Discovery of unusually stable reduced viologen via synergistic folding and encapsulation

Article abstract of DOI:10.1149/2.0711915jes

The preparation and identification of an unusually stable viologen radical has been demonstrated. The intramolecular interactions of bis-viologen species such as 1,1”-[1,2-phenylenebis(methylene)]bis[1’-methyl-4,4’-bipyridinium] (1) and 1,1”-(1,3-propaned

Full text of DOI:10.1149/2.0711915jes

Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
H825
0
013-4651/2019/166(15)/H825/10/$38.00 © The Electrochemical Society
Discovery of Unusually Stable Reduced Viologen via Synergistic
Folding and Encapsulation
1
2
2
3
Molly M. MacInnes,
Benjamin R. Cousineau, Sarah M. Youngs, Kumar Sinniah,
Joseph J. Reczek,2 and Stephen Maldonado
,z
1,4,∗,z
1
Department of Chemistry, University of Michigan, Ann Arbor, Michigan, USA
Department of Chemistry, Denison University, Granville, Ohio, USA
Department of Chemistry and Biochemistry, Calvin College, Grand Rapids, Michigan, USA
Program in Applied Physics, University of Michigan, Ann Arbor, Michigan, USA
2
3
4
The preparation and identification of an unusually stable viologen radical has been demonstrated. The intramolecular inter-
actions of bis-viologen species such as 1,1”-[1,2-phenylenebis(methylene)]bis[1’-methyl-4,4’-bipyridinium] (1) and 1,1”-(1,3-
propanediyl)bis[1’-methyl-4,4’-bipyridinium] (2) upon reduction and complexation by cucurbit[8]uril (CB[8]) result in the formation
of a unique monoradical species. Cyclic voltammetric responses for 1 and 2 in the presence of CB[8] showed a splitting of the degen-
−
−
erate 2 e redox wave (corresponding to the 4+/2+ transition of the bis-viologen) into two separate 1 e redox waves indicating the
stabilization of the 3+ form of the bis-viologens. Visible light absorbance showed that this species had a distinct electronic structure,
imparting a blue color unlike the 2+ state of the bis-viologen or the 1+ state of partially reduced methylviologen. The persistence
of the blue color over time, even in the presence of dissolved O2, illustrated a significant stabilization of these monoradical species
in comparison to any other partially reduced viologen radical. Molecular orbital calculations indicated that the origin of this stability
arises from the SOMO of the monoradical that features the electron density as buried between the two cofacial viologen units, hin-
dering electronic coupling with oxidants such as O2 and thereby providing enhanced stability for energy storage and electrochromic
applications.
©
2019 The Electrochemical Society. [DOI: 10.1149/2.0711915jes]
Manuscript submitted August 29, 2019; revised manuscript received October 16, 2019. Published November 15, 2019.
Viologens(1,1’-disubstituted-4,4’-bipyridiniums)areanimportant
class of redox-active compounds, often used as electron acceptors
in fundamental studies of photochemical, electrochemical, and pho-
toelectrochemical charge-transfer.1,2 Viologens are also widely em-
ployed as the redox-active component in rotaxane-based molecular
(1) and a propyl group (2) (Figure 1). These linkers space two
methyl viologen units by the similar short N-N’ distance of ap-
proximately 5 Å but differ in preorganization geometry and linker
rigidity. The presented electrochemical and spectroscopic data indi-
cate that when these viologen dimers are reduced in the presence
of cucurbit[8]uril (CB[8]), a unique and unusually stable monorad-
ical species is formed. The summary results show that specifically
the combination of the structure of the viologen dimer with subse-
quent encapsulation by CB[8] is required to form a stabilized, reduced
radical. Further, data are presented that distinguish these stabilized
species from viologen ‘pimers’ which have been extensively studied
3
–6
electronics, as the oxidant that forms reactive oxygen species in
7
–9
10–15
pesticides, asthechargestoragemediuminredoxflowbatteries,
16–19
and as the active material in electrochromic devices.
The appeal
of viologen compounds stems from two reasons. First, dicationic
2
+
•+
viologens (V ) readily form singly charged radical cations (V )
−
1,2
when reduced by 1e . The rate constant for this redox transition
is large because it involves minimal bond-length changes (i.e. the
inner-reorganization energy requirement is small). Second, this re-
2
,51–55
previously.
2
dox process occurs at a sufficiently negative standard potential that it
+
−
can mediate desirable, fuel-forming half reactions (e.g. 2H + 2e
Experimental
+
−
20,21
→
H
2
; CO
2
+ 2H + 2e → H
2 2
CO ).
Chemicals and materials.—Anhydrous dibasic sodium phos-
phate (Sigma Chemical Company), monobasic potassium phosphate
(99.6%, J.T. Baker, Inc.), sodium dithionite (≥82%, Sigma Aldrich),
cucurbit[7]uril hydrate (CB[7], Sigma Aldrich), cucurbit[8]uril hy-
drate (Sigma Aldrich), iodomethane (99%, Sigma Aldrich), 4,4’-
bipyridine (98%, Sigma Aldrich), DMF (≥99.8%, Sigma Aldrich),
α,α’-dibromo-o-xylene (97%, Sigma Aldrich), and deuterated water
(99.8 atom %, Acros Organics) were used as received. Water with a re-
sistivity >18.2 MΩ (Nanopure Barnstead Water Purification) was used
throughout. Glassy carbon electrodes were prepared by first polishing
with alumina (1 μm, 0.3 μm, and 0.05 μm alumina suspensions) on
felt and then sonicating in methanol for 30 minutes. A carbon cloth
electrode was used for bulk electrolysis and was prepared by thor-
oughly rinsing with clean water. Platinum mesh was used as a counter
electrode. For bulk electrolysis experiments, the platinum mesh was
isolated from the working electrode by a frit.
Arguably, the Achilles heel of viologens in redox-based energy
and electronic applications is the instability of their reduced forms in
For species like methyl viologen radicals (MV ), the rate
s ), effectively
at the diffusion limit. Accordingly, strategies that inhibit the parasitic
oxidation of reduced viologens with O are of significant importance.
2
One potential approach to stabilize reduced viologens is by em-
1
,22,23
•+
air.
8
−1 −1 24
constant for oxidation by O
2
is large (8 × 10 M
4
,25–34
bedding the viologen inside protective structures.
In this vein,
18,19,35–44
25,26,29,31,32
viologens have been inserted in polymers,
micelles,
Inclusion complexes between viologens
and cucurbit[n]urils (CB[n]), a class of rigid and water-soluble macro-
30,45–47
and macromolecules.
4
,27,28,33,48–50
cyclic host molecules, have also been examined.
The cu-
mulative observations suggest that although the apparent redox poten-
tials of the viologens can be tuned, the stability of their reduced forms
2
is minimal if O is not rigorously excluded.
In this work, we explore the concept that a hierarchical supramolec-
ular assembly can form structures that natively inhibit parasitic reac-
2
tion with O . Herein, we report the design, assembly, and electro-
Synthesis of N-Methyl-4,4’-bipyridinium iodide.—Iodomethane
1.20 mL, 19.2 mmol, 1.0 eq.) was added to 4,4’-bipyridine (2.9994 g,
chemical behaviors of covalently-linked viologen dimers complexed
with cucurbit[7]uril (CB[7]) and cucurbit[8]uril (CB[8]) in aque-
ous solution. To avoid confusion, in this report the term “dimer”
refers strictly to two viologens covalently bound through a defined
linker. Two specific linkers are explored, an ortho-benzyl group
(
1
9.2 mmol, 1.0 eq.) in CH Cl (50 mL). The solution was stirred at
2
2
room temperature for 24 hours. A bright yellow solid precipitated out.
The solid was filtered and any residual 4,4’-bipyridine washed away
with diethyl ether (3 × 10 ml), yielding N-Methyl-4,4’-bipyridinium
1
iodide (86.5%, 4.9497 g). H NMR of the solid showed >99% with
∗
no trace of starting material, and the collected solid was used directly
in the synthesis of 1 and 2.
Electrochemical Society Member.
E-mail: reczekj@denison.edu; smald@umich.edu
z
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H826
Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
h
g
i
f
f
N
N
N
N
e
d
N
N
e
d
N
c
b
N
c
b
c
b
N
N
a
a
a
2+
4+
4+
MV
1
2
O
O
N
O
Figure 1. Graphical depictions of 14 , 2 , CB[8],
+
4+
O
N
N
N
and CB[7]. Lower case letters correspond to labels
N
N
N
1
for signatures in H NMR data.
N
N
N
N
O
N
N
N
N
N
N
N
N
O
j, k
O
O
O
N
N
N
N
N
N
N
O
O
O
O
N
N
N
N
N
N
l
N
N
O
O
O
l
N
N
N
N
j, k
O
O
O
O
O
N
O
N
O
N
O
O
N
N
N
N
O
N
N
N
N
N
N
N
N
N N
N
N
O
N
O
N
N
O
O
CB[7]
CB[8]
Synthesis of 1.—N-Methyl-4,4’-bipyridinium iodide (0.6035 g,
.0 mmol, 2.2 eq.) was reacted with α,α’-dibromo-o-xylene (0.2432 g,
.92 mmol, 1.0 eq.) in DMF (30 mL). The solution was stirred at
0°C for 3 days with a rubber septum and vent needle. An orange
2 hours. The undissolved cucurbituril was filtered out using a 0.2 μm
2
0
6
nylon syringe filter. Based on NMR integrations, this resulted in a 1:1
4
+
4+
4+
ratio of 1 :CB[8]; a 1:1 ratio of 2 :CB[8]; a 1:30 ratio of 1 :CB[7];
4
+
and a 1:3 ratio of 2 :CB[7].
product precipitated out. The product was filtered and washed with
DMF (2 × 10 mL) to completely remove any residual N-Methyl-
High pressure liquid chromatography (HPLC).—Analytical
HPLC data was obtained on an Agilent HPLC 1100 with a DAD
detector, and a Microsorb MV 100-5-c18 column (150 × 4.6 mm,
5 um). The chromatography reported for 1 and 2 was performed under
2
1
4
,4’-bipyridinium iodide, yielding 1 (87.0%, 0.6893 g). H NMR
(
4
400 MHz, D
2
O) δ 9.194(d, J = 6.8 Hz, 4H), 9.112(d, J = 6.4 Hz,
H), 8.669(d, J = 6.8 Hz, 4H), 8.598(d, J = 6.8 Hz, 4H), 7.695(d, J =
9
.2 Hz, 2H), 7.423(d, J = 9.2 Hz, 2H), 6.229(s, 4H), 4.554(s, 6H) ppm;
isocratic condition with H O-methanol (90:10, v/v) as mobile phase,
1
3
−1
C NMR (400 MHz, D
2
O) δ 151.066, 149.571, 146.438, 145.901,
flow rate of 1mL min , and column temperature of T = 35°C.
131.451, 130.981, 130.837, 127.579, 126.908, 61.362, 48.531 ppm.
Analysis by analytical HPLC showed > 99% purity of 1 in the recov-
Isothermal titration calorimetry (ITC).—ITC data was obtained
on a Nano ITC Standard Volume instrument from TA Instruments.
Experiments were performed in 10 mmol sodium phosphate buffer and
all solutions were degassed for 25 min at room temperature prior to
running ITC experiments. The reference cell in the ITC was filled with
degassed ultrapure water (18.2 Mꢀ cm resistivity). Experiments
were run in multiple-injection mode with 20 injections (12 μL each)
with 1000 s injection intervals. The stir rate was 250 rpm. Once the
cell had been filled and the syringe loaded, stirring was turned on and
the apparatus was allowed to autoequilibrate. All data were recorded
with the TA Instruments software provided, ITCRun version 2.2.3.
The area under each peak was integrated, and the resulting data were
further analyzed using TA NanoAnalyze version 2.4.1, fitting to a 1:1
binding model of dimer to cucurbituril.
1
ered solid. A schematic depiction of the synthetic protocol and the H
13
NMR, C NMR, and HPLC spectra are presented in the Supporting
Information (Figures S1–S4, respectively).
−
1
Synthesis of 2.—N-Methyl-4,4’-bipyridinium iodide (0.6163 g,
2
0
3
.1 mmol, 2.2 eq.) was added to 1,3-dibromopropane (0.094 mL,
.93 mol, 1.0 eq.) in DMF (30 mL). The solution was stirred at 60°C for
days with a rubber septum and vent needle. A yellow product precip-
itated out. The product was filtered and washed with DMF (3 × 10 mL)
to completely remove any residual N-Methyl-4,4’-bipyridinium io-
1
dide, yielding 2 (35%, 0.258 g). H NMR (400 MHz, D
2
O) δ 9.098(d,
J = 6.8 Hz, 4H), 8.936(d, J = 6.8 Hz, 4H), 8.484(d, J = 6.8 Hz, 4H),
8
2
1
3
.408(d, J = 6.8 Hz, 4H), 4.873(t, J = 8.0 Hz, 4H), 4.368(s, 6H),
13
2
49.677, 146.409, 145.805, 127.522, 126.841, 58.306, 48.483, 31.905,
.859(m, J = 7.9 Hz, 2H) ppm; C NMR (400 MHz, D
O) δ 150.817,
Absorbance measurements.—UV-visible-NIR absorbance spec-
tra were obtained using a Varian Carey 5000 UV-vis-NIR spectropho-
tometer. Solutions of 0.25 mM viologen with and without approx-
imately 1 equivalent of cucurbituril were electrolyzed and immedi-
ately syringed into a 1-cm quartz cuvette sealed with a septum that
0.324 ppm. Analysis by analytical HPLC showed > 99% purity of 1
in the recovered solid. A schematic depiction of the synthetic protocol
and the H NMR, C NMR, and HPLC spectra are presented in the
Supporting Information (Figures S1–S4, respectively).
1
13
2
had been flushed with N . The supporting electrolyte was deaerated
Preparation of viologen-cucurbituril mixtures.—All viologen-
cucurbituril solutions were prepared by first dissolving the appropriate
0.075 M phosphate buffer (pH ∼7) using sodium monohydrogen phos-
phate and potassium dihydrogen phosphate. Backgrounds were taken
using plain electrolyte solutions. Absorbance changes upon expo-
amount of viologen in water (or D
2
O) or buffer solution. Cucurbituril
was then added in excess and the solution was stirred for at least
2
sure to O was monitored using an Analytik Jena SPECORD S600
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
H827
spectrophotometer. Solutions used for these experiments were 30 μM
8
6
4
2
0
viologen in 0.075 M phosphate buffer (pH 7). Solutions containing
cucurbituril were prepared by dissolving excess cucurbituril in the
buffered viologen solution, stirring for at least two hours and then
filtering through a 0.2 μm nylon syringe filter to remove undissolved
cucurbituril. Thesesolutionswereelectrolyzedto100%underastream
of nitrogen and then the electrodes and nitrogen stream were removed
and an immersion probe (Hellma Analytics) was quickly inserted.
Spectra were recorded using the immersion probe while stirring the
solution at a constant speed.
MV
MV + CB[7]
MV + CB[8]
-
2
Electrochemical measurements.—Electrochemical measure-
ments were taken using a CH Instruments 760C potentiostat. A
three-electrode cell was used with a supporting electrolyte of
deaerated 0.075 M phosphate buffer (pH∼7) as described above. A
platinum mesh was used as the counter electrode and Ag|AgCl was
used as the reference electrode. The working electrode for cyclic
4
2
0
1
1
1
+ CB[7]
+ CB[8]
voltammetry experiments was glassy carbon (geometric surface area
2
0
.071 cm ) and for electrolysis of bulk solutions was a carbon cloth
electrode.
1H NMR spectroscopy.—Measurements were conducted on a
Varian MR 400 instrument at room temperature and were referenced
against the solvent peak. Measurements were performed on 1 mM vi-
-2
2
ologen solutions in D O with and without approximately 1 equivalent
of cucurbituril. Sodium dithionite was used to reduce the solutions.
2
2
2
4
2
+ CB[7]
+ CB[8]
Molecular orbital calculations.—Molecular modeling was carried
out using Spartan ’16 for Macintosh. All structures were constructed
in Spartan and equilibrium geometries were calculated for each re-
dox state independently (4+, no unpaired elections; 3+, one unpaired
electron; 2+, two unpaired electrons) for dimers 1 and 2 and the cor-
responding complexes of the dimers and CB[8]. All geometry calcu-
lations were initiated from a randomized starting point and carried out
using density functional theory in water solvent with a B3LYP func-
0
-2
∗
tional at the 6–31G level. The calculated equilibrium structures were
then used in energy calculations to obtain orbital and spin-state infor-
mation using density functional theory in water solvent with a B97D
-
4
0.0
-0.2 -0.4 -0.6 -0.8
∗
∗
functional at the 6–311+G level. The B97D functional was used to
E / V vs E(Ag/AgCl)
achieve energy calculations that more adequately include dispersion
55
stabilization of the delocalized radicals.
Figure 2. Cyclic voltammetry of viologen/cucurbituril pairs. From top to bot-
tom: MV; 1; 2; MV + CB[7]; 1 + CB[7]; 2 + CB[7]; MV + CB[8]; 1 +
CB[8]; 2 + CB[8]. Concentrations of MV, 1, and 2¸were nominally 0.4, 0.2,
and0.2mM, respectively. Allvoltammetrywasperformedindeaerated0.075M
phosphate buffer at pH = 7 at a glassy carbon electrode and a scan rate of 0.05 V
Results
Dimer conformers.—Molecular modeling was performed to as-
certain the relevant conformers of 1 and 2. The pertinent results indi-
−
1
s
.
4
+
4+
cated that dimers 1 and 2 prefer to adopt extended conformations
in water due to charge-repulsion (Supporting Information, Figure S5).
2
+
2+
In contrast, the reduced dimers 1 and 2 assemble into folded struc-
tures that balance charge-repulsion with π-stacked stabilization of the
2
+
2+
radical species. While 1 and 2 adopt conformations with simi-
Table I. Voltammetric Features of Viologens With and Without
Cucurbiturils in Aqueous Solution.
lar distances between the viologen groups, the equilibrium geometry
2
+
structure calculated for the rigid o-xylene-linked 1 had near-perfect
2
+
alignment of overlapping viologen surfaces, while propyl-linked 2
Compound
E1/2a,b/V
ꢁEpa,c/V
displayed a significantly skewed relative orientation of the viologen
groups (Supporting Information, Figure S6).
MV
1
2
−0.638 ± 0.002
−0.424 ± 0.002
−0.466 ± 0.002
−0.663 ± 0.002
−0.540 ± 0.002
−0.484 ± 0.004
−0.192 ± 0.009
0.058 ± 0.002
0.036 ± 0.002
0.039 ± 0.002
0.065 ± 0.003
0.107 ± 0.002
0.061 ± 0.004
0.062 ± 0.009
4
+
,
Cyclic voltammetry.—The cyclic voltammetric responses for 1
24+
2+
, and MV in the presence and absence of cucurbit[7]uril (CB[7])
MV + CB[7]
1 + CB[7]
MV + CB[8]
1 + CB[8]
and CB[8] are shown in Figure 2. In the presented potential range,
MV , compound 1 , and compound 2 showed one single set of
waves corresponding to the formation of the MV radical ion and
p
the 1 and 2 diradical ions, respectively. The peak splitting (ꢁE )
2
+
4+
4+
•
+
2
+
2+
2
1
+
4+
values for MV and 1 were 0.058 ± 0.002 and 0.036 ± 0.002 V,
a
First most positive redox wave.
Midpoint between reduction and oxidation peak potentials with re-
respectively, and the absolute difference in the midpoint potentials
b
2
+
4+
(E1/2) between the cathodic and anodic waves for MV and 1 was
−
1
4
+
spect to E(Ag/AgCl) at a scan rate of 0.05 V s in deaerated 0.075 M
phosphate buffer (pH = 7).
0.214 V (Table I). The small ꢁE
p
value for 1 (Table I) suggested
4
+
that both viologen units in 1 were reduced at the same potential
c
Separation between reduction and oxidation peak potentials.
−
56
(i.e. a ‘degenerate’ set of two 1 e processes) and the positive shift
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H828
Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
ized by a factor of v−1/2. At a comparatively slow scan rate of
2
2
1
1
5
0
5
0
−
1
-
1
0.01 V s , two nominally-reversible redox waves were distinctly
0
.01 V s
−1
-
1
visible. At 0.1 V s , the peak current for the first reduction wave
0.1 V s
-
1
1 V s
decreased while the peak currents for the corresponding oxidation
−
1
wave and second redox process remained the same. At 1 V s
,
the peak current for the first reduction wave was minimal while the
other voltammetric features were nominally unchanged, consistent
with the first reduction process occurring at slower timescales than
the other electrochemical processes. Despite the less sharp voltam-
metric features, the electrochemical behavior of 2 in the presence of
CB[8] demonstrated a similar dependence with scan rate (Supporting
Information, Figure S7).
5
0
ν
-
5
10
15
-
-
0
.0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6
1
NMR spectroscopy.— H NMR spectra were collected to gain fur-
E /V vs E(Ag/AgCl)
ther insight on the complexation of 1 and 2 with CB[7] and CB[8].
Figure 4 shows the H NMR spectra of 1 and 2 in various oxidation
1
Figure 3. Cyclic voltammetric responses for 1 mM concentrations of 1 in the
presence of 1 mM of CB[8] at various scan rates, v. The y-axis represents the
measured currents divided by the employed scan rate. The electrolyte was a
deaerated 0.075 M phosphate buffer (pH 7).
states in the presence and absence of CB[7] and CB[8] in D
2
O. Upon
4+
complexation by CB[7], all proton signals for 1 moved to higher
field/lower chemical shift values (Figure 4a). The aromatic protons
of the viologen units and the bridging methylene peaks were shifted
most extensively whereas the peak corresponding to the methyl pro-
tons was affected only minimally. The presence of CB[8] in solution
observed in the voltammetry for 14+ compared to MV2+ was in accord
1
4+
caused qualitatively similar changes in the HNMR spectrum of 1
.
with an intramolecular dimerization reaction occurring after reduction
−
4+
25,57,58
These changes were consistent with the premise that in the 4+ oxi-
by 1e of each viologen unit in 1 in this electrolyte,
dation state, the viologen rings are spread apart in a “linear” dimer
4+
ꢀ
+
− Vꢀ+ → V
2+
conformation due to charge repulsion and the benzyl linker of 1
V
2
[1]
spends significant time inside of the cucurbituril cavity. Conversely,
the terminal methyl groups on the viologen units are predominantly
outside the cavities of the cucurbiturils since there is no possibility
The voltammetry of free 24 was similar to that of 1 . Accordingly,
+
4+
4
+
the primary redox response for 2 was assigned the 4+/2+ process,
5
6,59
4+
60
which is followed by dimerization.
We note that 2 was prone
of hydrogen bonding by those hydrogens with the C = O groups.
2
+
to decomposition back to the constituent viologen monomer units if
stored in air and exposed to light for prolonged periods of time. This
proclivity toward decomposition was not observed for 1
The paramagnetic character of 1 after reduction convoluted the pro-
ton signals of the viologens in the corresponding H NMR spectrum,
1
4
+
.
precluding definitive assignment of viologen orientation upon reduc-
4
+
4+
2+
1
Addition of CB[7] to solutions containing either 1 , 2 , or MV
tion. However, the significant broadening of the H NMR signals
altered the voltammetric responses. In all three cases, the peak cur-
rents were smaller, consistent with complexation of the viologens with
CB[7], which slowed diffusion in solution. Furthermore, in all three
cases the E1/2 values were shifted slightly negative, in accord with the
from CB[8] in the presence of reduced dimer was consistent with
tight association, and therefore close proximity, of folded 3+ and
2+ paramagnetic radical dimers within CB[8]. The 2+ dimer species
showed a more significant effect on the CB[8] protons, implying an
average closer proximity of radical character to the CB[8] protons. In
•
+
27,33,48
premisethatCB[7]willaccommodateonlyoneV unitatatime
4
+
and effectively eliminating the possibility of strong association be-
Figure 4b, the spectra for 2 with and without either CB[7] or CB[8]
tween the two V• units. For the oxidations of both 1 and 2 , the
+
2+
2+
1
are shown. The changes in the H NMR spectral features were gener-
peak current was comparatively attenuated in the presence of CB[7].
ally similar to those in Figure 4a. The signals for the aromatic peaks
of the viologen units and the methylene peaks of the propyl linker
were more strongly affected by the presence of cucurbituril than the
corresponding signatures of the methyl protons. These data suggest
that similar to 1 , the viologen units and propyl linker of 2 in the
4+ oxidation state are primarily inside the cucurbituril cavities upon
complexation.
2
+
+
The diminution was more pronounced for 2
.
4
4+
2+
Addition of CB[8] to solutions of either 1 , 2 , or MV effected
notably different changes in the voltammetric behaviors of each com-
2
+
4+
pound. For MV , the current again decreased due to slower diffusion
2
+
of the MV /CB[8] complex, but now the voltammetric wave shifted
toward more positive potentials (Table I). This shift was in accord with
•
+
a stabilization effect. Specifically, stacking of two separate MV con-
2
8,33
4+
tained within CB[8] forms the ‘pimer’ configuration.
For both 1
UV-visible-NIR spectroscopy.—Figure 5 displays photographs of
1 with and without CB[8] at various stages of reduction. The top left
4
+
and 2 , the voltammetric responses were considerably more complex.
4
+
The current magnitudes for both compounds were decreased and mul-
and bottom left panels show colorless solutions of 0.1 mM 1 and
4
+
4+
tipleredoxtransitionswereapparent. For1 , thevoltammetryshowed
0.1 mM 1 with an equivalent of CB[8], respectively. After adding ap-
two discrete, reversible redox waves with ꢁE values larger than that
p
proximately a half equivalent of sodium dithionite to these respective
solutions, the viologen solutions were partially reduced. In the ab-
sence of CB[8], the solution color was a pale purple. This same shade
of purple was observed upon adding one full equivalent of sodium
4
+
of 1 without CB[8]. The first wave occurred at the most positive
potential of any of the viologen solutions. Integration of the charge
passed across each wave observed in the presence of CB[8] showed
4
+
2+
that the charge for each wave was approximately equivalent. For 2
,
dithionite to yield a solution of 1 . In contrast, adding one half equiv-
4
+
the voltammetric response did not show two well-defined reversible
waves but faradaic current was passed across the same general range
of potentials. Additional voltammetric measurements were performed
alent of sodium dithionite to the solution of 1 with CB[8] caused
the solution to have a vivid blue hue. This solution, and the corre-
sponding solution containing 2 instead of 1, was confirmed by EPR
to contain a species with an unpaired electron (Supporting Informa-
tion). As expected for a delocalized and dynamic radical species, no
hyperfine coupling was observed in the EPR spectra. A more detailed
analysis of EPR spectra was not carried out on solutions having one
half equivalent of sodium dithionite to dimer as the exact concentra-
tion and conformation dynamics of the radical species could not be
controlled with sufficient precision. Upon adding a full equivalent of
sodium dithionite, the solution of 1 and CB[8] converted to the same
4
+
with a compound identical to 1 except that the linker featured the
two viologen units in the meta positions rather than ortho. In this case
4
+
4+
and in contrast to 1 (and 2 ), the presence of CB[8] did not result
in any appreciable shift in the voltammetric response, indicating that
the meta dimer was not similarly effected by CB[8].
4
+
The voltammetric response of 1 in the presence of CB[8] was
investigated further. Figure 3a shows the influence of scan rate (v)
on the observed voltammetric responses, with the currents normal-
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
H829
a)
b)
2
+
2
+
+
2
2
+ CB[8]
+ CB[8]
+ CB[7]
1
1
+ CB[8]
+ CB[8]
3
3+
4
+
4+
Figure 4. 1H NMR spectra of a) aqueous D2O solutions of 1 in
various oxidation states and in the presence of cucurbiturils and
b) aqueous D2O solutions of 2 in various oxidation states and in
the presence of cucurbiturils.
1
1
+ CB[7]
2
4
+
4+
j
l
k
+ CB[8]
j
l
k
2
+ CB[8]
4
+
4+
1
2
a
a
f
d,e b,c
d,e b,c
g h
f
9
8
7
6
5
4
9
8
7
6
5
4
3
Chemical shift /ppm
Chemical shift /ppm
purple color observed without CB[8] in solution. These colors were
also observed during bulk electrolyses of these solutions, indicating
the role of sodium dithionite was simply as a sacrificial reductant and
not as a complexing agent. In bulk electrolysis of 14 with CB[8], the
blue color was observed shortly after passing cathodic current, with
the color growing stronger until about half of the charge was passed
inclusion of CB[7] in solution effected no appreciable change in these
2+
2+
spectral features for either 1 or 2 . Addition of CB[8] also produced
little change to these spectra when the viologens were fully reduced to
the 2+ states. However, when the viologens were partially reduced to
the 3+ states in CB[8], a profound change in the spectra was observed.
+
3
+
In this case, the visible band was shifted from 536 nm to 623 nm for 1
4
+
2+
3+
that was required to completely reduce 1 to 1 . Upon passing fur-
and from 534 to 622 nm for 2 . The NIR band disappeared completely
from the spectrum of 1 and 2 . For these compounds, an increase
of absorption was noted near the cutoff, i.e. >1400 nm.
2
+
3+
3+
ther charge, the blue transitioned to the purple color of 1 . The same
4
+
color changes were observed in analogous experiments with 2 and
2
+
CB[8] but never for solutions of MV and CB[8].
Figure 6 shows the UV-vis-NIR spectra of 1 and 2 with and without
the addition of cucurbiturils after either partial (3+ oxidation state) or
Lifetime in aerated solutions.—During the colorimetric experi-
ments, a notable persistence of the blue color in air was observed
2
+
2+
2+
2+
full (2+ oxidation state) electrochemical reduction. Both 1 and 2
relative to the quick fading of the purple color for both 1 and 2
.
had broad asymmetric absorbances between 500 and 600 nm and a
second broad band around 850 nm, consistent with the known visible
Thisobservationwasinvestigatedfurther. Figure7presentstherelative
rates of change in the visible light absorbance spectra of the reduced
52,58,61,62
light absorbance properties of viologen ‘pimers’ in water.
The
viologens upon exposure to O with and without CB[8] present. In
2
Figure 5. Color variation in solutions of 1 with and without
CB[8] after addition of reducing equivalents, defined as one re-
−
ducing equivalent would reduce each viologen unit by 1e . Top
4
+
row: 1 (left); 1 after reduction with approximately half of the
electrons necessary to fully reduce the solution to 1 (middle);
after the solution has been fully reduced to 1 (right). Bottom
row: 1 with CB[8] (left); 1 with CB[8] after reduction with
2
+
2
+
1
4
+
approximately half of the electrons necessary to fully reduce the
2
+
solution to 1 with CB[8] (middle); 1 with CB[8] after the so-
2
+
lution has been fully reduced to 1 with CB[8] (right). The
solutions were 0.1 mM 1 in water. CB[8] was added in approxi-
mately 1 equivalent.
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H830
Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
3
2
2
2
+
+
+
+
3+
2+
2+
2+
1
1
1
1
+ CB[8]
+ CB[8]
+ CB[7]
2
2
2
2
+ CB[8]
+ CB[8]
+ CB[7]
4
00 600 800 1000 1200
Wavelength /nm
4
00 600 800 1000 1200
Wavelength /nm
2
+
2+
Figure 6. Absorbance spectra of 1 and 2 with and without cucurbiturils in solution after reduction to the 2+ or 3+ states. (left) 1 (black), 1 in the presence of
2
+
3+
2+
an equivalent of CB[7] (blue), 1 in the presence of an equivalent of CB[8] (green), and 1 in the presence of an equivalent of CB[8] (red). (right) 2 (black),
2
+
2+ 3+
in the presence of an equivalent of CB[7] (blue), 2 in the presence of an equivalent of CB[8] (green), and 2 in the presence of an equivalent of CB[8] (red).
2
4
+
Figure 7a, a 30 μM solution of only 1 was fully reduced by bulk
decays with a single exponential, the time constant for oxidation of
2
+
2+
electrolysis to 1 under a stream of nitrogen. Upon fully reducing the
solution, the nitrogen stream and electrodes were removed from the
cell and then the cell was exposed to air. The visible light absorbance
of the solution was monitored by taking a spectrum every 15 seconds.
Figure 7b demonstrates the same experiment done using a 30 μM so-
lution of 1 that also contained one equivalent of CB[8]. In Figure 7a,
the absorbance (λmax = 536 nm) decayed quickly over time, resulting
finally in a featureless spectrum. In Figure 7b, the initial absorbance
peak at λ = 536 nm decayed quickly but a new peak at λ = 623 nm
grew in with an isosbestic point at λ = 572 nm. The absorbance at
λ = 623 nm then decayed slowly. The final spectrum in this case
was also featureless.
1
by O
2
without CB[8] was 41 ± 2 s while in the presence of
CB[8] the measured time constant was 206 ± 1 s. Figures 7d–7f show
2+
the same data for analogous experiments performed with 2 . In the
absence of CB[8], the decay time constant was 16 ± 1 s but in the
presence of CB[8], the decay time constant was 126 ± 5 s. The same
4
+
+
experiments performed with MV showed no substantial influence
of CB[8] on the time for complete color fade with no absorbance
maximum ever observed at λ = 623 nm (i.e. 21 ± 1 s vs 45 ± 2 s,
Supporting Information, Figure S8).
Discussion
In Figure 7c, the black squares plot the absorbance over time at λ =
The cumulative data speak to four interrelated points. First, the
presented results demonstrate the formation of an atypical, reduced
viologen dimer radical through synergistic linker preorganization
and supramolecular assembly. Second, the data suggest this species
is electronically and electrochemically distinct from previously
5
36 nm of the spectrum in Figure 7a. The red circles separately show
the absorbance at λ = 623 nm in Figure 7b. In contrast to the full color
2
+
fade of 1 without CB[8] that occurred within 90 s, the absorbance at
λ = 623 nm persisted approximately 15 minutes. Fitting these spectral
a)
b)
c)
1
1
0.5
0.4
0.3
0.2
0.1
0.0
0.4
0.2
0.0
+ CB[8]
0
0
.4
.2
0.0
5
00
600
700
500
600
700
0
500
1000
Wavelength /nm
Wavelength /nm
Time /s
d)
e)
f)
0
0
0
0
0
.7
.6
.5
.4
.3
0
0
0
0
.3
.2
.1
.0
2
2
0
0
.3
.2
+ CB[8]
0.1
.0
0
5
00
600
700
500
600
700
0
500
1000
Wavelength /nm
Wavelength /nm
Time /s
2
+
Figure 7. a) Time-dependent absorbance spectra for 1 in aqueous solution recorded at 15 s intervals over 25 min. b) Time-dependent absorbance spectra for
2
+
1
in same solution but with one equivalent of CB[8]. c) Time-dependence of the absorbance maxima in Figure 7a at λ = 536 nm and in Figure 7b at λ = 623 nm.
2
+
2+
d) Time-dependent absorbance spectra for 2 in aqueous solution. e) Time-dependent absorbance spectra for 2 in the same solution but with one equivalent of
CB[8]. f) Time-dependence of the absorbance maxima in Figure 7d at λ = 534 nm and in Figure 7e at λ = 622 nm.
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
H831
well-explored π-stacked viologens. Third, these radical viologen
species exhibit significantly greater stability in the presence of O rela-
tive to other forms of reduced viologens. Fourth, the reduced viologen
radicals shown here have some parallels to other forms of viologen-
based materials but are unique in their simplicity of preparation and
dynamic hierarchical design. These points are discussed below.
The reductions of 1 and 2 in the presence of cucurbituril com-
plexing agents show evidence for reduced viologen dimers unlike
previously described π-stacked bis-radicals. In this regard, the pre-
sented data are clear that such species are not favored just because
of complexation in general. The cumulative observations in the pres-
ence of CB[7] show no unusual behavior or properties outside what
can be inferred from the aforementioned bis-viologen radicals and
2
27,50
prior reports of viologen insertion/de-insertion in CB[7].
Simply,
the small size of the CB[7] cavity is not able to accommodate two
Intra- and Intermolecular factors necessary to form stabilized,
reduced viologen dimer.—Singly reduced viologens can form du-
plexes in solution with strong overlap of the π orbitals between the
27,48,50
stacked aromatics.
Rather, the data imply strongly that CB[8] is
uniquely suited because it can host two reduced, tethered, and stacked
viologen units simultaneously to promote the formation of a new type
of stabilized radical duplex. To be clear, the inclusion of two separate
52,54,58,63
two units, often referred to as ‘pimers’.
The formation of these
2
−1
1
duplexes can have significant association constants (∼10 M ) since
2
+
reduced MV equivalents in CB[8] showed no evidence of a stable
3+ radical or otherwise remarkable properties. Previously reported
reduction of a viologen dimer connected with a longer flexible linker
(hexyl) in the presence of CB[8] also showed no evidence of a sta-
the π-π stacking allows for stabilization of the individual viologen
55,64
radicals.
Studies have shown that this stabilization is mostly at-
tributable to dispersion and delocalization of the two SOMO orbitals
across all surfaces of the stacked viologen moieties.55,65–67 Generally,
3
55,64–67
ble 3+ radical state. Accordingly, the data here are clear that both
such interactions have been described as ‘pancake bonding.’
Such delocalized bis-radical species form readily when viologens are
dimer pre-organization and complexation within a sufficiently large
cavity are necessary to make the putative blue viologen dimer in the
3+ oxidation state.
6
1,68
reduced electrochemically in sufficiently high concentrations
or
when reduced viologen units are in forced close proximity. This
52
4
+
A scheme describing the elementary electrochemical and chemical
steps describing the formation of the monoradical for 1 in the pres-
ence of CB[8] is proposed in Figure 8. Presumably, the formation of
the corresponding monoradical of 2 is similar, albeit with additional
competing equilibria for conformations that are not productive toward
the 3+ state. That is, this dimer is less rigid and can adopt more con-
formations that disfavor pairing. In this figure, vertical transitions de-
note chemical steps constituting either conformation or complexation
latter scenario is clearly evidenced in the voltammetry of free 1
4
+
and 2 by the notable positive shift of the voltammetric response for
the reduction from the 4+ to 2+ states compared to the reduction of
2
+
•+
MV to MV . The greater positive shift in redox potential as well
as the clearer voltammetry observed with 1 as compared to 2 suggests
that the preorganization of direct π-π overlap observed in modeling
the rigid ortho-benzyl linkage increases the likelihood of stabilized
2
+
π-π stacking between radical viologen units of 1 . The equilibrium
−
changes while horizontal transitions represent 1e electrochemical
geometry and flexibility of the propyl linker provides comparatively
2
+
charge transfers. Based on the collected voltammetry and calculations
of conformer stability (Supporting Information, Section S3), the equi-
less probability that the two reduced viologen units of 2 are well
positioned for delocalization of the bis-radical. The collective voltam-
4
+
2
+
2+
librium form of 1 has the viologen units spread far apart, i.e. as a
metric, NMR, and absorbance data for 1 and 2 are wholly in line
4
+
1
,52,61,68
linear conformer with K ꢀ 1. This linear conformation of 1 binds
with the extant literature on π-stacked bis-radical ‘pimers’.
2
Figure 8. Proposed elementary chemical and electrochemical
steps involved in the reduction of 1 in the presence of CB[8].
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H832
Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
∗
Figure 9. a) The relative energies of molecular orbitals from DFT calculations (Spartan 16,B97D, 6–311+G ) showing the 3+ and 2+ states of 1 both without
and within CB[8]. b) Graphical depiction of the sHOMO of 1 in the 3+ oxidation state and the sHOMO of 1 in the 2+ oxidation state. c) The spin densities of 1
inside CB[8] in the 3+ and 2+ oxidation states are shown, with CB[8] omitted for clarity. Side and top views of the 1 are shown, with the blue regions highlighting
the regions where the radical density is highest.
in a 1:1 ratio with CB[8] and the complex has a disassociation constant
tively burying the radical between the two viologen units. Essentially
an identical sHOMO orbital was calculated for energy-minimized 2
−
7
3+
K
1
= 8.3 × 10 M as measured by ITC (Figure S9). The reduction
of each viologen unit on unbound 1 is ‘degenerate’, i.e. the standard
in CB[8] (Supporting Information, Figure S10).
3
+
potentials are equivalent and the reduction processes are independent
Thespindensitymapofthecomplexbetween 1 andCB[8]further
supports the point that the transferrable electron is not readily acces-
sible. The density of unpaired spin is mostly located on the internal
π-surface of the folded dimer (Figure 9c). Conversely, the sHOMO of
of each other, as evidenced by the narrow ꢁE
p
in the voltammetry
4
+
56
4+
of 1 alone. Based on the shift of the voltammetric wave for 1
2
+
relative to the voltammetric wave of MV , the value of K
3
>> 1, i.e.
2
+
2+
1
changes conformation to stabilize the viologen units by co-facial
the 1 dimer in CB[8] has a node running between the two viologen
2
+
pairing. In the presence of CB[8], we infer the paired form of 1 par-
titions readily into CB[8] with K > K since the voltammetric wave
at relatively fast scan rates (> 1 V s ) for the 4+/2+ condition is
units at the π-π interface (Figure 9b). This aspect can also be visual-
ized in the corresponding spin density map of the complex between 12
+
4
1
−
1
and CB[8] bis-radical which shows significant unpaired spin density
on both the internal lower energy singlet HOMO-1 (sHOMO–1) and
external (sHOMO) viologen surfaces. These characteristics are further
4
+
shifted toward an even more positive potential. Reduction of 1 in
−1
the presence of CB[8] at slower scan rates (< 0.5 Vs ) allows access
to the significantly stabilized 4+/3+ transition, for which the folded
69–72
consistent with the ‘pancake bonding’ concept.
We posit that the
3
+
is likely very tightly associated with CB[8], combining to a re-
1
uniform distribution of electron density on both surfaces of the violo-
2
+
2+
markable positive shift in the 1e- reduction wave. The large value of
is also responsible for the rate independence of oxidation observed
gens for the sHOMO orbital of 1 (and 2 ) results in good electronic
K
4
coupling with oxidants and corresponding fast electron transfer.
2
+
for 1 in the presence of CB[8]. The preorganized structure requires
3
+
no conformational rearrangement to access the stable 1 complex,
aa ln lds cs oa ns er aq tue es n. tial 2+/3+ and 3+/4+ oxidation waves are observed at
Stability in the presence of O .—Based on the preceding discus-
2
3
+
3+
sion, we propose that the resultant stability of 1 and 2 in air is
largely a consequence of slower charge-transfer kinetics. The observed
stability against oxidation of the radical 3+ state of 1 in CB[8] by O
2
is
Electronic structure of reduced, viologen monoradical.—The
presented absorbance spectra clearly indicate that the electronic struc-
ture of the 3+ state is very different than that of the bis-radical form or
even of a singly reduced methyl viologen in organic solvents.52,62 To
gain insight into this empirical observation, molecular orbital calcula-
in accord with the proposed electronic structure of this species. Since
the CB[8] cavities force the viologen units in close proximity to form a
3
+
2+
tions were performed on energy minimized structures of 1 and 1
encapsulated in CB[8]. Figure 9 shows computations of the energetics
for the 3+ radical form of 1 relative to the 2+ bis-radical. In free solu-
tion both species yield equilibrium geometries favoring a partially-
folded state that balances charge-repulsion with radical dispersion
forces. For 1, this geometry leads to minimal overlap and perturbation
N
N
N
N
N
N
N
N
3
+
of the 1 singlet highest occupied molecular orbital (sHOMO) and
nd
predicts electrochemical near-degeneracy of the 2 reduction to the
2
+ state (Figure 9a). In the presence of CB[8], both reduced species
show strong complexation inside the host cavity. The resultant equilib-
rium geometry of the dimer is completely folded with a π-π stacking
distance < 3.6 Å. In this conformation within CB[8], the sHOMO of
3
+
1
shows a strong and stabilizing perturbation due to delocalization
of the radical across both viologen moieties, with perfectly symmet-
ric constructive overlap at the π-π interface (Figures 9a, 9b). This
constructive (bonding) symmetry suggests a significant stabilization
3
+
3+
of the complex between 1 and CB[8], with a majority of the 1
Figure 10. Graphical depictions of (left) cyclobis(paraquat-p-phenylene) and
(right) cyclobis(paraquat-o-phenylene).
sHOMO orbital density existing between the aromatic groups, effec-
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
H833
stable pocket where the radical electron resides, this structure inhibits
electronic coupling with external oxidants. In effect, this transferrable
electron becomes significantly less accessible to oxidants like O and
2
hence the rate of oxidation of the viologen dimer back to the 4+ state
is slowed. The viologen rings (and the CB[8] unit itself) represent
tunneling barriers that inhibit transfer of that specific electron. For
oxidation of the 2+ state, the viologen rings do not act as tunneling
barriers since the electron density is explicitly located on them. As a
lar pre-organization and complexation of much simpler-to-synthesize
viologens circumvents the need to prepare such cyclophanes. More
work is needed to determine whether design improvements can be
made to further enhance the stability of reduced viologens.
Conclusions
This work demonstrates that it is possible to form a stable, reduced
viologen radical in water through a combination of intramolecular
and supramolecular assemblies. The hierarchical reduction of pre-
organized viologen moieties and encapsulation of this folded species
inside the cavity of the rigid macrocyclic host CB[8] accesses a confor-
mation that specifically stabilizes a monoradical, 3+ cation species.
This reduced radical possesses a distinct visible light absorbance pro-
file that is unlike any ‘pimer’. Instead, its optical properties bear strong
resemblance to bis-viologen cyclophanes with ortho-benzyl linkers.
Unlike such cyclophanes which are difficult to synthesize, the com-
pounds shown here can be synthesized in almost quantitative yields
and the supramolecular assemblies form readily in water upon elec-
trochemical reduction. These encapsulated monoradicals are signifi-
result, O
2
can readily oxidize the 2+ state even when the dimer re-
sides inside of CB[8]. A further contrast along these lines is clear when
considering the spin density map for the energy-minimized conforma-
tion of two MV in CB[8]. In this case, the ‘pimer’ adopts a twisted
+
conformation (Supporting Information, Figure S11).
The peculiar absorbance maxima in Figures 7c and 7f can now
also be readily understood. When the predominant reduced forms in
2
+
2+
solution are 1 and 2 , even with complexation these species will
rapidly oxidize in air and the corresponding absorbance signature will
3
+
3+
quickly vanish. However, within CB[8], the 1 and 2 forms are
stable. They will accumulate and grow in concentration until all of
2
+
2+
1
and 2 is consumed. At long times, the concentration of these
species will decrease based on the rate constant for oxidation with O
Since these constants are likely comparatively slow when protected
2
cantly more stable in the presence of O , highlighting a unique reac-
2
.
tivity relative to other viologen species. Molecular modeling analy-
ses suggest the stability of this monoradical species results from the
central location of the SOMO between the viologen rings. Due to
the ease of preparation, these monoradical compounds are amenable
to further spectroscopic studies that detail their electronic structure.
More generally, these results suggest new design motifs for viologens
in applications where storing reduced equivalents of charge such as
electrochromic windows, redox flow batteries, and solar fuel storage
systems is critical.
3
+
3+
by CB[8], the absorbance signals for 1 and 2 slowly decay.
Relation to other viologen materials.—Although this report is
the first explicit use of pre-organization and complexation principles
to stabilize a reduced viologen, there are two areas that have some
parallels. The first area is redox active polymers based on viologen
12–14
moieties.
In the context of charge-storing catholytes, an organic
polymer with pendant side chains consisting of the ortho-benzyl vi-
14
ologen dimers has been reported. Unlike other poly-viologens, this
particular polymer featured a reduced form that was unusually stable in
air. The origins of this stability were not explored since that particular
material was not easy to oxidize back electrochemically and accord-
ingly not deemed readily useful for rapid charge/discharge cycling in
redox flow batteries. However, given the results shown here, a likely
reason for the observation of a stable, reduced state could be that the
reduction of the pendant ortho-benzyl viologen dimers formed simi-
lar 3+ viologen monoradicals. Instead of a complexation agent that
provided structure around the dimer, we posit that the folds and pock-
ets within the polymer could have provided similar steric conditions
to force viologen units in close proximity. At present, it is not clear
whether a different polymer structure can be specifically designed to
enhance the stability of the 3+ state but still afford release of the re-
ducing equivalent when desired. The presented data provide a basis
for designing such material.
Acknowledgments
We thank Adam Urbach (Trinity University) for an initial sample
of CB[8], Kyle Tsai for help with HPLC, and Jordan Fantini for help-
ful discussions. S. M. acknowledge generous support from the De-
partment of Energy (DE-SC0006628). M. M. M. acknowledges the
National Science Foundation (NSF) for support through the Graduate
Research Fellowship Program. J. J. R. acknowledges the Donors of
The American Chemical Society Petroleum Research Fund (52993-
UR1).
ORCID
Molly M. MacInnes https://orcid.org/0000-0001-9513-7402
Stephen Maldonado https://orcid.org/0000-0002-2917-4851
Similarly, fully (dually) tethered viologen dimers with the proper
structure might also enable the 3+ monoradical state. Specifically,
References
3
+
cyclophane compounds can have similar structures as 1 in CB[8]
but because both ends are linked by phenyl groups rather than
complexation. The most famous cyclophane is cyclobis(paraquat-p-
phenylene), i.e. “blue-box”, and is widely employed in catenanes and
1
2
. C. L. Bird and A. T. Kuhn, Chemical Society Reviews, 10, 49 (1981).
. P. M. S. Monk, The Viologens: Physicochemical Properties, Synthesis and Applica-
tions of the Salts of 4,4’-Bipyridine, John Wiley & Sons Ltd., West Sussex, England
(
1998).
73–75
rotaxanes.
The space between viologen units in para-“blue-box”
3. W. S. Jeon, A. Y. Ziganshina, J. W. Lee, Y. H. Ko, J. K. Kang, C. Lee, and K. Kim,
Angewandte Chemie, 115, 4231 (2003).
. Y. Song, X. Huang, H. Hua, and Q. Wang, Dyes and Pigments, 137, 229 (2017).
. H. R. Wessels, C. Slebodnick, and H. W. Gibson, Journal of the American Chemical
Society, 140, 7358 (2018).
6. B. Zhang, Y. Dong, J. Li, Y. Yu, C. Li, and L. Cao, Chinese Journal of Chemistry, 37,
69 (2019).
compounds is large enough to accommodate guest molecules but con-
versely precludes the possibility of forming a tight pocket to stabi-
lize the 3+ monoradical. Accordingly, these compounds have never
shown a stabilized reduced state. The ortho form of this cyclophane
is less studied. The small space between viologen units precludes
molecular guests and this molecule is exceedingly difficult to syn-
thesize with high yields. Its synthesis requires inordinately long re-
action times because of solubility incompatibilities of the precursors
4
5
2
7
. D. Bonneh-Barkay, W. J. Langston, and D. A. D. Monte, Antioxidants & Redox
Signaling, 7, 649 (2005).
8
. R. Castro, E. Moyano, and M. Galceran, Journal of Chromatography A, 914, 111
(2001).
76,77
9. M. Maroni, C. Colosio, A. Ferioli, and A. Fait, Toxicology, 143, 5 (2000).
0. L. Striepe and T. Baumgartner, Chemistry (2017).
1. T. Liu, X. Wei, Z. Nie, V. Sprenkle, and W. Wang, Advanced Energy Materials, 6
2016).
12. S. Sen, J. Saraidaridis, S. Y. Kim, and G. T. Palmore, ACS Appl. Mater. Interfaces, 5,
825 (2013).
3. G. Nagarjuna, J. Hui, K. J. Cheng, T. Lichtenstein, M. Shen, J. S. Moore, and
J. Rodriguez-Lopez, J. Am. Chem. Soc., 136, 16309 (2014).
4. M. Burgess, E. Chénard, K. Hernández-Burgos, G. Nagarjuna, R. S. Assary, J. Hui,
J. S. Moore, and J. Rodríguez-López, Chemistry of Materials, 28, 7362 (2016).
and low concentrations to prevent polymerization pathways.
The
1
1
available data on the electrochemical properties of this compound are
scarce, with only a few reports showing some optical data upon reduc-
(
63,78–80
tion or photoexcitation.
Still, these scant data are in accord with
7
the data shown here and suggest the electrochemical properties (and
stability in air) of ortho-“blue-box” could be similar. Experimental
verification of this hypothesis would prove insightful. However, the
larger take away of this work is that the combination of intramolecu-
1
1
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H834
Journal of The Electrochemical Society, 166 (15) H825-H834 (2019)
1
5. T. Janoschka, C. Friebe, M. D. Hager, N. Martin, and U. S. Schubert, ChemistryOpen,
, 216 (2017).
52. P. Neta, M.-C. Richoux, and A. Harriman, Journal of the Chemical Society, Faraday
6
Transactions 2: Molecular and Chemical Physics, 81, 1427 (1985).
53. J.-M. Lü, S. V. Rosokha, and J. K. Kochi, Journal of the American Chemical Society,
125, 12161 (2003).
54. A. Deronzier, B. Galland, and M. Vieira, Nouv. J. Chim., 6, 97 (1982).
55. M. R. Geraskina, A. S. Dutton, M. J. Juetten, S. A. Wood, and A. H. Winter,
Angewandte Chemie, 129, 9563 (2017).
56. D. E. Richardson and H. Taube, Inorganic Chemistry, 20, 1278 (1981).
57. P. Neta, M.-C. Richoux, and A. Harriman, J. Chem. Soc., Faraday Trans., 2, 81, 1427
(1985).
1
1
1
6. R. J. Mortimer, Electrochimica Acta, 44, 2971 (1999).
7. R. J. Mortimer, Annual Review of Materials Research, 41, 241 (2011).
8. D. C. Bookbinder and M. S. Wrighton, Journal of the Electrochemical Society, 130,
1080 (1983).
1
9. M. S. Simon and P. T. Moore, Journal of Polymer Science: Polymer Chemistry
Edition, 13, 1 (1975).
0. S. D. Tilley, S. Lany, and R. Van de Krol, Advances in Photoelectrochemical Water
2
Splitting: Theory, Experiment and Systems Analysis, Royal Society of Chemistry
(
2018).
58. F. Masaoki and N. Shun-ichi, Bulletin of the Chemical Society of Japan, 55, 513
(1982).
2
1. B. P. Sullivan, K. Krist, and H. Guard, Electrochemical and electrocatalytic reactions
of carbon dioxide, Elsevier (2012).
2. R. N. F. Thorneley, Biochimica et Biophysica Acta, 333, 487 (1974).
3. R. E. V. D. Leest, Electroanalytical Chemistry and Interfacial Electrochemistry, 43,
59. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamental and Applica-
tions, John Wiley & Sons, Hoboken, NJ (2001).
60. S. R. Peerannawar, V. V. Gobre, and S. P. Gejji, Computational and Theoretical
2
2
251 (1973).
Chemistry, 983, 16 (2012).
2
4. J. A. Farrington, M. Ebert, and E. J. Land, Journal of the Chemical Society, Faraday
Transactions 1: Physical Chemistry in Condensed Phases, 74, 665 (1978).
5. K. Hoshino, H. Sasaki, K. Suga, and T. Saji, Bull. Chem. Soc. Jpn., 60, 1521 (1987).
6. A. E. Kaifer and A. J. Bard, The Journal of Physical Chemistry, 89, 4876 (1985).
7. H.-J. Kim, W. S. Jeon, H. K. Young, and K. Kim, PNAS, 99, 5007 (2002).
8. W. S. Jeon, H.-J. Kim, C. Lee, and K. Kim, Chem. Commun., 1828 (2002).
9. J. Kostela, M. Elmgren, P. Hansson, and M. Almgren, Journal of Electroanalytical
Chemistry, 536, 97 (2002).
61. J. Mizuguchi and H. Karfunkel, Berichte der Bunsengesellschaft für physikalische
Chemie, 97, 1466 (1993).
62. C. Lee, Y. M. Lee, M. S. Moon, S. H. Park, J. W. Park, K. G. Kim, and S.-J. Jeon,
Journal of Electroanalytical Chemistry, 416, 139 (1996).
63. O. Enea, P. Crouigneau, J. Moser, M. Grätzel, and S. Huenig, New journal of chem-
istry, 15, 267 (1991).
2
2
2
2
2
64. D. W. Zhang, J. Tian, L. Chen, L. Zhang, and Z. T. Li, Chemistry–An Asian Journal,
10, 56 (2015).
3
3
3
3
3
3
0. R. Kaliappan, Y. Ling, A. E. Kaifer, and V. Ramamurthy, Langmuir, 25, 8982 (2009).
1. C.-W. Lee, O. M.-K., and J.-M. Jang, Langmuir, 9, 1934 (1993).
2. J. W. Park, S. H. Ko, and J.-Y. Park, Bull. Korean Chem. Soc., 13, 259 (1992).
3. A. E. Kaifer, W. Li, and S. Yi, Israel Journal of Chemistry, 51, 496 (2011).
4. L. Pospisil, R. Fuoco, and P. Papoff, J. Electroanal. Chem., 256, 83 (1988).
5. K. W. Willman and R. W. Murray, Journal of Electroanalytical Chemistry and
Interfacial Electrochemistry, 133, 211 (1982).
65. J. Huang and M. Kertesz, Journal of the American Chemical Society, 129, 1634
(2007).
66. K. Goto, T. Kubo, K. Yamamoto, K. Nakasuji, K. Sato, D. Shiomi, T. Takui,
M. Kubota, T. Kobayashi, and K. Yakusi, Journal of the American Chemical Society,
121, 1619 (1999).
67. K. Fukui, K. Sato, D. Shiomi, T. Takui, K. Itoh, K. Gotoh, T. Kubo, K. Yamamoto,
K. Nakasuji, and A. Naito, Synthetic metals, 103, 2257 (1999).
68. E. M. Kosower and J. L. Cotter, Journal of the American Chemical Society, 86, 5524
(1964).
69. J. L. Zafra, L. Qiu, N. Yanai, T. Mori, M. Nakano, M. P. Alvarez, J. T. L. Navarrete,
C. J. Gómez-García, M. Kertesz, K. Takimiya, and J. Casado, Angewandte Chemie
International Edition, 55, 14563 (2016).
36. X. Liu, K. G. Neoh, and E. T. Kang, Macromolecules, 36, 8361 (2003).
37. L. J. Kepley and A. J. Bard, Journal of The Electrochemical Society, 142, 4129 (1995).
38. D.-S. Guo, S. Chen, H. Qian, H.-Q. Zhang, and Y. Liu, Chemical Communications,
46, 2620 (2010).
39. R. J. Mortimer, A. L. Dyer, and J. R. Reynolds, Displays, 27, 2 (2006).
40. T. Nakahira and M. Graetzel, The Journal of Physical Chemistry, 88, 4006 (1984).
41. H. Sato and T. Tamamura, Journal of Applied Polymer Science, 24, 2075 (1979).
42. N. Oyama, T. Ohsaka, H. Yamamoto, and M. Kaneko, The Journal of Physical
Chemistry, 90, 3850 (1986).
70. K. E. Preuss, Polyhedron, 79, 1 (2014).
71. L. L. Miller and K. R. Mann, Accounts of Chemical Research, 29, 417 (1996).
72. M. Kertesz, Chemistry – A European Journal, 25, 400 (2019).
73. B. Odell, M. V. Reddington, A. M. Z. Slawin, N. Spencer, J. F. Stoddart, and
D. J. Williams, Angewandte Chemie International Edition in English, 27, 1547
(1988).
4
4
3. E. Dalton and R. W. Murray, The Journal of Physical Chemistry, 95, 6383 (1991).
4. A. Factor and G. Heinsohn, Journal of Polymer Science Part B: Polymer Letters, 9,
2
89 (1971).
5. M. Tomokazu, K. Tetsuya, A. Uichi, and O. Tetsuo, Chemistry Letters, 14, 1825
1985).
74. A. Trabolsi, N. Khashab, A. C. Fahrenbach, D. C. Friedman, M. T. Colvin, K. K. Cotí,
D. Benítez, E. Tkatchouk, J.-C. Olsen, and M. E. Belowich, Nature chemistry, 2, 42
(2010).
4
(
4
4
6. A. Yasuda, H. Mori, and J. Seto, Journal of Applied Electrochemistry, 17, 567 (1987).
7. V. Annibaldi and C. B. Breslin, Journal of Electroanalytical Chemistry, 832, 399
75. C. P. Collier, G. Mattersteig, E. W. Wong, Y. Luo, K. Beverly, J. Sampaio,
F. M. Raymo, J. F. Stoddart, and J. R. Heath, Science, 289, 1172 (2000).
76. W. Geuder, S. Hünig, and A. Suchy, Tetrahedron, 42, 1665 (1986).
77. W. Geuder, S. Hünig, and A. Suchy, Angewandte Chemie, 95, 501 (1983).
78. J. Moser and M. Graetzel, Journal of the American Chemical Society, 105, 6547
(1983).
(2019).
4
4
8. W. Ong, M. Gómez-Kaifer, and A. E. Kaifer, Organic Letters, 4, 1791 (2002).
9. S. Sun, R. Zhang, S. Andersson, J. Pan, D. Zou, B. Åkermark, and L. Sun, The Journal
of Physical Chemistry B, 111, 13357 (2007).
5
0. A. T. Buck, J. T. Paletta, S. A. Khindurangala, C. L. Beck, and A. H. Winter, Journal
of the American Chemical Society, 135, 10594 (2013).
79. X. Sun and Y.-K. Yang, Journal of the Chemical Society, Perkin Transactions, 2, 225
(1996).
5
1. M. J. Blandamer, J. A. Brivati, M. F. Fox, M. C. R. Symons, and G. S. P. Verman,
Trans. Faraday Soc., 63, 1850 (1967).
80. Y.-K. Yang and X. Sun, Journal of Photopolymer Science and Technology, 7, 193
(1994).
Downloaded on 2019-12-07 to IP 130.238.7.40 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).