Single-step versus stepwise two-electron reduction of polyarylpyridiniums: Insights from the steric switching of redox potential compression

Article abstract of DOI:10.1021/ja210024y

Contrary to 4,4′-dipyridinium (i.e., archetypal methyl viologen), which is reduced by two single-electron transfers (stepwise reduction), the 4,1′-dipyridinium isomer (so-called "head-to-tail" isomer) undergoes two electron transfers at apparently the same potential (single-step reduction). A combined theoretical and experimental study has been undertaken to establish that the latter electrochemical behavior, also observed for other polyarylpyridinium electrophores, is due to potential compression originating in a large structural rearrangement. Three series of branched expanded pyridiniums (EPs) were prepared: N-aryl-2,4,6-triphenylpyridiniums (Ar-TP), N-aryl-2,3,4,5,6-pentaphenylpyridiniums (Ar-XP), and N-aryl-3,5-dimethyl-2,4,6- triphenylpyridinium (Ar-DMTP). The intramolecular steric strain was tuned via N-pyridinio aryl group (Ar) phenyl (Ph), 4-pyridyl (Py), and 4-pyridylium (qPy) and their bulky 3,5-dimethyl counterparts, xylyl (Xy), lutidyl (Lu), and lutidylium (qLu), respectively. Ferrocenyl subunits as internal redox references were covalently appended to representative electrophores in order to count the electrons involved in EP-centered reduction processes. Depending on the steric constraint around the N-pyridinio site, the two-electron reduction is single-step (Ar = Ph, Py, qPy) or stepwise (Ar = Xy, Lu, qLu). This steric switching of the potential compression is accurately accounted for by ab initio modeling (Density Functional Theory, DFT) that proposes a mechanism for pyramidalization of the N_pyridinio atom coupled with reduction. When the hybridization change of this atom is hindered (Ar = Xy, Lu, qLu), the first reduction is a one-electron process. Theory also reveals that the single-step two-electron reduction involves couples of redox isomers (electromers) displaying both the axial geometry of native EPs and the pyramidalized geometry of doubly reduced EPs. This picture is confirmed by a combined UV-vis-NIR spectroelectrochemical and time-dependent DFT study: comparison of in situ spectroelectrochemical data with the calculated electronic transitions makes it possible to both evidence the distortion and identify the predicted electromers, which play decisive roles in the electron-transfer mechanism. Last, this mechanism is further supported by in-depth analysis of the electronic structures of electrophores in their various reduction states (including electromeric forms).

Full text of DOI:10.1021/ja210024y

Article
pubs.acs.org/JACS
Single-Step versus Stepwise Two-Electron Reduction of
Polyarylpyridiniums: Insights from the Steric Switching of Redox
Potential Compression
Jerome Fortage,†,‡ Cyril Peltier,§ Christian Perruchot,† Yohei Takemoto,∥ Yoshio Teki,∥ Fethi Bedioui,⊥
́
̂
Valerie Marvaud,‡ Gregory Dupeyre,† Lubomír Pospísil,# Carlo Adamo,§ Magdalena Hromadova,*,#
́
́
́
́
Ilaria Ciofini,*,§ and Philippe P. Laine*,†
́
†Universite
‡UPMC, Universite
75005 Paris, France
́
Paris Diderot, Sorbonne Paris Cite,
́
ITODYS, UMR 7086 CNRS, 15 rue Jean-Antoine de Baïf, 75013 Paris, France
ulaire, UMR 7201 CNRS, Case 42, 4 place Jussieu,
́
Paris 06, Institut Parisien de Chimie Molec
́
§
́
Ecole Nationale Superieure de Chimie de ParisChimie ParisTech, LECIME, UMR 7575 CNRS, 11 rue Pierre et Marie Curie,
́
75005 Paris, France
∥Department of Material Science, Graduate School of Science, Osaka City University, 3-3-138 Sugimoto, Sumiyoshi-ku,
Osaka 558-8585, Japan
⊥Universite
Chimique et Gen
́
Paris Descartes, Ecole Nationale Super
́
ieure de Chimie de ParisChimie ParisTech, Laboratoire de Pharmacologie
́
́
et
́
ique et d’Imagerie, UMR 8151 CNRS and U 1022 INSERM, 11 rue Pierre et Marie Curie, 75005 Paris, France
#J. Heyrovsky Institute of Physical Chemistry, v.v.i., Academy of Sciences of the Czech Republic, Dolejskova 3, 18223 Prague,
Czech Republic
S
* Supporting Information
ABSTRACT: Contrary to 4,4′-dipyridinium (i.e., archetypal
methyl viologen), which is reduced by two single-electron transfers
(stepwise reduction), the 4,1′-dipyridinium isomer (so-called
“head-to-tail” isomer) undergoes two electron transfers at
apparently the same potential (single-step reduction). A combined
theoretical and experimental study has been undertaken to
establish that the latter electrochemical behavior, also observed
for other polyarylpyridinium electrophores, is due to potential
compression originating in a large structural rearrangement. Three
series of branched expanded pyridiniums (EPs) were prepared: N-aryl-2,4,6-triphenylpyridiniums (Ar-TP), N-aryl-2,3,4,5,6-
pentaphenylpyridiniums (Ar-XP), and N-aryl-3,5-dimethyl-2,4,6-triphenylpyridinium (Ar-DMTP). The intramolecular steric
strain was tuned via N-pyridinio aryl group (Ar) phenyl (Ph), 4-pyridyl (Py), and 4-pyridylium (qPy) and their bulky 3,5-
dimethyl counterparts, xylyl (Xy), lutidyl (Lu), and lutidylium (qLu), respectively. Ferrocenyl subunits as internal redox
references were covalently appended to representative electrophores in order to count the electrons involved in EP-centered
reduction processes. Depending on the steric constraint around the N-pyridinio site, the two-electron reduction is single-step (Ar
= Ph, Py, qPy) or stepwise (Ar = Xy, Lu, qLu). This steric switching of the potential compression is accurately accounted for by
ab initio modeling (Density Functional Theory, DFT) that proposes a mechanism for pyramidalization of the Npyridinio atom
coupled with reduction. When the hybridization change of this atom is hindered (Ar = Xy, Lu, qLu), the first reduction is a one-
electron process. Theory also reveals that the single-step two-electron reduction involves couples of redox isomers (electromers)
displaying both the axial geometry of native EPs and the pyramidalized geometry of doubly reduced EPs. This picture is
confirmed by a combined UV−vis−NIR spectroelectrochemical and time-dependent DFT study: comparison of in situ
spectroelectrochemical data with the calculated electronic transitions makes it possible to both evidence the distortion and
identify the predicted electromers, which play decisive roles in the electron-transfer mechanism. Last, this mechanism is further
supported by in-depth analysis of the electronic structures of electrophores in their various reduction states (including
electromeric forms).
electronics2 and artificial photosynthesis,3 among others.
Therefore, there exists a real interest in gaining insights into
1. INTRODUCTION
At the nanoscale, and more specifically at the (intra)molecular
level, manipulating several electrons at the same time remains a
highly challenging task.1 In fact, this constitutes a goal of
Received: October 25, 2011
utmost importance in the related fields of molecular
Published: December 25, 2011
© 2011 American Chemical Society
2691
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
the manner by which certain redox-active molecules (in
particular, reducible electrophores) can handle multielectron
processes not only with respect to through-bond (electronic)
and through-space (electrostatic) interactions4−6 but also by
adapting their architecture.7,8 In this context, we have
undertaken the in-depth study of a series of pyridinium-based
model electrophores, hereafter referred to as branched
expanded pyridiniums (EPs).9−11 These EPs are capable of
undergoing multiple reduction processes in different manners,
and the unraveling of their peculiar structure−electrochemistry
relationships is ultimately expected to lead to a potentially
workable principle for electron storage.12−15
phenomenon, known as “potential compression” (case a) or
“potential inversion” (case b),8 is clearly the more likely. More
precisely, pyramidalization of the Npyridinio atom of the
pyridinium core, originating from a change in its hybridization
during reduction, is postulated (Figure 1).11
At the (intra)molecular level, as far as semirigid and
potentially fully conjugated systems are concerned, the
relationships between structure (more specifically conforma-
tion) and the dynamics of electron transfers, as well as
Figure 1. Pictorial representation of the type of structural distortion
assumed in the potential compression/inversion. X = CH, N, or
N+-Me, and Z is the charge of the electrophore in its native form
(Z = 1+, 2+).
intercomponent electronic coupling, are well established.16,17
A
typical instance is the conformational gating of photoinduced
processes within two-component systems conceived for charge
separation.18 The electrochemical properties of molecular
electrophores can also be varied conformationally, provided
that they have roughly the same intramolecular characteristics
as above. This is illustrated by the benchmark case of
bipyridinium isomers and derivatives (viologen-like species),
when the torsion angle about their interannular linkage is
gradually changed.19 Thereby, the critical energy of their lowest
unoccupied molecular orbital (LUMO), that governs their
reduction processes, is steadily changed. Only mild tuning of
electrochemical behavior is obtained in this way: the number of
electrons exchanged remains the same and the standard
potential(s) is(are) moderately shifted.20 In a further step of
molecular design, one may ask whether the redox properties
and, more specifically, the reduction regime of electrophores can
be drastically changed by structural factors, such as intra-
molecular steric hindrance.21 Achieving genuine steric switching
of electronic landscapes, giving rise to two distinguishable states
with totally different electrochemical features, would be the
target.22 The underlying working principle would rely on a
sterically controlled upheaval of redox-active molecular orbitals
(MOs) arising either from their very nature or from the way in
which incoming electron density is distributed among them
(case of redox isomers or electromers23).
Because of their great chemical versatility, combined with
the possibility of multielectron reduction, polyarylpyridinium
electrophores of the branched EP type9−11 offer a unique
opportunity to explore such structure−property relationships.
In this respect, the intriguing observation that the N-
pyridylium-2,4,6-triphenylpyridinium electrophore (qPy-TP)
is two-electron-reduced in a single-step process at −0.60 V
(vs SCE, in MeCN)11 merits our attention. The following
question arises: how can these two electrons be added at
apparently the same potential to such a small compact semirigid
electrophore, and the electrostatic (i.e., Coulombic) repulsion
be overcome? Indeed, except for metal ions, (single-step)
multielectron reduction is known to occur essentially in two
ways: (1) The redox-active sites do not experience electrostatic
interaction; however, in view of the molecular structures, this
explanation can reasonably be ruled out here. (2) There is an
electrostatic interaction but the system adapts to minimize it via
a structural rearrangement8 or by counterbalancing adverse
repulsion with some gain in electronic stability (e.g., resonance
energy/aromaticity).24 The second reduction thus becomes as
easy as (case a) or even easier than (case b) the first one. This
The existence of such a redox-driven structural upheaval is
primarily inferred on the basis of a computational study on the
two-electron-reduced hexaphenylpyridinium (Ph-XP).9 It also
comes from spectroscopic (Raman)25a and theoretical25b work
on the singlet-excited charge-transfer state of pyridinium-N-
phenolate betaine, B30. The present work aims to ascertain the
reality of this distortion (Figure 1) and to demonstrate that it
plays a critical role in the two-electron reduction of branched
EPs. This demonstration is based on a synthetic strategy
conceived to take advantage of the impact of intramolecular
steric hindrance on the reduction behavior. Thus, we propose
the mechanism by which the ability of some electrophores to
store one or two electron(s) at a given potential is switched. In
the following, we refer to “two elementary electron transfers
that occur apparently at the same potential” as a “single-step
two-electron” reaction.
2. MOLECULAR DESIGN AND METHODOLOGICAL
APPROACH
Comparison of the electrochemical behaviors of two-com-
ponent systems (dyads in Figure 2) consisting of a ferrocenyl
moiety (Fc) covalently linked via a phenylene spacer to a N-
methyl-2,4,6-triphenylpyridinium (Me-TP) or a 1,2,4,6-tetra-
phenylpyridinium (Ph-TP) moiety affords a first indication that
there might be interplay between steric hindrance and the
reduction processes within branched EPs.
A dramatic shrinking of the potential difference (ΔE) related
to the one-electron reductions is observed on going from Me-
TP (ΔE = 460 mV)11 to Ph-TP (ΔE = 160 mV),9 that is, on
replacing the methyl by an aryl group like phenyl at the Npyridinio
atom. To what extent this change of electrochemical behavior
is related purely to steric strain or whether there are also
electronic factors is one of the issues addressed here.
Our approach for unraveling the precise nature of structure−
property relationships within branched EPs relies on the steady
increase in the intramolecular crowding to produce steric gating
of the reduction regimes (two single-electron transfers versus
one two-electron process).26 The model electrophores
synthesized for this purpose are depicted in Chart 1 (see also
Supporting Information).27
Two types of steric constraint are systematically varied. On
one hand, the N-pyridinio aryl group (Ar) linkage is
encumbered by inserting methyls (R = H becomes R = Me
to give Ar = Ph/Xy, Py/Lu, and qPy/qLu; Chart 1) to lock the
electrophores in their axial conformation and, to some extent,
2692
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
block pyramidalization (Figure 1). On the other hand, the
bulkiness of the pyridinium core is steadily increased by
changing its β substituents R′ (i.e., H, Ph, and Me at positions 3
and 5), to give electrophores of types Ar-TP, Ar-XP, and Ar-
DMTP, respectively (Chart 1). This type of structural
modification is aimed at changing the energy barrier between
axial and pyramidal configurations and/or their relative sta-
bilities. Last, selected electrophores are derivatized with a
ferrocenyl group as an internal redox probe (one-electron
oxidation process related to the Fc+/Fc couple) to count the
number of electrons involved in each process.28,29 Syntheses
and full characterization of model electrophores and dyads are
given in the Supporting Information.
3. RESULTS AND DISCUSSION
3.1. Interplay between Structure and Electrochemical
Properties: Insights from Theory. We start with the results
of the initial molecular modeling study, carried out to
rationalize the already reported experimental findings,
especially those of the TP series bearing a N-pyridinio aryl
group lacking methyl substituents (Ar = Ph, Py, and qPy;
Chart 1).9,11 Geometry optimization was performed for the
electrophores in their native state ([TP]Z) as well as for their
singly ([TP](Z−1)) and doubly ([TP](Z−2)) reduced states, prior
to calculating the standard potentials for their first and second
reductions. The geometries obtained differ essentially in the
torsion angles only (in particular the interannular Ar−
pyridinium angle), which were found to decrease upon
reduction (planarization). We shall refer to these geometries
as axial (Figure 1). If these computed structures are not
surprising as such,18 they clearly do not conform to our above-
stated expectations (Figure 1) of an emerging pyramidalization
of the Npyridinio atom giving rise to distorted structures, referred
to as pyramidal structures (Figure 1), at least for what concerns
the doubly reduced species. Furthermore, the computed
reduction potentials (E([TP]Z/(Z−1)) and E([TP](Z−1)/(Z−2)))
were found to be very different. This finding is in contradiction
with the experimental observation of two very close one-
electron reductions (Ar = Ph, Figure 2)9 or even of an apparent
two-electron process (Ar = Py and qPy).11
Figure 2. Appending an aryl group at the N-pyridinio position of
EPs. Rotating disk electrode (RDE) voltammetry illustrates the
impact on electrochemical features (potential compression): super-
imposed RDE voltammogram traces of Fc-based dyads related to
Me-TP and Ph-TP, normalized with respect to one-electron redox
process of Fc+/Fc couple as internal reference (solvent is
acetonitrile, MeCN).
Chart 1. Changing Intramolecular Steric Strain within
Branched EPs while Counting the Electrons Involved in
a
Reduction Processes
On the other hand, for the crowded XP series (Chart 1)
bearing the same Ar groups without methyl substituents, like
Ph9 or Py and qPy (present work), structure optimization of
the doubly reduced electrophores ([XP](Z−2)) yielded a single
minimum corresponding to a pyramidalized geometry (Figure 3;
Table S1 in Supporting Information). In other words, the axial
form of [XP](Z−2) is unstable. Interestingly, for these XP types
of electrophores in their singly reduced state ([XP](Z−1)),
geometry optimization leads to axial structures ([XP(ax-
ial)](Z−1)) when one starts from their native axial geometry
(i.e., [XP(axial)]Z + 1 electron), whereas pyramidal structures
([XP(pyramid)](Z−1)) are obtained from the pyramidalized
geometry of the doubly reduced electron acceptors (i.e.,
[XP(pyramid)](Z−2) − 1 electron), as schematized in Figure 3.
This finding is a strong indication that there are at least two
wells (minima) in the potential energy surface (PES), i.e., two
possible forms for the one-electron-reduced XP electrophores
([XP](Z−1)), namely [XP(axial)](Z−1) and [XP(pyramid)](Z−1)
.
a
These redox isomers are referred to as electromers.23
Moreover, if it is assumed that [XP(axial)](Z−1) electromers
are the relevant intermediates, the reduction potentials
(E([XP]Z/(Z−1)) and E([XP](Z−1)/(Z−2))) are found to be
almost equal (within ca. 0.1 V) in the cases of Ar = Ph, Py,
Molecular structures and related labels of the tetraaryl- (TP),
hexaaryl- (XP), and mixed-branched 3,5-dimethyl-TP (DMTP)
pyridiniums synthesized and studied. Z is the electrophore charge in
its native form. Counteranions are BF4− for monocationic EPs (Z =
1+) and PF6− for dicationic EPs (Z = 2+).
2693
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Figure 3. Perfecting the computational procedure (geometry optimization) for Ar-XP and Ar-TP with Ar = Ph, Py, and qPy. Coordinate Q
represents the structural form of the electrophore for a given reduction state: “axial” or “pyramidal”. Closed circles indicate optimized structures
directly obtained from the native forms upon electron addition, whereas open circles denote optimized geometries derived from pyramidalized
doubly reduced forms. Redox isomers (electromers) are in brackets.
Figure 4. Optimized molecular geometries for the singly and doubly reduced forms of Ar-TP-type electrophores (representative case of Ar = qPy).
Pathways to [TP(pyramid)](Z−2) involving [TP(axial)](Z−1) or [TP(pyramid)](Z−1) as intermediates are equally relevant.
and qPy30 (“potential compression”; Table S1), in line with the
experimental single-step two-electron reductions (see section
3.2 below and ref 9). Note that the assumption that
[XP(pyramid)](Z−1) electromers are involved also leads to
almost equal values of potentials for these aryl groups, but with
the second reduction more anodic than the first one (“potential
inversion” 30), which also accounts well for the electrochemical
measurements. While electrochemistry does not discriminate
clearly between the two options, strict interpretation of the
computational results (Table S1), however, suggests that the
more likely explanation is a slight inversion of potential
(see section 3.2 below).
the 19 electrophores depicted in Chart 1, thus exploring
systematically axial and pyramidal geometries for singly and
doubly reduced species. All calculated data are collected in
Tables S1−S3 (Supporting Information).
3.2. Reduction Behavior of Branched EPs: Combined
Electrochemical and Theoretical Study. Parallel to the
theoretical study, the electrochemical behavior of the whole
series of electrophores and dyads (Chart 1) was investigated by
potentiodynamic methods including cyclic voltammetry (CV)
and hydrodynamic experiments (rotating disk electrode (RDE)
voltammetry). Table 1 lists the data (see also Supporting
Information).
Most important, here, is the fact that the consistency of
the theoretical and experimental findings for the Ar-XP series
(with Ar = Ph, Py, and qPy) relies on the assumption of an
intervening pyramidalized structure, at least for doubly reduced
species, hence the idea of considering this possibility for
electrophores of the TP series (Figures 1 and 2). Thus, when
this particular distortion was introduced into our computational
approach (Figure 3), the pyramidal geometries were indeed
found to correspond to real energy minima for both the singly
and the doubly reduced Ar-TP electrophores (Ar = Ph, Py, and
qPy; Figures 3 and 4; Table S2 in Supporting Information).
Two main trends can be seen. On one hand, when the aryl
group at the N-pyridinio position of the pyridinium core is
sterically demanding (i.e., Ar = Xy, Lu, and qLu; Chart 1), the
first reduction process involves one electron. This process is
most often reversible (the only exception is qLu-DMTP),
regardless of the level of steric constraint in the periphery of the
pyridinium platform (TP, XP, or DMTP; see Figure 5a,c and
Supporting Information). When the second one-electron
reduction is observed (TP series; see Supporting Information),
the process occurs at much higher potential and is systemati-
cally irreversible. On the other hand, when the aryl group has
no bulky substituent (i.e., Ar = Ph, Py, and qPy), the first
reduction process always involves two electrons (see Figure 5c
and Supporting Information) and its reversibility is sensitive to
the degree of crowding of the pyridinium platform (reversible
or irreversible; see Figure 5b and Supporting Information).
Also, similar values for reduction potentials (E([TP]Z/(Z−1)
)
and E([TP](Z−1)/(Z−2))) were derived (Table S2), this time in
accordance with experimental findings.9,11
This computational approach was therefore applied (Ex-
perimental Section) to assess the reduction potentials of each of
2694
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
a
Table 1. Reduction of Model Electrophores and
Ferrocenyl-Based Dyads (Chart 1) at Pt Electrode (vs SCE)
for MeCN Solutions (+0.1 M NBu4PF6) at Room
b
Temperature
entry
E1/2, V
n
E1/2, V
n
c
Me-TP
−1.07; rev.
−1.00; rev.
−1.06; rev.
−0.93; rev.
−0.92; rev.
−0.60; rev.
−0.70; rev.
−1.17 ; quasi-rev.
−1.16; rev.
−1.10 (Epc); irrev.
−0.73 (Epc); irrev.
−1.33 (Epc); irrev. 1 < n < 2
1
1
1
2
1
2
1
2
1
2
2
−1.53; rev.
−1.16; rev.
−1.83 (Epc); irrev.
1
1
1
d
Ph-TP
Xy-TP
c
Py-TP
Lu-TP
−1.70; irrev.
1
1
c
qPy-TP
qLu-TP
−1.23 (Epc); irrev.
d
e
Ph-XP
f
Xy-XP
Py-XP
qPy-XP
g
Ph-DMTP
Xy-DMTP
Py-DMTP
Lu-DMTP
f
f
−1.30; rev.
−1.20; quasi-rev.
−1.23; rev.
1
2
1
2
1
h
qPy-DMTP −0.84 (Epc); irrev.
qLu-DMTP −0.93 (Epc); irrev.
−1.40 (Epc); irrev.
1
a
b
With the exception of Lu-XP and qLu-XP: see ref 27. Unless
otherwise noted, E1/2 (vs SCE) is calculated as (Epa + Epc)/2, where
Epa and Epc are anodic and cathodic peak potentials measured by CV at
0.1 V s−1; n is the number of electrons involved in the redox process
determined by RDE voltammetry with Fc-based dyads (see text). For
c
d
more details, see Experimental Section. From ref 11. From ref 9.
e
f
Epc. Beyond the investigated potential window (below −1.90 V).
See refs 31 and 32. From the corresponding Fc-based dyad.
g
h
Note that the electrophore Ph-TP is a borderline case, since
two partly merged one-electron waves are observed instead of a
single two-electron wave (Figure 2).
The overall picture we get from the CV and RDE results (see
also Supporting Information), is the following: other things
being equal, the one- or two-electron nature of the first reduction
process is solely governed by the bulkiness of the N-pyridinio aryl
group. This picture is fully consistent with the assumption that
N-pyramidalization during reduction is responsible for the
potential compression (or inversion), which naturally vanishes
when it is prevented by steric hindrance of the aryl group.
As regards two-electron processes (Ar = Ph, Py, and qPy),
the relative degrees of reversibility/irreversibility depend on
steric parameters essentially varied via the β substituents on the
pyridinium core (Chart 1). Clearly, the reversibility is more
pronounced when there is more room for intramolecular
structural rearrangement. From the theoretical viewpoint, as
far as singly reduced species are concerned, both axial and
pyramidal electromeric structures are minima in the PES
(Figures 3 and 4 and Supporting Information). The same also
holds for the XP and DMTP series (Tables S1 and S3). For the
TP series, these electromers have similar stabilities (Table S2).
Regarding doubly reduced species, the two types of electro-
meric structures are found by calculation only in the case of
the TP family. For all three families (TP, XP, and DMTP), the
pyramidal structures are found to be by far the more stable
(Tables S1−S3). Clearly, the most likely explanation for the
different degrees of reversibility of two-electron reductions is
twofold: (1) different stabilities for the two singly reduced
electromers with respect to that of the final pyramidalized
doubly reduced species and (2) different interconversion
energy barrier heights.34
Figure 5. Impact of various steric constraints on electrochemical features.33
Superimposed representative voltammograms of Fc-based dyads containing
Ar-TP (R′ = H; black), Ar-XP (R′ = Ph; red), and Ar-DMTP (R′ = Me;
blue). (a) CV, Ar = Xy. (b) CV, Ar = Py. (c) RDE voltammetry, Ar = Xy
(solid lines) and Py (dashed-dotted lines). Plots of dyads are normalized
with respect to the one-electron redox process of the Fc+/Fc couple.
Concerning one-electron processes (Ar = Xy, Lu, and qLu),
the reversibility of the first reduction is in line with the slight
2695
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Figure 6. Representative optimized molecular geometries for the singly (top) and doubly (bottom) reduced forms of electrophores Ar-TP, Ar-XP,
and Ar-DMTP (representative case of Ar = Xy).
conformational changes computed for the one-electron-
reduced electrophores with bulky aryl groups. They all show
an axial structure closely resembling that of their native
forms as the only minimum of the PES (see Figure 6 and
Tables S1−S3). The systematic irreversibility of the second
one-electron reduction, when observed (only in the case of the
TP series), can be explained by the unexpected formation of
the severely and even locked pyramidalized structures
computed for the doubly reduced species (see Supporting
Information and Table S230). Less severely distorted structures,
referred to as bent structures (onset N-pyramidalization), are
also computed as the only stable products for the more
sterically crowded doubly reduced XP and DMTP (Figure 6).
In all cases, the driving force for the change of hybridization of
the Npyridinio atom is apparently sufficiently strong to overcome
the steric hindrance produced by the Me substituents of the
aryl group (Ar = Xy, Lu, and qLu).
To summarize, from this combined electrochemical and
theoretical study, one can derive the redox equations that
account accurately for the two observed reduction behaviors.
As regards the single-step two-electron reduction (“compres-
sion of potential”, ΔE > 0 with ΔE approaching 0, or “inversion
of potential”, ΔE(theory) ≤ 0 but ΔE(exptl) = 0, see Tables 1
and S1−S3)8 of branched expanded pyridiniums (EP):
energetic for Ar = Ph, Py, and qPy (see Tables S1−S3). As
regards the net electrochemical−distortion−electrochemical
process (herein referred to as an ECE process) related to the
single-step two-electron reduction, the question is now raised
as to whether the distortion is concerted35,36 with the first
elementary one-electron reduction or if the two electromers
interconvert after the first reduction. In the latter case, the
situation is as follows:
+e
Z
•
(Z−1)
[EP(axial)] ⎯→⎯ [EP (axial)]
E
1
(4)
•
(Z−1)
[EP (axial)]
•
(Z−1)
⇌ [EP (pyramid)]
Interconversion
(5)
(Z−1)
[EP•(pyramid)]
+e
(Z−2)
⎯→⎯ [EP(pyramid)]
E2
(6)
Then, as the pyramidalized doubly reduced electrophore is by
far the more stable, the electromer adopting this structure
([EP•(pyramid)](Z−1)) is therefore readily reduced, hence the
rapid consumption of [EP•(axial)](Z−1). This is reflected by the
disproportionation constant, which indicates that one-electron-
reduced electrophores are unstable. In the case of electron
acceptors of the TP type, contrary to XP and DMTP types of
electrophore, calculations indicate that the axial doubly reduced
electromers ([EP(axial)](Z−2)) also exist despite being much
less stable than the pyramidal ones, [EP(pyramid)](Z−2) (Table S2).
This opens up the possibility of another interconversion
equilibrium between electromers, this time involving [EP(pyr-
amid)](Z−2) and [EP(axial)](Z−2), although highly unfavorably
displaced for the latter. This second interconversion process,
along with the previous one, reflects the moderate steric strain
present within Ar-TP electrophores. Both equilibria are
consistent with the observed reversibility of two-electron
reduction when Ar = Ph, Py, and qPy (Table 1, Figure 5b,
and Supporting Information). Overall, the postulated fluxional
behavior32 related to the existence of electromers and
interconversion equilibria (eqs 4−6) appears as the most likely
explanation for the observed reduction behaviors, and is
therefore preferred to a hypothetical distortion concerted
Z
[EP(axial)]
+e
•
•
(Z−1)
⎯→⎯ [EP (axial)/EP (pyramid)]
E
1
(1)
(2)
(Z−1)
[EP•(axial)/EP•(pyramid)]
+e
(Z−2)
⎯→⎯ [EP(pyramid)]
E2
Disproportionation:
2[EP•(axial/pyramid)]
(Z−1)
(Z−2)
⇌ [EP(pyramid)]
ΔE = E2 − E1
+ [EP(axial)]Z
(3)
Calculations tell us that the two types of electromers,
[EP•(axial)](Z−1) and [EP•(pyramid)](Z−1), are roughly iso-
2696
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
with the first one-electron exchange, resembling that observed
in nitroalkanes, for instance.36,37 Based on computational
results, it follows that the single-step two-electron reduction
can be explained by a compression of potential in the case of TP
type electrophores (Table S2) and by an inversion of potential
for XP electrophores (Table S1). In the case of the DMTP
series, one can hardly decide between compression or inversion
of potential (Table S3).
Concerning the stepwise two-electron reduction (two
successive one-electron processes: ΔE > 0) within branched
EPs with Ar = Xy, Lu, and qLu, the computational study reveals
that there is no couple of electromers for the singly reduced
electrophores (see Tables S1−S3); hence, there is no possible
interconversion equilibrium. Consequently, the change of
hybridization is most likely concerted with the second one-
electron reduction.
+e
Z
•
(Z−1)
[EP(axial)] ⎯→⎯ [EP (axial)]
E
1
(4)
+e
•
(Z−1)
(Z−2)
[EP (axial)]
⎯→⎯ [EP(pyramid)]
E
2
(7)
3.3. Electrochromic Behavior of Branched EPs:
Combined Spectroelectrochemical and Theoretical
Study. UV−vis−NIR spectroelectrochemistry is the method
of choice for probing the identity of the reduced species formed
in solution during potentiodynamic measurements. For an
electrophore of a given reduction state (i.e., native, singly
reduced, or doubly reduced), different spectral signatures are
expected for its axial and pyramidal structures (if they exist),
based upon their differently extended π-delocalized systems.
Three representative electrophores of the TP-type were
investigated (Chart 1): Me-TP (Figure 7), qPy-TP (Figure 8),
and qLu-TP (Figure 9). In Figures 7−9 are superimposed
experimental plots and the calculated electronic transitions
(time-dependent DFT (TD-DFT) results; Supporting In-
formation, Tables S5−S7) corresponding to the forms (axial
or pyramidal) that match the experimental results. For the sake
of clarity, calculated transitions for the forms of reduced
electrophores that do not match the experimental spectra are
given in Supporting Information (Figures S2−S4).
Figure 7. UV−vis−NIR spectroelectrochemistry of Me-TP. Super-
imposed experimental plots (black, [Me-TP]+ native; red, [Me-TP]0
singly reduced; blue, [Me-TP]− doubly reduced) and corresponding
calculated electronic transitions: pyramidal (solid vertical lines) and
axial (dashed vertical lines). (a) [Me-TP]+ and [Me-TP]0. (b) [Me-
TP]0 and [Me-TP]−. Note that no energy minimum was found for
reduced axial forms (see Table S4).
In spite of the fact that energies of electronic transitions
calculated by our approach (Tables S5−S7) are often slightly
overestimated, they are in the range of accuracy expected for
the level of theory used39 and they reflect the experimental
spectra.
In the case of Me-TP, the structure optimizations performed
on the axial forms for both the one-electron and the two-
electron-reduced species all converge to the pyramidalized
geometry. Consequently, electronic transitions corresponding
only to the three chromophoric species are reported (Figure 7):
the axial native form, [Me-TP]+, and the one- and two-electron-
reduced pyramidal forms, [Me-TP]0 and [Me-TP]−, respec-
tively. The good agreement between experiment and theory
at the TD-DFT level (Figures 7 and S5) further supports
our inferences concerning Me-TP, and validates our present
approach.
In the case of qPy-TP, whether in its singly or doubly
reduced form, both the pyramidal and the axial redox isomers
correspond to minima on the PES. Therefore, the features of
up to five chromophoric species were calculated: [qPy-
TP(axial)]2+, [qPy-TP(axial)]+, [qPy-TP(pyramid)]+, [qPy-
TP(axial)]0, and [qPy-TP(pyramid)]0 (Table S6). Apart from
the [qPy-TP]2+ native electrophore, the calculated pattern of
electronic transitions was found to match properly the
experimental spectrum of the reduced electrophore in the
sole case of the pyramidal doubly reduced species [qPy-
TP(pyramid)]0 (see Figures 8 and S3), also computed to be
more stable than the axial one [qPy-TP(axial)]0 (Table S2).
Experimentally, no isosbestic point is observed during the
reduction of [qPy-TP]2+ to [qPy-TP(pyramid)]0 (Figure 8 and
Supporting Information). This is indicative of a reduction
process which is more complex than a mere direct transfer of a
couple of electrons. This observation is consistent not only with
an intervening interconversion equilibrium of singly reduced
electromers (eq 5 in section 3.2), but also with only a moderate
propensity of these latter ([qPy-TP]+) to disproportionate.
These findings further support the assumption that the N-
pyramidalization is correlated with the two-electron reduction
process, moreover involving intermediate electromeric species.
2697
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Figure 8. UV−vis−NIR spectroelectrochemistry of qPy-TP. Super-
imposed experimental plots (black, [qPy-TP]2+ native; blue, [qPy-
TP]0 doubly reduced) and corresponding calculated electronic
transitions: pyramidal (solid vertical lines) and axial (dashed vertical
lines). Calculated electronic features of the [qPy-TP]+ singly reduced
form (both axial and pyramidal) are given in Supporting Information.
Last, it is worth noting that the overall shape of the electronic
spectrum of [qPy-TP]0 is closely akin to that of [Me-TP]−,
indicating that the chromophoric entities are roughly the
same in both cases. Assuming that there is no σ or π dimeriza-
tion (pimerization),40 both the Me-TP and qPy-TP species
should lead, after two-electron reduction, to N-pyramidalized
carbanions.
A detailed analysis of the MOs involved in the electronic
transitions computed for the pyramidal doubly reduced species
of both Me-TP and qPy-TP further confirms this picture. Both
species have their HOMO localized on the pyridinium core and
their strongest electronic transitions are all of the HOMO−
(LUMO+n) type (Tables S5 and S6). The spectral features in
the visible−NIR region (λ > 400 nm) are the signature of the
distorted pyridinium core. Interestingly, in the case of doubly
reduced qPy-TP, the structural decoupling of the qPy aryl
group due to N-pyramidalization makes its contribution to the
HOMO practically negligible. Nonetheless, qPy mainly
contributes to the LUMO and (LUMO+2) orbitals of the
doubly reduced electrophore and participates in its chromo-
phoric activity (Supporting Information).
In the case of qLu-TP (Figure 9), it was possible to study the
electrochromic properties of the one-electron ([qLu-TP]+) and
two-electron ([qLu-TP]0) reduced species separately, in their
pure form. For [qLu-TP]+ (Figure 9a), the more diagnostic
spectral domain is the NIR: the experimental plot agrees well
with the two electronic transitions computed at 831 and 1370
nm38 for [qLu-TP(axial)]+, but does not match at all the set
calculated for [qLu-TP(pyramid)]+, whose computed lowest-
energy transition lies at 616 nm (with a very weak oscillator
strength, f, of about 0.021; Table S7). The structure in solution
of the singly reduced qLu-TP electrophore is therefore
unambiguously axial, in agreement with its computed greater
stability (Table S2). Furthermore, for the SOMO−LUMO
transition at 1370 nm,38 it is interesting to note that the SOMO
displays an important contribution centered on the lutidylium
Figure 9. UV−vis−NIR spectroelectrochemistry of qLu-TP. Super-
imposed experimental plots (black, [qLu-TP]2+ native; red, [qLu-TP]+
singly reduced; blue, [qLu-TP]0 doubly reduced) and corresponding
calculated electronic transitions: pyramidal (solid vertical lines) and
axial (dashed vertical lines). (a) [qLu-TP]2+ and [qLu-TP]+.38 (b)
[qLu-TP]+ and [qLu-TP]0.
substituent, which is actually made possible by the axial nature
of the electromer ([qLu-TP(axial)]+). For [qLu-TP]0, both
the pyramidal and the axial structures correspond to minima
on the PES (Table S2).30 However, comparison of the
experimental spectrum (Figure 9b) with the shapes of the
sets of electronic transitions computed for [qLu-TP(axial)]0
and [qLu-TP(pyramid)]0, shows that the pyramidal electromer
is the more likely in solution. Indeed, the strongest electronic
transition of [qLu-TP(axial)]0 is predicted at around 800 nm
(at 793 nm with f = 0.12 and at 762 nm with f = 0.42) along
with a second intense transition at 514 nm (f = 0.37), whereas
experimentally the red-edge absorption is situated at around
600 nm, with all the main significant bands positioned in
the UV−near-visible region (200−450 nm). On the other
hand, the dominant electronic transition computed for
[qLu-TP(pyramid)]0 (Table S7) is at ca. 400 nm (395 nm,
f = 0.58), with three weaker satellite transitions calculated
at 307 (f = 0.44), 473 (f = 0.34), and 546 nm (f = 0.10).
It follows that the spectrum recorded at the end of the
2698
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
second reduction is that of the pyramidal electromer,
[qLu-TP(pyramid)]0, thereby further substantiating the un-
expected finding that the energy barrier related to the steric
hindrance of the lutidylium group can be overcome by the gain
of energy associated with the hybridization change. This overall
picture is consistent with the computed relative stabilities: the
axial electromer is indeed the more stable for the singly reduced
form ([qLu-TP(axial)]+), whereas the pyramidal electromer
([qLu-TP(pyramid)]0) is the more stable bireduced species,
even though computed to be somewhat less stable30 and
almost isoenergetic with the axial one ([qLu-TP(axial)]0), in
line with the electrochemical findings of section 3.2 (i.e.,
irreversibility of the second reduction). It is also interesting
to note that, analogous to the doubly reduced Me-TP and
qPy-TP pyramidal species, in the case of the most stable
[qLu-TP(pyramid)]0 electromer the HOMO is essentially
localized on the pyridinium core, and the dominant transitions
observed are all of the HOMO-to-(LUMO+n) type (Support-
ing Information).
3.4. Integrated Overall Picture: Toward Steric Switch-
ing of Potential Compression/Inversion. To complete the
assessment of the observed potential compression (TP series)
or inversion (XP series), it is worth considering the possible
role of solvation that may be significant.8,41 Solvent stabilization
varies roughly as the square of the charge, according to the
Born treatment.8 Therefore, from the viewpoint of electro-
statics, one should distinguish between electrophores whose
charge decreases steadily upon reduction, as for dicationic head-
to-tail-bipyridinium species (Ar = qPy and qLu), going from +2
for the native state to zero for the fully reduced form, and the
monocationic electrophores whose charge goes from +1 to zero
and −1. The solvent used in the present study, MeCN, is an
electron-donating solvent that tends to stabilize cations and
destabilize anions. Solvent stabilization, therefore, is expected to
steadily decrease upon reduction of dicationic native electro-
phores. Also, neutral singly reduced monopyridinium-based
electrophores (i.e., monocationic native EPs) are hardly
stabilized by solvation; neither are their monoanionic doubly
reduced derivatives as compared to their monocationic native
parents. In fact, the same electrochemical behavior, namely
single-step two-electron reduction, is observed regardless of the
sequence of charge change. This persistence of the phenom-
enon of potential compression/inversion rules out any
determining role of the solvent in the observed electrochemical
process. Besides, solvation energy is not of the same order of
magnitude as that required for N-pyramidalization. Ion-pairing
effects are also ruled out by roughly the same reasoning, but
also in view of the chemical nature of the electrophores and the
counter-anions involved (PF6− or BF4−), which are poorly
nucleophilic.
Now that we have a clear-cut picture of the origin of single-
step two-electron reductions, the complete set of computed
data can be checked against available experimental data.
Graphical representations have been chosen to display the
relevance of the synthetic strategy (Chart 1) developed to
establish the existence of the N-pyramidalization as the key
structural rearrangement explaining the potential compression/
inversion and to further substantiate it. In Figure 10 are
superimposed computational results based on the assumption
of N-pyramidalization correlated with reduction (Tables S2 and
S4) and experimental measurements of reduction potentials,
along with the associated number of electrons (Table 1) for the
prototypical Ar-TP electrophore. Similar graphs for Ar-XP and
Figure 10. On/off switching of the potential compression for the Ar-
TP series of electrophores (configurational gating of reduction
processes), based on experimental data (Eexp; Table 1), calculated
data (Eth; Tables S2 and S4), and assuming that (1) there is an
interconversion equilibrium of electromers when Ar = Ph, Py, and qPy
(1st reduction, E°(red1)/(N−A); 2nd reduction, E°(red2)/(D−D))
and (2) the N-pyramidalization is somehow concerted with the second
one-electron reduction when Ar = Xy, Lu, and qLu (1st reduction,
E°(red1)/(N−A); 2nd reduction E°(red2)/(A−D)), as proposed in
section 3.2. N, A, and D refer to the native (axial), axial, and distorted
(pyramidal or bent) forms, respectively (see Supporting Information).
Ar-DMTP electrophores are given in Supporting Information
(Figures S7 and S8).
The whole idea was to show that by manipulating
appropriate structural parameters, it was possible to switch
on/off this compression or inversion of potential and thereby
to demonstrate the accuracy of our explanation (Figure 1). As
shown in Figure 10, there is an obvious agreement between
experimental and theoretical results that definitively validates
the quality of the modeling as well as the relevance of the
working hypotheses. Only a slight discrepancy between
experimental and calculated values of standard potential is
noted in the case of qPy-TP, which is ascribed to self-
interaction error.30,42
The radical change of electrochemical properties revealed by
their switchability also demonstrates that steric strain can
produce effects that compare well with those classically
obtained by modifying the strength of electron-withdrawing/
releasing substituents, referred to as electronic effects. In other
words, provided that electrophores undergo compression/
inversion of potential as a result of structural rearrangement,
steric tuning becomes competitive with electronic tuning of
their redox properties. Most importantly, we show here that
even the ability of an electrophore to store one or two electrons
at a given potential can be controlled and manipulated, which is
quite unusual.
3.5. Further Delineation of the Role of the N-Pyridinio
Group: Insights from Theory. MOs computed for various
model electrophores are analyzed in this section in order to
get a more precise although qualitative idea of the respective
roles of the N-pyridinio aryl groups and the pyridinium cores
(Figure 1) at the structural or electronic levels.
2699
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Comparison of the LUMOs of various electrophores in their
axial native forms reveals that the electron exchanged first is
attached to the core pyridinium or to the N-pyridinio aryl
group (Ar) depending on whether the electrophore is
monocationic (e.g., Py-TP) or dicationic (e.g., qPy-TP), as
expected11 (Figure 11). However, in the latter case, it is worth
where they are temporary stored. Note that under the (wrong)
assumption that electrophores remain axial upon reduction,
incoming electrons are finally stored (SOMO and HOMO) at
the site where they were first attached (LUMO); i.e., there is no
intramolecular shift of electron density (Figure 12). When
the aryl group is virtually redox-inactive (Ar = Ph, Xy) or
far less redox-active than the pyridinium core (Ar = Py, Lu),
that is, in the case of monocationic electrophores, incoming
electrons (of the first and second reductions) are directly
attached to the pyridinium core. These findings also hold
for other dicationic and monocationic electrophores of the
XP and DMTP types. In other words, if the aryl group of
branched EPs is not really involved in the electrophoric
properties (i.e., does not carry electron(s)), it may function
as an “electron antenna”.
In the case of heavily constrained electrophores (Ar-XP and
Ar-DMTP with Ar = Xy, Lu, and qLu), for which no single-step
two-electron reduction is observed, the situation can be quite
different (see Figure 13). Not only is N-pyramidalization
hindered, but also the electron-accepting strength of this
pyridinium core is weakened by slightly electron-donating
electronic effects of β substituents (i.e., Me and canted Ph),33 as
shown by the electrochemical study (Table 1, section 3.2).
As regards steric effects, inspection of doubly reduced species
is instructive (Figure 13, top). Thus, given the same lutidyl
aryl group (Lu), N-pyramidalization is computed for the XP
electrophore (pyramidalization is driven to completion)
whereas the backbone is only bent in the case of the DMTP
analogue (onset pyramidalization; see also Figure 6). This
finding is ascribed to the fact that the methyl substituents at
the β positions (R′ in Chart 1) are sterically more demanding
than the canted phenyl rings, due to their “ball-shaped” van der
Waals volume. These methyl substituents indirectly produce a
larger steric strain in the whole environment of the pyridinium
platform, leaving no room for a large structural rearrangement
like N-pyramidalization. The second significant observation is
that on going from Lu to qLu as the aryl group for the same XP
electrophore, the distortion changes from pyramidal to only
bent. This tendency is explained by the larger electron-
withdrawing ability of the qLu group, which promotes a certain
degree of intramolecular intersite electronic delocalization that
is expressed due to the synergistic assistance of the overall steric
crowding. Accordingly, qLu takes part in the electrophoric
function, unlike Lu, as revealed by the spatial extension of the
HOMO that affects the qLu moiety. The same delocalization
effect is evidenced for the DMTP analogues and is revealed by
Figure 11. Switching of the site of first reduction upon quaternization.
(a) LUMO of Py-TP (the pyridinium core is reduced first). (b)
LUMO of qPy-TP (the pendant pyridylium is reduced first);
isocontour = 0.025.
noticing that the Npyridinio atom of the pyridinium core is fully
involved in this LUMO. These different characteristics are
shared by all electrophores of the same charge.
To clarify the status of quaternized aryl groups (qPy or qLu),
which are naturally expected to play an electrophoric role, in
the sense of “charge/electron holders/carriers”, it is interesting
to scrutinize the relevant frontier orbitals (LUMO, SOMO,
and HOMO) of a representative species, qPy-TP, in its differ-
ent reduction states including electromeric axial and pyramidal
forms (Figure 12).
The first striking point is that, in all cases, whether for axial
or pyramidal forms including electromers, the LUMO remains
essentially located on the pyridylium aryl group. As a con-
sequence, the incoming electron is first attached to this part of
the electrophore, including in the second reduction of the
pyramidal electromer. Similar conclusions are reached by
close analysis of the frontier MOs of lutidylium derivatives
(Ar = qLu; Figure S6 in Supporting Information).
The second point concerns the SOMO and HOMO of
relevant pyramidal one- and two-electron-reduced electro-
phores, respectively. These MOs give an indication as to where
the electron density is after the first and second reduction.43 In
both cases, electrons are located essentially on the pyridinium
core. Thus, one could see these pyridylium or lutidylium aryl
groups as a kind of “antenna” for electrons, which are
subsequently channeled to the site of the pyridinium core,
Figure 12. Relevant frontier MOs of qPy-TP in axial and pyramidal geometries (isocontour = 0.025). For the shape of the lone pair of the
pyramidalized nitrogen atom in doubly reduced qPy-TP, see Figure S5 in Supporting Information.
2700
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Figure 13. HOMOs (top) and SOMOs (bottom) of electrophores Lu/qLu-XP and Lu/qLu-DMTP in their doubly and singly reduced forms,
respectively (isocontour = 0.025).
Figure 14. LUMOs and SOMOs of electrophores Ph-TP, Ph-XP, and Ph-DMTP in their native axial and distorted singly reduced forms
(isocontour = 0.025).
the less pronounced bending of the electrophore when Ar =
qLu than when Ar = Lu (Figure 13 top).
strain), the main difference between Ph and Py/qPy derivatives
stems from the electron-withdrawing ability of the latter aryl
groups.
The electrophoric role of the bulky quaternized aryl groups
(qLu) within these constrained electrophores is conserved in
both the axial one-electron-reduced and bent two-electron-
reduced structures (e.g., qLu-XP and qLu-DMTP; Figure 13).
It is worth recalling that this electrophoric contribution
vanishes for all three types of electrophores (i.e., TP, XP, and
DMTP) with a sterically less demanding aryl group like qPy, as
a consequence of the intervening pyramidalization (Figure 12
and Supporting Information), as previously shown.
On the basis of the foregoing, the borderline behavior of Ph-
TP, Ph-XP, and Ph-DMTP in their one-electron-reduced
states, in which the phenyl group is considered as redox-inactive
and almost neutral as regards electronic effects, can now be
rationalized. Compared to pyridyl and pyridylium analogues
(Ar = Py and qPy), which all adopt a pyramidal structure in one
of their singly reduced electromeric forms, not all the phenyl
derivatives follow this trend. The geometry of distorted
electromers is bent, pyramidal and loosely bent (almost axial)
for Ph-TP, Ph-XP, and Ph-DMTP, respectively (Figure 14).
Other things being equal (in particular intramolecular steric
This observation suggests the following explanation for the
onset pyramidalizations (bent structures): not enough electron
density is withdrawn (shifted) toward the Npyridinio atom to
drive its hybridization change to completion when Ar = Ph.
The spin density remains on the C4 position of the reduced
pyridinium core, as is usually the case.4g,44 Clearly, in the basic
case of Ph-TP, the steric strain is moderate and is normally
easily overcome. In the borderline case of Ph-XP, the role of
the steric strain on N-pyramidalization is revealed by the
complete hybridization change. This is assisted directly by the
phenyls on either side of the Npyridinio atom (as was
demonstrated for Me-PP and Ph-PP models; see note S1 in
Supporting Information) and indirectly by the β substituents on
the pyridinium core (R′ = Ph and Me; Chart 1). In the case of
Ph-DMTP, the further increase in steric strain overcomes the
weak driving force for the hybridization change postulated in
the case of Ar = Ph. There is therefore little pyramidalization, and
only slight bending is observed (Figure 14). In all three cases,
however, the doubly reduced electrophore is pyramidalized, giving
2701
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
rise to a compression of potential for Ph-TP with two closely lying
one-electron reductions (Figure 2),9 an inversion of potential with
a single-step two-electron reduction for Ph-XP,9 and a complex
electrochemical−chemical−electrochemical (ECE) behavior31 for
the particular Ph-DMTP.
To summarize, the following picture emerges from close
analysis of the active MOs and related optimized geometries of
model electrophores in their various reduction states:
involving a dynamic fluxional behavior32 between redox isomers
(electromers), is an important step forward.
From the present combined experimental and theoretical
study, the picture we obtain of branched expanded pyridiniums
is that of complex electrophores. The aryl group serves not only
for its structural hindrance and its electron-withdrawing ability;
it can also function as a kind of “electron antenna” 45 that
collects electrons (interfacial exchange site), while the
pyridinium core plays the role of the very electrophoric unit
(electron carrier) that ultimately traps the two electrons.
Overall, we introduce here a new functional paradigm: the
particularly fortunate association of the electrophoric pyridi-
nium motif with intramolecular steric hindrance about the
nitrogen atom that results in the phenomenon of N-
pyramidalization. We demonstrate that this latter confers
upon the thus-modified pyridinium its ability to undergo
apparently a two-electron process. The reorganization energy of
electromers subsequent to the first one-electron transfer (more
than) compensates adverse electrostatic factors, thereby
allowing a second one-electron transfer at the same potential,
which otherwise would occur at a sizably different potential (as
observed when the structural rearrangement is hindered).46
Moreover, when the intramolecular balance between steric
strain and electronic factors is properly adjusted, the two-
electron process becomes reversible.
• It is possible to establish a direct relationship between
the electrophoric character (electron carrier/holder) of a
group/component and its involvement in a mesomeric
effect, which is itself conditioned by the type of structure
(axial, bent or pyramidal) hosting this group. Typically,
as far as branched expanded pyridiniums are concerned,
qPy and qLu aryl groups are potentially electrophoric,
contrary to Ph, Xy, Py, and Lu groups. However, qPy and
qLu are effectively electrophoric in the case of axial and
bent molecular geometries only (i.e., not in pyramidal-
ized structures).
• In the case of electrophores adopting a pyramidal
structure (whether there be electromers or single
forms), these qPy and qLu groups contribute to the
properties of electrophores both by interfacing the
electron exchange (i.e., “by collecting incoming elec-
trons”) and by shifting this electron density toward the
Npyridinio atom of the core, like an antenna system.
• The increase in spin density on the site of the Npyridinio atom
is identified as the triggering event for the change of
hybridization, regardless of whether this electron density
comes from the “antenna” aryl group (Ar = qPy and
qLu) or from the directly reduced pyridinium core itself
(when Ar = Xy, Ph, Py, and Lu). In the latter case,
slightly electron-withdrawing Py and Lu groups are
better at localizing the spin density appropriately than Ph
and Xy groups.
• Whether this change of hybridization is indeed easily
achieved or not is a matter of a complex balance between
possibly conflicting structural and electronic factors.
Clearly, the steric crowding of the pyridinium core assists
the distortion, i.e. the change of hybridization of its
Npyridinio atom.
• The pyramidalization of electrophores in their singly
reduced form, whether or not as electromers, is crucial
for the observation of a single-step two-electron
reduction, as demonstrated by the steric switching of
electrochemical properties (Figure 10). This particular
distorted substructure is the only one able to handle two
electrons easily, that is, to minimize the adverse
Coulombic repulsion between these electrons, hence its
ability to undergo a second reduction at the same
potential as the first one.
Insights gained from the above analysis make it possible to
derive guidelines for future design of molecular devices meant
for multielectron handling, including storage.
Beyond the switchability of electrochemical behavior, other
applications of the interplay between redox and structural
properties can be considered, for example to produce
mechanical work at the molecular level (electromechanical
molecular actuators).20 Of particular interest in the context of
the present study is the possibility of harnessing these
electrophores that exhibit a large redox-driven structural
rearrangement with another functional element such as a
photoactuator (e.g., azobenzene) able to photoproduce a steric
strain on this electrophore. In doing so, not only the electron
storage ability of the modified electrophore could be controlled,
but also the triggering of the release or uptake of two electrons,
depending on whether the active electrophore is in its reduced
or native form (toward smart donors/acceptors of two
electrons).
Last, the existence of the N-pyramidalization induced by
reduction and the understanding of the manner by which this
configurational change can be monitored are of practical
interest for the design of molecular wires made up of
pyridinium subunits.4a−c,47,48
5. EXPERIMENTAL SECTION
5.1. Syntheses, Characterization and General Experimental
Details. Materials, syntheses of the EPs and related ferrocenyl dyads,
and full characterizations are provided in Supporting Information.
5.2. Computational Methods. The first reduction potential
toward a standard hydrogen electrode (SHE) was computed as
• As confirmed by the combined spectroelectrochemical/
TD-DFT study, the pyramidalized pyridinium core is the
only electrophoric component that is observed subse-
quent to two-electron reduction of the reference
electrophore, qPy-TP.
(ΔG − ΔG
)
nat
red
E
(V) =
+ E
SHE
1/2
F
F being the Faraday constant (23.06 kcal mol−1 V−1) and ESHE being
set to 4.44 V. The difference in free energies between the native and
reduced forms (Gibbs free energy, ΔG in kcal mol−1) computed under
standard conditions (T = 298.15 K and P = 1 atm) were used to
evaluate their relative stabilities as previously reported (see for instance
ref 49).
4. CONCLUSION
We have shown that the ability of branched expanded
pyridinium to accept (i.e., store) one or two electrons can be
sterically gated. Elucidation of the origin of the single-step two-
electron reduction, namely a potential compression/inversion
2702
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Analogously, the second reduction potential was computed as
and outlet openings allow filling the cell with degassed samples under
anaerobic conditions. Cyclic voltammograms obtained concurrently
with the UV−vis−NIR spectra were recorded using a potentiostat/
galvanostat Autolab 101 (Metrohm Autolab, Netherlands) at scan
rates of 10 mV s−1, while the spectra were sampled every 2 s using a
diode-array UV−vis−NIR spectrometer (Agilent, model 8453).
TBAPF6, (+99%) supporting electrolyte and MeCN solvent
(anhydrous, 99.8%) were supplied by Sigma-Aldrich and dried
before use.
−
−
)
(ΔG
red‐1e
− ΔG
F
red‐2e
E
(V) =
+ E
SHE
1/2
−
−
where ΔGred‑1e and ΔGred‑2e represent the Gibbs free energies of the
one- and two-electron-reduced electrophores, respectively.
All free energies were computed at the DFT level using the hybrid
exchange-correlation functional PBE050 in conjunction with a double-ζ
all-electron basis set.51
ASSOCIATED CONTENT
* Supporting Information
In order to assess the nature of the one- and two-electron-reduced
species, additional calculations were performed using a range-separated
hybrid functional (CAM-B3LYP52) and a larger basis set (the Pople 6-
31G(d) basis). The structures of native, singly reduced, and doubly
reduced forms of the Me-TP, qPy-TP, and qLu-TP systems were fully
optimized at this level of theory, and all stationary points so obtained
were characterized as minima by subsequent frequency calculations.
The spectral features (i.e., vertical absorption energies and
associated oscillator strengths) of each of these forms were computed
at the TD-DFT level using both the CAM-B3LYP and the PBE0
functional in conjunction with the larger (6-31+G(d)) basis set.
Data reported in the text and figures refer to TD-DFT calculations
performed using the PBE0 functional, since CAM-B3LYP yields quali-
tatively similar results in slightly poorer agreement with experimental
data, as expected. In all cases solvent effects were taken into account
using the Polarizable Continuum Model (PCM53) as implemented
in the Gaussian code,54 and MeCN was considered as solvent
(experimentally used solvent).
■
S
Complete refs 1 and 55; computational results including
optimized structures, calculated standard potentials, and
stabilities of reduced EPs along with their related electronic
properties (TD-DFT) and structures (involved frontier MOs);
full data related to spectroelectrochemical studies; selected
cyclic and RDE voltammograms; preliminary ESR data;
experimental details for the synthesis and characterization of
new compounds and precursors including 1H NMR (500
MHz) and 13C NMR (126 MHz) spectra as well as ESI mass
spectra. Cartesian coordinates of optimized structures and
corresponding total energies. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
philippe.laine@univ-paris-diderot.fr; ilaria-ciofini@
chimie-paristech.fr; hromadom@jh-inst.cas.cz
■
All calculations were performed using the Gaussian program
(Gaussian 09).55
5.3. Electrochemical Measurements. Electrochemical experi-
ments were carried out with a conventional three-electrode cell and a
PC-controlled potentiostat/galvanostat (Biologic VSP or Princeton
Applied Research Inc. model 263A). The working electrode was a
platinum electrode from Radiometer-Tacussel (area, 0.0314 cm2;
diameter, 2.0 mm) mounted in Teflon. Before each experiment, it was
carefully polished with 3 and 0.3 μm alumina pastes followed by
extensive rinsing with ultrapure Milli-Q water. Platinum wire was used
as the counter-electrode and a saturated calomel electrode (SCE) as
the reference. Electrolytic solutions, MeCN (Aldrich, anhydrous,
99.8%) containing 0.1 M tetrabutylammonium hexafluorophosphate
(TBAPF6, Aldrich, +99%) as supporting electrolyte, were routinely
deoxygenated by argon bubbling. The electrochemical properties were
determined using a monomer concentration of ca. 5 mM. All potential
values are given versus SCE. The reported numerical values (Table 1)
were corrected by using a dissolved Fc+/Fc couple as an internal
reference and by setting E1/2 (Fc+/Fc) equal to +0.380 V vs SCE in
MeCN.56 CV experiments were conducted at a scan rate of 0.1 V s−1
(except where specified in the text). RDE voltammetry experiments
were conducted at a scan rate of 0.05 V s−1 (except where specified in
the text) by rotating the disk electrode at 2000 rpm (Controvit device
from Radiometer-Tacussel, France). Square-wave voltammetry experi-
ments were performed with potential sweep rate of 100 mV s−1 (pulse
height, 25 mV; step height, 10 mV; frequency, 50 Hz).
ACKNOWLEDGMENTS
■
Prof. Sebastiano Campagna and his group, Dr. Sophie Griveau,
Prof. Cyrille Costentin, and Dr. John Lomas, are acknowledged
for stimulating discussions. J.F., V.M., G.D., C.A., I.C., and
P.P.L. are grateful to the French National Agency for Research
(NEXUS project ANR-07-BLAN-0277 and SWITCH project
ANR-2010-BLAN-712) for financial support.
REFERENCES
■
(1) Adams, D. A.; et al. J. Phys. Chem. B 2003, 107, 6668.
(2) (a) Joachim, C.; Ratner, M. A. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 8801. (b) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000,
408, 541.
(3) For selected recent reviews, see: (a) Gust, D.; Moore, T. A.;
Moore, A. L. Acc. Chem. Res. 2009, 42, 1890. (b) Magnuson, A.;
Anderlund, M.; Johansson, O.; Lindblad, P.; Lomoth, R.; Polivka, T.;
Ott, S.; Stensjo, K.; Styring, S.; Sundstrom, V; Hammarstrom, L. Acc.
̈
̈
̈
Chem. Res. 2009, 42, 1899. (c) Tinker, L. L.; McDaniel, N. D.;
Bernhard, S. J. Mater. Chem. 2009, 19, 3328. (d) Balzani, V.; Credi, A.;
Venturi, M. ChemSusChem 2008, 1, 26. (e) Eisenberg, R.; Gray, H. B.
Inorg. Chem. 2008, 47, 1697. (f) Benniston, A. C.; Harriman, A. Mater.
Today 2008, 11, 26. (g) Huyhn, M. H. V.; Meyer, T. J. Chem. Rev.
2007, 107, 5004. (h) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci.
U.S.A. 2006, 103, 15729. (i) McEvoy, J. P.; Brudvig, G. W. Chem. Rev.
2006, 106, 4455. (j) Konduri, R.; Ye, H.; MacDonnell, F. M.; Serroni,
S.; Campagna, S.; Rajeshwar, K. Angew. Chem., Int. Ed. 2002, 41, 3185.
(4) For selected examples related to the (multielectron) reduction of
so-called “(conjugation-)extended viologens” (Takahashi, K.; Nihira,
T.; Akiyama, K.; Ikegami, Y.; Fukuyo, E. J. Chem. Soc., Chem. Commun.
5.4. UV−Vis−NIR Spectroelectrochemical Measurements.
The prototype of the spectroelectrochemical optically transparent
thin-layer electrode (OTTLE) cell was developed in the J. Heyrovsky
́
57
́
Institute, and its construction was fully described by Krejci
̌
k et al.
The current modified version of the cell is commercially available from
the Department of Chemistry, University of Reading, UK.58 The steel
body of the cell uses the demountable Specac cell, model Omni-Cell
GS01800 (Specac Ltd., Orpington, UK).59 Two CaF2 windows form a
thin-layer cell, the thickness of which is given by a 0.19 mm poly-
ethylene spacer. The working and counter electrodes consist of two
platinum minigrids (wire diameter, 0.075 mm; 80 mesh woven) placed
8 mm apart. The reference electrode is a silver wire coated with a layer
of AgCl, placed 2 mm from the working electrode. The electrical
connections to a potentiostat are made through Pt wires melt-sealed in
the plastic spacer mentioned above. The light path is masked so that
the beam passes through the center of the working electrode. The inlet
1992, 620.), see: (a) Kolivoska, V.; Gal
Hromadova, M. Phys. Chem. Chem. Phys. 2011, 13, 11422. (b) Pospísil,
L.; Hromadova, M.; Fanelli, N.; Valasek, M.; Kolivoska, V.; Gal, M.
Phys. Chem. Chem. Phys. 2011, 13, 4365. (c) Pospísil, L.; Hromadova,
M.; Gal, M.; Valasek, M.; Fanelli, N.; Kolivoska, V. Collect. Czech.
́ ́
, M.; Pospísil, L.; Valasek, M.;
́
́
́
́
́
́
́
Chem. Commun. 2009, 74, 1559. (d) Casado, J.; Patchkovskii, S.;
Zgierski, M. Z.; Hermosilla, L.; Sieiro, C.; Moreno Oliva, M.;
2703
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
Lopez Navarrete, J. T. Angew. Chem., Int. Ed. 2008, 47, 1443.
71, 3481. (e) Benniston, A. C.; Harriman, A. Chem. Soc. Rev. 2006, 35,
169.
(e) Pospísil, L.; Fiedler, J.; Hromadova,
Pecka, J.; Michl, J. J. Electrochem. Soc. 2006, 153, E179. (f) Funston, A.;
Kirby, J. P.; Miller, J. R.; Pospísil, L.; Fiedler, J.; Hromadova, M.;
Gal, M.; Pecka, J.; Valasek, M.; Zawada, Z.; Rempala, P.; Michl, J.
́ ́ ́
M.; Gal, M.; Valasek, M.;
(17) (a) Laine,
M.; Amouyal, E. J. Am. Chem. Soc. 2002, 124, 1364. (b) Laine,
Bedioui, F.; Amouyal, E.; Albin, V.; Berruyer−Penaud, F. Chem.Eur.
J. 2002, 8, 3162. (c) Laine, P.; Amouyal, E. Chem. Commun. 1999, 935.
(18) (a) Laine, P. P.; Campagna, S.; Loiseau, F. Coord. Chem. Rev.
2008, 252, 2552. (b) Laine, P. P.; Bedioui, F.; Loiseau, F.; Chiorboli,
C.; Campagna, S. J. Am. Chem. Soc. 2006, 128, 7510. (c) Laine, P. P.;
Loiseau, F.; Campagna, S.; Ciofini, I.; Adamo, C. Inorg. Chem. 2006,
45, 5538. (d) Laine, P. P.; Ciofini, I.; Ochsenbein, P.; Amouyal, E.;
Adamo, C.; Bedioui, F. Chem.Eur. J. 2005, 11, 3711. (e) Ciofini, I.;
Laine, P. P.; Bedioui, F.; Adamo, C. J. Am. Chem. Soc. 2004, 126,
10763.
́
P.; Bedioui, F.; Ochsenbein, P.; Marvaud, V.; Bonin,
́
P.;
́
́
́
́
J. Phys. Chem. A 2005, 109, 10862. (g) Porter, W. W.; Vaid, T. P.;
Rheingold, A. L. J. Am. Chem. Soc. 2005, 127, 16559. (h) Porter, W. W.;
Vaid, T. P. J. Org. Chem. 2005, 70, 5028.
(5) For selected work dealing with the multielectron reduction of
fullerenes, see: (a) Diener, M. D.; Alford, J. M. Nature 1998, 393, 668.
(b) Weaver, M. J.; Gao, X. J. Phys. Chem. 1993, 97, 332. (c) Yang, Y.;
Arias, F.; Echegoyen, L.; Chibante, L. P. F; Flanagan, S.; Robertson, A.;
́
́
́
́
́
Wilson, L. J. J. Am. Chem. Soc. 1995, 117, 7801. (d) Xie, Q.; Per
́
ez-
Cordero, E.; Echegoyen, L. J. Am. Chem. Soc. 1992, 114, 3978.
(19) (a) Benniston, A. C.; Harriman, A.; Li, P.; Rostron, J. P.;
Harrington, R. W.; Clegg, W. Chem.Eur. J. 2007, 13, 7838. (b) Beer,
P. D.; Chen, Z.; Grieve, A.; Haggitt, J. J. Chem. Soc., Chem. Commun.
1994, 2413. (c) Kodaka, M.; Kubota, Y. J. Chem. Soc., Perkin Trans. 2
1999, 891. (d) Willner, I.; Ayalon, A.; Rabinovitz, M. New J. Chem.
1990, 14, 685. (e) Chen, P.; Curry, M.; Meyer, T. J. Inorg. Chem. 1989,
28, 2271. (f) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637.
(g) Thummel, R. P.; Lefoulon, F.; Chirayil, S.; Goulle, V. J. Org. Chem.
(6) For selected examples concerning oxidizable electrophores (i.e.,
potentially multielectron donors), that is, star-shaped semirigid and
coupled multicenter electrophores, see: (a) Diallo, A. K.; Absalon, C.;
Ruiz, J.; Astruc, D. J. Am. Chem. Soc. 2011, 133, 629. (b) Lambert, C.
Angew. Chem., Int. Ed. 2005, 44, 7337 and references therein.
(7) Han, Z.; Vaid, T. P.; Rheingold, A. L. J. Org. Chem. 2008, 73, 445.
(8) (a) Evans, D. H. Chem. Rev. 2008, 108, 2113 and references
therein. (b) Evans, D. H.; Hu, K. J. Chem. Soc., Faraday Trans. 1996,
92, 3983.
1988, 53, 4745. (h) Hunig, S.; Berneth, H. Top. Curr. Chem. 1980, 92, 1.
̈
(20) The same trends are also often observed for electrophores that
function as electron donors (oxidation regime). Examples are mainly
found in the field of electromechanical actuators, even though the
interplay between structural and redox properties is envisaged from
the reverse standpoint of the above-mentioned one. The tuning of the
structure is the mechanical output whereas the electrochemistry is the
input. Owing to the nature of the electroactive components involved
(e.g., oligothienyl or analogues), the HOMO is in this case primarily
impacted and thus, oxidation processes. (a) Marsella, M. J.; Reid, R. J.;
Estassi, S.; Wang, L.-S. J. Am. Chem. Soc. 2002, 124, 12507.
(b) Jousselme, B.; Blanchard, P.; Gallego-Planas, N.; Delaunay, J.;
Allain, M.; Richomme, P.; Levillain, E.; Roncali, J. J. Am. Chem. Soc.
2003, 125, 2888. (c) Jousselme, B.; Blanchard, P.; Levillain, E.;
Delaunay, J.; Allain, M.; Richomme, P.; Rondeau, D.; Gallego-Planas,
N.; Roncali, J. J. Am. Chem. Soc. 2003, 125, 1363. (d) Jousselme, B.;
Blanchard, P.; Gallego-Planas, N.; Levillain, E.; Delaunay, J.; Allain, M.;
Richomme, P.; Roncali, J. Chem.Eur. J. 2003, 9, 5297. (e) Jousselme,
B.; Blanchard, P.; Allain, M.; Levillain, E.; Dias, M.; Roncali, J. J. Phys.
Chem. A 2006, 110, 3488.
(21) For one such rare example, but dealing with the steric control of
the electron-donating ability of a molecule (oxidation regime), see:
Bellec, N.; Boubekeur, K.; Carlier, R.; Hapiot, P.; Lorcy, D.; Tallec, A.
J. Phys. Chem. A 2000, 104, 9750.
(22) Bistable electrophores typically made up of a bridging
photochromic switchable unit and two pyridinium termini, for
instance, are beyond the scope of the present work. Indeed, the
energy change of redox-active MOs not only originates from structural
modifications but also largely stems from alterations in the electronic
structure of the photochromic linker; see: (a) Gilat, S. L.; Kawai, S. H.;
Lehn, J.-M. Chem.Eur. J. 1995, 1, 275. (b) Gilat, S. L.; Kawai, S. H.;
Lehn, J.-M. J. Chem. Soc., Chem. Commun. 1993, 1439.
(9) Fortage, J.; Tuyer
F.; Campagna, S.; Griveau, S.; Bedioui, F.; Ciofini, I.; Laine,
Chem.Eur. J. 2010, 16, 11047.
(10) Peltier, C.; Adamo, C.; Laine,
̀
as, F.; Ochsenbein, P.; Puntoriero, F.; Nastasi,
́
P. P.
́
P. P.; Campagna, S.; Puntoriero,
F.; Ciofini, I. J. Phys. Chem. A 2010, 114, 8434.
(11) Fortage, J.; Peltier, C.; Nastasi, F.; Puntoriero, F.; Tuyer
Griveau, S.; Bedioui, F.; Adamo, C.; Ciofini, I.; Campagna, S.; Laine,
P. J. Am. Chem. Soc. 2010, 132, 16700.
(12) Prototypical complex for a single-electron storage: (a) Astruc,
D.; Hamon, J.-R.; Althoff, G.; Roman, E.; Batail, P.; Michaud, P.;
Mariot, J.-P.; Varret, F.; Cozak, D. J. Am. Chem. Soc. 1979, 101, 5445.
(b) Hamon, J.-R.; Astruc, D.; Michaud, P. J. Am. Chem. Soc. 1981, 103,
758.
(13) For strategies involving purely organic molecules for two-
electron storage, referred to as “super electron donors” (SEDs), see for
instance ref 4g and the following: (a) Murphy, J. A.; Garnier, J.; Park,
S. R.; Schoenebeck, F.; Zhou, S.-Z.; Turner, A. T. Org. Lett. 2008, 10,
1227. (b) Murphy, J. A.; Zhou, S.-Z.; Thomson, D. W.; Schoenebeck,
F.; Mahesh, M.; Park, S. R.; Tuttle, T.; Berlouis, L. E. A. Angew. Chem.,
Int. Ed. 2007, 46, 5178. (c) Richardson, R. D.; Wirth, T. Chem. Unserer
Z. 2008, 42, 190.
(14) For strategies relying on C−C bond forming/breaking, but
involving inorganic including organometallic (supra)molecular com-
pounds, also for two-electron storage, see: (a) Venkatesan, K.;
Blacque, O.; Fox, T.; Alfonso, M.; Schmalle, H. W.; Berke, H.
Organometallics 2004, 23, 1183. (b) Camp, C.; Mougel, V.; Horeglad,
̀
as, F.;
P.
́
́
́
P.; Pecaut, J.; Mazzanti, M. J. Am. Chem. Soc. 2010, 132, 17374.
(c) Floriani, C.; Solari, E.; Franceschi, F.; Scopelliti, R.; Belanzoni, P.;
Rosi, M. Chem.Eur. J. 2001, 7, 3052. (d) Franceschi, F.; Solari, E.;
Scopelliti, R.; Floriani, C. Angew. Chem., Int. Ed. 2000, 39, 1685.
(e) Rosi, M.; Sgamellotti, A.; Franceschi, F.; Floriani, C. Chem.Eur.
J. 1999, 5, 2914. (f) Franceschi, F.; Solari, E.; Floriani, C.; Rosi, M.;
Chiesi-Villa, A.; Rizzoli, C. Chem.Eur. J. 1999, 5, 708.
(23) (a) Bally, T. Nat. Chem. 2010, 2, 165. (b) Puschmann, F. F.;
Harmer, J.; Stein, D.; Ruegger, H.; de Bruin, B.; Grutzmacher, H.
̈
̈
Angew. Chem., Int. Ed. 2010, 49, 385.
(15) For selected examples of alternative strategies that rely on
nanoscale architectures like macrocycles or dendrimers decorated with
redox-active components, which could function as molecular batteries,
see: (a) Mayor, M.; Lehn, J.-M. J. Am. Chem. Soc. 1999, 121, 11231.
(b) Astruc, D. New J. Chem. 2011, 35, 764. and references therein.
(c) Ronconi, C. M.; Stoddart, J. F.; Balzani, V.; Baroncini, M.; Ceroni,
P.; Giansante, C.; Venturi, M. Chem.Eur. J. 2008, 14, 8365.
(16) Selected recent references: (a) Meylemans, H. A.; Damrauer, N.
H. Inorg. Chem. 2009, 48, 11161. (b) Hanss, D.; Wenger, O. S. Eur. J.
Inorg. Chem. 2009, 3778. (c) Indelli, M. T.; Chiorboli, C.; Flamigni, L.;
De Cola, L.; Scandola, F. Inorg. Chem. 2007, 46, 5630. (d) Benniston,
A. C.; Harriman, A.; Li, P.; Patel, P. V.; Sams, C. A. J. Org. Chem. 2006,
(24) Bendikov, M.; Wudl, F.; Perepichka, D. F. Chem. Rev. 2004, 104,
4891.
(25) (a) Hogiu, S.; Dreyer, J.; Pfeiffer, M.; Brzezinka, K.-W.;
Werncke, W. J. Raman Spectrosc. 2000, 31, 797. (b) Catalan, J.; Garcia
de Paz, J. L.; Reichardt, C. J. Phys. Chem. A 2010, 114, 6226.
(26) A similar strategy based on chemical substitution was, for instance,
applied by Hapiot, Lorcy and co-workers to investigate the potential
inversion observed in vinylogous tetrathiafulvalene derivatives: see ref 21.
(27) Amongst the 19 model electrophores and 13 ferrocenyl-based
dyads show in Chart 1, 14 electrophores and 9 dyads are new. These
were synthesized for the present study, with the exception of (q)Lu-
XP along with the associated ferrocenyl dyad, which could not be
2704
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705

Journal of the American Chemical Society
Article
obtained (Supporting Information). The other following compounds
are described in previous papers: Me-TP and its related dyad,11 Ph-
TP9,17,18 and its related dyad,9 Ph-XP and its related dyad,9 py-TP and
its related dyad,11 as well as qPy-TP.11
J. Electroanal. Chem. 1979, 105, 403. (c) Raghavan, R.; Iwamoto, R. T.
J. Electroanal. Chem. 1979, 102, 85.
(45) Ghaddar, T. H.; Wishart, J. F.; Thompson, D. W.; Whitesell, J.
K.; Fox, M. A. J. Am. Chem. Soc. 2002, 124, 8285.
(46) Lacoste, M.; Rabaa,
Saillard, J.-Y.; Le Beuze, A. J. Am. Chem. Soc. 1990, 112, 9548.
(47) (a) Valasek, M.; Pecka, J.; Jindrich, J.; Calleja, G.; Craig, P. R.;
̂
H.; Astruc, D.; Ardoin, N.; Varret, F.;
(28) Synthesized ferrocenyl-derivatized electrophores, referred to as
dyads, actually show two types of connection scheme: linkage of the
two components via the N-pyridinio phenyl group (cases of Ph-TP
and Ph-XP; see ref 9 and Figure 2) or via the phenyl at position 4 of
the pyridinium core (all other cases; Chart 1). Electrochemical
behaviors are not sensitive to this topological difference and allow
direct comparison of the various dyads, at least for RDE experiments.
(29) The covalent linking within these dyads allowed to avoid bias
inherent in different diffusion coefficients when studying bimolecular
systems (e.g., electrophore + Fc).
́
Michl, J. J. Org. Chem. 2005, 70, 405. (b) Berlin, Y. A.; Hutchison, G.
R.; Rempala, P.; Ratner, M. A.; Michl, J. J. Phys. Chem. A 2003, 107,
3970. (c) Schwab, P. F. H.; Levin, M. D.; Michl, J. Chem. Rev. 1999,
99, 1863.
(48) Hromadova,
́ ́ ́
M.; Kolivoska, V.; Gal, M.; Pospísil, L.; Sokolova,
R.; Valasek, M. J. Incl. Phenom. Macrocycl. Chem. 2011, 70, 461.
́
(49) Namazian, M.; Almodarresieh, H. A.; Noorbala, M. R.; Zare, H.
R. Chem. Phys. Lett. 2004, 396, 424.
(30) The slight discrepancy noted in the case of Ar = qPy is ascribed
to self-interaction error; refer to discussion in Supporting Information
that also deals with the associated case of qLu-TP.
(50) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158.
(51) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., Ed.; Plenum: New York, 1976; pp 1−28.
(52) Yanai, T.; Tew, D.; Handy, N. Chem. Phys. Lett. 2004, 393, 51.
(53) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(54) Scalmani, G.; Frisch, M. J.; Mennucci, B.; Tomasi, J.; Cammi, R.;
Barone, V. J. Chem. Phys. 2006, 124, 094107.
(55) Frisch, M. J. et al. Gaussian 09; Gaussian Inc.: Wallingford CT,
2010.
(56) Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta 2000, 298,
97.
(31) From CV and RDE experiments performed on the ferrocenyl
dyad related to Ph-DMTP, the number of exchange electrons involved
in the reduction process is 1.3 instead of 2, as expected from
calculations (see Supporting Information). This discrepancy does not
call into question the relevance of our present demonstration. In fact,
dynamic fluxionality and distortion are key features of these EP
electrophores, and it is recognized that the time scale of several
elementary processes involved in the complex ECE process of
potential compression may lead to a situation where the number of
electrons involved is between 1 and 2. For the in-depth study of a
molecule behaving in such an unusual manner, see ref 32 and
references therein.
(57) Krejcí
179.
̌
k, M.; Danek, M.; Hartl, F. J. Electroanal. Chem. 1991, 317,
(58) http://www.reading.ac.uk/web/FILES/chemistry/F_Hartl_
OTTLE_cells.pdf
(59) http://www.specac.com/products/liquid-transmission-cell/ft-ir-
(32) Felton, G. A. N.; Petro, B. J.; Glass, R. S.; Lichtenberger, D. L.;
Evans, D. H. J. Am. Chem. Soc. 2009, 131, 11290.
(33) See Note S2 in Supporting Information.
liquidtransmission-cell/530
(34) Due to the complexity and flexibility of the systems, the
evaluation of the barrier height represents a prohibitive task for the
used computational approach.
(35) Saveant, J.-M. Elements of Molecular and Biomolecular Electro-
́
chemistry. An Electrochemical Approach to Electron Transfer Chemistry;
Wiley-Interscience: Hoboken, NJ, 2006.
(36) Macías-Ruvalcaba, N. A.; Evans, D. H. Chem.Eur. J. 2007, 13,
4386.
(37) Kraiya, C.; Singh, P.; Evans, D. H. J. Electroanal. Chem. 2004,
563, 203.
(38) The full absorption spectrum of [qLu-TP]+ was recorded
between 250 and 1800 nm in a separate spectroelectrochemical
experiment, confirming the existence of the electronic transition
computed at 1370 nm. A very broad band with λmax at ca. 1250 nm was
indeed observed (Supporting Information). Only the onset of this NIR
band is shown in Figure 9.
(39) Jacquemin, D.; Perpet
Res. 2009, 42, 326.
̀
e, E.; Ciofini, I.; Adamo, C. Acc. Chem.
(40) (a) Itoh, M.; Kosower, E. M. J. Am. Chem. Soc. 1968, 90, 1843.
(b) Kimura, K.; Yamada, H.; Tsubomura, H. J. Chem. Phys. 1968, 48,
440. (c) Macías-Ruvalcaba, N. A.; Evans, D. H. J. Phys. Chem. C 2010,
114, 1285. (d) Macías-Ruvalcaba, N. A.; Evans, D. H. J. Phys. Chem. C
2007, 111, 5805. (e) Macías-Ruvalcaba, N. A.; Felton, G. A. N.; Evans,
D. H. J. Phys. Chem. C 2009, 113, 338.
(41) For an illustration of the dramatic effect that solvent can have on
the electrochemical behavior of certain electrophores, see: Hapiot, P.;
Kispert, L. D.; Konovalov, V. V.; Savea
123, 6669.
́
nt, J.-M. J. Am. Chem. Soc. 2001,
(42) Ciofini, I.; Adamo, C.; Chermette, H. J. Chem. Phys. 2005, 123,
121102.
(43) In this respect, electron density difference maps are usually the
more informative, since they provide a picture of the change in the
relaxed electron density. However, in the present case the overall
picture obtained from the analysis of these MOs is qualitatively the
same.
(44) (a) Kosower, E. M.; Cotter, J. L. J. Am. Chem. Soc. 1964, 86,
5524. (b) Yasukouchi, K.; Taniguchi, I.; Yamaguchi, H; Shiraishi, M.
2705
dx.doi.org/10.1021/ja210024y | J. Am. Chem.Soc. 2012, 134, 2691−2705