Tunneling versus hopping in mixed-valence oligo- p -phenylenevinylene polychlorinated bis(triphenylmethyl) radical anions

Article abstract of DOI:10.1021/ja1083859

Radical anions 1^-?-5^-?, showing different lengths and incorporating up to five p-phenylenevinylene (PPV) bridges between two polychlorinated triphenylmethyl units, have been prepared by chemical or electrochemical reductions from the corresponding diradicals 1-5 which were prepared using Wittig-Horner-type chemistry. Such radical anions enabled us to study, by means of UV-vis-NIR and variable-temperature electron spin resonance spectroscopies, the long-range intramolecular electron transfer (IET) phenomena in their ground states, probing the influence of increasing the lengths of the bridges without the need of using an external bias to promote IET. The temperature dependence of the IET rate constants of mixed-valence species 1 ^-?-5^-? revealed the presence of two different regimes at low and high temperatures in which the mechanisms of electron tunneling via superexchange and thermally activated hopping are competing. Both mechanisms occur to different extents, depending on the sizes of the radical anions, since the lengths of the oligo-PPV bridges notably influence the tunneling efficiency and the activation energy barriers of the hopping processes, the barriers diminishing when the lengths are increased. The nature of solvents also modifies the IET rates by means of the interactions between the oligo-PPV bridges and the solvents. Finally, in the shortest compounds 1 ^-? and 2^-?, the IET induced optically through the superexchange mechanism can also be observed by the exhibited intervalence bands, whose intensities decrease with the length of the PPV bridge.

Full text of DOI:10.1021/ja1083859

ARTICLE
pubs.acs.org/JACS
Tunneling versus Hopping in Mixed-Valence
Oligo-p-phenylenevinylene Polychlorinated Bis(triphenylmethyl)
Radical Anions
Vega Lloveras,† Josꢀe Vidal-Gancedo,† Teresa M. Figueira-Duarte,‡ Jean-Franc-ois Nierengarten,‡
Juan J. Novoa,§ Fernando Mota,§ Nora Ventosa,† Concepciꢀo Rovira,*,† and Jaume Veciana*,†
†Institut de Ciꢁencia de Materials de Barcelona (CSIC) and Networking Research Center on Bioengineering, Biomaterials and
Nanomedicine (CIBER-BBN), Campus Universitari de Bellaterra, E-08193 Cerdanyola, Spain
‡Laboratoire de Chimie des Matꢀeriaux Molꢀeculaires, Ecole Europꢀeenne de Chimie, Polymꢁeres et Matꢀeriaux, Universitꢀe de Strasbourg et
CNRS (UMR 7509), 25 rue Becquerel, 67087 Strasbourg, Cedex 2, France
§Departament de Química Física, Facultat de Química, Universitat de Barcelona, Av. Diagonal 47, 08208 Barcelona, Spain
S Supporting Information
b
ABSTRACT: Radical anions 1ꢀ•ꢀ5ꢀ•, showing different
lengths and incorporating up to five p-phenylenevinylene
(PPV) bridges between two polychlorinated triphenylmethyl
units, have been prepared by chemical or electrochemical
reductions from the corresponding diradicals 1ꢀ5 which were
prepared using WittigꢀHorner-type chemistry. Such radical
anions enabled us to study, by means of UVꢀvisꢀNIR and
variable-temperature electron spin resonance spectroscopies,
the long-range intramolecular electron transfer (IET) phenom-
ena in their ground states, probing the influence of increasing
the lengths of the bridges without the need of using an external
bias to promote IET. The temperature dependence of the IET rate constants of mixed-valence species 1ꢀ•ꢀ5ꢀ• revealed the
presence of two different regimes at low and high temperatures in which the mechanisms of electron tunneling via superexchange
and thermally activated hopping are competing. Both mechanisms occur to different extents, depending on the sizes of the radical
anions, since the lengths of the oligo-PPV bridges notably influence the tunneling efficiency and the activation energy barriers of the
hopping processes, the barriers diminishing when the lengths are increased. The nature of solvents also modifies the IET rates by
means of the interactions between the oligo-PPV bridges and the solvents. Finally, in the shortest compounds 1ꢀ• and 2ꢀ•, the IET
induced optically through the superexchange mechanism can also be observed by the exhibited intervalence bands, whose intensities
decrease with the length of the PPV bridge.
’ INTRODUCTION
bridge between the D/A moieties. The distance dependence of ET
rateconstants, kET, is oftendescribedby kET = k0 exp(ꢀβd), where
k0 is a kinetic pre-exponential factor, d isthe DꢀA center-to-center
distance, and β is the attenuation factor that depends primarily on
the nature of the molecular bridge. Theory predicts that another
regime may also exist wherein the distance dependence is very
weak and in which the molecular bridge essentially acts as an
incoherent molecular wire, yielding a reciprocal scaling of kET with
the distance.4ꢀ11 Therefore, beyond the range of direct DꢀA
electronic overlap, IET may occur either through a bridge-
mediated superexchange — also known as direct (nonresonant)
tunneling — between D/A electronic states or through an inco-
herent hopping. The relative contributions of these mechanisms
depend in part on the electronic nature of the bridge as well
Long-range electron transfer (ET) between donor (D) and
acceptor (A) molecules/groups where the separation between
the donor and acceptor greatly exceeds their spatial extent is a
subject of considerable interest for studies of ET mechanisms in
biological systems and in nanoscale circuits based on single
molecular devices.1ꢀ3 Electron transfer in donorꢀbridgeꢀ
acceptor (D-B-A) molecules has been actively investigated, in
part to determine which aspects of the bridge structure control
the intramolecular electron transfer (IET).4 Both the electronic
structure of the bridge and the effective distance between the
D/A centers are known to play a critical role in determining the
ease of the ET.4 However, very little is known about the
influence of external factors, like the solvent media and tem-
perature-induced perturbations, on the IET of nonperturbed
ground states of such systems. In most of the small D-B-A
systems, the ET rate scales exponentially with the length of the
Received: September 16, 2010
Published: March 29, 2011
r
2011 American Chemical Society
5818
dx.doi.org/10.1021/ja1083859 J. Am. Chem. Soc. 2011, 133, 5818–5833
|

Journal of the American Chemical Society
ARTICLE
Chart 1
as on the distance between the D/A centers.4 In general, tunneling
and hopping pathways operate in parallel, and the overall rate
constant of an IET will have contributions from both
mechanisms.12
symmetrical neutral R-B-R derivatives, are excellent candidates
for studying the different factors influencing IETs, like the length
and architectures of the bridges, as well as the temperature and
the solvent. Advantageously, IET in such compounds can be
studied without the need to perturb their ground states by
applying an external bias using the scanning tunneling micro-
scopy (STM) break junction technique or conducting atomic
Mixed-valence compounds of the type Rꢀ-B-R (or R-B-Rþ)
containing two linked redox centers in different formal oxidation
states, generated by the partial reduction (or oxidation) of
5819
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthetic Route for Diradicals 1ꢀ5
force microscopy (c-AFM), or by exciting the molecules using
optical pumpꢀprobe spectroscopy.4
in the excited state, while in our study the IET is observed in
the ground state.
Herein, we describe a series of structurally well-defined
diradicals 1ꢀ5 (Chart 1) in which oligo-p-phenylenevinylene
(oligo-PPV) bridges of different lengths are connecting two
polychlorotriphenylmethyl (PTM) radical units as redox
centers. Such PTM radical units are particularly interesting
not only because they have a large persistence, displaying high
thermal and chemical stabilities, but also because they are
electroactive species giving rise, either chemically or electro-
chemically, to the corresponding anions and cations.13 The
resulting anions are also quite stable species, while the cations
are highly reactive toward H2O, thus limiting their use.14 By
partial chemical or electrochemical reduction of diradicals
1ꢀ5, we generated the mixed-valence radical anions 1ꢀ•ꢀ5ꢀ•
and studied the optically and thermally induced IET processes
from the UVꢀvisꢀNIR spectra and variable-temperature
electron spin resonance (VT-ESR) spectra of these
species.15,16 The controlled molecular structure in this series
of compounds allows for probing the effects of the distance on
the rate at which ET occurs in these mixed-valence systems, as
well as the influence of solvent and temperature on IET
phenomenon and the molecular-wire behavior of the oligo-
PPVs. It is worth mentioning that only a few reports on
temperature control of ET through π-conjugated chains of
oligo-PPVs are known.9 Moreover, in many of those studies
IETs were investigated using femtosecond to nanosecond
optical pumpꢀprobe spectroscopy, where the ET takes place
’ RESULTS AND DISCUSSION
Diradicals. Synthesis. The synthetic route for preparing dir-
adicals 1ꢀ5 is based on the reaction of phosphonium bromide 15
with the suitable oligo-PPV dialdehydes 6ꢀ10 under Wittig
conditions followed by the in situ generation and oxidation of the
corresponding dicarbanions with p-chloranil (Scheme 1).
The substituted phosphonium bromide 1513c and dialdehydes
6, 8, and 1017 were prepared according to previously reported
procedures. Dialdehydes 7 and 9 were obtained for the first time
following the synthesis depicted in Scheme 2. The strategy
employed for the preparation of such dialdehydes is based upon
WittigꢀHorner-type chemistry. Thus, the treatment of aldehyde
17 with phosphonate 16 in THF in the presence of t-BuOK
afforded the targeted protected stilbene derivative 18, but as an
1
E:Z isomer mixture, as deduced from the H NMR spectrum.
The latter observation is quite surprising, as the reaction of
benzylic phosphonates with aromatic aldehydes under Wittig
ꢀHorner conditions is generally stereoselective, leading to the E
isomer only.17,18 In the present case, steric hindrance resulting
from the presence of the two octyloxy groups in the ortho
positions of the reactive groups in both 16 and 17 may explain
the lack of E:Z selectivity. The isomerization of the isomeric
mixture to the E derivative 18 was easily achieved by treatment
with a catalytic amount of iodine in refluxing toluene,
5820
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Scheme 2. Synthetic Route for Obtaining Dialdehydes 7 and 9a
a Reagents and conditions: (i) t-BuOꢀKþ/THF then I2 (cat.)/toluene; (ii) CF3CO2H/CH2Cl2/H2O; (iii) 16/t-BuOꢀKþ/THF then I2 (cat.)/toluene;
(iv) CF3CO2H/CH2Cl2/H2O.
Chart 2
Table 1. Redox Potentials of Compounds 1ꢀ14 in V vs Ag/
AgNO3 in CH2Cl2
compd
E0red/V
E0ox1/V
1
ꢀ0.18
ꢀ0.20
ꢀ0.22
ꢀ0.22
ꢀ0.18
1.14
0.96
0.86
0.82
0.80
1.79
1.04
1.03
0.85
0.80
1.05
0.93
0.83
0.78
2
3
4
5
6
7
8
and the compound (E)-18 was thus obtained in 78% yield.
Subsequent deprotection with CF3CO2H in CH2Cl2/H2O
afforded the dialdehyde 7 in 74% yield. Reaction of 7 with
phosphonate 16, under WittigꢀHorner conditions, followed by
treatment with a catalytic amount of iodine in refluxing toluene
then gave the all-E p-phenylenevinylene trimer 19 in 43% yield.
Deprotection by treatment with CF3CO2H in CH2Cl2/H2O
afforded dialdehyde 9.
The coupling of the substituted phosphonium bromide 15
with the suitable dialdehydes 6ꢀ10 was performed under Wittig
conditions to give the corresponding polychlorinated bis-
(triphenylmethanes). This reaction has been strongly stereose-
lective in other related compounds since it yields exclusively the
E/E isomers, as ascertained by NMR spectroscopy.13c Such
stereoselectivity is justified if we consider that the ylide derived
from the phosphonium bromide 15 is stabilized by the presence
of the polychlorinated aromatic ring. Actually, the natural pre-
ference of ylides with electron-withdrawing groups that stabilize
the betaine form has been experimentally shown to give mainly
the E isomer.19 In addition, the Z/E isomer distribution of the
Wittig products is also strongly influenced by the nature of the
9
10
11
12
13
14
ꢀ0.19
ꢀ0.20
ꢀ0.20
ꢀ0.18
base used in the preparation of the ylide.20 Indeed, potassium
tert-butoxide is the base of choice for maximizing the yields of E-
olefins.18,21
In order to optimize the yield of the desired diradical
compounds, the bis(triphenylmethane) precursors were not
isolated, and the synthesis of the diradical species 1ꢀ5 was
performed as a “one-pot” reaction by treatment of the corre-
sponding bis(triphenylmethane) precursor with an excess of
n-Bu4NþOHꢀ and subsequent oxidation of the resulting dianions
with p-chloranil. Apart from compound 1, diradicals 2ꢀ5 were
always isolated as the minor products and the corresponding
radical monoadducts 11ꢀ14 were the major ones, whatever the
reaction conditions. As already reported, the yields of Wittig-type
5821
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Table 2. UVꢀVis Data for Compounds 1ꢀ14 in CH2Cl2
compd
λmax/nm (10ꢀ3ε/Mꢀ1 cmꢀ1
)
1
321 (15.7), 368 (sh)
338 (21.1), 368 (sh)
341 (27.4), 369 (sh)
369 (sh)
388 (52.3)
389 (58.1)
390 (61.5)
389 (64.3)
389 (69.1)
418 (sh)
509 (11.8), 680 (3.8)
547 (12.2), 747 (4.9)
558 (sh), 778 (5.7)
579 (sh), 780 (6.2)
585 (sh), 782 (5.9)
2
434 (55.3)
454 (78.7)
462 (88.0)
475 (121.9)
403 (6.3)
3
4
5
334 (34.3), 369 (sh)
274 (16.0), 282 (13.3)
331 (17.3)
6
7
423 (27.1), 476 (sh)
446 (57.4)
462 (67.5)
470 (102.2)
427 (45.3)
450 (72.1)
462 (84.4)
473 (116.2)
8
275 (20.1), 282 (17.8), 341 (19.2)
336 (19.7)
9
10
11
12
13
14
333 (26.6)
334 (20.9), 367 (sh)
339 (22.2), 368 (sh)
333 (sh), 368 (sh)
333 (28.1), 368 (sh)
390 (42.6)
390 (44.2)
389 (46.1)
390 (47.8)
542 (7.2), 745 (2.1)
556 (sh), 776 (2.6)
574 (sh), 777 (2.3)
582 (sh), 780 (2.9)
reactions involving bis-aldehydes 6ꢀ10 diminish drastically with the
length of the conjugated system.17 It is also worth noticing that all the
obtained diradicals 1ꢀ5 are highly persistent and thermally stable in
the solid state and also in diluted solutions, even when exposed to air.
The solubility of diradicals 1ꢀ5 is very high in many solvents,
contrasting with the behavior of the previously reported diradical 20,
lacking the two octyloxy substituents (Chart 2).13a
The absorption spectra of diradicals 1ꢀ5 (Figure 1) and
monoradicals 11ꢀ14 also exhibit the characteristic πꢀπ* transi-
tions of the oligo-PPV units. In both series, this band is red-
shifted upon increasing the size of the π-conjugated bridging
system (Table 2). When comparing compounds 6ꢀ10 with the
corresponding monoradicals 11ꢀ14, the πꢀπ* transitions is
slightly red-shifted and the molar extinction coefficient larger. A
similar trend is observed when comparing monoradicals 11ꢀ14
with the corresponding diradicals 1ꢀ5 (see Figure S1, Support-
ing Information). This fact is likely the result of an enhanced
conjugation length of the PPV-type system due to the presence of
the terminal PTM radical group. On the other hand, the presence
of electronic delocalization is also observed in the two low-energy
radical bands that appear around 509ꢀ585 and 680ꢀ782 nm, in
which bathochromic shifts and enhanced absorptivities cause the
conjugation of the oligo-PPV bridges to increase. As previously
noted, increasing the number of PPV fragments in the mono-
radicals also gives rise to small red shifts and enhanced absorp-
tivities of these bands, although to a lesser extent than in the
diradicals.
It is worth mentioning that the changes in these low-energy
bands of the diradicals are much more pronounced when going
from diradical 1 to 3 than from 3 to 5. Such low-energy radical
bands are assigned to charge transfer from the electron-donating
PPV bridge to the electron-accepting PTM radical groups. This
assignment is confirmed by the absence of such absorptions in
the corresponding dianionic species generated from the radicals
(vide infra), where the acceptor character of PTM units is absent,
and is also supported by the solvent dependence shift, typical of
charge-transfer bands, exhibited by the low-energy radical bands
of diradical 1. On the other hand, the most characteristic radical
absorption, attributable to the HOMOꢀSOMO transition,24
appears in all radicals at 389 nm without showing any significant
difference with the size of the bridging conjugated system.
Electron Spin Resonance. X-band ESR spectra of diradicals
1ꢀ5 and monoradicals 11ꢀ14 were recorded in CH2Cl2/
toluene (1:1) solution in the temperature range of 140ꢀ300
K. In both cases, the ESR spectra at 300 K are practically identical
along the series (Table 3). As representative examples, the
experimental and simulated spectra of diradical 3 and mono-
radical 12 are shown in Figures 2 and 3, respectively.
Electrochemistry. The redox properties of compounds 1ꢀ14
were studied in solution by cyclic voltammetry (CV) (Table 1).
Cyclic voltammograms were recorded in CH2Cl2, with [(n-Bu)4N]
PF6 (0.1 M) as supporting electrolyte, a Pt wire as a working
electrode, and Ag/Agþ as the reference electrode. Monoradicals
11ꢀ14 show one reversible reduction process between ꢀ0.18 and
ꢀ0.20 V, corresponding to the reduction of the radical units to the
corresponding monoanions. For diradicals 1ꢀ5, the redox processes
of the two triphenylmethyl units are too close to be distinguished,
and only one reversible two-electron reduction process is observed
between ꢀ0.18 and ꢀ0.22 V. This result suggests the presence of
weak electronic interactions between the triphenylmethyl units of
diradicals 1ꢀ5, since in the case of a strong or moderate electronic
interaction between the two electroactive units, two electrochemical
waves are anticipated. The values of the redox potentials for the
reversible reduction process of all obtained monoradicals are similar
among them (see Table 1) and equal to those of the diradicals,
indicating a small influence of the π-conjugation of oligo-PPV
fragments on the SOMO levels of the radicals.
In oxidation, the cyclic voltammograms of monoradicals
11ꢀ14 and diradicals 1ꢀ5, as well as those of the corresponding
dialdehydes precursors 6ꢀ10, exhibit from one to three rever-
sible (or quasi-reversible) one-electron oxidation waves centered
on the oligo-PPV fragments.22 For the entire series of com-
pounds, the first oxidation potential decreases upon increasing
the conjugation length of the PPV core (Table 1).
Absorption Spectra. UVꢀvis absorption spectra of diradicals
1ꢀ5, dialdehydes 6ꢀ10, and monoradicals 11ꢀ14 were re-
corded in CH2Cl2, and the most relevant data are summarized in
Table 2. For compounds 6ꢀ10, the λmax of the absorption band
corresponding to the characteristic πꢀπ* transitions of oligo-
PPV units is progressively red-shifted from 403 to 470 nm when
the conjugation length is increased. At the same time, the molar
extinction coefficient is increased in the series, as classically
observed for such systems.22,23
5822
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Figure 1. UVꢀvis spectra of diradicals 1ꢀ5 in CH2Cl2.
Table 3. EPR Dataa for Diradicals 1ꢀ5 and Monoradicals
11ꢀ14 Obtained in CH2Cl2/Toluene (1:1) Solution
constants of monoradicals. As the isotropic hyperfine coupling
constants for diradicals 1ꢀ5 are half those found for the related
monoradicals 11ꢀ14 (see Table 3), we can conclude that the
two electrons in diradicals 1ꢀ5 are magnetically interacting, with
an electron exchange interaction constant, J, that fulfills the
condition |J| . |ai|.25 The isotropic giso values for all studied
mono- and diradicals are very close to that of the free electron
(Table 3), as already observed for other substituted PTM
radicals.13,14
compd T/K
g
aH
aC,R aC,arom ΔHpp
1
220 2.0025
0.88 (2 H)
14.8 6.5, 5.4
0.88b
0.32
0.36
0.40
0.40
2
220 2.0022 0.88 (2 H), 0.34 (2 H) 14.5 6.2, 5.1
220 2.0022 0.93 (2 H), 0.36 (2 H) 14.6 6.2, 5.1
220 2.0024 0.87 (2 H), 0.37 (2 H) 14.1 6.2, 5.3
230 2.0023 0.90 (2 H), 0.38 (2 H) 13.6 6.2, 5.4
3
4
5
Spectra of diradicals were also recorded in a frozen CH2Cl2/
toluene (1:1) mixture at 140 K in order to determine the absolute
values of zero-field-splitting parameters, |D/hc| and |E/hc|, that
arise from the dipolar magnetic interactions between the two
unpaired electrons. The values of these parameters for diradical 1
are |D/hc| = 3.9 ꢁ 10ꢀ4 cmꢀ1 (or D = 8.5 G) and |E/hc| ≈ 0,
which are equal to those of diradical 20, lacking the two octyloxy
substituents.26 Such an agreement reveals a similar dipolar
magnetic interaction between the two unpaired electrons of both
diradicals and a weak influence of the two electron-withdrawing
octyloxy susbtituents of the central phenyl ring on the distribu-
tion of the two unpaired electrons. From the zero-field-splitting
parameter |D/hc|, one may also estimate an effective interspin
separation, R, assuming that D = 3g2β2/2R3,26ꢀ28 strictly valid
only for localized diradicals, is applicable to the delocalized
diradicals studied here. Indeed, one may estimate an effective
distance of 15 Å for diradical 1, which is somewhat smaller than
the nominal separation of 18.8 Å between the two methyl carbon
atoms calculated from optimized structures using AM1 semi-
empirical minimizations. This result is in agreement with a
certain degree of electronic delocalization for 1. For the rest of
diradicals (2ꢀ5), their D values are either very small (D e 1 G)
or undetectable by conventional continuous-wave ESR, as ex-
pected from the large distances between the two methyl carbon
atoms, estimated by theoretical calculations as 18.8, 25.4, 31.9,
38.2, and 44.7 Å for 1ꢀ5, respectively.28 In line with such large
distances is the fact that any of the forbidden ΔMs = 2 transitions
were observed for the studied diradicals at half-field in the ESR
spectra at 140 K.
11
12
13
14
1ꢀ•
2ꢀ•
3ꢀ•
4ꢀ•
5ꢀ•
220 2.0020 1.80 (1 H), 0.70 (1 H) 29.1 12.7, 10.2 0.56
220 2.0022 1.75 (1 H), 0.70 (1 H) 28.9 12.7, 10.2 0.62
220 2.0022 1.80 (1 H), 0.70 (1 H) 29.0 12.5, 10.1 0.65
240 2.0023 1.80 (1 H), 0.70 (1 H) 29.3 12.5, 10.2 0.69
200 2.0025
1.70 (1 H)
1.0
205 2.0022 1.65 (1 H), 0.72 (1 H)
220 2.0024 1.70 (1 H), 0.70 (1 H)
220 2.0024 1.68 (1 H), 0.72 (1 H)
240 2.0023 1.70 (1 H), 0.70 (1 H)
0.85
0.70
0.80
0.80
a g, Lande factor ((0.0003); aX, hyperfine coupling constant (in gauss)
with atom X; ΔHpp, peak-to-peak line width (in gauss). b The larger line
width of 1 is ascribed to the larger dipolar interaction in this diradical.
At 300 K, monoradicals showed ESR spectra with two
symmetrical main lines originated by the hyperfine coupling of
the unpaired electron with the hydrogen atom of the ethylene
group next to the PTM unit. Several weak satellite lines corre-
sponding to the hyperfine couplings with the naturally abundant
13C nuclei of the PTM unit (Table 3) were also observed. When
the temperature was lowered, the main lines narrowed, thus
making it possible to distinguish an additional hyperfine coupling
with another hydrogen atom of the ethylene moiety, with a
smaller hyperfine coupling constant, yielding a spectrum con-
sisting of four main lines (see Figure 3b and Table 3). Spectra of
diradicals 1ꢀ5 at 300 K are different since they exhibit three
overlapped symmetrical main lines, which are each split into two
lines when the temperature is lowered (see Figure 2b), with
isotropic hyperfine coupling constants that are approximately
half those of the monoradicals. The 13C hyperfine coupling
constants in the biradicals are also half the values of the coupling
Radical Anions. Generation of Radical Anions. Mixed-valence
radical anions 1ꢀ•ꢀ5ꢀ•were generated by the chemical reduc-
tion of diradicals 1ꢀ5 with metallic Cu under conditions
5823
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Figure 2. Experimental and simulated isotropic solution EPR spectra of diradical 3 in CH2Cl2/toluene (1:1) solution at 300 (a) and 220 K (b).
Figure 3. Experimental and simulated isotropic solution EPR spectra of monoradical 12 in CH2Cl2/toluene (1:1) solution at 300 (a) and 220 K (b).
5824
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Figure 4. Top: Evolution of the UVꢀvis during the course of the reduction of diradical 2 in CH2Cl2 with metallic Cu. Inset shows the intervalence band
at the NIR region of radical anion 2ꢀ•. Bottom: Experimental intervalence band (solid line) of 2ꢀ•, which was deconvoluted by means of two Gaussian
functions (dashed lines). Notice that the sum of the dashed lines (dotted line) closely matches the experimental spectrum.
Table 4. UVꢀVis Data for Anions 12ꢀꢀ52ꢀ and 11ꢀꢀ14ꢀ in
previously optimized for the reduction of (2,3)-(6,7)-bis(1,4-
dioxo-1,4-dihydrobenzo)tetrathiafulvalene.16d,27 Thus, the re-
CH2Cl2 with [(Ph)4P]Br (0.1 M)
duction was performed on stirred 1 ꢁ 10ꢀ4ꢀ2.5 ꢁ 10ꢀ4
M
compd
λmax/nm (10ꢀ3ε/Mꢀ1 cmꢀ1
)
solutions of 1ꢀ5 in CH2Cl2 with [(Ph)4P] Br (0.1 M) as
supporting electrolyte and using a wire of metallic copper as
the reducing chemical agent. This process takes place smoothly
and gradually, and it can be stopped at any point just by removing
the copper wire from the solution. The reduction progress was
followed continuously by UVꢀvisꢀNIR spectroscopy until the
dianions were fully generated. As an example, Figure 4 illustrates
the evolution of the absorption spectra during the course of the
reduction of diradical 2, where the formation of the radical anion
12ꢀ
22ꢀ
32ꢀ
42ꢀ
52ꢀ
11ꢀ
12ꢀ
13ꢀ
14ꢀ
525 (44.3)
526 (48.0)
526 (53.8)
526 (57.1)
612 (45.9)
609 (52.4)
606 (56.9)
604 (58.2)
602 (66.2)
606 (28.3)
603 (31.9)
601 (33.6)
599 (36.0)
421 (42.5)
445 (69.6)
457 (79.0)
473 (115.8)
421 (38.2)
447 (65.7)
459 (78.1)
472 (111.3)
525 (27.8)
525 (31.4)
525 (32.3)
2
ꢀ• followed by the dianion 22ꢀ is observed. The observation of
some well-defined isosbestic points during the course of the
reduction indicates that no byproducts due to decompositions
are originated during this smooth process. For comparison
purposes, the monoanions 11ꢀꢀ14ꢀ were also generated from
the radicals 11ꢀ14 using the same reduction procedure. All
relevant spectroscopic data of the final products —the dianions
produced by the oligo-PPV moieties are blue-shifted with respect to
those exhibited by the neutral radicals and their intensity decreases,
showing that, in the mono- and dianions, the π-conjugation of the
PPV-type moiety is less extended than in their corresponding radicals.
The mono- and dianions show two new broad and intense
bands between 525 and 612 nm, one of which always appears
centered at 525 nm in all mono- and dianions, while the least
energetic one is blue-shifted from 612 to 602 nm when going
from 12ꢀ to 52ꢀ. In the largest anionic species 14ꢀ and 52ꢀ, both
with five PPV units, only the least energetic band was observed
since the other band is probably inside the intense band
generated by the PPV moiety.
1
2ꢀꢀ52ꢀ and monoanions 11ꢀꢀ14ꢀ —are collected in Table 4.
In all reduction reactions of radicals, the initial band centered
around 389 nm, characteristic of the PTM radical chromophore,
decreases until it disappears completely when all radical centers are
reduced to anions. Also the least energetic band of the mono- and
diradicals, centered between 680 and 782 nm, completely disappears
when the corresponding anions are formed. The last observation
supports the assignment of this band to charge-transfer absorption
from the PPV electron-donor fragment to the PTM electron-acceptor
units. In all the resulting mono- and dianions, the absorption bands
Optically Induced Intramolecular Electron Transfer in
1ꢀ•ꢀ5ꢀ•. During the reduction of diradicals 1 and 2, together
with the above-described absorptions, a weak and broad band
appears in the NIR region (see Figure 4 inset) whose intensity
5825
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Table 5. IVT Absorption Data of 1ꢀ•, 2ꢀ•, and 20ꢀ•, Acquired
Using Chemical Reduction with Cu in CH2Cl2 with [(Ph)4P]
Br (0.1 M), along with the Effective Electronic Couplings and
Separations of the Two Redox Sites
the bridge and the PTM units (vide infra) and, consequently, a
slight decrease of the effective electronic coupling between the
terminal sites.
Thermally Induced Intramolecular Electron Transfer in
1ꢀ•ꢀ5ꢀ•. We also studied the thermally activated IET process
in 1ꢀ•ꢀ5ꢀ• by means of VT-ESR spectroscopy. For this purpose,
the electrochemical reduction of diradicals 1ꢀ5 was performed
until an almost complete reduction to 12ꢀꢀ52ꢀ was achieved.
This procedure is advantageous since it avoids interferences from
the diradical precursors and the resulting solutions contain,
besides a small concentration of the desired mixed-valence
1ꢀ•ꢀ5ꢀ• species, a large concentration of the ESR-silent
compda ν max/cmꢀ1 εmax/Mꢀ1 cmꢀ1 Δν1/2/cmꢀ1 R/Å Vab/cmꢀ1
h
1ꢀ•
2ꢀ•
20ꢀ•
7090
7553
6349
451
264
677
3150
2110
2940
18.8
25.4
18.8
109
52
123
a Data for 20ꢀ• were taken from ref 13c.
increases until a complete formation of the mixed-valence radical
anion species 1ꢀ•and 2ꢀ• is achieved. From that point the intensity
of this band decreases as the radical anion is progressively trans-
formed into the dianion. This optical transition is expected for
mixed-valence compounds that belong to Robin and Day’s class II28
in which an optically induced IET phenomenon, produced by a
bridge-mediated superexchange (tunneling) mechanism, takes
place. For such optical transitions, when the distance between the
two electroactive centers increases, the intensity of the intervalence
transition (IVT) band decreases and at the same time is shifted
toward shorter wavelengths, as is indeed observed for 1ꢀ• and 2ꢀ•,
which show IVT bands at 1420 and 1380 nm, respectively. For the
mixed-valence species 3ꢀ•, 4ꢀ•, and 5ꢀ•, no IVT bands were
observed, even after a careful deconvolution of their spectra. This
result is not surprising because the calculated distances between the
two methyl carbon atoms are too large to detect any IVT band.29 In
fact, IVT bands are increasingly difficult to detect when the distance
between the two redox centers approaches 20ꢀ25 Å because of the
exponential decay of its intensity.4 As a rule of thumb, the tunneling
is not effective over distances greater than ∼25 Å.30
1
2ꢀꢀ52ꢀ species, thereby providing good spectral resolutions.
The ESR data of all radical anions 1ꢀ•ꢀ5ꢀ• in CH2Cl2 at low
temperatures are gathered in Table 3. All compounds display
four main lines at low temperature in their ESR spectra, which are
very similar to those of the related monoradicals 11ꢀ14 (see
Table 3). These characteristics are consistent with the coupling
of the unpaired electron of these species with the two hydrogen
atoms of the ethylene moiety directly linked to the PTM radical
unit, showing at the same time that these radical anions are at the
“slow-exchange” regime under such conditions. By contrast, at
room temperature and at intermediate temperatures, the spectral
features differ notably from those observed at low temperature,
indicating that a dynamic IET takes place on the ESR time scale.
To perform a complete VT-ESR study with compounds
1ꢀ•ꢀ5ꢀ•, we selected CH2Cl2 and 1,2-dichlorobenzene as the
solvents of choice because both solvents solubilize all radical
anions, as well as the electrolyte in which they are generated, and
both have low dielectric constants, enabling ESR spectra to be
obtained with good resolution. In addition, the melting and
boiling points of both solvents are very different, ensuring a large
temperature range for variable-temperature measurements.
Thus, the VT-ESR study in CH2Cl2 in the temperature range
of 200ꢀ310 K allowed us to follow completely the IET processes
in 1ꢀ•, 2ꢀ•, and 3ꢀ•. As an example, we show in Figure 5 the VT-
ESR spectra of the mixed-valence species 2ꢀ• (see also Figure
S2). The ESR spectrum of this radical anion at 200 K displays two
symmetrical lines arising from the hyperfine coupling with only
one hydrogen atom of the ethylene moiety. This result demon-
strates that, at this temperature, the unpaired electron of radical
anion 2ꢀ• is localized on only one half of the molecule on the
ESR time scale, being therefore at the “slow-exchange” regime.
When the temperature increases, a new central line gradually
emerges between the two initial ones. This evolution is consis-
tent with the increase of the IET rate between the two sites going
from the “slow-exchange” limit at 200 K to the “fast-exchange”
regime at 300 K. In the latter region the unpaired electron of 2ꢀ•
In order to determine the effective electronic coupling, Vab,
between the two equivalent electroactive centers of radical
anions 1ꢀ• and 2ꢀ•, we applied the equation developed by
Hush,31 as was previously done for the radical anion 20ꢀ•.13,26
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vab ¼ ½2:05 ꢁ 10ꢀ2
ε
maxνmaxΔν1=2ꢂR
ð1Þ
Here, Vab is the tunneling matrix element (in cmꢀ1), εmax the
maximum extinction coefficient (in Mꢀ1 cmꢀ1), νmax the transi-
h
tion energy (in cmꢀ1), Δν1/2 the full-width at half-height
h
(in cmꢀ1) of the IVT band, and R the effective separation of the
two redox sites (in Å), which can be estimated from semiempirical
calculations. The resulting data, shown in Table 5, were obtained
after correction for the comproportionation equilibrium and by
deconvolution of the experimental spectra (see Figure 4).
The effective electronic coupling values obtained for the
1
mixed-valence species 1ꢀ• and 2ꢀ• are 109 and 52 cmꢀ1
,
is coupled with two H nuclei of both ethylene groups, both
respectively, showing a large decrease of the tunneling matrix
element as the effective distance between the redox sites in-
creases. This is the normal behavior expected for a tunneling
mechanism, mediated by superexchange through the bridge,
whose intensity decreases exponentially with the distance be-
tween the centers. On the other hand, the Vab value for 1ꢀ• is
somewhat lower than that previously reported for the nonsub-
stituted related radical anion 20ꢀ•.26 This result suggests that
steric and electronic effects of the two octyloxy substituents on
the six-member ring slightly influence on the effective electronic
coupling. Probably steric hindrance between the bulky octyloxyl
substituents and the chlorine atoms on the neighboring phenyl
rings of PTM units is at the origin of a loss of coplanarity among
becoming equivalent nuclei due to the high rate of the ET. The
resulting spectrum displays three main lines with a hyperfine
coupling constant half the value of that at low temperature. The
simulation of ESR spectra32 at different temperatures yielded
the IET rate constants, kET, that vary from 8.1 ꢁ 106 sꢀ1 at 200 K
to 2.1 ꢁ 108 sꢀ1 at 310 K (Figure 5). On going from 1ꢀ• to 3ꢀ•,
the rate constants kET at a given temperature decrease, indicating
that the IET process is slowed as the distance between the
electroactive centers increases. In accordance, species 4ꢀ• and
5ꢀ•, with even larger bridges, appear as totally localized species in
the range of temperatures allowed by CH2Cl2.
The shape evolution trend with temperature of the ESR
spectra of anion radicals 1ꢀ•ꢀ5ꢀ• in 1,2-dichlorobenzene was
5826
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Figure 5. VT-ESR spectra of 2ꢀ• in CH2Cl2 with 0.1 M of [(Ph)4P]Br. Experimental spectra in the temperature range of 200ꢀ310 K (left) and
simulated spectra (right) with different IET rates.
Figure 6. VT-ESR spectra of 4ꢀ• in 1,2-dichlorobenzene with 0.1 M [(Ph)4P]Br. Experimental spectra in the temperature range of 260ꢀ420 K (left)
and simulated spectra (right) with different IET rates.
similar to that obtained in CH2Cl2 (see Figures 6, S3, and S4).
Nevertheless, in this case it was possible to record spectra at
very elevated temperatures and accordingly to follow IET
processes with higher rate constants. As an example, for
compound 1ꢀ•, CH2Cl2 allowed a maximum rate of only
6.3 ꢁ 108 sꢀ1, while with 1,2-dichlorobenzene the maximum
rate was 10.0 ꢁ 108 sꢀ1. 1,2-Dichlorobenzene also made it
possible to follow the thermally activated IET process of the
mixed-valence species 4ꢀ• (Figure 6) and to observe a small
change in the IET rate constants of the mixed-valence species
5
ꢀ• on going from 300 to 420 K (see Figure S4), because this
compound is at the “slow- exchange” regime below 300 K.
Experimental ESR spectra of all five mixed-valence species at
320 K in 1,2-dichlorobenzene (see Figure S5) clearly demon-
strate that the rate constants decrease with the oligo-PPV
bridge lengths.
5827
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
In order to understand the IET mechanism operating in anion
radicals 1ꢀ•ꢀ5ꢀ•, we studied the temperature dependence of the
kET values obtained from the simulation of the VT-ESR spectra
of 1ꢀ•ꢀ3ꢀ• in CH2Cl2 and 1ꢀ•ꢀ4ꢀ• in 1,2-dichlorobenzene.
Unfortunately, the lack of experimental data available for 5ꢀ•,
even in 1,2-dichlorobenzene, prevented its inclusion in this
study. As already mentioned, in mixed-valence systems the IET
may occur either through a bridge-mediated superexchange (or
direct tunneling) process between donor and acceptor electronic
states or through an incoherent hopping process; both mecha-
nisms can operate in parallel, with the overall rate constant
receiving contributions from both mechanisms.12 The tempera-
ture dependences of the two pathways can be used to clarify the
relative importance of each, since the tunneling mechanism in
principle should not be dependent on temperature, while the
hopping process must follow a classical Arrhenius relation. The
Arrhenius plots log(kET) vs 1/T of the kET values obtained from
the VT-ESR spectra for 1ꢀ•ꢀ3ꢀ• in CH2Cl2 and for 1ꢀ•ꢀ4ꢀ• in
1,2-dichlorobenzene are depicted in Figure 7, and the associated
kinetic data, such as the activation energy barriers, Ea, and
exponential prefactors, A, are gathered in Table 6. The Arrhenius
plots in CH2Cl2 can be separated into two distinct temperature
regions: the low-T region at T < 240 K for 1ꢀ•, <270 K for 2ꢀ•,
and <280 K for 3ꢀ•, and the high-T region at higher tempera-
tures. In the latter region there is a stronger temperature
dependence with larger slopes, which suggests a dominant
thermally activated hopping process. In this region, the activation
energy barriers and the exponential prefactors decrease as the
length of the bridges increases (Table 6). In the low-T region the
three radical anions exhibit lower values for both Ea and A
parameters than in the high-T regime. Arrhenius plots with VT-
ESR data of 1ꢀ• and 2ꢀ• in 1,2-dichlorobenzene also show the
existence of two distinct temperature regimes (Figure 7,
bottom), while the two regimes are not distinguished for 3ꢀ•
and 4ꢀ•, indicating the existence of only the high-T one. In the
latter regime the Ea and A values for 1ꢀ•ꢀ4ꢀ• also decrease as the
length of the bridge increases (Table 6).
The presence of two temperature regions in both solvents
could be explained in two alternative ways: (a) the existence of
two parallel conventional thermally activated IET processes with
different kinetic parameters A and Ea, associated with different
conformations of the compounds present in solution in the two
temperature regions, or (b) the presence at low-T of a dominant
Figure 7. Top: Arrhenius plots of the IET rate constants (kET) for 1ꢀ•
,
2
ꢀ•, and 3ꢀ• in CH2Cl2. Bottom: Arrhenius plots of the IET rate
constants (kET) for 1ꢀ•ꢀ4ꢀ• in 1,2-dichlorobenzene. Continuous lines
correspond to Arrhenius plots for high-temperature regime data, while
the discontinuous lines correspond to analogous plots for the low-
temperature regime data.
Table 6. Kinetic Dataa in the Two Temperature Regions Obtained from VT-ESR Experiments for IET Processes of Mixed-Valence
Species 1ꢀ•ꢀ4ꢀ• in CH2Cl2 and 1,2-Dichlorobenzene
T-region
parameter
1•ꢀ
2•ꢀ
3•ꢀ
4•ꢀ
CH2Cl2
high-T
A
(1.1 ( 0.4) ꢁ 1013
5.8 ( 0.6
(6.6 ( 1.3) ꢁ 1011
(2.5 ( 1.2) ꢁ 1010
4.1 ( 1.3
Ea
n; R2
5.0 ( 0.3
4; 0.998
5; 0.998
3; 0.995
low-T
A
(1.4 ( 0.8) ꢁ 1011
3.7 ( 1.1
(4.6 ( 0.8) ꢁ 1010
3.6 ( 0.3
(7.0 ( 4.7) ꢁ 108
2.2 ( 1.3
Ea
n; R2
4; 0.983
4; 0.999
3; 0.976
C6H4Cl2
high-T
low-T
A
(6.4 ( 3.8) ꢁ 1013
7.6 ( 1.1
(2.7 ( 0.6) ꢁ 1012
(5.1 ( 1.6) ꢁ 109
3.5 ( 0.5
(4.7 ( 0.6) ꢁ 108
1.9 ( 0.2
Ea
n; R2
6.3 ( 0.3
4; 0.995
6; 0.998
7; 0.981
8; 0.984
A
5.3 ( 0.9 ꢁ 1010
3.7 ( 0.4
1.0 ( 0.2 ꢁ 1011
4.3 ( 0.3
Ea
n; R2
3; 0.999
5; 0.998
a A in sꢀ1; Ea in kcal molꢀ1. n is the number of temperatures used in each fitting.
3
5828
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
are involved. The assignment of a hopping mechanism for the longest
compounds is additionally supported by the linear trend of kET versus
1/d observed for the largest compounds (see Figure 8 inset). This
kind of transition from tunneling to hopping in long conjugated
molecules has also been observed recently for wires with a different
nature.36
As the nature of solvents may also modulate the IET rate
constants of mixed-valence compounds,37 we also decided to
study such an influence using the radical anion 1ꢀ•. Therefore,
we performed an ESR study of the thermally activated IET
process of 1ꢀ• in different solvents. The solvents were chosen
from the 11 different classifications of the most common organic
solvents, published elsewhere,38 in order to maximize the repre-
sentation of all types of solute/solvent interactions. For this
solvent choice the solubility of the involved species and the
electrolyte was also taken into account, as well as other solvent
properties such as the dielectric constants, which should be low
in order to obtain a good ESR spectra resolution, and the boiling
and freezing temperatures of solvents. From all initially tested
solvents, only seven had the characteristics to obtain good
enough ESR spectra. The resulting VT-ESR spectra of 1ꢀ• in
these solvents at different temperatures, as well as the associated
IET rate constants obtained by simulation of the experimental
spectra, are shown in Figures S6ꢀS10. The kinetic data obtained
from Arrhenius plots of the simulated rate constants in the high-
T range, where the hopping mechanism is dominant for all
solvents, are given in Table S1. There are strong differences in the
activation energy barriers, Ea, depending on the nature of
solvents. In order to find the origin of such differences, a linear
solvation energy relationship (LSER) analysis was performed
with the Ea data. The LSER analysis relates the energy changes
involved in a given molecular process that takes place in solution
with several specific and nonspecific properties of the solvents in
which the process is carried on.39ꢀ41 The LSER model describing
the Ea for 1ꢀ• shows that the IET process is highly sensitive to
changes of the polarity/polarizability of the solvent media (π*)
and to the hydrogen-bond donor (HBD) acidities of the solvent
media (R), while other solvent parameters are unimportant (see
Table S2). Thus, increasing the solvent polarity/polarizability
lowers the hopping process, while increasing the HBD acidity
enhances this process. This result suggests that the main inter-
actions of solvents influencing the IET through hopping in 1ꢀ•
occur through the oligo-PPV bridges, where dipolar and HB
interactions change, in opposite directions, the energy levels of
the bridge modifying the electron hopping rates.
Figure 8. Dependence of the IET rate constants on the bridge lengths
for 1ꢀ•ꢀ5ꢀ• in 1,2-dichlorobenzene at 300 K. Inset: Dependence of kET
vs 1/d for such compounds.
tunneling process activated by the thermal population of specific
conformations. Indeed, internal rotations may create geometries
with smaller barriers to tunneling, and the process itself can
appear to be temperature dependent. So, we should not neglect
bridge dynamics33 when analyzing the temperature dependence
of ET, as the IET could be gated by torsional motions between
the donor and the first bridge phenyl ring, as well as that
occurring between the vinyl group and the phenylene linked to
it.34 Lowering the torsional barrier between the donor and the
stilbenyl bridge allows more efficient electron transport between
the donor and acceptor. Theoretical calculations (vide infra)
confirmed the rich conformational space accessible for such
radical anions at a relatively low energy cost.
The operative mechanism for all studied radical anions was
unraveled by studying the dependence on the bridge lengths of
the kET constants at one fixed temperature at which we have
reliable experimental rate constants. Figure 8 depicts the plot of
ln(kET) versus d for 1ꢀ•ꢀ5ꢀ• in 1,2-dichlorobenzene at 300 K. In
such mixed-valence molecules there appear to be two main ET
regimes under these conditions. Thus, for the two shortest
bridges there is a sharp decrease in the electron-transfer rate
constants with the distance, similarly to that for compounds
exhibiting an exponential dependence on the distance, like kET
=
k0 exp(ꢀβd), where IET occurs through a tunneling mechanism
via superexchange.9,10 In contrast, for the largest bridges the
dependence of kET on dꢀ1 points out that the IET occurs via
hopping.
Theoretical Studies of Radical Anions 1ꢀ•ꢀ4ꢀ•. In order to
study the conformational space accessible for the studied radical
anions and its influence on their electronic structures, we
explored at the CASSCF/3-21G level, using various active
spaces, the shape of the potential energy surface of 1ꢀ•. This
study was made as a function of both the dihedral angles that
define the coplanarity of the three central six-membered rings
(two angles, ang1 and ang2, which define the coplanarity of the
left and right rings of PTM units, respectively) and the most
relevant distances that define the symmetric/asymmetric char-
acteristics of the molecule. These calculations indicated that the
most relevant geometric parameters defining the localization/
delocalization of the electronic density were the ang1 and ang2
dihedral angles. Therefore, we decided to explore in a more
systematic way the shape of the potential energy surface as a
function of these two dihedral angles. These calculations were
done at the ROHF/3-21G level, in order to lower the
For the largest bridges (3ꢀ•ꢀ5ꢀ•), there is a change in
mechanism regime since kET constants decrease monotonically
with the length of the bridge. A fit for the two shortest bridges
yields a β value of 0.140 Åꢀ1, which is very similar to the reported
values for similar oligo-PPV bridges in other compounds,4,9 while
for the three largest bridges we obtained a value for the
attenuation factor of 0.041 Åꢀ1. Therefore, in the shortest
bridges, the contribution of the tunneling mechanism should
be very significant, since the distance dependence is too abrupt to
be consistent with an ET limited by hopping,35 whereas in the
longest compounds, where a very small energy gap between the PTM
radical electron-acceptor unit and the PPV electron-donor bridge is
anticipated (see the Absoprtion Spectra section), the preferred
pathway seems to be hopping where the donor-bridge-localized states
5829
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
Figure 9. Shape of the potential energy surface computed for 1ꢀ• as a
function of the two dihedral angles that define the coplanarity of the
central six-membered rings. Negative values of ang1 correspond to
conrotatory movements, while positive values refer to disrotatory
movements. Angles are given in degrees and energies in kcal/mol
relative to the lowest energy point (placed at 38 and 45° for ang1 and
ang2, respectively).
Figure 10. Shapes of the SOMO of 1ꢀ• from ROHF calculations for the
indicated (ang1,ang2) points. The ang1 angle corresponds to the left-
side dihedral, while ang2 is for the right-side angles.
Only close to the (0,0) region is the electron distribution mostly
located in the three central six-membered rings (which at this
point are coplanar). However, such points are located in a high-
energy region (ROHF calculations using the B3LYP geometries
indicate that this region is located 55 kcal/mol above the
minimum), but reoptimization of these geometries would give
much smaller energy increments. The (0,0) region is the
only zone of the potential energy surface where the electron
distribution is symmetrically located in the central three six-
membered rings.
As complementary information, we also computed the relative
energy and electronic structure of the first excited state at points
(45,45), (0,45), and (45,90). The calculations were done at the
CASSCF(7,7)/3-21G level. The first excited state is located
between 33 and 36 kcal/mol above the ground state for the
same point, depending on the angles. The electronic distribution
is symmetrical and highly delocalized over the central part of the
molecule. Thus, this is a surface that will be accessible photo-
chemically, but not thermally, being responsible for the optically
induced IVT (vide supra).
We also computed the effect of increasing the number of PPV
units that connect the two PTM units. The SOMO shapes
calculated at the ROHF/3ꢀ-•21G level for the (45,45) point for
radical anions 2ꢀ• and 3 exhibit an asymmetry (see Figure
S12), as also occurs for 1ꢀ• (see Figure 10). For radical anion
4ꢀ•, the electronic distribution is found to be symmetrically
delocalized over the central six-membered rings. A more detailed
investigation of all studied radical anions, taking into account the
presence of solvents, is required for a more quantitative evalua-
tion of the trends, as has been observed for other mixed-valence
systems.41 Nevertheless, the overall trend is that the symmetric
delocalized state becomes more stable as the number of PPV
units is increased.
computational cost while preserving the quality of computations.
For each value of ang1 and ang2, the geometry was the optimum
B3LYP/3-21G geometry after changing only the dihedral angles.
No further optimization was done, given the small differences in
the distances found between the B3LYP and CASSCF optimum
geometries.
The ROHF potential energy surface for the dihedral rotation is
shown in Figure 9. The surface is almost symmetrical with
respect to the 0 of the ang1 dihedral angles (the small asymmetry
is due to the lack of symmetry of the B3LYP/3-21G optimum
geometry, which has a C1 symmetry). Although not plotted, it is
also symmetric with respect to the 0 of the ang2 dihedral angle.
The lowest energy point inthispotentialsurface isplaced atang1=
38° and ang2 = 45°. The three maxima at the points (ang1,ang2)
= ((90, 0) and (0,0) are consequences of the steric repulsions
induced by the Cl atoms when the central rings are forced to
be coplanar. Other relevant points are the ((90,45) and the
((45,90) points, which correspond to the lowest energy path-
ways for the rotation of these dihedral angles in that potential
energy surface. We also analyzed the changes in the electronic
structure in the most relevant points of the potential energy
surface of Figure 9. At the ROHF level, this involves simply
analyzing the shape of the SOMO orbital, while at the CASSCF
level one has to look into all active orbitals. In both cases we
found that the most relevant change is a variation of the degree of
localization/delocalization of the electronic density, which
changes depending on the region of the surface evaluated. The
changes detected at the ROHF level and at the CASSCF level are
similar. Figure 10 plots the shape of the SOMO (in the CASSCF
calculations this is selected as the orbital having an occupation
closest to 1) for the characteristic points indicated before.
Combining these orbital changes with Figure 10, which gives
us an idea about the regions of the conformational space that
can be energetically visited depending on the temperature, one
can know what kind of electron localization will be observed
experimentally at room temperature. At room temperature,
radical 1ꢀ• can thermally visit any point of the region around
((37,45), where the electron distribution is mostly located in
one of the two PTM halves and the central six-membered ring.
’ CONCLUSIONS
Conjugated diradicals 1ꢀ5 with different lengths, incorporat-
ing from one to five PPV units between two polychlorinated
triphenylmethyl (PTM) radical units, have been prepared using
5830
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
WittigꢀHorner-type chemistry. Chemical and electrochemical
reduction of the diradicals 1ꢀ5 gave the series of radical anions
1ꢀ•ꢀ5ꢀ• and dianions 12ꢀꢀ52ꢀ, which were fully characterized.
To the best of our knowledge, radical anions 1ꢀ•ꢀ5ꢀ• enabled to
study for the first time, by means of UVꢀvisꢀNIR and ESR
spectroscopies, the long-range IET phenomena in the ground
state in a series of mixed-valence species with increasing lengths
and without using any external perturbation. The temperature
dependence of the IET rate constants of mixed-valence species
1ꢀ•ꢀ5ꢀ• revealed the presence of two different regimes at low
and high temperatures. In each of these regimes a different IET
mechanism dominates. Thus, at higher temperatures the domi-
nant mechanism is the thermally activated hopping one, while at
low temperatures electron tunneling via superexchange is the
operative mechanism. Both mechanisms occur to different
extents depending on the length of the mixed-valence com-
pounds. Thus, the lengths of the oligo-PPV bridges have a
notable influence on the activation energy barriers of the hopping
processes, diminishing the barriers as the lengths are increased.
The nature of solvent also modifies the IET rates when they are
controlled by the hopping mechanism by means of the interac-
tions between the PPV bridges and the solvents. In the shortest
compounds 1ꢀ• and 2ꢀ•, optically induced IET through a
superexchange mechanism was also observed, both mixed-va-
lence species exhibiting intervalence bands whose intensities
decreased with the length of the PPV fragment as the electronic
coupling between the PTM units does.
Ag/AgNO3 reference electrode. The experiments were carried out in dry
CH2Cl2 containing 0.1 M [(n-Bu)4N]PF6 as supporting electrolyte.
Deoxygenation of the solutions was achieved by bubbling argon for at
least 5 min, and the working electrode was cleaned after each run. The
cyclic voltammograms were recorded with a scan rate increasing from
0.05 to 1.00 V sꢀ1. Ferrocene was used as an internal reference for
calibrating the redox potentials and checking the reversibility of the
electrochemical setup. The theoretical redox potential for FeCp2 vs
Agþ/Ag in CH2Cl2 was assumed to be 0.46 V.42
Synthesis of Compounds. Compounds 6, 8, 10, 16, and 17 were
prepared according to previously reported procedures,17 while the
preparation of compounds 1ꢀ5 and 18 is given in the Supporting
Information.
Theoretical Calculations. B3LYP and CASSCF calculations
were done using the appropriate options in the Guassian03 program.43
The B3LYP functional is the three-parameter functional of Becke, a
hybrid functional that mixes HartreeꢀFock exchange with the
LeeꢀYangꢀParr exchange-correlation functional (a class of gradient-
corrected exchange and correlation functionals). The CASSCF(n,m)
calculations (complete active space multiconfiguration self-consistent
field computation on (n,m) active space) were obatined by combining
the n electrons and the m orbitals of the active space in all possible
forms.44 The smallest active space selected in this study was the (3,3)
active space, obtained by selecting the SOMO, SOMOꢀ1, and LUMO
orbitals in a UB3LYP computation. The remaining active spaces were
obtained by increasing the number of electrons and orbitals in an
appropriate way, the largest one being a (7,7) active space. With
this active space we confirmed the validity of the computations
done with smaller active spaces. Geometry optimizations were done
using the usual procedures within Gaussian; in the B3LYP and CASSCF
cases all analytical gradients were used. The minimum condition in
full minimizations was confirmed by looking at the vibrational
frequencies.
’ EXPERIMENTAL SECTION
General Procedures. All reactions were carried out under Ar and
using solvents which were dried carefully by routine procedures. For
column chromatography, silica gel 60 (230ꢀ400 mesh, 0.040ꢀ0.063
mm) was purchased from E. Merck. Thin-layer chromatography (TLC)
was performed on glass sheets coated with silica gel 60 F254 purchased
from E. Merck, visualized by UV light. NMR spectra were recorded on a
Bruker AC 200 (200 MHz) or a Bruker AM 400 (400 MHz) instrument
with solvent peaks as reference. Mass measurements were carried out on
a Bruker BIFLEX matrix-assisted laser desorption/ionization time-of-
flight (MALDI-TOF) mass spectrometer. A saturated solution of 1,8,9-
trihydroxyanthracene (dithranol, Aldrich, EC 214-538-0) in CH2Cl2
was used as a matrix. For diradicals 1ꢀ5 and monoaldehydes 6ꢀ10, no
matrix was used, or in some cases nitroanthracene. FAB mass spectra
were obtained on a VG AutoSpec instrument, using m-nitrobenzyl
alcohol as a matrix. FT-IR spectra were recorded on a Nicolet Impact
410 spectrophotometer using KBr disks. Elemental analyses were
performed by the analytical service at the Institut Charles Sadron,
Strasbourg, and in the Servei d’Anꢁalisi Química de la UAB.
UVꢀvisꢀnear-IR spectra were taken on a Varian Cary 5 E spectro-
photometer. ESR spectra were obtained on an X-band spectrometer
(Bruker ESP 300 E) equipped with a field-frequency (F/F) lock
accessory and built-in NMR gaussmeter. A rectangular TE102 cavity
was used for the measurements. The signal-to-noise ratio of spectra was
increased by accumulation of scans using the F/F lock accessory to
guarantee large field reproducibility. Precautions to avoid undesirable
spectral distortions and line broadenings, such as those arising from
microwave power saturation and magnetic field overmodulation, were
also taken. To avoid dipolar line broadening from dissolved oxygen,
solutions were always carefully degassed with pure argon. The CV
measurements were performed on an EG&G PAR 263A potentiostat/
galvanostat controlled by a personal computer and driven by dedicated
software. CV was performed with a conventional three-electrode con-
figuration consisting of platinum working and auxiliary electrodes and a
’ ASSOCIATED CONTENT
S
Supporting Information. Complete refs 1 and 43; synthe
b
sis of compounds 6, 8, 10, 16, and 17; comparative UVꢀvis
spectra of 2, 7, and 11; VT-ESR spectra of radical anions 1ꢀ• and
3
ꢀ• in CH2Cl2; VT-ESR spectra of radical anions 1ꢀ• and 3ꢀ• in
1,2-dichlorobenzene; VT-ESR spectra of radical anions 1ꢀ• ꢀ3ꢀ•
and 5ꢀ• in 1,2-dichlorobenzene; ESR spectra at 300 K of radical
anions 1ꢀ•ꢀ5ꢀ• in 1,2-dichlorobenzene; VT-ESR spectra of
radical anion 1ꢀ• in anisole, ethyl acetate, toluene, chloroform,
and benzonitrile; theoretical calculations of the conformational
space and the electronic structure of radical anions 1ꢀ•ꢀ4ꢀ•;
and a detailed study of the influence of solvents on IET of
radical anion 1ꢀ• by the linear solvation energy relationship.
This material is available free of charge via the Internet at
http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
vecianaj@icmab.es; cun@icmab.es
’ ACKNOWLEDGMENT
The authors thank EU Large Project ONE-P (FP7-NMP-
2007-212311), Instituto de Salud Carlos III through “Acciones
CIBER”, the MICINN, Spain (CTQ2006-06333/BQU and
CTQ2010-19501/BQU), and Generalitat de Catalunya (2009S-
GR158). The team from Universitat de Barcelona thanks the
5831
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
MICINN Spanish (MAT2008-02032/MAT-FEDER Fonds co-
financed, and UNBA05-33-001) for support and CESCA and
BSC for computer time.
Vidal-Gancedo, J.; Veciana, J.; Rovira, C.; Hudhomme, P. Angew. Chem.,
Int. Ed. 2003, 42, 2765.
(17) Langa, F.; Gomez-Escalonilla, M. J.; Rueff, J.-M.; Figueira-
Duarte, T. M.; Nierengarten, J.-F.; Palermo, V.; Samor, P.; Rio, Y.;
Accorsi, G.; Armaroli, N. Chem.—Eur. J. 2005, 11, 4405.
(18) M€ullen, K.; Wegner, G. Electronic Materials: the Oligomer
Approach, Wiley-VCH: Weinheim, 1998.
(19) House, H. O. Modern Synthetic Reactions, 2nd ed.; Benjamin:
Menlo Park, CA, 1972; p 608.
(20) (a) Wurst, K.; Elsner, O.; Schottenberger, H. Synlett 1995, 833.
(b) Rodríguez, J.-G.; O~nate, A.; Martin-Villamil, R. M.; Fonseca, I.
J. Organomet. Chem. 1996, 513, 71.
(21) Fitjer, L.; Quabeck, U. Synth. Commun. 1985, 15, 855.
(22) Peeters, E.; van Hal, P. A.; Knol, J.; Brabec, C. J.; Sariciftci, N. S.;
Hummelen, J. C.; Janssen, R. A. J. J. Phys. Chem. B 2000, 104, 10174.
(23) Armaroli, N.; Accorsi, G.; Gisselbrecht, J. P.; Gross, M.;
Krasnikov, V.; Tsamouras, D.; Hadziioannou, G.; Gomez-Escalonilla,
M. J.; Langa, F.; Eckert, J. F.; Nierengarten, J.-F. J. Mater. Chem. 2002,
12, 2077.
’ REFERENCES
(1) Adams, D. M.; et al. J. Phys. Chem. B 2003, 107, 6668.
(2) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(3) Isied, S. S.; Ogawa, M. Y.; Wishart, J. F. Chem. Rev. 1992, 92, 381.
(4) (a) Electron Transfer in Chemistry; Balzani, V., Ed.; Wiley:
Weinheim, 2001; Vols. IꢀV. (b) Fox, M. A., Ed. Chem. Rev. 1992, 92,
365ꢀ490. (c) Paddon-Row, M. N. Acc. Chem. Res. 1994, 27, 18. (d)
Schab, P. F. H.; Levin, M. D.; Michl, J. Chem. Rev. 1999, 99, 1863. (e)
Segal, D.; Nitzan, A.; Davis, W. B.; Wasielewski, M. R.; Ratner, M. A. J.
Phys. Chem. B 2000, 104, 3817. (f) Giese, B. Acc. Chem. Res. 2000,
33, 631. (g) Launay, J. P. Chem. Soc. Rev. 2001, 30, 386. (h) Giese, B.;
Amaudrut, J.; Kohler, A. K.; Spormann, M.; Wessely, S. Nature 2001,
412, 318. (i) Wold, D. J.; Haag, R.; Rampi, M. A.; Frisbie, C. D. J. Phys.
Chem. B 2002, 106, 2813. (j) Nakamura, T.; Fujitsuka, M.; Araki, Y.; Ito,
O.; Ikemoto, J.; Takimiya, K.; Aso, Y.; Otsubo, T. J. Phys. Chem. B 2004,
108, 10700. (k) Engelkes, V. B.; Beebe, J. M.; Frisbie, C. D. J. Am. Chem.
Soc. 2004, 126, 14287. (l) Conwell, E. M. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 8795. (m) Choi, S. H.; Kim, B.; Frisbie, C. D. Science 2008,
320, 1482. (n) Lu, Q.; Liu, K.; Zhang, H.; Du, Z.; Wang, X.; Wang, F.
ACS Nano 2009, 3, 3861.
(24) Ballester, M.; Riera, J.; Casta~ner, J.; Rodriguez, A.; Rovira, C.;
Veciana, J. J. Org. Chem. 1982, 47, 4498.
(25) Electronic Paramagnetic Resonance; Weil, J. A., Bolton, J. R., Eds.;
Wiley and Sons: New York, 1994.
(26) Lloveras, V.; Vidal-Gancedo, J.; Ruiz-Molina, D.; Figueira-
Duarte, T. M.; Nierengarten, J.-F.; Veciana, J.; Rovira, C. Faraday
Discuss. 2006, 131, 291.
(5) Ratner, M. A.; Jortner, J. In Molecular Electronics; Jortner, J.,
Ratner, M., Eds.; Blackwell: Oxford, 1997; p 5.
(6) Davis, W. B.; Wasielewski, M. R.; Ratner, M. A.; Mujica, V.;
Nitzan, A. J. Phys. Chem. A 1997, 101, 6158.
(7) Okada, A.; Chernyak, V.; Mukamel, S. J. Phys. Chem. A 1998,
102, 1241.
(27) Dumur, F.; Gautier, N.; Gallego-Planas, N.; Sahin, Y.; Levillain,
E.; Mercier, N.; Hudhomme, P.; Masino, M.; Girlando, A.; Lloveras, V.;
Vidal-Gancedo, J.; Veciana, J.; Rovira, C. J. Org. Chem. 2004, 69, 2164.
(28) Robin, M.; Day, P. Adv. Inorg. Radiochem. 1967, 10, 247.
(29) It is well known that the unpaired electron of highly chlorinated
triphenylmethyl radicals and the negative charge of their corresponding
carbanions are mainly localized on the R carbons.14 Therefore, the two R
carbon atoms of the radical anions 1ꢀ•ꢀ5ꢀ• can be considered in a first
approximation as the electron active sites for the ET phenomena. It must also
be noticed that, as far as IET phenomena are concerned, the main distance to
be considered is the through-bond distance between the two electron active
sites; i.e., the sum of bond lengths corresponding to the conjugated pathway.
However, as shown by ESR, the average distance between the two spins—
the average interspin separation—differs from the nominal separation, either
through-space or through-bond, between the two radical R carbon atoms
obtained from the AM1-minimized structures, due to the different extents of
delocalization in both diradicals.
(30) McCreery, R. L. Chem. Mater. 2004, 16, 4477.
(31) Hush, N. S. Coord. Chem. Rev. 1985, 64, 135.
(32) (a) Heinzer, J. J. Mol. Phys. 1971, 22, 167. (b) Heinzer, J.
Quantum Chemistry Program Exchange 1972, No. 209.
(33) (a) Petterson, K.; Kyrychenko, A.; R€onnow, E.; Ljungdahl, T.;
Martensson, J.; Albinsson, B. J. Phys. Chem. A 2006, 110, 310. (b) Eng,
M. P.; Ldungdahl, T.; Martensson, J.; Albinsson, B. J. Phys. Chem. B
2006, 110, 6483. (c) Scott, A. M.; Ricks, A. B.; Calvin, M. T.; Wasieleski,
M. R. Angew. Chem., Int. Ed. 2010, 49, 2904.
(8) Pollard, W. T.; Felts, A. K.; Friesner, R. A. Adv. Chem. Phys. 1996,
93, 77.
(9) (a) Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1992,
114, 6227. (b) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski,
M. R. Nature 1998, 396, 60. (c) Davis, W. B.; Ratner, M. A.; Wasielewski
J. Am. Chem. Soc. 2001, 123, 7877. (d) Giacalone, F.; Segura, J. L.;
Martín, N.; Guldi, D. M. J. Am. Chem. Soc. 2004, 126, 5340.
(e) Giacalone, F.; Segura, J. L.; Martín, N.; Raney, J.; Guldi, D. M.
Chem.—Eur. J. 2005, 11, 4819. (f) Lancaster, K.; Odom, S. A.; Jones,
S. C.; Thayumanavan, S.; Marder, S. R.; Brꢀedas, J.-L.; Coropceanu, V.;
Barlow, S. J. Am. Chem. Soc. 2009, 131, 1717.
(10) (a) Weiss, E. A.; Ahrens, M. J.; Sinks, L. E.; Gusev, A. V.; Ratner,
M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 5577. (b) Malak,
R. A.; Gao, Z.; Wishart, J. F.; Isied, S. S. J. Am. Chem. Soc. 2004,
126, 13888.
(11) Goldsmith, R. H.; Sinks, L. E.; Kelley, R. F.; Betzen, L. J.; Liu,
W.; Weiss, E. A.; Ratner, M. A.; Wasielewski, M. R. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 3540.
(12) Paddon-Row, M. N. In Electron Transfer in Chemistry; Balzani,
V., Ed.; Wiley-VCH: Weinheim, 2001; Vol. III, Part 2, p 179.
(13) (a) Bonvoisin, J.; Launay, J.-P.; Rovira, C.; Veciana, J. Angew.
Chem., Int. Ed. Engl. 1994, 33, 2106. (b) Sedo, J.; Ruiz, D.; Vidal-
Gancedo, J.; Rovira, C.; Bonvoisin, J.; Launay, J.-P.; Veciana, J. Adv.
Mater. 1996, 8, 748. (c) Rovira, C.; Ruiz-Molina, D.; Elsner, O.; Vidal-
Gancedo, J.; Bonvoisin, J.; Launay, J.-P.; Veciana, J. Chem.—Eur. J. 2001,
7, 240. (d) Elsner, O.; Ruiz-Molina, D.; Vidal-Gancedo, J.; Rovira, C.;
Veciana, J. Nano Lett. 2001, 1, 117.
(14) (a) Ballester, M. Acc. Chem. Res. 1985, 297, 131. (b) Veciana, J.;
Riera, J.; Casta~ner, J.; Ferrer, N. J. Organomet. Chem. 1985, 297, 131. (c)
Armet, O.; Veciana, J.; Rovira, C.; Riera, J.; Casta~ner, J.; Molins, E.; Rius,
J.; Miravitlles, C.; Olivella, S.; Brichfeus, J. J. Phys. Chem. 1987, 91, 5608.
(15) (a) Smart, J. C.; Pinsky, B. L. J. Am. Chem. Soc. 1980, 103, 1009.
(b) Kramer, J. A.; Hendrickson, D. N. Inorg. Chem. 1980, 19, 3330.
(16) (a) Rak, S. F.; Miller, L. L. J. Am. Chem. Soc. 1992, 114, 1388.
(b) Nelsen, S. F.; Ismagilov, R. F.; Trieber, D. A., II Science 1997,
278, 846. (c) Lindeman, S. V.; Rosokha, S. V.; Sun, D.; Kochi, J. K. J. Am.
Chem. Soc. 2002, 124, 843. (d) Gautier, N.; Dumur, F.; Lloveras, V.;
(34) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem.
Soc. 2001, 123, 7877.
(35) Sikes, H. D.; Smalley, J. F.; Dudek, S. P.; Cook, A. R.; Newton,
M. D.; Chidsey, C. E. D.; Feldberg, S. W. Science 2001, 291, 1519.
(36) (a) Choi, S. H.; Risko, C.; Ruiz-Delgado, M. C.; Kim, B.;
Brꢀedas, J.-L.; Frisbe, C. D. J. Am. Chem. Soc. 2010, 132, 4358. (b) Hines,
T.; Diez-Perez, I.; Hihath, J.; Liu, H.; Wang, Z.-S.; Zhao, J.; Zhou, G.;
M€ullen, K.; Tao, N. J. Am. Chem. Soc. 2010, 132, 11658.
(37) Nielsen, S. F.; Trieber, D. A.; Ismagilov, R. F.; Teki, Y. J. Am.
Chem. Soc. 2001, 123, 5684.
(38) Ventosa, N.; Ruiz-Molina, D.; Sedo, J.; Rovira, C.; Tomas, X.;
Andre, J.-J.; Bieber, A.; Veciana, J. Chem.—Eur. J. 1999, 5, 3533.
(39) Taft, R. W.; Abboud, J. L. M.; Kamlet, M. J.; Abraham, M. H.
J. Solution Chem. 1985, 14, 153.
(40) Kamlet, M. J.; Taft, R. W. J. Chem. Soc., Perkin Trans. 1979,
2, 337.
5832
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833

Journal of the American Chemical Society
ARTICLE
(41) Kamlet, M. J.; Doherty, R. M.; Carr, P. W.; Mackay, D.;
Abraham, M. H.; Taft, R. W. Environ. Sci. Technol. 1988, 22, 503.
(42) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.
(43) Frisch, M. J.; et al. Gaussian03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(44) Hegarty, D.; Robb, M. A. Mol. Phys. 1979, 38, 1795.
5833
dx.doi.org/10.1021/ja1083859 |J. Am. Chem. Soc. 2011, 133, 5818–5833