A comprehensive study of extended tetrathiafulvalene cruciform molecules for molecular electronics: Synthesis and electrical transport measurements

Article abstract of DOI:10.1021/ja509937k

Cruciform-like molecules with two orthogonally placed π-conjugated systems have in recent years attracted significant interest for their potential use as molecular wires in molecular electronics. Here we present synthetic protocols for a large selection o

Full text of DOI:10.1021/ja509937k

Article
pubs.acs.org/JACS
A Comprehensive Study of Extended Tetrathiafulvalene Cruciform
Molecules for Molecular Electronics: Synthesis and Electrical
Transport Measurements
Christian R. Parker,† Edmund Leary,*,‡,§ Riccardo Frisenda,⊥ Zhongming Wei,*,†,∥ Karsten S. Jennum,†
Emil Glibstrup,† Peter Bæch Abrahamsen,† Marco Santella,†,∥ Mikkel A. Christensen,†
Eduardo Antonio Della Pia,† Tao Li,† Maria Teresa Gonzalez,§ Xingbin Jiang,# Thorbjørn J. Morsing,†
Gabino Rubio-Bollinger,‡ Bo W. Laursen,† Kasper Nørgaard,† Herre van der Zant,*,⊥ Nicolas Agrait,‡,§
and Mogens Brøndsted Nielsen*,†
†Department of Chemistry & Center for Exploitation of Solar Energy & Nano-Science Center & Danish-Chinese Center for
Nano-Electronics, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark
‡Laboratorio de Bajas Temperaturas, Departamento de Física de la Materia Condensada Mod
Madrid, E-28049, Madrid, Spain
́ ́
ulo 3, Universidad Autonoma de
§IMDEA-Nanoscience, Campus de Cantoblanco, Calle Faraday 9, Ciudad Universitaria de Cantoblanco, E-28049 Madrid, Spain
⊥Kavli Institute of Nanoscience, Delft University of Technology, 2600 GA Delft, The Netherlands
∥Sino-Danish Centre for Education and Research (SDC), Niels Jensens Vej 2, DK-8000 Aarhus C, Denmark
#National Center for Nanoscience and Technology, Beijing 100190, P. R. China
S
* Supporting Information
ABSTRACT: Cruciform-like molecules with two orthogonally
placed π-conjugated systems have in recent years attracted
significant interest for their potential use as molecular wires in
molecular electronics. Here we present synthetic protocols for
a large selection of cruciform molecules based on oligo-
(phenyleneethynylene) (OPE) and tetrathiafulvalene (TTF)
scaffolds, end-capped with acetyl-protected thiolates as
electrode anchoring groups. The molecules were subjected to
a comprehensive study of their conducting properties as well as
their photophysical and electrochemical properties in solution.
The complex nature of the molecules and their possible
binding in different configurations in junctions called for
different techniques of conductance measurements: (1)
conducting-probe atomic force microscopy (CP-AFM) meas-
urements on self-assembled monolayers (SAMs), (2) mechanically controlled break-junction (MCBJ) measurements, and (3)
scanning tunneling microscopy break-junction (STM-BJ) measurements. The CP-AFM measurements showed structure−
property relationships from SAMs of series of OPE3 and OPE5 cruciform molecules; the conductance of the SAM increased with
the number of dithiafulvene (DTF) units (0, 1, 2) along the wire, and it increased when substituting two arylethynyl end groups
of the OPE3 backbone with two DTF units. The MCBJ and STM-BJ studies on single molecules both showed that DTFs
decreased the junction formation probability, but, in contrast, no significant influence on the single-molecule conductance was
observed. We suggest that the origins of the difference between SAM and single-molecule measurements lie in the nature of the
molecule−electrode interface as well as in effects arising from molecular packing in the SAMs. This comprehensive study shows
that for complex molecules care should be taken when directly comparing single-molecule measurements and measurements of
SAMs and solid-state devices thereof.
HOMO−LUMO gaps. Detailed investigations of molecular
conductivity as a function of molecular structure and molecule-
electrode anchoring groups are particularly important for the
INTRODUCTION
■
Development of organic molecules as components for
molecular electronics devices has attracted interest as a means
of achieving miniaturization of integrated electronic circuits.1 In
particular, nanometer-sized π-conjugated molecules are suitable
as wires owing to their delocalized electrons and small
Received: June 6, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
development of the field. Such structure−property relationships
have been established from studies of single-molecule junctions
and self-assembled monolayers (SAMs), and they reveal that
the molecular conductance depends on several parameters such
as molecular length, HOMO−LUMO gap, molecular geometry,
quantum interference phenomena, and the nature and positions
of anchoring groups.1,2
Chart 1. Cruciforms Classified by Number of DTFs along
the OPE (in red) and As Termini (in blue)
Tetrathiafulvalene (TTF) has played a key role in molecular
electronics since Aviram and Ratner3 in 1974 suggested that a
molecular rectifier could be obtained by linking together this
donor and the electron-acceptor tetracyanoquinodimethane via
a nonconjugated bridge. Thus, several TTF-based intra-
molecular charge-transfer compounds have been prepared and
some tested in junctions.4 Owing to the three redox states of
TTF (0, +1, +2), several TTF-based wires, switches, and
memory devices have also been prepared.5 Oligo-
(phenyleneethynylene)s (OPEs) present another class of π-
conjugated molecules employed as wires.2 Thiols/thiolates,
conveniently protected as thioacetates, are often used for
anchoring such molecules to metal electrodes,6 but also
alkylsulfides and disulfides act as anchoring groups at metal
surfaces/junctions.6,7
We have in recent years combined OPEs and TTF in
cruciform-like structures in which dithiafulvenes (DTFs) are
placed perpendicularly to an OPE backbone, hence incorporat-
ing an extended TTF (exTTF).8 π-Conjugated cruciform
motifs have in general attracted interest for their electronic and
optical properties.9 Acetylenic benzene-extended TTF building
blocks were developed which allowed for stepwise construction
of OPE-TTF cruciforms.8a−d,f An enhanced stability of these
modules relative to other acetylenic motifs10 presented a major
advantage. Previously, conducting probe (CP) atomic force
microscopy (AFM) measurements on SAMs on gold of two
cruciform molecules with thioacetate anchoring groups revealed
an increased conductance relative to simple OPEs.8c Moreover,
we have contacted an OPE5-TTF cruciform in a three-terminal
geometry and showed gate-controlled conductance switching.8e
In 2013 we reported a new class of exTTF cruciforms in which
SAc-functionalized DTFs were incorporated as termini.8d,f
These molecules are composed of two perpendicularly oriented
exTTFs. The DTF-thiolate anchoring groups seemed to form
stable, highly conducting molecular junctions in break-junction
experiments. Yet, significantly lower conductances were
obtained by CP-AFM measurements on SAMs. To elucidate
in detail how the charge-transport properties are influenced by
the molecular structure, anchoring groups, and the exper-
imental technique, we designed a large selection of cruciform
motifs in which the number and placement of DTF units along
an OPE is systematically changed.
The target molecules (and reference molecules) can be
divided into six classes (Chart 1). Class [x.y] corresponds to
molecules with x DTFs as termini of the OPE and y DTFs
placed orthogonally to the OPE. Class [0.0] with no DTFs is
represented by the simple OPE3 and OPE5 wires 12a and 2,8c
while class [0.1] in total has one DTF unit, placed along the
wire (OPE3s 3, 4, and OPE5 5). Molecules with one additional
DTF along the OPE ([0.2]) are represented by OPE3s 6−9
and OPE5s 10 and 11. We have previously described the
syntheses of OPE3-bisDTFs 68a,b and 88d and of the OPE5-
bisDTF 10.8b We found that 8 containing no SAc end groups
showed no binding in break junctions, nor did it form SAMs on
gold. To explore the influence of distortion of the π-system
from planarity, we designed OPE3 9 with Me groups at the
exocyclic fulvene C atoms. Compounds 12 and 13 only have
DTFs as termini, one at each end of the wire ([2.0]).
Compounds 14 ([2.1]) and 15 ([2.2]) have DTF termini and,
in addition, one and two DTFs, respectively, along the wire.
Compounds 12−15 have different substitution patterns of the
DTF termini (SCH2CH2CN, SMe, SAc). Synthesis and studies
of compounds of structure 15 build upon a recent
communication.8d Conductance studies were performed using
(1) CP-AFM measurements on SAMs, (2) mechanically
controlled break-junction (MCBJ) experiments on single
molecules, and (3) scanning tunneling microscopy break-
junction (STM-BJ) experiments on single molecules.
RESULTS AND DISCUSSION
■
Synthesis. The DTF unit is in general introduced from
suitably alkyne-functionalized derivatives of benzaldehyde in a
Wittig reaction. Chart 2 shows suitable building blocks (16−
19) for the preparation of OPE-DTFs, that is, OPEs with one
DTF group. The synthesis of 16, 18, and 19, provided in
Supporting Information (SI), starts out from triflation of 2,5-
Chart 2. Arene Building Blocks
B
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Synthesis of OPE3-DTFs and OPE5-DTF ([0.1]) (Pr = n-Pr)
dihydroxybenzaldehyde followed by Sonogashira cross-cou-
pling11 reactions with suitable alkynes. The corresponding
CO2Me derivatives12,13 (instead of CHO) of 16 and 18 can
also be used as precursors as the CO2Me group is readily
converted to the desired aldehyde functionality by a reduction
followed by oxidation protocol (see SI). For synthesis of 17, we
refer to Wang et al.,14 and for an alternative route to 18, we
refer to Cade et al.15
Scheme 3. Synthesis of Asymmetrical OPE3-DTF ([0.1])
with Only One SAc End Group
With compound 18 in hand, the OPE3-DTF 3 and OPE5-
DTF 5 can be prepared according to Scheme 1. First, 18 was
treated with the phosphorus ylide generated from the known
phosphonium salt 20,16 affording the DTF compound 21.
Desilylation accompanied by transesterification, enhancing
solubility (CO2Me to CO2Pr, Pr corresponds to n-Pr
everywhere), gave the product 22, which was treated with
either 23,8b 24, or 25 under Sonogashira conditions to furnish
OPE3s 3, 26, and 27. Desilylation of 27 to 28 followed by a
coupling with 23 using the PdCl2(PPh3)2/CuI catalyst system
gave the SAc end-capped OPE5-DTF 5. To our surprise,
changing the catalyst system to Pd(dba)3/Xphos/CuI gave the
product 29, resulting from acetylation of the acetylides. While
acetylation reactions to our knowledge have not been effected
by this specific catalyst system, they were reported under other
catalyst conditions.17 An alternative protocol for constructing
the OPE3-DTF core is shown in Scheme 2. Here the core was
coupling with phenylacetylene, affording the OPE2 derivative
34. While we were not able to convert 33 to the OPE2-DTF
35, compound 34 was successfully converted to 35. Finally,
desilylation followed by cross-coupling between the inter-
mediate (36) and 23 gave 4.
The OPE3-bisDTF 7 with NH2 end groups was prepared
according to Scheme 4. First the known conversion of 37 to 38
Scheme 2. Synthesis of OPE3-DTF ([0.1]) with tert-Butyl-
Protected Thiol End Groups
Scheme 4. Synthesis of OPE3-bisDTF ([0.2]) with NH2 End
Groups
first prepared by treating the ditriflate 16 with the known
alkyne 30.6b The product 31 was subjected to a Wittig reaction
with 20 affording the OPE3-DTF 32 with tert-butylthio end
groups. Such groups can be converted to thioacetates in the
presence of DTF units8a and can also be used directly as
electrode anchoring groups (when weak binding is preferred).7c
The unsymmetrical OPE3-DTF 4 was prepared according to
Scheme 3. Compound 17 has both an aldehyde and a triflate
functionality, which allows it to be either subjected to a Wittig
reaction, affording the DTF compound 33, or a Pd-catalyzed
was performed;8b 38 was then coupled with p-iodoaniline (39),
giving target molecule 7. We also prepared the related Boc-
protected derivative of 7 from the Boc-protected p-iodoani-
line18 (see SI), but attempts of removing the Boc groups
afterward were, however, not successful. Instead we isolated a
compound, which from spectroscopic and mass data seemed to
be the product where HCl had added to each of the fulvenes.
Synthesis of OPE3-bisDTF 9 is shown in Scheme 5. First, the
known dialdehyde 408a was treated with MeMgBr, and the
resulting alcohol (mixture of stereoisomers, used crude) was
C
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
exTTFs 12a and 12b (Scheme 7); we previously used the same
method to prepare 12e (from 50 and 51e).8f Removal of one
Scheme 5. Synthesis of OPE3-bisDTF ([0.2]) with Methyl
Substituents at the Fulvene Units
Scheme 7. Synthesis of Extended TTFs ([2.0])
cyanoethyl group at each DTF by treatment with CsOH
generated dithiolates, which were acetylated to give products
12c and 12d.
OPE3s 13−15 with DTF termini were made by a
combination of phosphite-mediated couplings and Sonogashira
reactions, recently used to prepare derivatives 15a−d.8d,f The
aryl iodides 52a,b,e (Scheme 8) are key building blocks for the
oxidized by PCC. The diketone 41 was treated with the
phosphonate ester 42,19 deprotonated by NaHMDS, to afford
the product 43 in a Horner−Wadsworth−Emmons reaction.20
The structure was confirmed by X-ray crystallographic analysis
(Figure 1),21 which showed the two DTFs to be almost
Scheme 8. Synthesis of OPE3s with DTF Termini ([2.0],
a
[2.1], [2.2])
Figure 1. ORTEP plots showing molecular structure of 43 (two
views).
perpendicular to the plane of the central benzene ring, in
contrast to the planar structures of 378a and 8.8d Desilylation/
transesterification followed by coupling with 23 finally gave 9.
Incorporation of two DTFs along an OPE5 at two different
benzene rings was performed according to Scheme 6. First, 19
was subjected to a Wittig reaction to form the DTF compound
45, which was then subjected to a sequence of desilylation (and
transesterification) and Sonogashira reactions to ultimately give
11.
The next objective was to prepare molecular wires with DTF
termini. Phosphite-mediated coupling between benzaldehyde
derivatives and 4,5-dialkylthio-1,3-dithiole-2-thiones was pre-
viously shown to provide an efficient route to phenyl-DTFs.8f
Mulla and Zhao22 have also recently used this strategy for
incorporating DTFs as termini of OPEs. Heating terephtha-
laldehyde 50 and thiones 51a23 and 51b24 in P(OEt)3 gave
a
Yields/conditions: Table 1.
Sonogashira reactions. Compound 52a was previously prepared
from 4-iodobenzaldehyde and 4,5-bis(2′-cyanoethylthio)-1,3-
dithiole-2-thione using P(OEt)3.8f Cyanoethyl deprotections
followed by methylation/acetylation of the intermediate
thiolates gave methylated and acetylated derivatives (see SI).
Synthesis of the rather unstable alkyne coupling partner 53,
incorporating solubilizing pentyl groups, is shown in SI. From
these key building blocks, together with the diynes 22 and 38
Scheme 6. Synthesis of OPE5-bisDTF ([0.2])
D
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
and p-iodobenzaldehyde (24), differently substituted deriva-
tives of 13−15 were prepared according to Scheme 8 and Table
1. Different stereoisomers were obtained for DTFs containing
two different substituents (Z/E).
Table 1. Conditions and Yields for Scheme 8
starting
entry material(s)
conditions
product
yield
a
1
2
3
53 + 52a PdCl2(PPh3)2/CuI; NEt3
54 + 51a P(OEt)3; 110 °C
13a
13a
13b
11%
80%
50%
53 + 52b Pd(OAc)2/CuI/dppf/PPh3; NEt3/
PhMe; sonication 45 °C
Figure 2. UV−vis absorption spectra of 1 [0.0], 3 [0.1], 6 [0.2], 9
[0.2], and 13a [2.0] in CH2Cl2.
4
5
6
7
8
9
26 + 51a P(OEt)3; 110 °C
26 + 51b P(OEt)3; 110 °C
55 + 51a P(OEt)3; 110 °C
38 + 52b Pd(PPh3)4/CuI; NH(i-Pr)2
55 + 51b P(OEt)3; 100 °C
14a
14b
15a
15b
15b
15e
81%
37%
b
72%
54%
66%
54%
The broken conjugation in nonplanar 9 [0.2] causes a blueshift
in the end-absorption relative to that of 6. Comparison of
OPE3s 6 [0.2] and 13a [2.0] shows that having two DTFs
along the OPE induces a larger redshift than two DTF termini.
OPE5 10 [0.2] exhibits a longest-wavelength absorption
maximum (427 nm) red-shifted relative to that of 11 (360
nm), with DTFs at different benzene rings.
The photoluminescence spectra showed broad and unre-
solved bands, and the emission peaks are red-shifted for the
OPE-DTFs (SI). The DTF unit also causes the compounds to
be only weakly fluorescent relative to the parent OPEs 1 and 2.
Fluorescence quenching was recently reported for other DTF-
end-capped OPE/OPV oligomers.22
38 + 52e PdCl2(PPh3)2/CuI; THF/NEt3;
sonication 40 °C
10
11
12
13
14
15
13a
13b
14a
15a
15a
15b
CsOH; AcCl; DMF/PrOH
KOt-Bu; AcCl; DMF
13c
13d
14c
15b
15c
15d
36%
64%
68%
87%
53%
19%
CsOH; AcCl; DMF/PrOH
CsOH; MeI; DMF/PrOH
CsOH; AcCl; DMF/PrOH
c
c
c
CsOH; AcCl; DMF/PrOH
a
b
Initially: NH(i-Pr)2/THF. A yield of 63% was previously reported
under similar conditions; ref 8d. Ref 8d.
c
We have previously8d reported synthesis of 15a by a
phosphite coupling and converted this compound to either
15b or 15c; the former could subsequently be converted to
15d. Scheme 8 shows a more direct route to 15b by
incorporating the DTF termini from the precursor 38.
Compounds 13a,b and 14a,b can also be synthesized by two
other methods, starting from bis-terminal alkynes 53, 22, and
38, which in the first method (Table 1, entries 1, 3, 7, 9) were
coupled with DTF-functionalized iodobenzene (52a,b,e). This
coupling was quite slow and only gave poor to moderate yields
(11−54%) and a number of byproducts. We tried to couple the
thioacetate 52d with 38 aiming at a direct synthesis of 15d, but
the reaction did not go to completion and gave several
byproducts. A more successful method coupled p-iodobenzal-
dehyde 24 to the alkynes 53, 22, and 38; these reactions were
noticeably faster when the amine was used as neat solvent.
Dialdehydes 54, 26, and 55 were then coupled with thiones
51a,b using P(OEt)3 (Table 1, entries 2, 4−6, 8) to give
products in 70−80% yields.
As for 12a, removal of one cyanoethyl group on each DTF in
13−15a was achieved by CsOH. While methylation of the
intermediate thiolate worked in high yield (87%; entry 13),
yields were generally low for acetylations (19−68%; entries
10−12, 14, 15). With a large excess of AcCl, the desired
product was not obtained. For 14 and 15 with central DTFs, a
reaction with HCl seemed to occur. Finally, we note that
attempts of preparing derivatives of 15 and 52 with two SAc
groups at each DTF were not successful; only unstable
products were obtained.
Electrochemistry. The redox properties of representative
compounds were studied by cyclic voltammetry and differential
pulse voltammetry at concentrations of ca. 1 mM in CH2Cl2 +
0.1 M Bu4NPF6 (data listed in SI). Most of the compounds
were oxidized irreversibly at a scan rate of 100 mV/s, as
previously seen for DTF compounds due to radical
dimerization.25 At higher scan rates (1 V/s) the reversibility
became slightly better for 37, and two waves were discernible in
the forward and backward scans (see SI). By substituting one of
the SiMe3 groups of 37 for a Si(i-Pr)3, we found previously8b
partially reversible oxidations at half-wave potentials of 0.56 and
0.67 V vs Fc/Fc+ at a scan rate of 100 mV/s. The bulky Si(i-
Pr)3 may help to slow down dimerization, but better
reversibility may also be a concentration effect. Thus, for a
very dilute sample of 3, the first oxidation seemed to approach
reversibility. More electron-rich DTFs with alkylthio sub-
stituents (12a, 15a) were easier to oxidize than DTFs with ester
groups (6, 8). ExTTF 43 in which the fulvene H’s are replaced
by Me’s, preventing radical dimerization, exhibited a reversible
two-electron oxidation (see SI). This oxidation (E = +0.72 V vs
Fc/Fc+) is significantly anodically shifted (by 0.2 V) relative to
that of 37 (Epa = +0.54 V), presumably due to the orthogonal
arrangement of the two DTFs, also rendering them
independent redox centers.
CP-AFM Measurements on SAMs. High-quality SAMs
could be grown on Au substrate of most of the OPEs (dissolved
in Et3N/THF) with SAc end groups through the optimization
of growth conditions (following a general protocol).8c,26 Most
of the OPE SAMs were free of pinholes and defects and densely
packed with high coverage as checked by electrochemistry. The
experimental thickness of the SAMs of OPE3 1 (1.78 nm),
OPE3-DTF 3 (1.96 nm), OPE3-bisDTF 6 (2.36 nm), OPE5 2
(3.03 nm), and OPE5-bisDTF 10 (3.10 nm) were determined
from XPS data. Conductances were measured by CP-AFM in
vertical structure junctions. This method was previously shown,
UV−vis Absorption and Fluorescence Spectroscopy.
The optical properties were studied in CH2Cl2 (all data are
included in SI). Addition of DTFs along the OPE results in a
new red-shifted longest-wavelength absorption band as revealed
by comparing OPE3s 1 [0.0], 3 [0.1], and 6 [0.2] (Figure 2)
(and from OPE3s 13a [2.0], 14a [2.1], and 15a [2.2], see SI).
E
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
using a Pt AFM tip, to be efficient for measuring transport
properties of SAM-based molecular junctions trapping several
hundred molecules.27 In the present work, a Au-coated
conductive tip was used as top electrode forming a Au/SAM/
Au vertical junction. The same tip was used for all
measurements ensuring a constant contact area in the
junctions. This allows for comparison of relative conductance
between the different SAMs but not directly with the single-
molecule conductance since the actual number of molecules in
the junctions is unknown.
10 and 11 [0.2]. This behavior indicates that electron transport
in the SAMs is dominated by the molecular HOMO, which is
raised in energy by DTF units. The OPE5 series also reveals
that the position of the DTFs along the wire has an influence;
both OPE5 10 and 11 have two DTFs ([0.2]), but when they
are on the same benzene ring in the middle of the backbone
(10), a higher conductance is obtained. By changing the
position of these electron-donating substituents, we make small
changes in the HOMO, again supporting this as an important
transport channel in these SAM experiments. A strong effect of
the DTF termini is also observed; 12d (“OPE1” [2.0]) exhibits
a conductance 2 orders of magnitude larger than the other
molecules, although it is of comparable length to the OPE3s
(but shows a very broad range in conductance values).
Similarly, it is evident that substituting aryl groups in the
OPE5s for DTF termini increases the conductance (OPE3 15d
vs OPE5 10 and OPE3 13d vs OPE5 2). The effect of DTF
termini within the OPE3 series is more difficult to generalize
(blue arrows in Figure 3). Proceeding from OPE3 1 to 13d
results in an increased conductance although the length of the
molecule is increased, while OPE3 15d with two DTF termini
shows a reduced conductance relative to the shorter OPE3 6.
The influence of distorting the DTF units from coplanarity is
revealed by comparing [0.2] OPE3s 6 and 9; the conductance
of nonplanar 9 is only ca. 65% that of 6. OPE3 4 with only one
SAc did not form good SAMs (low coverage density), and we
could not grow good SAMs of 15e.
The results are collected in Table 2 and averaged I−V curves
are shown in SI. The conductance values were the average data
Table 2. CP-AFM Measurements on SAMs
a
compound
DTFs [x.y] conductance (nS) average resistance (Ω)
OPE3 1
[0.0]
[0.0]
[0.1]
[0.1]
[0.2]
[0.2]
[0.2]
[0.2]
[2.0]
[2.0]
[2.2]
1.49 0.48
0.58 0.35
2.85 2.23
0.76 0.60
5.56 2.81
3.61 2.75
2.28 1.53
1.65 0.91
143.6 91.3
2.06 1.06
4.06 1.23
6.71 × 108
1.72 × 109
3.51 × 108
1.32 × 109
1.80 × 108
2.77 × 108
4.39 × 108
6.06 × 108
6.96 × 106
4.85 × 108
2.46 × 108
OPE5 2
OPE3 3
OPE5 5
OPE3 6
OPE3 9
OPE5 10
OPE5 11
“OPE1” 12d
“OPE3” 13d
“OPE3” 15d
a
Errors are the standard deviation of the mean.
Single-Molecule Conductivity Studies. The conductance
of 1, 3, 6, 7, 9, 10, and 15c was measured using the STM-BJ
(Madrid) and MCBJ (Delft) methods. For the STM between
5000 and 10000 traces were recorded per compound and
1000−2000 for the MCBJ. Figure 4 shows representative
of about 200 measurements and determined over a small bias
range of 0.1 V. We notice that the conductance decreases with
the length of the molecules. If the number of DTFs are the
same, an OPE5 shows lower conductance than the related
OPE3 (OPE5 10 vs OPE3 6 and OPE5 5 vs OPE3 3). Previous
measurements using a Pt AFM tip showed that the
conductance increased (by a factor of ca. 9) when proceeding
from OPE3 1 to OPE3-bisDTF 7 or from OPE5 2 to OPE5-
bisDTF 10.8c Measurements on the large selection of OPE3
and OPE5 molecules support this trend for both series as
depicted in Figure 3. For the OPE3 series, we observe that the
conductance increases as the number of DTFs along the wire
increases. Thus, as indicated by the red arrows, for the OPE3
series the conductance increases in the sequences: 1 [0.0] < 3
[0.1] < 6 [0.2] and 13d [2.0] < 15d [2.2]. For the OPE5 series,
the same influence of the DTFs is observed: 2 [0.0] < 5 [0.1] <
Figure 4. Representative individual G vs z traces for molecules 6, 7,
15c, and 10 (more examples in SI) with plateau regions highlighted in
red. OPEs 1, 3, 6 and 9 show similar plateau profiles to 6, and only a
typical trace of 6 is presented. For 7 the junctions often form after the
gold contact rupture. Lower gains were used for the traces of 6 and 7
than for 15c and 10, which places the noise level at 10−6 G0 and 10−7
G0, respectively.
conductance vs distance traces (G(z)) for 6, 7, 15c, and 10
(STM data). At values close to 1 G0, small plateaus in G,
followed by a sharp drop, indicate the final breakdown of the
gold contact. Below this, longer plateaus indicate molecular
junction formation, where one or a few molecules may be
bound between the electrodes. The plateaus for all compounds
generally fluctuate up to 1 order of magnitude from beginning
to end, with discrete jumps between different conductance
states visible as the junction is elongated. At the end, a sharp
drop in current indicates breakdown of the junction. 1D and
Figure 3. Conductances of SAMs as a function of the OPE length and
number and position of DTF units.
F
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. Left: 1D normalized histograms of 1 (a), 3 (b), 6 (c), 9 (d) measured in Madrid (red) and Delft (black) with no data selection. Blue
histogram in (d) contains only selected plateaus of 9 (STM) to amplify the low molecular signal. Right: 2D normalized histograms using MCBJ [1
(e), 3 (f), 6 (g), 9 (h)]; STM [(1 (i), 3 (j), 6 (k), 9 (l)].
Figure 6. 2D histograms of the traces containing plateaus of 1, 3, 6, 9, 7, 15c, and 10 from the STM-BJ technique. All histograms have been plotted
with the same x−y axes, apart from 10 due to the longer plateaus and lower conductance.
2D histograms were built from the total recorded data (Figure
5). We also carried out a selection of only the traces containing
plateaus (from STM data), the 2D histograms of which are
shown in Figure 6 (see SI for 1D histograms). Figure 5a−d
shows the 1D conductance histograms for molecules 1, 3, 6,
and 9 without data selection; each bin is equal to log (G/G0) =
0.02 for the STM data and 0.04 for the MCBJ data. Histograms
were normalized according to ref 28.
Comparing the results between the two setups in the two
different laboratories, we find that the 1D histograms for
molecules 1 and 3 (Figure 5a,b) are generally very similar. The
peaks located close to log(G/G0) = −4 are centered at
practically the same values for both molecules in both setups.
This value is also very similar to another OPE3 we previously
studied containing two hexyloxy substituents on the central
benzene ring.28 The lower peak height of 3 compared to 1 is
also reproduced in both setups and is a result of a lower
percentage of plateaus obtained for 3 (with the STM we find
that 26% of traces show plateaus for 1 and 18% for 3). The
normalized 2D histograms of 1 and 3 from the MCBJ setup are
shown in Figure 5e,f and those of the STM in Figure 5i,j. The
color scale indicates the number of counts normalized to the
number of curves used to build the histograms, and to aid
comparison between the two sets of data we use the same range
for both. From the STM data, the mean plateau length for 1
and 3 was found to be 1.28 and 1.25 nm, respectively (see SI
for plateau length distributions). We generally expect an initial
gap of approximately 0.5 nm due to relaxation of one or two
gold atoms in the electrodes, putting the mean real
interelectrode spacing very close to the S−S distance of 1
and 3 (both 2.07 nm). The 95th percentile values were 1.84
and 1.87 nm, which are only slightly less than the molecular
length and represent the few junctions in which the snap-back
is very small. The use of this particular color scale also shows
that no plateaus were found for 6 with the MCBJ, while a
definite signal is obtained by STM, albeit weaker than for 1 and
3. The percentage of plateaus found for 6 with the STM was
12% in the run shown in Figure 5c. Several samples were
measured with the STM, also giving weak signals, and we had
to test several different areas to produce the results shown in
Figure 5c. Clearly, statistically speaking, this molecule is harder
to trap than 1 and 3, so it is reasonable to assume that in the
MCBJ, where only a single pair of electrodes can be used, there
is a much smaller probability of wiring than in the STM.
Molecule 9 gives a weak signal in both setups, which is not so
clear looking at the 1D histogram (Figure 5d, 9% plateaus).
The signal is, however, clearly visible in the 2D histograms
thanks to the choice of contrast, which enables low counts to be
viewed in blue/green (Figure 5g,k). In order to emphasize the
weak signal in the 1D histogram, we have plotted only the
selected plateaus from the STM data, which shows that the
peak is located in roughly the same position as for 1, 3, and 6.
G
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Molecule 9 is similar in structure to 6 but has a Me group on
each of the DTFs. The fact that a signal is seen for 9 but not for
6 in the MCBJ study may suggest that the Me groups play a
role in helping junction formation. In the STM results,
however, we find roughly the same percentage of plateaus for
both molecules 6 and 9. As it is hard, however, to control the
precise conditions that might influence the junction formation
probability on any given sample (especially the distribution of
molecules on the surface), we cannot place too much emphasis
on the final overall percentage. We suggest that while the
number of experimental runs (i.e., number of separate samples
prepared) is too low to draw solid conclusions about the
efficacy of the additional methyl groups, they may slightly help
to form junctions incorporating a TTF unit.
Overall, there seems to be a clear tendency, which is seen in
both setups, for the junction formation probability to decrease
with the number of DTF side groups (the highest percentages
found with the STM method were 26%, 18%, 12% and 9% for
molecules 1, 3, 6, and 9, respectively). This strongly suggests
that the DTF groups hinder the formation of junctions, most
likely due to their additional S atoms. These can potentially
interact with the gold to reduce the mobility of the molecule,
which would make it harder for it to diffuse toward the
junction. Additionally, this could also make it more difficult to
suspend the molecule freely between the contacts. The
similarity in the position of the conductance peaks of molecules
1, 3, 6, and 9 shows that there is, however, negligible electronic
influence of the DTF units, at least in the low bias regime, with
the most probable conductance lying between 1.5 and 1.8 ×
10−4 G0 for all molecules. The histograms have a broadness of
approximately 1 order of magnitude at half-maximum, which
can be explained as a result of different molecular binding
geometries, and the potential of having more than one wired
molecule. As such factors can vary slightly from one experiment
to another, this limits the degree to which peak positions can be
compared. Hence, the observed 20% variation in the most
probable conductance seen for each OPE3 compound is well
within that naturally expected based on these factors alone. We
conclude that any electronic effect of the DTF groups is less
than this variation.
We find subtle differences between the results of the two
setups including shorter 1 G0 plateaus and fewer counts in the
region immediately after the gold contact breakage for the
MCBJ technique. The shorter 1 G0 plateaus result in a less
prominent peak in the histograms in Figure 5a−d. This is a
known phenomenon related to the separation speed of the
electrodes.29 The rate of retraction in the STM was
approximately 40 and 5 nms−1 in the MCBJ. Despite the
shorter 1 G0 plateaus, however, the stretching length of the
molecules does not show the same consistent trend. This can
be noticed in the 1D histograms in Figure 5a−d; the ratio of
the molecular peak heights is much more similar than for the 1
G0 peaks. This points to an insensitivity of the molecular
junction lifetime within the range of retraction speeds tested.
We have observed similar behavior in amine terminated
OPEs.30 It is also clear that more pure-tunneling traces are
seen in the STM, and with greater counts in the region between
log(G/G0) = 0 and −2. This may also be related to the speed of
the measurement as the slower it is, the more the gold
electrodes can retract through relaxation of the gold atoms.
We now focus on the other compounds for which a signal
was found with the STM, but not for the MCBJ. This includes
7, 15c, and 10. To highlight the differences, the 2D histograms
built from only the traces containing plateaus are shown in
Figure 6. We note that 15c was previously studied by MCBJ8f
using CH2Cl2 as solvent rather than dichlorobenzene. Under
these conditions, a clear conductance plateau around 10−3 G0
was seen, and it is likely the junctions here contains an
aggregate of molecules. Calculations previously showed that the
different E/Z isomers should have the same transmission at the
Fermi level.8d A noticeable decrease in the conductance is
found for 7 (amine anchors) (Figure 6e). The 1D histogram
peak maximum is centered at log(G/G0) = −4.6 (2.5 × 10−5
G0), almost 1 order of magnitude lower than for 6. This value is
consistent with previously reported values for amine terminated
OPEs.30 Between 8 and 20% of traces contained a conductance
plateau depending on the experimental run, which is similar to
6, but much lower than for the nonfunctionalized OPE diamine
(50%).30 This agrees with the results of 6, supporting the
observation that DTF units decrease junction formation
probability.
Molecule 15c (Figure 6f) gives a very weak signal, only 3% of
traces showed a significant plateau. The low success rate in the
break-junction experiment is once again consistent with the
presence of a large amount of sulfur hindering the diffusion of
the molecule into the junction. The plateau length is not,
however, significantly longer than that observed for molecules
1, 3, 4, 6−9, despite the longest S−S distance in the molecule
(2.9 nm), which is 0.8 nm more than in 1, 3, 4, 6−9. The 1D
histogram peak is centered at log(G/G0) = −4.6, which is
similar to 7 and suggests that rather than binding at the
terminal S atoms, the molecule possibly adopts a configuration
in which it binds through the face of a DTF unit to each
electrode. This would explain both the shorter than expected
plateaus and reasonably high conductance. For alternative
binding scenarios (and calculated S−S distances), see SI.
In comparison, for OPE5 10 (Figure 6g) we find a high rate
of junction formation (similar to OPE3 1). In one experimental
run we found a rate of 24%, which is higher by at least a factor
of 2 compared to all the runs with the shorter exTTF-
containing OPEs. A likely reason is that the greater flexibility of
the long OPE5 prevents the DTFs from interacting strongly
with gold. The conductance of 10 is significantly lower than
that of 6, with the peak in the 1D histogram centered at log(G/
G0) = −6.1. This is consistent with a previous report of an
OPE52f and suggests the molecules indeed bind through their
terminal S atoms. The mean plateau length of 2.1 nm is also
consistent with the molecule binding in this way (adding the
gold retraction results in a real spacing of 2.5−2.7 nm, and the
calculated S−S distance is 3.3 nm).
Taking the conductance of 1 and 2 we can extract a β-value,
which represents the degree of electronic attenuation of the
OPE series under the present conditions. This yields a value of
3.7 nm−1, close to the 3.4 nm−1 found by Kaliginedi et al.31 for
similar OPE molecules. We also find the same value for
molecules 6 and 10 with the exTTF groups. This is, however,
almost twice that of the 2 nm−1 found by photoinduced charge-
transfer studies of TTF-OPE-C60 triads.32 The lower β-value of
the donor-bridge-acceptor system may be a result of the
different tunnel-barrier heights in both systems, determined by
the gold electrodes and donor/acceptor units, respectively.33 In
a study of OPE wires in a large-area junction device a still lower
β-value of 1.5 nm−1 was found. In this study, Valkenier et al.26
used gold electrodes, with a layer of PEDOT:PSS placed
between the OPE layer and the top gold contact. Plausibly, the
origin of the low value here could lie in the uncertainty in
H
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 3. Summary of Conductances
compound
DTFs [x.y] greatest S−S distance (nm) CP-AFM conductance (nS) STM-BJ conductance (nS) STM-BJ conductance, log(G/G0) max
OPE3 1
OPE5 2
OPE3 3
OPE3 6
OPE3 9
OPE5 10
OPE3 15
[0.0]
[0.0]
[0.1]
[0.2]
[0.2]
[0.2]
[2.2]
2.0
3.4
2.0
2.0
2.0
3.4
3.4
1.49 0.48
0.58 0.35
2.85 2.23
5.56 2.81
3.61 2.75
2.28 1.53
4.06 1.23 (15d)
5.1 0.05
0.062 0.0003
4.0 0.2
−3.82 0.44
−6.03 0.3
−3.85 0.5
−3.80 0.64
−3.74 0.56
−6.14 0.41
−4.54 0.67 (15c)
2.7 0.5
10.6 0.14
0.039 0.0006
0.74 0.03 (15c)
a
For the CP-AFM data the conductance is the mean value, and the error is the standard deviation. For the STM-BJ we quote both (1) the maximum
of the linear histogram representation, where the error is the error in the fit to a single Gaussian, and (2) the maximum in the log histogram (also
fitted with a Gaussian), where the error is half the width at half-maximum. The peak maxima for 1, 3, 6, and 9 remain approximately identical in the
log(G/G0) histogram but drop slightly in the linear histograms. The reason for this is the broadening, which occurs to both higher and lower values,
can be seen in Figure S2 (SI) in the log histograms. This naturally shifts the linear maximum down but fails to capture the nature of the broadening.
Hence, for the comparison of molecules using the BJ method, we prefer the log representation.
packing density of molecules in the junction and the thickness
of the layer. It is also possible that the formation of a layer shifts
the frontier orbitals relative to the isolated molecule case due to
a difference in the polarizability of the environment,34 for which
there is also evidence in single-molecule junctions.35
which is a “push down” process with the tip load on top of the
SAMs, a group of molecules is measured in the junctions; with
the molecules being slightly tilted under the push (load force)
from the CP-AFM tip. The tilt actually places the center of the
molecule closer to the electrodes, so its influence on the charge
transport would be greater than in the BJ measurement. We
also note that this result is unlikely to be related to the total
number of molecules in the junction as the DTFs contain
reasonably bulky ester groups that should actually slightly
reduce the density and, hence, the conductance of the layer.
Based on this alone, there would have to be 3−4 times the
density of molecules of 6 compared to 1 to explain the larger
conductance. This seems physically unrealistic, supporting an
explanation based on the transport pathway through the
molecules. In the SAM molecular junctions, the molecular
backbones are densely enough packed, such that the total
current density could also contain contributions from both
intramolecular (through-bond tunneling) and intermolecular
(wire-to-wire tunneling) transport. We suggest that the
combination of greater proximity of the central groups to the
electrode, poorer top contact, and the greater likelihood of
through-space or intermolecular transport pathways can
account for the different conductance trend found for the
SAM-based measurement compared to the single molecule
result.
Overall, the comparison suggests that molecules can show
different properties depending on their environment. Based on
the conductance trend found by CP-AFM for 1, 3 and 6, we
can assume that transport is dominated by the molecular
HOMOs. On the other hand, the BJ results show no such
trend, and this suggests that the HOMOs do not contribute
significantly to the transport along the backbone. The upshot of
this is that one technique alone cannot be used to predict the
properties of molecular junctions in all environments. This has
important consequences for the field of molecular electronics,
where it has generally been assumed that studying the
conductance at the fundamental level of just one molecule is
useful for explaining the properties of larger molecular
ensembles. Our work shows this particular concept may not
be universally valid. Further studies are needed on a wider
range of molecules to determine the extent of how electrical
properties do or do not translate from the single- to the many-
molecule case.
The fact that we find essentially the same conductances and
β-values for our exTTF molecules (6 and 10) as for the
respective analogues without DTF groups (1 and 2) shows
there is little influence of substitution, at least in the low bias
regime, toward transport across single molecules in BJ
experiments. Without further analysis, it is difficult to pin
down the precise reasons for this, however it is possible that the
orbitals responsible for transport have a low contribution from
the DTF groups. Despite the fact that the DTF units are
expected to contribute strongly to the HOMO, the main
transport orbital is not necessarily this orbital. Transport could
in fact be dominated by a lower lying level, such as the
HOMO−2, which may be more delocalized through the
molecule, providing better electronic coupling between the two
electrodes.
The lower conductance found for the SAM in the OPE3
series (except for 6) compared to the BJ measurements (see
summary in Table 3) seems at first counterintuitive considering
the high number of molecules expected to be present in each
CP-AFM junction. This can, however, be convincingly
explained by assuming that the top SAc remains intact in the
SAM, while both SAc groups are likely cleaved in the BJ.28 This
would not only break the chemical contact of a Au−S direct
bond but would also increase the length of the junction by the
additional nonconjugated acetyl group at the interface to the
Au tip.
A further interesting result of our study is the different trends
seen for the same OPE3 backbone (1, 3, 6) with different
numbers of DTF side groups between CP-AFM and BJ. The
trend in the CP-AFM data follows the order G6 > G3 > G1,
while the BJ data show the molecules to have essentially the
same conductance. To explain this we refer to the nature of the
top contact in the CP-AFM. The probable lack of a covalent
bond between the tip and the molecules could conceivably
change the transport pathway through the layer. In the BJ, the
main pathway most likely occurs through the entire molecule,
from sulfur to sulfur. In the CP-AFM, the poorer top contact,
combined with the known tilt of thiol monolayers, may force
the current to take an alternative path. The poorer contact will
decrease through-bond tunneling and increase through-space or
intermolecule pathways. In the CP-AFM device geometry,
I
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
number 606728), and Spanish MICINN/MINECO through
the programs MAT2011-25046 are thanked for support.
CONCLUSIONS
■
By a combination of Sonogashira, Wittig, and phosphite-
mediated reactions, a large selection of cruciform-like OPE-
TTF molecules was prepared. Thioacetate electrode anchoring
groups were incorporated either as OPE end groups or as
peripheral substituents on DTF termini. CP-AFM measure-
ments on SAMs showed the conductance to increase with (1)
decreasing length of the molecule, (2) an increasing number of
orthogonally placed DTFs along the wire, and (3) by replacing
aryl units of the OPE by DTFs. In contrast, both STM-BJ and
MCBJ single-molecule measurements did not show any
significant change in conductance by having DTFs along the
OPE. Instead, the probability of trapping a molecule decreased
systematically with the number of orthogonally placed DTFs
within the OPE3 series (1, 3, 6); these probably offer
alternative binding sites, reducing the mobility of the molecules
for diffusing toward the junction. Despite this, the mean
junction separation at breakdown was virtually identical across
this series, indicating that within the junction the DTFs do not
cause significant instability. Notably, measurements on SAMs
provide significantly lower conductances than those on single
molecules. One reasonable explanation is the different
connecting geometries of the device. In the CP-AFM
measurements on SAMs, we most likely have a junction
structure corresponding to Au−S−OPE−SAc ||| Au, where the
top sulfur is not covalently linked to Au (indicated by |||
symbol), while the BJs in most cases likely correspond to Au−
S−OPE−S−Au (covalent anchoring to both electrodes). The
lack of a covalent top linkage may be responsible for the
different trends seen for the single-molecule and SAM studies
by altering the conductance pathway across the junction. The
different outcome not only brings to the forefront the challenge
of characterizing conducting properties of complex molecules
containing several sulfur atoms as potential binding sites but
also stresses the importance of the interface with regard to the
observed trends within series of molecules. It is very gratifying
that CP-AFM/SAM studies carried out in different laboratories
(Beijing/Au tip vs Copenhagen/Pt tip (previous work)) reveal
similar trends as do BJ studies carried out in different
laboratories (Madrid/STM-BJ vs Delft/MCBJ).
REFERENCES
■
(1) (a) Robertson, N.; McGowan, C. A. Chem. Soc. Rev. 2003, 32,
96−103. (b) Benniston, A. C. Chem. Soc. Rev. 2004, 33, 573−578.
(c) Nørgaard, K.; Bjørnholm, T. Chem. Commun. 2005, 1812−1823.
(d) Weibel, N.; Grunder, S.; Mayor, M. Org. Biomol. Chem. 2007, 5,
2343−2353. (e) Klajn, R.; Stoddart, J. F.; Grzybowski, B. A. Chem. Soc.
Rev. 2010, 39, 2203−2237.
(2) (a) Tour, J. M.; Rawlett, A. M.; Kozaki, M.; Yao, Y.; Jagessar, R.
C.; Dirk, S. M.; Price, D. W.; Reed, M. A.; Zhou, C.-W.; Chen, J.;
Wang, W.; Campbell, I. Chem.Eur. J. 2001, 7, 5118−5134.
(b) Mayor, M.; Weber, H. B.; Reichert, J.; Elbing, M.; von Hanisch,
̈
C.; Beckmann, D.; Fischer, M. Angew. Chem., Int. Ed. 2003, 42, 5834−
́
5838. (c) Huber, R.; Gonzalez, M. T.; Wu, S.; Langer, M.; Grunder, S.;
Horhoiu, V.; Mayor, M.; Bryce, M. R.; Wang, C.; Jichati, R.;
Schonenberger, C.; Calame, M. J. Am. Chem. Soc. 2008, 130, 1080−
̈
1084. (d) Vonlanthen, D.; Mishchenko, A.; Elbing, M.; Neuburger, M.;
Wandlowski, T.; Mayor, M. Angew. Chem., Int. Ed. 2009, 48, 8886−
8890. (e) Kaliginedi, V.; Moreno-García, P.; Valkenier, H.; Hong, W.;
́
García-Suarez, V. M.; Buiter, P.; Otten, J. L. H.; Hummelen, J. C.;
Lambert, C. J.; Wandlowski, T. J. Am. Chem. Soc. 2012, 134, 5262−
5275. (f) Frisenda, R.; Perrin, M. L.; Valkenier, H.; Hummelen, J. C.;
van der Zant, H. S. J. Phys. Status Solidi B 2013, 250, 2431−2436.
(3) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 277−283.
(4) (a) Segura, J. L.; Martín, N. Angew. Chem., Int. Ed. 2001, 40,
1372−1409. (b) Perepichka, D. F.; Bryce, M. R.; Pearson, C.; Petty,
M. C.; McInnes, E. J. L.; Zhao, J. P. Angew. Chem., Int. Ed. 2003, 42,
4636−4639. (c) Ho, G.; Heath, J. R.; Kondratenko, M.; Perepichka, D.
F.; Arseneault, K.; Pez
́
olet, M.; Bryce, M. R. Chem.Eur. J. 2005, 11,
2914−2922. (d) Leroy-Lhez, S.; Baffreau, J.; Perrin, L.; Levillain, E.;
Allain, M.; Blesa, M.-J.; Hudhomme, P. J. Org. Chem. 2005, 70, 6313−
6320.
(5) (a) Lou, Y.; Collier, 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, 519−525. (b) Giacalone, F.;
́
Herranz, M.; Gruter, L.; Gonzalez, M. T.; Calame, M.; Schonenberger,
̈
̈
̂
C.; Arroyo, C. R.; Rubio-Bollinger, G.; Velez, M.; Agraït, N.; Martín,
N. Chem. Commun. 2007, 4854−4856. (c) Leary, E.; Higgins, S. J.; van
Zalinge, H.; Haiss, W.; Nichols, R. J.; Nygaard, S.; Jeppesen, J. O.;
Ulstrup, J. J. Am. Chem. Soc. 2008, 130, 12204−12205. (d) Liao, J.;
Agustsson, J. S.; Wu, S.; Schonenberger, C.; Calame, M.; Leroux, Y.;
̈
Mayor, M.; Jeannin, O.; Ran, Y.-F.; Liu, S.-X. Nano Lett. 2010, 10,
759−764.
(6) (a) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.;
Whitesides, G. M. Chem. Rev. 2005, 105, 1103−1169. (b) Stuhr-
Hansen, N.; Sørensen, J. K.; Moth-Poulsen, K.; Christensen, J. B.;
Bjørnholm, T.; Nielsen, M. B. Tetrahedron 2005, 61, 12288−12295.
(c) Nørgaard, K.; Nielsen, M. B.; Bjørnholm, T. In Functional Organic
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures, characterization data, and NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
Materials; Muller, T. J. J., Bunz, U. H. F., Eds.; Wiley-VCH: Weinheim,
̈
Germany, 2007; pp 353−392.
(7) (a) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105,
4481−4483. (b) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546−
AUTHOR INFORMATION
■
Corresponding Authors
edmund.leary@imdea.org
zmwei@iccas.ac.cn
H.S.J.vanderZant@tudelft.nl
mbn@kiku.dk.
558. (c) Kubatkin, S.; Danilov, A.; Hjort, M.; Cornil, J.; Bred
́
as, J.-L.;
d, P.; Bjørnholm, T. Nature 2003, 425,
̊
Stuhr-Hansen, N.; Hedegar
698−701. (d) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.;
Whitesides, G. M. Chem. Rev. 2005, 105, 1103−1169. (e) Meisner, J.
S.; Ahn, S.; Aradhya, S. V.; Krikorian, M.; Parameswaran, R.;
Steigerwald, M.; Venkataraman, L.; Nuckolls, C. J. Am. Chem. Soc.
2012, 134, 20440−20445.
Notes
The authors declare no competing financial interest.
(8) (a) Sørensen, J. K.; Vestergaard, M.; Kadziola, A.; Kilsa, K.;
̊
Nielsen, M. B. Org. Lett. 2006, 8, 1173−1176. (b) Jennum, K.;
Vestergaard, M.; Pedersen, A. H.; Fock, J.; Jensen, J.; Santella, M.; Led,
ACKNOWLEDGMENTS
■
J. J.; Kilsa, K.; Bjørnholm, T.; Nielsen, M. B. Synthesis 2011, 539−548.
̊
The European Union seventh Framework Programme (FP7/
2007-2013) under the grant agreement no. 270369 (“ELFOS”),
University of Copenhagen, FOM, NWO/OCW, ERC
(advanced ERC grant), (Mols@Mols), “MOLESCO” (Project
(c) Wei, Z.; Li, T.; Jennum, K.; Santella, M.; Bovet, N.; Hu, W.;
Nielsen, M. B.; Bjørnholm, T.; Solomon, G. C.; Laursen, B. W.;
Nørgaard, K. Langmuir 2012, 28, 4016−4023. (d) Parker, C. R.; Wei,
Z.; Arroyo, C. R.; Jennum, K.; Li, T.; Santella, M.; Bovet, N.; Zhao, G.;
J
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Hu, W.; van der Zant, H. S. J.; Vanin, M.; Solomon, G. S.; Laursen, B.
W.; Nørgaard, K.; Nielsen, M. B. Adv. Mater. 2013, 25, 405−409.
(e) Fock, J.; Leijnse, M.; Jennum, K.; Zyazin, A. S.; Paaske, J.;
Lamert, C. J.; Wandlowski, T. J. Am. Chem. Soc. 2012, 134, 5262−
5275.
(32) Atienza, C.; Martín, N.; Wielopolski, M.; Haworth, N.; Clark,
T.; Guldi, D. M. Chem. Commun. 2006, 3202−3204.
(33) Adams, D. M.; Brus, L.; Chidsey, C. E. D.; Creager, S.; Creutz,
C.; Kagan, C. R.; Kamat, P. V.; Lieberman, M.; Lindsay, S.; Marcus, R.
A.; Metzger, R. M.; Michel-Beyerle, M. E.; Miller, J. R.; Newton, M.
D.; Rolison, D. R.; Sankey, O.; Schanze, K. S.; Yardley, J.; Zhu, X. J.
Phys. Chem. B 2003, 107, 6668−6697.
Hedegard, P.; Nielsen, M. B.; van der Zant, H. S. J. Phys. Rev. B 2012,
̊
86, 235403. (f) Schou, S. S.; Parker, C. R.; Lincke, K.; Jennum, K.;
Vibenholt, J.; Kadziola, A.; Nielsen, M. B. Synlett 2013, 24, 231−235.
(9) See for example: Zucchero, A. J.; McGrier, P. L.; Bunz, U. H. F.
Acc. Chem. Res. 2010, 43, 397−408.
(10) Andersson, A. S.; Qvortrup, K.; Torbensen, E. R.; Mayer, J.-P.;
(34) Torrente, I. F.; Franke, K. J.; Pacual, J. I. J. Phys.: Condens. Matter
2008, 20, 184001.
Gisselbrecht, J.-P.; Boudon, C.; Gross, M.; Kadziola, A.; Kilsa, K.;
̊
Nielsen, M. B. Eur. J. Org. Chem. 2005, 3660−3671.
(35) Leary, E.; Hobenreich, H. H.; Higgins, S. J.; van Zalinge, H.;
̈
(11) Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975,
16, 4467−4470.
Haiss, W.; Nichols, R. J.; Finch, C. M.; Grace, I.; Lambert, C. J.;
McGrath, R.; Smerdon, J. Phys. Rev. Lett. 2009, 102, 086801.
(12) CO2Me analogue of 16 − known compound: Zhang, Q.; Shi,
C.; Zhang, H.-R.; Wang, K. K. J. Org. Chem. 2000, 65, 7977−7983.
(13) CO2Me analogue of 18 − known compound: Jacobsen, M. F.;
Ravnsbæk, J. B.; Gothelf, K. V. Org. Biomol. Chem. 2010, 8, 50−52.
(14) Wang, L.-G.; Zhan, T.-G.; Zhao, X.; Jiang, X.-K.; Li, Z.-T.
Tetrahedron 2012, 68, 5303−5310.
(15) Cade, I.; Long, N. J.; White, A. J. P.; Williams, D. J. J. Organomet.
Chem. 2006, 691, 1389−1401.
(16) Sato, M.; Gonnella, N. C.; Cava, M. P. J. Org. Chem. 1979, 44,
930−933.
(17) (a) Baxendale, I. R.; Schou, S. C.; Sedelmeier, J.; Ley, S. V.
Chem.Eur. J. 2010, 16, 89−94. (b) Egi, M.; Azechi, K.; Saneto, M.;
Shimizu, K.; Akai, S. J. Org. Chem. 2010, 75, 2123−2126.
(18) Hu, L.-Y.; Ryder, T. R.; Rafferty, M. F.; Taylor, C. P.; Feng, M.
R.; Kuo, B.-S.; Lotarski, S. M.; Miljanich, G. P.; Millerman, E.; Siebers,
K. M.; Szoke, B. G. Bioorg. Med. Chem. 2000, 8, 1203−1212.
(19) Baffreau, J.; Dumur, F.; Hudhomme, P. Org. Lett. 2006, 8,
1307−1310.
(20) (a) Wadsworth, W. S., Jr. Org. React. 1977, 25, 73−253.
(b) Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 863−927.
(21) Crystal data for 50: C34H38O8S4Si2, Mr = 759.11, orthorhombic,
Pbca, a = 20.3534(10), b = 8.3596(5), c = 22.6457(11), V = 3853.1(4),
T = 122(1) K, μ(MoKα) = 0.355, 2θmax= 53.46°, 49188 reflections
measured, 4097 unique reflections (Rint = 0.1405), R1(I > 4σ(I)) =
0.0426, wR2(all data) = 0.0981, S = 1.043. CCDC 1005219 contains
the supplementary crystallographic data for this paper. These data can
be obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
(22) Mulla, K.; Zhao, Y. J. Mater. Chem. C 2013, 1, 5116−5127.
(23) Svenstrup, N.; Rasmussen, K. M.; Hansen, T. K.; Becher, J.
Synthesis 1994, 809−812.
(24) Simonsen, K. B.; Svenstrup, N.; Lau, J.; Simonsen, O.; Mørk, P.;
Kristensen, G. J.; Becher, J. Synthesis 1996, 407−418.
́
(25) (a) Salle, M.; Belyasmine, A.; Gorgues, A.; Jubault, M.; Soyer, N.
Tetrahedron Lett. 1991, 32, 2897−2900. (b) Lorcy, D.; Carlier, R.;
Robert, A.; Tallec, A.; Le Magueres, P.; Ouahab, L. J. Org. Chem. 1995,
60, 2443−2447. (c) Frere, P.; Skabara, P. J. Chem. Soc. Rev. 2005, 34,
69−98.
̀
̀
(26) Valkenier, H.; Huisman, E. H.; van Hal, P. A.; de Leeuw, D. M.;
Chiechi, R. C.; Hummelen, J. C. J. Am. Chem. Soc. 2011, 133, 4930−
4939.
(27) (a) Choi, S. H.; Kim, B.; Frisbie, C. D. Science 2008, 320, 1482−
́
1486. (b) Guedon, C. M.; Valkenier, H.; Markussen, T.; Thygesen, K.
S.; Hummelen, J. C.; van der Molen, S. J. Nat. Nanotechnol. 2012, 7,
305−309.
́
(28) Gonzalez, M. T.; Leary, E.; García, R.; Verma, P.; Herranz, M.
́
A.; Rubio-Bollinger, G.; Martín, N.; Agraït, N. J. Phys. Chem. C 2011,
115, 17973−17978.
(29) Tsutsui, M.; Shoji, K.; Taniguchi, M.; Kawai, T. Nano Lett. 2008,
8, 345−349.
́
́
(30) Gonzalez, M. T.; Díaz, A.; Leary, E.; García, R.; Herranz, M. A.;
Rubio-Bollinger, G.; Martín, N.; Agraït, N. J. Am. Chem. Soc. 2013,
135, 5420−5426.
(31) Kaliginedi, V.; Moreno-García, P.; Valkenier, H.; Hong, W.;
́
García-Suarex, V. M.; Buiter, P.; Hotten, J. L. H.; Hummelen, J. C.;
K
dx.doi.org/10.1021/ja509937k | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX