Detecting Mechanochemical Atropisomerization within an STM Break Junction

Article abstract of DOI:10.1021/jacs.7b10542

We have employed the scanning tunneling microscope break-junction technique to investigate the single-molecule conductance of a family of 5,15-diaryl porphyrins bearing thioacetyl (SAc) or methylsulfide (SMe) binding groups at the ortho position of the phenyl rings (S2 compounds). These ortho substituents lead to two atropisomers, cis and trans, for each compound, which do not interconvert in solution under ambient conditions; even at high temperatures, isomerization takes several hours (half-life 15 h at 140 °C for SAc in C₂Cl₄D₂). All the S2 compounds exhibit two conductance groups, and comparison with a monothiolated (S1) compound shows the higher group arises from a direct Au-porphyrin interaction. The lower conductance group is associated with the S-to-S pathway. When the binding group is SMe, the difference in junction length distribution reflects the difference in S-S distance (0.3 nm) between the two isomers. In the case of SAc, there are no significant differences between the plateau length distributions of the two isomers, and both show maximal stretching distances well exceeding their calculated junction lengths. Contact deformation accounts for part of the extra length, but the results indicate that cis-to-trans conversion takes place in the junction for the cis isomer. The barrier to atropisomerization is lower than the strength of the thiolate Au-S and Au-Au bonds, but higher than that of the Au-SMe bond, which explains why the strain in the junction only induces isomerization in the SAc compound.

Full text of DOI:10.1021/jacs.7b10542

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Article
Detecting mechanochemical atropisomerization within an STM breakjunction
Edmund Leary, Cecile Roche, Hua-Wei Jiang, Iain Grace, M. Teresa González, Gabino
Rubio-Bollinger, Carlos Romero-Muñiz, Yaoyao Xiong, Qusiy Al-Galiby, Mohammed
Noori, Maria A. Lebedeva, Kyriakos Porfyrakis, Nícolas Agraït, Andrew Hodgson,
Simon J. Higgins, Colin J. Lambert, Harry L. Anderson, and Richard J. Nichols
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b10542 • Publication Date (Web): 20 Dec 2017
Downloaded from http://pubs.acs.org on December 20, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 11
Journal of the American Chemical Society
1
2
3
4
5
6
Detecting mechanochemical atropisomerization within an
STM breakjunction
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
#
,
†
‡
‡
§
||
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Edmund Leary,* Cécile Roche, HuaꢀWei Jiang, Iain Grace, M. Teresa González, Gabino Rubioꢀ
┴
┼
‡
§
§
Bollinger, Carlos RomeroꢀMuñiz, Yaoyao Xiong, Qusiy AlꢀGaliby, Mohammed Noori, Maria A.
ʃ
ʃ
||,┴
#,
†
†
Lebedeva, Kyriakos Porfyrakis, Nicolás Agrait, Andrew Hodgson, Simon J. Higgins, Colin J.
§
‡
#,
†
Lambert,* Harry L. Anderson,* and Richard J. Nichols*
†
Department of Chemistry, Donnan and Robert Robinson Laboratories, University of Liverpool, Liverpool L69 7ZD, U.K.
Surface Science Research Centre and Department of Chemistry, University of Liverpool, Oxford Street, Liverpool L69
#
3
BX, UK
‡
Department of Chemistry, Oxford University, Chemistry Research Laboratory, Oxford OX1 3TA, U.K.
Department of Materials, University of Oxford, OX1 3PH, UK
ʃ
§
Department of Physics, Lancaster University, Lancaster, U.K.
||
Instituto Madrileño de Estudios Advanzados (IMDEA), Calle Faraday 9, Campus Universitario de Cantoblanco, 28049 Maꢀ
drid, Spain.
┴
Departamento de Física de la Materia Condensada, and Instituto “Nicolás Cabrera”, Universidad Autónoma de Madrid, Eꢀ
2
8049, Madrid, Spain.
┼
Departamento de Física Teórica de la Materia Condensada, Universidad Autónoma de Madrid, Eꢀ28049 Madrid, Spain.
KEYWORDS Mechanochemistry, STM, molecular electronics, single molecule conductance, porphyrins.
ABSTRACT: We have employed the scanning tunneling microscope breakꢀjunction technique to investigate the singleꢀmolecule
conductance of a family of 5,15ꢀdiaryl porphyrins bearing thioacetyl (SAc) or methylsulfide (SMe) binding groups at the orthoꢀ
position of the phenyl rings (S2 compounds). These ortho substituents lead to two atropisomers, cis and trans, for each compound,
which do not interconvert in solution under ambient conditions; even at high temperatures, isomerization takes several hours (halfꢀ
life 15 h at 140 °C for SAc in C Cl D ). All the S2 compounds exhibit two conductance groups and comparison with a monothioꢀ
2
4
2
lated (S1) compound shows the higher group arises from direct Auꢀporphyrin interaction. The lower conductance group is associatꢀ
ed with the SꢀtoꢀS pathway. When the binding group is SMe, the difference in junction length distributions reflects the difference in
SꢀS distance (0.3 nm) between the two isomers. In the case of SAc, there is no discernible differences between the plateau length
histograms of the two isomers, and both show maximal stretching distances well exceeding their calculated junction lengths. Conꢀ
tact deformation accounts for part of the extra length, but the results indicate that cisꢀtoꢀtrans conversion takes place in the junction
for the cis isomer. The barrier to atropisomerization is lower than the strength of the thiolate AuꢀS and AuꢀAu bonds, but higher
than that of the AuꢀSMe bond, which explains why the strain in the junction only induces isomerization in the SAc compound.
Introduction
Porphyrins are a class of organic heterocycles that garner
much attention in the fields of nanoscience and molecular
electronics. They have extensive πꢀconjugation and strong
optical absorption in the visible spectrum, leading to promiꢀ
Molecular electronics is an extremely exciting area of nanoꢀ
technology, which aims to develop new electronic devices
which operate at the single molecule level.
and controlling the conformation of molecules in single moleꢀ
cule junctions (SMJs) is a key challenge on the road to funcꢀ
tional electronic devices based on individual molecules. It is
an important consideration for rigid molecules, such as those
1
,2,3,4
Understanding
9
nent applications in lightꢀharvesting systems and dyeꢀ
10
sensitized solar cells. To study how charge transfer takes
place at the level of individual molecules, photoꢀphysical exꢀ
periments such as transient absorption spectroscopy have been
11
5
,6
used, but wiring individual molecules between a pair of elecꢀ
trodes and studying the conductance under an applied electriꢀ
based on conjugated πꢀsystems, and also for biomolecules
7
8
with weaker internal hydrogenꢀbonds such as DNA. Here, by
studying a family of porphyrin compounds in which we speꢀ
cifically vary the anchor groups and conduct a detailed platꢀ
eauꢀlength analysis, we show it is possible to probe the conꢀ
formation of the molecules between a pair of nanoꢀelectrodes,
and moreover show that atropisomerization takes place due to
the strain imposed on the molecules during tip retraction.
cal bias should lead to new insights and potential new applicaꢀ
12,13
tions.
There are a growing number of experiments on indiꢀ
14,15
vidually wired porphyrins using STM
(BJ) based methods.
and breakꢀjunction
16,17,18
The development of such singleꢀ
molecule conductance techniques allows, in particular, the
19,13
20
21,22
conductance,
IV behavior and thermopower
to be
probed at the singleꢀmolecule level. Previous studies on porꢀ
phyrin singleꢀmolecule junctions have investigated factors
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 11
2
3,15
such as number of oligomer units,
the influence of the cenꢀ
26,27
16
24,25
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tral ion, the type of binding group
and also its position
2
8
and image charge effects.
The challenge of relating the conductance vs. distance (Gz)
traces from breakꢀjunction (BJ) experiments is particularly
important for molecules with large πꢀsystems like porphyrins.
With their extended πꢀsurface, they have the potential to interꢀ
act with gold electrodes in several ways, either through the
anchor groups placed at the ends of the molecule, or via the
2
9
face of the πꢀsystem. Such Auꢀπ interactions, which have
30
been explored previously in for example oligoenes, oliꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29,6
31
go(phenylene ethynylene)s, and fullerenes, are effectively
considered to ‘shortꢀcircuit’ the wire relative to the endꢀtoꢀend
pathway. It is therefore of great importance to know how molꢀ
ecules orient, and hence interact with the electrodes, in order
to relate their conductance properties with structurally differꢀ
ent compounds. The initial objective of the work presented
here was to understand the interaction of sulfurꢀsubstituted
porphyrins with gold electrodes, so as to measure charge
Figure 1 Structures of the compounds investigated, and Xꢀray
structure of S2-trans-SMe with coordinated pyridine (yellow,
S; red, Zn; grey, C; light blue, N; white, H).
3
2
transport through larger porphyrin nanorings.
A few previous studies on porphyrins have shown examples
1
6
dipyrromethane with ꢀSAc or ꢀSMe orthoꢀsubstituted benzalꢀ
dehydes, followed by metalation with zinc(II) acetate. This
procedure yielded a mixture of cisꢀ and transꢀatropisomers,
S2-cis-SAc and S2-trans-SAc, and S2-cis-SMe and S2-trans-
SMe, respectively. The isomers were separated by HPLC and
are stable at room temperature. The cis and trans isomers have
of G profiles in the form of 2Dꢀhistograms, but there is a
z
lack of extensive investigations of this type for porphyrins. We
synthesized and isolated different conformational isomers of
5
,15ꢀdiarylporphyrins bearing either thioacetate (RSAc) bindꢀ
ing groups, which form thiolate (RS) junctions with gold, or
methyl sulfide (RSMe). These anchor groups are attached to
the aryl rings at the ortho position (relative to the porphyrin
group) which gives rise to two possible stereoisomers, a cis
and a trans form, due to hindered rotation around the phenylꢀ
porphyrin bond (Fig. 1). Our initial hypothesis was that during
1
very similar H NMR spectra with identical coupling patterns
and very small differences in chemical shifts. A preliminary
assignment was obtained by comparing their retention times
on HPLC (normal phase). The cis isomer is expected to have a
larger dipole moment than the trans and was therefore asꢀ
signed as the product with the longer retention time. For the ꢀ
SMe isomers the assignment was confirmed by Xꢀray crystalꢀ
lography on the isolated trans isomer (Figure 1). For the ꢀSAc
isomers, the following experiment was performed to confirm
junction elongation, the G profile would differ noticeably
z
between the isomers due to different preferred geometries of
the molecules within the junction. We also predicted the averꢀ
age junction breakdown distance should differ by about 0.3
nm, that is by the difference in S to S distance between cis and
trans configurations.
3
2
the assignment: Starting from a strapped porphyrin with both
thiol groups tethered into a cis arrangement by a bisꢀthioester
linker, the strap was cleaved by addition of methylamine, and
the resulting thiol groups were acylated to ꢀSAc with acetic
anhydride. A single isomer, corresponding to the slowerꢀ
running isomer on HPLC, was formed, thereby confirming the
assignment of the cis isomer.
Our main findings are as follows. Two distinctive conductꢀ
ance groups were found for all S2 compounds, and we assign
the highest (G1) to a direct Auꢀporphyrin interaction. This was
confirmed by testing a monothiolated (S1) compound, which
showed only highꢀconductance plateaus, similar to G1 of the
S2. We thus associate the lower conductance events (G2) with
a predominantly SꢀtoꢀS conductance pathway. For the lowꢀ
conductance plateaus, the initial hypothesis of shorter plateaus
for the cis isomer was confirmed for the SMe terminated comꢀ
pounds. This was not, surprisingly, the case for SAc, where we
found no appreciable difference between cis and trans isomers
in all runs. For SMe, all junction lengths were less than the
calculated values, whereas for SAc we found that plateaus
could exceed this length by 5–8 Å. To explain this behavior,
we have to take into account the AuꢀS binding energy, which
is strong enough not only to deform the electrodes but also to
force the isomerization of the cis to the trans configuration.
The AuꢀSMe binding energy is too low to affect these proꢀ
cesses.
The monoꢀfunctionalized reference compound S1-Ph-SAc
was synthesized by condensation of dipyrromethane with a
mixture of unsubstituted benzaldehyde and ꢀSAc orthoꢀ
substituted benzaldehyde. The four porphyrin products formed
in this reaction (unsubstituted 5,15ꢀdiphenylporphyrin, the
desired monoꢀfunctionalized product, and the pair of bisꢀ
functionalized cisꢀ and transꢀatropisomers) were easily sepaꢀ
rated by column chromatography, and the monoꢀ
functionalized product was metalated to give S1-Ph-SAc.
Break Junction Experiments
To form molecular junctions, we used the breakꢀjunction
method as described previously with a homeꢀbuilt scanning
33
tunneling microscope (STM). For details of sample preparaꢀ
Experimental Methods
Synthesis
tion and further methodology, please see the Supporting Inꢀ
formation (SI). Briefly, a gold STM tip is repeatedly driven in
and out of contact with a gold surface and the current moniꢀ
tored as a function of the distance travelled. At values close to
Detailed synthetic routes and procedures are presented in
the SI. The S2 compounds were obtained by condensation of
1 G (the quantum of conductance) small plateaus in G folꢀ
0
2
ACS Paragon Plus Environment

Page 3 of 11
Journal of the American Chemical Society
traces. For OPE3ꢀdithiols this drops to about 12–26 %.
3
3,34
lowed by a sharp drop indicate the final breakdown of the
We
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
metallic junction. Below this, and following the initial jumpꢀ
outꢀofꢀcontact (JOC), we observe exponential tunneling if no
molecule is wired or, in the case a molecule is wired, conductꢀ
ance plateaus develop which stand out from the exponential
background. We recorded thousands of opening/closing cycles
per experiment, and we repeated each measurement with fresh
electrodes to test for potential variation in the gold breaking
dynamics. In all experiments we use a tipꢀsample bias of 0.23
V.
conclude that junction formation for the porphyrin compounds
is to a certain extent influenced by the porphyrin unit itself.
This is understandable given that porphyrins are well known
to bind strongly to metal surfaces, and the lower percentages
of junctions are consistent with a reduced surface mobility
compared to OPEs.
35
As can be seen in Fig. 2 for the majority of S2 compounds,
two groups of plateaus can be seen, an upper group (labeled
G1) in which the plateaus are generally shorter, and a lower
group (G2) with correspondingly longer plateaus. Using our
analysis algorithm, we have further subdivided these traces
depending on whether they contain a G1 or a G2 plateau, or
both (Fig. 3). We have collated the plateau percentage data in
Table S2. We find that G1 and G2 plateaus can occur either
independently or together within the same trace, and we find
that G2 plateaus are generally the most probable, regardless of
isomer or binding group. Another significant finding is that
traces containing both a G1 and a G2 plateau occur more freꢀ
quently for the trans isomers than the cis, both for SAc and
SMe. This will be discussed later in the text. It is also remarkꢀ
able that the cis compounds can bind at all to the electrodes,
given the position of the sulfur atoms. We note, however, that
To process the data, we run an algorithm that separates moꢀ
lecular junctions from tunnelingꢀonly junctions, and from this
we create 2D density plots (histograms) of the molecular conꢀ
ductance at all electrode separations. To calculate the plateau
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
lengths for each individual G trace, we measure the distance
z
between two points in each trace, the first just after the AuꢀAu
junction cleavage (log(G/G ) = –0.5) and the other being the
0
last point in the trace which occurs one order of magnitude
below the main conductance (determined from the 1D histoꢀ
gram peak). Finally, we add 0.4 nm to account for the initial
jumpꢀoutꢀofꢀcontact (please see SI figure S26 for further deꢀ
tails). This is the minimum jump we expect to occur, and
means we do not overestimate the calibrated distance. Each
individual length is then plotted in a histogram to show the
total distribution.
36
the conformational flexibility of porphyrins is wellꢀknown,
and a 10ꢀ15 degree outꢀofꢀplane distortion could more easily
allow the sulfurs to bind to gold, as long as the electrodes are
quite sharp, as is the case in BJs.
Results and Discussion
The conductance for each group is given in Table 1, where
we have fitted a Gaussian to the peak of the 1D histogram
generated from the G1/G2ꢀonly traces, and here we quote the
maximum value. There is a clear difference in the conductance
for both G1 and G2 between SAc termination and the SMe
equivalents, with the SAc values approximately a factor four
higher than SMe (for the 1D histograms we refer to Fig. S22).
We find, however, no discernable differences in the conductꢀ
ance between cis and trans isomers with the same binding
group. We note, however, that G1 for both S2-cis-SMe and
S2-trans-SMe is more weakly defined than the SAc anaꢀ
logues.
First, we separate the plateauꢀcontaining traces from the
tunnelingꢀonly traces, which allows us to determine the perꢀ
centage of total junctions displaying molecular junction forꢀ
mation. These are shown in Fig. 2, and the tunnelingꢀonly
traces can be found in Fig. S21. For all S2 compounds (SAc
and SMe) we find this percentage falls between 7–14 %, with
no obvious bias towards one termination or isomer. These
rates are generally lower than typically found for oliꢀ
go(phenylene ethynylene), OPE, based compounds measured
under similarly solventꢀfree conditions. For either OPE3ꢀ
diamines or methyl sulfides (where 3 describes the number of
phenyl groups) plateaus are found in generally 25–60 % of all
Figure 2. 2D histograms of the “plateauꢀcontaining” traces recorded using a tipꢀsample bias of 0.23 V. The number of traces in
each histogram with the percentage of the total is A: 1288 (10.5 %), B: 510 (7.5 %), C: 1199(13.4 %), D: 1271 (9.3 %), E: 1186
19.5%).
(
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 11
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. 2D histograms of the traces separated into various groups for each compound. (A) S2-trans-SMe, G1; (B) S2-trans-
SMe, G2; (C) S2-trans-SMe, G1& G2; (D) S2-cis-SMe, G1; (E) S2-cis-SMe, G2; (F) S2-cis-SMe, G1& G2; (G) S2-trans-SAc,
G1; (H) S2-trans-SAc, G2; (I) S2-trans-SAc, G1& G2; (J) S2-cis-SAc, G1; (K) S2-cis-SAc, G2; (L) S2-cis-SAc, G1& G2; where
G1 refers to the high conductance and G2 the low conductance plateau.
The plateau lengths for each group are given in Table S3
G1) and S4 (G2). The distance between the last gold atoms of
ecule is either pulled along the electrodes the furthest, to a
smaller JOC or to the extrusion of gold atoms. We have
checked the longest plateaus, and see a tendency for the JOC
to be less for these junctions. There is still, however, an appreꢀ
ciable initial retraction of about 0.4 nm (see Fig. S26 in the SI
for details of the JOC calibration). Generally, we find the JOC
varies between 0.4 to 0.6 nm over the entire measurement, and
as it is impossible to calibrate each individual junction sepaꢀ
rately, we simply calibrate all junctions using the lowerꢀ
determined distance. This means we will underestimate the
real electrode separation in some junctions, but importantly we
(
each pyramidal electrode attached to the sulfur groups (as
generated by our DTF calculations) is 1.5/1.6 nm for S2-trans-
SAc/SMe, 1.2/1.3 nm for S2-cis-SAc/SMe. Based on this, we
expect plateaus for the trans isomers to be, on average, about
0.3 nm longer than for cis isomers. The experimentally measꢀ
ured plateau lengths can be fitted quite well with a single
Gaussian distribution, from which we can define a mostꢀ
probable stretching value (the peak maximum, L ) and a maxꢀ
p
th
imum stretching value (the 95 percentile, L ). The maximum
9
5
stretching value may be related to junctions in which the molꢀ
4
ACS Paragon Plus Environment

Page 5 of 11
Journal of the American Chemical Society
*
G
G1
G
G2
G
G2
(
log(G/G
0
))
(log(G/G
0
))
0
(log(G/G ))
1
2
3
4
5
6
7
S2-trans-SMe
S2-cis-SMe
S2-trans-SAc
S2-cis-SAc
–3.2
–3.0
–2.6
–2.5
–4.6
–4.6
–4.1
–4.1
–4.8
–4.8
–4.4
–4.4
Table 1. The mean 1D logꢀhistogram peak maxima for each
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
comꢀpound measured over two separate runs. G * is the valꢀ
ue at maximum junction extension (see SI Figure S24ꢀ25).
G2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
do not overestimate the longest junctions. Finally, we quote
all distances as the mean over two separate experimental runs.
Looking at the SMe terminated isomers first, we determined
L (L ) for S2-trans-SMe to be 0.90 (1.18) nm for G1 and
p
95
1
.10 (1.42) nm for G2. For S2-cis-SMe these values were 0.76
Figure 4. Calibrated G2 plateau length distributions for the
average of two runs per compound. L denotes the displaceꢀ
ment length as recorded in the measurement. We add 0.4 nm
to account for the JOC. Vertical lines represent the calculated
junction lengths between the final Au atoms.
(0.99) nm for G1 and 0.92 (1.22) nm for G2. It is clear that for
both isomers the L95 values of G2 correspond very well with
the expected maximum AuꢀAu separations for a single juncꢀ
tion. We also find the difference in plateau length between cis
and trans isomers (the difference in L (L ) is 0.18 (0.23) nm),
p
95
which is close to the anticipated value of 0.3 nm. This allows
us to associate with confidence G2 plateaus to junctions in
which transport takes place between the two sulfur atoms
bound to each electrode. This behavior is akin to that seen for
the family of alkanediamines studied by Arroyo et al, where
the longest plateaus did not exceed the calculated junction
be stressed that, especially for S2-cis-SMe, G1 is less wellꢀ
defined, which may be a result of the weaker interaction with
gold by the SMe group. It is interesting to note that the ratio of
conductance between G1SMe and G1SAc is the same as with
G2SMe and G2SAc, indicating that the anchor groups still play a
role in determining the overall conductance despite not fully
contributing to the current pathway in G1.
37
length.
Turning to the SAc terminated isomers next, in which the
acetyl group is known to cleave off upon exposure to gold
The observation of a separate conductance group correꢀ
sponding to an Auꢀporphyrin direct interaction independently
corroborates the low overall percentage of plateaus for both
SAc and SMe compounds. This interpretation also allows us to
understand the observation that for both cis isomers (SMe and
SAc) we consistently find fewer junctions with a transition
from G1 to G2 compared to the trans counterparts. The mean
percentage of these traces for the transꢀSMe isomer (over two
runs) is 7.7 %, which compares to just 2.4 % for the cisꢀSMe
(see Table S2 for a breakdown of each run). For the transꢀSAc
we find 19.9 % compared to only 9.0 % for the cisꢀSAc. Our
explanation for this is the following. In the case of the cis isoꢀ
mers, if the central porphyrin ring binds to one electrode, with
one S group attached to the other, the second S group cannot
easily attach to the same electrode as the porphyrin ring. This
means that once the Auꢀporphyrin contact breaks due to
stretching, the whole molecule must reorient itself in order for
the second S group to bind. The scenario is not the same for
the trans isomers, where the natural orientation of the moleꢀ
cule in the junction favors attachment of both S groups, even
when there is a direct Auꢀporphyrin interaction. We have atꢀ
tempted to represent these scenarios is Fig. S23.
3
8
under the conditions used here, we determined L (L ) for
p
95
S2-trans-SAc to be 0.95 (1.38) nm for G1 and 1.40 (2.02) nm
for G2. For S2-cis-SAc these values were 1.02 (1.41) nm for
G1 and 1.47 (2.02) nm for G2. These values are surprising for
several reasons. First, we see that there is hardly any differꢀ
ence in plateau length distributions between cis and trans isoꢀ
mers. Secondly, the L95 values greatly exceed the calculated
maximum lengths by 0.5 nm in the case of the trans isomer
and by 0.8 nm in the case of the cis isomer. Even the L value
p
of G2 for the cis isomer is greater than the 1.2 nm calculated
junction length, and so there is clearly a different junction
breaking mechanism at work during junction evolution of the
thiolates, again in agreement with the findings of Arroyo et
37
al.
Fig. 2E shows the 2D histogram of a monoꢀSAc (S1) variꢀ
ant, which cannot form sulfurꢀtoꢀsulfur bound junctions. As
we only observe one conductance group in this case, which
occurs at the same conductance as G1 for the S2 variants, this
proves that the current in these junctions must follow a path
other than the SꢀS pathway. For the S1, the calibrated maxiꢀ
mum plateau length (L ) corresponds well to the calculated
9
5
distance between a gold atom attached to the sulfur and the
opposing edge of the porphyrin ring (both 1.1 nm). Taking
into account the short nature of these plateaus, we conclude
that G1 involves a direct coupling between the porphyrin and
one of the electrodes, in effect bypassing one of the sulfur
contacts in the S2 compounds. This might include either the
central zinc ion, the πꢀsystem or a combination of both. This
distance is slightly shorter than measured for G1 for the S2
SAc compounds, which may be due to the presence of two S
atoms in the S2 case. The S1 plateau lengths are about the
same as we find for the S2 SMe compounds, although it must
We now return to understanding the mechanism underlying
the qualitatively different behavior of G2 for both cis and
trans SAc compounds compared with the SMe counterparts.
The plateau length distributions (L) for G2 are shown in Fig.
4. We begin by noting that OPE3ꢀdithiols generally show
measurably longer plateaus than the methylsulfide and amine
terminated analogues, by 0.2 to 0.3 nm, despite having practiꢀ
6
cally the same molecular length. This value is in reasonable
agreement with the difference in L and L95 we find between
p
G2 of S2-trans-SMe and S2-trans-SAc (0.3 and 0.5 nm reꢀ
spectively). We can understand this in terms of the stronger
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 11
AuꢀS thiolate bond compared with the weaker dative interacꢀ
Using these values, we can estimate the barrier at room
temperature to compare with the BJ experiments, which we
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tion of the SMe. This implies that under tension, the electrodes
will deform elastically by an amount similar to that of an Auꢀ
Au atomic junction (0.2–0.4 nm). Conversely, S2-cis-SAc
shows significantly longer plateaus than S2-cis-SMe, with Lp
and L95 values greater by 0.5 and 0.8 nm respectively. These
differences are hard to explain by just considering deformation
of the contacts at room temperature. The extra length can,
however, be accounted for satisfactorily if we assume the cis
isomer undergoes isomerization to the trans isomer during
elongation of the junction. As this would be a mechanicallyꢀ
induced effect, S2-trans-SAc could not be converted to S2-
cis-SAc. The overlap in the plateau length distributions of the
two isomers would suggest that this happens in most junctions,
otherwise a greater proportion of shorter plateaus would be
present, which we do not find. In the following we provide
further evidence for atropisomerization using a kinetic and
theoretical analysis.
‡
–1
find ∆G = 129 (± 19) kJ mol (1.34 eV) for SAc, and the
–1
same value 129 (± 18) kJ mol (1.29 eV) for SMe. Okuno et
al calculated the reaction path and potential energy barrier for
phenyl ring rotation in orthoꢀmethoxyphenylporphyrin, which
is a similar structure to our SMe/SAc compounds. They found
–1
a barrier of 105 kJ mol (1.09 eV), which is comparable to the
42
values we have measured.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Theoretical Modeling
To explain the conductance measurements, we perform abꢀ
initio based quantum transport calculation using the Gollum
43
code. Firstly, the molecules in Figure 1 were relaxed to their
optimum geometries using the density functional theory code
44
SIESTA. A doubleꢀζ polarized basis set was utilized along
with normꢀconserving pseudopotentials and the exchange corꢀ
relation functional was described by the generalized gradient
A kinetic analysis of the cis/trans atropisomerization equiꢀ
librium in solution (Figure 5) reveals that the rotational barriꢀ
ers around the porphyrinꢀaryl bond in ꢀSAc and ꢀSMe comꢀ
pounds are very similar (see also SI pages S18ꢀS20 for further
details). Neither the SAc nor the SMe sets of compounds
isomerize at room temperature, however at 140 °C in d2ꢀ
45
approximation. The fineness of the real space grid was deꢀ
fined by a mesh cutꢀoff of 150 Rydbergs and a force tolerance
of 0.01 eV/Å was used to determine the ground state geomeꢀ
try. The aryl rings have a torsion angle of 69° for the SAc and
SMe anchor groups and 63° for the nonꢀfunctionalized ring, in
keeping with a statistical analysis of conformations in crystal
1
,1,2,2ꢀtetrachloroethane their halfꢀlives drop to 15 and 11 h,
46
structures. The difference in the ground state energy for S2-
trans-SAc and S2-cis-SAc is 0.005 eV (0.003 eV for SMe).
respectively (see SI for details). Under these conditions, we
measure free energies of activation for the rotation around a
‡
–1
porphyrinꢀaryl bond of ∆G = 145.8 (± 0.4) kJ mol and 144.7
To calculate the electronic conductance, we attach the molꢀ
ecules to gold electrodes, which consist of six layers of (111)
gold each containing 25 atoms and a pyramid of 11 gold atoms
and extract a Hamiltonian describing this extended molecule
using SIESTA. For the SAc anchor group, the optimum bindꢀ
ing geometry to the gold tip via the terminal sulfur atom is
calculated to be a distance of 2.45 Å and an AuꢀSꢀC angle of
120° with a binding energy (B.E) of 1.6 eV. The B.E agrees
–1
(± 0.6) kJ mol for the SAc and SMe compounds, respectiveꢀ
ly. The rotational barriers are higher than reported values for
3
9
40
41
zinc porphyrins with ester, amide, or cyano orthoꢀ
substituents, which is consistent with the larger size of the
sulfur atom compared to oxygen, nitrogen, and carbon atoms.
Our results are also in agreement with molecular models
showing that the barrier to rotation originates mostly from the
bulk of the sulfur atom rather than its methyl or acetyl substitꢀ
uents. They suggest that the difference in molecular junction
behavior between S2-cis-SAc and S2-cis-SMe is due to the
difference in bond strength between the molecule and the gold
electrodes: for S2-cis-SMe the molecular junction breaks beꢀ
fore isomerization occurs, while for S2-cis-SAc the SꢀAu bond
is strong enough to enable a strainꢀinduced isomerization in
4
7
well with similar molecules calculated elsewhere. For SMe
anchor groups, the binding distance is slightly larger at 2.55 Å
and the angle is 117° with a much smaller B.E of 0.35 eV. The
binding for the G1 conductance group where one gold tip is
attached to the core porphyrin unit, the optimum location is
Figure 5 Atropisomerization equilibrium between cis and
trans compounds (R = Ac, Me). The isomers are stable in
solution at room temperature but exchange upon heating or
due to strain in a molecular junction.
the molecular junction. We used the Eyring equation to estiꢀ
mate the enthalpy and entropy of activation from the rates of
‡
atropisomerism at 120 and 140 °C. This analysis gives ∆H =
–
1
‡
–1
–1
8
5 (± 15) kJ mol and ∆S = –146 (± 38) J mol K for SAc
‡ –1 ‡
–1
and ∆H = 88 (± 14) kJ mol and ∆S = –138 (± 36) J mol
–1
K
for SMe, in C Cl D . Changing the solvent to d6ꢀDMSO
2 4 2
gave faster rates of isomerization, although kinetic experiꢀ
ments were less reproducible in this solvent, possibly due to
competing decomposition.
Figure 6. (Top) Contact geometries for the SAc series of molꢀ
ecules (left) G1, G2ꢀcis (middle) and G2ꢀtrans (right). (Botꢀ
tom) Zeroꢀbias transmission coefficient T(E) vs electron enerꢀ
gy for SAc anchor group (left) and SMe anchor group (right).
6
ACS Paragon Plus Environment

Page 7 of 11
Journal of the American Chemical Society
found to be directly above the central zinc atom, with a bindꢀ
ing distance of 2.9 Å and a B.E of 0.15 eV. The zeroꢀbias
transmission coefficient T(E) for each of these contact geomeꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tries obtained using the Gollum code are shown in Figure 6.
The transmission data show that the curves for the S2-cis-
SAc and S2-trans-SAc the T(E) curves are almost identical,
with the DFTꢀpredicted Fermi energy E sitting close to the
F
HOMO resonance of the molecule, which is typical behavior
for thiol anchor groups. The value of log(G/G ) = log T(E ) is
0
F
–
2.8 for cis and –2.9 for trans. For the S1-Ph-SAc the conꢀ
ductance is higher with a value of –2.4. These values are
greater than the experimental measurements in Table 1. The
overestimation arises due to known problems with DFT in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
48
predicting the correct position of the Fermi energy. In pracꢀ
tice E usually lies closer to the middle of the HOMOꢀLUMO
F
gap, which in our case is approximately E – E = 0.5 eV. From
F
the value of T(E) at this energy, we find that log G/G is –3.98
0
for S2-cis-SAc, –4.02 for S2-trans-SAc and –3.2 for S1-Ph-
SAc which is in much better agreement with experiment. The
calculated goldꢀgold separation for each of these three juncꢀ
tions is 1.5 nm (S2-trans-SAc), 1.2 nm (S2-cis-SAc) and 0.96
nm (S1-Ph-SAc).
The transmission curves for the SMe terminated molecules
show similar behavior (Figure 6) with the cis and trans conꢀ
figurations possessing almost identical transmission curves.
Again the goldꢀgold separations are different (1.3 nm cis, 1.6
nm trans). The DFTꢀpredicted Fermi energy now lies close to
the LUMO resonance, which would lead to conductances
which are too high. Shifting this towards the middle of the gap
Figure 7. Different configurations and transitions of the moleꢀ
cules between the electrodes. A to A represents the transition
from group 1 (G1) to group 2 (G2). B1 to B2 represents atroꢀ
pisomerization of S2ꢀorthoꢀcis to S2ꢀorthoꢀtrans. C1 to C2
describes the SMeꢀanchored cis junction, which cleaves before
isomerization can occur.
1
2
to an energy E – E = –0.5 eV yields conductance values
F
log(G/G ) of –4.43 for S2-cis-SMe and –4.50 for S2-trans-
0
SMe which are lower than the conductance values for the thiol
series in agreement with the measured values in Table 1.
We have carried out a theoretical analysis of a possible tranꢀ
sition state geometry using a representative isolated molecule.
For details of the methods used please see SI Section 5. This
was carried out by rotating one aryl group stepwise and by
evaluating the energy of the system at each step. For each
point, freezing some atoms was necessary in order to prevent
the system from regaining the original structure. By performꢀ
ing such an analysis, a barrier of 1.77 eV was obtained, and
The energy barrier for rotation of the end aryl group about
the porphyrin axis is large due to the steric effects of the addiꢀ
tional thiol and SMe groups in the ortho position. A detailed
description of the energy landscape of these molecules can be
found in the SI (Section 3). If the porphyrin is constrained to
retain its optimum geometry, the energy barrier is extremely
high. However, if the porphyrin is allowed to deform within
0
we show the transition structure (at α ≈ 145 ) in Fig. S32
42
the junctions, then the energy barrier is significantly lower.
(lower right panel). This possible energy pathway is in good
Also, the B.E of the goldꢀgold bond at the apex of a pyramid
of gold atoms is calculated to be 1.62 eV which is slightly
larger than the gold thiol bond. For thiols experimentally, the
mean breakdown force of individual junctions has been measꢀ
agreement with the experimentally determined value, although
we are still overestimating the value somewhat due to the reꢀ
strictions imposed.
4
9
50
This mechanism indicates that the porphyrin adopts signifiꢀ
cantly nonꢀplanar conformations inside the junction, which is
something seldom considered, especially for conjugated πꢀ
systems. Such deviations from nonꢀplanarity may be important
when trying to understand the conductance of long organic
molecules and also flexible bioꢀmolecules such as proteins and
DNA. Due to the size of the barrier, we do not expect thermal
fluctuations or bias voltage (0.23 V used here) to play a signifꢀ
icant role in the isomerization process. The observed relation
between isomerization and moleculeꢀelectrode binding group
strongly implies this is a mechanicallyꢀdriven phenomenon.
Although an exact value for the barrier cannot be extracted
from the experiments, we can place upper and lower limits of
ured as close to 1.5 nN,
(the same as that for an AuꢀAu
bond, implying breakdown of thiolated junctions could occur
either between AuꢀAu or AuꢀS bonds). For SMe, the junction
5
1
breakdown force has been measured to be 0.5 nN, For cisꢀtoꢀ
trans conversion to be induced by the junction, the amount of
energy needed to break the bond at the contact must be larger
than the barrier to atropisomerization. Based on the B.E calcuꢀ
lations, which give a larger B.E for the thiol than the measured
rotational barrier, but a lower B.E in the case of SMe, as well
as the experimental force measurements, it seems perfectly
plausible that during the stretching of thiol junctions, more
work is required to break the AuꢀAu/AuꢀS bond than to overꢀ
come the rotational barrier, hence isomerization can take
place. This clearly is not the case for SMe, where the junction
breaks before isomerization can occur (this is represented
schematically in Fig. 7).
1
.6 eV and 0.35 eV using the B.E of AuꢀS and AuꢀSMe
groups respectively. This fits nicely with the value from our
kinetic analysis of 1.34 eV (S2-trans-SAc). To obtain greater
precision would require studying other binding groups, with a
greater range of B.E.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 11
We do not observe an obvious transition in any of the indiꢀ group is related to shorter conductance path which does not
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
vidual traces or the 2D histogram for the isomerization of S2-
cis-SAc. This could be for several reasons. First, the isomeriꢀ
zation most likely occurs on time scales much too short to be
detected by our instrumentation (molecular motions are generꢀ
ally on the picoꢀsecond scale). This would explain why the
conductance we observe for cis/trans isomers is the same
involve both sulfur atoms. The plateau lengths suggest binding
to the central porphyrin unit. Our theoretical study fully supꢀ
ports these observations, and we can reproduce G1 and G2
using the experimentally inferred binding locations. We also
show theoretically that the conductance does not depend on
the trans or cis configuration even though the electrode sepaꢀ
rations differ.
(
within error). Despite the implication that the torsion angle
between one of the aryl groups and the central porphyrin ring
of S2-cis-SAc varies more during stretching compared to S2-
trans-SAc, which would be expected to produce a change in
For the SꢀS pathway junctions (G2) we found consistent
differences in the breakdown distance distributions for the
SMe compounds, the cis isomer showing overall shorter platꢀ
eaus than the trans. We found the maximum observed breakꢀ
down distance fits well to the calculated junction length for
each compound, and practically no junctions exceed this valꢀ
ue. In contrast, the SAc compounds showed no noticeable
difference in breakdown distance, and both often exceed the
calculated fullyꢀstretched value, which for the cis isomer was
up to 0.8 nm beyond this value. In order to reconcile this, deꢀ
formation of the contacts explains part of the ‘extra’ distance,
but an additional mechanism must be at work. We investigated
the potential energy barrier to ring rotation for the outer pheꢀ
nyl groups, carrying out a kinetic analysis of the cis/trans atꢀ
ropisomerization equilibrium in solution. From this we estiꢀ
52
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
conductance, the likely quick nature of this process is not
expected to produce a noticeable change in the average distriꢀ
bution of torsion angles across the junction. Also, individual
traces fluctuate by about one order of magnitude in conductꢀ
ance due to changes in moleculeꢀelectrode contact, and this
3
3
may mask traces of the isomerization process. We notice too
that all wired cisꢀthiol molecules seem to be isomerized, which
we infer due to the absence of a higher proportion of shorter
plateaus for S2ꢀcisꢀSAc compared to the trans.
Another interesting, and perhaps surprising, outcome is that
the cisꢀconfiguration can bind to the electrodes at all. Steric
hindrance between the Hꢀatoms of the phenyl ring and gold
might be expected to strongly reduce the possibility of wiring.
We note, however, that the flexibility of the porphyrin, espeꢀ
cially in the presence of gold as we have shown, can quite
conceivably lead to more favorable conformations that allow
AuꢀS binding at both ends of the molecule. This highlights
once again that understanding conformational degrees of freeꢀ
dom inside molecular junctions is key for a sound interpretaꢀ
tion of results.
–1
mated the room temperature barrier to be 129 kJ mol (1.34
eV) for both SAc and SMe. Our theoretical calculations
showed that the AuꢀS (for thiolate connection) and AuꢀAu
bond energies are greater than this (1.60 and 1.62 eV respecꢀ
tively), whereas the AuꢀSMe bond is significantly weaker
(0.35 eV). We infer, therefore, that for the cis SAc isomer,
strainꢀinduced cis to trans isomerization takes place during the
stretching of a single molecule junction. This process does not
occur for the SMe analogue. For isomerization to occur, the
porphyrin must adopt significantly nonꢀplanar conformations
within the junction, which stresses the importance of considerꢀ
ing molecularꢀdeformation inside junctions.
The method we have outlined for controlling molecular conꢀ
formation using different anchor groups could be applied to
any system with one, or several, internal conꢀformational barꢀ
riers. We envisage this could be used, perhaps, to shed more
light on the effect of unraveling of Hꢀbonded molecules or the
unfolding of DNA, peptides or proteins on their transport
properties. Having weaker binding groups would allow moꢀ
lecular conformations to persist under stretching, whereas
stronger binding groups would force the molecule to unravel,
allowing for a detailed comparison of both situations.
ASSOCIATED CONTENT
The Supporting Information is available free of charge via the
Internet at http://pubs.acs.org. This includes all synthetic proceꢀ
dures and analysis, details of the kinetic study of atropisomerizaꢀ
tion, Xꢀray crystal structures, general methods for the BJ study
plus further analysis, theoretical binding energy and groundꢀstate
energy analyses. The DOI for the STM data set is:
We further point out that knowledge of the anchor group
binding strength and internal atropisomerization barriers can
be considered complementary, and in principle the variation of
one could be used to elucidate the other to a reasonable degree
of accuracy. This might lead to more elaborate studies in
which anchor group binding strengths can be assessed through
internal reference barriers, complimenting the force measureꢀ
1
0.17638/datacat.liverpool.ac.uk/398.
AUTHOR INFORMATION
Corresponding Authors
5
0,53
ments currently performed.
*
*
E.Leary@liverpool.ac.uk;
Conclusions
harry.anderson@chem.ox.ac.uk;
In conclusion, we have studied a family of porphyrin comꢀ
pounds in which the binding groups (either SAc or SMe) are
located ortho to the central porphyrin on the outer phenyl
rings, giving rise to two distinct atropisomers, cis and trans.
For all S2 compounds, two conductance groups were observed
*
*
c.lambert@lancaster.ac.uk
nichols@liverpool.ac.uk;
Author Contributions
(
an upper G1 and a lower G2), the conductance values of
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the manuꢀ
script.
which do not depend significantly on the isomerization of the
molecule. The SMe anchor gave consistently lower conductꢀ
ance values than SAc, by about a factor four. Replacing one of
the terminal PhSAc groups with a Ph inhibited the lower G2
plateaus, but still displayed the upper G1, showing that this
8
ACS Paragon Plus Environment

Page 9 of 11
1
Journal of the American Chemical Society
16)
(
Liu, Z.ꢀF.; Wei, S.; Yoon, H.; Adak, O.; Ponce, I.;
Jiang, Y.; Jang, W.ꢀD.; Campos, L. M.; Venkataꢀ
raman, L.; Neaton, J. B. Nano Lett. 2014, 14, 5365.
Li, Z.; Park, T.ꢀH.; Rawson, J.; Therien, M. J.;
Borguet, E. Nano Lett. 2012, 12, 2722.
Acknowledgements
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RJN, SH EL and AH acknowledge the UK Engineering and Physꢀ
ical Sciences Research Council (EPSRC) for the funding of this
research under grant numbers EP/M014169/1, P/M029522/1 and
EP/M005046/1. CL, IG, QAG acknowledge the European Comꢀ
mission (EC) FP7 ITN “MOLESCO” (project no. 606728) and the
(17)
(18)
Noori, M.; Aragonès, A. C.; Di Palma, G.; Darwish,
N.; Bailey, S. W. D.; AlꢀGaliby, Q.; Grace, I.; Amaꢀ
bilino, D. B.; GonzálezꢀCampo, A.; DíezꢀPérez, I.;
Lambert, C. J. Sci. Rep. 2016, 6, 37352.
o
EPSRC (grant n s. EP/M014452/1 and EP/N017188/1). HLA, CR
and HWJ thank the EPSRC (grant EP/M016110/1), UK National
Mass Spectrometry Facility at Swansea University for recording
mass spectra and Dr. Amber L. Thompson for help with Xꢀray
crystal structure refinements. KP and ML acknowledge the
EPSRC grant EP/K030108/1. We thank Linda A. Zotti for useful
discussion.
(
19)
20)
Chen, F.; Hihath, J.; Huang, Z.; Li, X.; Tao, N. Annu
Rev Phys Chem 2007, 58, 535.
Widawsky, J.; Kamenetska, M.; Klare, J.; Nuckolls,
C.; Steigerwald, M.; Hybertsen, M.; Venkataraman, L.
Nanotechnology 2009, 20, 434009.
Reddy, P.; Jang, S.ꢀY.; Segalman, R. A.; Majumdar,
A. Science 2007, 315, 1568.
Evangeli, C.; Gillemot, K.; Leary, E.; González, M.
T.; RubioꢀBollinger, G.; Lambert, C. J.; Agraït, N.
Nano Lett. 2013, 13, 2141.
Sedghi, G.; Sawada, K.; Esdaile, L. J.; Hoffmann, M.;
Anderson, H. L.; Bethell, D.; Haiss, W.; Higgins, S.
J.; Nichols, R. J. J. Am. Chem. Soc. 2008, 130, 8582.
Li, Z.; Smeu, M.; Park, T.ꢀH.; Rawson, J.; Xing, Y.;
Therien, M. J.; Ratner, M. A.; Borguet, E. Nano Lett.
2014, 14, 5493.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
(21)
Notes
(22)
(23)
(24)
The authors declare no competing financial interests.
References
(
1)
Su, T. A.; Neupane, M.; Steigerwald, M. L.; Venkataꢀ
raman, L.; Nuckolls, C. Nat. Rev. Mater. 2016, 1,
1
6002.
Van Dyck, C.; Ratner, M. A. J. Phys. Chem. C 2017,
21, 3013.
Perrin, M. L.; Burzurí, E.; van der Zant, H. S. Chem.
Soc. Rev. 2015, 44, 902.
Cuevas, J. C.; Scheer, E. Molecular electronics: an
introduction to theory and experiment; World Scienꢀ
tific, 2010.
(25)
Li, Z.; Smeu, M.; Ratner, M. A.; Borguet, E. J. Phys.
Chem. C 2013, 117, 14890.
(2)
(3)
(4)
1
(
(
26)
27)
Li, Z.; Borguet, E. J. Am. Chem. Soc. 2011, 134, 63.
Kiguchi, M.; Takahashi, T.; Kanehara, M.; Teranishi,
T.; Murakoshi, K. J. Phys. Chem. C 2009, 113, 9014.
Perrin, M. L.; Verzijl, C. J.; Martin, C. A.; Shaikh, A.
J.; Eelkema, R.; Van Esch, J. H.; Van Ruitenbeek, J.
M.; Thijssen, J. M.; Van Der Zant, H. S.; Dulić, D.
Nat. Nanotechnol. 2013, 8, 282.
Haiss, W.; Wang, C.; 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.
Meisner, J. S.; Kamenetska, M.; Krikorian, M.; Steiꢀ
gerwald, M. L.; Venkataraman, L.; Nuckolls, C. Nano
Lett. 2011, 11, 1575.
Gillemot, K.; Evangeli, C.; Leary, E.; La Rosa, A.;
González, M. T.; Filippone, S.; Grace, I.; Rubioꢀ
Bollinger, G.; Ferrer, J.; Martín, N.; Lambert, C. J.;
Agraït, N. Small 2013, 9, 3812.
Roche, C.; Luo, Q.; GilꢀRamírez, G.; Jiang, H.ꢀW.;
Kohn, D. R.; Xiong, Y.; Thompson, A. L.; Anderson,
H. L. J. Org. Chem. 2017, 82, 7446.
(28)
(5)
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.
Miguel, D.; Alvarez de Cienfuegos, L.; Martínꢀ
Lasanta, A.; Morcillo, S. P.; Zotti, L. A.; Leary, E.;
Bürkle, M.; Asai, Y.; Jurado, R.; Cárdenas, D. J. J.
Am. Chem. Soc. 2015, 137, 13818.
(29)
(30)
(31)
(6)
(
7)
8)
Wang, L.; Gong, Z.; Li, S.; Hong, W.; Zhong, Y.;
Wang, D.; Wan, L. Angew. Chem. 2016, 128, 12581.
Bruot, C.; Xiang, L.; Palma, J. L.; Li, Y.; Tao, N. J.
Am. Chem. Soc. 2015, 137, 13933.
(
(9)
Li, L.ꢀL.; Diau, E. W.ꢀG. Chem. Soc. Rev. 2013, 42,
291.
(32)
(10)
Mathew, S.; Yella, A.; Gao, P.; HumphryꢀBaker, R.;
Curchod, B. F.; AshariꢀAstani, N.; Tavernelli, I.;
Rothlisberger, U.; Nazeeruddin, M. K.; Grätzel, M.
Nat. Chem. 2014, 6, 242.
(33)
Parker, C. R.; Leary, E.; Frisenda, R.; Wei, Z.; Jenꢀ
num, K. S.; Glibstrup, E.; Abrahamsen, P. B.; Santelꢀ
la, M.; Christensen, M. A.; Della Pia, E. A.; Li, T.;
Gonzalez, M. T.; Jiang, X.; Morsing, T. J.; Rubioꢀ
Bollinger, G.; Laursen, B. W.; Nørgaard, K.; van der
Zant, H.; Agrait, N.; Nielsen, M. B. J. Am. Chem. Soc.
(
11)
12)
Armaroli, N.; Accorsi, G.; Song, F.; Palkar, A.; Echeꢀ
goyen, L.; Bonifazi, D.; Diederich, F. ChemPhysChem
2
005, 6, 732.
(
Xu, B.; Tao, N. Science 2003, 301, 1221.
Nichols, R. J.; Haiss, W.; Higgins, S. J.; Leary, E.;
Martin, S.; Bethell, D. Phys. Chem. Chem. Phys.
2010, 12, 2801.
2
014, 136, 16497.
(13)
(34)
Teresa Gonzalez, M.; Leary, E.; Garcia, R.; Verma,
P.; Angeles Herranz, M.; RubioꢀBollinger, G.; Martin,
N.; Agrait, N. J. Phys. Chem. C 2011, 115, 17973.
Jung, T. A.; Schlittler, R. R.; Gimzewski, J. K. Nature
1997, 386, 696.
(
14)
15)
Bennett, N.; Xu, G.; Esdaile, L. J.; Anderson, H. L.;
Macdonald, J. E.; Elliott, M. Small 2010, 6, 2604.
Sedghi, G.; GarcíaꢀSuárez, V. M.; Esdaile, L. J.; Anꢀ
derson, H. L.; Lambert, C. J.; Martín, S.; Bethell, D.;
Higgins, S. J.; Elliott, M.; Bennett, N. Nat. Nanotechꢀ
nol. 2011, 6, 517.
(35)
36)
(
(
Senge, M. O.; MacGowan, S. A.; O’Brien, J. M.
Chem. Commun. 2015, 51, 17031.
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 11
(
37)
Arroyo, C. R.; Leary, E.; CastellanosꢀGomez, A.; Ruꢀ
bioꢀBollinger, G.; Teresa Gonzalez, M.; Agrait, N. J.
Am. Chem. Soc. 2011, 133, 14313.
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.
(46)
(47)
Anderson, H. L. Chem. Commun. 1999, No. 23, 2323.
Zotti, L. A.; Kirchner, T.; Cuevas, J.ꢀC.; Pauly, F.;
Huhn, T.; Scheer, E.; Erbe, A. Small 2010, 6, 1529.
Quek, S. Y.; Khoo, K. H. Acc. Chem. Res. 2014, 47,
3250.
Li, X.; He, J.; Hihath, J.; Xu, B.; Lindsay, S.; Tao, N.
J. Am. Chem. Soc. 2006, 128, 2135.
Nef, C.; Frederix, P. L. T. M.; Brunner, J.; Schönenꢀ
berger, C.; Calame, M. Nanotechnology 2012, 23,
365201.
Aradhya, S. V.; Meisner, J. S.; Krikorian, M.; Ahn, S.;
Parameswaran, R.; Steigerwald, M. L.; Nuckolls, C.;
Venkataraman, L. Nano Lett. 2012, 12, 1643.
Mishchenko, A.; Vonlanthen, D.; Meded, V.; Buerkle,
M.; Li, C.; Pobelov, I. V.; Bagrets, A.; Viljas, J. K.;
Pauly, F.; Evers, F.; Mayor, M.; Wandlowski, T. Nano
Lett. 2010, 10, 156.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(38)
(48)
(49)
(50)
(39)
(40)
(41)
Fujimoto, T.; Umekawa, H.; Nishino, N. Chem. Lett.
1
992, 21, 37.
Freitag, R. A.; Whitten, D. G. J. Phys. Chem. 1983,
7, 3918.
8
Hatano, K.; Kawasaki, K.; Munakata, S.; Iitaka, Y.
Bull. Chem. Soc. Jpn. 1987, 60, 1985.
Okuno, Y.; Kamikado, T.; Yokoyama, S.; Mashiko, S.
J. Mol. Struct. Theochem 2002, 594, 55.
Ferrer, J.; Lambert, C. J.; GarcíaꢀSuárez, V. M.; Manꢀ
rique, D. Z.; Visontai, D.; Oroszlany, L.; Rodríguezꢀ
Ferradás, R.; Grace, I.; Bailey, S.; Gillemot, K. New J.
Phys. 2014, 16, 093029.
Soler, J.; Artacho, E.; Gale, J.; Garcia, A.; Junquera,
J.; Ordejon, P.; SanchezꢀPortal, D. J. Phys.ꢀCondens.
Matter 2002, 14, 2745.
(51)
(52)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(42)
(43)
(53)
Xu, B.; Xiao, X.; Tao, N. J. J. Am. Chem. Soc. 2003,
125, 16164.
(
44)
45)
(
Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev.
Lett. 1996, 77, 3865.
10
ACS Paragon Plus Environment

Page 11 of 11
Journal of the American Chemical Society
TOC graphic:
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
11
ACS Paragon Plus Environment