Tuning delocalization in the radical cations of 1,4-bis[4-(diarylamino) styryl]benzenes, 2,5-bis[4-(diarylamino)styryl]thiophenes, and 2,5-bis[4-(diarylamino)styryl]pyrroles through substituent effects

Article abstract of DOI:10.1021/ja3023048

Radical cations have been generated for 10 bis[4-(diarylamino)styryl]arenes and heteroarenes to investigate the effect of the electron-richness of the terminal groups and of the bridging (hetero)arene on delocalization. The intervalence charge-transfer ba

Full text of DOI:10.1021/ja3023048

Article
pubs.acs.org/JACS
Tuning Delocalization in the Radical Cations of 1,4-Bis[4-
(diarylamino)styryl]benzenes, 2,5-Bis[4-
(diarylamino)styryl]thiophenes, and 2,5-Bis[4-
(diarylamino)styryl]pyrroles through Substituent Effects
Stephen Barlow,* Chad Risko, Susan A. Odom, Shijun Zheng, Veaceslav Coropceanu, Luca Beverina,
Jean-Luc Bredas, and Seth R. Marder
́
Center for Organic Photonics and Electronics and School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta,
Georgia 30332-0400, United States
S
* Supporting Information
ABSTRACT: Radical cations have been generated for 10 bis[4-
(diarylamino)styryl]arenes and heteroarenes to investigate the
effect of the electron-richness of the terminal groups and of the
bridging (hetero)arene on delocalization. The intervalence charge-
transfer bands of these radical cations vary from weak broad
Gaussians, indicative of localized class-II mixed-valence species, to
strong relatively narrow asymmetric bands, characteristic of
delocalized class-III bis(diarylamino) species, to narrow symmetric
bands in cases where the bridge contribution to the singly occupied
molecular orbital is largest. Hush analysis of these bands yields estimates of the electronic coupling varying from 480 cm−1
(electron-poor bridge, most electron-rich terminal aryl groups) to 1000 cm−1 (electron-rich bridge, least electron-rich termini) if
the diabatic electron-transfer distance, Rab, is equated to the N−N separation. Computational and electron spin resonance (ESR)
evidence for displacement of the diabatic states into the bridge (reduced Rab) suggests that these values are underestimates and
that even more variation is to be expected through the series. Several dications have also been studied. The vis−NIR absorption
of the dication of (E,E)-1,4-bis{4-[bis(4-n-butoxyphenyl)amino]styryl}-2,5-dicyanobenzene is seen at an energy similar to that of
the strongest band in the spectrum of the corresponding weakly coupled monocation, with approximately twice the absorptivity,
and its ESR spectrum suggests essentially noninteracting radical centers. In contrast, the electronic spectra of class-III
monocations show no clear relationship to those of the corresponding dications, which ESR reveals to be singlet species.
catalyzed chemistry developed by Buchwald and Hartwig,13,14
INTRODUCTION
Organic mixed-valence (MV) species consist of two or more
linked organic redox centers with different formal oxidation
■
and since, if appropriately substituted, they form moderately
stable radical cations under mild oxidation conditions.15,16
states.1,2 The electronic structure of a two-site MV species can
In many organic and inorganic MV systems, the redox
centersfor example, metal atoms or triarylaminesare often
be described in terms of the interaction of two different diabatic
sufficiently remote from one another to preclude direct overlap
states in which the oxidation or reduction is localized on either
of the two redox centers. The electronic coupling, V, between
these two states, which can be evaluated from characteristics of
between orbitals localized on these centers. Electronic coupling
can be considered as the result of interaction of bridge-based
orbitals with the “redox group” orbitals.17 In π-conjugated
the intervalence charge-transfer (IVCT) absorption character-
istics following Hush,3 is often stronger in organic MV species
organic systems it is the highest filled and lowest empty bridge
than for their inorganic counterparts.4 Moreover, some organic
orbitals that are generally most important for oxidized and
reduced species, respectively. The strength of the resultant
MV species are relevant to materials chemistry: for example, the
electronic coupling is, therefore, expected to increase with
charge carriers in bis(triarylamine)-based hole-transport
materials, such as the widely used 4,4′-bis(phenyl-m-
increased interaction between local redox-center and bridge
tolylamino)biphenyl, TPD,5−8 are the corresponding MV
orbitals, this interaction increasing as the bridge-based orbitals
radical cations,9,10 and the absorption characteristics of MV
approach the redox-center orbitals in energy, as shown
schematically in Figure 1. A number of studies support this
bis(triarylamine) radical cations generated by photoinduced
electron transfer may be useful in optical limiting.11,12 In
expectation, including several bis(triarylamine) MV systems.
For example, the couplings found for the radical cations of
addition to their relevance to materials chemistry, triarylamine
redox centers are also attractive candidates for the study of MV
phenomena since a wide variety of systems can be synthesized
in a reasonably straightforward manner, especially using Pd-
Received: March 14, 2012
Published: June 11, 2012
© 2012 American Chemical Society
10146
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
Here we investigate the extent to which the electronic
coupling in the radical cations of 1,4-bis[4-(diarylamino)-
styryl]benzenes and of related 2,5-bis[4-(diarylamino)styryl]-
heteroarenes 1−10 (Chart 1) varies as the electron-richness of
the terminal aryl groups and of the bridging 1,4-phenylene, 2,5-
thienylene, or pyrrole-2,5-diyl groups is varied.
Chart 1. Stuctures of Bis[(diarylamino)styryl]benzenes,
Bis[(diarylamino)styryl]thiophenes, and
Bis[(diarylamino)styryl]pyrroles Discussed in This Work
Figure 1. Schematic showing the anticipated effects of raising the
energy of the highest occupied bridge-based orbital relative to those of
terminal donors on (left) the highest filled orbitals of a neutral species
and (right) the instantaneous spin distribution in the corresponding
radical cation.
bis(diphenylamino)-terminated biphenyls or benzenes exceed
those of their bis(di-p-anisylamino)-terminated ana-
logues,10,18−20 demonstrating the role of the electron-richness
of the end group. The role of the bridge21 is illustrated by work
on the radical cations of bis{[4-(di-p-anisylamino)phenyl]-
ethynyl}arenes with varying arene cores22,23 and by the strong
couplings observed in species incoporating thiophene groups in
the bridge.24 Inevitably, increasing coupling via bridging orbitals
will lead to an increasing contribution of bridge orbitals to the
singly occupied molecular orbital (SOMO); in other words, the
oxidation (or reduction) acquires increased bridge character, or
the redox center becomes displaced from its formal position
into the bridge with a concomitant reduction in the diabatic
electron-transfer distance, Rab. Indeed, for bis(diarylamino)
species, values of Rab considerably reduced from the nitrogen−
nitrogen separation are suggested by computational estimates25
or are required for consistency between alternative exper-
imentally based estimates of the electronic coupling, V.24−26
Furthermore, density functional theory (DFT) calculations,
supported by electron spin resonance (ESR) measurements,
suggest that replacing the stilbene bridge of a bis(diarylamino)
radical cation with a more electron-rich 1,2-dithienylethene
bridge leads to an increase in the spin density on the bridge at
the expense of that on the nitrogen atoms and the terminal aryl
groups.24 In the limiting case, the bridge contribution to the
SOMO becomes sufficiently large that the radical ion is better
described as the result of a bridge-based redox process (as
shown schematically in Figure 1) than as being a true MV ion,
as previously reported for several inorganic examples27,28 and
recognized as a possibility in bis(triarylamine) MV systems.23
RESULTS AND DISCUSSION
■
Compounds Investigated. The syntheses of 1−
925,26,29−34 have previously been described. The terminal
redox centers vary, in order of increasing electron-richness,
from unsubstituted and alkyl-substituted 4-(diphenylamino)-
phenyl groups to the more electron-rich 4-[bis(4-
alkoxyphenyl)amino]phenyl groups,35 while the arenes/hetero-
arenes in the center of the bridge vary considerably in electron-
richness from dicyanobenzene to an N-alkylpyrrole.36 The new
compound 10 incorporates the even more electron-rich N-
alkyl-3,4-dialkoxypyrrole bridging group and was synthesized
from the Horner reaction between the appropriate diethyl
[(diarylamino)benzyl]phosphonate and pyrrole-2,5-dialdehyde
(see the Supporting Information for details). Figure 2 shows
the extent to which the electron-richness of end and bridging
groups is varied among compounds 1−10, as approximated by
the highest occupied molecular orbital (HOMO) energies
calculated at the B3LYP/6-31G(d) level for the triarylamine
10147
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
surprising since electronic coupling generally makes only a
small contribution to ΔE1/2,
38 while electrostatic contributions,
which can be important in systems where the charges are
localized over a relatively small number of atoms and/or where
the two redox centers are in close proximity, are likely to be
small in the present systems.39
+/0
The E1/2 values reflect the overall electron-richness of the
molecule and are consistent with expectations based on the
electron-withdrawing or -donating character of the bridging and
terminal substituents and on the relative ionization potentials of
benzene, thiophene, and pyrrole.36 The trends in E1/2 also
+/0
agree well with those seen in the calculated HOMO energies
(Figure 2). Moreover, comaparison of the electrochemical
potentials of 6−10 to those of model compounds in which the
terminal diarylamino groups are replaced with tert-butyl groups
indicates that the difference in potential between diamines and
model compounds decreases as the core of the bridge is varied
from dialkoxybenzene to dialkoxythiophene to pyrrole to
dialkoxypyrrole (see Table S1 in the Supporting Information),
suggesting increasing bridge character to the oxidation.
Figure 2. HOMO energies calculated at the B3LYP/6-31G(d) level of
theory for compounds 1−10 and for the constituent end group (E1−
E3) and bridging group (B1−B7) fragments (some alkyl groups have
been replaced by methyl groups in the calculations).
Robin and Day Classification. The lowest energy features
of the electronic spectra of the radical cations (selected
examples are shown in Figure 3; the remainder are shown in
(E1−E4) and divinylbenzene or divinylheterocycle (B1−B7)
fragments. Solutions containing the corresponding radical
cations and, in some cases (where solubility and stability
allowed, specifically for 1, 4, 5, 6, 7, and 8), the dications were
obtained in dichloromethane using tris(4-bromophenyl)-
+/0
+/0
aminium hexachloroantimonate (E1/2 = 0.70 V vs FeCp2
in CH2Cl237) as the oxidant. UV−vis−NIR and ESR data for
1+, 4+, 42+, and 72+ have previously been published as parts of
different studies.25,26,33
Electrochemistry. The redox potentials of the neutral
species in dichloromethane are summarized in Table 1
Table 1. Electrochemical Half-Wave Potentials (V vs
FeCp2+/0, CH2Cl2/0.1 M [nBu4N]+[PF6]−) for 1,4-Bis[4-
(diarylamino)styryl]benzenes and 2,5-Bis[4-
a
(diarylamino)styryl]heteroarenes
compound
+/0
2+/+
3+/2+
Figure 3. IVCT bands for selected bis[(diarylamino)styryl]benzene
and bis[(diarylamino)styryl]heterocycle radical cations in CH2Cl2.
b
b
b
b
b
b
b
b
1
2
+0.26
+0.51
+0.44
+0.20
+0.31
+0.24
+0.05
+0.15
+0.94
−
3
+1.06
+0.81
+0.97
+0.91
Figure S2 in the Supporting Information) are assigned to IVCT
transitions. As shown in Table 2, those of 1+−4+ are
approximately symmetric Gaussians with widths at half-height,
ν1/2, exceeding the Hush limit of the following equation:
̅
4
5
6
c
7
+0.70
8
+0.97
+0.58
+0.73
ν1/2[Hush] = 4 (ln 2)λkBT
̅
(1)
9
−0.18
−0.26
+0.01
10
−0.02
where λ is the reorganization energy, which, for an MV
compound in which charge is primarily localized on one of two
equivalent redox centers, is equal to the IVCT absorption
a
b
All processes were reversible. Separation between first and second
oxidations not resolved. 4+/3+ couple observed at +1.05 V.
c
maximum, νmax, thus suggesting that these species are Robin-
̅
(electrochemical data for 4 and 6−8 in dichloromethane have
previously been reported,25,34 while potentials in other solvents
have also been reported for some of the other species29).
Generally, the first and second electrons are removed at similar
potentials, and the separation, ΔE1/2 = E1/22+/+ − E1/2+/0, is not
resolvable using cyclic voltammetry, the only exceptions being
the pyrrole-based species 9 and 10, for which ΔE1/2 = 0.19 and
0.24 V, respectively; low to unresolvable values of ΔE1/2 have
been reported for other bis(triarylamine)s18 and are not
and-Day40 class-II MV compounds (as previously noted for 1+
and 4+ in refs 26 and 25, respectively). The IVCT band of 5+
has neither the symmetric Gaussian line shape typical of class-II
systems nor the characteristic shape of class-III bis-
(triarylamine) MV IVCT bands (vide infra), presumably
reflecting a species poised on the class-II/class-III borderline.
The IVCT line shapes for 6+−8+ are typical of those found for
class-III bis(triarylamine)s,10,18,19,25 i.e., narrower than the
Hush limit expected for a class-II system (eq 1) and strongly
10148
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
Table 2. Experimental Parameters Characterizing the Intervalence Absorptions of the Radical Cations of
Bis[(diarylamino)styryl]benzenes and Bis[(diarylamino)styryl]heterocycles in CH2Cl2 with TD-DFT Gas-Phase Values in
a
Italics Along with Estimates of the Electronic Coupling
b
c
εmax/(103 M−1
cm−1
ν1/2[obsd]/
ν1/2[Hush]/
ν1/2[high]/
ν1/2[high]/
VNN
/
Vclass III /
̅
̅
̅
̅
νmax/cm−1
)
cm−1
cm−1
ν1/2[low]
ν1/2[Hush]
μge/D
cm−1
cm−1
̅
̅
̅
b
d
d
d
1+
2+
3+
4+
5+
6+
7+
8+
9+
10+
7450
[4684]
[5649]
[5962]
[5300]
[6173]
6644
6.1
14.2
17.9
15.5
21.3
46.3
45.8
53.0
27.6
41.1
4490
4150
3550
3450
3760
3350
3530
3620
3840
4200
4410
1
1.08
5.78
7.91
[21.0]
480
480
670
700
640
810
960
1000
920
940
5460
5300
6130
4860
5400
5660
6390
7630
8420
4400
3980
0.96
1.14
1.21
1.23
[19.3]
[19.0]
[21.1]
[19.1]
17.2
11.3
b
d
d
d
4310
1
1.15
10.3
11.8
13.6
13.9
12.9
10.4
3380
1980
2450
1980
3090
1920
2.15
1.50
1.81
1.41
1.25
1.01
1.38
0.67
0.97
0.60
0.82
0.44
2700
2830
3200
3820
4210
6280
16.6
7273
13.9
6946
16.8
8018
9.65
14.3
a
TD-DFT values for 1+−5+ are calculated for symmetrical DFT geometries, whereas the experimental data suggest that these cations belong to class-
b
II; accordingly, the calculated νmax and μge values for those compounds cannot be directly compared to experiment and are given in brackets. From
eq 4 using experimental values of νmax and μge and equating Rab to the N−N distance in the AM1 geometries of the neutral species. From eq 3 using
the experimental values of νmax. Appears to be approximately symmetrical and assumed to be so due to overlap with other bands on the high-energy
̅
c
̅
d
̅
side.
asymmetric; the asymmetry has been explained in terms of
coupling of the electron-transfer coordination to symmetric
vibrational modes.19,41−43 Compared to the IVCT bands of
6+−8+, that of 9+ is less asymmetric and is broader, although
still narrower than the Hush limit, and that of 10+ is relatively
narrow and essentially symmetrical.44 The solvatochromic
behavior of the IVCT bands of selected examples is also
broadly consistent with the assignments to classes II and III
(see the Supporting Information for details).
IVCT Bands and Electronic Coupling. Compound 1
combines the most electron-rich terminal aryl groups and the
most electron-poor bridging arylene group, and the IVCT band
of 1+ is the weakest of those studied, in terms of both
absorptivity, εmax, and transition dipole, μge. For compound 4+,
the IVCT band is lower in energy and stronger than that of 1+,
which has similar electron-rich end groups but a more electron-
poor bridge. The increased strength can be attributed to
increased electronic coupling, V, between the two diabatic
states (vide infra). For a class-II system, the absorption
maximum is equal to the reorganization energy associated with
the intramolecular electron transfer, λ; the external (solvent)
contribution, λ0, to λ is related to the adiabatic electron-transfer
distance, R12, according to
The structures of 1+−10+ obtained from AM1/CI (CI =
configuration interaction) calculations also suggest varying
degrees of delocalization (see the Supporting Information for
full details).45 The structures of 1+−4+ are markedly unsym-
metrical. In the case of 1+, the oxidized amine is neither
perfectly planar nor coplanar with the adjacent portion of the
bridge; for this amine the N−Cbridge bonds are longer and the
terminal N−Caryl bonds shorter than those in the AM1
geometry of the corresponding neutral compound, and the
distyrylbenzene bridge retains the twisted character seen in the
neutral species, suggesting that the end groups for this cation
play a more important role than the bridge in stabilizing the
positive charge. The AM1/CI structures of 2+ and 3+ are
qualitatively similar to that we have previously discussed for
4+:25 one of the amines is coplanar with the adjacent stilbene
unit of the bridge and exhibits N−Cbridge and N−Caryl bonds
shorter and longer, respectively, than those for the neutral
species, while the rest of the structure resembles that of the
neutral species. In contrast, the calculations suggest that
oxidation of 5−10 results in both amino N−Cbridge bonds
shortening and all four terminal N−Caryl bonds lengthening,
more planar distyryl(hetero)arene bridges, and planarization at
both amine centers. However, except in the case of 9+, the C−
N bond changes are more significant at one of the amine
centers and the other amine center is somewhat twisted from
coplanarity with the bridge, perhaps reflecting some residual
tendency of the computational method toward overlocaliza-
tion.46,47 Nonetheless, the trends in the AM1/CI geometries are
broadly consistent with the experimental results, with the
compounds that experimentally appear to belong to class-II
having the least planar and least symmetrical structures, while
those that show class-III-like IVCT bands have structures that
are more extensively planarized and more symmetrical.
e2
4π 2a
1
1
2a2
1
R12
1
⎠ n
1
D
⎛
⎜
⎞
⎟
⎛
⎜
⎝
⎞
⎟
⎠
λ0 =
+
−
−
2
⎝
(2)
1
where e is the electronic charge, a1 and a2 are effective radii of
the two redox centers, and n and D are the solvent refractive
index and dielectric constant, respectively.48 Accordingly, the
shift to lower IVCT transition energy between 1+ and 4+ may
indicate a reduced value of R12. Indeed, the AM1/CI
calculations referred to above suggest that R12 decreases
substantially (from 16.8 to 7.9 Å, Table S3 in the Supporting
Information) between these two species.49 Such a decrease in
R
12 can be attributed to (i) a reduced diabatic electron-transfer
distance, Rab, associated with increased delocalization of the
diabatic states into the more electron-rich bridge and/or (ii)
more ground-state delocalization between the two diabatic
states associated with the increased V. A similar increase in μge
and bathochromic shift of νmax and decrease in the AM1/CI
̅
estimate of R12 are seen on reducing the electron-richness of
the terminal aryl groups, rather than increasing the electron-
richness of the bridge, in 2+ or 3+ (Table 2; Table S3 and
Figure 4, Supporting Information).
In class-III MV compounds, the IVCT absorption maximum
is a direct measure of the electronic coupling50 according to
2V = νm
(3)
̅
ax
The experimental trends are consistent with the expected
effects on V of varying the electron-richness of the terminal and
10149
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
Figure 4. Representations of the three highest occupied orbitals of 1 and 9, along with the HOMOs for fragments for the triarylamine and bridge
groups, obtained at the B3LYP/6-31G(d) level of theory.
bridging groups. Thus, there are successive blue shifts in νmax
for 9 and 10 the largest coefficients are on the pyrrole ring
(Figure 4). Thus, although the calculations do confirm that the
lowest energy transitions in each case have significant amine−
amine IVCT character, they also suggest that, for the more
strongly delocalized species, the band has significant amine →
bridge charge-transfer character. The important role played by
the bridge in the IVCT bands of species such as 9+ means that
the two-site Hush model is of limited applicability: a complete
treatment of these species would require consideration of the
couplings between three sitesthe two terminal redox sites
and a bridge redox siteusing an approach similar to that used
in ref 23. However, parametrizing such a three-site model
experimentally requires reliable assignment of, and reliable
extraction of transition energies and transition dipole moments
from, both electronic transitions between the resultant three
potential surfaces. Accordingly, and to facilitate direct
comparison with species such as 1+, which are much better
approximated as two-site systems, in this study we limit our
discussion to effective couplings extracted using a two-site
treatment.
̅
from 6+ to 8+ to 10+, consistent with the expected increase in
electron-richness from dialkoxybenzene to dialkoxythiophene
to dialkoxypyrrole; the blue shift seen between 7+ and 8+ can be
attributed to the reduced electron-richness of the end groups,
and the blue shift seen between 9+ and 10+ can be attributed to
both the more electron-rich bridge and the less electron-rich
end groups in the latter species. These shifts are all reproduced
by time-dependent density functional theory (TD-DFT)/
B3LYP calculations (Table 2), suggesting, as in some of our
previous studies,24,51 that although inappropriate for class-II
systems,45 this methodology can provide insight into the
electronic properties of delocalized (class-III-like) species.52
The TD-DFT-calculated transitions for these class-III species
are, in each case, well-described as SOMO − 1 → SOMO
transitions, these orbitals corresponding closely to the (HOMO
− 1)s and HOMOs, respectively, of the neutral species. For all
compounds, the HOMO − 1 and HOMO (Figure 4) can be
regarded as combinations of the local HOMOs of the two
triarylamine portions of the molecule: the HOMO − 1 is
essentially localized on the two triarylamine moieties, whereas
the HOMO is an antibonding combination of the HOMO of
the divinylarene/heterocycle of the bridge with the local
HOMOs of the two triarylamines.53 The relative bridge and
triarylamine contributions to the HOMO vary considerably
among the series: in the case of 1 the bridge contributions are
minor and result primarily from the vinylene groups, whereas
Finally, 5+ shows the lowest energy IVCT maximum of any
of the compounds investigated here. This is consistent with its
assignment to either class-II or class-III; in the class-II case one
would anticipate reduced R12 vs 4+ due to the less electron-rich
terminal aryl groups, whereas if 5+ belongs to class-III, the
apparent coupling would be expected to be smaller than that of
6+, in which the bridge is more electron-rich.
10150
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
As noted above, the electronic coupling, V, between the two
redox sites of a class-III system can be obtained directly from
the IVCT absorption maximum. However, for both class-II and
III species, V can be estimated from characteristics of the IVCT
using the Hush expression
compounds. We have previously compared DFT-calculated and
experimental values of ESR hyperfine coupling constants to
establish that DFT gives a good description of the spin-density
distribution in the radical cations of some bis(diarylamino)
derivatives of thiophene-based bridges.24 Such an approach is
slightly less straighforward here in that the experimental ESR
spectra (Figure 5; Figure S6, Supporting Information) display
μ νm
eRab
̅
ax
ge
V =
(4)
where e is the elementary charge and Rab is the diabatic
electron-transfer distance, i.e., the effective separation between
the two redox sites in the absence of electronic coupling.54 The
values of VNN given in Table 2 were obtained from eq 4 using
μge and νmax taken from the experimental spectra and using the
̅
nitrogen−nitrogen distance, RNN, for Rab. The values generally
increase down the series 1+−10+ as the bridge becomes more
electron-rich, varying by a factor of ca. 2 between extremes, and
are consistently larger for the experimentally class-III-like
compounds 6+−10+ than for the class-II species 1+−4+. This is
broadly consistent with expectations based upon the electron-
richness of terminal and bridging groups. The slight
inconsistencies (for example, the experimentally indistinguish-
able estimates for 1+ and 2+) may reflect the (variable)
discrepancy between Rab and RNN (vide infra); uncertainties in
the determination of μge may also play a role, whereas in the
case of 5+ it is unclear how best to apply Hush theory to the
unusual line shape of the IVCT band.
Figure 5. Left: first-derivative ESR spectra of selected diamine radical
cations in CH2Cl2. Right: line widths of ESR spectra of diamine radical
cations obtained from experiment (open circles) and simulation based
on DFT hyperfine coupling constants and a 0.5 G line broadening
(solid squares).
However, as noted in the Introduction, displacement of the
diabatic states into the bridge means that Rab is generally
considerably less than RNN, meaning that use of RNN leads to
underestimates of V. Indeed, we have previously reported
little to no resolvable coupling and, in many cases, are
structureless Gaussian absorptions: the previously reported
room-temperature ESR spectra of 1+ and 4+ are characterized
by poorly resolved coupling to two 14N centers (consistent with
rapid intramolecular electron transfer between the two redox
centers),26 while spectra of 5+, 6+, 7+, 8+, and 10+ acquired
under the same conditions show an even less well-resolved
structure. However, there is considerable variation in the overall
line width of the ESR signals from cation to cation. In principle,
intermolecular exchange between cations and neutral species
can lead to narrowing of spectra; however, such effects give rise
to a Lorentzian line shape26,58 and are not observed in related
systems at comparable concentrations,24,26,58 so it is assumed
that the line widths are determined by the unresolved hyperfine
coupling constants. Accordingly, hyperfine constants obtained
from B3LYP/6-31G(d) calculations were used in combination
with a constant Gaussian line broadening to simulate spectra
using WinSIM;59−61 the overall widths at half-height of the
integrated spectra were then compared with those of the
experimental integrated spectra. It should be noted that in all
cases the calculated coupling constants were obtained for
symmetrical DFT geometries, whereas NIR data (vide supra)
and, for 1+, low-temperature ESR data26 indicate some of the
cations are class-II species, meaning that the calculations cannot
describe the instantaneous spin distribution within these
cations; however, we have previously shown that DFT can
nonetheless give a reasonable description of the average
coupling constants and, therefore, spin distribution resulting
from intermolecular exchange.26 As shown in Figure 5, both
experimental and simulated line widths decrease in the order 1+
> 4+ > 5+ > 6+; experiment and simulations indicate similar line
widths for 6+, 7+, and 8+ but disagree on the details, and finally,
both experiment and simulation show an increased line width
for 10+ relative to 7+ and 8+. The calculations indicate that
computational estimates of Rab for 4+ that are 46−61% of
25
RNN
,
while we have used variable-temperature (VT) ESR to
26
estimate a value for 1+ that is 39% of RNN
.
For the species
with class-III-like line shapes, coupling can also be estimated
using eq 3; these values are given in Table 2 and indicate that
use of RNN both underestimates V and underestimates the
variation of V between the class-III-like members of the series.
The values of Vclass III obtained from eq 3 are larger by a factor
of ca. 3 (6+−8+) or 4 (9+ and 10+) than the values of VNN
,
which may be compared to the factors of ca. 1.5−2 seen for
class-III species with shorter bridges.24,25 We have also
computationally estimated Rab for the present series (Table
S3, Supporting Information), both from TD-DFT transition
dipole moments, obtained at the symmetric DFT geo-
metries,41,55 and from the relationship among μab, μge, and
Δμ12,56,57 using experimental values of μge and AM1/CI values
of Δμ12: the first approach gives values of Rab that vary from ca.
R
NN/2 for 1+ to ca. RNN/3 for 10+, while the second approach
suggests variation from close to RNN (1+) to ca. RNN/4 (10+).
However, while the values estimated vary considerably with the
specific approach, Rab is found to generally decrease with more
electron-rich bridging groups or more electron-poor end
groups; i.e., the factors that lead to increased V also lead to
reduced Rab. Therefore, these computational estimates also
suggest that use of RNN leads not only to underestimates of V,
but also to underestimates of the variation of V between the
most weakly and most strongly coupled examples.
Monocation Electron Spin Resonance Spectra and
Spin Distribution. As noted above, the SOMOs of the more
delocalized examples are increasingly concentrated on the
bridging groups, raising the question of the extent to which
these radical cations can still be regarded as diamine MV
10151
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
Table 3. Two Alternative Partitionings of the DFT-Calculated Spin Densities between Terminal and Bridging Groups of
Diamine Radical Cations
these trends in line width are primarily due to decreasing
coupling to the diamine 14N nuclei (AN is calculated to decrease
from 3.44 G for 1+ to 1.81 G for 10+), opposed by increasing
other weakly coupled bis(triarylamine) radical cations, where
they have previously been assigned to localized transitions
within the Ar2N+ units.18 Figure 6 shows that, for 1, the dication
spectrum is dominated by an absorption at essentially the same
energy as the “Ar2N+” absorption of the monocation with
1
coupling to bridge nuclei (most dramatically the vinylene H
nuclei and, in the case of 10+, the pyrrole 14N nucleus, for
which AN = −1.44 G); moreover, DFT is seen to provide a
reasonable description of the spin distribution in these cations,
consistent with the observation (vide supra) that TD-DFT
transition energies for the delocalized examples are in
reasonable agreement with experiment.
Table 3 summarizes how the DFT spin densities are
distributed. The oxidation of 1+ is the most strongly
triarylamine-localized, with 84% of the spin density on these
groups, but nonetheless shows some contribution from the
divinyldicyanobenzene bridge, consistent with the reduced
values of Rab implied by some of the computational estimates
(see above and the Supporting Information) and by previous
VT-ESR studies.26 The spin densities on the triarylamine
moieties generally decrease down the series, with all
experimentally class-III-like radical cations 6+−10+ showing
reduced triarylamine spin density relative to their class-II
counterparts (1+−4+). Moreover, the successively increased
bridge character of the oxidation seen among 6+, 8+, and 10+
(dialkoxybenzene vs dialkoxythiophene vs dialkoxypyrrole) is
consistent with electrochemical comparisons to model
compounds (vide supra; Table S1, Supporting Information).
While there is no clear-cut criterion on what can and cannot be
considered an MV compound, it is clear that MV character
decreases down the series (1+−10+) at the expense of bridge-
centered oxidation. However, in all cases, apart from that of
10+, more than half the spin is located on the triarylamine units,
indicating significant retention of MV character.
Electronic and Electron Spin Resonance Spectra of
the Dications. For selected examples we have also obtained
vis−NIR and ESR spectra of the corresponding dications. In
addition to the low-energy bands assigned to IVCT discussed in
the previous sections, the localized radical cations 1+−4+ (as
well as the putatively borderline cation 5+) all exhibit strong
absorptions at 13000−15000 cm−1 (see Figures S2 and S3 in
the Supporting Informaiton and the figures in refs 25 and 26) at
energies similar to those of the absorptions of isolated
triarylamine radical cations62−64 and of bands observed for
Figure 6. Vis−NIR absorption spectra of monocations (solid lines)
and dications (broken lines) of 1 (above) and 8 (below) in CH2Cl2.
The strong absorptions at high energy in the monocation spectra are
due to the presence of an excess of the neutral species. Note the
different vertical scales in the two graphs.
10152
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
approximately twice the absorptivity,65 suggesting the dication
can be regarded as two more-or-less noninteracting triarylami-
nium units. This is consistent with the location of the 1n+
system toward the weakly coupled end of the class-II spectrum,
where the redox centers are almost independent, so as
described by Robin and Day,40 apart from a weak IVCT
band, the electronic spectrum of the MV species exhibits
features characteristic of both the fully oxidized and reduced
species (i.e., dication and neutral, respectively, in the present
series). On the other hand, the spectrum of 42+ is not so clearly
related to that of 4+, being dominated by a strong absorption at
9300 cm−1 (Figure 2 of ref 25), similar to that seen for 52+−82+
(9500−10000 cm−1; Figure 6b; Figues S3−S5, Supporting
Information), indicating deviation from the picture of non-
interacting triarylaminium units. However, while 42+ also
exhibits an “Ar2N+” absorption in a position similar to that of
the monocation, 52+−82+ do not.66
Several limiting magnetic situations can be envisaged for
dications of diamines: (a) strong ferromagnetic coupling
resulting in an S = 1 configuration, (b) strong antiferromagnetic
coupling resulting in an S = 0 situation, and (c) two
independent S = 1/2 radical centers (i.e., a vanishing energy
gap between singlet and triplet). The first scenario is typically
favored for systems in which the relevant orbitals are
orthogonal, such as those with 1,3-phenylene bridges, whereas
in the present systems with 1,4-phenylene (or pyrrole-2,5-diyl
or 2,5-thienylene) bridges, singlet (b), “biradical” (c), or
intermediate situations might be expected, depending on the
degree of interaction between the redox centers. For several of
the bis[(diarylamino)styryl] compounds for which we were
able to obtain reasonably stable dication solutions, we
estimated room-temperature spin concentrations by double
integration of the first-derivative ESR spectra. We have
previously reported that 72+ is a singlet (ESR, NMR,
comparison of bond lengths in the crystal structure with
those calculated for various electronic configurations)33 and
that solutions of 42+ have a spin concentration according to
Figure 7. Comparison of electronic coupling in the monocation (solid
squares), estimated using eq 4 and the N−N separation, with the
number of spins per dication molecule (open circles), estimated by
ESR, for some bis[(diarylamino)stryryl]benzenes. The dotted lines are
provided as a guide to the eye but have no physical significance.
bis[(diarylamino)styryl]arene/heteroarene radical cations.
While a few previous studies have recognized the importance
of relative bridge and end-group orbital energies in organic MV
compounds, the present work includes more extensive
variations of these relative energetics, resulting in a large
variation in the delocalization behavior and a more
comprehensive understanding of how this behavior can be
controlled: a combination of experimental and computational
data show that these cations can be tuned from weakly coupled
strongly unsymmetrical class-II MV ions (1+), via more strongly
coupled class-II species (e.g., 4+) and an apparently borderline
case (5+), to delocalized symmetrical class-III MV-type species
(e.g., 6+) and to symmetrical species that might arguably be
better described as “bridge-based” (10+). Using a two-state
Hush analysis of the lowest energy NIR transitions, the
apparent electronic couplings between the redox centers, V,
vary by at least a factor of ca. 2 when Rab is assumed to be
roughly constant. However, orbitals, spin densities (validated
by comparison of DFT-simulated and experimental ESR
spectra), and transition dipoles obtained from TD-DFT
calculations indicate increasing displacement of the redox
centers into the bridge as the bridge electron-richness is
increased relative to that of the terminal groups, thus further
accentuating the variation in V. The corresponding dications
vary from diradicals, consisting of two weakly interacting
triarylaminium moieties, to singlets (i.e., strong antiferromag-
netic coupling), the trends in magnetic coupling in the
dications paralleling those in electronic coupling in the radical
cations.
ESR that corresponds to 1.2
0.1 S = 1/2 centers per
molecule.25 We found that 52+, 62+, and 82+ are also essentially
diamagnetic (consistent with the similarity of the electronic
spectra of 52+−82+, vide supra).67 On the other hand, the spin
concentration measured by ESR for 12+ essentially corresponds
to the “biradical” limit, consistent with the dication vis−NIR
spectrum described above, which also suggests largely
independent triarylaminium centers. Figure 7 compares the
number of spins measured by room-temperature ESR for the
dications with the electronic couplings estimated for the
corresponding MV radical monocations using eq 4 and taking
Rab = RNN. The figure clearly indicates that the trend in
magnetic coupling in these dications parallels that in electronic
coupling in the corresponding MV cations. We have previously
made similar observations in a series of bis(di-p-anisylamino)-
terminated oligo(phenylenevinylene)s with varying bridge
length (compound 4 being a member of that series).25,68 In
both the previously studied and present series, for the species
where V is sufficiently strong to lead to class-III MV behavior in
the monocation, the magnetic coupling is sufficiently strong to
result in a room-temperature singlet configuration for the
dication.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details for synthesis, mono- and dication
characterization, and quantum chemical calculations, additional
electrochemical data, figures showing additional complete
mono- and dication spectra, solvatochromic data for selected
cations, computational estimates of geometric and electron-
transfer distances and of electronic coupling, computed
geometric parameters, and XYZ coordinates and absolute
energies (hartrees) for the DFT-optimized geometries of the
neutral and radical cation species. This material is available free
of charge via the Internet at http://pubs.acs.org.
SUMMARY
■
We have varied the relative electron-richness of the terminal
aryl and bridging arene/heteroarene groups in a series of
10153
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
(22) Lambert, C.; Noll, G.; Schelter, J. Nat. Mater. 2002, 1, 69.
(23) Lambert, C.; Amthor, S.; Schelter, J. J. Phys. Chem. A 2004, 108,
6474.
̈
AUTHOR INFORMATION
■
Corresponding Author
stephen.barlow@chemistry.gatech.edu
(24) Odom, S. A.; Lancaster, K.; Beverina, L.; Lefler, K. M.;
Thompson, N. J.; Coropceanu, V.; Bred
S. Chem.Eur. J. 2007, 13, 9637.
(25) Barlow, S.; Risko, C.; Chung, S.-J.; Tucker, N. M.; Coropceanu,
V.; Jones, S. C.; Levi, Z.; Bredas, J. L.; Marder, S. R. J. Am. Chem. Soc.
2005, 127, 16900.
(26) Lancaster, K.; Odom, S. A.; Jones, S. C.; Thayumanavan, S.;
Marder, S. R.; Bredas, J. L.; Coropceanu, V.; Barlow, S. J. Am. Chem.
́
as, J.-L.; Marder, S. R.; Barlow,
Notes
The authors declare no competing financial interest.
́
ACKNOWLEDGMENTS
■
This work was supported by the National Science Foundation,
through the Science and Technology Center Program (Grant
DMR-0120967) and a Graduate Research Fellowship to S.A.O.,
and by the Office of Naval Research, through the DARPA
MORPH Program (Grant N-00014-06-1-0897). Part of the
computational resources have been provided by Georgia Tech's
Center for Computational Molecular Science and Technology,
which is funded through a NSF CRIF award (Grant No. CHE-
0946869) and by the Georgia Institute of Technology. We also
thank S. Thaymanavan and Michael D. Levin for the syntheses
of 1−3 and John R. Reynolds for a gift of an intermediate used
to obtain 10.
́
Soc. 2009, 131, 1717.
(27) Kaim, W. Inorg. Chem. 2011, 50, 9752 and references therein..
(28) Roussel, P.; Cary, D. R.; Barlow, S.; Green, J. C.; Varret, F.;
O’Hare, D. Organometallics 2000, 19, 1071.
(29) Pond, S. J. K.; Rumi, M.; Levin, M. D.; Parker, T. C.; Beljonne,
D.; Day, M. W.; Bredas, J. L.; Marder, S. R.; Perry, J. W. J. Phys. Chem.
A 2002, 106, 11470.
(30) Frederiksen, P. K.; Jørgensen, M.; Ogilby, P. R. J. Am. Chem. Soc.
2001, 123, 1215.
(31) Kauffman, J. M.; Moyna, G. J. Org. Chem. 2003, 68, 839.
(32) Rumi, M.; Ehrlich, J. E.; Heikal, A. A.; Perry, J. W.; Barlow, S.;
Hu, Z.; McCord-Maughon, D.; Parker, T. C.; Rockel, H.;
̈
Thayumanavan, S.; Marder, S. R.; Beljonne, D.; Bred
Chem. Soc. 2000, 122, 9500.
́
as, J.-L. J. Am.
REFERENCES
■
(1) Hankache, J.; Wenger, O. S. Chem. Rev. 2011, 111, 5138.
(2) Heckmann, A.; Lambert, C. Angew. Chem., Int. Ed. 2012, 51, 326.
(3) Hush, N. S. Prog. Inorg. Chem. 1967, 8, 391.
(4) Nelsen, S. F.; Tran, H. Q.; Nagy, M. A. J. Am. Chem. Soc. 1998,
120, 298.
(5) Tang, C. W.; Van Slyke, S. A. Appl. Phys. Lett. 1987, 51, 913.
(6) Kido, J.; Kimura, M.; Nagai, K. Science 1995, 267, 1332.
(7) Bulovic, V.; Gu, G.; Burrows, P. E.; Forrest, S. R.; Thompson, M.
E. Nature 1996, 380, 29.
(33) Zheng, S.; Barlow, S.; Risko, C.; Kinnibrugh, T. L.; Khrustalev,
V. N.; Antipin, M. Y.; Tucker, N. M.; Timofeeva, T.; Coropceanu, V.;
Jones, S. C.; Bred
1812.
́
as, J.-L.; Marder, S. R. J. Am. Chem. Soc. 2006, 128,
(34) Zheng, S.; Beverina, L.; Barlow, S.; Zojer, E.; Fu, J.; Padilha, L.
A.; Fink, C.; Kwon, O.; Yi, Y.; Shuai, Z.; Van Stryland, E. W.; Hagan,
D. J.; Bred
(35) Kwon, O.; Barlow, S.; Odom, S. A.; Beverina, L.; Thompson, N.
J.; Zojer, E.; Bredas, J.-L.; Marder, S. R. J. Phys. Chem. A 2005, 109,
9346.
́
as, J.-L.; Marder, S. R. Chem. Commun. 2007, 1372.
́
(8) Borsenberger, P. M.; Weiss, D. S. Organic Photoreceptors for
Xerography; Marcel Dekker: New York, 1998.
(36) The ionization energies of pyrrole and thiophene are 8.9 and 8.2
eV, respectively (Sell, J. A.; Kuppermann, A. Chem. Phys. Lett. 1978,
61, 355), although a smaller difference might be expected between
pyrrole and dialkoxythiophene. We have previously noted that
alkylpyrrole derivatives are more readily oxidized than their
dialkoxythiophene analogues (see ref 34 and Zheng, S.; Leclercq, A.;
Fu, J.; Beverina, L.; Padilha, L. A.; Zojer, E.; Schmidt, K.; Barlow, S.;
Luo, J.; Jiang, S.-H.; Jen, A. K.-Y.; Yi, Y.; Shuai, Z.; Van Stryland, E. W.;
(9) Littleford, R. E.; Paterson, M. A. J.; Low, P. J.; Tackley, D. R.;
Jayes, L.; Dent, G.; Cherryman, J. C.; Brown, B.; Smith, W. E. Phys.
Chem. Chem. Phys. 2004, 6, 3257.
(10) Low, P. J.; Paterson, M. A. J.; Puschmann, H.; Goeta, A. E.;
Howard, J. A. K.; Lambert, C.; Cherryman, J. C.; Tackley, D. R.;
Leeming, S.; Brown, B. Chem.Eur. J. 2004, 10, 83.
(11) Spangler, C. W. J. Mater. Chem. 1999, 9, 2013.
(12) Huang, C.; Sartin, M. M.; Siegel, N.; Cozzuol, M.; Zhang, Y.;
Hales, J. M.; Barlow, S.; Perry, J. W.; Marder, S. R. J. Mater. Chem.
2011, 21, 16119.
́
Hagan, D. J.; Bredas, J.-L.; Marder, S. R. Chem. Mater. 2007, 19, 432).
(37) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877.
(38) Sutton, J. E.; Taube, H. Inorg. Chem. 1981, 20, 3125.
(39) Ribou, A. C.; Launay, J.-P.; Sachtleben, M. L.; Li, H.; Spangler,
C. W. Inorg. Chem. 1996, 35, 3735.
(13) Wolfe, J. P.; Wagaw, S.; Buchwald, S. L. J. Am. Chem. Soc. 1996,
118, 7215.
(14) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7217.
(15) Seo, E. T.; Nelson, R. F.; Fritsch, J. M.; Marcoux, L. S.; Leedy,
D. W.; Adams, R. N. J. Am. Chem. Soc. 1966, 88, 3498.
(16) Dapperheld, S.; Steckhan, E.; Grosse-Brinkhaus, K.-H.; Esch, T.
Chem. Ber. 1991, 124, 2557.
(40) Robin, M. B.; Day, P. Adv. Inorg. Chem. Radiochem. 1967, 10,
247.
(41) Coropceanu, V.; Malagoli, M.; Andre,
Chem. Soc. 2002, 124, 10519.
(42) Coropceanu, V.; Malagoli, M.; Andre,
Chem. Phys. 2001, 115, 10409.
́
J. M.; Bred
́
as, J. L. J. Am.
́
J. M.; Bred
́
as, J. L. J.
(17) Brunschwig, B. S.; Creutz, C.; Sutin, N. Chem. Soc. Rev. 2002,
31, 168.
(43) For examples of other work recognizing the importance of
symmetric modes in the IVCT line shapes of delocalized and nearly
delocalized MV systems, see: Reimers, J. R.; Hush, N. S. Chem. Phys.
1996, 208, 177. Piepho, S. B. J. Am. Chem. Soc. 1988, 110, 6319.
Borras-Almenara, J. J.; Coronado, E.; Ostrovsky, S. M.; Palii, A. V.;
́
Tsukerblat, B. S. Chem. Phys. 1999, 240, 149. Bailey, S. E.; Zink, J. I.;
Nelsen, S. F. J. Am. Chem. Soc. 2003, 125, 5939.
(18) Lambert, C.; Noll, G. J. Am. Chem. Soc. 1999, 121, 8434.
(19) Szeghalmi, A. V.; Erdmann, M.; Engel, V.; Schmitt, M.; Amthor,
̈
S.; Kriegisch, V.; Noll, G.; Stahl, R.; Lambert, C.; Leusser, D.; Stalke,
̈
D.; Zabel, M.; Popp, J. J. Am. Chem. Soc. 2004, 126, 7834.
(20) Coropceanu, V.; Gruhn, N. E.; Barlow, S.; Lambert, C.;
́
Durivage, J. C.; Bill, T. G.; Noll, G.; Marder, S. R.; Bredas, J.-L. J. Am.
̈
Chem. Soc. 2004, 126, 2727.
(44) We have previously observed that the IVCT of a delocalized
bis(diarylamino)dithienopyrrole radical cation shows a band shape
rather different from those of analogues with bithiophene or
dithienothiophene bridges (ref 24).
(45) Optimization of the structures of mixed-valence species typically
gives highly methodology-dependent results; electron self-interaction
errors in DFT methods tend to afford delocalized structures for open-
shell systems, while the neglect of electron correlation effects in
(21) Other features of bridging groups besides their local orbital
energies can play a role in determining electronic coupling. For
example, the radical cation of 9,10-bis[bis(4-methoxyphenyl)amino]
anthracene radical cation is less strongly coupled than its p-phenylene-
bridged analogue due to the disruption of oribital overlap through
steric effects: Lambert, C.; Risko, C.; Coropceanu, V.; Schelter, J.;
Amthor, S.; Gruhn, N. E.; Durivage, J. C.; Bred
́
as, J. L. J. Am. Chem.
Soc. 2005, 127, 8508.
10154
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155

Journal of the American Chemical Society
Article
Hartree−Fock (HF) methods often leads to localized solutions: Bally,
T.; Borden, W. T. In Reviews in Computational Chemistry; Lipkowitz,
K., Boyd, D. B., Eds.; Wiley-VCH: New York, 1999; Vol. 13, p 1.
Pacchioni, G.; Frigoli, F.; Ricci, D. Phys. Rev. B 2000, 63, 054102.
Kobko, N.; Masunov, A.; Tretiak, S. Chem. Phys. Lett. 2004, 392, 444.
Lundberg, M.; Siegbahn, P. E. M. J. Chem. Phys. 2005, 122, 224103.
Indeed, UB3LYP/6-31G(d) structures (see the Supporting Informa-
tion) of 1+−10+ are all symmetric, i.e., suggesting all the species belong
to class-III, whereas UHF-AM1 results suggest all belong to class-II.
Calculations employing AM1 coupled with CI, to include some degree
of electron correlation, can, in some cases, more usefully distinguish
between delocalized and localized systems; for example, consistent
with experiment, class-II and class-III structures are predicted by
AM1/CI calculations for 4+ and its shorter homologue, the (E)-4,4′-
bis[bis(4-methoxyphenyl)amino]stilbene cation, respectively (ref 25).
Recently, hybrid DFT/HF approaches have been used to successfully
model (de)localization phenomena in organic MV systems: Renz, M.;
Theilacker, K.; Lambert, C.; Kaupp, M. J. Am. Chem. Soc. 2009, 131,
exact within the two-state model regardless of the magnitude of V
(Creutz, C.; Newton, M.; Sutin, N. J. Photochem. Photobiol., A 1994,
82, 47), at least for cases where only antisymmetric vibrational modes
are coupled to the electron transfer. As first pointed out by Hush
(Hush, N. S. ACS Symp. Ser. 1982, 198, 301), however, eq 4 will
overestimate the coupling in cases where the electron transfer is also
coupled to symmetric modes. As discussed in ref 50, neglect of
coupling to symmetric modes is likely to have small effects in the
current systems. In particular, these effects are likely to be insignificant
compared to the effects of overestimating Rab discussed in the text.
(55) Piepho, S. B.; Krausz, E. R.; Schatz, P. N. J. Am. Chem. Soc.
1978, 100, 2996.
(56) Reimers, J. R.; Hush, N. S. J. Phys. Chem. 1991, 95, 9773.
(57) Cave, R. J.; Newton, M. D. Chem. Phys. Lett. 1996, 249, 15.
(58) Veregin, R. P.; Harbour, J. R. J. Phys. Chem. 1990, 94, 6231.
(59) Duling, D. R. J. Magn. Reson., Ser. B 1994, 104, 105.
(60) http://www.niehs.nih.gov/research/resources/software/tox-
pharm/tools/index.cfm (accessed June 4, 2012).
(61) The widths of the simulated spectra depend on the value of the
Gaussian line broadening chosen for convolution with the DFT
hyperfine constants. However, for small broadenings, sufficient to
obscure hyperfine coupling, but much smaller than the overall line
width, the line broadening does not affect the pattern of variation
between compounds.
16292. Kaupp, M.; Renz, M.; Parthey, M.; Stolte, M.; Wurthner, F.;
̈
Lambert, C. Phys. Chem. Chem. Phys. 2011, 13, 16973.
(46) On the other hand, we cannot rule out the possibility that 6+−8+
and 10+ lie in Meyer’s “solvent-delocalized” intermediate class-II/III,
which can be difficult to experimentally distinguish from class-III:
Demadis, K. D.; Hartshorn, C. M.; Meyer, T. J. Chem. Rev. 2001, 101,
2655.
(62) Walter, R. I. J. Am. Chem. Soc. 1966, 88, 1923.
(63) Schmidt, W.; Steckhan, E. Chem. Ber. 1980, 113, 577.
(64) Robinson, J.; Osteryoung, R. A. J. Am. Chem. Soc. 1980, 102,
4415.
(47) Indeed, the adiabatic electron-transfer distance, R12, indicated
for the AM1/CI structures for 1+ is considerably greater than the
diabatic distance, Rab, deduced in ref 26 from comparison of optical
and VT-ESR data, again suggesting that the AM1/CI structures are
overlocalized.
(65) In the case of 3+ and 4+, an additional band between the IVCT
and “Ar2N+” absorptions may correspond to a charge transfer from the
highest π-bridge-based orbital to the radical cation center (see refs 22
23, and 25).
(48) Meyer, T. J. Acc. Chem. Res. 1978, 11, 94.
(49) The decrease in λ0 predicted between 1+ and 4+ from eq 2 using
the AM1/CI values of R12 and assuming the invarient donor and
acceptor radii in both cases (2970 cm−1) is somewhat larger than the
experimental shift (1320 cm−1). However, the AM1/CI values of R12
are likely overestimates (see ref 47), while the varied deviations of
both Rab and R12 from the geometric N−N distance mean that
approximating the redox centers in both cases as triarylamines is
inappropriate.
(66) For these more delocalized monocations, the bands in this
region are somewhat red-shifted from those seen for the class-II
species, as previously seen for other class-III bis(diarylamino) MV
compounds (refs 10 , 18 , and 25), and perhaps having a rather
different origin (ref 10).
(67) Trace radical signals can arise from some small paramagnetism
in the dication, from trace monocation arising from imperfect
stoichiometry in the oxidation or from decomposition of the dication,
or from trace oxidizing agent arising from imperfect stoichiometry.
(68) For examples of previous studies comparing magnetic coupling
in doubly oxidized (or reduced) organic biradicals and the
corresponding singly oxidized (or reduced) MV systems, see ref 25.
Also see: Nelsen, S. F.; Ismagilov, R. F.; Teki, Y. J. Am. Chem. Soc.
1998, 120, 2200. Shultz, D. A.; Fico, R. M.; Bodnar, S. H.; Kumar, K.;
Vostrikova, K. E.; Kampf, J. W.; Boyle, P. D. J. Am. Chem. Soc. 2003,
125, 11761. Correlations between magnetic and electronic coupling in
homovalent and MV species, respectively, have also been seen in
systems with metal-based redox centers; for example: Fabre, M.;
Bonvoisin, J. J. Am. Chem. Soc. 2007, 129, 1434.
(50) In the case of significant coupling of the electron-transfer
reaction coordinate to symmetric vibrational modes, the coupling is
overestimated using eq 3 by a quantity equal to half the relaxation
energy associated with the symmetric modes: Coropceanu, V.;
́
Boldyrev, S. I.; Risko, C.; Bredas, J. L. Chem. Phys. 2006, 326, 107 .
Although the asymmetry of the IVCT bands of 6+−8+ is, as previously
discussed for other bis(triarylamine) MV systems (refs 19 , 41, and 42)
attributable to the effects of coupling to symmetric modes, the strength
of this coupling decreases rapidly as the size of the system is increased
(refs 20 , 41, and 51), and the overall IVCT line widths seen for 6+−8+
are also consistent with much less significant contributions than
previously found in shorter species with benzene, biphenyl, and tolane
bridges. Thus, the effect of coupling to symmetric modes on V in the
current systems has been neglected.
(51) Risko, C.; Coropceanu, V.; Barlow, S.; Geskin, V.; Schmidt, K.;
Gruhn, N. E.; Marder, S. R.; Bred
7959.
́
as, J.-L. J. Phys. Chem. C 2008, 112,
(52) The calculations fail to correctly reproduce the relative positions
of the IVCT maxima of 6+ and 7+ and of 8+ and 9+; however, in both
of these comparisons there are opposing effects expected, those of
bridging and end groups in the first case and of changing the bridging
heteroatom and removing alkoxy substituents from the bridge in the
latter.
(53) In the case of 9 the HOMO − 2 can be well-described as the
corresponding bonding combination, whereas none of the orbitals of 1
can be approximated in such a straightforward fashion (although the
HOMO − 2 of 1 still has a considerable contribution from the local
bridge HOMO).
(54) Although eq 4 was originally derived at the perturbation limit of
weak electronic coupling (ref 3), it was subsequently shown to be
10155
dx.doi.org/10.1021/ja3023048 | J. Am. Chem. Soc. 2012, 134, 10146−10155