Single-Molecule Conductance in Atomically Precise Germanium Wires

Article abstract of DOI:10.1021/jacs.5b08155

While the electrical conductivity of bulk-scale group 14 materials such as diamond carbon, silicon, and germanium is well understood, there is a gap in knowledge regarding the conductivity of these materials at the nano and molecular scales. Filling this gap is important because integrated circuits have shrunk so far that their active regions, which rely so heavily on silicon and germanium, begin to resemble ornate molecules rather than extended solids. Here we unveil a new approach for synthesizing atomically discrete wires of germanium and present the first conductance measurements of molecular germanium using a scanning tunneling microscope-based break-junction (STM-BJ) technique. Our findings show that germanium and silicon wires are nearly identical in conductivity at the molecular scale, and that both are much more conductive than aliphatic carbon. We demonstrate that the strong donor ability of C-Ge I-bonds can be used to raise the energy of the anchor lone pair and increase conductance. Furthermore, the oligogermane wires behave as conductance switches that function through stereoelectronic logic. These devices can be trained to operate with a higher switching factor by repeatedly compressing and elongating the molecular junction.

Full text of DOI:10.1021/jacs.5b08155

Article
pubs.acs.org/JACS
Single-Molecule Conductance in Atomically Precise Germanium
Wires
Timothy A. Su,†,§ Haixing Li,‡,§ Vivian Zhang,† Madhav Neupane,† Arunabh Batra,‡
Rebekka S. Klausen,†,⊥ Bharat Kumar,†,∥ 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
ABSTRACT: While the electrical conductivity of bulk-scale
group 14 materials such as diamond carbon, silicon, and
germanium is well understood, there is a gap in knowledge
regarding the conductivity of these materials at the nano and
molecular scales. Filling this gap is important because
integrated circuits have shrunk so far that their active regions,
which rely so heavily on silicon and germanium, begin to
resemble ornate molecules rather than extended solids. Here
we unveil a new approach for synthesizing atomically discrete
wires of germanium and present the first conductance
measurements of molecular germanium using a scanning tunneling microscope-based break-junction (STM-BJ) technique.
Our findings show that germanium and silicon wires are nearly identical in conductivity at the molecular scale, and that both are
much more conductive than aliphatic carbon. We demonstrate that the strong donor ability of C−Ge σ-bonds can be used to
raise the energy of the anchor lone pair and increase conductance. Furthermore, the oligogermane wires behave as conductance
switches that function through stereoelectronic logic. These devices can be trained to operate with a higher switching factor by
repeatedly compressing and elongating the molecular junction.
However, adding an additional Ge−Ge σ-bond in the
germanium wires has essentially the same effect on the
magnitude of conductance as adding an additional Si−Si σ-
bond in silicon wires. The Au−Gen−Au junction switches to a
high-conductance (G) state only when both terminal Me−S−
CH2−GeMe2− dihedral angles are twisted into ortho
configurations, suggesting a possible route to an “AND” logic
gate that operates by means of a stereoelectronic effect.
Furthermore, we find that the repeated compression−
elongation of the Au−Gen−Au junction causes the device to
switch with a higher conductance magnitude with each
successive cycle; these cycles effectively mold the electrodes
into an optimal morphology for mechanically manipulating the
geometry of the molecule in the junction.
INTRODUCTION
■
Here we describe both the first deterministic synthesis of long,
linear permethyloligogermanes and the first single-molecule
conductance measurements on these atomically precise strands
of germanium (Figure 1). An understanding of charge transport
in atomically precise, molecular-scale germanium will inform
the transport issues and opportunities in ever-smaller silicon−
germanium integrated circuits.1 Moreover, nanoscale forms of
germanium such as germanene,2,3 germanium quantum dots,4
and germanium nanowires5−9 are emerging as new materials
with interesting electronic properties.
RESULTS AND DISCUSSION
■
Synthesis. In order to study the conductance properties of
molecular germanium, we needed to develop a new synthesis
that was robust and produced linear-chain oligogermane wires
in a stepwise fashion. Though germanium oligomers have been
known since 1925,10 methods for the facile synthesis and
isolation of long (n > 6) linear oligogermanes have not been
The synthesis that is developed here allows us to easily
functionalize these germanium wires with aurophilic methyl-
thiomethyl groups on their termini to allow connection to gold
electrodes. Using the scanning tunneling microscope-based
break-junction (STM-BJ) method, we find that each Gen
oligomer is more conductive than its Sin isostructure due to the
enhanced interaction in the contact group between the sulfur
lone pair and the more strongly donating C−Ge σ-bond.
Received: August 3, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. (a) STM-BJ experimental diagram of Au−Ge10−Au junction. (b) DFT-optimized structures for the Ge1−Ge10 series with backbone
dihedrals held at 180°. Ge, green; C, gray; S, yellow; H omitted for clarity.
reported.11−15 Wurtz coupling between Me3GeCl and
Me2GeCl2 has been used to synthesize long oligogermanes
(≤Ge10Me22), but these reactions give low-yielding mixtures
that are difficult to separate and products that are not
functionalized.16
Scheme 1. Iterative Synthesis of Permethyloligogermanes
Terminated with Methylthiomethyl End Groups
a
Here we introduce a new method for synthesizing an entire
series of permethyloligogermanes (from n = 1−10) that are
terminated with methylthiomethyl contacts in high yields with
easy purification. Our approach was inspired by the iterative
synthesis of α,ω-diphenylpermethyloligogermanes (Ph-
[GeMe2]n-Ph) for n = 1−5 pioneered by Kumada and co-
workers.17 They observed low yields for this reaction due to the
tendency of Ge−Ge bonds to disproportionate in the presence
of strong nucleophiles such as dimethylphenylgermyl lithium.
We were able to prevent oligogermane disproportionation by
attenuating the strength of the nucleophile via transmetalation
with MgBr2, thereby accessing high yields (62−87%) of the
longer oligomers. This simple adaptation has enabled us to
achieve the first deterministic synthesis of germanium
oligomers with n > 6. Using this new approach we have
synthesized the longest linear oligogermane (Ge11Me22Ph2)
reported to date (see Supporting Information (SI) for details).
Scheme 1 illustrates our iterative approach to synthesizing
the [GeMe2]n oligogermanes (n = 1−10) terminated by
electrode-binding methylthiomethyl end groups (Ge1−Ge10,
Figure 1). We reduce chlorodimethylphenylgermane with
lithium metal17 and then transmetalate with MgBr2 to access
dimethylphenylgermyl magnesium bromide 1. We grow α,ω-
diphenyloligogermane 2 outward, two germanium subunits at a
time, by treating α,ω-dichlorooligogermane 3 with 2 equiv of 1.
After purification of 2, protiodegermylation of α,ω-diphenyl-
oligogermane 2 under acidic conditions furnishes the chain-
extended α,ω-dichlorooligogermane 3.18 We functionalize α,ω-
dichlorooligogermane 3 with electrode contacts (CH2SMe) by
reacting 3 with methylthiomethyl lithium19 to furnish the final
α,ω-bis(methylthiomethyl)oligogermanes Ge1−Ge10.
a
The n+2 oligomer growth cycle is depicted in blue. Reagents and
conditions: (a) i. Li, THF; ii. MgBr2, THF. (b) THF, 0 °C; for n = 3−
10, yields are 62−87%. (c) i. CF3SO3H, CH2Cl2; ii. Et3N·HCl, Et2O;
for n = 3−10, yields are 65−94%. (d) n-BuLi, TMEDA, CH3SCH3,
THF; for n = 1−10, yields are 40−84%.
formed by repeatedly breaking and forming point contacts
between Au tip and substrate electrodes in a dilute (0.10−1.00
mM) solution of the oligogermane in 1,2,4-trichlorobenzene
under ambient conditions. After the Au−Au point contact is
broken, the aurophilic thiomethyl groups21 on the oligo-
germane bind the electrodes to form a Au−Gen−Au junction.
Conductance is measured across the gap as a function of tip−
substrate displacement, and the resulting traces reveal
molecule-dependent plateaus signifying junction formation
with conductance values below G0 (2e2/h), the quantum of
conductance describing a single Au−Au atomic contact.22 The
junction breaks once the distance between the electrodes
becomes too large for the molecule to bridge this gap. We form
and break thousands of molecular junctions and analyze all
measured traces using logarithm-binned one-dimensional (1D)
and two-dimensional (2D) histograms. 1D histograms provide
a distribution of all measured conductance values from all
traces; 2D histograms sum all conductance values while
retaining relative displacement information.23
Conductance Measurements. We measured the single-
molecule conductance of oligogermanes Ge1−Ge10 with the
STM-BJ technique.20 In brief, Au−Gen−Au junctions are
The 1D histograms for the conductance measurements of
Ge1−Ge10 are shown in Figure 2a. The peak positions
demonstrate a clear exponential decrease in conductance as the
B
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. (a) Logarithm-binned 1D conductance histograms of Ge1−Ge10, each comprising thousands of measurement traces. (b) Conductance
peak positions from the simple elongation measurements of C4−C8, Si1−Si10, and Ge1−Ge10 (panel a) plotted against effective molecular length
(L, in Å) give the decay constant β. βC = 0.74 0.03 Å−1, βSi = 0.39 0.01 Å−1, and βGe = 0.36 0.01 Å−1.
molecular length increases from n = 1−10. We obtain β for the
Gen series in Figure 2b by plotting the conductance peaks from
Figure 2a against the effective molecular length (L) on a
semilog plot and fitting a line through these points, with G =
R−1 e−βL. For every molecule described here, L is defined as the
distance between the two distal methylene carbons in the
density functional theory (DFT)-optimized structures, R is an
effective contact resistance, and β is the conductance decay
parameter. β describes the extent to which conductance in a
given oligomeric material decreases as the number of repeat
units increases, and depends strongly on the extent by
electronic delocalization (conjugation) in the backbone. We
find that conductances from Ge1−Ge10 follow an exponential
here. Drenth and co-workers characterized the ultraviolet
absorption in a series of peralkylated silicon and germanium
oligomers (n = 2−6).36 By measuring the transition energy as a
function of n, they elucidated the resonance stabilization energy
between adjacent atoms and ultimately demonstrated that
silane and germane oligomers share the same strength of
conjugation between their σ-bonding orbitals. Boberski et al.
and Okano et al. studied σ-conjugation by probing the
electrochemical oxidation potentials of permethylated silicon
and germanium oligomers.37,38 The data from these two studies
demonstrated that the first oxidation potential decreased by
roughly the same extent in silanes and germanes with increasing
n. We find the same to be true for the molecules studied here
(Figure S1). Importantly, these single-molecule measurements
demonstrate that the numerous studies devoted to under-
standing the nature of orbital interactions can serve as a
predictive tool for designing molecular electronic components.
Though the Gen and Sin series are near-identical in β value,
the conductance of each germanium oligomer is ∼1.5 times
higher than that of its silicon counterpart; this difference in
conductance arises from subtle differences in the way H2C−
GeMe2 and H2C−SiMe2 σ-bonds interact with the S pπ lone
pair at each termini. C−Ge σ-bonds are more electron donating
than C−Si σ-bonds because they are higher in energy and more
diffuse.39−44 The molecular orbital most relevant to con-
ductance in these systems is the HOMO, which features strong
S lone pair orbital character.30,45−48 The destabilizing
interaction between the filled S lone pair and the more
strongly donating C−Ge σ-bond raises the HOMO energy
closer toward the Au Fermi level, thereby increasing
conductance.45,47,49
This line of reasoning is consistent with our DFT
calculations of Si1 and Ge1 (see SI for computational details).
Ge1 (−5.49 eV) possesses a higher HOMO energy than Si1
(−5.56 eV) in conformations where the S lone pair and C−
Si(Ge) σ-bond are coplanar. Our calculations are supported by
Glass et al., who demonstrated that the ionization energy of
organosulfides is inversely proportional to the donor ability of
the σ-bond in the β-position relative to the S atom.50 We also
find evidence for these destabilizing effects in our estimations of
the HOMO energy from cyclic voltammetry studies on the Sin
and Gen series (Figure S1). As n increases, the HOMO
energies in both Gen and Sin rise with a similar slope but are
offset by a roughly constant energy value. These results
demonstrate that we can utilize the differential donor ability of
σ-bonds to tune the energetics of the linker lone pair, and
therefore, conductance.
in L with βGe = 0.36
0.01 Å−1 (0.74
0.02 n−1). This
signifies that charge transport in molecular germanium (up to
Ge10) occurs through a coherent tunneling mechanism.24
Figure S1 plots the conductances of all molecules against n.
We find that the decay constant of oligogermanes is very
similar to that of oligosilanes (βSi = 0.39 0.01 Å−1, 0.75
0.01 n−1) and oligo-p-phenylenes25 but much shallower than
that of alkanes (βC = 0.74 0.03 Å−1, 0.94 0.05 n−1). Our
experimental findings are consistent with calculations of
Matsuura on permethylated n = 2−6 oligomers of carbon,
silicon, and germanium with −CH2SH linker end groups.26
Matsuura calculated βC = 0.71 Å−1 for alkanes, βSi = 0.31 Å−1
for oligosilanes, and βGe = 0.34 Å−1 for oligogermanes. The β
values determined from these computations are remarkably
close to what we observe experimentally.
Because the repeat units in the alkane, silane, and germane
oligomers are single atomic units, β can serve as an index for
comparing the intrinsic electrical σ-conductivity of the group 14
elements at the single-bond level. The near-identical β values
for silanes and germanes suggest that Si−Si and Ge−Ge σ-
bonds are similar in charge transport ability at the molecular
scale, and that both are much more conductive than C−C σ-
bonds. This is perhaps not surprising based on the relationship
between conductance and conjugation or through the
molecule.27 Alkanes demonstrate a high β value because of
the weak delocalization in the C−C σ-backbone.28 We
previously established that the shallow β value29,30 observed
for oligosilanes is related to the strong conjugation in the Si−Si
bonds, which are much larger and more strongly interacting
compared to C−C σ-bond orbitals.31−35 Since oligosilanes and
oligogermanes are known to display similar degrees of σ-
conjugation based on the spectroscopic and electrochemical
studies of group 14 oligomers from the past several decades, it
is not surprising that they also show near-identical β values
C
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
We also note that there is an appreciable odd−even effect on
conductance that exists for the Gen series in Figure 2b, where
the even-numbered oligomers fall below the decay line (error is
within the size of the marker). The odd−even effect is also
manifested in the peak shapes of the odd and even oligomers
the odd oligomers are much sharper in conductance width than
the even ones (Figures S3 and S4). This alternant trend seems
to occur in the Sin series as well, although with a smaller
magnitude. These odd−even effects may stem from the strong
vicinal hyperconjugation in germanes, and are the subject of
further study.51
Stereoelectronic Switching. 2D histogram analysis
(Figure S5) and tunnel coupling calculations (Figure S6,
details in SI) suggest that each Gen oligomer acts as a
mechanically triggered switch that operates with a stereo-
electronic mechanism analogous to what was previously
described for Si1−Si10.30 In brief, elongating the Au−Gen−
Au junction changes the lowest energy configuration of the two
terminal Me−S−CH2−GeMe2− dihedrals from the sterically
favored anti (A) conformation to the mechanically favored
ortho (O) conformation (Figure 3). The conductance-switching
the σ-conjugation in the oligogermane backbone is slightly
stronger than in the oligosilane backbone, since the dihedral
angles are closer to anti. This would account for the lower β
value that we observe experimentally for the Gen series and the
higher tunnel coupling ratio that we calculate for the high- and
low-G states in oligogermanes versus oligosilanes (Figure S6).
Following this line of reasoning, we predict that structurally
constraining the σ-backbone to ω = 180° in group 14 wires may
enable even lower β values and higher switching factors to be
observed. Tamao and co-workers have synthesized such
conformationally locked oligosilanes and have demonstrated
increased σ-conjugation in the context of charge transfer57 and
absorption58 studies.
Our tunnel coupling calculations also demonstrate that, in
principle, the Gen switches are capable of behaving as AND
logic gates that follow the truth table in Figure 3, where the
high-G state is only observed under the condition that both
terminal dihedral angles are in the ortho state. We have not yet
experimentally realized such a logic device because we can only
indirectly set the dihedral configuration by regulating the inter-
electrode distance; developing a precise method for controlling
each dihedral input independently will enable the creation of a
true stereoelectronic logic device.
Junction Training. In the simple elongation experiments
described above, we continually widen the electrode gap until
the molecular junction breaks. For all Gen oligomers,
conductance switches from a low- to high-G state (Figure 4a)
in the final ∼2 Å of elongation with a factor that varies slightly
from molecule to molecule (Figures S5 and S7). For instance,
Au−Ge5−Au junctions demonstrate a switching factor
Figure 3. (a) Molecular structure of Ge6 with two terminal Me−S−
CH2−GeMe2− dihedrals ω1 and ω2. (b) Newman projections of the
mechanically favored ortho (left) and sterically favored anti (right)
dihedral configurations. (c) Standard representation of an AND logic
gate with ω1 and ω2 as inputs for stereoelectronic switching. (d) Truth
table for observing either the low-conducting (L) or high-conducting
(H) state based on tunnel coupling calculations.
event likely embodies a shift in molecular geometry from the
low-conducting A-A and O-A configurations to the high-
conducting O-O configuration. The conjugation in the Gen σ-
backbone should enable this mechanism of switching because it
electronically couples the two terminal Me−S−CH2−GeMe2−
dihedrals.
George et al. modeled the relationship between conforma-
tion and conductance in hydrogenated oligosilanes and found
that conductance (and σ-conjugation) is maximized when the
Si−Si−Si−Si dihedrals have an anti (ω = 180°) geometry.52
Michl and co-workers53,54 demonstrated that permethyloligo-
silanes settle into transoid (ω = 160−175°)55 rather than anti
minimum energy geometries due to steric effects from the
methyl groups.56 In our previous calculations on the Au−Si4−
Au system, we found that the internal Si−Si−Si−Si backbone
geometry maintains a relatively constant dihedral angle that
averages around ω = 168°. Here we find that the Ge−Ge−Ge−
Ge dihedral in the Au−Ge4−Au system remains relatively
constant with an average of ω = 172° (Table S4). We
hypothesize that the tetragermane dihedral is closer to anti than
the tetrasilane because the Ge−Ge bonds (2.51 Å) are longer
than the Si−Si bonds (2.38 Å), which would reduce the steric
repulsion from the methyl groups. These results suggest that
Figure 4. (a) Individual traces of Ge5 demonstrate switching from low
to high conductance as the molecular junction is elongated. (b) 2D
histogram of five successive 2 Å compression−elongation cycles for
the Au−Ge5−Au junction measured using a modified piezo ramp
(black line). Switching is induced between low- and high-G states
when the junction is pushed and pulled consecutively. The 2D
histogram is constructed from all traces sustaining a junction over all
ten hold sections. (c) 1D histograms compiled from each hold period
after compressions (gray lines) and hold periods after elongations (red
lines) from panel b. (d) The conductance peak positions from the 1D
histograms in panel c plotted against the hold section. The ratio of
conductance between the compression (low G) and elongation (high
G) peaks gives the switching factor for a single cycle. The switching
factors upon compression from cycles 1−5 are 1.6, 2.3, 2.9, 3.3, and
3.4, respectively.
D
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(defined as the ratio of the conductance of the high-G state to
that of the low-G state) of 1.6 upon the initial elongation
(Figure S7). Here we demonstrate that we can increase the
switching factor beyond this initial ratio by repeatedly
compressing and elongating the molecular junction with a
modified piezo ramp (Figure 4b).
situations where direct contact between the electrodes is
undesirable.
CONCLUSION
■
In this study we developed a new method for synthesizing long
(>2 nm) molecular germanium wires with atomic precision that
utilizes the attenuated nucleophilicity of germylmagnesium
species. To demonstrate the utility of this system, we
performed the first single-molecule conductance measurements
on atomically defined molecular germanium wires. As in the
silicon series, the germanium series behave as stereoelectronic
switches activated by stretching or compressing our junction.
Consecutive compression−elongation cycles train the molec-
ular junction to both exhibit a higher switching factor and
distinguish molecular geometries of disparate electronic
character with more accuracy. Each Gen oligomer demonstrates
a heightened conductance relative to its Sin congener due to
the stronger destabilizing interaction between the S lone pair
and methylene−Ge σ-bond in the contact group. And yet,
adding an extra Ge−Ge σ-bond to the Gen series has essentially
the same effect on conductance magnitude as adding an extra
Si−Si σ-bond to the Sin series.
Instead of elongating the junction to rupture after we fully
extend the Au−Gen−Au junction, we perform push−pull cycles
by first compressing the electrode gap by 2 Å, holding the
electrodes fixed for 50 ms, widening the electrode gap by 2 Å,
then holding for 50 ms. Figure 4b shows a 2D histogram
describing five consecutive push−pull cycles for the Au−Ge5−
Au junction. Figure 4c compiles the conductance measured
during each hold section into 1D histograms; the resulting
conductance peak values are plotted in Figure 4d. Figure 4d
shows that the first compression gives a switching factor of 1.6,
and the final compression gives a switching factor of 3.4. Each
successive push−pull cycle increases the switching factor and
sharpness of switching, training the junction to perform
increasingly better as a switch. We also find that the final
switching factor for Ge5 (3.4) is larger than for Si6 (2.5)
(Figure S8); this difference is likely related to the slightly
stronger σ-conjugation in the germane backbone that we
discussed in the subsection above.
The conductance trends in Figure 4 demonstrate that this
junction training effect arises from the decreasing conductance
of the low-G state. In each successive cycle, the elongation
(high-G) conductance peak value remains constant whereas the
compression (low-G) peak value decreases steadily (Figure 4d).
Figure 4c reveals that in the first cycle, the compressed low-G
state is broad in conductance and features a significant amount
of high-G character. Every subsequent cycle narrows the
conductance distribution by decreasing the amount of high-G
character in the compressed low-G state; this in turn decreases
the conductance of the low-G peak and maximizes the
switching factor with each additional cycle.
As we described before, fully stretching the molecular
junction imposes mechanical strain on the system.30 The data
here suggest that, for initial compression events, the system
frequently relaxes by reorganizing the atomic electrode
environment while sustaining the high-conducting O-O
junction geometry. Repeatedly stretching and compressing
the junction likely shapes the malleable electrodes into a
structure where electrode surface reorganization is no longer
the dominant relaxation pathway; instead, the system
compensates for the changing interelectrode distance by
twisting the junction’s terminal dihedrals into a shorter, less
conductive geometry. These cycles mechanically anneal59 the
electrodes into an optimal morphology for stereoelectronic
switching and enable the junction to distinguish molecular
conformations with more clarity.
Trouwborst et al. previously reported a similar junction
training technique to create reproducible Au−Au point contacts
from disordered electrode environments.60 In that study, the
authors organized the tip atoms by performing consecutive 1−2
Å compression−elongation cycles near the point of contact
between two Au atoms at cryogenic temperatures, and in this
way allowed the electrode atoms to probe different geometries
and find the most stable configuration. It seems likely that we
are organizing the electrodes here in much of the same way at
room temperature, but are doing so with a molecular tether;
this method might then serve as a valuable approach to
organizing the atomic arrangement of the electrode surface in
More broadly, these results show that Si and Ge molecular
wires possess essentially the same length-dependent con-
ductivity due to the similar extent of σ-conjugation in these
systems. The intrinsic electrical conductivity of bulk silicon is
the same order of magnitude as bulk germanium at room
temperature, and both are many orders of magnitude more
conductive than bulk diamond.61 The periodic trends in
molecular conductivity that we observe here therefore mirror
group 14 conductivity trends in solid-state materials; this is not
the case for stochastically grown Si and Ge nanowires where
reported conductivities vary by many orders of magnitude due
to a number of extrinsic factors.6−9 We envision that the atomic
precision in the molecular wires here will enable their use as
reliable platforms for studying effects such as strain and doping
in electronic materials, to not only probe the fundamental
nature of these effects in bulk systems but also to inform the
design of next-generation electronic circuit materials.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b08155.
Supplemental figures, synthetic procedures, character-
ization of compounds, STM-BJ measurement details, and
computational details (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
*mls2064@columbia.edu
*lv2117@columbia.edu
*cn37@columbia.edu
Present Addresses
⊥R.S.K.: Department of Chemistry, The Johns Hopkins
University, Baltimore, MD 21218, USA
∥B.K.: IBM Watson Research Center, Physical Sciences, 1101
Kitchawan Rd., Route 134, P.O. Box 218, Yorktown Heights,
NY 10598, USA
Author Contributions
§T.A.S. and H.L. contributed equally to this work.
E
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(30) Su, T. A.; Li, H.; Steigerwald, M. L.; Venkataraman, L.;
Nuckolls, C. Nat. Chem. 2015, 7, 215.
(31) Miller, R. D.; Michl, J. Chem. Rev. 1989, 89, 1359.
(32) Tsuji, H.; Terada, M.; Toshimitsu, A.; Tamao, K. J. Am. Chem.
Soc. 2003, 125, 7486.
(33) Tamao, K.; Tsuji, H.; Terada, M.; Asahara, M.; Yamaguchi, S.;
Toshimitsu, A. Angew. Chem. 2000, 112, 3425.
(34) Imhof, R.; Antic, D.; David, D. E.; Michl, J. J. Phys. Chem. A
1997, 101, 4579.
(35) Tsuji, H.; Michl, J.; Tamao, K. J. Organomet. Chem. 2003, 685, 9.
(36) Drenth, W.; Noltes, J. G.; Bulten, E. J.; Creemers, H. M. J. C. J.
Organomet. Chem. 1969, 17, 173.
(37) Boberski, W. G.; Allred, A. L. J. Organomet. Chem. 1975, 88, 65.
(38) Okano, M.; Mochida, K. Chem. Lett. 1990, 701.
(39) Wierschke, S. G.; Chandrasekhar, J.; Jorgensen, W. L. J. Am.
Chem. Soc. 1985, 107, 1496.
(40) Lambert, J. B.; Zhao, Y.; Emblidge, R. W.; Salvador, L. a.; Liu,
X.; So, J.-H.; Chelius, E. C. Acc. Chem. Res. 1999, 32, 183.
(41) Nguyen, K. A.; Gordon, M. S.; Wang, G. T.; Lambert, J. B.
Organometallics 1991, 10, 2798.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank James L. Leighton for insightful discussions and
Brandon Fowler for mass spectrometry characterization. T.A.S.
is supported by the NSF Graduate Research Fellowship under
grant no. 11-44155. H.L. is supported by the Semiconductor
Research Corporation and New York CAIST program. We
thank the NSF for the support of these studies under grant no.
CHE-1404922.
REFERENCES
■
(1) Markoff, J. IBM Discloses Working Version of a Much Higher-
Capacity Chip. New York Times 2015 (July 9), B2.
(2) Ni, Z.; Liu, Q.; Tang, K.; Zheng, J.; Zhou, J.; Qin, R.; Gao, Z.; Yu,
D.; Lu, J. Nano Lett. 2012, 12, 113.
(3) Bianco, E.; Butler, S.; Jiang, S.; Restrepo, O. D.; Windl, W.;
Goldberger, J. E. ACS Nano 2013, 7, 4414.
(4) Heath, J. R.; Shiang, J. J.; Alivisatos, A. P. J. Chem. Phys. 1994,
(42) Lambert, J. B.; Wang, G. T.; Teramura, D. H. J. Org. Chem.
1988, 53, 5422.
(43) Eaborn, C.; Pande, K. C. J. Chem. Soc. 1960, 1566.
(44) C−Ge bonds are better at stabilizing β-carbocation
intermediates in the context of chemical reactivity for this very
reason, accelerating reaction rates by 1−2 orders of magnitude relative
to their organosilane congeners (refs 42 and 43).
(45) Klausen, R. S.; Widawsky, J. R.; Su, T. A.; Li, H.; Chen, Q.;
Steigerwald, M. L.; Venkataraman, L.; Nuckolls, C. Chem. Sci. 2014, 5,
1561.
(46) Meisner, J. S.; Ahn, S.; Aradhya, S. V.; Krikorian, M.;
Parameswaran, R.; Steigerwald, M.; Venkataraman, L.; Nuckolls, C.
J. Am. Chem. Soc. 2012, 134, 20440.
(47) Engelkes, V. B.; Beebe, J. M.; Frisbie, C. D. J. Am. Chem. Soc.
2004, 126, 14287.
(48) Su, T. A.; Widawsky, J. R.; Li, H.; Klausen, R. S.; Leighton, J. L.;
Steigerwald, M. L.; Venkataraman, L.; Nuckolls, C. J. Am. Chem. Soc.
2013, 135, 18331.
(49) Venkataraman, L.; Park, Y. S.; Whalley, A. C.; Nuckolls, C.;
Hybertsen, M. S.; Steigerwald, M. L. Nano Lett. 2007, 7, 502.
(50) Glass, R. S.; Block, E.; Gruhn, N. E.; Jin, J.; Lorance, E.; Zakai,
U. I.; Zhang, S. Z. J. Org. Chem. 2007, 72, 8290.
(51) Song, L.; Lin, Y.; Wu, W.; Zhang, Q.; Mo, Y. J. Phys. Chem. A
2005, 109, 2310.
101, 1607.
(5) Heath, J. R.; Legoues, F. K. Chem. Phys. Lett. 1993, 208, 263.
(6) Gu, G.; Burghard, M.; Kim, G. T.; Dusberg, G. S.; Chiu, P. W.;
Krstic, V.; Roth, S.; Han, W. Q. J. Appl. Phys. 2001, 90, 5747.
(7) Mahenderkar, N. K.; Liu, Y.; Koza, J. A.; Switzer, J. A. ACS Nano
2014, 8, 9524.
̈
(8) Barth, S.; Kolesn
Chem. Mater. 2011, 23, 3335.
́ ́
ik, M. M.; Donegan, K.; Krstic, V.; Holmes, J. D.
(9) Hanrath, T.; Korgel, B. A. J. Phys. Chem. B 2005, 109, 5518.
(10) Morgan, G. T.; Drew, H. D. K. J. Chem. Soc., Trans. 1925, 127,
1760.
(11) Amadoruge, M. L.; Weinert, C. S. Chem. Rev. 2008, 108, 4253.
(12) Weinert, C. S. Dalt. Trans. 2009, 1691.
(13) Roewe, K. D.; Rheingold, A. L.; Weinert, C. S. Chem. Commun.
2013, 49, 8380.
(14) Wagner, H.; Baumgartner, J.; Muller, T.; Marschner, C. J. Am.
Chem. Soc. 2009, 131, 5022.
̈
(15) The hexagermane is the longest structurally characterized linear
oligomer and was independently synthesized by Roewe et al. (ref 13)
via Ge4Ph8 ring-opening and hydrogermolysis, as well as by Wagner et
al. (ref 14) via germanium shuttling.
(16) Mochida, K.; Hata, R.; Chiba, H.; Seki, S.; Yoshida, Y.; Tagawa,
S. Chem. Lett. 1998, 263.
(17) Kumada, M.; Sakamoto, S.; Ishikawa, M. J. Organomet. Chem.
1969, 17, 235.
(18) Ruehl, K. E.; Matyjaszewski, K. J. Organomet. Chem. 1991, 410,
1.
(19) Peterson, D. J. J. Org. Chem. 1967, 32, 1717.
(20) Xu, B.; Tao, N. J. Science 2003, 301, 1221.
(21) 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.
(52) George, C. B.; Ratner, M. A.; Lambert, J. B. J. Phys. Chem. A
2009, 113, 3876.
(53) Piqueras, M. C.; Crespo, R.; Michl, J. J. Phys. Chem. A 2003, 107,
4661.
(54) Fogarty, H. A.; Imhof, R.; Michl, J. Proc. Natl. Acad. Sci. U. S. A.
2004, 101, 10517.
(55) Michl, J.; West, R. Acc. Chem. Res. 2000, 33, 821.
(56) We study permethylated oligosilanes and -germanes because
they are much more stable than their hydrogenated structural
analogues.
(57) Shibano, Y.; Sasaki, M.; Tsuji, H.; Araki, Y.; Ito, O.; Tamao, K. J.
Organomet. Chem. 2007, 692, 356.
(58) Fukazawa, A.; Tsuji, H.; Tamao, K. J. Am. Chem. Soc. 2006, 128,
6800.
(59) Sabater, C.; Untiedt, C.; Palacios, J. J.; Caturla, M. J. Phys. Rev.
Lett. 2012, 108, 205502.
(22) Agraït, N.; Rodrigo, J.; Vieira, S. Phys. Rev. B: Condens. Matter
Mater. Phys. 1993, 47, 12345.
(23) Kamenetska, M.; Koentopp, M.; Whalley, A.; Park, Y.;
Steigerwald, M.; Nuckolls, C.; Hybertsen, M.; Venkataraman, L.
Phys. Rev. Lett. 2009, 102, 126803.
(24) Salomon, A.; Cahen, D.; Lindsay, S.; Tomfohr, J.; Engelkes, V.
B.; Frisbie, C. D. Adv. Mater. 2003, 15, 1881.
(60) Trouwborst, M.; Huisman, E.; Bakker, F.; van der Molen, S.; van
Wees, B. Phys. Rev. Lett. 2008, 100, 175502.
(61) Madelung, O. Semiconductors: Data Handbook, 3rd ed.; Springer:
Berlin, 2004.
(25) Venkataraman, L.; Klare, J. E.; Nuckolls, C.; Hybertsen, M. S.;
Steigerwald, M. L. Nature 2006, 442, 904.
(26) Matsuura, Y. J. Appl. Phys. 2014, 115, 043701.
(27) Venkataraman, L.; Klare, J. E.; Nuckolls, C.; Hybertsen, M. S.;
Steigerwald, M. L. Nature 2006, 442, 904.
(28) Sandorfy, C. Can. J. Chem. 1955, 33, 1337.
(29) Klausen, R. S.; Widawsky, J. R.; Steigerwald, M. L.;
Venkataraman, L.; Nuckolls, C. J. Am. Chem. Soc. 2012, 134, 4541.
F
DOI: 10.1021/jacs.5b08155
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX