Correlations between molecular structure and single-junction conductance: A case study with oligo(phenylene-ethynylene)-type wires

Article

pubs.acs.org/JACS

Correlations between Molecular Structure and Single-Junction

Conductance: A Case Study with Oligo(phenylene-ethynylene)-Type

Wires

Veerabhadrarao Kaliginedi,^†Pavel Moreno-García,^†,‡,⊗Hennie Valkenier,^§,⊥,⊗Wenjing Hong,^†,⊗

Víctor M. García-Suarez,^∥,#,⊗Petra Buiter,^§Jelmer L. H. Otten,^§Jan C. Hummelen,*^,§,⊥

́

Colin J. Lambert,*^,∥and Thomas Wandlowski*^,†

^†Department of Chemistry and Biochemistry, University of Berne, Freiestrasse 3, CH-3012 Berne, Switzerland

^‡Instituto de Física, Benemer

́ ́

ita Universidad Autonoma de Puebla, Apartado Postal J-48, Puebla 72570, Mexico

⊥

^§Stratingh Institute for Chemistry and Zernike Institute for Advanced Materials, University of Groningen, Nijenborgh 4, 9747 AG

Groningen, The Netherlands

^∥Department of Physics, Lancaster University, Lancaster LA1 4YB, United Kingdom

^#Departamento de Física, Universidad de Oviedo and CINN (CSIC), ES-33007 Oviedo, Spain

S

* Supporting Information

ABSTRACT: The charge transport characteristics of 11

tailor-made dithiol-terminated oligo(phenylene-ethynylene)

(OPE)-type molecules attached to two gold electrodes were

studied at a solid/liquid interface in a combined approach

using an STM break junction (STM-BJ) and a mechanically

controlled break junction (MCBJ) setup. We designed and

characterized 11 structurally distinct dithiol-terminated OPE-

type molecules with varied length and HOMO/LUMO

energy. Increase of the molecular length and/or of the

HOMO−LUMO gap leads to a decrease of the single-junction

conductance of the linearly conjugate acenes. The exper-

imental data and simulations suggest a nonresonant tunneling mechanism involving hole transport through the molecular

HOMO, with a decay constant β = 3.4 0.1 nm⁻¹and a contact resistance R_c= 40 kΩ per Au−S bond. The introduction of a

cross-conjugated anthraquinone or a dihydroanthracene central unit results in lower conductance values, which are attributed to a

destructive quantum interference phenomenon for the former and a broken π-conjugation for the latter. The statistical analysis of

conductance−distance and current−voltage traces revealed details of evolution and breaking of molecular junctions. In particular,

we explored the effect of stretching rate and junction stability. We compare our experimental results with DFT calculations using

the ab initio code SMEAGOL and discuss how the structure of the molecular wires affects the conductance values.

processes triggered by an electric potential,^22,23the mechanical

INTRODUCTION

Due to the intrinsic limitations of silicon-based electronics, the

■

motion of a probe electrode,²⁴or light excitation.^6,25,26

Furthermore, molecular wires are essential components for

idea of using organic molecules as functional units in electronic

connecting functional units.⁵Oligo(phenylene-ethynylene)

(OPE) compounds represent a particular unique family of

molecular wires. They are fully π-conjugated rigid rod-like

molecules with a HOMO−LUMO gap of ∼3 eV. Their

structure and functional properties are tunable over a wide

range of parameters.^27,28

devices has received increasing interest.¹The detailed under-

standing of charge transport through single-molecule junctions

is a key prerequisite for the design and development of

molecular electronic devices. The implementation of (single)

molecules and tailored molecular assemblies in electronic

circuits requires the optimization of their structures toward

desired functionalities and a reliable method of wiring them

into nanoscale junctions.²⁻¹⁰In recent years, important

correlations between transport properties and molecular

structure have been demonstrated from the investigation of

functional properties of single molecules and self-assembled

monolayers.¹¹⁻¹³Examples of reported functionalities illus-

trated by single molecules include diodes,¹⁴⁻¹⁶transistors,^17,18

memory effects,¹⁹⁻²¹and various switching phenomena, such as

Various experimental approaches have been employed to

investigate charge transport properties of OPE-type derivatives

in molecular ensembles and at the single-molecule level.^10,11

The former include crossed wire configurations,²⁹nanopores,³⁰

Au colloid arrays,³¹current−probe atomic force microscopy

Received: December 10, 2011

Published: February 21, 2012

5262

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

Figure 1. Structures of the OPE derivatives investigated. R = CH₃(CH₂)₅O.

(CP-AFM),³²⁻³⁴tuning fork-based techniques,³⁵liquid metal

junctions,³⁶and molecular layers sandwiched between litho-

graphically prepared contact electrode assemblies,³⁷as well as a

variety of electrochemical platforms.³⁸⁻⁴⁰Conductance experi-

ments with single or a few OPE-type molecules have been

addressed in matrix-isolation experiments,⁴¹mechanically

controlled break junctions (MCBJ)^19,42−45and STM break

junctions (STM-BJ),⁴⁶⁻⁵⁴and related techniques. The

experimental approaches mentioned above differ in the

formation of reproducible contacts between the molecules

and probe electrodes.⁵⁵For the reliability of the data it is very

important to demonstrate that the main features in “single”-

junction conductance experiments do not suffer from

“experimental artifacts”. In particular, STM-BJ and MCBJ

techniques provide access to measurements of a large number

of individual conductance−distance and/or current−voltage

traces under a wide range of conditions, enabling a robust and

statistically relevant data analysis. Many configurations can be

sampled and characterized quantitatively.

HOMO level.⁶²The influence of torsion angle φ was studied in

several series of biphenyl derivatives. The authors found that

the conductance follows a cos²φ correlation.^58,63,64However,

correlations between junction conductance, junction evolution,

junction stability, and molecular structure were not clearly

addressed. In particular, no comprehensive comparison of such

results as obtained from two or more different experimental

techniques (under the same experimental conditions) has been

reported for rod-like molecules of the OPE family.

In the present work we investigate charge transport

properties of 11 tailor-made dithiol-terminated OPE molecules

(Figure 1) attached to two gold electrodes in complementary

STM-BJ and MCBJ experiments. The family consists of

structurally distinct molecules with varied lengths and

energetics (HOMO−LUMO gap). The role of cross-con-

jugated building blocks is addressed as well. The statistical

analysis of individual conductance−distance and current−

voltage traces revealed details of the evolution and breaking

of single molecular junctions. In particular, we explored the

effect of stretching rate and junction stability. Finally, we

compare our experimental results with DFT calculations using

the ab initio code SMEAGOL, and discuss how the molecular

junctions evolve upon stretching and how the structures of the

molecular wires affect the junction transport characteristics.

Due to different experimental techniques and data analysis

procedures adopted by different groups, there are distinct

differences in absolute values of conductances reported for

single molecules and molecular assemblies. However, the search

for trends in structure−conductance (and/or reactivity)

correlations appears to be a very reliable and fruitful

concept.^3,5,9,56One of the key properties investigated is the

length dependence of the molecular conductance. Systematic

experimental and theoretical studies with thiol-terminated

aliphatic and aromatic molecular wires suggested a nonresonant

tunneling process as the main transport mechanism.^5,57

EXPERIMENTAL SECTION

■

Chemicals and Synthesis. The molecular wires 1−7 were

synthesized following the methodology developed by Stuhr-Hansen

et al.^65,66In short, the core of the molecular wire was constructed in a

Sonogashira cross-coupling reaction, followed by conversion of the

two tert-butyl-protected terminals (1−9A) into acetyl-protected

dithiols (Scheme 1). The isolated yields of this step varied with the

solubility of the compounds, ranging from 43% for 3 up to 91% for 1.

A slightly different protocol was used for 8 and 9.⁶⁷These air-stable

compounds 1−9 can be deprotected to form dithiolates in situ.³⁷

The synthesis of compounds 1 and 5 was reported by Stuhr-

Hansen⁶⁶and Mayor et al.⁴³Detailed procedures and spectroscopic

data of the molecules 1, 3, 4, 6, and 7 are given in the Supporting

Information (SI). For the synthesis and characterization of 8 and 9 we

refer to refs 56 and 68. The hexyloxy-substituted OPE rods 10 and 11

were obtained according to literature routes described by Zhou et al.⁶⁹

All chemicals used in the synthesis were purchased from Aldrich,

Acros, or Alfa-Aesar and employed as received, unless stated otherwise.

Conductance Measurements. The transport characteristics in

single-molecule junctions were studied using STM-BJ and comple-

mentary MCBJ measurements, both in solution and at room

However, the tunneling decay parameter β in G = A e^−βL

,

with G and L as the molecular conductance and length,

respectively, varied considerably. Examples of π-conjugated

aromatic rods include a systematic study of −NH₂-terminated

OPEs with β = 2.0 nm⁻¹.⁵¹Venkatamaran et al. reported a

larger value of 4.0 nm⁻¹for oligophenylenes, in good

agreement with calculations.⁵⁸Alkoxy substituents at the

central phenyl ring, which are often used to increase the

solubility of OPE molecules, do not appear to influence the

conductance values.⁴⁴Smaller β values were found for

oligoynes,⁵⁹Zn-porphyrin-containing wires,⁶⁰and oligothio-

phenes.⁶¹Related studies on a series of conjugated molecules

have shown that the conductance decreases upon increasing the

oxidation potential, i.e., upon lowering the energy of the

5263

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

Production, 500 nm s⁻¹) with a resolution of 5 nm per step and a

moving distance of up to 1 cm, and a piezo stack on top (8 μm moving

distance). The tunneling current between the two ends of the “broken

wire”, which act at a given bias voltage (typically between 0.020 and

0.200 V) as working electrodes WE1 and WE2, was utilized as a

feedback signal. The pushing process started with the stepper motor.

Once a current decrease was detected, which represents the breaking

of the gold−gold contact, the stepper motor was paused, and the z-

motion control was switched to the piezo stack. The voltage output of

the piezo stack communicates with an onboard trigger. The latter

senses the transport respective tunneling current and leads to the

conductance. When the conductance reaches the noise threshold (G <

10⁻⁸G₀) or the high contact limit (set to 10G₀), the voltage ramp for

the piezo stack stops for a preset time of 0.5 s, and subsequently

decreases or increases, respectively. The cycle is repeated more than

2000 times to obtain statistically relevant data.

The MCBJ controlling unit is based on a lab-built bipotentiostat

with two bipolar tunable logarithmic I−V converters as current

measuring units, which are operated by a custom-designed micro-

controller.⁷⁴The unit provides three analog signals. The first one

controls the potential of WE1, which is particularly important for

advanced electrochemical experiments with the MCBJ setup. The

second one controls the voltage difference between the two working

electrodes WE1 and WE2 (bias voltage V_bias), which drives the current

through the two gold electrodes for the conductance measurements.

The third channel controls the voltage output of the piezo stack in the

range of 0−50 V, allowing the displacement of the piezo stack up to 8

μm with rates ranging from 3 to 3000 nm s⁻¹, which translates into

lateral pulling (pushing) rates between the two gold leads of 0.1−100

nm s⁻¹, with a positional precision better than 0.05 nm. The distance

between the two gold electrodes in the MCBJ setup was calibrated

with complementary STM-BJ experiments assuming that the tunneling

decay is identical under the same experimental conditions.

The entire setup was placed in two Faraday boxes, one for the

mechanical unit and the other for the controller unit in order to avoid

electronic cross-talk between the different functional parts. The full

technical description of the setup is reported by Hong et al.⁷³

The sample templates were spring steel sheets (30 mm × 10 mm ×

0.2 mm), which were cleaned in boiling 25% nitric acid (Merck AG,

65%, pro analysis) and Milli-Q water (18.2 MΩ cm⁻¹, 2 ppb TOC),

and dried in a stream of argon. A gold wire (Goodfellows 99.999%,

100 μm diameter) was fixed on these sheets with two drops of

preheated epoxy (Stycast 2850FT with catalyst 9) and cured overnight

at 60 °C. The freely suspended part of the gold wire (less than 500

μm) was notched with a scalpel blade to fabricate a constriction point.

The as-prepared sample sheets were cleaned in boiling Milli-Q water

for 15 min, rinsed with isopropanol (Aldrich, pro analysis), dried with

argon, and mounted on the sample holder of the MCBJ setup. Finally,

the precleaned Kel-F liquid cell was installed on top of the sample with

a Kalrez O-ring to avoid leakage of the solution. The closed liquid cell

was flushed with argon through an inert cycling system to remove

oxygen, and the test-molecule-containing solution was pumped into

the cell in several cycles. Finally, the input and output valves for

solution and gas purging were closed, and the experiment started.

Deprotection of the compounds 1−7 was carried out in situ by

adding tetrabutylammonium hydroxide just before the start of the

transport experiment in the STM configuration. For comparison, we

also carried out complementary studies with molecules 1−11

employing both the STM-BJ and MCBJ setups in the absence of

deprotecting agents.

Scheme 1. Schematics of the General Synthesis of Molecular

Wires 1−9

a

A represents the acetyl-protected and B the tert-butyl-protected

molecular wires.

temperature. The STM-BJ technique repeatedly traps molecules

between a sharp STM tip and an atomically flat sample.^70,71The

MCBJ approach is based on the formation and breaking of molecular

junctions between notched and freely suspended gold wires supported

on insulated sheets of spring steel.^72,73

The STM-BJ measurements were carried out with a Molecular

Imaging PicoSPM housed in an all-glass argon-filled chamber and

equipped with a dual preamplifier⁷⁴capable of recording currents in a

wide range of 1 pA to 150 μA with high resolution. The nonamplified

low-current signal was fed back to the STM controller, preserving the

STM imaging capability. The current−distance measurements were

performed with a separate, lab-built analog ramp unit. For further

technical details we refer to our previous work.^63,64

The sample electrodes were Au(111) disks, 2 mm height and 10

mm in diameter, or gold single-crystal bead electrodes. The Au(111)

substrates were flame-annealed prior to use. A freshly prepared

solution containing typically a 0.1 mM concentration of the respective

molecule (1−11) in a mixture of 1,3,5-trimethylbenzene (TMB,

Aldrich, p.a.) and tetrahydrofuran (THF, Aldrich, p.a.), volume ratio

4:1, was added to a Kel-F flow-through liquid cell mounted on top of

the sample. The STM tips were uncoated, electrochemically etched

gold wires (Goodfellow, 99.999%, 0.25 mm diameter), capable of

imaging with atomic resolution.

The following protocol was applied after assembling a new

experiment: The tip was brought to a preset tunneling position

typically defined by i_T= 50−100 pA and a bias voltage V_bias= 0.10 V,

followed by imaging the substrate. Current−distance measurements

were performed at a fixed lateral position with the STM feedback

switched off and the vertical tip movement controlled by the ramp unit

described above. The measuring cycle was performed in the following

way: The controlling software drove the tip toward the adsorbate-

modified surface. The approach was stopped when a predefined upper

current limit was reached (typically 10 μA or <10G₀, with G₀being the

fundamental conductance quantum, 77.5 μS). After a short delay

(∼100 ms) ensuring tip relaxation and the formation of stable

contacts, the tip was retracted by 2−5 nm until a low current limit of

∼10 pA was reached. The approaching and withdrawing rates were

varied from 2 to 160 nm/s. The entire current−distance traces were

recorded with a digital oscilloscope (Yokogawa DL 750, 16 bit, 1 MHz

sampling frequency) in blocks of 186 individual traces. Up to 2000

traces were recorded for each set of experimental conditions to

guarantee the statistical significance of the results. For each molecule

the data were acquired at three different bias voltages of 0.065, 0.100,

and 0.170 V.

Details of the data analysis are reported in the SI.

Theoretical Methods. The electronic and transport properties of

the OPE derivatives were obtained using the ab initio code

SMEAGOL,^75,76which employs the Hamiltonian provided by the

density functional theory (DFT) code SIESTA,⁷⁷in combination with

the nonequilibrium Green's function formalism. SIESTA uses

nonconserving pseudopotentials to account for the core electrons,

and a linear combination of pseudoatomic orbitals to span the valence

states. The calculations used a single-ζ basis set for the gold leads,

which included the s- and d-orbitals in the valence. An energy cutoff of

The MCBJ experiments are based on the opening and closing of

nanogaps formed by notching a freely suspended, horizontally

supported gold wire (99.999%, Goodfellow, 100 μm diameter) in a

sample-molecule-containing solution, as controlled by the vertical

movement of a pushing rod. The motion control is based on a

combination of a stepper motor (Accu-coder 95511, Encoder

5264

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

200 Ry was chosen to define the real-space grid, which is used to

represent the density and the potential, and to calculate the

Hamiltonian and the overlap matrix elements. The local density

approximation (LDA)⁷⁸was employed to account for exchange and

correlation effects. The Fermi distribution function was smoothed with

a temperature of 100 K to improve the convergence of the results. The

molecular coordinates were relaxed in the isolated molecule until the

forces were smaller than 0.05 eV/Å, and then included into the

“extended molecule” without further relaxation. The system was made

periodic along the perpendicular directions x and y, and only the

gamma point was used. The “extended molecule” is represented by the

dithiolated OPE compound bridged to five atomic gold layers to

account for specific properties of the contact region such as adsorption

geometry, molecular conformation, etc. Each layer contains 18 atoms,

i.e., 6 and 3 atoms along each lattice vector on the (111) surface. This

arrangement was necessary to avoid overlaps with molecular images

along the perpendicular directions (see also SI).

SMEAGOL divides the entire nanoscale junction into three parts:

the left and the right bulk electrodes, simulated by three gold layers

grown along the (111) direction, and the “extended molecule”.

SMEAGOL uses the Hamiltonian derived from SIESTA to calculate

self-consistently the density matrix, the transmission coefficients T(E)

of electrons from the left to the right lead, and the I−V characteristics.

Corrected positions of the Fermi energy relative to the HOMO

were obtained by applying a scissors-type operator to the bare DFT

results (SAINT⁷⁹⁻⁸¹) that moves the occupied and the unoccupied

molecular levels downward and upward, respectively. The spectral

adjustment was chosen to fit the position of such levels as obtained

from a combination of experimental UV photoelectron spectroscopy

(UPS), UV/vis absorption, and electrochemical data. Further details

about the theoretical method and the energy level alignment are given

in the SI.

Figure 2. (A−C) Conductance measurements of 2 employing the

STM-BJ approach. (A) Typical conductance distance traces recorded

at V_bias= 0.1 V. (B) 1D conductance histogram and (C) 2D histogram

generated from 2000 individual curves. (D−F) Conductance measure-

ments of 2 employing the MCBJ approach. (D) Typical conductance

distance traces measured at V_bias= 0.1 V. (E) 1D conductance

histogram and (F) 2D conductance histograms generated from 2000

curves.

RESULTS AND DISCUSSION

■

OPE Compound 2: Conductance Measurements and

Molecular Junction Evolution. Conductance Measure-

ments Based on Stretching Traces. Figure 2A displays typical

conductance−distance stretching traces, plotted on a loga-

rithmic scale, as recorded for the deprotected compound 2 in

TMB/THF by the STM-BJ upon application of a bias voltage

V_bias= 0.10 V between the gold STM tip and the Au(111)

substrate. After the formation of the contact between the tip

and the substrate, the tip was withdrawn with a rate v = 58 nm

s⁻¹. All curves show initially a stepwise decrease of the

conductance from ∼10G₀to 1G₀, G₀= 2e²/h = 77.5 μS being

the quantum of conductance, with quantized conductance steps

occurring at integer multiples of G₀. Structurally the evolution

of the conductance corresponds to the decrease of the number

of gold atoms in the constriction upon junction elongation until

the contact with only one gold atom in the cross-section is

formed. Subsequently, the current abruptly decreases by several

orders of magnitude (“jump-out-of-contact”) upon stretching,

and additional features are typically observed at G < 10⁻³G₀

before the conductance reaches the noise level of ∼10⁻⁷G₀.

We distinguish three types of measured curves: 25−30% of

all traces exhibited after the “jump-out-of-contact” a bare

tunneling current, exponentially decaying until the noise level is

reached (type I, black traces in Figure 2A; type I curves are

attributed to traces with no molecules trapped between the tip

and substrate); 5−10% of all curves are rather noisy and exhibit

fluctuations, which are typically attributed to mechanical

instabilities (type II, gray); and the remaining 60−70% of all

traces recorded show well-defined plateaus after breaking of the

gold contact (type III, blue traces in Figure 2A). The plateaus

occur in a narrow conductance range around 10⁻⁴G₀and are

characterized by a slight decrease of the conductance with

stretching distance Δz, before it drops abruptly. In 50% of these

traces, short plateaus with conductances around 10⁻⁵G₀were

also observed. Both types of conductance plateau were absent

in control experiments in the absence of compound 2. We

assign these plateaus to the conductance of junctions formed by

molecules bridging the gap between the two gold electrodes.

Without any data selection, 2000 conductance−distance

traces, measured under the same conditions, were analyzed

further. Figure 2B displays the one-dimensional (1D) histogram

of the logarithm of the conductance log(G/G₀)⁸²constructed

from all data points of 2000 curves of 2 (V_bias= 0.10 V). The

sharp peak in the histogram around 0 represents the

conductance of a single-atom gold−gold contact.⁸³The 1D

histogram displays a prominent peak G_Hwith a maximum at 1.8

× 10⁻⁴G₀(13.9 nS) and a shoulder at lower G in the entire

range between 0 and −6 (in units of of log(G/G₀)), the latter

being the lower limit of a reliably measurably conductance in

our STM-BJ setup (which is indicated as “noise” in Figure

2B,C). The shoulder was deconvoluted as a small peak centered

at G_L= 9.7 × 10⁻⁶G₀(0.76 nS) by fitting Gaussians to the

histogram. We attribute G_Hand G_Lto the two most probable

conductance states of a single molecule 2 bound to two gold

leads. This assignment is supported by a two-dimensional (2D)

histogram^64,84constructed as follows. First, we normalized all

individual conductance traces to a common distance scale by

assigning Δz = 0 at G = 0.7G₀.⁶⁴This procedure is justified by

the sharp drop in conductance just after G₀. The conductance−

distance histogram was then constructed by counting the

5265

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

occurrence of [log(G/G₀);Δz] pairs in a 2D field. Further

details of this analysis procedure are described in the SI.

2D histograms constructed from traces such as in Figure 2A

(Figure 2C) show features of gold quantum contacts at G ≥ G₀,

a second main cloud-like feature at [0 ≤ Δz ≤ 1.25 nm; 10⁻³

G₀≥ G ≥ 2 × 10⁻⁵G₀] with a center around 1.2 × 10⁻⁴G₀, and

a satellite feature at [1.10 nm ≤ Δz ≤ 1.25 nm; 10⁻⁵G₀≥ G ≥

10⁻⁶G₀] centered around 9.7 × 10⁻⁶G₀. The latter two features

correspond to the conductance peaks with maxima at G_Hand

G_L, as observed in the 1D histogram (Figure 2B). Note that the

high-conductance feature H starts immediately after breaking of

the gold−gold contact, while the low-conductance range L

starts only after feature H at Δz ≈ 1.10 nm. The experimentally

observed decrease of the conductance in region H with

increasing Δz (Figure 2C) might reflect a continuous decrease

of the tilt angle between the “trapped” molecular wire 2 and the

surface normal upon pulling.⁵²

Figure 3. (A) Conductance histograms of 2 constructed from STM-BJ

measurements with stretching rates of 1.45 (black), 58 (green), and

145 nm s⁻¹(blue), V_bias= 0.1 V. (B) Most probable molecular

junction conductances G_Hand G_Lvs the logarithm of the stretching

rate v. The star symbols represent an artifact related to the noise level

of our STM-BJ setup.

Complementary experiments were carried out with the

acetyl-protected compound 2 in the STM-BJ and the MCBJ

setups. The results of the MCBJ experiments are summarized in

Figure 2D−F. The amount of curves displaying features of

molecular junctions increases up to 80% of all measured curves,

which we attribute to the overall higher mechanical stability of

our MCBJ setup. We also note that the MCBJ setup has an

electrical noise level of ∼10⁻⁸G₀(indicated as “noise” in Figure

2E,F) and thus allows measurements of lower conductances

than the STM-BJ setup. Individual traces (Figure 2D), the 1D

histogram (Figure 2E), and the 2D histogram (Figure 2F)

demonstrate a single peak with a maximum at 1.3 × 10⁻⁴G₀(10

nS) and a cloud-like feature at [0 ≤ Δz ≤ 1.30 nm; 10⁻³G₀≥ G

≥ 3 × 10⁻⁵G₀] corresponding to the high conductance of

molecular junctions, in good agreement with the STM-BJ

experiments. No low-conductance feature was observed in

experiments with the acetyl-protected molecular wire 2. A

Figure 4. Characteristic lengths distributions for the stretching of

molecular junctions of 2 from STM-BJ experiments, and analyzed in

the range (A) from G = 0.7G₀to G = 10⁻⁵G₀(high-conductance

range) and (B) from G = 0.7G₀to G = 10⁻⁶G₀(high- and low-

conductance range), with stretching rates of 145 (black), 58 (green),

and 1.45 nm s⁻¹(blue), V_bias= 0.10 V. (C) Most probable

characteristic length Δz^L* as a function of the stretching rate ln v.

(D) Dependence of the characteristic length for the stretching of

gold−gold junctions (in G = 1.1G₀to G = 0.7G₀) on ln v. The

experimental conditions are the same as in panel C. (E) Characteristic

length distribution for the stretching of molecular junctions with 2 (in

G = 0.7G₀to G = 10⁻⁶G₀) measured in MCBJ experiments with 2 nm

s⁻¹, V_bias= 0.1 V.

similar observation was also reported by Gonzal

́

ez et al.⁵⁴

However, the low-conductance junctions reappeared when the

aromatic solvent TMB was replaced by the nonpolar aliphatic

solvent decane (see SI).

Finally, we comment that the high-conductance value G_Hfor

compound 2 reported here is comparable with values published

previously by Tao and co-workers (6−13 nS),^46,85,86Xing et al.

(10 nS),⁵⁰and Liu et al. (3.6 nS)⁴⁷as based on STM-BJ

measurements. Haiss et al.⁵²found ∼10 nS in their I(t) and I(s)

experiments.^52,87Huber et al.⁴⁴and Wu et al.⁴⁵obtained values

of ∼9 nS by an MCBJ technique.

The stability of molecular junctions formed by compound 2

attached to two gold electrodes was explored in the STM-BJ

setup by varying the stretching rate v from 1 to 145 nm s⁻¹. The

1D histograms (Figure 3A) demonstrate that the most probable

conductance values G_Hand G_Lare rather independent of the

stretching rate (Figure 3B). However, the width (fwhm) of the

corresponding histogram peaks increases slightly with decreas-

ing v. This trend might reflect contributions of thermal

fluctuations, atomic-scale variations in the geometry of the

molecule−electrode contacts, conformation changes in the

molecule, and other factors.⁸⁸

A distinct rate dependence was obtained in STM-BJ

experiments for the distance Δz over which the molecular

junction could be stretched before breakdown (Figure 4).

The plot was constructed on the basis of a statistical analysis

of all experimental conductance versus displacement traces (cf.

1D histogram in Figure 3 and the corresponding 2D histograms

in SI, section 5) in 0.7G₀(the displacement Δz is set to zero at

this point for each individual trace) up to a lower limit G_min

,

typically 1 order of magnitude below the most probable high-

(G_H) or low (G_L)-conductance states, respectively. Technical

details of this approach were described in our previous

publication⁶²and are briefly outlined in the SI. Figure 4A,B

illustrates distributions of Δz^Hand Δz^Ldetermined from all

experimental traces as measured at the three different rates v.

The distributions show two peaks labeled as T and M. The T

peak centered at Δz ≈ 0.3−0.5 nm results from traces showing

only through-space tunneling contributions, as verified in

experiments with adsorbate-free solutions. The second peak

labeled M reflects properties of a “true” molecular junction.

The maximum of the M peak as determined by Gaussian fits

represents the most probable characteristic distance Δz* over

which a junction can be stretched. With an increase of the

stretching rate from 1.45 to 145 nm s⁻¹, Δz* increases for the

high-conductance feature from Δz^H* = 0.90 to 1.50 nm (Figure

5266

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

4A). Extending the analysis to the low-conductance region until

breakdown is observed, Δz^L* ranges from 1.10 to 1.70 nm. The

difference Δz^L* − Δz^H* amounts to ∼0.20 nm and is rather

independent of the stretching rate. We also calculated the

junction formation probability from the relative areas of the

peaks T and M from the displacement histograms in Figure 4B,

as defined by Gaussians, and obtained 75%, 85%, and 65% for

stretching rates of 1.45, 58, and 145 nm s⁻¹, respectively.

The experimentally observed dependence Δz^L* vs ln v for

the overall stretching process of 2 (Figure 4C) indicates a

thermally activated bond-breaking process.⁸⁸⁻⁹¹The values of

Δz^L* for 2 and Δz^Au* are rather independent of the loading

rate below 60 and above 140 nm s⁻¹, which reflects quasi-

equilibrium conditions and could be attributed to a

spontaneous (due to the thermal fluctuations) or adiabatic

(external force that overcomes the binding energy barrier)

breakdown regime.^89,90At intermediate stretching rates, a

region of a distinct logarithmic dependence of Δz^Au* on ln v is

observed. Similar trends were reported before by Huang et al.⁸⁹

for the breaking of octanedithiol junctions formed between two

gold electrodes. The authors reported values of Δz* = 0.1 and

0.2 nm in the limits of low and high stretching rates,

respectively, rather similar to their and our results for the

breaking of monatomic gold contacts (Figure 4D). This

comparative study let the authors suggest that the breakdown

of octanedithiol-based molecular junctions takes place most

likely at gold−gold bonds. This hypothesis is supported by

single-molecule electromechanical studies using CP-AFM.⁹²In

case of OPE molecules, the values of Δz^L* are distinctly larger

(Figure 4C) in the two quasi-equilibrium stretching regimes, as

compared to the breaking of a gold−gold atomic contact, but

still smaller than the length of the molecule. Clearly, the

formation and subsequent breakdown of junctions containing

OPE-type molecules involve processes such as picking up the

molecule from the surface,⁹³coordination and/or conformation

changes,⁷¹and/or sliding along the contacting leads before

rupture.²⁴Further details will be addressed in the Discussion.

The higher mechanical stability of our MCBJ setup,⁷³as

compared to a typical STM-BJ experiment (see SI for details),

allows resolving simultaneously both quasi-equilibrium states as

characterized by a low and a higher characteristic stretching

length Δz* (Figure 4E), even at rather low rates of opening/

closing the nanocontact junction. The corresponding histo-

grams (see SI) show a clear trend toward a more probable

higher characteristic length Δz* with increasing stretching rate.

Current−Voltage Measurements. The high mechanical

stability of the MCBJ setup also offers the possibility to

acquire simultaneously I−V_biascurves during the slow opening

and closing of a molecular junction with a typical rate of 0.5 nm

s⁻¹(Figure 5). The bias voltage was ramped between −1.00

and +1.00 V at 40 V s⁻¹,⁷³allowing us to measure 40−50 I−

V_biascurves per individual trace. The curves measured during

the different stages of extension correspond to gold−gold

contacts (Figure 5A), molecular junctions (Figure 5B), and the

tunneling response through the solvent (Figure 5C). The zero-

bias conductances (Figure 5D) were calculated from the slopes

of the linear parts of the I−V_biascurves in the range from −0.30

to +0.30 V. The 1D log-conductance histogram as constructed

from ∼1000 individual data points is displayed in Figure 5E.

The estimated most probable junction conductance of 2 in

region G_His in good agreement with data shown in Figure 2E,

which supports the reliability of our I−V_biasexperiments.

Sensitivity limitations in the setup for I−V_biasmeasurements at

Figure 5. Typical I−V_biascurves measured (A) for gold−gold contacts,

(B) for molecular junctions and (C) in the tunneling regime during a

single stretching trace of 2, V_bias= 0.10 V, pulling rate v = 0.5 nm s⁻¹.

(D) Zero-bias conductance as calculated from the slopes of individual

I−V_biastraces during 4000 stretching sequences. (E) Conductance

histograms built from zero-bias conductance of 4000 I−V_biascurves

(red) and from the conductance−distance traces (black) simulta-

neously measured with the MCBJ setup.

larger values of V_biasprevented us from obtaining reliable data

in the G_Lregion. We also note that some molecular junctions

are broken before being completely stretched (examples are the

two left traces in Figure 5D), which we attributed to enhanced

thermal fluctuations of the junction (“junction heating”) at high

bias voltages.

Figure 6. (A) 2D histogram of 4000 I−V_biascurves measured for 2

with the MCBJ setup in the high-conductance region of molecular

junctions. (B) Statistically significant I−V_biastrace (orange dots)

calculated from 4000 individual I−V_biascurves of 2 and DFT-based I−

V_biascurve (blue line). (C) Computed transmission trace and SAINT-

adjusted value of E_F(for details see Theoretical Methods and SI).

To obtain the most probable I−V_biascharacteristics of the

high-conductance molecular junction, we analyzed statistically

the I−V_biascurves with zero-bias conductance between 2 ×

10⁻⁵G₀and 10⁻³G₀.^73,94,95Figure 6A shows the 2D I−V_bias

histogram constructed from 4000 I−V_biastraces. For each value

of the bias voltage we determined the maximum of the

corresponding current distribution (obtained from Gaussians

fits) as the most probable current.^73,95The resulting “statisti-

cally significant” I−V_biascurve is plotted as orange dots in

Figure 6B. The curve shows fluctuations, but no characteristic

resonance features. We note that the conductance of molecular

5267

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

junctions in the nonresonant tunneling regime is controlled by

(i) the position of the dominant transport MO relative to the

metal Fermi level and (ii) the broadening of this MO due to the

coupling to the metal leads.⁵Comparison of the experimentally

most probable I−V_biascurve of 2 (dots in Figure 6B) with the

computed I−V_biascharacteristics (blue solid line in Figure 6B),

as obtained from ab initio generated transmission curves

(SMEAGOL, for details see SI), allowed us to adjust the scale

of E_HOMO− E_Fas obtained from simulations. The current was

calculated by integrating the respective transmission curve in

the corresponding bias window. The Fermi level that gave the

best agreement was slightly higher in energy (0.4 eV, blue

vertical line) than the original Fermi level.

The agreement is rather good despite the fact that

nonequilibrium effects, such as the movement of resonances

under an applied bias voltage, were not taken into account

(these and other effects, such as bias-induced destruction of

resonances, start to dominate at higher voltages). The I−V_bias

curve is typical of a system with the Fermi level in the

HOMO−LUMO gap and close to a resonance; i.e., the current

is ohmic at low voltages but changes its slope and increases

more sharply at higher voltages. Figure 6C illustrates the

SAINT-corrected transmission curve (see SI for more details),

which demonstrates that the resonance closest to the Fermi

level E_Fin the modified energy scale represents a HOMO

feature, indicating that HOMO-mediated hole transport is

likely to dominate the transport behavior of 2 attached to two

gold leads. This conclusion is in agreement with previous work

on dithiol-terminated OPE rods attached to gold leads⁵³and is

directly supported by thermopower measurements in phenyl-

dithiol-containing heterojunctions.⁹⁶

Figure 7. 1D conductance histograms of 1−9 and 11, measured in the

STM-BJ setup (1−6) or the MCBJ stage (7−9, 11), at V_bias= 0.1 V.

The STM-BJ- and MCBJ-generated histograms are both based on the

analysis of 2000 individual traces.

Effect of Molecular Structure on Conductance. The

Most Probable Conductance. Following the experimental

approaches and data analyzing strategies introduced for the

reference compound 2, we designed and characterized 11

dithiol-terminated OPE-type molecules with varied length,

HOMO−LUMO energy, and conjugation of the π-system. We

take the distance between two sulfur atoms obtained from DFT

calculations of the relaxed extended molecules (see Theoretical

Methods and Table S2 in SI for details) as a transport-relevant

molecular length, L_m. The values of L_mrange between 1.32 (1)

and 3.39 nm (11). The HOMO−LUMO gap was estimated

experimentally from UV/vis absorption spectra of the tert-

butyl-terminated dithiols. All compounds are fully π-con-

jugated, except the anthraquinone-based cross-conjugated

compound 8 and the dehydroanthraquinone derivative 9 with

broken π-conjugation. Tables S1 and S2 (SI) summarize

characteristic geometric and energetic parameters of these

molecules.

In the following we discuss key experimental results obtained

with a stretching rate of either 58 (STM-BJ) or 2 nm s⁻¹

(MCBJ). These rates ensure that all experiments are carried out

in the “quasi-equilibrium” region of the spontaneous break-

down (cf. Figure 4). Original data for all investigated

compounds are given in the SI.

Figure 7 displays the 1D log-conductance histograms for 10

compounds. Figure 8A and Table S3 (SI) summarize all

characteristic conductance data. The data plotted in Figure 7A

demonstrate that the main conductance features of G_Htype

shift to lower values with increasing molecular length. In

particular, the most probable conductance decreases from 1.46

× 10⁻³G₀(1) to 10⁻⁶G₀(11). Comparing the data for

compound 2 and the molecular wires included in Figure 7B

Figure 8. (A) Experimental conductances G_Hand G_Las a function of

the molecular length L_m(the inset shows the low-conductance G_Las a

function of L_m). (B) High-conductance G_Hplotted vs L_m(E_F

−

E_HOMO

)

^1/2. (C) DFT-based G_Hvalues (black circles and black line) as

a function of L_m. For comparison, also the experimental values are

given as open circles and blue line. The numerical conductance values

and molecular lengths are summarized in Tables S2 and S3 (SI).

Computational details are given in the Theoretical Methods section

and in the SI.

reveals a similar trend with increasing HOMO−LUMO gap.

Nearly identical most probable conductance data were obtained

in the STM-BJ and the MCBJ experiments, except for wires 6−

9 and 11. Reliable data for these molecules were accessible only

in the MCBJ configuration because of sensitivity limitations of

the STM setup in the low-bias regime. Compounds 2 and 10

gave identical results, suggesting that the hexyloxy substitution

of the central phenyl rings has no significant effect on the

transport close to the Fermi level. This hypothesis is supported

by DFT calculations reported by Martin et al.⁸⁷

5268

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

Comparing our results in the high-conductance region G_Hto

those reported by other authors reveals that our values for 1

and 5 are much higher (1.46 × 10⁻³G₀vs 0.54 × 10⁻⁴G₀by

Beebe et al.³⁴) or lower (0.47 × 10⁻³G₀vs 7.7 × 10⁻³G₀by

Reichert et al.⁴²). These differences may be partially related to

the different experimental techniques used as well as to

environmental conditions chosen. The situation is more

consistent with dithiolated OPE derivatives with three coupled

phenyl rings, such as 2 and 10, where conductances ranging

between 5.0 × 10⁻⁵G₀and 1.8 × 10⁻⁴G₀were pub-

lished.^{44,45,50−52,85}Comparing further the dithiol-terminated

wires 1 and 2 with the corresponding diamines^47,51,95shows

higher single-junction conductance data for the dithiol

derivatives. The same trend was also found for substituted

benzene and biphenyl derivatives.^{48,50,58,63,97}Lu et al. reported

the opposite for oligo(p-phenylene-ethynylene) rods with more

than three phenyl rings, based on a combined STM-BJ and CP-

AFM study.⁵¹These authors attributed their observations to an

intrinsic change of the transport mechanism from tunneling to

hopping.

monotonic increase of the most probable conductances with

decreasing HOMO−LUMO gap.

Molecules 4 and 5 thus do not follow the general G_Hvs L_m

trend displayed as the blue line in Figure 8A. The latter,

however, is only a part of the more general dependence that

can be expressed by the equation for tunneling over a distance

L through a rectangular energy barrier with the height Φ (in the

low bias limit):¹⁰²

−1 −1/2

G ≈ exp(−10.12 nm eV

Φ L)

(1)

For the through-molecule tunneling we attribute the

tunneling distance L to the molecular length L_m, and the

energy barrier Φ to the difference between the Fermi energy of

the gold electrodes E_Fand the energy of the closest molecular

orbital. The simulations of charge transport in various

molecules bound via thiol anchors to gold electrodes

demonstrate unequivocally that the transport proceeds by

HOMO-assisted hole-tunneling (see also Theory section).⁵

50,52

Consequently, we set Φ = E_F− E_HOMO

.

We estimated the

Figure 8 summarizes the dependencies of the most probable

conductances G_H(high) and G_L(low) versus the molecular

lengths L_m. The analysis of the high-conductance state G_Hof

the single molecular junctions formed by the linearly

conjugated compounds 1−3, 6, 7, 10, and 11 demonstrates

an exponential distance dependence with a decay parameter

values of E_HOMOby using DFT calculations. However, since

DFT underestimates the HOMO−LUMO gap, the energy

levels were adjusted on the basis of UPS measurements of the

respective anthraquinone derivatives (see SI for details). E_F=

5.0 eV was taken from the work reported by Veenstra et al.¹⁰³

To consider both the length and the influence of the

tunneling barrier, we replotted the log(G_H/G₀) data of all 11

β_GH= 3.4

0.1 nm⁻¹. No indication for a change of the

compounds versus L_m(E_F− E_HOMO)

^1/2(Figure 8B). Except for

transport mechanism from tunneling to hopping, as recently

discussed for a series of diamino-substituted OPE derivatives,

was found.⁵¹By extrapolating the G_Hversus L_mdependence to

L_m= 0, we determined an effective contact resistance for two

Au−S bonds equal to 80 kΩ, or 40 kΩ per Au−S site. This

value is in agreement with literature data reported for Au−

dithiolate−Au junctions.^57,98

Furthermore, extrapolation to L_m= 0.619 nm, the S−S

distance in benzenedithiol, yields a conductance of 0.012G₀, in

good agreement with literature data.^2,50,86,97The value of β_GH

found in our study is identical to those reported in

electrochemical electron transfer experiments with ferrocene-

terminated OPE wires by Creager et al.⁴⁰and Newton et al.,³⁹

and larger than those determined in CP-AFM experiments (2.1

nm⁻¹by Liu et al.³³) as well as for OPEs with amine anchoring

groups (2 nm⁻¹by Lu et al.⁵¹) and carbodithiolate linkers (0.5

nm⁻¹by Xing et al.⁵⁰). Tomfohr et al.⁹⁹analyzed electron

transport in OPE-type oligomers on the basis of a Landauer

formalism in combination with a band structure analysis and

obtained β_th= 2.3 nm⁻¹for a planar conformation. These

authors also demonstrated that β_thincreases with increasing

torsion angle between adjacent phenyl rings.

The decay parameter β_GHobtained in this experimental OPE

study compares with that of other typical molecular bridges as

follows: saturated alkanes^57,70,71(7−10 nm⁻¹) > oligophen-

yls^100,101(PP, 3.5−5 nm⁻¹) > OPE (this work) > oligo-

(phenylene-vinylenes)^39,57(OPV, 1.7−1.8 nm⁻¹) > oligothio-

phene⁶¹(OT, 1 nm⁻¹) > oligoynes⁵⁹(OY, 0.6 and 0.06 nm⁻¹).

This trend reflects an enhanced electronic communication

across the molecular bridge with increasing delocalization of the

adjacent π-system and lowering of the HOMO−LUMO gap.³⁹

We also note that molecules having the basic motif of

compound 2 with a fixed S−S distance but varying in the

structure of the central aromatic core, such as 2, 4, and 5, lead

to a decrease of the HOMO−LUMO gap (Tables S1 and S2)

in the sequence 2 > 4 > 5. Figures 7B and 8A demonstrate a

8 and 9, the data points of the nine fully conjugated OPE

analogues display a clear linear dependence. The fit is

represented by the blue line in Figure 8B. Thus, the simple

tunneling eq 1 containing both the energetic and the geometric

parameters describes the conductance variation. There are,

however, deviations from this linear behavior.¹⁰⁴In particular,

compounds 8 and 9 have the same “skeleton” as 7, but a

different electronic structure. The conductance of 9 is lower

than that of 7 due to the broken π-conjugation in the center of

the molecule, which cannot be accounted for just by the

variation of the HOMO level. The cross-conjugated anthraqui-

none-based compound 8 has an even lower conductance.

Thygesen et al.^105,106demonstrated in recent simulation studies

that the cross-conjugation in 8 results in a destructive quantum

interference effect, leading to a sharp drop in the transmission

probability of the electrons at the Fermi energy.¹⁰⁶

The inset in Figure 8A also shows the length dependence of

the most probable conductances G_Lin the low-conductance

region. This conductance feature was observed for six

compounds only. While the conductances for the shorter

molecular wires 1, 2, 4, and 5 vary only slightly from 10⁻⁵G₀

(2) to 2.6 × 10⁻⁵G₀(4), a strong decrease is observed for the

fused linearly conjugated rods 6 and 7. The absence of the G_L

feature for the compounds 3 and 11 might reflect the additional

conformation degrees of freedom due to phenyl ring rotations⁹⁹

as well as the steric influence of the hexyloxy substituents in the

latter molecule. The second distance decay parameter

determined for compounds 2, 6, and 7 (inset in Figure 8A)

amounts to β_GL= 14.6 0.35 nm⁻¹, which is significantly larger

than β_GHas well as decay parameters for other typical molecular

bridges reported so far.

2D Histograms, Stretching Length, and Tilt Angle.

Inspection of the 2D conductance−displacement histograms

reveals a slight but distinct decrease of the conductance during

the stretching for all investigated compounds (see Figure 2 and

5269

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

Figures S7−S19, SI). For compounds 1, 2, 4−7, and 10, a

second feature with lower conductances evolves in experiments

carried out in TMB/THF. The feature is even more

pronounced when decane is used as a solvent, and when

working with deprotected molecules. The possible origin of the

low-conductance feature will be addressed further.

Following the strategy presented above for compound 2, we

analyzed the characteristic length distributions upon stretching

in the high-conductance range, and up to the low-conductance

region. They were defined as distances over which the tip

moves while the junction conductance drops from 0.7G₀to

0.1G_Hfor Δz^Hor from 0.7G₀to the noise level (10⁻⁶G₀for

STM-BJ and 10⁻⁸G₀for MCBJ measurements) for Δz^L. Pulling

rates of 58 and 2 nm s⁻¹were employed in STM-BJ and MCBJ

experiments, respectively.

All length distribution histograms for compounds 1 and 3−

11 are presented in the SI. The most probable characteristic

lengths Δz^H* and Δz^L* increase with molecular length L_m

(Table S3). For experiments with two conductance features

resolved, the difference Δz^L* − Δz^H* amounts to 0.20−0.35

nm rather independently of molecule and solvent. This value is

close to the diameter of a gold atom and might indicate that the

pulling of an additional gold atom is involved in the junction

transition from the high to the low-conductance state.

In order to estimate the absolute distances between the two

gold electrodes in the most probable configurations prior to

breaking, we need to consider the “snap-back”, i.e., the fast

relaxation of gold electrodes upon breaking a monatomic gold−

gold contact. This effect is in particular responsible for the

sharp decrease of the current from ∼1G₀into the molecular

conductance region. To estimate the snap-back distance, we

extrapolated the tail of the tunneling current measured after

breaking the gold−gold contact to G₀. We obtained a distance

correction Δz_corr= 0.5 nm (for further details see SI and refs

24, 83, and 107). Figure 9A,B illustrates the snap-back-

corrected characteristic lengths for the high and for the low-

conductance states, z^H* = Δz^H* + Δz_corrand z^L* = Δz^L*+

Δz_corr, as a function of L_m. z^H* is smaller than L_mfor all

molecules investigated. On the other hand, the snap-back-

corrected electrode separation in a completely stretched

junction, z^L* = Δz^L* + Δz_corr, is not always smaller than L_m

(cf. molecules 1, 4, and 5, Figure 9B and Table S3).

We used z^H* to estimate the tilt angle θ of the molecules in

the high-conductance state (vs the normal to electrode surface)

from the simple geometric relation θ = arccos(z^H*/L_m). The

results are presented in Figure 9C. They demonstrate that θ

increases with L_m. The tilt angle is zero for 1, indicating that the

molecule is in an upright position in the high-conductance

state, while longer molecules are inclined. This observation also

suggests that the longer OPE wires most probably slide along

the gold contacts upon formation of a single molecular

junction, and breaking occurs, even under conditions of a

maximum possible stretching, not in a fully upright position but

at an angle to the surface normal (see Table S3).

Finally we would like to comment on the origin of the low-

conductance features. They have rather peculiar properties: (i)

They were observed as shoulders and low-density clouds in the

conductance histograms only for 6 compounds out of 11. (ii)

They are more pronounced in decane than in TMB. (iii) The

conductances G_Lare independent of molecular length for the

short molecules, and for long molecules they decrease with a

distance decay parameter being higher than that for any type of

molecular bridge studied so far (including aliphatic wires). (iv)

For some molecules the electrode separation in the low-

conductance state is longer than the molecular length L_m. We

can account for these trends only by attributing the low-

conductance features to the conductance through π−π-stacked

molecules.^45,53Stacking is particular promoted by the higher

density of π-electrons, as in compounds 4−7, which indeed

show low-conductance features, as well as by nonaromatic

solvents. However, the low-conductance feature for the π-

electron-deficient molecule 1 may as well be accompanied by

the pulling out of a gold atom from the tip or the surface.

Ab Initio Transport Calculations for the Three Series

of Molecules. To provide further insight into the

experimentally observed trends and the evolution of the

molecular junctions formed between the various OPE

derivatives coupled to gold electrodes, we performed

theoretical calculations using a combination of DFT and a

nonequilibrium Green's functions formalism (for further details

see SI).

The STM-BJ configuration was represented by an extended

molecule with one surface atom bound to a three-fold hollow

(H) site of the Au(111) surface (Au−S distance ∼0.21 nm)

and the other one to a gold adatom, which was contacted to the

surface in a H site (Au−S distance ∼0.24 nm) following the

(111) direction of growth (see Figure S20). This geometry was

found to be energetically the most stable on the basis of LDA

calculations.^52,53We also considered adatom−adatom (A-A),

hollow−hollow (H-H), and top-top (T-T, not shown)

geometries as well as the role of the tilt angle θ of the

molecular axis to the surface normal and the torsion angle φ

upon rotation of the central phenyl ring. Figure 10 illustrates

selected configurations (for further details see SI). We

computed the energy-dependent transmission T(E) using the

ab initio code SMEAGOL^75,76for the description of the

electron transport in various “extended molecule” config-

urations. T(E) represents the probability for electrons injected

with an energy E from one electrode to be transmitted through

Figure 9. Most probable characteristic length data z* as extracted from

the analysis of STM-BJ (black circles) and MCBJ (blue circles) data of

1 to 11 in the conductance ranges (A) 0.7G₀< G < 0.1 G_H(z^H*, A)

and (B) 0.7G₀< G < 10⁻⁶G₀(STM-BJ experiments) or 0.7G₀< G <

10⁻⁸G₀(MCBJ experiments) (z^L*, B). (C) Tilt angle θ of molecules

1−11 in the high-conductance state prior to contact breaking as a

function of the molecular length.

5270

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

Figure 10. Four representative configurations of the extended

molecule used in the calculations: side views of (A) hollow−hollow

(H-H) and (B) adatom−adatom (A-A) and (C) top view of H-H with

a tilt angle of 80° with respect to the normal to the surface. (D) The

same configuration as in (C), but with a torsion angle of 38° of the

central ring.

Figure 11. SAINT-corrected transmission curves calculated with

SMEAGOL for 1−9 and 11. The hexoxy-side groups in 11 were not

included into the calculations.

hole transport behavior in all wires. The relative position of

(E_HOMO− E_F) upon the SAINT correction is in good

agreement with the estimated shift, based on the analysis of

the experimentally obtained I−V_biascharacteristics in the

molecular plateau region (Figure 6).

The transmission coefficients T(E) of 1−3 and 11 (and 10,

which is not shown) reveal that as the length increases, the

HOMO−LUMO gap decreases (Figure 11A), since for longer

molecules the electrons can more easily delocalize along the

molecule backbone and the repulsion between levels decreases.

At the same time the transmission T(E_F) at the Fermi energy

(E = E_F) near the middle of the gap decreases exponentially.

This second effect dominates the decrease in the HOMO−

LUMO gap and produces an exponential reduction of the

transmission near the HOMO resonance, in agreement with

the experimental results (cf. Figure 8A).

Comparing molecules 2, 4, and 5 (Figure 11A,B) reveals a

decrease of the HOMO−LUMO gap due to an increase of the

electron delocalization in the central aromatic moiety, which

decreases the separation between levels. In this case T(E_F) does

not decrease exponentially, since all molecules have roughly the

same length (Tables S1 and S2). The net effect is therefore an

increase of the T(E_F) due to the presence of the nearby

HOMO resonance.

In the series of molecules 2, 6, and 7(Figure 11A,C) both the

reduction of the HOMO−LUMO gap and the increase of the

molecular length are important. The LUMO shifts downward

and the conductance in the gap decreases exponentially. As a

consequence T(E_F) decreases slightly from 2 to 6 to 7.

We have plotted in Figure 8C the zero-bias conductance, as

calculated from the SAINT-corrected transmission at the Fermi

level, as a function of the molecular length. The graph

the molecular junction. The conductance G is defined as the

transmission T(E_F) evaluated at the Fermi level E_Fin units of

the conductance quantum G₀. Molecular states of the OPEs

appear as resonance peaks in the transmission spectrum.

Qualitative trends in families of molecules are known to be

correctly reproduced by transport calculations based on bare

DFT-derived mean-field Hamiltonians. However, quantitative

agreement between experiment and theory is hampered by the

bandgap problem (DFT underestimates the HOMO−LUMO

gap) and other problems related to the LDA functional, such as

self-interaction errors.^78,108

In the bare DFT calculations of the family of the linearly

conjugated OPE derivatives 1−11, the Fermi level is found to

be typically pinned to the HOMO orbital, so that the

conductance at E_Fis overestimated. To resolve this problem,

a semiempirical self-energy and screening energy correction has

been introduced by a scissor-like operator (spectral adjustment

in nanoscale transport, SAINT)⁷⁹⁻⁸¹prior to the transport

calculations.

The SAINT operator moves the occupied molecular levels

(HOMO) to lower and the unoccupied levels (LUMO) to

higher energies, respectively. The derived mean-field Hamil-

tonian was constructed in such a way that (i) the HOMO

agrees with UPS experiments of OPE derivatives attached to

gold films and (ii) that the HOMO−LUMO gap matches the

optical HOMO−LUMO gaps, which were estimated from the

onsets of the UV/vis absorption spectra (Table S1). Figure 11

shows plots of the SAINT-corrected transmission curves of 1 to

11 for the adatom−hollow (A-H) upright geometry. The

resonance closest to the Fermi level (about 1 eV lower than E_F)

corresponds to a HOMO feature indicating a HOMO-mediated

5271

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

demonstrates that the experimental trend is correctly

reproduced by the theoretical calculations. More importantly,

the quantitative zero-bias conductance values are also close to

the experimental values, which is not typical of DFT-based

calculations without corrections. In addition, the decay constant

obtained for series 1−3, 6, 7 (and 11) is 2.76 nm ⁻¹, in good

agreement with the experimental value 3.4 nm⁻¹. We note that

the calculation of 11 was carried out without incorporating the

hexyloxy-substituents, which are considered to have a negligible

effect on the transport close to the Fermi level, as it was shown

recently for methoxy and tert-butyl substitutions in OPEs.⁵³

Additional calculations following the same strategy as outlined

above reveal that the H-H configuration leads to similar

conclusions, however with higher junction conductances. The

lowest overall conductances were found for the A-A

coordination geometry (shown in SI). The best agreement

between experimental and computed data in the high-

conductance range G_Hwas obtained for the A-H configuration,

which may be also considered as a reasonable representation of

the STM-BJ geometry. However, contributions of other Au−S

coupling configurations including those of “phase-on” edge

sites during the evolution of the molecular junction between

the gold leads and the dithiolated OPE-type molecules cannot

be excluded,^{53,63,71,109,110}as the 1D and 2D conductance

histograms show a non-negligible width in the distribution

around the most probable conductance values (Figure 2 and

SI). Finally, we comment that the distinct deviations of 8 and 9

from the length-dependence trend of the OPE wires

investigated could be attributed to the relatively high position

of the LUMO level resulting from the broken π-conjugation in

9 and the appearance of a destructive quantum interference in 8

as demonstrated in recent studies^{105,106,111,112}and by the

transmission curves plotted in Figure 11D. Thygesen et al.

demonstrated that the quantum interference effect is rather

insensitive to details of the contact geometry. It is a property of

the symmetry of the molecular orbitals and, as a consequence,

its overall appearance is not severely influenced by the exact

position of the HOMO and LUMO levels, respectively. T(E_F)

is predicted to show without and with the application of a

scissors operator, such as in the SAINT correction used in our

approach, a distinct minimum leading to a conductance at E_F

which is much smaller than that for the “simple” anthracene rod

7. This trend is indeed observed experimentally (see Figure

8A).

is sensitive to θ, and increases significantly for large angles. (A

similar trend was proposed by Haiss et al.⁵²) These results are

illustrated in Figure 12. The behavior can be explained by

Figure 12. SAINT-corrected transmission curves obtained for

molecule 2 in (A) a straight (0°) and tilted (80°) in A-H configuration

and (B) an A-A configuration.

considering that the main orbitals on the sulfur that contribute

to the HOMO resonance and couple to the π-orbitals in the

carbon backbone (x and y p-orbitals) have by symmetry zero

overlap with the s-orbital of the gold adatom (tip) or the

hollow site atoms (substrate), and a very small overlap with the

more localized d-orbitals when the molecule is straight.^113−115

Therefore, they give a very small coupling. When the molecule

is rotated, however, such orbitals become nonorthogonal with

the gold orbitals and increase the overlap. This effect enhances

the coupling, or in other words, increases the density of states

at E_F, as can be seen in the dramatic increase of the width of the

HOMO resonance when the angle increases (Figure 12).

Despite the success, there are however unresolved questions

related to the evolution of the conductance with the stretching

distance between the STM tip and the surface. From the 2D

conductance (log(G/G₀)) versus displacement (Δz) histo-

grams (Figure 2 and SI), it is clear that as Δz increases, the

high-conductance G_Hportion of the histogram decreases

toward lower values. Furthermore, the relative distance at

which this region terminates is smaller than L_m, the length of

the molecule. Figure 9A,B illustrates that this trend is valid for

the most probable characteristic length in the high- (z^H*) as

well as in the low-conductance range (z^L*), with corrections for

the snap-back distance Δz_corr= 0.5 nm. Figure 9C shows that

the tilt angles of the molecular in the junction in the high-

conductance state are increasing with increasing molecular

length L_m. These experimental results imply that the OPE wires

are tilted with respect to the surface, and that the angle θ

relative to the surface normal decreases as the tip retracts.

Therefore we also carried out calculations with a rather large

angle for the prototype wire 2. We found that the transmission

Figure 13. Summary of the entire stretching process, with the two

limiting configurations: (A) tilted molecule connected in an A-H

configuration at the beginning and (B) a straight molecule connected

in an A-A configuration at the end of the stretching process.

From the experimental observation of tilt angles and

characteristic lengths (z^H*, z^L*), we can draw the following

picture of the evolution of the junction as the tip is retracted

from the surface (Figure 13). For small relative distances Δz,

the OPEs typically have a large tilt angle θ with respect to the

surface normal, which is basically determined by the snap-back

distance of Δz_corr≈ 0.5 nm. As Δz increases, the most probable

value of θ and the most probable conductance decrease. At

5272

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

intermediate distances, a range of angles are encountered,

leading to a spread of conductance values of the junctions. We

also note that fluctuations of the torsion angle φ between the

phenyl rings could add to the spread in the main conductance

region G_H(see SI for transmission curves). Upon further

displacement, typically (z^H* − z^L*) = 0.2−0.3 nm, several OPE

wires (1, 2, 4−7) show clear evidence for the evolution of a

low-conductance region G_Lon the right of the main

conductance region in the 2D conductance histograms (see

SI). For molecule 1, θ is estimated to be zero, which indicates

that the molecule is completely elongated in the junction. The

probability of forming π−π-stacks in the molecular junction is

rather low for 1 due to the low π-electron density in the

molecule. Further, based on the snap-back-corrected character-

istic length z^L*, it is clear that the low-conductance feature

observed in the histogram can be attributed to the pulling out

of a gold atom from the electrode. This case would correspond

to the molecule also coupled to an adatom on the surface,

whose presence would give a longer distance relative to the

surface normal and a lower conductance when the molecule

assumes small angles. Indeed the difference in conductance

between the A-H and the A-A configurations is close to the

experimentally observed trends. Note, however, that the results

are rather similar to those of the A-H configuration when the

angles are large, which seems to indicate that such angles are

not favored in this configuration and at those distances. The

two limiting configurations of the pulling process are displayed

in Figure 13, namely the A-H with a tilt angle and the straight

A-A, which give rise to the highest and the lowest values of

conductance, respectively. Molecules 4−7 contain larger π-

electron cores, which can facilitate π−π-stacking. For these

molecules there is a high probability to form supramolecular

π−π-stacking junctions, which may represent the low-

conductance state. Comparative experiments of mono- and

dithiolated OPE rods suggest that supramolecular π−π-stacks

could form indeed.^44,52Both modes, the A-A configuration as

well as the intermolecular π−π-stacking, may represent the final

stage of molecular junction breakage, or may occur even

coupled.

Experimentally we found high (G_H) and low (G_L)-

conductance values for the OPEs 1, 2, 4−7. Theoretical

calculations suggest that the appearance of the low-conductance

feature originates from a change in surface coordination from a

adatom−hollow to an adatom−adatom geometry with the

molecular rod assuming a straight orientation to the surface

normal for smaller molecules (1, 2). The involvement of

stacking interactions in this mechanism cannot be excluded for

the molecules with larger π-electron core (4−7), and are one of

the targets of our current work.

In summary, we have developed, employing a combined

experimental and theoretical approach, a rather uniform

description of the formation and breaking of OPE-type

molecular rods attached to gold electrodes. We demonstrated

the sensitivity of our complementary STM-BJ and MCBJ

approach, and developed clear correlations between molecular

structure features and conductance, which shall serve as a

guidance for future research on designing and developing

nanoscale-based molecular circuits.

ASSOCIATED CONTENT

* Supporting Information

■

S

Details on the molecular synthesis, the experimental setup, and

data analysis as well as the DFT-type transport calculations.

This material is available free of charge via the Internet at

http://pubs.acs.org.

AUTHOR INFORMATION

Corresponding Author

j.c.hummelen@rug.nl; c.lambert@lancaster.ac.uk; thomas.

wandlowski@dcb.unibe.ch

■

Author Contributions

^⊗These authors contributed equally to this work.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS

■

This work was supported by the Swiss National Science

Foundation (200021-124643; NFP 62), EC FP7 ITN

“FUNMOLS” Project No. 212942, DFG under SPP 1243,

and the University of Bern. The authors thank I. Pobelov and

A. Mishchenko for discussions and critical reading of the

CONCLUSIONS

■

We studied experimentally (STM-BJ and MCBJ techniques)

and theoretically the conductance of a family of OPEs with

thiol linker groups. 1D and 2D conductance histograms showed

well-defined conductance features of the junctions formed. We

observed experimentally and theoretically that the junction

conductance decreases with molecular length and increases

with a decrease in (E_F− E_HOMO), the difference between E_F

and the nearest molecular orbital, which is E_HOMOin the

present case. The exponential decay of the conductance with

molecular length and the value of the decay constant β indicate

that the charge transport through these molecules follows a

nonresonant tunnelling mechanism. The SAINT-corrected

DFT calculations with a adatom−hollow contact geometry

for the high-conductance state G_Hshow an excellent agreement

with the experiment, both qualitatively and quantitatively. We

also conclude that the most probable breaking event of a

stretched molecular junction occurs in a tilted geometry.

Distinct deviations from the experimental correlation of

conductance versus distance L_mfor the cross-conjugated wire

8 and OPE 9 are attributed to a destructive quantum

interference and to the broken π-conjugation, respectively.

manuscript, and Prof. S. Hoger for providing molecules 10 and

̈

11. H.V. acknowledges NanoNed, funded by the Dutch

Ministry of Economic Affairs (project GMM. 6973). V.M.G.-

́

S. thanks the Spanish Ministerio de Ciencia e Innovacion for a

Ramon y Cajal Fellowship (RYC-2010-06053) and the project

FIS2009-07081.

REFERENCES

■

(1) Aviram, A.; Ratner, M. Chem. Phys. Lett. 1974, 29, 277−283.

(2) Chen, F.; Hihath, J.; Huang, Z. F.; Li, X. L.; Tao, N. J. Annu. Rev.

Phys. Chem. 2007, 58, 535−564.

(3) McCreery, R. L.; Bergren, A. J. Adv. Mater. (Weinheim, Ger.)

2009, 21, 4303−4322.

(4) Moth-Poulsen, K.; Bjornholm, T. Nat. Nanotechnol. 2009, 4,

551−556.

(5) Cuevas, J. C.; Scheer, E. Molecular Electronics: An Introduction to

Theory and Experiment; World Scientific: Singapore, 2010.

(6) van der Molen, S. J.; Liljeroth, P. J. Phys.: Cond. Matt. 2010, 22,

133001−133030.

(7) Li, C.; Mishchenko, A.; Pobelov, I.; Wandlowski, T. Chimia 2010,

64, 383−390.

5273

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

(8) Nichols, R. J.; Haiss, W.; Higgins, S. J.; Leary, E.; Martin, S.;

Bethell, D. Phys. Chem. Chem. Phys. 2010, 12, 2801−2815.

(9) Metzger, R. M. Unimolecular and Supramolecular Electronics II;

Topics in Current Chemistry 313; Springer Verlag: Berlin, 2012.

(10) Tao, N. J. Nat. Nanotechnol. 2006, 1, 173−181.

(11) Song, H.; Reed, M. A.; Lee, T. Adv. Mater. (Weinheim, Ger.)

2011, 23, 1583−1608.

(37) 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.

(38) Cai, L. T.; Skulason, H.; Kushmerick, J. G.; Pollack, S. K.; Naciri,

J.; Shashidhar, R.; Allara, D. L.; Mallouk, T. E.; Mayer, T. S. J. Phys.

Chem. B 2004, 108, 2827−2832.

(39) Newton, M. D.; Smalley, J. F. Phys. Chem. Chem. Phys. 2007, 9,

555−572.

(40) Creager, S.; Yu, C. J.; Bamdad, C.; O’Connor, S.; MacLean, T.;

Lam, E.; Chong, Y.; Olsen, G. T.; Luo, J. Y.; Gozin, M.; Kayyem, J. F. J.

Am. Chem. Soc. 1999, 121, 1059−1064.

(41) Donhauser, Z. J.; Mantooth, B. A.; Kelly, K. F.; Bumm, L. A.;

Monnell, J. D.; Stapleton, J. J.; Price, D. W.; Rawlett, A. M.; Allara, D.

L.; Tour, J. M.; Weiss, P. S. Science 2001, 292, 2303−2307.

(42) Reichert, J.; Ochs, R.; Beckmann, D.; Weber, H. B.; Mayor, M.;

von Lohneysen, H. Phys. Rev. Lett. 2002, 88, 176804.

(43) Mayor, M.; Weber, H. B.; Reichert, J.; Elbing, M.; von Hanisch,

C.; Beckmann, D.; Fischer, M. Angew. Chem., Int. Ed. 2003, 42, 5834−

5838.

(12) Weibel, N.; Grunder, S.; Mayor, M. Org. Biomol. Chem. 2007, 5,

2343−2353.

(13) Tour, J. M. Chem. Rev. 1996, 96, 537−553.

́

(14) Díez-Perez, I.; Hihath, J.; Lee, Y.; Yu, L. P.; Adamska, L.;

Kozhushner, M. A.; Oleynik, I. I.; Tao, N. J. Nature Chem. 2009, 1,

635−641.

(15) Metzger, R. M.; Chen, B.; Hopfner, U.; Lakshmikantham, M. V.;

Vuillaume, D.; Kawai, T.; Wu, X. L.; Tachibana, H.; Hughes, T. V.;

Sakurai, H.; Baldwin, J. W.; Hosch, C.; Cava, M. P.; Brehmer, L.;

Ashwell, G. J. J. Am. Chem. Soc. 1997, 119, 10455−10466.

(16) Elbing, M.; Ochs, R.; Koentopp, M.; Fischer, M.; von Hanisch,

C.; Weigend, F.; Evers, F.; Weber, H. B.; Mayor, M. Proc. Natl. Acad.

Sci. U.S.A. 2005, 102, 8815−8820.

(44) Huber, R.; Gonzal

Horhoiu, V.; Mayor, M.; Bryce, M. R.; Wang, C. S.; Jitchati, R.;

́

ez, M. T.; Wu, S.; Langer, M.; Grunder, S.;

(17) Song, H.; Kim, Y.; Jang, Y. H.; Jeong, H.; Reed, M. A.; Lee, T.

Nature 2009, 462, 1039−1043.

Schonenberger, C.; Calame, M. J. Am. Chem. Soc. 2008, 130, 1080−

̈

1084.

(18) Xu, B. Q.; Xiao, X. Y.; Yang, X. M.; Zang, L.; Tao, N. J. J. Am.

Chem. Soc. 2005, 127, 2386−2387.

(45) Wu, S. M.; Gonzal

́

ez, M. T.; Huber, R.; Grunder, S.; Mayor, M.;

Schonenberger, C.; Calame, M. Nat. Nanotechnol. 2008, 3, 569−574.

̈

(19) Lortscher, E.; Ciszek, J. W.; Tour, J.; Riel, H. Small 2006, 2,

973−977.

(46) Xiao, X. Y.; Nagahara, L. A.; Rawlett, A. M.; Tao, N. J. J. Am.

Chem. Soc. 2005, 127, 9235−9240.

(20) Green, J. E.; Choi, J. W.; Boukai, A.; Bunimovich, Y.; Johnston-

Halperin, E.; DeIonno, E.; Luo, Y.; Sheriff, B. A.; Xu, K.; Shin, Y. S.;

Tseng, H.-R.; Stoddart, J. F.; Heath, J. R. Nature 2007, 445, 414−417.

(21) Lee, J.; Chang, H.; Kim, S.; Bang, G. S.; Lee, H. Angew. Chem.,

Int. Ed. 2009, 48, 8501−8504.

(47) Liu, K.; Wang, X.; Wang, F. ACS Nano 2008, 2, 2315−2323.

(48) Hybertsen, M. S.; Venkataraman, L.; Klare, J. E.; Cwhalley, A.;

Steigerwald, M. L.; Nuckolls, C. J. Phys.: Cond. Matt. 2008, 20, 374115.

(49) Weibel, N.; Mishchenko, A.; Wandlowski, T.; Neuburger, M.;

Leroux, Y.; Mayor, M. Eur. J. Org. Chem. 2009, 6140−6150.

(50) Xing, Y.; Park, T.-H.; Venkatramani, R.; Keinan, S.; Beratan, D.

N.; Therien, M. J.; Borguet, E. J. Am. Chem. Soc. 2010, 132, 7946−

7956.

(22) Blum, A. S.; Kushmerick, J. G.; Long, D. P.; Patterson, C. H.;

Yang, J. C.; Henderson, J. C.; Yao, Y. X.; Tour, J. M.; Shashidhar, R.;

Ratna, B. R. Nat. Mater. 2005, 4, 167−172.

(23) Choi, B. Y.; Kahng, S. J.; Kim, S.; Kim, H.; Kim, H. W.; Song, Y.

J.; Ihm, J.; Kuk, Y. Phys. Rev. Lett. 2006, 96, 156106.

(24) 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, 230−234.

(51) Lu, Q.; Liu, K.; Zhang, H. M.; Du, Z. B.; Wang, X. H.; Wang, F.

S. ACS Nano 2009, 3, 3861−3868.

(52) Haiss, W.; Wang, C. S.; Grace, I.; Batsanov, A. S.; Schiffrin, D. J.;

Higgins, S. J.; Bryce, M. R.; Lambert, C. J.; Nichols, R. J. Nat. Mater.

2006, 5, 995−1002.

(25) van der Molen, S. J.; Liao, J.; Kudernac, T.; Agustsson, J. S.;

Bernard, L.; Calame, M.; van Wees, B. J.; Feringa, B. L.;

(53) Martin, S.; Grace, I.; Bryce, M. R.; Wang, C.; Jitchati, R.;

Batsanov, A. S.; Higgins, S. J.; Lambert, C. J.; Nichols, R. J. J. Am.

Chem. Soc. 2010, 132, 9157−9164.

Schonenberger, C. Nano Lett. 2009, 9, 76−80.

̈

(26) He, J.; Chen, F.; Liddell, P. A.; Andreasson, J.; Straight, S. D.;

Gust, D.; Moore, T. A.; Moore, A. L.; Li, J.; Sankey, O. F.; Lindsay, S.

M. Nanotechnology 2005, 16, 695−702.

́

(54) Gonzalez, M. T.; Leary, E.; Garcia, R.; Verma, P.; Angeles

Herranz, M.; Rubio-Bollinger, G.; Martin, N.; Agrait, N. J. Phys. Chem.

C 2011, 115, 17973−17978.

(27) Jenny, N. M.; Mayor, M.; Eaton, T. R. Eur. J. Org. Chem. 2011,

4965−4983.

(55) Akkerman, H. B.; de Boer, B. J. Phys.: Cond. Matt. 2008, 20,

013001.

(28) Baumgardt, I.; Butenschoen, H. Eur. J. Org. Chem. 2010, 1076−

1087.

(56) Valkenier, H. Molecular Conductance: Synthesis, Assembly,and

Electrical Characterization of π-Conjugated Wires and Switches. Ph.D.

Thesis, Rijksuniversiteit, Groningen, July, 2011.

(29) Kushmerick, J. G.; Holt, D. B.; Pollack, S. K.; Ratner, M. A.;

Yang, J. C.; Schull, T. L.; Naciri, J.; Moore, M. H.; Shashidhar, R. J.

Am. Chem. Soc. 2002, 124, 10654−10655.

(57) Liu, H. M.; Wang, N.; Zhao, J. W.; Guo, Y.; Yin, X.; Boey, F. Y.

C.; Zhang, H. ChemPhysChem 2008, 9, 1416−1424.

(58) Venkataraman, L.; Klare, J. E.; Nuckolls, C.; Hybertsen, M. S.;

Steigerwald, M. L. Nature 2006, 442, 904−907.

(30) Chen, J.; Reed, M. A.; Rawlett, A. M.; Tour, J. M. Science 1999,

286, 1550−1552.

(31) Liao, J.; Bernard, L.; Langer, M.; Schonenberger, C.; Calame, M.

̈

(59) Wang, C. S.; Batsanov, A. S.; Bryce, M. R.; Martin, S.; Nichols,

Adv. Mater. (Weinheim, Ger.) 2006, 18, 2444−2447.

(32) Rawlett, A. M.; Hopson, T. J.; Nagahara, L. A.; Tsui, R. K.;

Ramachandran, G. K.; Lindsay, S. M. Appl. Phys. Lett. 2002, 81, 3043−

3045.

R. J.; Higgins, S. J.; García-Suar

Soc. 2009, 131, 15647−15654.

(60) Sedghi, G.; García-Suar

́

ez, V. M.; Lambert, C. J. J. Am. Chem.

́

ez, V. M.; Esdaile, L. J.; Anderson, H. L.;

Lambert, C. J.; Martin, S.; Bethell, D.; Higgins, S. J.; Elliott, M.;

Bennett, N.; Macdonald, J. E.; Nichols, R. J. Nat. Nanotechnol. 2011, 6,

517−523.

(33) Liu, K.; Li, G.; Wang, X.; Wang, F. J. Phys. Chem. C 2008, 112,

4342−4349.

(34) Beebe, J. M.; Kim, B.; Frisbie, C. D.; Kushmerick, J. G. ACS

Nano 2008, 2, 827−832.

(61) Yamada, R.; Kumazawa, H.; Noutoshi, T.; Tanaka, S.; Tada, H.

Nano Lett. 2008, 8, 1237−1240.

(35) Fan, F. R. F.; Yang, J. P.; Cai, L. T.; Price, D. W.; Dirk, S. M.;

Kosynkin, D. V.; Yao, Y. X.; Rawlett, A. M.; Tour, J. M.; Bard, A. J. J.

Am. Chem. Soc. 2002, 124, 5550−5560.

(62) Quinn, J. R.; Foss, F. W. Jr.; Venkataraman, L.; Breslow, R. J.

Am. Chem. Soc. 2007, 129, 12376−12377.

(63) Mishchenko, A.; Vonlanthen, D.; Meded, V.; Burkle, M.; Li, C.;

Pobelov, I. V.; Bagrets, A.; Viljas, J. K.; Pauly, F.; Evers, F.; Mayor, M.;

Wandlowski, T. Nano Lett. 2010, 10, 156−163.

(36) Fracasso, D.; Valkenier, H.; Hummelen, J. C.; Solomon, G. C.;

Chiechi, R. C. J. Am. Chem. Soc. 2011, 9556−9563.

5274

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Journal of the American Chemical Society

Article

(64) Mishchenko, A.; Zotti, L. A.; Vonlanthen, D.; Burkle, M.; Pauly,

F.; Cuevas, J. C.; Mayor, M.; Wandlowski, T. J. Am. Chem. Soc. 2011,

133, 184−187.

(65) Stuhr-Hansen, N. Synth. Commun. 2003, 33, 641−646.

(66) Stuhr-Hansen, N.; Sørensen, J. K.; Moth-Poulsen, K.;

Christensen, J. B.; Bjørnholm, T.; Nielsen, M. B. Tetrahedron 2005,

61, 12288−12295.

(67) Blaszczyk, A.; Elbing, M.; Mayor, M. Org. Biomol. Chem. 2004, 2,

2722−2724.

(68) van Dijk, E. H.; Myles, D. J. T.; van der Veen, M. H.;

Hummelen, J. C. Org. Lett. 2006, 8, 2333−2336.

(69) Zhou, C. Z.; Liu, T. X.; Xu, J. M.; Chen, Z. K. Macromolecules

2003, 36, 1457−1464.

(70) Xu, B. Q.; Tao, N. J. J. Science 2003, 301, 1221−1223.

(71) Li, C.; Pobelov, I.; Wandlowski, T.; Bagrets, A.; Arnold, A.;

Evers, F. J. Am. Chem. Soc. 2008, 130, 318−326.

(72) Muller, C. J.; Vanruitenbeek, J. M.; Dejongh, L. J. Phys. Rev. Lett.

1992, 69, 140−143.

(73) Hong, W.; Valkenier, H.; Meszaros, G.; Manrique, D. Z.;

Mishchenko, A.; Putz, A.; Garcia, P. M.; Lambert, C. J.; Hummelen, J.

C.; Wandlowski, T. Beilstein J. Nanotechnol. 2011, 2, 699−713.

(74) Meszaros, G.; Li, C.; Pobelov, I.; Wandlowski, T. Nano-

technology 2007, 18, 424004.

(97) Chen, F.; Li, X.; Hihath, J.; Huang, Z.; Tao, N. J. Am. Chem. Soc.

2006, 128, 15874−15881.

(98) Chen, F.; Li, X. L.; Hihath, J.; Huang, Z. F.; Tao, N. J. J. Am.

Chem. Soc. 2006, 128, 15874−15881.

(99) Tomfohr, J.; Sankey, O. F. J. Chem. Phys. 2004, 120, 1542−

1554.

(100) Wold, D. J.; Haag, R.; Rampi, M. A.; Frisbie, C. D. J. Phys.

Chem. B 2002, 106, 2813−2816.

(101) Ishida, T.; Mizutani, W.; Aya, Y.; Ogiso, H.; Sasaki, S.;

Tokumoto, H. J. Phys. Chem. B 2002, 106, 5886−5892.

(102) Beebe, J. M.; Kim, B.; Gadzuk, J. W.; Frisbie, C. D.;

Kushmerick, J. G. Phys. Rev. Lett. 2006, 97, 026801.

(103) Veenstra, S. C.; Stalmach, U.; Krasnikov, V. V.; Hadziioannou,

G.; Jonkman, H. T.; Heeres, A.; Sawatzky, G. A. Appl. Phys. Lett. 2000,

76, 2253−2255.

(104) Visontai, D.; Grace, I. M.; Lambert, C. J. Phys. Rev. B 2010, 81,

035409.

(105) Markussen, T.; Schioumltz, J.; Thygesen, K. S. J. Chem. Phys.

2010, 132, 224104.

(106) Markussen, T.; Stadler, R.; Thygesen, K. S. Nano Lett. 2010,

10, 4260−4265.

(107) Yanson, A. I.; Bollinger, G. R.; van den Brom, H. E.; Agrait, N.;

van Ruitenbeek, J. M. Nature 1998, 395, 783−785.

(108) Toher, C.; Filippetti, A.; Sanvito, S.; Burke, K. Phys. Rev. Lett.

2005, 95, 146402.

(75) Rocha, A. R.; García-Suar

Ferrer, J.; Sanvito, S. Phys. Rev. B 2006, 73, 085414.

(76) Rocha, A. R.; García-Suarez, V. M.; Bailey, S. W.; Lambert, C. J.;

́

ez, V. M.; Bailey, S.; Lambert, C.;

(109) Ulrich, J.; Esrail, D.; Pontius, W.; Venkataraman, L.; Millar, D.;

Doerrer, L. H. J. Phys. Chem. B 2006, 110, 2462−2466.

(110) Paulsson, M.; Krag, C.; Frederiksen, T.; Brandbyge, M. Nano

Lett. 2009, 9, 117−121.

(111) Wang, C.; Bryce, M. R.; Gigon, J.; Ashwell, G. J.; Grace, I.;

Lambert, C. J. J. Org. Chem. 2008, 73, 4810−4818.

(112) Papadopoulos, T. A.; Grace, I. M.; Lambert, C. J. Phys. Rev. B

2006, 74, 193306.

́

Ferrer, J.; Sanvito, S. Nat. Mater. 2005, 4, 335−339.

(77) Soler, J. M.; Artacho, E.; Gale, J. D.; Garcia, A.; Junquera, J.;

Ordejon, P.; Sanchez-Portal, D. J. Phys.: Cond. Matt. 2002, 14, 2745−

2779.

(78) Perdew, J. P.; Zunger, A. Phys. Rev. B 1981, 23, 5048−5079.

(79) Mowbray, D. J.; Jones, G.; Thygesen, K. S. J. Chem. Phys. 2008,

128, 111103.

(113) Bratkovsky, A. M.; Kornilovitch, P. E. Phys. Rev. B 2003, 67,

115307.

(80) García-Suar

053026.

́

ez, V. M.; Lambert, C. J. New J. Phys. 2011, 13,

(114) Dunbar, T. D.; Cygan, M. T.; Bumm, L. A.; McCarty, G. S.;

Burgin, T. P.; Reinerth, W. A.; Jones, L.; Jackiw, J. J.; Tour, J. M.;

Weiss, P. S.; Allara, D. L. J. Phys. Chem. B 2000, 104, 4880−4893.

(115) Liu, Z.; Wang, X.; Dai, K.; Jin, S.; Zeng, Z.-C.; Zhuang, M.-D.;

Yang, Z.-L.; Wu, D.-Y.; Ren, B.; Tian, Z.-Q. J. Raman Spectrosc. 2009,

40, 1400−1406.

(81) Quek, S. Y.; Choi, H. J.; Louie, S. G.; Neaton, J. B. Nano Lett.

2009, 9, 3949−3953.

(82) Gonzal

́

ez, M. T.; Wu, S.; Huber, R.; van der Molen, S. J.;

Schonenberger, C.; Calame, M. Nano Lett. 2006, 6, 2238−2242.

̈

(83) Agrait, N.; Yeyati, A. L.; van Ruitenbeek, J. M. Phys. Rep.: Rev.

Sec. Phys. Lett. 2003, 377, 81−279.

(84) Martin, C. A.; Ding, D.; Sørensen, J. K.; Bjørnholm, T.; van

Ruitenbeek, J. M.; van der Zant, H. S. J. J. Am. Chem. Soc. 2008, 130,

13198−13199.

(85) Weibel, N.; Blaszczyk, A.; von Haenisch, C.; Mayor, M.;

Pobelov, I.; Wandlowski, T.; Chen, F.; Tao, N. Eur. J. Org. Chem.

2008, 136−149.

(86) Xiao, X. Y.; Xu, B. Q.; Tao, N. J. Nano Lett. 2004, 4, 267−271.

(87) Martin, S.; Grace, I.; Bryce, M. R.; Wang, C. S.; Jitchati, R.;

Batsanov, A. S.; Higgins, S. J.; Lambert, C. J.; Nichols, R. J. J. Am.

Chem. Soc. 2010, 132, 9157−9164.

(88) Huang, Z.; Chen, F.; Bennett, P. A.; Tao, N. J. Am. Chem. Soc.

2007, 129, 13225−13231.

(89) Huang, Z.; Xu, B.; Chen, Y.; Di Ventra, M.; Tao, N. Nano Lett.

2006, 6, 1240−1244.

(90) Evans, E. Annu. Rev. Biophys. Biomol. Struct. 2001, 30, 105−128.

(91) Evans, E.; Ritchie, K. Biophys. J. 1997, 72, 1541−1555.

(92) Xu, B. Q.; Xiao, X. Y.; Tao, N. J. J. Am. Chem. Soc. 2003, 125,

16164−16165.

(93) Lafferentz, L.; Ample, F.; Yu, H.; Hecht, S.; Joachim, C.; Grill, L.

Science 2009, 323, 1193−1197.

(94) Lortscher, E.; Weber, H. B.; Riel, H. Phys. Rev. Lett. 2007, 98,

̈

176807.

(95) Widawsky, J. R.; Kamenetska, M.; Klare, J.; Nuckolls, C.;

Steigerwald, M. L.; Hybertsen, M. S.; Venkataraman, L. Nanotechnology

2009, 20, 434009.

(96) Malen, J. A.; Doak, P.; Baheti, K.; Tilley, T. D.; Segalman, R. A.;

Majumdar, A. Nano Lett. 2009, 9, 1164−1169.

5275

dx.doi.org/10.1021/ja211555x | J. Am. Chem. Soc. 2012, 134, 5262−5275

Article Doi

Correlations between molecular structure and single-junction conductance: A case study with oligo(phenylene-ethynylene)-type wires

DOI: 10.1021/ja211555x

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja211555x

Full text of DOI:10.1021/ja211555x

Products guided by the article

R&D Labs maybe for 1364344-54-6

Relevant to this article

Hot Product