The Viologen Cation Radical Pimer: A Case of Dispersion-Driven Bonding

Article abstract of DOI:10.1002/anie.201704959

The π bonds between organic radicals have generated excitement as an orthogonal interaction for designing self-assembling architectures in water. A systematic investigation of the effect of the viologen cation radical structure on the strength and nature of the pimer bond is provided. A striking and unexpected feature of this π bond is that the bond strength is unchanged by substitution with electron-donating groups or withdrawing groups or with increased conjugation. Furthermore, the interaction is undiminished by sterically bulky N-alkyl groups. Theoretical modeling indicates that strong dispersion forces dominate the interaction between the radicals, rationalizing the insensitivity of the bonding interaction to substituents: The stacking of polarizable π radicals leads to attractive dispersion forces in excess of typical dispersion interactions of small molecules and helps overcome the Coulombic repulsion of bringing two cationic species into contact.

Full text of DOI:10.1002/anie.201704959

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: The Viologen Cation Radical Pimer: A Case of Dispersion-Driven
Bonding
Authors: Margarita Geraskina, Andrew S. Dutton, Mark Juetten,
Samuel Wood, and Arthur Winter
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201704959
Angew. Chem. 10.1002/ange.201704959
Link to VoR: http://dx.doi.org/10.1002/anie.201704959
http://dx.doi.org/10.1002/ange.201704959

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
The Viologen Cation Radical Pimer: A Case of Dispersion-Driven
Bonding
Margarita R. Geraskina, Andrew S. Dutton, Mark J. Juetten, Samuel A. Wood,[b] and Arthur H.
Winter*[a]
Abstract: Pi bonds between organic radicals have generated
excitement as an orthogonal interaction for designing self-
assembling architectures in water. Here, we provide a systematic
investigation of the effect of the viologen cation radical structure on
the strength and nature of the pimer bond. A striking and unexpected
feature of this pi bond is that the bond strength is unchanged by
substitution with electron-donating groups or withdrawing groups or
with increased conjugation. Furthermore, the interaction is
undiminished by sterically bulky N-alkyl groups. Theoretical
modelling indicates that strong dispersion forces dominate the
interaction between the radicals, rationalizing the insensitivity of the
bonding interaction to substituents: The stacking of polarizable pi
radicals leads to attractive dispersion forces in excess of typical
dispersion interactions of small molecules and helps overcome the
Coulombic repulsion of bringing two cationic species into contact.
The ability of open-shell species to form weak self-association
complexes has attracted attention for a variety of applications. In
particular, stabilized organic pi radicals can associate to form
van der Waals complexes, sigma dimers, or multi-centered pi-
bonded dimers. The pi bond in particular is a fascinating one for
its unusual multicenter bonding pattern that brings atoms closer
than the van der Waals radii but longer than a conventional
sigma bond (>2.8 Å), while straddling the knife edge between
van der Waals interactions and conventional chemical bonds in
strength and properties.
Viologen cation radicals are a well-known class of thermally
stable organic radicals that can be prepared simply via chemical
or electrochemical reduction of synthetically-accessible
bipyridinium dications.[1] Within polar solvents that can screen
the intermolecular Coulombic repulsion between like ions, these
radicals dimerize to form such pi-bonded dimers, referred to
whimsically as ‘pimers’. Pi dimerization has been investigated
for related pi radicals, such as neutral pi radicals (e.g.
phenalenyl radicals),[2] pi anion radicals (e.g. tetracyanoethylene
anion radical),[3] and cation radicals (e.g. biphenylene cation
radicals),[4] but there are no detailed investigations into the effect
of the radical structure on the strength and nature of the
viologen cation radical pi bond. The ability to switch this
interaction on and off using redox control make this interaction a
versatile one for applications such as molecular data storage
systems,[5] photochromic materials,[6] molecular machines,
electrochromic devices,[7] spintronics,[8] and organic spin
crossover materials.[9]
Figure 1. Library of viologen dications 1-19 we synthesized as precursors to
the viologen cation radicals.
[a]
[b]
Dr. A. Winter
Department of Chemistry
Iowa State University
Dilution binding studies. To better understand the nature of
this viologen cation radical pi bond, we synthesized and
performed binding studies with the one-electron reduced cation
radical of the three major groups of viologen derivatives depicted
in Figure 1: N,N’-disubstituted viologens (1-12), N,N’-
disubstituted core-substituted viologens (13-16), and N-mono-
substituted viologens (17-19). To compare the strength of the pi
interactions, association constants were determined for each
radical cation by performing dilution binding studies in buffered
water solutions at pH=7.0. The binding isotherms determined by
Ames, IA, 50011, US
E-mail: winter@iastate.edu
M. Geraskina, A. Dutton, M. Juetten, S. Wood
Department of Chemistry
Iowa State University
Ames, IA, 50011, US
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
Table 1. Apparent association constants for 1-18 determined by UV-Vis
spectroscopy and EPR spectroscopy in aqueous solution (298 K, average
values of 3 independent runs).
UV-Vis or EPR spectroscopy were fit to obtain to association
constants (insets in Figures 2 and 3; see ESI for details).
UV-Vis and EPR titrations rule out disproportionation
reactions and sigma dimers. We determined the apparent
association constants using both UV-Vis dilution studies
following the pi dimer, which has a characteristic absorption
band in the IR where the free radical cation does not absorb,
and electron paramagnetic resonance (EPR) spectroscopy,
following the monomeric free radical. We used both methods to
UV-Vis
EPR
Ka (pH 7.0, M-1)
Ka (pH 7.0, M-1)
1
2
1.3 * 103
2.3 * 103
5.0 * 103
2.8 * 103
6.0 * 104
1.7 * 104
1.2 * 103
2.2 * 103
3.4 * 103
5.3 * 103
3.6 * 102
9.0 * 102
9.1 * 103
1.8 * 104
3.7 * 103
1.5 * 102
3.9 * 101
2.9 * 101
1.2 * 103
7.4 * 102
2.3 * 102
1.4 * 103
3.1 * 103
n.d.
rule out
a hypothetical disproportionation reaction as an
additional contributor to the radical—dimer equilibrium, as well
as possible sigma dimers. (See Scheme 1). If such sigma
dimers or
a
disproportionation reaction were significant
contributors to the aqueous equilibrium, we would obtain
different apparent binding constants by following the EPR-active
radical than obtaining the binding constants following the growth
of the EPR-inactive pi dimer. As can be seen in Table 1, while
there is some variation between the values determined by the
two methods, no systematic variation can be seen in the
apparent association constants determined by EPR or UV-Vis
spectroscopy. These experiments rule out major contributions to
the room temperature aqueous equilibrium by these species.
3
4
5
6
7
n.d.
8
n.d.
9
n.d.
10
11
12a
13
14
15
16
17
18
n.d.
Scheme 1. Considered viologen cation radical binding equilibria and
hypothetical disproportionation reaction. Computations (and EPR hyperfine
coupling) indicate that the charge and spin of the viologen radical cation are
delocalized over both rings (which are coplanar), but only one canonical
structure is depicted for ease of notation.
n.d.
Figure 3 shows the UV-Vis spectra of solutions of the reduced
dications for representative viologen derivatives. As can be
seen in all spectra, at low concentrations only a band at ~600
nm, attributable to the monomeric radical, is observed, while at
higher concentrations a broad absorbance in the near-infrared
grows in that can be attributable to the absorbance of the pi
dimer.
Adding donating (e.g. 14, 15) or withdrawing groups (13) to
the core rings has little measurable effect on the binding affinity,
nor does increasing the conjugation with additional aryl rings,
substituted with either strong electron-donating (8) or
withdrawing groups (7, 10, 11). Mono-substituted viologens 17-
19 exhibit complex behavior, which is detailed in the ESI.
Another unexpected result was that the radical cation of 4,
despite its bulky isopropyl groups, has an association constant
unchanged from that of the radical cation of the dimethyl
derivative of 1. Viologens substituted with benzyl groups (5, 6)
have slightly higher affinity, which can be attributed to the
hydrophobic effect. With two aryl groups, each containing two
powerful withdrawing nitro groups (12a) the radical cations show
a slight decrease of binding affinity, but to a first approximation,
n.d.
1.1 * 104
2.0 * 103
2.3 * 103
n.d.
n.d.
n.d.
substituents have
a negligible effect on the strength of
pimerization of viologen cation radicals.
N.D. = Not determined due to poor solubility/stability of the formed radical
(see ESI for specific discussion).
This article is protected by copyright. All rights reserved.

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
The first-contact van der Waals (vdW) dimers are probably
spin-unpaired and indistinguishable from the free radicals by
EPR.
Our EPR experiments are blind to any equilibrium
between the free radical and the first-contact vdW dimer if the
spins are unaltered. The UV-Vis experiments are also blind
unless the vDW dimer has a unique absorption band, which we
do not observe.
Likely, the vdW dimer is a transient
intermediate separating the pimer from the free radicals that is
not significantly thermally populated at equilibrium.
Viologen cation radicals do not prefer social self-sorting.
To test whether viologen cation radicals prefer to bind to
themselves (narcissistic self-sorting) or to different cation
radicals (social self-sorting), we mixed equal amounts of hetero-
substituted viologens containg the p-methoxyphenyl donating
group 8 and the electron withdrawing group p-nitrophenyl 10 (at
pH 7), and determined an apparent dilution binding constant.
Individual solutions of the same concentration have association
constant value of 2.2 × 103 M-1 and 5.3 × 103 M-1 for compound 8
and 10, respectively. However, fits of the binding isotherm for
the mixed species was the average value (3.7 × 103 M-1) of the
binding constant for the self-association, indicating that the
different radicals show no energetic preference to engage in
social self-sorting.
Computational studies indicate large dispersion forces
dominate the pi bond. It is virtually a truism of discussions of
bonding that structure influences bond strength from electronic
and steric effects, so why is this pi bond insensitive to
substituent effects? To evaluate the nature of the pimer
interaction, we turned to computational analysis. First we
attempted to identify
a computational method that could
Figure 2. Example of EPR binding studies in pH 7 buffer solution for selected
compounds (top) and their nonlinear fitting curve the dimer concentration vs
total concentration of viologen derivative (bottom): (a) 1 (100 - 4000 µM). (b) 2
(100 - 4000 µM). (c) 4 (100 - 4000 µM). (d) 3 (100 - 4000 µM).
reproduce the binding free energy of the parent methyl viologen
radical cation. As can be seen Figure 4, all DFT methods
lacking dispersion corrections give nonsensical predictions for
the binding free energy.
All these methods predict the
pimerization to be unfavorable by 17 kcal/mol or more. Without
dispersion corrections, the Coulombic repulsion dominates and
no bonding is predicted.
Only functionals including recently-developed dispersion
corrections accurately reproduce the binding free energy.
Comparison of B98 and B97D is informative, as these methods
represent the same functional but B97D includes a dispersion
corrected energy, shows a difference in binding free energy of
~20 kcal/mol! The B97D method exactly reproduces the
experimental binding free energy with the SMD aqueous
solvation model, so we used this method for all future
computations.
These calculations suggest that dispersion
forces play a major role in stabilizing the pi dimer.
Deviation from expt.
DGaq,298 kcal mol-1
Dimerization Structure 1
6-31+G(d,p)
25
20
15
10
5
cc-pVDZ
0
!5
!10
Figure 4. Similar to Schreiner, et al.,[10] absolute deviations from the averaged
experimental binding energies determined via UV-Vis and EPR at pH 7 and
pH 9.6 (4.0 kcal/mol) for 1 computed at various levels of theory. Red bars
indicate cc-pVDZ basis set and gold bars 6-31+G(d,p).
Figure 3. Representative UV-Vis dilution binding studies in pH 7 buffer
solution: (a) 1 (20 - 1000 µM). (b) 10 (50 - 600 µM). (c) 13. (d) 8. (e) 11. (f) 12a.
Inset: nonlinear fitting curve at maximum absorbance of the dimer vs total
concentration of the solution. g) cuvette pictures of 100 μM solutions (pH 7) of
(left to right): 1, 6, 9, 12a.
This article is protected by copyright. All rights reserved.

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
wavefunction into more familiar Lewis-type canonical
hyperconjugative bonding interactions. The sum of all the orbital
interactions between the two viologen cation radical rings is 5.0
kcal/mol using this method. Thus, the covalent interaction can
be estimated as ~4-5 kcal/mol. Given that the experimental
binding free energy is ca. -4 kcal/mol, the pimer would not be
formed without this covalent character. But the dispersion
correction is 3-4 times as large and represent a much larger
effect.
What about covalent interactions? Inspection of the HOMO and
LUMO indicate that there is an interaction of the pi SOMOs to
generate the unusual pi bond from stacked overlap of the SOMO
pi orbitals (Figure 5), which explains the closed-shell singlet
nature of the pimer. How much is this covalent interaction
worth?
To determine the magnitude of the covalent interaction, we
evaluated two different computational probes.
First, we
computed the singlet-triplet gap at the singlet geometry, which is
-4.6 kcal/mol in favor of the singlet state. Inspection of the
SOMOs for the triplet state shows that the two radical electrons
are disjoint, with one SOMO on one of the viologen radical
The pi bond has a flat rotational potential energy surface,
but large barriers for bond compression to make sigma
dimers. To evaluate the directionality of the pi bond, PES scans
cations and the other SOMO on the other viologen cation radical. were computed for the parent methyl viologen radical cation
Thus, the exchange integral should be vanishingly small, and
dimer. (See Figure 5a). These scans indicate several minima
on an extremely flat torsional potential energy surface. This flat
torsional PES stands in sharp contrast to the neutral phenalenyl
pimer, which has preferred torsional angles and large barriers
for rotation.[11] An unusual feature of these radical interactions is
their dynamic, fluxional nature. Assuming a diffusion-limited rate
constant for formation of the radical dimer, the off rate constant
can be estimated as ~106 sec-1. Thus, rapid flux on the
microsecond time scale, in addition to torsional rotations on an
even faster timescale, are anticipated. Consequently, it is
perhaps misleading to describe these radical interactions with
discrete geometries due to their highly dynamic nature.
Furthermore, the flat torsional PES rationalizes the insensitivity
of the pi bond strength to sterically bulky N-alkylation
substituents, as the molecule can rotate to a canted or
perpendicular geometry with little energetic cost.
the singlet-triplet gap should represent
approximation of the covalent bonding energy.
a
reasonable
With regard to the directionality of the bond along the pi
bonding axis, PES scans were performed from the varying
possible sigma dimers (Figure 6). The sigma dimers considered
were the 4,4’ sigma dimer, the 2,4’ sigma dimer, and the 2,2’
sigma dimer. These sigma dimers are minima on the potential
energy surface but are much higher in energy than the pimer
minimum and are not anticipated to be populated at room
temperature, consistent with our experiments described above.
Thus, the viologen cation radical pimer has little torsional
angular dependence, but a significant distance dependence on
the pi axis.
Figure 6. Relaxed scans of the sigma dimerization computed with B97D/6-
31+G(d,p) in the SMD aqueous model. The 4,4’ sigma dimer opens up to a
local pimer minimum that is not the global minimum.
Figure 5. (a) Relaxed scan of the torsional potential surface of methyl viologen
radical cation, 1, (B97D/6-31+G(d,p), SMD aqueous solvation). Minima can
be seen at parallel, canted, and perpendicular geometries. (b) HOMO and (c)
LUMO shown for the optimized structure of 1.
Conclusion. In conclusion, we have systematically
investigated the influence of substituents on the pimerization of
viologen cation radicals and found that the strength of the pi
dimer is essentially insensitive to substituents. London
dispersion forces are generally considered to be weak forces in
As another evaluation of the magnitude of the covalent
interaction, we used Weinhold's Natural Bond Orbital (NBO)
analysis. NBO analysis attempts to deconvolute the
This article is protected by copyright. All rights reserved.

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
Uchida, T. Kubo, M. Kertesz, Journal of the American Chemical
Society 2014, 136, 18009-18022.
a. J. J. Novoa, P. Lafuente, R. E. Del Sesto, J. S. Miller,
Angewandte Chemie International Edition 2001, 40, 2540-2545; b.
J. Jakowski, J. Simons, Journal of the American Chemical Society
solution for small molecules, but the recent inclusion of
dispersion corrections into DFT methods has led to a re-
evaluation of their importance, including for a number of
unexpected bonding interactions.[12] All molecular complexes will
experience van der Waals forces, but here the exceptionally
polarizable pi radicals lead to attractive London dispersion
forces in excess of what would normally be anticipated for
molecules of this size, which leads to a dispersion-dominated pi
bond that helps overcome the cation-cation Coulombic repulsion
of bringing two charged species into union. This suggests that
neutral pi radicals (or other highly polarizable species) could
make ultra-strong dispersion bonds, which we are currently
pursuing.
[3]
[4]
2003, 125, 16089-16096; c.
J.-F. Penneau, L. L. Miller,
Angewandte Chemie International Edition in English 1991, 30,
986-987.
J.-M. Lü, S. V. Rosokha, J. K. Kochi, Journal of the American
Chemical Society 2003, 125, 12161-12171.
[5]
[6]
M. Graetzel, Accounts of Chemical Research 1981, 14, 376-384.
a. N. Mercier, European Journal of Inorganic Chemistry 2013,
2013, 19-31; b. D. Meyerstein, J. Rabani, M. S. Matheson, D.
Meisel, The Journal of Physical Chemistry 1978, 82, 1879-1885.
a. F. Dickert, Angewandte Chemie (International ed.) 1999, 38,
2456-2457; b. P. M. S. Monk, Journal of Electroanalytical
Chemistry 1997, 432, 175-179; c. K. Sato, T. Nakahodo, H.
Fujihara, Chemical Communications 2011, 47, 10067-10069; d. D.
M. DeLongchamp, M. Kastantin, P. T. Hammond, Chemistry of
Materials 2003, 15, 1575-1586; e. V. Jain, M. Khiterer, R.
Montazami, H. M. Yochum, K. J. Shea, J. R. Heflin, ACS Applied
Materials & Interfaces 2009, 1, 83-89; f. R. Nakajima, Y. Yamada,
T. Komatsu, K. Murashiro, T. Saji, K. Hoshino, RSC Advances
2012, 2, 4377-4381.
[7]
Acknowledgements
Financial support from NSF (CHE-1464956) and ACS PRF
(PRF55820-ND4) are gratefully acknowledged. We thank Sarah
Cady and the Iowa State Chemical Instrumentation Facility for
assistance with the EPR measurements.
[8]
[9]
a. S. A. Wolf, D. D. Awschalom, R. A. Buhrman, J. M. Daughton,
S. von Molnár, M. L. Roukes, A. Y. Chtchelkanova, D. M. Treger,
Science 2001, 294, 1488-1495; b. I. Žutić, J. Fabian, S. Das
Sarma, Reviews of Modern Physics 2004, 76, 323-410.
a.
M. J. Juetten, A. T. Buck, A. H. Winter, Chemical
Keywords: viologen cation radical • London dispersion • stable
free radicals • non-covalent binding • k
Communications 2015, 51, 5516-5519; b. A. T. Buck, J. T.
Paletta, S. A. Khindurangala, C. L. Beck, A. H. Winter, Journal of
the American Chemical Society 2013, 135, 10594-10597; c. M. R.
Geraskina, A. T. Buck, A. H. Winter, The Journal of Organic
Chemistry 2014, 79, 7723-7727.
A. A. Fokin, L. V. Chernish, P. A. Gunchenko, E. Y. Tikhonchuk, H.
Hausmann, M. Serafin, J. E. P. Dahl, R. M. K. Carlson, P. R.
Schreiner, Journal of the American Chemical Society 2012, 134,
13641-13650.
Z.-h. Cui, H. Lischka, H. Z. Beneberu, M. Kertesz, Journal of the
American Chemical Society 2014, 136, 5539-5542.
J. P. Wagner, P. R. Schreiner, Angewandte Chemie International
Edition 2015, 54, 12274-12296.
[1]
[2]
P. M. S. Monk, The viologens: physicochemical properties,
synthesis, and applications of the salts of 4,4'-bipyridine, Wiley,
1998.
[10]
a. D. Small, V. Zaitsev, Y. Jung, S. V. Rosokha, M. Head-Gordon,
J. K. Kochi, Journal of the American Chemical Society 2004, 126,
13850-13858; b. S. Suzuki, Y. Morita, K. Fukui, K. Sato, D.
Shiomi, T. Takui, K. Nakasuji, Journal of the American Chemical
Society 2006, 128, 2530-2531; c. 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, J. Casado, Angewandte Chemie
International Edition 2016, 55, 14563-14568; d. Z. Mou, K.
[11]
[12]
This article is protected by copyright. All rights reserved.

10.1002/anie.201704959
Angewandte Chemie International Edition
COMMUNICATION
COMMUNICATION
A Piece of Pimer. Experimental and
computational investigations of the
pimer bond between viologen cation
radicals show that the bonding
Margarita R. Geraskina, Andrew S.
Dutton, Mark J. Juetten, Samuel A.
Wood, Arthur H. Winter*
Page No. – Page No.
interaction is dominated by
exceptionally strong dispersion forces
from ultra-polarizable pi radicals.
The Viologen Cation Radical Pimer: A
Case of Dispersion-Driven Bonding
This article is protected by copyright. All rights reserved.