramethyldisilane (5b) to yield α,ω-
diphenyloligosilanes 4a and 4b (Scheme 1B). Protiodesilylation
with HCl/AlCl3 yielded α,ω-dichlorooligosilanes 5c and 5d,
These promising initial experiments inspired considerable
interest in the development of polymeric silanes as electronic
materials; however, the challenges in silane synthesis ultimately
hindered the widespread study of oligo- and polysilanes after
the 1980s.7−9 High-molecular-weight polymers are obtained via
sodium-mediated Wurtz-type coupling of dichlorosilane mono-
mers. The harsh conditions limit the functional group tolerance
Received: December 29, 2011
Published: February 21, 2012
© 2012 American Chemical Society
4541
dx.doi.org/10.1021/ja211677q | J. Am. Chem. Soc. 2012, 134, 4541−4544

Journal of the American Chemical Society
Communication
conductance values below G0, molecule-dependent plateaus
were observed (Figure 2A inset). Thousands of traces were
measured and used to generate two-dimensional (2D)
conductance−displacement histograms without data selec-
tion.20,21 Figure 2A shows such a 2D histogram measured for
Si1 that was generated with 27 000 traces [2D histograms for
the other compounds are shown in the Supporting Information
(SI)]. We see an intense peak around 10−4G0 that extends for a
distance of ca. 0.5 nm, indicating that single-molecule junctions
are formed reproducibly with this molecule and can be
elongated over that distance.
Scheme 1. Synthetic Scheme for α,ω-Bis(4-
methylthio)phenyloligosilanes Si1−5: (a) Arylsilylation
Reaction; (b) Synthesis of α,ω-Dichlorooligosilanes 5c−e
One-dimensional conductance histograms generated using
logarithm bins for the five silanes studied are shown in Figure
2B. A broad peak in each histogram at a molecule-dependent
conductance value was observed (black arrows). The peak
broadness is attributed to conductance variation from junction
to junction, primarily due to variations in the Au−S−C
torsional angle.22,23 The peak positions are plotted in Figure 2C
on a semilogarithmic scale against L, the effective length of the
silane (the distance between the carbons para to the methylthio
substituent). We see that the conductance decreased
exponentially with increasing L for the first four silanes in
this series (i.e., G ∼ e−βL with β = 0.27 0.01 Å−1).24 The
decay constant measured here is different than the value
measured in a study of photoinduced electron transfer in
porphyrin−silane−fullerene dyads (β = 0.16 Å−1).25 Excited-
state relaxation and low-bias molecular conduction are
fundamentally different processes. This β value compares
favorably to β values observed for conjugated olefins, which
range from 0.17−0.27 Å−1.23,26 The statistically determined
conductance of the pentasilane Si5 fell below this line, as can be
seen in Figure 2C. We attribute this small decrease in
conductance for Si5 to its slightly twisted backbone structure,27
as determined from density functional theory (DFT)
calculations (Figure 1D), which decreases the “conjugation”
of this molecule, similar to gauche defects in alkane
backbones.28 Notably, the difference in through-space length
between trans-Si5 and the twisted conformer shown in Figure
1D is insignificant.
which could then be coupled with the aryl lithiates or with
silyllithium 3 to yield pentasilane 4c.16 The arylsilylation
reaction efficiently yielded the bis(4-(methylthio)phenyl)-
silanes in 60−80% yield. These compounds are air-stable solids
that are soluble in common organic solvents, including
hydrocarbon, aromatic, and chlorinated solvents. The crystal
structure of disilane Si2 (Figure 1B) confirmed that the Si chain
adopts a fully staggered trans conformation, which is expected
to have the optimal orbital overlap for electronic communica-
tion.17,18 It is also important to note that there was no
detectable decomposition or oxidation during the break-
junction measurements.
The conductance of these oligosilanes Si1−5 was measured
using a scanning tunneling microscope-based break-junction
(STM-BJ) technique.12,19 This was carried out using a gold tip
and substrate to form and break gold point contacts repeatedly
in solutions of the target compounds (1 mM) in 1,2,4-
trichlorobenzene. The conductance (current/voltage) was
measured across the Au tip/substrate pair as a function of
the tip/substrate separation. The conductance traces showed
plateaus at integer multiples of G0, the quantum of
conductance, that correspond to multiples of Au atoms. At
Figure 2. (a) 2D conductance histogram for Si1 showing a clear conductance peak that extends over a distance of 0.5 nm relative to the break of the
G0 contact. The histogram binning is 100 per decade of conductance for the y-axis and 0.0079 nm for the x-axis. The scale bar shows points per trace.
The inset shows individual conductance traces. (b) Logarithm-binned conductance histograms generated using a bin size of 100/decade for
compounds Si1−5. (c) Conductance peak values of the single-molecule junctions for the Sin (black) and Cn (red) series as functions of the effective
molecular length (L, in Å), defined as the distance between the carbons para to the methylthio substituent. L was determined from crystal structures
Chem3D and DFT calculations. The values thus obtained varied insignificantly (<3%). The measured decay constants β were 0.27 0.01 Å−1 for
Sin and 0.68 0.05 Å−1 for Cn.
4542
dx.doi.org/10.1021/ja211677q | J. Am. Chem. Soc. 2012, 134, 4541−4544

Journal of the American Chemical Society
Communication
Figure 2B shows that in addition to a strong molecule-
dependent peak, all of these histograms had a less intense peak
at ca. 10−3G0 (blue arrows in Figure 2A,B). The 2D histogram
indicates that this peak corresponds to conductance features
occurring immediately after the single-atom contact is broken at
a short tip/sample displacement. Although the origin of this
high-conductance feature is not fully understood, control
experiments using molecules with only one terminal methyl
sulfide or no anchor groups were conducted to test whether the
molecule interacts with gold through the Si backbone, in
analogy with an oligoene potentiometer.23 Neither molecule
showed clear conductance signatures (see the SI for synthesis
and conductance histograms), demonstrating that in fact both
sulfides are required for junction formation. Another possibility
is that a cis conformation of the molecule in which the
thioanisole moieties are stacked could be formed at a short
electrode displacement, leading to a higher conductance with a
short molecular plateau length. In support of this hypothesis,
we offer the fact that conductance through the π−π stack in
cyclophanes has been recently reported29 and that this
hypothesis accounts for the observation of the same 10−3G0
peak in an analogous alkane series (see below).
These conductance results confirm the prediction first
advanced several decades ago that while silanes are structural
analogues of alkanes, their electronic properties are far more
similar to those of conjugated, unsaturated hydrocarbons.1 The
conductance of oligosilanes Si1−5 relative to that of alkanes
C2−4 inspired us to investigate electronic communication
between the terminal sulfur atoms in these molecules. We
performed DFT-based calculations (B3LYP/6-31G**) on
Si1−5 and found in each case that the highest occupied
molecular orbital (HOMO) is distributed evenly across the
silicon chain, the π systems of the arenes, and the pπ orbitals of
the two sulfur atoms.30 The HOMO of Si4 is depicted in Figure
3 as a representative example. The contrast with the carbon
analogue C4 is dramatic: there is essentially no electron density
along the C−C backbone, and the HOMO is localized
exclusively on the thioanisole fragments. In contrast, the
HOMO of a butadiene terminated with 4-(methylthio)phenyl
groups (6) is fully delocalized, as in the silane, indicating that
the two sulfur pπ orbitals couple much more strongly across the
Si−Si and CC bonds than the C−C bond.
To elucidate the greater coupling observed in the Si−Si and
CC materials, we separated the molecular conductors into
their constituent components: the 4-(methylthio)phenyl end
groups and the conducting oligomer.31,27b Computational
analysis of each part was conducted (see Table S3 in the SI).
We found that the HOMOs of tetrasilane and butadiene are
delocalized over the respective C and Si atoms. The HOMO
level of 4-methylthioanisole lies at −5.49 eV with respect to
vacuum, and it is a close combination of the pπ lone pair on S
and the π space on the ring. We found that of the three
oligomers investigated, the tetrasilane (HOMO = −5.96 eV)
and butadiene (HOMO = −5.60 eV) lie closest in energy to
that of the 4-methylthioanisole, suggesting that extensive
mixing of the thioanisole and Si−Si orbitals [σ(Si−Si)−π
conjugation] occurs; moreover, Figure 3 demonstrates that the
orbital mixing is geometrically accessible.32 The HOMO of
hexane lies at the substantially lower energy of −8.29 eV,
resulting in very minimal mixing of the σ framework and the
thionanisole orbitals [negligible σ(C−C)−π conjugation]. As a
result, the HOMO of C4 is localized on the thioanisole. Our
results indicate that uniform delocalization of the HOMO over
the entire molecule improves the conductance. Consistent with
this observation, a siloxane variant of Si2 was synthesized in
which the Si chain is disrupted with an O atom; no
conductance peak was observed (see the SI for the synthesis,
conductance histogram, and HOMO calculation). We also note
in passing that differential orbital coupling may account for the
difference in contact resistance observed for alkanes and
oligosilanes, as seen in Figure 2C.
We have presented a quantitative comparison of the
conductive properties of oligosilanes and alkanes that was
enabled by straightforward synthetic access to silanes
functionalized with aurophilic contact groups. We have found
that saturated Si nanowires display a very slow decay of
conductance with length; the decay constant is comparable to
that of polyacetylene fragments (oligoenes) and considerably
less than those of alkane structural analogues. These results,
coupled to our theoretical studies, highlight the important role
of σ-bond conjugation in charge transport through Si−Si σ
bonds. Currently the application of Si oligomers and polymers
in electronic devices is limited,8 while applications of
conjugated carbon-based materials such as polyacetylene are
ubiquitous.33 The materials described herein have promising
physical properties that are complementary to those of rigid,
conjugated systems, including conformational flexibility, air and
thermal stability, and high solubility in common organic
solvents. We expect that the desirable physical properties and
high conductance of the materials described herein will result in
the development of novel molecular devices and nanoscale
architectures.
ASSOCIATED CONTENT
* Supporting Information
■
S
Synthetic methods and procedures, characterization data,
electrochemical and spectroscopic characterization, computa-
tional data, and crystallographic data (CIF). This material is
available free of charge via the Internet at http://pubs.acs.org.
Figure 3. B3LYP/6-31G**-calculated HOMOs of (a) tetrasilane Si4,
(b) butane C4, and (c) 1,4-bis(4-(methylthio)phenyl)buta-1,3-diene
(6).
AUTHOR INFORMATION
Corresponding Author
lv2117@columbia.edu; cn37@columbia.edu
■
4543
dx.doi.org/10.1021/ja211677q | J. Am. Chem. Soc. 2012, 134, 4541−4544

Journal of the American Chemical Society
Communication
(23) Meisner, J. S.; Kamenetska, M.; Krikorian, M.; Steigerwald, M.
L.; Venkataraman, L.; Nuckolls, C. Nano Lett. 2011, 11, 1575.
(24) (a) Griffiths, D. J. Introduction to Quantum Mechanics, 2nd ed.;
Pearson Prentice Hall: Upper Saddle River, NJ, 2005; Chapter 8.
(b) Salomon, A.; Cahen, D.; Lindsay, S.; Tomfohr, J.; Engelkes, V. B.;
Frisbie, C. D. Adv. Mater. 2003, 15, 1881.
(25) Sasaki, M.; Shibano, Y.; Tsuji, H.; Araki, Y.; Tamao, K.; Ito, O. J.
Phys. Chem. A 2007, 111, 2973.
(26) Visoly-Fisher, I.; Daie, K.; Terazono, Y.; Herrero, C.; Fungo, F.;
Otero, L.; Durantini, E.; Silber, J. J.; Sereno, L.; Gust, D.; Moore, T. A.;
Moore, A. L.; Lindsay, S. M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
8686.
(27) (a) Albinsson, B.; Antic, D.; Neumann, F.; Michl, J. J. Phys.
Chem. A 1999, 103, 2184. (b) Schepers, T.; Michl, J. J. Phys. Org.
Chem. 2002, 15, 490.
(28) Jones, D. R.; Troisi, A. J. Phys. Chem. C 2007, 111, 14567.
(29) Schneebeli, S. T.; Kamenetska, M.; Cheng, Z. L.; Skouta, R.;
Friesner, R. A.; Venkataraman, L.; Breslow, R. J. Am. Chem. Soc. 2011,
133, 2136.
(30) Pitt, C. G.; Bock, H. J. Chem. Soc., Chem. Commun. 1972, 28.
(31) Bock, H.; Ensslin, W. Angew. Chem., Int. Ed. Engl. 1971, 10, 404.
(32) Pitt, C. G.; Toren, E. C.; Carey, R. N. J. Am. Chem. Soc. 1972,
94, 3806.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
STM-BJ experiments were supported as part of the Center for
Re-Defining Photovoltaic Efficiency Through Molecular-Scale
Control, an Energy Frontier Research Center funded by the
U.S. Department of Energy (DOE), Office of Science, Office of
Basic Energy Sciences under Award DE-SC0001085. L.V.
thanks the Packard Foundation for support. We acknowledge
Yi Rong and Ged Parkin (Columbia University) for single-
crystal X-ray diffraction of compound Si2. The National
Science Foundation (CHE-0619638) is thanked for acquisition
of an X-ray diffractometer. We acknowledge Roger A.
Lalancette (Department of Chemistry, Rutgers University) for
the Si4 X-ray data and solution of the structure. We also
acknowledge support by NSF-CRIF Grant 0443538 for part of
the purchase of the X-ray diffractometer.
REFERENCES
■
(1) West, R.; Carberry, E. Science 1975, 189, 179.
(2) Gilman, H.; Atwell, W. H.; Schwebke, G. L. J. Organomet. Chem.
1964, 2, 369.
(33) Heeger, A. J. Chem. Soc. Rev. 2010, 39, 2354.
(3) Hausser, K. W.; Kuhn, R.; Smakula, A.; Hoffer, M. Z. Phys. Chem.,
Abt. B 1935, 29, 371.
(4) Gilman, H.; Morris, P. J. J. Organomet. Chem. 1966, 6, 102.
(5) Carberry, E.; West, R.; Glass, G. E. J. Am. Chem. Soc. 1969, 91,
5446.
(6) Traven, V. F.; West, R. J. Am. Chem. Soc. 1973, 95, 6824.
(7) Miller, R. D.; Michl, J. Chem. Rev. 1989, 89, 1359.
(8) Feigl, A.; Bockholt, A.; Weis, J.; Rieger, B. Adv. Polym. Sci. 2011,
235, 1.
(9) David, L. D.; Djurovich, P. I.; Stearley, K. L.; Srinivasan, K. S. V. J.
Am. Chem. Soc. 1981, 103, 7352.
(10) Jones, R. G.; Holder, S. J. Polym. Int. 2006, 55, 711.
(11) Benfield, R. E.; Cragg, R. H.; Swain, A. C. J. Chem. Soc., Chem.
Commun. 1992, 112.
(12) (a) Venkataraman, L.; Klare, J. E.; Nuckolls, C.; Hybertsen, M.
S.; Steigerwald, M. L. Nature 2006, 442, 904. (b) Park, Y. S.; Whalley,
A. C.; Kamenetska, M.; Steigerwald, M. L.; Hybertsen, M. S.; Nuckolls,
C.; Venkataraman, L. J. Am. Chem. Soc. 2007, 129, 15768.
(13) Xu, B. Q.; Tao, N. J. J. Science 2003, 301, 1221.
(14) (a) Sen, P. K.; Ballard, D.; Gilman, H. J. Organomet. Chem.
1968, 15, 237. (b) Chernyavskii, A. I.; Yu, D.; Larkin, D. Y.;
Chernyavskaya, N. A. J. Organomet. Chem. 2003, 679, 17. (c) Gilman,
H.; Inoue, S. J. Org. Chem. 1964, 29, 3418.
(15) (a) George, M. V.; Peterson, D. J.; Gilman, H. J. Am. Chem. Soc.
1960, 82, 403. (b) Fleming, I.; Roberts, R. S.; Smith, S. C. J. Chem.
Soc., Perkin Trans. 1 1998, 1209. (c) Shibano, Y.; Sasaki, M.; Tsuji, H.;
Araki, Y.; Ito, O.; Tamao, K. J. Organomet. Chem. 2007, 692, 356.
(16) Fleming, I.; Henning, R.; Parker, D. C.; Plaut, H. E.; Sanderson,
P. E. J. J. Chem. Soc., Perkin Trans. 1 1995, 317.
(17) Tamao, K.; Tsuji, H.; Terada, M.; Asahara, M.; Yamaguchi, S.;
Toshimitsu, A. Angew. Chem., Int. Ed. 2000, 39, 3287.
(18) George, C. B.; Ratner, M. A.; Lambert, J. B. J. Phys. Chem. A
2009, 113, 3876.
(19) Venkataraman, L.; Klare, J. E.; Tam, I. W.; Nuckolls, C.;
Hybertsen, M. S.; Steigerwald, M. L. Nano Lett. 2006, 6, 458.
(20) Martin, C. A.; Ding, D.; Sorensen, J. K.; Bjornholm, T.; van
Ruitenbeek, J. M.; van der Zant, H. S. J. J. Am. Chem. Soc. 2008, 130,
13198.
(21) Quek, S. Y.; Kamenetska, M.; Steigerwald, M. L.; Choi, H. J.;
Louie, S. G.; Hybertsen, M. S.; Neaton, J. B.; Venkataraman, L. Nat.
Nanotechnol. 2009, 4, 230.
(22) Park, Y. S.; Widawsky, J. R.; Kamenetska, M.; Steigerwald, M. L.;
Hybertsen, M. S.; Nuckolls, C.; Venkataraman, L. J. Am. Chem. Soc.
2009, 131, 10820.
4544
dx.doi.org/10.1021/ja211677q | J. Am. Chem. Soc. 2012, 134, 4541−4544