Electrochemical Single-Molecule Transistors with Optimized Gate Coupling

Article abstract of DOI:10.1021/jacs.5b08431

Electrochemical gating at the single molecule level of viologen molecular bridges in ionic liquids is examined. Contrary to previous data recorded in aqueous electrolytes, a clear and sharp peak in the single molecule conductance versus electrochemical potential data is obtained in ionic liquids. These data are rationalized in terms of a two-step electrochemical model for charge transport across the redox bridge. In this model the gate coupling in the ionic liquid is found to be fully effective with a modeled gate coupling parameter, ξ, of unity. This compares to a much lower gate coupling parameter of 0.2 for the equivalent aqueous gating system. This study shows that ionic liquids are far more effective media for gating the conductance of single molecules than either solid-state three-terminal platforms created using nanolithography, or aqueous media.

Full text of DOI:10.1021/jacs.5b08431

Subscriber access provided by CMU Libraries - http://library.cmich.edu
Article
Electrochemical single-molecule transistors with optimized gate coupling
Henrry M. Osorio, Samantha Catarelli, Pilar Cea, Josef B. G. Gluyas, Frantisek Hartl,
Simon J. Higgins, Edmund Leary, Paul J. Low, Santiago Martin, Richard J. Nichols,
Joanne Tory, Jens Ulstrup, Andrea Vezzoli, Qiang Zeng, and David Costa Milan
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b08431 • Publication Date (Web): 21 Oct 2015
Downloaded from http://pubs.acs.org on October 25, 2015
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 40
Journal of the American Chemical Society
1
2
Electrochemical single-molecule transistors with optimized gate coupling
3
4
5
6
7
8
9
Henrry M. Osorio,(a) Samantha Catarelli,(b) Pilar Cea,(a,c) Josef B. G. Gluyas,(d) František
Hartl,(e) Simon J. Higgins,(b) Edmund Leary,(b) Paul J. Low,(d) Santiago Martín,(a,f) Richard J.
Nichols,(b)* Joanne Tory,(e) Jens Ulstrup,(g) Andrea Vezzoli,(b) David C. Milan(b) and Qiang
Zeng.(e)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(a) Departamento de Química Física, Facultad de Ciencias, Universidad de Zaragoza,
50009, Spain.
(b) Department of Chemistry, University of Liverpool, Crown Street, Liverpool, L69 7ZD,
United Kingdom.
(c) Instituto de Nanociencia de Aragón (INA) and Laboratorio de microscopias avanzadas
(LMA), edificio i+d Campus Rio Ebro, Universidad de Zaragoza, C/Mariano Esquillor, s/n,
50018 Zaragoza, Spain.
(d) School of Chemistry and Biochemistry, University of Western Australia, 35 Stirling
Highway, Crawley 6009, Australia
(e) Department of Chemistry, University of Reading, Whiteknights, Reading, RG6 6AD,
UK
(f) Instituto de Ciencias de Materiales de Aragón (ICMA), Universidad de Zaragoza-CSIC,
50009 Zaragoza, Spain.
(g) Department of Chemistry and NanoDTU, Technical University of Denmark, DK2800
Kgs. Lyngby, Denmark.
* Corresponding author e-mail address: nichols@liv.ac.uk
ABSTRACT
Electrochemical gating at the single molecule level of viologen molecular bridges in ionic
liquids is examined. Contrary to previous data recorded in aqueous electrolytes, a clear and
sharp peak in the single molecule conductance versus electrochemical potential data is
obtained in ionic liquids. These data are rationalized in terms of a two-step electrochemical
ꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 40
1
2
model for charge transport across the redox bridge. In this model the gate coupling in the
ionic liquid is found to be fully effective with a modeled gate coupling parameter, ξ, of unity.
This compares to a much lower gate coupling parameter of 0.2 for the equivalent aqueous
gating system. This study shows that ionic liquids are far more effective media for gating the
conductance of single molecules than either solid-state three-terminal platforms created
using nanolithography, or aqueous media.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords: spectro-electrochemistry, electrochemical double layer, STM, molecular
electronics, charge transfer, viologens, ionic liquids.
Introduction
Molecular switching, for instance between low and high conductance states, has been a
theme of ongoing interest in molecular electronics and it has been shown in recent years
that this can be achieved even down to the single molecule level.1 Molecular analogues of
field effect transistors have been achieved, in which the conductance state of a molecular
junction can be electrostatically modulated by a third (gate) terminal. In solid-state devices
this has to be achieved by bringing the gate terminal into close proximity to the molecule that
is wired between two contacting electrodes (source and drain). Since the separation of the
source and drain electrodes will be of molecular dimensions, it is a considerable challenge to
bring the gate electrode reliably into close enough proximity to gate the molecular junction
electrostatically. Nevertheless, capacitive coupling of the molecular junction to the gate
electrode has been achieved in solid-state platforms either using nanolithographically
fabricated electrodes, or mechanically controlled break junctions constructed on a substrate
bearing an electrostatic ‘back gate’ electrode. If the gate is perfectly electrostatically coupled
to the junction, then frontier HOMO and LUMO molecular orbitals of the bridge would shift in
unison with the applied gate voltage, to give a gate coupling parameter, here called ξ, of 1
(i.e. a 0.1 V shift of the gate voltage produces a 0.1 eV shift in the molecular orbital energy).
Although relatively large coupling parameters have been achieved for individual devices
ꢁ
ACS Paragon Plus Environment

Page 3 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(e.g. ~0.2),2 the nanofabrication processes do not generally lend themselves to precise and
reproducible placement of the gate electrode, and ξ is typically much lower than this, and
moreover exhibits significant variation from device to device. An alternative approach is to
use liquid gates, with control of the electrochemical double layer.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
“Electrolyte gating” relies on the formation of electrochemical double layers along the
source-drain channel. Since electrochemical double layers can have nanometer thicknesses
separating electronic surface charge and ions in solution, even at modest electrolyte
concentrations, they provide both an effective and a highly reproducible gating platform.
Moreover, the gate voltage is readily controlled through variation of the electrochemical
potential. Electrolyte gating has been applied to a wide variety of bulk and nanostructured
materials including nanocrystalline thin filmsꢃ5 carbon nanotubes,3,4 and solid-state
semiconductor devices in which the source-drain channel is a solid semiconductor or even a
superconductor.6,7 Such devices are typically referred to as electrochemical double layer
transistors (EDLTs). In 1984 White et al. reported on the fabrication of a polypyrrole
derivatized microelectrode array that operated as a transistor through electrochemical
control of the oxidation of the bulk conducting polymer film by a gating electrode.8 More
recently, this electrochemical gating concept has been extended to the single molecular
bridge level, in which the electrochemical potentials and resultant double layers tune
molecular orbital energy levels. In the case of redox active molecular wires, this can be used
to switch the molecule between accessible redox states. A schematic illustration of
electrolyte gating of a single-molecule bridge is shown in Figure 1. The single molecule is
tethered between the STM tip and substrate surface, which can be viewed as the drain and
source electrodes respectively. The counter and reference electrode combination provides
the gate electrode in this 4-electrode bipotentiostat setup, with independent electrochemical
potential control of the substrate (working electrode 1, or “source”) and the STM tip (working
electrode 2, or “drain”). The potential difference between these two electrodes is the bias
ꢂ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 40
1
2
3
4
5
6
voltage applied to the molecular junction. Figure 1B illustrates that the redox group is gated
by the close proximity of electrolyte ions
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: (A) Electrochemical single molecule gating using an electrochemical STM with
bipotentiostat control of the electrochemical potential of the substrate and STM tip. The
single molecule bridge shown here (12+, see below) is derived from 1,1'-bis(6-
(acetylthio)hexyl)-4,4'-bipyridinium hexafluorophosphate. (B) Illustrates the close proximity of
the electrolyte ions which are “gating” the redox group (the electrolyte gating concept).
There is now a growing number of examples of electrochemical gating of single molecules,
using the single-molecule electrochemical transistor-like configuration where the molecular
bridge is tethered between substrate and STM tip, both of which are typically made of gold.9-
33 Viologen molecular wires were the first electrochemical redox system studied in this
single-molecule junction configuration, with the molecule anchored to the source and drain
ꢄ
ACS Paragon Plus Environment

Page 5 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
electrodes through chemisorbed thiol contacting groups.10 This redox bridge system (the
V2+/V+. redox system, with V=viologen) has continued to be of interest due to its highly
reversible electrochemical reduction processes.11-13 Other examples of such “wired” single-
molecule electrochemical junctions include molecular wires containing pyrrolo-
tetrathiafulvalene (pTTF),14,15 oligoanilines,16-18 oligothiophenes,19 unsubstituted and
substituted oligo(phenylene ethynylenes) (OPEs),20-22 ferrocene,23 anthraquinone,24-26
perylene tetracarboxylic bisimides (PBI),27-31 carotenoids 32 and benzodifuran.33 More
recently, the electrolyte gating of non-redox active molecules has also been examined.34-36
Electrolyte gating where the redox active molecule is attached to the substrate but not the
STM tip, or occasionally vice versa37-47 has also been demonstrated. In this configuration,
electron tunneling through the adsorbed molecule as a function of electrochemical potential
with the STM tip located above the adsorbed layer is monitored.37 This can be referred to as
an electrochemical scanning tunneling spectroscopy configuration, and systems thus studied
include, for example, metal complexes,37,40,41,43,44 metalloproteins,48-50 DNA,51 52 and
viologens13,46. Albrecht et al. pioneered electrochemical scanning tunneling spectroscopy
experiments conducted in ionic liquid environments. 53 There have also been studies of
electrolyte gating in large area planar molecular junctions,54 as well as devices featuring
single nanoparticles55 or fabricated nanoelectrode gaps.56,57
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A comparison of the electrochemical gate voltage behavior reported in the single molecule
studies mentioned above clearly indicates that the resulting gating response is highly
variable. In some cases, in a plot of single-molecule conductance against applied
electrochemical potential a peak is seen as the voltage is adjusted through the reversible
potential for the redox switching.14,15,28,29,33,58,59 In other cases, no defined peak in the
molecular conductance versus electrochemical potential is observed, but rather a
monotonically rising conductance spanning many hundreds of millivolts; an early example of
this latter behavior is the electrochemical gating of the conductance of viologen molecular
wires in aqueous solution.10,13 In this case the conductance gradually rises as the
ꢅ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 40
1
2
electrochemical potential is taken from positive values where the molecule is in its oxidized
(bipyridinium) state, to potentials negative of its first reduction;10-13 the term “soft gating” was
coined to refer to this phenomenon.12 Similar broad conductance rise or broad conductance
gating peaks feature in other electrochemical single-molecule conductance switching
systems, such as those based on perylene bisimides,28,29,31,58 benzodifuran,33
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
anthraquinone-based norbornylogous bridges24 25 and metalloproteins.59 60
In this present study we examine electrochemical conductance gating of viologen-based
single-molecule wire systems in ionic liquids and compare the response with data from
aqueous media. The data recorded in ionic liquid reported here fit within the two-step
(hopping) mechanism introduced by Kuznetsov and Ulstrup with a high gating coupling
parameter of ξ = 1. We refer here to such effective gating with a high gate coupling
parameter as “hard gating”. When the same model is applied to the data from similar studies
of 12+ in aqueous solution, a much lower gate coupling parameter of ξ = 0.2 is obtained,
which accounts for the broad rise in conductance with electrochemical potential.10,12,13 This
difference in gating behavior between the two electrolytes arises from the gating provided by
the double layers, which are expected to be very different in ionic liquids, which represent
“highly concentrated and dense” 61 electrolyte media when compared with dilute aqueous
electrolytes (the electrolytic properties of ionic liquids have been reviewed elsewhere 61).
UV/visible and IR spectroelectrochemical experiments in both aqueous and ionic liquid
electrolytes show that the electronic structure of the reduced viologen is not changed by the
ionic liquid electrolyte. The present study shows that ionic liquid electrochemical gating of
single-molecule conductance is far more effective than gating that is generally achieved
either in solid-state three-terminal platforms created using nanolithography or gating with
aqueous double layers.
■ RESULTS
The viologen based molecular wires used in this study (Chart 1) contained either classical
thiol contacting groups (12+), or the more esoteric trimethylsilylethynyl contacting groups
ꢆ
ACS Paragon Plus Environment

Page 7 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(22+). The acetyl-protected precursor 3•[PF6]2 was synthesized by the route shown in
Scheme S1 (Supporting Information) and is an adaptation of the synthesis of bromide salts
of 32+ described elsewhere. 62 Monolayers of 12+ form in situ upon exposure of gold
substrates to solutions of the acetyl protected precursor 3•[PF6]2 (see Methods) for STM
investigations.14 Compound 22+ was synthesized as the bis tetrafluoroborate salt (Scheme 1).
Sonogashira cross-coupling63-65 of ethynyltrimethylsilane (7) with 4-iodobenzyl alcohol (8),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
66
gave 4-(trimethylsilyl)ethynyl-benzyl alcohol (9). Bromination of compound 9 with PPh3•Br2
gave 4-(trimethylsilyl)ethynyl-benzyl bromide (10). The target viologen (22+) was formed by
condensation of 4,4’-bipyridine with 10 under reflux in acetonitrile and isolated following
exchange of the bromide counter ions with ammonium tetrafluoroborate.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 40
1
2
3
Chart 1: The viologen based molecular wires (12+ and 22+) used in the STM studies
4
5
described within. Under the experimental conditions used the counter ions are determined
6
7
8
by the electrolytic medium. 3•[PF6]2 is the acetyl protected precursor of 12+ with PF6
-
9
counter-ions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Synthesis of 2•[BF4]2, yields are given in parentheses.
Cyclic voltammetry of the initial electrochemical reduction of 3•[PF6]2,the thioacetate
protected form of 12+, in the ionic liquid 1-butyl-3-methylimidazolium triflate (BMIM-OTf) to its
radical cation is shown in Figure 2. The redox wave centered at 0 V arises from ferrocene
(Fc) added to the electrolyte for voltammetric calibration.67 The peak-to-peak separation for
the viologen (32+ ↔ 3+.) redox couple is ~ 80 mV at low scan rates and increases with
increasing scan rate. This behavior points to quasi-reversibility of the viologen system in
ionic liquid electrolyte. Further voltammetric data are provided in the supporting information.
The calculated diffusion constant from these voltammetric data for 3•[PF6]2 in BMIM-OTf is D
= 3.41 x 10-8 cm2 s-1 and a heterogeneous rate constant for the 32+ ↔ 3+. redox couple of khet
= 6 x 10-4 cm s-1 has been estimated (see Supporting Information).
ꢇ
ACS Paragon Plus Environment

Page 9 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: Cyclic voltammograms recorded for 3 x 10-3 M of the thioacetate protected form of
12+ (3•[PF6]2) in BMIM-OTf for 11 different scan rates and Au(111) working electrodes. The
redox wave on the left corresponds to the viologen dication/radical cation redox couple,
while the redox wave centered at 0 V arises from ferrocene added to the electrolyte as
internal standard. The CVs were recorded at 0.05 V/s, 0.1 V/s and then in 0.1 V intervals to
1 V/s.
ꢈ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3 Left: conductance histograms recorded for molecular bridges formed from 12+ in
ionic liquid electrolyte as a function of the marked electrochemical potential (versus Fc/Fc+ at
0.0 V). Right: The single-molecule conductance of 12+ versus electrochemical potential.
Figure 3 shows conductance histograms for molecular junctions formed from 12+ recorded in
BMIM-OTf using an STM and the I(s) technique (see Experimental). These conductance
histograms were recorded for the marked electrochemical potentials (versus Fc/Fc+) and for
a constant bias voltage of 0.6 V. The peak of the conductance histograms shows a marked
dependence on the electrochemical potential, as plotted in Figure 3 (right). This plot shows
data recorded for 13 separate conductance histograms for 13 different electrochemical
potentials. We have also recorded the cyclic voltammetric response (Figure 4a),
conductance histograms (Figure 4b) and plotted electrochemically gated single-molecule
conductance data (Figure 4c) for 22+, which features the trimethylsilylethynyl anchoring
groups, in the same ionic liquid BMIM-OTf. 68 69 Similar to 12+, compound 22+ also shows a
single-molecule conductance maximum centered close to the equilibrium redox potential.
ꢀꢉ
ACS Paragon Plus Environment

Page 11 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4 (A): Cyclic voltammograms recorded for the first reduction of 2•[BF4]2 in BMIM-OTf.
The CVs were run at 0.1 V s-1 intervals between 0.1 and 1.0 V s-1. (B) Conductance
histograms recorded for molecular bridges formed from 22+ in BMIM-OTf electrolyte as a
function of the marked electrochemical potential (versus Fc/Fc+ at 0.0 V). (C): Single
molecule conductance of 22+ versus electrode potential in BMIM-OTf.
As indicated in the Introduction and described in greater detail below, in aqueous electrolyte
1
2+ exhibits a gradual sigmoidal increase in the molecular conductance with applied gate
potential over a wide electrochemical potential range of ~1 V. This strongly contrasts the
bell-shaped conductance response of both 12+ and 22+ in electrochemically gated systems in
ionic liquid electrolytes. In order to examine whether the electrolytic medium gives rise to
differences in the electronic structure or conformation of the redox-active viologen cores
upon reduction, which would be important considerations in rationalizing different gating
behavior in the two media, we have studied the reduction of 2•[BF4]2 by UV/visible and IR
spectroelectrochemistry in ionic liquid electrolytes. UV/visible spectroscopy is a very
ꢀꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 40
1
2
3
4
5
6
7
8
9
sensitive probe of the electronic and structural properties of the viologen radical cation, and
UV/visible spectroelectrochemical studies of the viologen redox system in aqueous solution
are very well described in the literature.70 On the other hand we are not aware of any reports
of UV/visible and IR spectroelectrochemistry in ionic liquid electrolytes for viologen
derivatives, or indeed any other electrochemical system. Details of the thin-layer
spectroelectrochemical setup with ionic liquids and the stepwise reduction of 2•[BF4]2 to the
corresponding radical cation and neutral viologen are given in the Experimental section and
Supporting Information. UV/visible spectra recorded for 2•[BF4]2 in dry BMIM-OTf during the
reduction of the dication (colorless) to the intensely colored radical cation are presented in
Figure 5a, showing that reduction is accompanied by the growth of intense absorption
envelopes at ~400 nm and ~610 nm. Due to insufficient solubility of 2•[BF4]2 and 3•[PF6]2 in
aqueous media, and for the purposes of comparison, the reduction of methyl viologen
(11•[Cl]2, Chart 2) has been investigated by UV/visible spectroelectrochemistry in both
aqueous and ionic liquid media (Figure 5B and supporting information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chart 2. The model methyl viologen dichloride, 11•[Cl]2
The comparison of the viologen core reduction shown in Figures 5a (2•[BF4]2 in BMIM-OTf)
and Figure 5B (11•[Cl]2 in water) shows the growth of bands in very similar positions in the
two different media. The spectroelectrochemically generated UV-Vis spectrum of the one-
electron reduced species 11+ in water (Figure 5B) agree well with literature spectra of the
methyl viologen radical cation,70 and is also similar to the UV-Vis spectrum of a reduced
viologen appended to a triosmium cluster in MeCN.71 The electronic absorption properties of
the V2+/V+. redox system therefore appear to be largely unaffected by the surrounding
medium, with the only significance difference being the appearance of dimerization in the
ꢀꢁ
ACS Paragon Plus Environment

Page 13 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
aqueous solution (see Figure 5 caption). This is further supported by UV/vis
spectroelectrochemistry of 11•[Cl]2 in BMIM-OTf which showed that the spectra of 2+ and 11+
were near identical when recorded in the same medium (see Supporting Information).
Further monitoring of the reduction of 2•[BF4]2 in BMIM-OTf by IR spectroscopy showed
formation of a strong band at 1634 cm–1 (Figure 6) in good agreement with the ν(C=C) band
reported72 for a viologen radical cation. The intensity of the ν(C≡C) band at 2160 cm–1 in 22+
also increases slightly as the radical cation 2+ is formed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Reversible UV-Vis spectral changes accompanying the reduction of (A) ca 2 mM
2•[BF4]2 in BMIM-OTf to its stable radical cation within an optically transparent thin layer
electrochemical (OTTLE) cell (l ~ 0.2 mm) at T = 293 K. (B) Reversible UV-Vis spectral
ꢀꢂ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 40
1
2
3
4
5
6
7
8
9
changes accompanying the reduction of 2 mM 11•[Cl]2 in D2O/Bu4NBr to a radical cation and
its subsequent dimerization within an OTTLE cell at T = 293 K. As the doublet around ~380
nm grows there is an inversion in intensities. This and the growing absorption at ~870 nm
are ascribed to dimerization of the radical cation in the aqueous electrolyte. These features
are not observed in BMIM-OTf, see Figure S17. (The colors are to help distinguish the
curves recorded during the reduction).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Reversible IR spectral changes accompanying the reduction of 10 mM 2•[BF4]2 in
BMIM-OTf to its stable radical cation within an OTTLE cell at T = 293 K.
■ DISCUSSION
A gradual sigmoidal increase in the molecular conductance has been observed previously in
studies of electrochemically gated conductance of 12+ in aqueous electrolytes, and attributed
to a ‘soft’ electrochemical gating of the conductance response.10-13 In contrast, the single-
ꢀꢄ
ACS Paragon Plus Environment

Page 15 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
molecule conductance data of 12+ and 22+ in ionic liquid both show a bell-shaped
conductance versus potential form, with a clear maximum around the formal electrode
potential of the V2+/V+. couple.10-13 This bell-shaped conductance versus potential form fits
well to the two-step hopping model of Kuznetsov and Ulstrup with partial vibrational
relaxation. The essence of the two-step process is that at, or close to, the equilibrium
potential, both the oxidized and reduced forms of the molecular bridge contribute
comparably to the current flow across the molecular junction. This model has been widely
applied to charge transport across redox active molecular junctions and has also been used
to justify enhanced current flow with redox active molecules in the tip-to-sample gap of an
electrochemical STM.12,14,31,40-44,47,58,59,73,74 In this model charge is transported (“hops”)
through either the reduced or oxidized state of the molecule. In the first step the redox center
has to reorganize through environmental and internal fluctuations so that the relevant
molecular orbitals approach the Fermi level of one of the electrodes. Charge is transferred
onto the redox center through a Frank Condon type transition. This is followed, in the
adiabatic limit, by partial vibrational relaxation and subsequent charge transfer to the other
electrode. In this partially relaxed state many electrons or holes can transfer across
contributing to the enhancement in the junction current seen close to the reversible potential.
By contrast, in the weak-coupling (diabatic) limit complete vibrational relaxation occurs
before the electron/hole transfers to the second electrode and current enhancement is
minimal. A schematic diagram illustrating the 2-step adiabatic Kuznetsov- Ulstrup used here
is shown in Scheme 2. This illustrates the electrochemical nature of this process with step
(1) representing the electron transfer from the left metal contact onto the redox group
(reduction), step (2) representing the molecule and environment partially relaxing in the
adiabatic limit and step (3) representing the electron transfer from the redox group
(oxidation) onto the right contact.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀꢅ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2
This illustrates the 2-step adiabatic Kuznetsov- Ulstrup model used here,
showing the two gold electrodes (“left” and “right”) and the redox states of the molecular
bridge in between. This proceeds with electron transfer from the left metal contact onto the
redox group following pre-organization at and around the redox center The rate of electron
transfer is represented by the electrochemical rate constant k1 (step 1). The now reduced
redox group and surrounding environment partially relax (step 2). This is followed by electron
transfer from the redox group (re-oxidation) onto the right contact with the electrochemical
rate constant k2 (step 3), resulting in a cascade of electron transfer between the electrodes
through the partially relaxed state of the redox bridge.
In contrast to these two-step hopping mechanisms, resonant tunneling corresponds to the
transferring charge populating the HOMO or LUMO on the bridge for such short time periods
that no vibrational relaxation in the intermediate state occurs. Since both of these levels are
displaced from the Fermi levels by the reorganization energy, this means that for resonant
tunneling the current maximum is shifted by the reorganization (free) energy from the
equilibrium potential. Since reorganization energies in such systems typically exceed a few
hundred meV, and our ionic liquid data shows the maximum molecular conductance close to
the V2+/V+. reversible potential, we conclude that the data are best represented by a two-step
ꢀꢆ
ACS Paragon Plus Environment

Page 17 of 40
Journal of the American Chemical Society
1
2
hopping model. Furthermore, the considerable current enhancement points to the strong
coupling (adiabatic) limit being more applicable. For these reasons we use below the
adiabatic limit of the two-step model to describe the conductance versus electrochemical
potential data. This consequently gives a maximum in the current versus electrochemical
potential relation, with the maximum close to the equilibrium potential.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The absence of a maximum for molecular conductance versus electrochemical potential in
the aqueous data previously reported would indicate that resonance of the gold Fermi levels
with the redox center is not attained for the single molecule conductance configuration in the
aqueous electrolyte.12 In this case, it has been previously argued that electron transfer in
aqueous media is instead by superexchange through a significantly off-resonant and
consequently empty redox level of the viologen (V2+) moiety. The off-resonant oxidized state
(V2+) was thus taken to dominate the single-molecule conductance behavior of the viologen
system in aqueous electrolytes. The gradual increase in conductance of the viologen system
in aqueous electrolyte was rationalized with this superexchange model coupled with “soft-
gating” to account for the conductance rise as the electrochemical potential is made more
negative.12 The soft-gating was attributed to thermally accessible configurational fluctuations
of the molecule and its environment which bring the molecular bridge into non-equilibrium
configurations which promote more facile electron tunneling across the molecular junction.12
Leary et al. suggested that twisting about the central C-C bond of the V2+ moiety could be an
important low-frequency mode in this respect.15
In our present study we offer another explanation based on our present observations that on
changing the aqueous electrolyte to an ionic liquid electrolyte, the behavior changes from
soft gating to hard gating. Our model below attributes the change in behavior to differences
in the electrolyte gating properties of the ionic liquid compared to the aqueous electrolytes.
This view is supported by the UV/vis spectroscopic data presented above which shows that
the electronic absorption properties of viologen redox systems appear to be largely
unaffected by the surrounding medium.
ꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 40
1
2
To model the data we take the expression for the two-step hopping model as given by
Kuznetsov and Ulstrup73:
3
4
5
6
7
8
9
ꢁꢔꢕ
ꢔ
ꢖꢗꢘꢙ
ꢏ
ꢑꢒꢓꢉ
ꢐ
ꢚꢛꢜ
ꢊꢋ
ꢀꢁꢂꢃ ꢄ ꢀꢅꢆꢇꢈꢉ
ꢏ
ꢐ
Equation (1)
ꢡ
ꢧ
ꢠ ꢢꢣꢤꢥꢦ ꢨ
ꢥꢠꢩꢪ
ꢌꢍꢎ
ꢖꢗꢘꢙ
ꢫꢬꢭ
ꢝꢞꢟꢃꢉ
ꢐ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Here
charge transport following the two step adiabatic model. In this equation
reorganization energy (both inner and outer-sphere), the Boltzmann constant,
voltage, the temperature, the charge on an electron. The modeling parameters
represent the fraction of the electrochemical potential and the fraction of the bias voltage
dropped at the redox site, respectively. The parameter is referred to in the introduction as
the “gating parameter” and it is a focus here for justifying the differences between the data
j
enh is the enhanced current following across the molecular junction as a result of
is the total
bias the bias
and
λ
k
V
T
e
ξ
γ
ξ
recorded in aqueous and ionic liquid electrolytes. An expression for
elsewhere73:
j0 can be found
ꢀꢅ ꢮ ꢯꢰꢱꢁꢲꢲꢳꢴꢵ
Equation (2)
With
ꢰ ꢄ ꢯꢶꢷꢸꢹꢟ ꢉꢻꢼ ꢾ ꢿ ꢼ ꢾ ꢐꣁꢺ Equation (3)
ꢺ
ꢺ
ꢽ
ꢽ
ꣀ ꣀ
Where the
of electronic states in the metal electrodes near the Fermi level. The subscripts
to the left and right electrodes, respectively. eff is the effective nuclear vibrational
κ
terms refer to the electron transmission coefficient and the
ρ
terms the density
L
and refer
R
ω
frequencies. From these terms a numerical form can be obtained for
information) which is used in the modeling.
jenh (see supporting
Figure 7a shows modeling of the conductance versus overpotential data recorded for 12+ in
ionic liquid with the numerical form of equation (1). A gate coupling parameter of = 1 is
used in this fitting with = 1.3 eV and = 0.5 (fitting errors are 0.2, 0.03 eV and 0.1 in ξ
ξ
λ
γ
±
±
±
,
λ
and
γ, respectively. See supporting information for plot of conductance versus potential
with error bars). Figure 7b shows the 12+ data for aqueous phosphate buffer electrolytes and
ꢀꢇ
ACS Paragon Plus Environment

Page 19 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
the modeling also using equation (1). Notice here that the electrochemical axis scale in
Figure 7b covers twice the voltage range of Figure 7a, which highlights by comparison how
broad the conductance rise is in aqueous electrolytes. This data from aqueous
electrolytes,10,12 has not been previously modeled, and the present study provides
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
opportunity to re-examine these existing data. The fitting parameters obtained here are ξ =
0.2, λ = 0.83 eV and
γ = 0.2 (with fitting errors of ± 0.15, ± 0.03 eV and ± 0.1, respectively).
As discussed elsewhere,14 a correlation is expected between the parameters
ξ
and γ in the
narrow tunneling gap in which the Debye length is comparable to the tunneling gap width.
For instance, at high ionic strength a high value for
ξ (e.g. 1) might be expected to correlate
with a low value for
γ
(e.g. 0.1).14 This is clearly not the case here and this discrepancy14
may be associated with the large size of the ionic liquid ions and the structuring and
“layering” behavior which is known to take place for ionic liquid double layers.61
Figure 7
Single molecule conductance data (blue circles) and fitting with equation 1
(solid blue lines), for 12+ in BMIM-OTf (a) and aqueous electrolytes (b). Experimental data in
(b) was recorded in 0.1 M phosphate buffer solution and is taken from refs. 10,12
The single molecule conductance versus electrochemical potential data for 22+ has also
been modeled (see Supporting Information for data fitting). A design feature of 22+ is the
incorporation of TMSE (TMSE, -C≡C-SiMe3) anchoring groups. TMSE groups for contacting
to gold have been used in a number of previous molecular electronics studies of either
molecular films or single molecules.68,69,75 The TMSE groups have a relatively large footprint
ꢀꢈ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 40
1
2
3
4
5
6
7
8
9
and this, coupled with the nonlinear geometry of 22+, is used here to hinder π-stacking and
prevent formation of multi-molecular junctions in the gap. In particular, it is recognized in the
literature that viologen dications and their radical cations have a propensity to stack in
mixed-valence viologen salts.76 Reassuringly the electrochemical switching data for 22+
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(Figure 4b) is similar to that for 12+. The fitting parameters obtained from the 22+ data are
ξ =
1, λ = 1.3 eV and
γ = 0.5 (with fitting errors of ± 0.35, ± 0.1 eV and ± 0.1, respectively).
These are comparable to the fitting parameters for 12+ in ionic liquid, showing again the
highly effective gating in ionic liquid with
ξ = 1.
■ CONCLUSIONS
Single-molecule electrochemical gating of viologen molecular bridges, studied here for the
first time in ionic liquids, is very different from the behavior previously observed for aqueous
electrolytes. In BMIM-OTf strong electrolyte gate coupling is observed with a calculated gate
coupling efficiency of 100 %. Gate coupling efficiency for the viologen system in aqueous
electrolytes is considerably lower at 20 %. This demonstrates the effectiveness of ionic
liquids for single molecule electrolyte gating. Since the gating coupling is controlled by
electrostatic interactions between the double layer gate and the redox bridge we note that its
efficiency might be expected to depend on both these components, and therefore vary
between redox systems. However, we note that a similar trend can be seen in the data for a
redox active pyrrolo-tetrathiafulvalene bridge, 14 with a gate coupling of ξ=0.8-1 in ionic liquid
compared to a lower value (ξ=0.5) in aqueous solutions.
UV/vis spectra show that the electronic absorption properties of viologen redox systems
appear to be largely unaffected by the surrounding medium, supporting the notion that the
differences observed between ionic liquid and aqueous environments arises from the
electrolyte screening rather than electronic structure differences of the viologen core in these
different media. Both the ionic liquid and aqueous electrolyte single molecule data for the
viologen systems could be fitted with the standard two-step hopping model of Kuznetsov and
ꢁꢉ
ACS Paragon Plus Environment

Page 21 of 40
Journal of the American Chemical Society
1
2
Ulstrup with partial vibrational relaxation. The surprising high values of reorganization energy
3
4
5
and uncorrelated bias voltage (γ) and electrochemical (ξ) potential drop parameters shows
6
7
8
9
that new models will be eventually required to deal with the complexities of the molecular
level structuring and screening properties of ionic liquid electrolytes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
■ EXPERIMENTAL METHODS
N,N’-Di-(4-(Trimethylsilylethynyl)benzyl)-4,4’-bipyridinium bis(tetrafluoroborate)
(2•[BF4]2)
A mixture of 4,4ʹ-bipyridine (100 mg, 640 µmol), compound 10 (428 mg, 1.60 mmol),
acetonitrile (30 mL) and water (3 mL) was heated at reflux for 20 h and then allowed to cool
to ambient temperature. After cooling, the creamy, powdery precipitate was isolated by
filtration and the filtrate discarded. The collected precipitate (350 mg) was mixed with
ammonium tetrafluoroborate (420 mg, 4.01 mmol) in absolute ethanol (53 mL) and water
(11 mL), and the mixture heated to reflux until a transparent, pale yellow solution was
obtained, and then allowed to cool to ambient temperature. Cooling resulted in crystallisation
of a white solid which was isolated by filtration and dried on the filter to afford 2•[BF4]2 in 82%
yield as a white crystalline solid (260 mg, 369
(v/v)): = 0.18 (s, 18 H, Si(C 3)3), 5.81 (s, 4 H, C
2 H, -3/ -5, C6H4), 8.40 (m, 2 H,
(224.5 MHz, CD3CN/D2O 2:1 (v/v)):
= –0.2 (Si( H3)3), 65.1 (C 2), 97.2 (C≡
-5, C6H4), 133.6 (
µ
mol). 1H-NMR (699.7 MHz, CD3CN/D2O 2:1
δ
H
H2), 7.44 (m, 2 H,
H-3/
H-5, C6H4), 7.49 (m,
H
H
H
-3/
H-5, C5H4N), 9.01 (m, 2 H, H-2/H
-6, C5H4N). 11B-NMR
δ
= –1.4. 13C-NMR (175.9 MHz, CD3CN/D2O 2:1 (v/v)):
δ
C
H
C
-Si), 104.9 (
C≡C-Si), 125.2 (
C-4, C6H4), 128.4 (
C-
3/
2/
C
-5, C5H4N), 130.5 (
-6, C5H4N), 151.2 (
C
-3/
C
C
-2/ -6, C6H4), 134.1 (
C
C-1, C6H4), 146.6 (C-
C
C
-4 , C5H4N). 19F-NMR (564.3 MHz, CD3CN/D2O 2:1 (v/v)):
4). 29Si-NMR (139.0 MHz, CD3CN/D2O 2:1
= −17.1. ESI-MS: m/z (%) 617 (17) [M – BF4]+, 343 (80), 187 (100). Anal. calcd for
δ = –
151.24 (m, 0.8 B, 11 4), –151.19 (m, 0.2 B, 10
BF BF
(v/v)):
δ
C34H38B2F8N2Si2 C: 57.97; H: 5.44; N: 3.98 found C: 57.83; H: 5.46; N: 4.02.
4-(Trimethylsilylethynyl)benzyl alcohol (9).
ꢁꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 40
1
2
Ethynyltrimethylsilane (923 mg, 9.40 mmol), 4-iodobenzyl alcohol (2.00 g, 8.55 mmol)
,
3
4
5
PdCl2(PPh3)2 (60.0 mg, 85.5 µmol) and copper(I) iodide (16.0 mg, 85.5 µmol) were dissolved
6
7
8
9
in triethylamine (50 mL) at 20 °C and the reaction mixture was stirred at this temperature for
16 h. The solvent was then removed under reduced pressure and the residue was purified
by flash column chromatography (silica gel; eluent, hexanes/ethyl acetate (80:20 (v/v))) to
give a yellow oil. This oil was crystallised from hexanes (5 mL, cooling at –20 °C) to afford 3
in 67% yield (1.17 g, 8.55 mmol) as a colorless crystalline solid. 1H-NMR (699.7 MHz
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CDCl3):
δ
= 0.25 (s, 9 H, Si(C
H
3)3), 4.69 (s, 2 H, C
H
2), 7.30 (d, 2 H, 3JH-H = 8.3 Hz,
H
-3/
= 0.1
-2/ -6,
= −17.8.
H
-5,
C6H4), 7.46 (d, 2 H, 3JH-H = 8.3 Hz,
H-2/
H
-6, C6H4). 13C-NMR (175.9 MHz, CDCl3):
δ
(Si(
CH3)3), 65.1 (
CH2), 94.4 (C≡
C-Si), 105.0 (
C≡C-Si), 122.5 (
C-1, C6H4), 126.8 (
C
C
C6H4), 123.3 (
C-3/
C-5, C6H4), 141.3 (
C
-4, C6H4). 29Si-NMR (139.0 MHz, CDCl3):
δ
EI-MS: m/z (%) 204 (22) [M+], 189 (100) [M+ − CH3]. 1H and 13C NMR data were consistent
with those reported in the literature.77
4-(Trimethylsilylethynyl)benzyl bromide (10)
Bromine (1.10 g, 6.87 mmol of Br2) was added dropwise at 0 °C to a stirred solution of
triphenylphosphine (1.89 g, 7.15 mmol) in dichloromethane (20 mL). On completion of this
addition, a solution of compound 9 (1.17 g, 5.72 mmol) and imidazole (467 mg, 6.87 mmol)
in dichloromethane (7 mL) was added to the solution of bromine triphenylphosphine complex
via cannula at 0 °C. Subsequently, the reaction mixture was allowed to warm to ambient
temperature, stirred for 3 h at this temperature and concentrated in vacuo. Diethyl ether
(30 mL) was added to the residue and the flask was placed into an ultrasound bath until a
homogeneous beige precipitate in a pale yellow solution was obtained. The precipitate was
removed by filtration and discarded. The filtrate was concentrated under reduced pressure,
hexanes (30 mL) were added to the residue and the flask was once again placed into an
ultrasound bath until a homogeneous white precipitate in a pale yellow solution was
obtained. The precipitate was removed by filtration and the filtrate concentrated to give a
pale yellow oil; this crystallised on standing to afford 10 in 90% yield (1.38 g, 5.17 mmol) as
ꢁꢁ
ACS Paragon Plus Environment

Page 23 of 40
Journal of the American Chemical Society
1
2
3
a yellow tinged crystalline solid. 1H-NMR (699.7 MHz CDCl3):
δ
= 0.25 (s, 9 H, Si(CH3)3),
4
5
6
7
4.46 (s, 2 H, C
H
2), 7.32 (d, 2 H, 3JH-H = 8.3 Hz,
H
-3/
= 0.1 (Si(
H
-5, C6H4), 7.43 (d, 2 H, 3JH-H = 8.3 Hz,
H-2/
H
-6, C6H4). 13C-NMR (175.9 MHz, CDCl3):
δ
CH3)3), 33.1 (
CH2), 95.4 (C≡C-Si),
8
9
104.6 (
C≡C-Si), 123.4 (
C-1, C6H4), 129.1 (
C-2/
C-6, C6H4), 132.5 (
C-3/
C-5, C6H4), 138.1 (
C-4,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C6H4). — 29Si-NMR (139.0 MHz, CDCl3):
δ
= −17.6. EI-MS: m/z (%) 268 (25) [M+], 253 (73)
[M+ − CH3], 187 (100) [M+ − Br]. 1H and 13C NMR data were consistent with those reported in
the literature.78
Cyclic Voltammetry
1-Butyl-3-methylimidazolium triflate, BMIM-OTf (IoLiLyt, 99%), was dried for 24 hours under
vacuum prior to use (120 ºC). Cyclic voltammetry of the thioacetate protected form of 12+
(3•[PF6]2) was recorded for Au(111) substrate. This gold electrode was immersed for 24 h in
a solution of 3•[PF6]2 (7.5 mM in methanol) then it was cleaned with methanol and Milli-Q
water (Millipore), and dried in a stream of N2 gas. Pt wires were used as counter and quasi
reference electrodes. Before measurements a drop of diluted ferrocene was added to use as
internal reference. Measurements were carried out in dried BMIM-OTf and under N2 flow.
The concentration of 3•[PF6]2 in BMIM-OTf was 3x10-3 M.
Cyclic voltammetry of 2•[BF4]2 in BMIM-OTf was recorded for a 7.5 mM solution of the
analyte in BMIM-OTf. This solution was prepared by first dissolving 2•[BF4]2 in methanol, and
then adding this solution to the ionic liquid. The methanol was removed by heating the
solution at ~100°C for 1.5 h with a nitrogen sparge. The working electrode was a Au(111)
bead, with a hanging meniscus, while the counter and reference electrodes were 0.5 mm Pt
wire. All potentials were determined against an internal Fc/Fc+ reference added after the
experiment.
Single-Molecule Conductance Measurements
ꢁꢂ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 40
1
2
3
The STM based
I
(s
) method as described in references 10,79 has been used to obtain single-
4
5
6
7
8
molecule conductance values as a function of electrode potential. In brief, an STM tip is
approached close to the gold surface so that the target molecule can bind between the gold
9
tip and the substrate. The tip is then rapidly retracted and a current (I) versus distance (s)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
curve recorded. Hundreds of such junction making and breaking curves are analyzed
statistically in histograms to yield the conductance. To avoid ambiguity caused by the
inclusion of retraction traces where no junction is formed, we constructed conductance
histograms from retraction traces displaying plateaus and steps, which are synonymous with
molecular junction formation. The STM- ) measurements were performed with flame-
I(s
annealed Au substrate. For the 12+ measurements under electrochemical potential control
the substrate was immersed for 5 minutes in a solution of 3•[PF6]2 (7.5 mM in methanol),
then it was cleaned with methanol and Milli-Q water (Millipore) and dried in a stream of N2. It
has been previously shown that adsorbed layers of thiols can be formed by direct adsorption
from the corresponding thioacetate protected molecules in ionic liquid solutions.14
Measurements were carried out in N2 atmosphere using the described substrate and in a 1.5
mM solution of 3•[PF6]2 in the ionic liquid (BMIM-OTf). Pt wires were used as counter and
reference electrodes. The STM tip was an etched gold wire. Set point parameters were I0
=20 nA and Ut = 0.6 V.
Spectroelectrochemistry
UV-Vis and IR spectroelectrochemical experiments were carried out using an optically
transparent thin layer spectroelectrochemical (OTTLE) cell 80 equipped with CaF2 optical
windows, Pt minigrid working and auxiliary electrodes, and an Ag wire pseudo-reference
electrode. UV-Vis spectroelectrochemistry was recorded with a Scinco S3100 diode array
spectrophotometer. IR spectroelectrochemical data were recorded using a Bruker Vertex
70v FT-IR spectrometer equipped with a DLaTGS detector working at the resolution of 1 cm–
1. The spectroelectrochemical measurements were monitored using thin-layer cyclic
ꢁꢄ
ACS Paragon Plus Environment

Page 25 of 40
Journal of the American Chemical Society
1
2
voltammetry performed with an EmStat-3 potentiostat (PalmSens BV, The Netherlands)
operated with the PSTrace v.4.2 software.
3
4
5
6
7
8
9
Solutions of 2•[BF4]2 in BMIM-OTf contained 2 mM (UV-Vis spectroelectrochemistry) or 10
mM (IR spectroelectrochemistry) 2•[BF4]2 and were prepared as follows. BMIM-OTf (IoLiLyt,
99%) was heated in a Schlenk vessel under stirring for >18 h at 120 ºC under high vacuum
2•[BF4]2 was partly dissolved in a few drops of ethanol and hot BMIM-OTf was added with a
vacuum tight Hamilton syringe to give a homogenous solution. The resulting solution was
sparged with dry argon at 120 ºC for ca 2 h.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Supporting Information Available: General synthetic methods and synthetic details for 12+,
N
,
N
’-Di-(6-(thioacetyl)hexyl)-4,4’-bipyridinium bis(hexafluorophosphate) (3•[PF6]2),
(6-(thioacetyl)hexyl)-4,4’-bipyridinium dibromide (3•[Br]2) and -(6-bromohexyl)
ethanethioate. Electrochemistry of 2(BF4 2 in acetonitrile electrolytes and of 2(BF4)2 in ionic
N,N’-Di-
S
)
liquid, the two step model equations and 22+ conductance data, break-off distance data,
diffusion coefficients and heterogeneous rate constants, collected UV-Vis and IR spectra,
plots with error bars for 12+ and 22+ data and bias voltage data. This material is available free
of charge via the Internet at http://pubs.acs.org.
Additional supporting data files at DOI: http://datacat.liverpool.ac.uk/id/eprint/62
■ ACKNOWLEDGEMENTS
R.J.N. and S.J.H. thank EPSRC for funding (grant EP/H035184/1 and EP/K007785/1). S.M.
and P.C. are grateful for financial assistance from Ministerio de Economía y Competitividad
from Spain in the framework of projects CTQ2012-33198 and CTQ2013-50187-EXP and
support from DGA and Fondos Feder for funding Platon research group. P.J.L. thanks
EPSRC for funding and also gratefully acknowledges support from the Australian Research
Council (DP 140100855) and the award of a Future Fellowship (FT120100073). H.M.O is
grateful for financial assistance from Secretaría Nacional de Educación Superior, Ciencia,
ꢁꢅ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 40
1
2
Tecnología e Innovación from Ministerio de Educación (Ecuador). F.H. thanks EPSRC for
funding (EP/K00753X/1) and acknowledges the University of Reading for the support of the
Reading Spectroelectrochemistry center (University Project D14-015 Spectroelectrochemical
Cells).
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
REFERENCES
ꢊꢀꢋ
ꢌꢍꢎ ꢏꢐꢑ aꢒꢓꢐꢎꢃ {ꢔ Wꢔꢕ [ꢖꢓꢗꢐꢑꢒꢘ ꢃ tꢔ Wꢀꢁꢂꢃꢄꢅ ꢀ tꢆꢇꢈꢉꢊꢈꢋ/ꢀꢃꢌꢍꢃꢈꢍꢌ aꢄꢎꢎꢍꢂ ꢀꢁꢂꢁ
ꢃ ꢏꢏꢃ ꢀꢂꢂꢉꢉꢀꢔ
ꢊꢁꢋ
{ꢒꢎ"ꢃ Iꢔꢕ Yꢖ%ꢃ òꢔꢕ Wꢍꢎ"ꢃ òꢔ Iꢔꢕ Wꢐꢒꢎ"ꢃ Iꢔꢕ wꢐꢐꢏꢃ aꢔ !ꢔꢕ [ꢐꢐꢃ Çꢔ bꢄꢎꢁꢂꢍ ꢀꢁꢁꢃ ꢐꢑꢏꢃ ꢀꢉꢂꢈꢔ
ꢃ
ꢊꢂꢋ
ꢀꢁꢁꢂ
ꢊꢄꢋ
Yꢑ*"ꢐꢑꢃ aꢔꢕ .*ꢖꢘꢐꢓꢍꢍꢑꢃ aꢔ wꢔꢕ b*--.ꢍ*%ꢐꢑꢃ Çꢔꢕ {/ ꢒꢎꢐꢎ.ꢐꢑ"ꢐꢑꢃ /ꢔꢕ Cꢒꢑꢑꢒꢃ [ꢔ !ꢒꢒꢅꢓ tꢆꢇꢈꢓ [ꢍꢎꢎꢓ
ꢃ ꢔꢕꢃ ꢀꢁꢈꢀꢔ
wꢒ-ꢐꢎ.ꢓꢍꢘꢘꢃ {ꢔꢕ òꢍꢖ- ꢃ òꢔꢕ tꢍꢑ2ꢃ Wꢔꢕ Dꢒꢑꢐꢃ Wꢔꢕ {ꢍ4ꢒꢎꢒꢌꢍꢃ ëꢔꢕ a/9*ꢐꢎꢃ tꢔ [ꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢁꢀꢃ ꢏꢃ
ꢇꢆꢈꢔ
ꢊꢅꢋ
ꢊꢆꢋ
aꢄꢎꢍꢂꢓ ꢀꢁꢂꢁ
ꢊ ꢋ
aꢐ*ꢓꢐꢎ2ꢍ%7ꢃ 9ꢔ !ꢔ Wꢓ tꢆꢇꢈꢓ /ꢆꢍ ꢓ . ꢂꢃꢃꢃ
òꢐꢃ Wꢔ Çꢔꢕ Lꢎꢒ*ꢐꢃ {ꢔꢕ Yꢒ.ꢍ9ꢍ- ꢖꢃ Yꢔꢕ Yꢍ-ꢍ ꢍꢑꢍꢃ òꢔꢕ ò*ꢍꢎꢃ Iꢔ Çꢔꢕ { ꢖ%ꢒꢘꢍꢎꢖꢃ Iꢔꢕ L:ꢍ-ꢍꢃ òꢔ bꢄꢎꢓ
ꢃ ꢀꢁꢅꢔ
ò*ꢍꢎꢃ Iꢔꢕ { ꢖ%ꢒꢘꢍꢎꢖꢃ Iꢔꢕ Ç-*2ꢍ4ꢍ2ꢖꢃ !ꢔꢕ h ꢘꢒ%ꢒꢃ !ꢔꢕ Yꢍ:ꢍ-ꢍ2ꢖꢃ aꢔꢕ L:ꢍ-ꢍꢃ òꢔ !ꢌ&ꢓ Cꢁꢃꢊꢎꢓ
aꢄꢎꢍꢂꢓ ꢀꢁꢁꢃ "%ꢃ ꢀꢉꢄꢆꢔ
í ꢖꢘꢐꢃ Iꢔ {ꢔꢕ Yꢖꢘꢘꢓꢐ-ꢐꢎꢃ Dꢔ tꢔꢕ íꢑꢖ" ꢘꢒꢎꢃ aꢔ {ꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢂꢃꢄꢅ
bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔ ){ꢉꢃ*ꢅꢍ aꢀꢅꢍꢊꢁꢅꢄꢂ 9ꢅꢍꢊꢎꢂꢀꢊꢆꢍ ꢉꢈꢎꢂꢇ ,ꢉꢎꢆꢉꢃ ꢄꢃ {Ça) ꢉꢃ
ꢃ "#$ꢃ ꢇꢂꢀꢔ
ꢃ %
ꢃ
ꢊꢇꢋ
ꢊꢈꢋ
ꢃ "#ꢑꢃ ꢅꢂ ꢅꢔ
)9ꢅꢍꢊꢎꢂꢀꢊꢄꢎꢄꢅꢇꢈꢉꢈ ꢋ Çꢆꢍꢀꢂꢍꢎꢉꢊꢄꢅ Cꢀꢁꢃꢌꢄꢎꢉꢀꢃꢈ ꢄꢃꢌ aꢀꢌꢍꢅ 9.ꢒꢍꢂꢉ ꢍꢃꢎꢈ)/ !ꢌ&ꢄꢃꢊꢍꢈ ꢉꢃ 9ꢅꢍꢊꢎꢂꢀꢊꢆꢍ ꢉꢊꢄꢅ
{ꢊꢉꢍꢃꢊꢍ ꢄꢃꢌ 9ꢃ*ꢉꢃꢍꢍꢂꢉꢃ*/ ëꢀꢅꢁ ꢍ óLë/ tꢁ3ꢅꢉꢈꢆꢍꢂꢈ íꢉꢅꢍꢇꢋë/I/ 9ꢌꢉꢎꢀꢂꢈ [ꢓ Yꢉ3ꢅꢍꢂ/ wꢓ/ꢓ !ꢅ8ꢉꢂꢍ/ 5ꢓaꢓ Yꢀꢅ3
ꢄꢃꢌ Wꢓ [ꢉꢒ8ꢀ,ꢈ8ꢉꢓ
ꢊꢀꢉꢋ
Wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢆ
ꢊꢀꢀꢋ [ꢖꢃ ùꢔ Iꢔꢕ tꢒ.ꢐꢓꢒꢌꢃ Lꢔꢕ Iꢍꢎꢃ .ꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔꢕ .ꢓꢍ-4/492ꢃ !ꢔꢕ aꢍ9ꢒꢑꢃ aꢔ bꢄꢃꢀꢎꢍꢊꢆꢃꢀꢅꢀ*ꢇ ꢀꢁꢁꢇ
"ꢕꢃ ꢉꢄꢄꢉꢀꢇꢔ
ꢊꢀꢁꢋ Iꢍꢖ--ꢃ íꢔꢕ !ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ ꢌꢍꢎ ùꢍꢓꢖꢎ"ꢐꢃ Iꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ .ꢐꢘ ꢐꢓꢓꢃ 5ꢔꢕ Iꢒꢐ.ꢐꢎꢑꢐꢖ/ ꢃ Iꢔꢕ {/ ꢖ??ꢑꢖꢎꢃ 5ꢔ
Wꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ Y*4ꢎꢐꢘ-ꢒꢌꢃ !ꢔ aꢔꢕ ù ꢍꢎ"ꢃ Wꢔꢕ / ꢖꢃ vꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ Wꢓ tꢆꢇꢈꢓ /ꢆꢍ ꢓ . ꢀꢁꢁꢇ """ꢃ ꢆ ꢉꢂꢔ
ꢊꢀꢂꢋ [ꢖꢃ ùꢔꢕ Iꢍꢎꢃ .ꢔꢕ aꢐ-4ꢍꢑꢒ-ꢃ Dꢔꢕ tꢒ.ꢐꢓꢒꢌꢃ Lꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔꢕ .ꢓꢍ-4/492ꢃ !ꢔꢕ aꢍ9ꢒꢑꢃ aꢔ Cꢄꢂꢄꢌꢄꢇ
5ꢉꢈꢊꢁꢈꢈꢓ ꢀꢁꢁ "$"ꢃ ꢀꢁꢀꢔ
ꢊꢀꢄꢋ Yꢍ9ꢃ bꢔ Wꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ Wꢐ77ꢐ-ꢐꢎꢃ Wꢔ hꢔꢕ [ꢐꢍꢑ9ꢃ 9ꢔꢕ [9/ꢒꢒ7-ꢃ Wꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔ Wꢓ ! ꢓ
/ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢂꢀ "$ꢐꢃ ꢀꢆꢇꢀ ꢔ
ꢊꢀꢅꢋ [ꢐꢍꢑ9ꢃ 9ꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ ꢌꢍꢎ ùꢍꢓꢖꢎ"ꢐꢃ Iꢔꢕ Iꢍꢖ--ꢃ íꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ b9"ꢍꢍꢑꢏꢃ {ꢔꢕ Wꢐ77ꢐ-ꢐꢎꢃ Wꢔ hꢔꢕ
Üꢓ-ꢘꢑ*7ꢃ Wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢄ "$#ꢃ ꢀꢁꢁꢉꢄꢔ
/ ꢐꢎꢃ Cꢔꢕ Iꢐꢃ Wꢔꢕ b*/2ꢒꢓꢓ-ꢃ /ꢔꢕ wꢒ.ꢐꢑꢘ-ꢃ Çꢔꢕ Yꢓꢍꢑꢐꢃ Wꢔ 9ꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢁꢈ
Iꢐꢃ Wꢔꢕ / ꢐꢎꢃ Cꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔꢕ b*/2ꢒꢓꢓ-ꢃ /ꢔ !ꢒꢒꢅꢓ tꢆꢇꢈꢓ [ꢍꢎꢎꢓ ꢀꢁꢁꢇ %#ꢃ ꢉ ꢁꢀꢀꢁꢔ
Iꢐꢃ Wꢔꢕ Cꢒꢑ4ꢍꢎꢖꢃ 9ꢔ {ꢔꢕ bꢍ"ꢍ ꢍꢑꢍꢃ [ꢔ !ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔ Wꢀꢁꢂꢃꢄꢅ ꢀ tꢆꢇꢈꢉꢊꢈꢋ/ꢀꢃꢌꢍꢃꢈꢍꢌ
ꢏ#ꢃ ꢂ ꢄꢀꢁꢉꢔ
ó*ꢃ .ꢔ vꢔ vꢔꢕ [ꢖꢃ óꢔ [ꢔ [ꢔꢕ óꢖꢍꢒꢃ óꢔ òꢔ òꢔꢕ {ꢍ2ꢍ"*/ ꢖꢃ Iꢔꢕ Çꢍꢒꢃ bꢔ Wꢔ Wꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢁꢈ
óꢖꢍꢒꢃ óꢔ òꢔꢕ bꢍ"ꢍ ꢍꢑꢍꢃ [ꢔ !ꢔꢕ wꢍ:ꢓꢐꢘꢘꢃ !ꢔ aꢔꢕ Çꢍꢒꢃ bꢔ Wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢈ "ꢏꢔꢃ ꢈꢁꢂꢅꢔ
Iꢐꢃ Wꢔꢕ C*ꢃ vꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔꢕ /ꢖ-4ꢐ2ꢃ Wꢔ íꢔꢕ Çꢒ*ꢑꢃ Wꢔ aꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁ "ꢏꢕꢃ ꢀꢄꢇꢁꢇꢔ
Yꢒꢓꢖꢌꢒ-2ꢍꢃ ëꢔꢕ aꢒꢑꢐꢎꢒCDꢍꢑ/ꢖꢍꢃ tꢔꢕ Yꢍꢓꢖ"ꢖꢎꢐꢏꢖꢃ ëꢔꢕ Iꢒꢎ"ꢃ íꢔꢕ aꢍ9ꢒꢑꢃ aꢔꢕ íꢐꢖ.ꢐꢓꢃ bꢔꢕ
""#ꢃ ꢉꢈꢔ
óꢖꢍꢒꢃ óꢔ òꢔꢕ .ꢑ*ꢎꢐꢃ 5ꢔꢕ Iꢐꢃ Wꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔꢕ Dꢒꢑ%ꢍꢎꢃ /ꢔ .ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔ /ꢆꢍ ꢓ tꢆꢇꢈꢓ ꢀꢁꢁ
Iꢍꢖ--ꢃ íꢔꢕ ꢌꢍꢎ ùꢍꢓꢖꢎ"ꢐꢃ Iꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ .ꢐꢘ ꢐꢓꢓꢃ 5ꢔꢕ Iꢒ.ꢐꢎꢑꢐꢖ/ ꢃ Iꢔꢕ {/ ꢖ??ꢑꢖꢎꢃ 5ꢔ Wꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ
ꢃ
"ꢏ:ꢃ ꢀꢅꢁꢈꢄꢔ
ꢃ
ꢃ
ꢃ
ꢃ
ꢃ
ꢊꢀꢆꢋ
ꢊꢀ ꢋ
ꢊꢀꢇꢋ
ꢃ :ꢃ ꢅꢉꢂꢔ
ꢃ
aꢄꢎꢎꢍꢂ ꢀꢁꢁꢄ
ꢃ
ꢊꢀꢈꢋ
ꢊꢁꢉꢋ
ꢊꢁꢀꢋ
ꢊꢁꢁꢋ
ꢃ :ꢃ ꢀꢄꢈꢀꢔ
ꢃ
ꢃ
íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔ 9ꢅꢍꢊꢎꢂꢀꢊꢆꢉ ꢓ !ꢊꢎꢄ ꢀꢁꢂꢆ
ꢊꢁꢂꢋ
ꢃ
ꢃ
$ꢏꢑꢃ ꢀꢂꢇꢔ
ꢁꢆ
ACS Paragon Plus Environment

Page 27 of 40
Journal of the American Chemical Society
1
2
3
ꢊꢁꢄꢋ
5ꢍꢑ:ꢖ- ꢃ bꢔꢕ 5ꢖꢐ4Ctꢐꢑꢐ4ꢃ Lꢔꢕ D*ꢒꢃ {ꢔ òꢔꢕ Çꢍꢒꢃ bꢔ Wꢔꢕ Dꢒꢒꢏꢖꢎ"ꢃ Wꢔ Wꢔꢕ tꢍꢏꢏꢒꢎCwꢒ:ꢃ aꢔ bꢔ Wꢓ tꢆꢇꢈꢓ
4
5
6
7
8
9
/ꢆꢍ ꢓ / ꢀꢁꢂꢀ
ꢊꢁꢅꢋ 5ꢍꢑ:ꢖ- ꢃ bꢔꢕ 5ꢖꢐ4Ctꢐꢑꢐ4ꢃ Lꢔꢕ 5ꢍ {ꢖꢓꢌꢍꢃ tꢔꢕ Çꢍꢒꢃ bꢔ Wꢔꢕ Dꢒꢒꢏꢖꢎ"ꢃ Wꢔ Wꢔꢕ tꢍꢏꢏꢒꢎCwꢒ:ꢃ aꢔ bꢔ !ꢃ*ꢍ,ꢓ
/ꢆꢍ ꢓ Lꢃꢎꢓ 9ꢌꢓ ꢀꢁꢂꢀ :"ꢃ ꢂꢁꢉꢂꢔ
ꢊꢁꢆꢋ .ꢍ" ꢐꢑꢎꢐꢗꢍꢏꢃ aꢔꢕ ù ꢍꢒꢃ óꢔꢕ hꢑꢎ-ꢒꢃ Yꢔ .ꢔꢕ C*ꢐꢐ"ꢃ aꢔꢕ aꢒꢑꢐꢎꢒCDꢍꢑ/ꢖꢍꢃ tꢔꢕ w*ꢏꢎꢐꢌꢃ !ꢔ ëꢔꢕ
Yꢍꢓꢖ"ꢖꢎꢐꢏꢖꢃ ëꢔꢕ ëꢐ-4ꢘꢐꢑ"ꢒ%ꢃ {ꢔꢕ I*ꢍꢎ"ꢃ /ꢔꢕ Iꢒꢎ"ꢃ íꢔꢕ .ꢑꢒꢐ2%ꢍꢎꢎꢃ tꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔꢕ Ç 9"ꢐ-ꢐꢎꢃ Yꢔ {ꢔꢕ
.ꢑ9/ꢐꢃ aꢔ wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢂꢅ "$ꢑꢃ ꢀ ꢈꢁꢁꢔ
ó*ꢃ .ꢔ vꢔꢕ óꢖꢍꢒꢃ óꢔ òꢔꢕ òꢍꢎ"ꢃ óꢔ aꢔꢕ ùꢍꢎ"ꢃ [ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢈ
[ꢖꢃ óꢔꢕ Iꢖ ꢍꢘ ꢃ Wꢔꢕ / ꢐꢎꢃ Cꢔꢕ aꢍ-*ꢏꢍꢃ Çꢔꢕ ùꢍꢎ"ꢃ [ꢔꢕ Çꢍꢒꢃ bꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢇ
[ꢖꢃ óꢔ [ꢔꢕ ó*ꢃ .ꢔ vꢔꢕ óꢖꢍꢒꢃ óꢔ òꢔꢕ òꢍꢎ"ꢃ óꢔ aꢔꢕ ùꢍꢎ"ꢃ [ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔ Cꢄꢂꢄꢌꢄꢇ 5ꢉꢈꢊꢁꢈꢈꢓ ꢀꢁꢁ
[ꢖꢃ /ꢔꢕ {ꢘꢐ7ꢍꢎꢐꢎ2ꢒꢃ ëꢔꢕ [ꢖꢎꢃ aꢔCWꢔꢕ Iꢒꢎ"ꢃ íꢔꢕ í*ꢐꢑꢘ ꢎꢐꢑꢃ Cꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔ tꢆꢇꢈꢉꢊꢄ {ꢎꢄꢎꢁꢈ
ꢏ:#ꢃ ꢁꢄꢅꢇꢔ
[ꢖꢃ /ꢔꢕ aꢖ- / ꢐꢎ2ꢒꢃ !ꢔꢕ [ꢖꢃ ùꢔꢕ tꢒ.ꢐꢓꢒꢌꢃ Lꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔꢕ [ꢖꢃ óꢔ vꢔꢕ í*ꢐꢑꢘ ꢎꢐꢑꢃ Cꢔꢕ .ꢍ"ꢑꢐꢘ-ꢃ !ꢔꢕ
9ꢌꢐꢑ-ꢃ Cꢔ Wꢀꢁꢂꢃꢄꢅ ꢀ tꢆꢇꢈꢉꢊꢈꢋ/ꢀꢃꢌꢍꢃꢈꢍꢌ aꢄꢎꢎꢍꢂ ꢀꢁꢁꢄ ꢏ#ꢃ ꢂ ꢄꢀꢁꢁꢔ
ꢊꢂꢁꢋ ëꢖ-ꢒꢓ9CCꢖ- ꢐꢑꢃ Lꢔꢕ 5ꢍꢖꢐꢃ Yꢔꢕ Çꢐꢑꢍ4ꢒꢎꢒꢃ òꢔꢕ Iꢐꢑꢑꢐꢑꢒꢃ /ꢔꢕ C*ꢎ"ꢒꢃ Cꢔꢕ hꢘꢐꢑꢒꢃ [ꢔꢕ 5*ꢑꢍꢎꢘꢖꢎꢖꢃ 9ꢔꢕ {ꢖꢓ.ꢐꢑꢃ Wꢔ
ꢃ ""ꢑꢃ ꢁꢀꢉꢈꢂꢔ
ꢃ
ꢃ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢊꢁ ꢋ
ꢊꢁꢇꢋ
ꢊꢁꢈꢋ
ꢊꢂꢉꢋ
ꢃ
"ꢏꢔꢃ ꢁꢂꢇꢆꢔ
"ꢏ%ꢃ ꢀꢀꢅꢂꢅꢔ
"$"ꢃ ꢀꢀꢀꢔ
ꢃ
ꢃ
{ꢀꢅꢉꢌꢉ .ꢋ.ꢄꢈꢉꢊ {ꢀꢅꢉꢌ {ꢎꢄꢎꢍ tꢆꢇꢈꢉꢊꢈ ꢀꢁꢂꢆ
ꢊꢂꢀꢋ
ꢃ
ꢃ
Wꢔꢕ {ꢐꢑꢐꢎꢒꢃ [ꢔꢕ D*-ꢘꢃ 5ꢔꢕ aꢒꢒꢑꢐꢃ Çꢔ !ꢔꢕ aꢒꢒꢑꢐꢃ !ꢔ [ꢔꢕ [ꢖꢎꢏ-ꢍ9ꢃ {ꢔ aꢔ tꢂꢀꢊꢓ bꢄꢎꢓ !ꢊꢄꢌꢓ {ꢊꢉꢓ Üꢓ{ꢓ!ꢓ ꢀꢁꢁ
"#$ꢃ ꢇꢆꢇꢆꢔ
ꢃ
ꢊꢂꢂꢋ
{ꢔCóꢔꢕ .ꢒꢑ"*ꢐꢘꢃ 9ꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢂꢅ
ꢊꢂꢄꢋ [ꢖꢃ /ꢔꢕ aꢖ- / ꢐꢎ2ꢒꢃ !ꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔ Üꢃꢉ ꢀꢅꢍꢊꢁꢅꢄꢂ ꢄꢃꢌ {ꢁꢒꢂꢄ ꢀꢅꢍꢊꢁꢅꢄꢂ 9ꢅꢍꢊꢎꢂꢀꢃꢉꢊꢈ Lꢉ<
/ꢆꢍ ꢉꢈꢎꢂꢇ ꢄꢃꢌ tꢆꢇꢈꢉꢊꢈ aꢍꢍꢎ ꢄꢎ aꢍꢎꢄꢅꢋaꢀꢅꢍꢊꢁꢅꢍ Lꢃꢎꢍꢂ ꢄꢊꢍꢈ ꢀꢁꢂꢀ $"$ꢃ ꢀꢁꢀꢔ
ꢊꢂꢅꢋ /ꢍ7ꢒ44ꢖꢃ .ꢔꢕ / ꢐꢎꢃ vꢔꢕ 5ꢍꢑꢍꢎ/ꢐꢘꢃ tꢔꢕ Yꢒꢘꢖ*"ꢍꢃ aꢔꢕ .*44ꢐꢒꢃ aꢔꢕ bꢐꢍꢘꢒꢎꢃ Wꢔ .ꢔꢕ b*/2ꢒꢓꢓ-ꢃ /ꢔꢕ
ëꢐꢎ2ꢍꢘꢍꢑꢍ%ꢍꢎꢃ [ꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢂꢅ "ꢐꢃ ꢀꢄꢉꢉꢔ
ꢊꢂꢆꢋ .ꢑꢒꢒ2ꢐꢃ wꢔ Wꢔꢕ Wꢖꢎꢃ /ꢔꢕ {4*%-2ꢖꢃ 5ꢔ {ꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ aꢍꢒꢃ .ꢔCíꢔꢕ Ç 9"ꢐ-ꢐꢎꢃ Yꢔ {ꢔꢕ {/ :ꢍꢑ4ꢍ/ ꢐꢑꢃ
íꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢂꢈ ":ꢃ ꢁ ꢅꢔ
Çꢍꢒꢃ bꢔ Wꢔ tꢆꢇꢈꢓ wꢍ&ꢓ [ꢍꢎꢎꢓ ꢂꢃꢃ
[ꢖꢃ ùꢔꢕ [ꢖꢃ Iꢔꢕ / ꢐꢎꢃ {ꢔꢕ Cꢑꢒꢐ ꢓꢖ/ ꢃ Çꢔꢕ òꢖꢃ /ꢔꢕ {/ ꢒꢐꢎꢐꢎ.ꢐꢑ"ꢐꢑꢃ /ꢔꢕ /ꢍꢓꢍ%ꢐꢃ aꢔꢕ 5ꢐ/*ꢑꢘꢖꢎ-ꢃ {ꢔꢕ [ꢖ*ꢃ
ꢃ
"$ꢑꢃ ꢇꢇꢆ ꢔ
ꢃ
ꢃ
ꢃ
ꢊꢂ ꢋ
ꢊꢂꢇꢋ
ꢀꢁꢁꢅ
ꢊꢂꢈꢋ
ꢃ ꢔꢑꢃ ꢄꢉꢆꢆꢔ
íꢍ--ꢐꢓꢃ wꢔ !ꢔꢕ /ꢑꢐꢏꢒꢃ Dꢔ aꢔꢕ C*ꢖꢐꢑꢐꢑꢃ wꢔ wꢔꢕ Cꢐꢓꢏ ꢐꢖ%ꢃ 5ꢔ [ꢔꢕ Dꢒꢑ%ꢍꢎꢃ /ꢔ .ꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ
"ꢏꢑꢃ ꢁꢈꢅꢔ
ù ꢍꢎ"ꢃ Wꢔ 5ꢔꢕ / ꢖꢃ vꢔ Wꢔꢕ !ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ Y*4ꢎꢐꢘ-ꢒꢌꢃ !ꢔ aꢔꢕ Dꢑ*..ꢃ aꢔꢕ Iꢍꢎ-ꢐꢎꢃ !ꢔ Dꢔꢕ íꢍ/2ꢐꢑ.ꢍꢑꢘ ꢃ
ꢃ
Iꢔꢕ íꢐꢓꢖꢎꢏꢐꢑꢃ !ꢔ /ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ 9ꢅꢍꢊꢎꢂꢀꢊꢆꢉ ꢓ !ꢊꢎꢄ ꢀꢁꢁꢈ
ꢃ
:#ꢃ ꢂꢀꢄꢂꢔ
ꢊꢄꢉꢋ
ꢄꢂꢉꢔ
ꢊꢄꢀꢋ
ꢊꢄꢁꢋ
ꢊꢄꢂꢋ
!ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ D*/2ꢖꢍꢎꢃ !ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔꢕ ëꢒ-ꢃ Wꢔ Dꢔ L999 Çꢂꢄꢃꢈꢄꢊꢎꢉꢀꢃꢈ ꢀꢃ bꢄꢃꢀꢎꢍꢊꢆꢃꢀꢅꢀ*ꢇ ꢀꢁꢁꢈꢃ ꢐꢃ
!ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ D*/2ꢖꢍꢎꢃ !ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔꢕ ëꢒ-ꢃ Wꢔ Dꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢁꢈ
!ꢓꢐ--ꢍꢎꢏꢑꢖꢎꢖꢃ !ꢔꢕ {ꢍꢓꢐꢑꢎꢒꢃ aꢔꢕ Cꢑꢍ..ꢒꢎꢖꢃ {ꢔꢕ Cꢍ//ꢖꢃ tꢔ !ꢒꢒꢅꢓ tꢆꢇꢈꢓ [ꢍꢎꢎꢓ ꢀꢁꢁꢈ
!ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ aꢒꢘ Ctꢒ*ꢓ-ꢐꢎꢃ Yꢔꢕ / ꢑꢖ-ꢘꢐꢎ-ꢐꢎꢃ Wꢔ .ꢔꢕ D*/2ꢖꢍꢎꢃ !ꢔꢕ .ꢗꢒꢑꢎ ꢒꢓ%ꢃ Çꢔꢕ ëꢒ-ꢃ Wꢔ Dꢔꢕ
"$"ꢃ ꢁꢆꢅꢔ
!ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ D*/2ꢖꢍꢎꢃ !ꢔꢕ Y*4ꢎꢐꢘ-ꢒꢌꢃ !ꢔ aꢔꢕ ëꢒ-ꢃ Wꢔ Dꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁ ꢃ "ꢏꢕꢃ
ꢃ :ꢃ ꢀꢄꢅꢀꢔ
ꢃ
ꢕꢑꢃ ꢂꢔ
Üꢓ-ꢘꢑ*7ꢃ Wꢔ Cꢄꢂꢄꢌꢄꢇ 5ꢉꢈꢊꢁꢈꢈꢓ ꢀꢁꢁ
ꢊꢄꢄꢋ
ꢃ
ꢀ ꢀꢂꢁꢔ
ꢊꢄꢅꢋ
ꢊꢄꢆꢋ
ꢊꢄ ꢋ
ꢊꢄꢇꢋ
5ꢍꢌꢖ-ꢃ Wꢔ Wꢔꢕ tꢐꢘꢐꢑ-ꢃ .ꢔꢕ óꢖꢃ íꢔ Wꢀꢁꢂꢃꢄꢅ ꢀ tꢆꢇꢈꢉꢊꢈꢋ/ꢀꢃꢌꢍꢃꢈꢍꢌ aꢄꢎꢎꢍꢂ ꢀꢁꢁꢄ
tꢒ.ꢐꢓꢒꢌꢘꢃ Lꢔ ëꢔꢕ [ꢖꢃ ùꢔ Iꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢄ "$#ꢃ ꢀꢆꢉꢄꢅꢔ
w*ꢏꢎꢐꢌꢃ !ꢔ ëꢔꢕ tꢒ.ꢐꢓꢒꢌꢃ Lꢔ ëꢔꢕ íꢍꢎꢏꢓꢒ:-2ꢖꢃ Çꢔ Wꢓ 9ꢅꢍꢊꢎꢂꢀꢄꢃꢄꢅꢓ /ꢆꢍ ꢓ ꢀꢁꢂꢂ ꢑꢑ#ꢃ ꢂꢉꢁꢔ
/ ꢖꢃ vꢔ Wꢔꢕ ù ꢍꢎ"ꢃ Wꢔ 5ꢔꢕ Wꢐꢎ-ꢐꢎꢃ tꢔ {ꢔꢕ / ꢑꢖ-ꢘꢐꢎ-ꢐꢎꢃ Iꢔ 9ꢔ aꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ Cꢄꢂꢄꢌꢄꢇ 5ꢉꢈꢊꢁꢈꢈꢉꢀꢃꢈ
ꢃ ꢏ#ꢃ ꢂ ꢄꢀꢁꢂꢔ
ꢃ
ꢃ
ꢀꢁꢁ
ꢃ
"$"ꢃ ꢀꢇꢀꢔ
ꢊꢄꢈꢋ
/ ꢖꢃ vꢔ Wꢔꢕ Cꢍꢑꢌꢐꢑꢃ hꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ tꢂꢀꢊꢍꢍꢌꢉꢃ*ꢈ ꢀ ꢎꢆꢍ bꢄꢎꢉꢀꢃꢄꢅ !ꢊꢄꢌꢍ ꢇ ꢀ {ꢊꢉꢍꢃꢊꢍꢈ ꢀ ꢎꢆꢍ
Üꢃꢉꢎꢍꢌ {ꢎꢄꢎꢍꢈ ꢀ ! ꢍꢂꢉꢊꢄ ꢀꢁꢁꢈꢃ "#ꢏꢃ ꢀꢆꢁꢉꢂꢔ
ꢊꢅꢉꢋ
5ꢐꢓꢓꢍ tꢖꢍꢃ 9ꢔ !ꢔꢕ / ꢖꢃ vꢔꢕ Wꢒꢎꢐ-ꢃ 5ꢔ 5ꢔꢕ aꢍ/ꢏꢒꢎꢍꢓꢏꢃ Wꢔ 9ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔꢕ 9ꢓꢓꢖꢒꢘꢘꢃ aꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢂꢂ
ꢃ
""ꢃ ꢀ ꢆꢔ
ꢊꢅꢀꢋ
Iꢍꢎ-ꢐꢎꢃ !ꢔ Dꢔꢕ {ꢍꢓꢌꢍꢘꢒꢑꢐꢃ tꢔꢕ Yꢍꢑꢓ-ꢐꢎꢃ Yꢔ Yꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ íꢐꢎ"ꢐꢓꢃ Wꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ tꢆꢇꢈꢓ /ꢆꢍ ꢓ
/ꢆꢍ ꢓ tꢆꢇꢈꢓ ꢀꢁꢂꢆ
ꢃ
":
ꢃ
ꢆꢔ
ꢊꢅꢁꢋ
{ꢍꢓꢌꢍꢘꢒꢑꢐꢃ tꢔꢕ ùꢐꢎ"ꢃ 5ꢔꢕ Yꢍꢑꢓ-ꢐꢎꢃ Yꢔ Yꢔꢕ / ꢖꢃ vꢔꢕ íꢐꢎ"ꢐꢓꢃ Wꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ /ꢆꢍ ꢒꢆꢇꢈꢊꢆꢍ ꢀꢁꢂꢆ
ꢃ
"ꢐꢃ
ꢁꢀꢉꢀꢔ
ꢁ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 40
1
2
3
ꢊꢅꢂꢋ
!ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ aꢒꢘ Ctꢒ*ꢓ-ꢐꢎꢃ Yꢔꢕ / ꢑꢖ-ꢘꢐꢎ-ꢐꢎꢃ Wꢔ .ꢔꢕ Iꢗꢐꢓ%ꢃ Wꢔꢕ .ꢗꢒꢑꢎ ꢒꢓ%ꢃ Çꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔ Wꢓ ! ꢓ
4
5
6
7
8
9
/ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁ
ꢊꢅꢄꢋ
ꢊꢅꢅꢋ
ꢊꢅꢆꢋ
ꢃ
"ꢏꢕꢃ ꢆꢅ ꢄꢔ
Çꢑꢍꢎꢃ 9ꢔꢕ 5*ꢍꢘꢖꢃ aꢔꢕ í ꢖꢘꢐ-ꢖꢏꢐ-ꢃ Dꢔ aꢔꢕ wꢍ%7ꢖꢃ aꢔ !ꢔ Cꢄꢂꢄꢌꢄꢇ 5ꢉꢈꢊꢁꢈꢈꢓ ꢀꢁꢁ
Dꢖꢘꢘꢖꢎ-ꢃ 5ꢔ Lꢔꢕ .ꢐꢘ ꢐꢓꢓꢃ 5ꢔꢕ {/ ꢖ??ꢑꢖꢎꢃ 5ꢔ Wꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔ bꢄꢎꢁꢂꢍ ꢀꢁꢁꢁ ꢐ#ꢕꢃ ꢆ ꢔ
Iꢐꢃ Iꢔ óꢔꢕ [ꢖꢃ óꢔ [ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔꢕ bꢍ"ꢍ ꢍꢑꢍꢃ [ꢔ !ꢔꢕ !%ꢓꢍꢎꢖꢃ Lꢔꢕ Ç-*ꢖꢃ wꢔ tꢆꢇꢈꢓ wꢍ&ꢓ .ꢓ ꢀꢁꢁꢆꢃ ꢑꢕꢃ
ꢃ "$"ꢃ ꢀꢈ ꢔ
ꢃ
ꢉꢄꢅꢂꢉꢁꢔ
ꢊꢅ ꢋ
Iꢐꢃ Iꢔ óꢔꢕ ù *ꢃ Wꢔ {ꢔꢕ Çꢍꢒꢃ bꢔ Wꢔꢕ bꢍ"ꢍ ꢍꢑꢍꢃ [ꢔ !ꢔꢕ !%ꢓꢍꢎꢖꢃ Lꢔꢕ Ç-*ꢖꢃ wꢔ Wꢓ ! ꢓ /ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢁꢂ
ꢃ
"ꢏ$
ꢃ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢂꢉꢔ
ꢊꢅꢇꢋ
5ꢖꢐ4Ctꢐꢑꢐ4ꢃ Lꢔꢕ [ꢖꢃ ùꢔ Iꢔꢕ D*ꢒꢃ {ꢔ òꢔꢕ aꢍꢏꢏꢐꢎꢃ /ꢔꢕ I*ꢍꢎ"ꢃ Iꢔ [ꢔꢕ / ꢐꢃ òꢔ Yꢔꢕ òꢍꢎ"ꢃ óꢔ aꢔꢕ ùꢍꢎ"ꢃ [ꢔꢕ
ꢃ ꢉꢄꢄꢔ
!ꢑꢘꢐ-ꢃ Wꢔ aꢔꢕ 5ꢖꢐ4Ctꢐꢑꢐ4ꢃ Lꢔꢕ Dꢒꢑꢒ-ꢘꢖ4ꢍꢃ tꢔ bꢄꢃꢀ [ꢍꢎꢎꢓ ꢀꢁꢂꢀ
5ꢐꢓꢓꢍ tꢖꢍꢃ 9ꢔ !ꢔꢕ / ꢖꢃ vꢔꢕ aꢍ/ꢏꢒꢎꢍꢓꢏꢃ Wꢔ 9ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ Wꢔꢕ Wꢒꢎꢐ-ꢃ 5ꢔ 5ꢔꢕ 9ꢓꢓꢖꢒꢘꢘꢃ aꢔ bꢄꢃꢀꢈꢊꢄꢅꢍ ꢀꢁꢂꢀ
Çꢍꢒꢃ bꢔ Wꢔ !/{ bꢄꢃꢀ ꢀꢁꢂꢀꢃ ꢑ
ꢊꢅꢈꢋ
ꢊꢆꢉꢋ
ꢃ "ꢏꢃ ꢁꢆ ꢈꢔ
ꢃ
ꢐ
ꢃ ꢀꢉꢆꢔ
ꢊꢆꢀꢋ
ꢊꢆꢁꢋ
ꢊꢆꢂꢋ
ꢊꢆꢄꢋ
ꢊꢆꢅꢋ
ꢊꢆꢆꢋ
ꢊꢆ ꢋ
ꢊꢆꢇꢋ
Cꢐꢏꢒꢑꢒꢌꢃ aꢔ ëꢔꢕ Yꢒꢑꢎ9- ꢐꢌꢃ !ꢔ !ꢔ /ꢆꢍ ꢓ wꢍ&ꢓ ꢀꢁꢂꢅ
Dꢖꢘꢘꢖꢎ-ꢃ 5ꢔ Lꢔꢕ .ꢐꢘ ꢐꢓꢓꢃ 5ꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ {/ ꢖ??ꢑꢖꢎꢃ 5ꢔ Wꢔ !ꢌ&ꢓ aꢄꢎꢍꢂꢓ ꢂꢃꢃꢃ
{ꢒꢎꢒ"ꢍ- ꢖꢑꢍꢃ Yꢔ Wꢓ hꢂ*ꢄꢃꢀ ꢍꢎꢓ /ꢆꢍ ꢓ ꢀꢁꢁꢀ ꢑ:$ꢃ ꢄꢆꢔ
{ꢒꢎꢒ"ꢍ- ꢖꢑꢍꢃ Yꢔꢕ Çꢒ ꢏꢍꢃ òꢔꢕ Iꢍ"ꢖ ꢍꢑꢍꢃ bꢔ Çꢍꢎꢂꢄꢆꢍꢌꢂꢀꢃ [ꢍꢎꢎꢓ ꢂꢃꢇꢈꢃ ꢄꢄꢆ ꢔ
/ ꢖꢎ/ ꢖꢓꢓꢍꢃ wꢔꢕ bꢍꢗꢐꢑꢍꢃ /ꢔ /ꢆꢍ ꢓ wꢍ&ꢓ ꢀꢁꢁꢇ "#ꢔꢃ ꢇ ꢄꢔ
.ꢍꢑꢎꢐꢘꢘꢃ 5ꢔ {ꢔꢕ {/ ꢍ*-ꢃ {ꢔ 9ꢔ hꢂ*ꢓ [ꢍꢎꢎꢓ ꢀꢁꢂꢂ "$ꢃ ꢄꢉꢁꢉꢔ
/ꢒꢎꢎꢐꢓꢓ9ꢃ bꢔ Dꢔꢕ Dꢐꢖ"ꢐꢑꢃ íꢔ 9ꢔ /ꢆꢍ ꢓ wꢍ&ꢓ ꢂꢃꢃ %ꢑꢃ ꢇ
aꢍꢑD*ꢐ-CDꢒꢎ4ꢍꢓꢐ4ꢃ {ꢔꢕ ò*?ꢖꢘꢃ 5ꢔ {ꢔꢕ Iꢒ:ꢍꢑꢏꢃ Wꢔ !ꢔ Yꢔꢕ aꢍꢑꢘꢖꢎꢃ {ꢔꢕ h-ꢒꢑꢖꢒꢃ Iꢔ aꢔꢕ Dꢍꢑ/ꢖꢍC{*ꢍꢑꢐ4ꢃ
ꢐꢏꢃ ꢂꢂꢇꢔ
tꢐꢑꢍꢃ Dꢔꢕ aꢍꢑꢘꢖꢎꢃ {ꢔꢕ .ꢍꢓꢓꢐ-ꢘꢐꢑꢒ-ꢃ [ꢔ aꢔꢕ Iꢒ7ꢐꢃ !ꢔ Wꢔꢕ [ꢒ:ꢃ tꢔ Wꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ /ꢐꢍꢃ tꢔ /ꢆꢍ ꢉꢈꢎꢂꢇꢋ
ꢄ 9ꢁꢂꢀꢒꢍꢄꢃ Wꢀꢁꢂꢃꢄꢅ ꢀꢁꢂꢁ "ꢑꢃ ꢀꢂꢂꢈꢇꢔ
Yꢖ%ꢃ Iꢔ Wꢔꢕ Wꢐꢒꢎꢃ íꢔ {ꢔꢕ Yꢒꢃ òꢔ Iꢔꢕ Yꢖ%ꢃ Yꢔ tꢂꢀꢊꢓ bꢄꢎꢓ !ꢊꢄꢌꢓ {ꢊꢉꢓ Üꢓ{ꢓ!ꢓ ꢀꢁꢁꢀ
ëꢐꢑ"ꢐꢐꢑꢃ Cꢔ íꢔꢕ Yꢓꢐꢌꢐꢑꢓꢍꢍꢎꢃ /ꢔ Wꢔꢕ aꢍꢘꢒ*-ꢐ2ꢃ tꢔꢕ Çꢒ:ꢑꢖꢐꢃ aꢔꢕ {ꢘ*?2ꢐꢎ-ꢃ 5ꢔ Wꢔꢕ Iꢍꢑꢘꢓꢃ Cꢔ 9ꢔ Lꢃꢀꢂ*ꢓ
ꢐꢐꢃ ꢀꢂꢀꢈꢔ
tꢒꢑꢘꢐꢑꢃ íꢔ íꢔꢕ ëꢍꢖꢏꢃ Çꢔ tꢔ Wꢓ hꢂ*ꢓ /ꢆꢍ ꢓ ꢀꢁꢁꢈ
ù ꢍꢎ"ꢃ Wꢔ 5ꢔꢕ Y*4ꢎꢐꢘ-ꢒꢌꢃ !ꢔ aꢔꢕ aꢐꢏꢌꢐꢏꢐꢌꢃ Lꢔ Dꢔꢕ / ꢖꢃ vꢔ Wꢔꢕ !ꢓ.ꢑꢐ/ ꢘꢃ Çꢔꢕ Wꢐꢎ-ꢐꢎꢃ tꢔ {ꢔꢕ Üꢓ-ꢘꢑ*7ꢃ
"#ꢕꢃ ꢁ ꢂ ꢔ
!ꢑꢘꢐ-ꢃ Wꢔ aꢔꢕ [ꢒ7ꢐ4Caꢍꢑꢘꢖꢎꢐ4ꢃ aꢔꢕ Dꢖꢑꢍ*ꢏꢐꢘꢃ !ꢔꢕ 5ꢖꢐ4Ctꢐꢑꢐ4ꢃ Lꢔꢕ {ꢍꢎ4ꢃ Cꢔꢕ Dꢒꢑꢒ-ꢘꢖ4ꢍꢃ tꢔ Wꢓ ! ꢓ
/ꢆꢍ ꢓ {ꢀꢊꢓ ꢀꢁꢂꢀ "$ꢐꢃ ꢁꢉꢁꢀꢇꢔ
aꢖꢓꢓꢍꢑꢃ 5ꢔꢕ ëꢐꢎ2ꢍꢘꢍꢑꢍ%ꢍꢎꢃ [ꢔꢕ 5ꢒꢐꢑꢑꢐꢑꢃ [ꢔ Iꢔ Wꢓ tꢆꢇꢈꢓ /ꢆꢍ ꢓ / ꢀꢁꢁꢇ
[ꢐ.ꢓꢍꢎ/ꢃ bꢔꢕ aꢐꢑ/ꢖꢐꢑꢃ bꢔꢕ Çꢒ%ꢍꢃ hꢔꢕ Yꢍ--ꢖ.ꢍꢃ !ꢔ Iꢔꢕ ùꢒꢑꢖꢎꢍꢃ [ꢔꢕ !*.ꢍꢎC{ꢐꢎ4ꢖꢐꢑꢃ tꢔꢕ tꢍ-D*ꢖꢐꢑꢃ /ꢔ
ꢐ%ꢃ ꢀꢉꢁ ꢁꢔ
Iꢖꢑꢍꢒ2ꢍꢃ {ꢔꢕ Iꢖꢑꢍꢘꢍꢃ Yꢔꢕ { ꢖꢒꢎꢒ9ꢍꢃ aꢔ !ꢃ*ꢍ,ꢓ /ꢆꢍ ꢓ Lꢃꢎꢓ 9ꢌꢓ ꢀꢁꢁꢅ
[ꢐꢌꢐꢎꢘꢖ-ꢃ bꢔꢕ wꢍ:ꢍ- ꢏꢐ ꢃ !ꢔ aꢔ aꢔꢕ 9ꢓꢏꢐꢑꢃ Lꢔ !ꢔꢕ òꢍꢎ"ꢃ Wꢔ Iꢔꢕ 5ꢍ--ꢃ !ꢔꢕ {ꢒꢘꢖꢑꢖꢒ*C[ꢐꢌꢐꢎꢘꢖ-ꢃ /ꢔ
/ꢆꢍ ꢓ aꢄꢎꢍꢂꢓ ꢀꢁꢁꢅ "ꢑꢃ ꢀꢄꢈꢂꢔ
ꢊ ꢈꢋ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ Iꢍꢖ--ꢃ íꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ [ꢐꢍꢑ9ꢃ 9ꢔꢕ aꢍꢑꢘꢖꢎꢃ {ꢔꢕ .ꢐꢘ ꢐꢓꢓꢃ 5ꢔ tꢆꢇꢈꢓ /ꢆꢍ ꢓ /ꢆꢍ ꢓ
tꢆꢇꢈꢓ ꢀꢁꢂꢁ "ꢏꢃ ꢁꢇꢉꢀꢔ
ꢊꢇꢉꢋ YꢑꢐꢗEF2ꢃ aꢔꢕ 5ꢍꢎG2ꢃ aꢔꢕ Iꢍꢑꢘꢓꢃ Cꢔ Wꢓ 9ꢅꢍꢊꢎꢂꢀꢄꢃꢄꢅꢓ /ꢆꢍ ꢓ ꢂꢃꢃꢂ
ꢃ ""ꢐꢃ ꢁꢈ ꢇꢔ
ꢃ
""ꢃ ꢂ ꢔ
ꢃ
ꢃ
ꢃ
ꢃ
ꢔ
ëꢔ aꢔꢕ bꢖ/ ꢒꢓ-ꢃ wꢔ Wꢔꢕ Iꢖ""ꢖꢎ-ꢃ {ꢔ Wꢔꢕ /ꢐꢍꢃ tꢔꢕ [ꢒ:ꢃ tꢔ Wꢔ 5ꢄꢅꢎꢀꢃ Çꢂꢄꢃꢈꢄꢊꢎꢉꢀꢃꢈ ꢀꢁꢂꢆ
ꢊꢆꢈꢋ
ꢃ
ꢃ
ꢊ ꢉꢋ
ꢊ ꢀꢋ
/ꢆꢍ ꢓ ꢀꢁꢁꢈ
ꢊ ꢁꢋ
ꢊ ꢂꢋ
ꢃ %%ꢃ ꢅꢉꢉ ꢔ
ꢃ
ꢃ ꢔ#ꢃ ꢅꢉꢁꢇꢔ
Wꢔ /ꢆꢍ ꢓ wꢍ&ꢓ ꢀꢁꢁꢄ
ꢊ ꢄꢋ
ꢃ
ꢃ
ꢊ ꢅꢋ
ꢊ ꢆꢋ
ꢃ """ꢃ ꢀ ꢆꢂꢅꢔ
/ꢆꢍ ꢓ /ꢀ ꢁꢃꢓ ꢀꢁꢂꢆ
ꢃ
ꢊ
ꢋ
ꢃ ꢐ$ꢃ ꢂꢇꢀꢄꢔ
ꢊ ꢇꢋ
ꢃ
ꢃ
ꢃ
$"ꢔꢃ ꢀ ꢈꢔ
ꢁꢇ
ACS Paragon Plus Environment

Page 29 of 40
Journal of the American Chemical Society
1
2
3
TOC GRAPHIC
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢁꢈ
ACS Paragon Plus Environment