Effects of dipolar interactions on linear and nonlinear optical properties of multichromophore assemblies: A case study

Article abstract of DOI:10.1002/chem.200500910

Interchromophore interactions in flexible multidipolar structures for nonlinear optics were addressed by a combined experimental and theoretical study on two series of one-, two-, and three-chromophore systems in which identical push-pull chromophores are assembled through covalent and flexible linkers in close proximity. The photophysical and nonlinear optical properties (quadratic hyperpolarizability) of the multichromophore systems were investigated and compared to those of the monomeric chromophores. Multimers have larger dipole moments than their monomeric analogues, that is, the dipolar subchromophores self-orientate within the multimeric structures. This effect was found to depend on the intersubchromophore distance in a nontrivial manner, which confirms that molecular engineering of such flexible systems is more complex than in completely geometrically controlled systems. Electric-field-induced second-harmonic generation (EFISHG) measurements in solution revealed increased figures of merit as compared to the monomeric analogue. This effect increases with increasing number and polarity of the individual subchromophores in the nanoassembly and increasing spacing between dipolar subchromophores. Experimental results are interpreted by a theoretical model for interacting polar and polarizable chromophores. The properties of multidipolar assemblies are shown to be related to the relative orientation of chromophores, which is imposed by interchromophore interactions. The supramolecular structure is thus a result of self-organization. The proposed theoretical model was also used to predict the properties of multichromophore structures made up of more polar and polarizable push-pull chromophores, and showed that stronger interchromophore interactions can heavily affect the individual optical responses. This suggests new routes for engineering highly NLO responsive multichromophore systems.

Full text of DOI:10.1002/chem.200500910

FULL PAPER
DOI: 10.1002/chem.200500910
Effects of Dipolar Interactions on Linear and Nonlinear Optical Properties of
Multichromophore Assemblies: A Case Study
Francesca Terenziani, Olivier Mongin, Claudine Katan,
Bharath Kumar Goud Bhatthula, and Mireille Blanchard-Desce*[a]
In memory of Rüdiger Wortmann
Abstract: Interchromophore interac-
tions in flexible multidipolar structures
for nonlinear optics were addressed by
a combined experimental and theoreti-
cal study on two series of one-, two-,
and three-chromophore systems in
which identical push–pull chromo-
phores are assembled through covalent
and flexible linkers in close proximity.
The photophysical and nonlinear opti-
cal properties (quadratic hyperpolariza-
bility) of the multichromophore sys-
tems were investigated and compared
to those of the monomeric chromo-
phores. Multimers have larger dipole
moments than their monomeric ana-
logues, that is, the dipolar subchromo-
phores self-orientate within the multi-
meric structures. This effect was found
to depend on the intersubchromophore
distance in a nontrivial manner, which
confirms that molecular engineering
of such flexible systems is more com-
plex than in completely geometrically
controlled systems. Electric-field-in-
duced second-harmonic generation
(EFISHG) measurements in solution
revealed increased figures of merit as
compared to the monomeric analogue.
This effect increases with increasing
number and polarity of the individual
subchromophores in the nanoassembly
and increasing spacing between dipolar
subchromophores. Experimental results
are interpreted by a theoretical model
for interacting polar and polarizable
chromophores. The properties of multi-
dipolar assemblies are shown to be re-
lated to the relative orientation of
chromophores, which is imposed by in-
terchromophore
interactions.
The
supramolecular structure is thus
a
result of self-organization. The pro-
posed theoretical model was also used
to predict the properties of multichro-
mophore structures made up of more
polar and polarizable push–pull chro-
mophores, and showed that stronger in-
terchromophore interactions can heavi-
ly affect the individual optical respons-
es. This suggests new routes for engi-
neering highly NLOresponsive multi-
chromophore systems.
Keywords: chromophores · fluores-
cence · nonlinear optics · self-organ-
ization · theoretical models
Introduction
applications relies both on enhancing the response at the
molecular level and on the design and elaboration of supra-
molecular structures having optimized responses, by taking
advantage of symmetry and environmental effects. As the
synthetic approach evolves from the molecular to the supra-
molecular level, theoretical and interpretative models are
needed to guide design at all stages. This is a critical issue,
not only because real-world applications often involve con-
densed-phase systems ranging from solid-state devices to
soft condensed matter (including biological systems), in
which the individual molecular entities will interact, but also
in a more fundamental vein. In fact, while structure–proper-
ty relationships are quite actively and successfully investigat-
ed at the molecular level,[5] supramolecular structure–prop-
erty relationships remain an interesting open question,
which deserves further analysis.[1,6,7]
Molecular materials are among the most promising candi-
dates for advanced applications in electronics[1] and photon-
ics,[2,3] and organic probes exhibiting high nonlinear optical
(NLO) responses attract much attention in the field of bio-
logical imaging.[4] The search for good candidates for these
[a] Dr. F. Terenziani, Dr. O. Mongin, Dr. C. Katan,
Dr. B. K. G. Bhatthula, Dr. M. Blanchard-Desce
Synth›se et ElectroSynth›se Organiques (CNRS, UMR 6510)
Institut de Chimie, UniversitØ de Rennes 1
Campus Scientifique de Beaulieu
Bât 10 A, 35042 Rennes Cedex (France)
E-mail: mireille.blanchard-desce@univ-rennes1.fr
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3089

A great variety of structures can be generally classified as
“supramolecules”, in spite of their possible different natures,
ranging from crystals to films to multimolecular size-control-
led assemblies to aggregates and multibranched or dendritic
structures. However, all these architectures share the basic
feature of being characterized by interactions between the
“molecular bricks”. These interactions can be electrostatic
in nature, originate from orbital overlap, or even involve
active connectors between the components. While much ex-
perimental work has been devoted to the study of supramo-
lecules,[6] interpretative models are still scanty and relevant
only to a few specific types of interacting systems.
A seminal theoretical work on the role of intermolecular
interactions on the second-order optical properties of chro-
mophoric molecular assemblies, which already dates back to
more than ten years ago,[8] demonstrated that the hyperpo-
larizability of p-nitroaniline dimers and trimers strongly de-
pends on the relative molecular orientation of the chromo-
phores and on intermolecular chromophore–chromophore
interactions. This important result, confirmed by others,[9,10]
suggested some guidelines for the optimized design of chro-
mophoric assemblies with large hyperpolarizabilities.[11]
However, up to now, the still-unsolved problem is the
ability to predict the properties of a supramolecular archi-
tecture starting from the properties of its building blocks.
The benefit of such a bottom-up approach would be consid-
erable, offering the possibility of taking advantage of the
available knowledge on molecular properties and opening
the way to the design of knowledge-based materials. Ab
initio methods[12] and semiempirical calculations[13,14] are val-
uable tools for this aim, but at the expense of increased
computational effort with increasing size of the system, es-
pecially when excited states, solvents or vibrational effects
become important. In this respect, the study and modelling
of size-limited and geometry-controlled molecular assem-
blies is not only important for the intrinsic properties of
these structures, but also is a playground for modelling
supramolecular interactions and testing proposed models. In
fact, these assemblies have the basic ingredients of larger
and more complex structures, with the advantage of being
particularly suitable from the modeling point of view.
Many supramolecular structures show enhanced NLOre-
sponses, but interpretative models have not often been ap-
plied or sufficiently tested. However, several (experimental)
studies on different multichromophore systems showed that
interchromophore interactions can play a significant role. In
particular, increased SHG figures of merit for multichromo-
phore structures have been reported in the literature.[28–31]
For example, donor–p–acceptor-functionalized calix[4]ar-
enes showed up to 2.5-fold amplified figures of merit (per
subchromophore) compared to the corresponding reference
compounds.[32] A striking example of increasing the figure of
merit is a chromophore-functionalized polymer that shows
35-fold response per subchromophore with respect to the
free chromophore.[33] In these cases, the increased response
was mainly due to a high degree of alignment of the func-
tionalising chromophores: the hyperpolarizability of the
supramolecular structures could be interpreted well by the
oriented-gas model, even though, in the case of the polymer,
a major contribution to the global dipole moment is made
by the main polymer chain.[33] In other examples, the re-
sponse of multichromophore assemblies shows small[34] or
more sizeable deviations[35,36] from additive behavior. These
effects were ascribed to dipole–dipole coupling between
subchromophores in functionalized cyclodextrins[34] and den-
drimers,[36–38] and also to possible p–p interactions in func-
tionalized calix[4]arenes.[35,39] Another striking example of
the role of interchromophore interactions is observed for cy-
clophane derivatives in which two dipolar push–pull stil-
benes that are maintained in a forced relative orientation in-
teract via through-space interactions.[40,41] All these studies
point to the importance of structural parameters (geometry,
distance, nature of the chromophores) in multichromophore
structures and call for further analysis.
To address the problem of interaction effects on the prop-
erties of molecular multidipolar assemblies, we studied two
series of homologous compounds, obtained by grafting one,
two or three active dipolar chromophores onto a central
phenyl core. In these model systems in which push–pull
chromophores are assembled via covalent flexible linkers in
close proximity, dipolar interactions between individual sub-
chromophores are expected to influence both the conforma-
tion and the optical properties of the multichromophore
nanoassemblies. From this perspective, chromophores with
large dipole moments were selected. Such flexible multimer
series provide valuable models for the investigation of con-
finement effects on the linear and nonlinear optical (NLO)
responses of push–pull chromophores. Here we report their
synthesis, spectroscopic characterization (absorption, fluo-
rescence, solvatochromism) and experimental determination
of their second-order NLOproperties. Complementary to
previously reported experimental and theoretical studies,
our goal was to investigate how 1) the number of sub-
chromophores, 2) the nature of the dipolar subchromo-
phores (i.e., polarization and polarizability) and 3) geomet-
rical constraints (such as fixed distance between specific
parts of subchromophores) can influence the “supramolec-
ular” effect.
Simple models, based on a few states[15–17] and on the
choice of well-focussed interactions[18] are also of great im-
portance. These are, for example, the basic ingredients of
the most widely used approach for interacting molecules:
the excitonic model.[19] This picture is apparently oversimpli-
fied, but it works well in describing the properties of aggre-
gates of molecules interacting through their transition
dipole moments,[20] and it has also been successfully applied
to model multipolar structures for NLO, in the case of weak
interactions.[21] However, the same model is inadequate for
strongly interacting dipolar entities,[22–24] as are often present
in architectures optimized for second- or third-order NLO.
In some cases the presence of nonelectrostatic intermolecu-
lar interactions (e.g., exchange interactions) is also a source
of deviations from the predictions of the excitonic
model.[25–27]
3090
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
The linear and nonlinear optical properties of the two
series were interpreted by a recently proposed model which
Results
describes the responses of interacting polar molecules start-
ing from the properties of the single chromophore.[18,42] We
demonstrate that interactions between subchromophores are
responsible for their arrangement within the “supramole-
cules” and thus fix the responses of the assemblies. For the
studied series, this is true even though interchromophore
distances are too large to sizeably affect the individual re-
sponses of each constituent subchromophore. We also show,
through simulations that, at shorter distances or for more
polar and/or polarizable subchromophores, the geometry
and optical response of the multichromophore assemblies
are highly and nontrivially affected by subchromophore in-
teractions. This opens the route towards cooperative multi-
chromophore assemblies in which the NLOresponses of the
individual subchromophore components would also be en-
hanced and much more efficient molecular assemblies
would result.
Model systems: Two series of size-controlled multichromo-
phore assemblies were synthesized by functionalising a cen-
tral phenyl core with two or three push-pull chromophores
with the same electron-releasing amino group and by two
different electron-withdrawing moieties:
a nitrophenyl
group (Series I) or a dicyanovinyl moiety (Series II). The
chromophores were grafted to the central core through flex-
ible connectors that act as “passive” spacers between the
active chromophores. This allows electrostatic forces to be
singled out as the source of intermolecular interactions. In
addition, the chosen spacers maintain spatial proximity be-
tween subchromophores and thus allow through-space
dipole–dipole interaction to take place while providing the
required flexibility for orientational freedom.
Three types of dimers were prepared by grafting the two
subchromophores onto the central core in ortho, meta or
para positions. This allows the distance between subchromo-
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3091

M. Blanchard-Desce et al.
phores to be modulated and the effect of interchromophore
distance on the properties of the multichromophoric assem-
blies to be studied. The spectroscopic and second-order
NLOproperties of the dimers and trimers in each series
were studied and compared with the properties of the corre-
sponding monomeric model compounds obtained by attach-
ing a single subchromophore to the same central core
through the same connector. The effects of intermolecular
interactions on the geometry and properties of the assem-
blies were deduced and rationalized.
Synthesis: Graftable chromophoric building blocks 3 and 6b
were synthesized from the same precursor, that is, aldehyde
1 (Scheme 1). The NLO-phore 3 was prepared by Knoeve-
nagel condensation of 1 with malononitrile (2), whereas 6b
was obtained in a three-step sequence involving protection
of the alcohol functionality of 1 as an acetal, Horner–Wads-
worth–Emmons condensation with phosphonate 5 and de-
Scheme 2. Synthesis of the two series of multichromophoric assemblies.
Reaction conditions: NEt3, CH2Cl2, 208C, 4 h, then reflux, 1 h.
Figure 2 compares the spec-
tra of the compounds of series
I (a) and II (b) in CHCl3.
Spectra in the other solvents
are available as Supporting In-
formation. Slight effects on ab-
sorption
and
fluorescence
bands are recognizable, due to
subchromophoric interactions.
In particular, a small blue shift
is observed going from the mo-
nomer to the multimers in the
case of both series, whereas a
decrease of oscillator strength
(per subchromophore) is ob-
served in the case of series II
(see also Table 2). The de-
Scheme 1. Synthesis of graftable chromophoric bricks 3 and 6b. a) 2, EtOH, reflux, 18 h; b) dihydropyran,
PPTS, CH2Cl2, 208C, 15 h; c) 5, NaH, THF, 208C, 16 h; d) HCl, CH2Cl2/EtOH, reflux, 17 h.
protection of the acetal (Scheme 1). By esterifying both al-
cohols (6b and 3) with acyl chlorides 8a–c and 9, two series
of di- and trichromophore assemblies (I and II, respectively)
were obtained. Monochromophore models I-1 and II-1 were
obtained by reaction of 6b and 3 with benzoyl chloride
(Scheme 2).
crease in maximum extinction coefficient is accompanied by
a slight broadening of the spectrum. This (inhomogeneous)
broadening on going from the monomer to the trimer is
most probably linked to the larger conformational disorder
in the multimers.
Fluorescence quantum yields were measured by using flu-
orescein in 0.1n NaOH as reference (fref =0.90).[43] Com-
pounds of series I have significant quantum yields in tolu-
ene, which vary from about 80% for the monomer down to
about 20% for the trimer. In CHCl3 and acetone the quan-
tum yields are very low (about 3 and 1%, respectively) for
all members of the series. Series II has very weak quantum
yields, always lower than 1%.
Linear optical properties: The UV/Vis absorption and fluo-
rescence spectra of I-1, I-2O, I-2P and I-3 were recorded in
toluene, chloroform and acetone. Spectra of II-1, II-2M, II-
2P and II-3 were recorded in toluene, CHCl3 and dimethyl
sulfoxide (DMSO). A fourth solvent (triacetin) was used for
I-1. The solvatochromic absorption and emission properties
of I-1 and II-1 are reported in Figure 1 and Table 1, while
the solvatochromism of dimers and trimers is available as
Supporting Information. The solvatochromic behavior of
each compound in a series is nearly the same, characterized
by a red shift of both absorption and fluorescence bands
with increasing solvent polarity, an effect which is more pro-
nounced for fluorescence spectra.
Dipole moments: Dipole moments m of all the products of
series I and of II-1 were determined from dielectric and re-
fractive index measurements in CHCl3, by means of the
Debye formula.[44] The measured values are reported in
Table 4. Compound II-1 has a larger dipole moment than I-
1, even though the length of the dipole can be estimated to
3092
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
Figure 1. Absorption and fluorescence solvatochromism of the monomer-
Figure 2. Absorption (per subchromophore) and fluorescence spectra in
ic models: a) I-1; b) II-1.
chloroform: a) series I; b) series II.
Table 2. Absorption and fluorescence properties of analogous com-
pounds from series I and II in CHCl3: maximum absorption wavelength
labs, oscillator strength f, emission wavelength lem at the maximum, and
fluorescence quantum yield f.
Table 1. Absorption and fluorescence properties of I-1 and II-1 in differ-
ent solvents: maximum absorption wavelength labs, oscillator strength f,
emission wavelength lem at the maximum, and fluorescence quantum
yield f.
labs [nm]
f
lem [nm]
f
Solvent
labs [nm]
f
lem [nm]
f
I-1
440
438
436
438
433
430
430
430
1.6
3.2
3.3
4.8
2.0
4.0
3.8
5.1
750
745
740
735
465
465
465
470
0.031
0.036
0.025
0.032
0.002
0.002
0.002
0.004
toluene
CHCl3
acetone
toluene
CHCl3
433
440
439
425
433
1.6
1.6
1.7
1.6
2.0
585
750
775
450
465
0.84
I-2O
I-2P
I-3
I-1
0.031
0.015
0.001
0.002
II-1
II-1
II-2M
II-2P
II-3
DMSO440
2.1
490
0.004
be shorter. This suggests a more pronounced charge-transfer
character of II-1 in the ground state, probably due to the
stronger electron-withdrawing character of the dicyanovinyl
group with respect to the nitro group.
Values for series I allow the conclusion that all the multi-
mers have a large dipole moment (from 12 to 15 D), which
indicates a nonrandom spatial organization of the subchro-
mophores grafted to the central core. In particular, by as-
suming a V-shaped configuration for the dimers, and a
calyx-like shape for the trimers (see Figure 3), the angle
formed by each subchromophore with the main symmetry
axis of the assembly can be estimated. Under the assump-
tion that the dipole moment of each subchromophore in the
Table 3. Parameters [eV] of the two-state model (see text) that allow re-
production of the experimental absorption and fluorescence spectra of I-
1 and II-1. The three values for eor correspond to toluene, CHCl3, and
acetone, respectively, for I-1; and to toluene, CHCl3, and DMSO, respec-
tively, for II-1.
p
z0
2t
wv
ev
G
eor
ec
I-1
II-1
1.3
1.0
0.6
1.1
0.12
0.16
0.52
0.21
0.06
0.05
0/0.35/0.40
0.16/0.27/0.43
0.15
–
assembly is not strongly affected by intermolecular interac-
tions, the following angles were calculated: 458 for I-2O, 358
for I-2M, 408 for I-2P and 558 for I-3.
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3093

M. Blanchard-Desce et al.
studied products is typical of transitions involving an excited
charge-transfer state having a larger dipole moment than
the ground state.[48] A rough assessment of the value of Dm
(the difference between the dipole moment of the excited
and ground states) can be derived from b(0) by using the
two-level model, starting from the measured values of the
transition frequency and dipole moment.[46] This estimate
leads to Dm values of 9.1 and 3.6 D for I-1 and II-1, respec-
tively. This also confirms the assumption of a more pro-
nounced charge-transfer character for the ground state of II-
1 with respect to I-1.
The spectra of II-1 can be reproduced by the two-state
model with Holstein coupling to an effective molecular vi-
bration and solvation interaction. The model is detailed in
reference [49] and summarized in the Experimental Section.
The fit allows estimation of the model parameters: the
Figure 3. Schematic representation of the supramolecular structures.
Zigzag lines correspond to the linkers between the central core and the
active subchromophores. Subchromophores are represented by rods bear-
ing electron-donor (D) and electron-acceptor (A) groups. For dimers and
trimers C2v and C3v symmetries are assumed, respectively; a is the angle
formed by each subchromophore with respect to the main symmetry axis
(dashed line).
energy difference between the neutral and the zwitterionic
p
basis states 2z0, the mixing matrix element À 2t, the vibra-
tional frequency wv and the vibrational relaxation energy ev
of the effective coupled mode. To reproduce the spectra,
Gaussian band shapes were chosen with half-width at half-
maximum G. Another parameter is needed to describe the
solvent polarity: the solvent relaxation energy eor, which is
the only parameter that is allowed to change from one sol-
vent to another. Spectra of II-1 can be reproduced by fixing
the parameters listed in Table 3. To quantitatively reproduce
the absorption intensity, the value m0 =20.5 D was fixed for
the dipole moment of the zwitterionic basis state. The fit of
the spectra is shown in Figure 4b. These parameters corre-
spond to a degree of charge transfer 1 in the ground state
ranging from 0.15 in toluene to 0.16 in DMSO.
The model, when allowed to fully account for molecular
polarizability at all orders, also estimates different degrees
of charge transfer for the molecule when slow coordinates
(vibrational and solvation degrees of freedom) are in equi-
librium with the electronic distribution relevant to the
ground or excited state.[49,50] This mechanism of interaction
is typically nonlinear, since the configuration of slow de-
grees of freedom depends on the molecular electronic distri-
bution, which in turn is affected by the configuration of slow
degrees of freedom. This leads to the following degrees of
charge transfer 1 in the Franck–Condon ground state, which
is relevant to the fluorescence process: 0.20, 0.22 and 0.25 in
toluene, CHCl3 and DMSO, respectively. This means that in
Second-order NLO properties: The second-order NLOre-
sponses of all products were measured in CHCl3 by the elec-
tric-field-induced second-harmonic generation (EFISHG)
technique, operating at 1.9 mm.[45] This method, which can
only be used for solutes having a nonvanishing dipole
moment, provides access to the product mb(2w), where b is
the vector component of the b tensor in the direction of the
dipole moment (z), usually called bz. In the following all the
reported and discussed b values correspond to this quantity,
except where explicitly specified. The b(2w) values can be
calculated provided the dipole moments are known. The
corresponding static values b(0) can also be estimated by
using the two-state model[46] [Eq. (1)]
ðwa2bsÀw2Þðw2absÀ4w2Þ
bð0Þ ¼
bð2wÞ
ð1Þ
w4abs
The EFISHG figure of merit mb(0) is then accessible. All
results are reported in Table 4 according to the X conven-
tion as defined in reference [47]. The multichromophore
strategy leads to amplification of the figure of merit by up
to a factor of four for the trimer of series II, with a net gain
with respect to the increased molecular weight. Indeed all
multimers show higher normalized figure of merit mb(0)/M
than the corresponding mono-
mer. Interestingly, the en-
hancement seems to increase
with increasing number of sub-
chromophores in the nanoas-
sembly and with increasing po-
larity of the subchromophores.
Table 4. Dipole moments and hyperpolarizabilities of all compounds of series I and some compounds of series
II. The last two columns report calculated values. Values in parentheses correspond to the ratio between the
response of the relevant multimer and the response of the monomer.
Compound
m
mb(2w)
mb(0)/M
b(2w)
mb(2w)calcd
b(2 w)calcd
[D]
[10À48 esu]
[10À48 esumolgÀ1
]
[10À30 esu]
[10À48 esu]
[10À30 esu]
I-1
8.7
390
0.61
0.62
0.68
0.77
0.75
0.35
0.36
0.52
45
58 (1.3)
56 (1.2)
68 (1.5)
83 (1.8)
19
258
153
I-2O
I-2M
I-2P
I-3
II-1
II-2M
II-3
12.3 (1.4)
14.2 (1.6)
13.2 (1.5)
15.1 (1.7)
9.7
720 (1.8)
790 (2.0)
900 (2.3)
1260 (3.2)
180
488 (1.9)
578 (2.2)
632 (2.4)
733 (2.8)
214
209 (1.4)
228 (1.5)
239 (1.6)
255 (1.7)
67
Discussion
–
–
350 (1.9)
715 (4.0)
–
–
544 (2.5)
831 (3.9)
107 (1.6)
130 (1.9)
Solvatochromism: The solvato-
chromic behavior of all the
3094
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
ter: the conformational relaxation energy ec. The first effect
of this coupling is an increased Stokes shift, because the
equilibrium conformational coordinate is different for the
ground and excited states. In addition, a Boltzmann distribu-
tion of conformations must be accounted for at finite tem-
perature, since the conformational coordinate is character-
ized by very low frequency (typically in the far-infrared/mi-
crowave region) that leads to an additional inhomogeneous
broadening of the absorption and emission bands.
By fixing the parameters in Table 3, the spectra of I-1 in
all solvents can be reproduced.[52] The value of m0 was fixed
to 33 D. The corresponding degrees of charge transfer are
lower than for II-1, on the order of 0.05 in all the solvents.
This is again a confirmation of the weaker electron-with-
drawing power of the nitro group compared to the dicyano-
vinyl group. The solvatochromic effect is also smaller in ab-
sorption and larger in fluorescence in series I than in series
II, and this points to the more neutral nature of the chromo-
phore in series I.
Interchromophore effect: Interchromophore interactions for
the two studied series are purely electrostatic in nature. The
spacers connecting each subchromophore to the central
phenyl core are “passive”, that is, they prevent extension of
conjugation from one subchromophore to another. More-
over, the estimated distance between the grafted subchro-
mophores is large enough to hinder overlap between molec-
ular orbitals of different subchromophores. In addition, in
both cases interactions are weak, as demonstrated by the
small variations of absorption and fluorescence spectra on
going from the monomers to the dimers to the trimer. This
is due to their strongly dominant neutral nature and also to
the molecular architecture. In the chosen multichromophore
structures, chromophores are linked to the central core
through one end, while the other end is completely free, so
that the subchromophores have the possibility to adopt the
relative orientation that minimises repulsive interactions.
This leads to noncentrosymmetrically self-organized struc-
tures, as proven by the large dipole moments measured even
in the case of trimers and para dimers. This intrinsic asym-
metry results in significant second-order nonlinearity, that is,
even in the case of weak interaction, when the individual
optical responses of the subchromophores within the nano-
assembly are not much affected, the dipolar interchromo-
phore interactions play an important role by controlling
their relative orientation and thus the overall architecture of
the nanoassembly.
Figure 4. Calculated absorption and fluorescence solvatochromism of the
monomeric models: a) I-1; b) II-1. Parameters in Table 3 were used.
the geometry relevant to the fluorescence process, the mole-
cule has a more polar and polarizable ground state than in
the geometry relevant to the absorption process.
The fit of the spectra of I-1 is somewhat more difficult.
Spectra in Figure 1a show that the Stokes shift is very large
even in an almost nondipolar solvent like toluene. This sug-
gests the presence of another coupled degree of freedom,
different from a molecular vibration and from the solvation
coordinate. Due to the nonrigid structure of the relevant
chromophore, this coordinate is probably a conformational
degree of freedom, so that fluorescence likely stems from a
so-called twisted intramolecular charge transfer state. This
hypothesis is indirectly confirmed by spectra recorded in tri-
acetin, a very viscous solvent (dot-dashed lines in Fig-
ure 1a): in fact, even though triacetin is more polar than
CHCl3, the Stokes shift in this solvent is smaller than in
chloroform, and this suggests that the high viscosity hinders
to some extent conformational reorganization in the excited
state. Conformational coordinates are different in nature
from molecular vibrations or from solvation degrees of free-
dom: while the last-named coordinates affect diagonal terms
in the Hamiltonian, a conformational coordinate modulates
the mixing matrix element (the charge transfer integral) be-
tween the basis states.[51] This type of coupling was already
modelled to successfully describe the spectra of phenol blue,
another flexible chromophore.[49] Here we use this extension
of the model, by introducing the relevant additional parame-
By using a simple electrostatic model, it is possible to plot
the dependence of the interaction energy on the angle
formed by the subchromophores. The dipole approximation
could be used, but it would not permit cases in which the
chromophores have different lengths (e.g., series I and
series II) to be distinguished. Hence, here we prefer to
assign positive/negative charges to the donor/acceptor ends
and model the interactions by estimating the interchromo-
phore distance (at the donor ends) and the length of each
chromophore (between donor and acceptor groups). We as-
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3095

M. Blanchard-Desce et al.
sumed a V-shaped configuration for dimers (C2v charge sym-
metry) and a calyx-like shape for trimers (C3v charge sym-
metry), as sketched in Figure 3. We fixed the distance be-
tween the donor ends at 13, 15, and 17 Š respectively for
ortho, meta and para positions, and imposed a length of 10
and 8 Š, respectively, for the chromophores in series I and
II. By minimising the electrostatic energy of interaction, we
predict the following angles formed by each subchromo-
phore with respect to the principal symmetry axis of the as-
sembly: 468 for I-2O, 418 for I-2M, 388 for I-2P, 558 for I-3;
408 for II-2O, 368 for II-2M, 338 for II-2P, and 478 for II-3.
The deviation between the two series results from the differ-
ence in chromophore length. The estimated angles compare
well with the angles derived from the experimental dipole
moments of (multi)chromophores of series I.[53] This simple
procedure is already a clear demonstration that interchro-
mophore electrostatic interactions determine the relative
orientation of subchromophores and hence the supramolec-
ular conformation of the assembly.
Figure 5. Calculated absorption spectra (per subchromophore) of series
II (chloroform). Parameters in Table 3 were used. For II-3 inhomogene-
ous broadening has not been taken into account, and a half-width at half-
maximum of 0.11 eV has been imposed.
Note that the strength of the electrostatic interaction that
can be estimated between subchromophores in the assem-
blies is at least one order of magnitude higher than the in-
teraction between each subchromophore and the static elec-
tric field applied for the EFISHG experiment (typically on
the order of 106 VmÀ1). For this reason the geometry of the
multichromophores can be considered as independent of the
applied electric field, and the nanoassemblies can be regard-
ed as independent units that interact with the field itself via
their global dipole moment.
mophore length), so that multichromophore effects are also
weak: subchromophore polarity only varies by 1 and 2%,
respectively, for the trimers of series I and II with respect to
the monomeric analogue. The main characteristics of the
spectra are reproduced, such as the small blue shift of the
band and the small decrease in oscillator strength from the
monomers to the multimers. Experimental spectra of dimers
and trimer are somewhat additionally broadened, but this
can be associated with the higher geometrical disorder for
the multimers. These calculated spectra are confirmation
that the fixed geometrical parameters are reliable.
To further validate the proposed approach, in the follow-
ing we show spectra calculated for dimers and trimers, as
well as the calculated b values for all members of the two
series. The extension of the two-state Mulliken model to de-
scribe interacting chromophores has been recently pro-
posed.[18,54] This approach is different from the standard ex-
citonic model, since it allows the properties of the “supra-
molecule” to be calculated by starting from the parameters
estimated for the isolated (or solvated) chromophore unit
and from the geometry of the assembly. This is due to the
fact that the molecular (hyper)polarizability is fully account-
ed for, so that environmental effects are self-consistently
taken into account. Another noteworthy difference to the
standard excitonic approach is the fact that all electrostatic
interactions are considered, not only those arising from the
coupling of the transition dipole moments; the latter ap-
proximation is strictly valid only for nondipolar molecules.
This model for interacting polar and polarizable chromo-
phores has been further extended to account for molecular
vibrations and solvation,[42,55] to allow comparison with ex-
perimental measurements. Along these lines, the calcula-
tions reported hereafter were performed.
A further test of the chosen parameters and of the pro-
posed interpretation is given by comparison of experimental
and calculated b values. Calculated mb(2w) and b(2w) are
reported in Table 4. The agreement is not quantitative with
respect to the individual values, but the trend within each
series is well reproduced, as shown by the ratios between
the response of the multimers and the response of the corre-
sponding monomer (values in parentheses in Table 4). The
nonquantitative agreement of the individual values is mainly
due to the evaluation of the dipole moment in the ground
and excited states. Both the transition energy and the transi-
tion dipole moment are well reproduced (as demonstrated
by the good agreement between experimental and calculat-
ed spectra). On the contrary, the ground-state dipole mo-
ments fixed by the chosen molecular parameters strongly
underestimate the measured values: 1.6 D and 3.2 D calcu-
lated for I-1 and II-1, versus experimental values of 8.7 and
9.7 D, respectively. This quite large discrepancy is somewhat
surprising, since the calculated (permanent) dipole moments
allow the solvatochromic behavior of all compounds to be
correctly reproduced. A possible reason may be related to
the fact that the dipole moments estimated from the fit of
the spectra are related to the polarizable (and hence respon-
sive) part of the skeleton only, while the measured values
correspond to the total dipole moments. Furthermore, the
contribution of the connecting spacers to the total dipole
moment may be nonnegligible, and their contribution is
Figure 5 shows spectra calculated for series II. For the
multimers, the same parameters as deduced for the mono-
mer were used, with additional geometrical parameters, that
is, interchromophore distances and angles, as obtained by
minimising the interactions. In this case interactions are
weak (interchromophore distances are about twice the chro-
3096
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
more important for smaller dipole moments of the grafted
chromophore (more important for series I than for series
II). The same problem also afflicts the excited state, for
which calculations overestimate the dipole moment. Despite
these discrepancies, we observe good agreement between
the experimental and calculated values of the ratio between
the nonlinear responses of the multimers and of the corre-
sponding monomer (Table 4).
Some reports in the literature suggest that strong interac-
tions between chromophores can lead to large deviations
from simple additive contributions of active chromophores.
The analysis here is a case study, relevant to the limit of
weak interactions; but the adopted model can be exploited
to predict the properties of strongly interacting chromo-
phores, since it fully takes into account the molecular polar-
izability. In fact, dipolar chromophores are highly responsive
to any change in the environment (as demonstrated by their
pronounced and nontrivial solvatochromism), so that the
properties of multichromophore systems are expected to
strongly depend on relative intermolecular orientation and
distance.[8,13,18] The stronger the interactions and the more
polarizable the chromophores, the larger these “environ-
mental” effects will be.[18] Figure 6 shows examples of what
The sizeable increase in the b value from ortho dimer to
meta dimer to para dimer (series I) indicates that the dis-
tance between subchromophores influences the self-arrange-
ment of the dipolar subchromophores within the nanoas-
sembly and thus its second-order optical response. The non-
random spatial self-organization results in an increase in b
response from the monomer to the multimers by a factor on
the order of 1.5–2. This, in combination with the increased
dipole moment in the dimers and trimers, leads to an am-
plification of the figure of merit mb by up to a factor of four.
The amplification is more important for series II than for
series I, because of the smaller angles imposed by interchro-
mophore interactions, which result from the difference in
length of the corresponding monomers. Interestingly, for
both series the trimer shows improved performance as com-
pared to the dimeric analogues. Hence, for the same inter-
chromophore distance (i.e., when comparing trimers with
meta dimers), the multichromophore strategy provides im-
proved second-order NLOresponse with increasing number
of subchromophores in the nanoassembly.
Conclusions and Perspectives
We have addressed the properties of multidipolar structures
by experimental measurements of linear and nonlinear opti-
cal properties, and with a recently proposed theoretical
model for interacting polar and polarizable molecules. Mul-
tichromophore assemblies were obtained by grafting one,
two or three active chromophores onto a central passive
core. In the studied cases, responses of multimers are almost
additive, in the sense that an oriented-gas model can ac-
count for the properties of the assemblies. More important-
ly, we demonstrated that the relative orientation of subchro-
mophores is dictated by intermolecular interactions, so that
interactions are the driving force in determining the proper-
ties of the assembly, even when they are fairly weak. Inter-
branch interactions are thus responsible for nonrandom self-
arrangement of the subchromophores grafted to the central
core, which lead to an increased dipole moment of the
supramolecular assemblies with respect to the monomeric
analogues. The increase in dipole moment has a twofold ad-
vantage. It is responsible for better molecular orientation
and thus a higher value of the global order parameter,
which possibly results in higher bulk susceptibilities. More-
over, at the microscopic level this effect can lead to in-
creased figures of merit mb per subchromophore: in our case
an increase of up to 30% (for compound II-3) was ob-
served.
Figure 6. Angle and distance dependence of the SHG figure of merit, cal-
culated for a dimer (C2v symmetry) of a) C16H33Q-3CNQ molecules in
CHCl3 (parameters estimated in reference [56]) and b) molecules that are
in the cyanine limit when isolated.
is expected for more polar and polarizable interacting chro-
mophores. Figure 6a shows the figure of merit (absolute
value) calculated for a dimer (C2v symmetry) of molecules
that are almost zwitterionic when isolated in solution. Pa-
rameters are relevant to the chromophore C16H33Q-3CNQ,
as reported in reference [56]. In this case, at an intermolecu-
lar distance of about 10 Š and a corresponding equilibrium
intermolecular angle (ca. 428), the response of the dimer is
almost three times the response of the isolated molecule, be-
cause polarization effects sum up to oriented-gas effects in
determining the properties. For this same dimer, a figure of
merit five times larger than that of the monomer is expected
for parallel molecules at an intermolecular distance of 8 Š
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3097

M. Blanchard-Desce et al.
sufficient for describing the basic physics of these molecules. The theoret-
ical approach adopted here is based on the two-state Mulliken model,[57]
extended to account for Holstein coupling to molecular vibrations and
solvation interaction. For a single chromophore in solution the relevant
Hamiltonian reads as Equation (3),[50]
(without polarizability effects, the gain factor could not be
larger than four). Figure 6b is relevant to push–pull chromo-
phores at the cyanine limit when isolated. This means that
their first hyperpolarizability is vanishing. However, due to
intermolecular interactions, large SHG figures of merit can
be obtained. This is a striking example of the importance of
polarizability effects on the properties of interacting sys-
tems: even starting from nonresponsive molecules, highly re-
sponsive assemblies could be achieved. Hence, the same
model molecule, arranged in assemblies of different fashions
(relative distance, orientation, etc.) can behave in quantita-
tively and qualitatively different ways. The ability to correct-
ly take into account all these ingredients in the design and
synthesis of supramolecular structures in a knowledge-based
(or bottom-up) approach will allow geometrical degrees of
freedom to be tailored to maximise the desired responses. In
this respect, we are currently working on the engineering of
multichromophore systems made up of strongly polar and/or
polarizable chromophores, to exploit their self- or pre-ar-
rangement, as well as a possible self-improving mechanism
of the desired property.
ꢀ
ꢁ
X
pffiffiffi
pffiffiffiffiffiffi
1
2
^
^
^
h ¼ 2 z01À 2tsx þ
ðwi2Qi2 þ Pi2ÞÀ 2eiwiQi1
ð3Þ
i
where z0 and t are the parameters of the two-state model: 2z0 is the
energy difference between the neutral and zwitterionic basis states, and
pffiffiffi
ˆ
À
2t the mixing matrix element (charge-transfer integral); 1 is the oper-
ˆ
ator that counts electrons on the acceptor site; and sx is one of the Pauli
ˆ
spin operators. The expectation value 1=h1i measures the weight of the
zwitterionic structure in the ground state, and hence represents the mo-
lecular polarity. In fact, following Mulliken, a dipole moment m0 can be
associated to the zwitterionic basis state, neglecting all other contribu-
tions. In this way, the ground-state dipole moment is given by m01, and
the excited state dipole moment by m0(1À1). Transition energy and
pffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
dipole moments are respectively given by ꢀhwCT
m0 1ð1À1Þ.
=
2t/ 1ð1À1Þ and mCT
=
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
In the above Hamiltonian the summation runs over coupled modes (of
coordinate Qi, momentum Pi and frequency wi), that is, coordinates along
which the potential energy surfaces relevant to the two basis states are
displaced. The parameter ei measures the relaxation energy of the rele-
vant mode. Coupled modes can be molecular vibrations, or also a solva-
tion coordinate (in the case of polar solvents). Here we reproduce experi-
mental spectra by use of a single coupled molecular vibration (of fre-
quency wv and relaxation energy ev) and a solvation coordinate (of relax-
ation energy is eor). The frequency of the solvation degree of freedom
(which is very low, typically in the far-infrared/microwave region) need
not be fixed, since this mode can be treated as a classical coordinate.[58]
At finite temperature a Boltzmann distribution along this coordinate
must be accounted for, which is the origin of inhomogeneous broadening
of spectra in polar solvents.[49, 58] Along the vibrational coordinate, the po-
tential energy surfaces relevant to the two basis states are chosen as har-
monic and with the same frequency. However, as a result of diagonaliza-
tion, potential energy surfaces are no longer harmonic for the ground
and excited states, and they are softened and hardened, respectively.[58] In
the adiabatic approximation, the electronic problem corresponds again to
a two-state problem, but with the parameter z0 self-consistently renor-
malized: z0!z0À(ev+eor)1 if slow degrees of freedom are in equilibrium
for the ground state, and z0!z0À(ev+eor)(1À1*) if slow degrees of free-
dom are in equilibrium for the excited state.[50, 58] The self-consistency cor-
responds to the system (hyper)polarizability. One of the consequences of
the electronic polarizability is just the different degree of charge transfer
1 in the case of different geometries of slow variables: a ground-state po-
larity 1 relevant to the absorption process and a ground state polarity 1*
relevant to the fluorescence process. This allows nonspecular absorption
and fluorescence spectra to be accounted for.[49,50]
Experimental Section
Spectroscopic measurements: UV/Vis spectra were recorded on a Jasco
V-570 spectrophotometer on about 10À5 m solutions. Fluorescence meas-
urements were performed on dilute solutions (ca. 10À6 m) with an Edin-
burgh Instruments (FLS 920) spectrometer in photon-counting mode. For
series II, Raman emission lines of the solvent were subtracted from the
emission signal. Fluorescence quantum yields were measured by using
fluorescein in 0.1n NaOH as standard (quantum yield f=0.90).[43] Sol-
vents were of spectroscopic grade. All measurements were carried out at
room temperature.
Dipole moment measurements: Permanent dipole moments were deter-
mined by means of a WTW dipole meter (type DM01) with a capaci-
tance cell for liquids and a capacitance-measuring electronic bridge.
NLO measurements: Measurements of
b were performed by the
EFISHG technique.[45] The EFISHG experiment allows determination of
the mean microscopic hyperpolarizability g0 [Eq. (2)].
g0 ¼ gðÀ2 w; w, w, 0Þ þ mbðÀ2w; w, wÞ=ð5 kTÞ
ð2Þ
The first term is the scalar part of the cubic hyperpolarizability tensor,
whereas the second originates from the partial orientation of the perma-
nent dipole moment m in the static field. The orientational contribution is
usually assumed to be the predominant component in the case of polar
charge-transfer molecules. The product mb(2w) is thus directly inferred.
EFISHG measurements were conducted with a Q-switched Nd3+:YAG
laser emitting pulse trains and operating with the first Stokes radiation,
at 1.907 mm, of the YAG 1.064 mm emission generated in a hydrogen
Raman cell. These experiments were performed by using, for each mole-
cule, solutions of increasing concentration in chloroform. Measurements
were calibrated relative to the pure solvent. The experimental accuracy
does not exceed 5%. The reported b values follow the X convention as
defined in reference [47].
The vibrational problem is solved exactly (for the ground and excited
electronic states) on the basis of a harmonic oscillator of frequency wv
centered at the relevant equilibrium position.[59] The vibrational matrix is
diagonalized on a truncated basis sufficient to achieve convergence.[60] As
for semiempirical models, all parameters are fixed so as to best repro-
duce experimental spectral features. Spectra were calculated by associat-
ing to each vibronic transition a Gaussian lineshape.
For flexible chromophores another degree of freedom describing confor-
mational motion must be accounted for. Such a mode tunes the charge-
pffiffiffi
transfer integral À 2t, and is characterized by a very low frequency (typ-
ically in the microwave region). The electronic Hamiltonian taking into
account this mode reads as Equation (4),[49]
Theoretical model and calculations: Push–pull chromophores are mole-
cules made up of an electron-donor and an electron-acceptor group con-
nected by a p-conjugated bridge. They are characterized by a low-energy
transition (in the visible region) with high intensity, linked to charge
transfer from the donor to the acceptor moiety. For this reason they are
usually and successfully described as resonating between two basis states:
the neutral and the zwitterionic structure, so that a two-level model is
D2
pffiffiffi
ð4Þ
^
^
h ¼ 2 z01À 2ðt þ DÞsx þ
ec
where D is the conformational coordinate measuring the deviation of t
from its reference value, and ec the corresponding relaxation energy; the
3098
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
simple hypothesis that a harmonic potential is associated to this mode
has been made. Although different in nature, the conformational coordi-
nate can be treated along the same lines as the solvation degree of free-
dom, that is, in the adiabatic approximation and in the classical limit. The
effects of this Peierls-like coupling (increased Stokes shift and additional
inhomogeneous broadening) are discussed in the text.
Melting points were determined on an Electrothermal IA9300 digital
melting point instrument. NMR: Bruker ARX 200 (1H: 200.13 MHz, 13C:
50.32 MHz) or Avance AV 300 (1H: 300.13 MHz, 13C: 75.48 MHz), in
CDCl3 solutions; 1H chemical shifts d are given in ppm relative to TMS
as internal standard, J values in Hz and 13C chemical shifts relative to the
central peak of CDCl3 at 77.0 ppm. High-resolution mass spectra were re-
corded at the Centre RØgional de Mesures Physiques de l’Ouest
(C.R.M.P.O., Rennes), using a Micromass MS/MS ZABSpec TOF instru-
ment with EBE TOF geometry; liquid secondary ion mass spectrometry
(LSIMS) was performed at 8 kV with Cs+ in m-nitrobenzyl alcohol
(mNBA).
Intermolecular interactions in “dense” systems can be inserted into the
model by using the molecular Hamiltonian for each chromophoric entity
and adding the interaction term. The form of the Hamiltonian is then
Equation (5),[18,54]
X
X
X
1
2
13C NMR spectra are available as Supporting Information.
^ ^
Vij1i1j
H ¼
hi þ
ð5Þ
4-[Hexyl(2-hydroxyethyl)amino]benzaldehyde (1) was prepared from
hexylaniline, analogously to reference [64]; 1H NMR (200.13 MHz,
CDCl3): d=9.69 (s, 1H), 7.69 and 6.72 (AA’XX’, JA,X =9.0 Hz, 4H), 3.85
(dt, J=5.8 Hz, 2H), 3.59 (t, J=5.8 Hz, 2H), 3.42 (t, J=7.8 Hz, 2H), 1.85
(t, J=5.8 Hz, 1H), 1.62 (m, 2H), 1.37–1.29 (m, 6H), 0.90 ppm (t, J=
6.5 Hz, 3H); 13C NMR (50.32 MHz, CDCl3): d=190.2, 152.9, 132.2, 124.8,
111.0, 59.7, 52.7, 51.6, 31.5, 26.7, 26.6, 22.6, 13.9 ppm; HRMS (LSIMS+,
i
i
j
6
where Vij measures the interaction between zwitterionic species at posi-
tions i and j, and the two summations run over chromophores in the as-
sembly. This Hamiltonian is fairly general: throughout this paper we
single out electrostatic forces as the only source of interaction, and
choose to assign to each chromophore a fixed length and +/Àe charges
to the donor/acceptor ends for the zwitterionic basis states. Electrostatic
interactions are thus inserted as interactions between charges rather than
between dipoles: this allows the length of the chromophores to be taken
into account and interactions between spatially close molecules to be de-
scribed. Since experimental measurements on multichromophore assem-
blies were made in solution, the interaction terms Vij are screened by the
squared refractive index at optical frequencies[61] (n2 =2.1 for CHCl3).
Here dimeric and trimeric multichromophores are considered with sym-
metries of C2v and C3v, respectively. The length of the chromophore, the
interchromophore distance and the angle formed with the main axis of
symmetry determine the values of Vij. A few predictions on assemblies
made up with different chromophores are also presented.
mNBA): m/z calcd for C15H23NO2 [M +]: 249.1728; found: 249.1724.
C
2-{[4-[Hexyl(2-hydroxyethyl)amino]phenylmethylene}propanedinitrile (3).
A
solution of 1 (0.500 g, 2.0 mmol) and malononitrile (2) (0.132 g,
2.0 mmol) in anhydrous ethanol (40 mL) was refluxed for 18 h. After
cooling, the solvent was evaporated under reduced pressure and the
crude product was purified by column chromatography (CH2Cl2/Et2O95/
5) to yield 3 (0.490 g; 82%); m.p. 112–1138C; 1H NMR (200.13 MHz,
CDCl3): d=7.80 and 6.72 (AA’XX’, JA,X =9.3 Hz, 4H), 7.45 (s, 1H), 3.88
(dt, J=5.6 Hz, 2H), 3.62 (t, J=5.6 Hz, 2H), 3.46 (t, J=7.8 Hz, 2H), 1.64
(m, 2H), 1.55 (t, J=5.6 Hz, 1H), 1.33 (m, 6H), 0.91 ppm (t, J=6.4 Hz,
3H); 13C NMR (50.32 MHz, CDCl3): d=157.7, 153.1, 134.0, 119.1, 116.0,
114.9, 111.7, 71.1, 59.8, 52.6, 51.8, 31.5, 26.8, 26.5, 22.5, 13.9 ppm; HRMS
+
C
The Hamiltonian is written on the basis obtained by the direct product of
the 2N electronic states (N=number of chromophores) multiplied the vi-
brational states. The dimension of the problem is (2M)N, where M is the
number of vibrational states introduced for each chromophore.[18] Eigen-
values and eigenvectors obtained by diagonalization were used to calcu-
late spectra of multimers of series II, obtained with M=8 (large enough
to give converged results). Calculated spectra of multimers of series I are
not presented, since the large M value needed in this case (M=18) leads
to a too large a numerical problem.
(LSIMS , mNBA): m/z calcd for C18H23N3O [M +]: 297.1841; found:
297.1840.
4-{Hexyl[2-(tetrahydro-2H-pyran-2-yloxy)ethyl]amino}benzaldehyde (4):
Pyridinium p-toluenesulfonate (PPTS, 0.579 g, 2.3 mmol) was added to a
solution of
3 (2.85 g, 11.4 mmol) and 3,4-dihydro-2H-pyran (DHP,
2.10 mL, 23.0 mmol) in anhydrous CH2Cl2 (56 mL). The mixture was stir-
red at 208C for 15 h. Water was added, and the organic layer separated
and dried (Na2SO4). After evaporation of the solvent, the crude product
was purified by column chromatography (CH2Cl2/AcOEt, gradient from
100/0 to 90/10) to yield 4 (3.78 g; 99%); 1H NMR (200.13 MHz, CDCl3):
d=9.71 (s, 1H), 7.70 and 6.71 (AA’XX’, JA,X =9.1 Hz, 4H), 4.58 (m, 1H),
3.94–3.36 (m, 8H), 1.82–1.42 (m, 8H), 1.33 (m, 6H), 0.90 ppm (t, J=
6.6 Hz, 3H); 13C NMR (50.32 MHz, CDCl3): d=189.8, 152.6, 132.0, 124.8,
110.8, 99.0, 64.5, 62.1, 51.4, 50.5, 31.5, 30.4, 26.7, 26.5, 25.2, 22.5, 19.3,
The values of the first static hyperpolarizability b are obtained through
sum-over-states expressions at the experimental frequency (2w=1.3 eV).
In particular, the vector component of the b tensor in the direction of the
dipole moment (z) is calculated [Eq. (6)],
13.9 ppm; HRMS (LSIMS+, mNBA): m/z calcd for C20H31NO3 [M +]:
333.2304; found: 333.2306.
X
C
1
3
bz
¼
ðbzii þ bizi þ biizÞ
ð6Þ
i
N-Hexyl-4-[(1E)-2-(4-nitrophenyl)ethenyl]-N-{2-[(tetrahydro-2H-pyran-2-
yl)oxy]ethyl}aminobenzene (6a): NaH (0.188 g, 60% dispersion in miner-
al oil) was added to a solution of 4 (0.981 g, 2.94 mmol) and diethyl (4-ni-
trobenzyl)phosphonate (5, 0.885 g, 3.24 mmol) in dry THF (25 mL). The
mixture was stirred at 208C for 16 h. After addition of water (50 mL),
the organic layer was separated and dried (Na2SO4). After evaporation
of the solvent, the crude product was purified by column chromatography
where i runs over Cartesian coordinates (x, y, z). If Kleinman symmetry
(valid out of resonance) is imposed,[62] this expression reduces to bz =
ꢁibzii. Kleinman symmetry was exploited in the present calculations after
checking its validity in the relevant regime. The symmetry of the assem-
blies is also exploited, so that the following expressions are needed
[Eq. (7) and (8)]
1
(CH2Cl2) to yield 6a (0.794 g; 60%); H NMR (200.13 MHz, CDCl3): d=
bz ¼ bzzz þ bzyy for C2v þ Kleinman symmetry
bz ¼ bzzz þ 2bzyy for C3v þ Kleinman symmetry
ð7Þ
ð8Þ
8.17 and 7.55 (AA’XX’, JA,X =8.8 Hz, 4H), 7.41 and 6.69 (AA’XX’, JA,X =
8.8 Hz, 4H), 7.19 (d, J=16.2 Hz, 1H), 6.89 (d, J=16.2 Hz, 1H), 4.60 (m,
1H), 3.95–3.78 (m, 2H), 3.66–3.43 (m, 4H), 3.37 (t, J=7.7 Hz, 2H), 1.88–
1.48 (m, 8H), 1.38–1.28 (m, 6H), 0.90 ppm (t, J=6.5 Hz, 3H); 13C NMR
(50.32 MHz, CDCl3): d=148.6, 145.6, 145.0, 133.6, 128.5, 125.9, 124.1,
123.5, 121.0, 111.6, 99.1, 64.8, 62.2, 51.4, 50.5, 31.6, 30.6, 27.0, 26.7, 25.3,
22.6, 19.4, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
M=8 molecular vibrations per molecular site were used for calculations
in both series.[63] Throughout this paper b corresponds to the quantity bz.
Moreover, the X convention, as defined in ref. [47], is adopted.
C27H36N2O4 [M +]: 452.2675; found: 452.2662.
C
Synthesis
2-[Hexyl{4-[(1E)-2-(4-nitrophenyl)ethenyl]phenyl}amino]ethanol (6b): 1m
HCl (0.8 mL) was added to a solution of 6a (0.710 g, 1.57 mmol) in
CH2Cl2/EtOH (12 mL, 1/1). The mixture was refluxed for 17 h. After
cooling, water (24 mL) was added, the mixture was made slightly basic
with aqueous NaHCO3 and extracted with CH2Cl2. The combined organic
General methods: All reactions involving air- or water-sensitive com-
pounds were carried out under argon. Solvents were generally dried and
distilled prior to use. Reactions were monitored by TLC on Merck 60
F254 aluminum sheets precoated with silica gel. Column chromatography:
Merck silica gel Si 60 (40–63 mm, 230–400 mesh), unless otherwise noted.
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3099

M. Blanchard-Desce et al.
layers were dried (Na2SO4), the solvent evaporated and the crude prod-
uct purified by column chromatography (CH2Cl2/AcOEt, gradient from
100/0 to 97/3) to yield 6b (0.569 g; 98%); m.p. 111–1128C; 1H NMR
(200.13 MHz, CDCl3): d=8.17 and 7.55 (AA’XX’, JA,X =8.8 Hz, 4H), 7.41
and 6.72 (AA’XX’, JA,X =9.0 Hz, 4H), 7.19 (d, J=16.3 Hz, 1H), 6.90 (d,
J=16.3 Hz, 1H), 3.83 (dt, J=5.9 Hz, 2H), 3.54 (t, J=5.9 Hz, 2H), 3.37
(t, J=7.7 Hz, 2H), 1.64 (t, J=5.9 Hz, 1H), 1.61 (m, 2H), 1.32 (m, 6H),
0.90 ppm (t, J=6.6 Hz, 3H); 13C NMR (50.32 MHz, CDCl3): d=
148.8,145.8, 144.9, 133.4, 128.5, 126.0, 124.2, 124.1, 121.5, 112.2, 60.1, 52.9,
51.6, 31.6, 26.8, 26.7, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z
27.1, 26.7, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
C52H58N4O8 [M +]: 866.4255; found: 866.4264.
C
Tris[2-{hexyl[4-{(1E)-2-(4-nitrophenyl)ethenyl}phenyl]amino}ethyl] 1,3,5-
benzenetricarboxylic ester (I-3): Reaction of 1,3,5-benzenetricarbonyl tri-
chloride (9, 9.9 mg, 37.3 mmol) with 6b (62 mg, 168.3 mmol), as described
for I-2O, afforded I-3 (39.0 mg; 83%); 1H NMR (200.13 MHz, CDCl3):
d=8.81 (s, 3H), 8.15 and 7.52 (AA’XX’, JA,X =8.8 Hz, 12H), 7.43 and
6.76 (AA’XX’, JA,X =8.9 Hz, 12H), 7.16 (d, J=16.3 Hz, 3H), 6.88 (d, J=
16.3 Hz, 3H), 4.54 (t, J=6.2 Hz, 6H), 3.76 (t, J=6.2 Hz, 6H), 3.39 (t, J=
7.4 Hz, 6H), 1.62 (m, 6H), 1.35–1.27 (m, 18H), 0.89 ppm (t, J=6.4 Hz,
9H); 13C NMR (50.32 MHz, CDCl3): d=164.7, 148.2, 145.8, 144.8, 134.7,
133.3, 131.0, 128.6, 126.0, 124.4, 124.1, 121.6, 111.9, 62.6, 51.3, 49.0, 31.6,
27.1, 26.7, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
calcd for C22H28N2O3 [M +]: 368.2100; found: 368.2100.
C
2-{Hexyl[4-{(1E)-2-(4-nitrophenyl)ethenyl}phenyl]amino}ethyl
benzoic
ester (I-1): Et3N (57 mL) and benzoyl chloride (7) (28 mL, 0.241 mmol)
were added to a solution of 6b (60 mg, 0.163 mmol) in dry CH2Cl2
(0.75 mL, stabilized with amylene) under argon. The mixture was stirred
at 208C for 4 h and then refluxed for 1 h. Water was added and the mix-
ture was extracted with CH2Cl2. The extracts were dried (Na2SO4), the
solvent evaporated and the crude product purified by column chromatog-
raphy (CH2Cl2) to yield I-1 (69.5 mg; 90%); 1H NMR (200.13 MHz,
CDCl3): d=8.17 and 7.55 (AA’XX’, JA,X =9.0 Hz, 4H), 8.05–8.00 (m,
2H), 7.62–7.53 (m, 1H), 7.49–7.40 (m, 2H), 7.44 and 6.77 (AA’XX’,
C75H84N6O12 [M +]: 1260.6147; found: 1260.6142.
C
2-{[4-(2,2-dicyanoethenyl)phenyl]hexylamino}ethyl benzoic ester (II-1):
Reaction of 3 (57 mg, 0.192 mmol) with benzoyl chloride (7, 54 mg,
0.384 mmol), as described for I-1, with subsequent purification by column
chromatography (heptane/CH2Cl2 60/40 then 40/60), afforded II-1
(75.6 mg; 98%); m.p. 72–738C; 1H NMR (200.13 MHz, CDCl3): d=7.99
(m, 2H), 7.82 and 6.79 (AA’XX’, JA,X =9.4 Hz, 4H), 7.63–7.55 (m, 1H),
7.46 (s, 1H), 7.48–7.41 (m, 2H), 4.52 (t, J=6.3 Hz, 2H), 3.83 (t, J=
6.3 Hz, 2H), 3.48 (t, J=7.7 Hz, 2H), 1.66 (m, 2H), 1.36–1.29 (m, 6H),
0.90 ppm (m, 3H); 13C NMR (50.32 MHz, CDCl3): d=166.2, 157.8, 152.6,
133.8, 133.2, 129.5, 129.3, 128.4, 119.4, 115.8, 114.7, 111.7, 72.0, 61.3, 51.4,
49.1, 31.4, 26.9, 26.4, 22.4, 13.9 ppm; HRMS (LSIMS+, mNBA): m/z
JA,X =8.9 Hz, 4H), 7.20 (d, J=16.2 Hz, 1H), 6.91 (d, J=16.2 Hz, 1H),
4.50 (t, J=6.3 Hz, 2H), 3.76 (t, J=6.3 Hz, 2H), 3.41 (t, J=7.7 Hz, 2H),
1.65 (m, 2H), 1.35–1.30 (m, 6H), 0.90 ppm (t, J=6.4 Hz, 3H); 13C NMR
(50.32 MHz, CDCl3): d=166.5, 148.3, 145.8, 144.9, 133.5, 133.1, 129.8,
129.6, 128.6, 128.4, 126.0, 124.2, 124.1, 121.5, 111.9, 61.9, 51.4, 49.3, 31.6,
27.2, 26.7, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
calcd for C25H27N3O2 [M +]: 401.2103; found: 401.2093.
C
Bis[2-{[4-(2,2-dicyanoethenyl)phenyl]hexylamino}ethyl] 1,2-benzenedicar-
boxylic ester (II-2O): Reaction of phthaloyl chloride (8a, 13.0 mg,
64 mmol) with 3 (58 mg, 195 mmol), as described for I-2O, with subse-
quent purification by column chromatography (heptane/CH2Cl2 30/70),
afforded II-2O (37.9 mg; 82%); m.p. 76–778C; 1H NMR (200.13 MHz,
C29H32N2O4 [M +]: 472.2362; found: 472.2359.
C
Bis[2-{hexyl[4-{(1E)-2-(4-nitrophenyl)ethenyl}phenyl]amino}ethyl]
1,2-
benzenedicarboxylic ester (I-2O): Et3N (51 mL) and 6b (80 mg, 217 mmol)
were added to a solution of phthaloyl chloride (8a, 16.7 mg, 82.3 mmol)
in dry CH2Cl2 (1.0 mL, stabilized with amylene) under argon. The mix-
ture was stirred at 208C for 4 h and then refluxed for 1 h. Water was
added and the mixture was extracted with CH2Cl2. The extracts were
dried (Na2SO4), the solvent evaporated and the crude product purified
by column chromatography (CH2Cl2 then CH2Cl2/AcOEt 95/5) to yield I-
2O (60.4 mg; 85%); 1H NMR (200.13 MHz, CDCl3): d=8.16 and 7.52
(AA’XX’, JA,X =8.9 Hz, 8H), 7.68 and 7.55 (AA’XX’, 4H), 7.39 and 6.71
CDCl3): d=7.78 and 6.75 (AA’XX’,
JA,X =9.1 Hz, 8H), 7.67–7.53
(AA’XX’, 4H), 7.44 (s, 2H), 4.46 (t, J=6.2 Hz, 4H), 3.79 (t, J=6.2 Hz,
4H), 3.43 (t, J=7.8 Hz, 4H), 1.62 (m, 4H), 1.35–1.25 (m, 12H), 0.89 ppm
(t, J=6.2 Hz, 6H); 13C NMR (50.32 MHz, CDCl3): d=167.2, 157.8, 152.5,
133.9, 131.5, 131.4, 128.9, 119.6, 115.7, 114.7, 111.8, 72.7, 62.0, 51.5, 49.0,
31.5, 27.0, 26.5, 22.6, 14.0; HRMS (LSIMS+, mNBA): m/z calcd for
C44H48N6O4 [M +]: 724.3737; found: 724.3728.
C
(AA’XX’,
JA,X =8.8 Hz, 8H), 7.15 (d, J=16.3 Hz, 2H), 6.86 (d, J=
Bis[2-{[4-(2,2-dicyanoethenyl)phenyl]hexylamino}ethyl] 1,3-benzenedicar-
boxylic ester (II-2M): Reaction of isophthaloyl chloride (8b, 13.0 mg,
64 mmol) with 3 (58 mg, 195 mmol), as described for I-2O, with subse-
quent purification by column chromatography (heptane/CH2Cl2 30/70),
afforded II-2M (43.2 mg; 93%); m.p. 115–1168C; H NMR (200.13 MHz,
CDCl3): d=8.57 (s, 1H), 8.15 (d, J=7.8 Hz, 2H), 7.80 and 6.79 (AA’XX’,
16.3 Hz, 2H), 4.43 (t, J=6.4 Hz, 4H), 3.70 (t, J=6.4 Hz, 4H), 3.34 (t, J=
7.6 Hz, 4H), 1.59 (m, 4H), 1.31–1.24 (m, 12H), 0.87 ppm (t, J=6.6 Hz,
6H); 13C NMR (75.48 MHz, CDCl3): d=167.4, 148.2, 145.8, 144.8, 133.4,
131.7, 131.3, 128.9, 128.6, 126.0, 124.2, 124.1, 121.5, 111.8, 62.4, 51.4, 49.0,
31.6, 27.1, 26.6, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
1
C52H58N4O8 [M +]: 866.4255; found: 866.4261.
C
JAX =9.1 Hz, 8H), 7.51 (t, J=7.8 Hz, 1H), 7.44 (s, 2H), 4.53 (t, J=6.1 Hz,
4H), 3.83 (t, J=6.1 Hz, 4H), 3.45 (t, J=7.9 Hz, 4H), 1.65 (m, 4H), 1.31
(m, 12H), 0.88 ppm (t, J=6.7 Hz, 6H); 13C NMR (50.32 MHz, CDCl3):
d=165.3, 157.8, 152.7, 134.0, 133.9, 130.8, 130.0, 128.8, 119.5, 115.7, 114.7,
111.8, 72.3, 61.8, 51.4, 49.0, 31.4, 27.0, 26.5, 22.5, 13.9 ppm; HRMS
Bis[2-{hexyl[4-{(1E)-2-(4-nitrophenyl)ethenyl}phenyl]amino}ethyl]
1,3-
benzenedicarboxylic ester (I-2M): Reaction of isophthaloyl chloride (8b)
(7.6 mg, 37.4 mmol) with 6b (40 mg, 108.6 mmol), as described for I-2O,
afforded I-2M (30.7 mg; 95%); 1H NMR (200.13 MHz, CDCl3): d=8.67
(t, J=1.7 Hz, 1H), 8.20 (dd, J=7.5 Hz, 1.7 Hz, 2H), 8.17 and 7.54
(AA’XX’, JA,X =8.9 Hz, 8H), 7.53 (t, J=7.5 Hz, 1H), 7.44 and 6.76
(LSIMS+, mNBA): m/z calcd for C44H48N6O4 [M +]: 724.3737; found:
C
724.3738.
(AA’XX’,
JA,X =8.8 Hz, 8H), 7.19 (d, J=16.2 Hz, 2H), 6.90 (d, J=
Bis[2-{[4-(2,2-dicyanoethenyl)phenyl]hexylamino}ethyl] 1,4-benzenedicar-
boxylic ester (II-2P): Reaction of terephthaloyl chloride (8c, 13.0 mg,
64 mmol) with 3 (58 mg, 195 mmol), as described for I-2O, with subse-
quent purification by column chromatography (heptane/CH2Cl2 30/70),
afforded II-2P (41.1 mg; 89%); m.p. 155–1568C; 1H NMR (200.13 MHz,
CDCl3): d=8.02 (s, 4H), 7.80 and 6.80 (AA’XX’, JAX =8.9 Hz, 8H), 7.45
(s, 2H), 4.55 (t, J=5.9 Hz, 4H), 3.85 (t, J=5.9 Hz, 4H), 3.47 (t, J=
7.5 Hz, 4H), 1.66 (m, 4H), 1.33 (m, 12H), 0.90 ppm (m, 6H); 13C NMR
(50.32 MHz, CDCl3): d=165.4, 157.8, 152.7, 133.9, 133.5, 129.6, 119.6,
115.7, 114.7, 111.9, 72.6, 61.9, 51.4, 49.1, 31.5, 27.0, 26.5, 22.5, 13.9 ppm;
16.2 Hz, 2H), 4.52 (t, J=6.4 Hz, 4H), 3.76 (t, J=6.4 Hz, 4H), 3.40 (t, J=
7.6 Hz, 4H), 1.63 (m, 4H), 1.35–1.27 (m, 12H), 0.89 ppm (t, J=6.3 Hz,
6H); 13C NMR (50.32 MHz, CDCl3): d=165.6, 148.3, 145.9, 144.9, 134.0,
133.4, 130.8, 130.4, 128.7, 128.6, 126.0, 124.3, 124.1, 121.6, 111.9, 62.3,
51.4, 49.2, 31.7, 27.2, 26.7, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z
calcd for C52H58N4O8 [M +]: 866.4255; found: 866.4245.
C
Bis[2-{hexyl[4-{(1E)-2-(4-nitrophenyl)ethenyl}phenyl]amino}ethyl]
1,4-
benzenedicarboxylic ester (I-2P): Reaction of terephthaloyl chloride (8c,
14.4 mg, 70.9 mmol) with 6b (80 mg, 217 mmol), as described for I-2O, af-
forded I-2P (56.9 mg; 93%); 1H NMR (200.13 MHz, CDCl3): d=8.18
HRMS (LSIMS+, mNBA): m/z calcd for C44H48N6O4 [M +]: 724.3737;
C
found: 724.3732.
and 7.56 (AA’XX’,
JAX =8.9 Hz, 8H), 8.06 (s, 4H), 7.43 and 6.75
(AA’XX’, AX =8.9 Hz, 8H), 7.20 (d, J=16.3 Hz, 2H), 6.91 (d, J=
J
Tris[2-{[4-(2,2-dicyanoethenyl)phenyl]hexylamino}ethyl] 1,3,5-benzenetri-
carboxylic ester (II-3): Reaction of 1,3,5-benzenetricarbonyl trichloride
(9, 17.0 mg, 64 mmol) with 3 (87.9 mg, 296 mmol), as described for I-2O,
with subsequent purification by column chromatography (heptane/
CH2Cl2 30/70), afforded II-3 (23.4 mg; 35%); m.p. 156–1578C; 1H NMR
16.3 Hz, 2H), 4.51 (t, J=6.1 Hz, 4H), 3.76 (t, J=6.1 Hz, 4H), 3.39 (t, J=
7.6 Hz, 4H), 1.63 (m, 4H), 1.35–1.25 (m, 12H), 0.89 ppm (t, J=6.4 Hz,
6H); 13C NMR (50.32 MHz, CDCl3): d=165.6, 148.3, 145.8, 144.9, 133.7,
133.4, 129.6, 128.6, 126.0, 124.3, 124.1, 121.6, 111.9, 62.4, 51.3, 49.2, 31.6,
3100
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102

Effects of Dipolar Interactions on Optical Properties of Multichromophore Assemblies
FULL PAPER
(200.13 MHz, CDCl3): d=8.69 (s, 3H), 7.80 and 6.81 (AA’XX’, JA,X
=
[25] A. L. Thompson, K. M. Gaab, J. Xu, C. J. Bardeen, T. J. Martínez, J.
Phys. Chem. A 2004, 108, 671–682.
[26] S. Tretiak, S. Mukamel, Chem. Rev. 2002, 102, 3171–3212.
[27] C. Katan, F. Terenziani, O. Mongin, M. H. V. Werts, L. Porr›s, T.
Pons, J. Mertz, S. Tretiak, M. Blanchard-Desce, J. Phys. Chem. A
2005, 109, 3024–3037.
[28] E. Hendrickx, C. Boutton, K. Clays, A. Persoons, S. van Es, T. Bie-
mans, B. Meijer, Chem. Phys. Lett. 1997, 270, 241–244.
[29] H.-J. Deussen, E. Hendrickx, C. Boutton, D. Krog, K. Clays, K.
Bechgaard, A. Persoons, T. Bjørnholm, J. Am. Chem. Soc. 1996, 118,
6841–6852.
[30] M. Kauranen, T. Verbiest, C. Boutton, M. N. Teerenstra, K. Clays,
A. J. Schouten, R. J. M. Nolte, A. Persoons, Science 1995, 270, 966–
969.
[31] P. J. A. Kenis, O. F. J. Noordman, S. Houbrechts, G. J. van Hummel,
S. Karkema, F. C. J. M. van Veggel, K. Clays, J. F. J. Engbersen, A.
Persoons, N. F. van Hulst, D. N. Reinhoudt, J. Am. Chem. Soc. 1998,
120, 7875–7883.
[32] P. J. A. Kenis, E. G. Kerver, B. H. M. Snellink-Ruꢁl, G. van Hummel,
S. Harkema, M. C. Flipse, R. H. Woudenberg, J. F. J. Engbersen,
D. N. Reinhoudt, Eur. J. Org. Chem. 1998, 1089–1098.
[33] T. Verbiest, C. Samyn, C. Boutton, S. Houbrechts, M. Kauranen, A.
Persoons, Adv. Mater. 1996, 8, 756–759.
9.2 Hz, 12H), 7.44 (s, 3H), 4.58 (t, J=5.9 Hz, 6H), 3.87 (t, J=5.9 Hz,
6H), 3.47 (t, J=7.8 Hz, 6H), 1.67 (m, 6H), 1.34 (m, 18H), 0.90 ppm (t,
J=6.5 Hz, 9H); 13C NMR (50.32 MHz, CDCl3): d=164.4, 157.9, 152.8,
134.8, 134.0, 130.7, 119.7, 115.8, 114.7, 111.8, 72.5, 62.3, 51.2, 48.9, 31.5,
27.0, 26.6, 22.6, 14.0 ppm; HRMS (LSIMS+, mNBA): m/z calcd for
C63H70N9O6 [M+H]+: 1048.5449; found: 1048.5459.
Acknowledgement
F.T. acknowledges support by a Marie Curie Intra-European Fellowship
within the 6th European Community Framework Programme.
We
acknowledge
financial
support
from
DGA
(grant
N8.00.34.070.00.470.75.653). B.K.G.B. received a postdoctoral fellowship
from the French Minist›re de l’Education Nationale, de l’Enseignement
SupØrieur et de la Recherche. We are grateful to Prof. I. Ledoux and
Prof. J. Zyss for access to EFISHG facilities. Part of the calculations was
supported by the “Centre Informatique National de l’Enseignement Su-
pØrieur” (CINES-France).
[34] E. D. Rekaꢂ, J.-B. Baudin, L. Jullien, I. Ledoux, J. Zyss, M. Blan-
chard-Desce, Chem. Eur. J. 2001, 7, 4395–4402.
[35] E. Kelderman, W. A. J. Starmans, J. P. M. van Duynhoven, W. Ver-
boom, J. F. J. Engbersen, D. N. Reinhoudt, Chem. Mater. 1994, 6,
412–417.
[1] J. R. Heath, M. A. Ratner, Phys. Today 2003, 56, 43, and references
therein.
[2] T. C. Lin, S. J. Chung, K. S. Kim, X. P. Wang, G. S. He, J. Swiatkie-
wicz, H. E. Pudavar, P. N. Prasad in Polymers for Photonics Applica-
tions II (Ed.: K.-S. Lee), Springer, Berlin, 2003, pp. 157–193.
[3] P. N. Prasad, B. A. Reinhardt, Chem. Mater. 1990, 2, 660–669.
[4] M. Blanchard-Desce, C. R. Phys. 2002, 3, 439–448.
[5] S. R. Marder, B. Kippelen, A. K.-Y. Jen, N. Peyghambarian, Nature
1997, 388, 845–851.
[6] Special Issue on Supramolecular Chemistry and Self-Assembly: Sci-
ence 2002, 295.
[7] C. E. Dykstra, Adv. Chem. Phys. 2003, 126, 1–40.
[8] S. Di Bella, M. A. Ratner, T. J. Marks, J. Am. Chem. Soc. 1992, 114,
5842–5849.
[36] S. Yokoyama, T. Nakahama, A. Otomo, S. Mashiko, J. Am. Chem.
Soc. 2000, 122, 3174–3181.
[37] H. Ma, A. K.-Y. Jen, Adv. Mater. 2001, 13, 1201–1205.
[38] E. J. H. Put, K. Clays, A. Persoons, T. Biemans, C. P. M. Luijkx,
E. W. Meijer, Chem. Phys. Lett. 1996, 260, 136–141.
[39] E. Brouy›re, A. Persoons, J.-L. BrØdas, J. Phys. Chem. A 1997, 101,
4142–4148.
[40] G. P. Bartholomew, G. C. Bazan, J. Am. Chem. Soc. 2002, 124, 5183–
5196.
[41] G. P. Bartholomew, I. Ledoux, S. Mukamel, G. C. Bazan, J. Zyss, J.
Am. Chem. Soc. 2002, 124, 13480–13485.
[9] T. Hamada, J. Chem. Phys. 1996, 100, 8777–8781.
[10] J. M. Stout, C. E. Dykstra, J. Am. Chem. Soc. 1995, 117, 5127–5132.
[11] T. Verbiest, S. Houbrechts, M. Kauranen, K. Clays, A. Persoons, J.
Mater. Chem. 1997, 7, 2175–2189.
[42] A. Painelli, F. Terenziani, L. Angiolini, T. Benelli, L. Giorgini,
Chem. Eur. J. 2005, 11, 6053–6063.
[43] J. N. Demas, G. A. Crosby, J. Phys. Chem. 1971, 75, 991–1024.
[44] P. Debye, Polar Molecules, Dover, New York, 1945.
[45] B. F. Levine, C. G. Bethea, J. Chem. Phys. 1975, 63, 2666–2682.
[46] J. L. Oudar, D. S. Chemla, J. Chem. Phys. 1977, 66, 2664–2668.
[47] A. Willetts, J. E. Rice, D. M. Burland, D. P. Shelton, J. Chem. Phys.
1992, 97, 7590–7599.
[48] C. Reichardt, Chem. Rev. 1994, 94, 2319–2358.
[49] B. Boldrini, E. Cavalli, A. Painelli, F. Terenziani, J. Phys. Chem. A
2002, 106, 6286–6294.
[12] B. Champagne, D. M. Bishop, Adv. Chem. Phys. 2003, 126, 41–92.
[13] A. Datta, S. K. Pati, J. Chem. Phys. 2003, 118, 8420–8427.
[14] M. Guillaume, E. Botek, B. Champagne, F. Castet, L. Ducasse, Int.
J. Quantum Chem. 2002, 90, 1378–1387.
[15] M. Drobizhev, Y. Stepanenko, Y. Dzenis, A. Karotki, A. Rebane,
P. N. Taylor, H. L. Anderson, J. Am. Chem. Soc. 2005, 127, 15352–
15353.
[16] W.-H. Lee, H. Lee, J.-A. Kim, J.-H. Choi, M. Cho, S.-J. Jeon, B. R.
Cho, J. Am. Chem. Soc. 2001, 123, 10658–10667.
[50] A. Painelli, F. Terenziani, Chem. Phys. Lett. 1999, 312, 211–220.
[51] M. M. Tautounji, M. A. Ratner, J. Phys. Chem. A 2000, 104, 8566–
8569.
[52] Spectra were calculated without taking into account the inhomoge-
neous broadening effect due to the solvent. This has no influence on
the position and integrated intensity of the bands.
[17] M. Barzoukas, M. Blanchard-Desce, J. Chem. Phys. 2000, 113, 3951–
3959.
[18] F. Terenziani, A. Painelli, Phys. Rev. B 2003, 68, 165405.
[19] V. M. Agranovich, M. D. Galanin, Electronic Excitation Energy
Transfer in Condensed Matter, North-Holland Publishing Company,
Amsterdam, 1982.
[53] For dimers and trimer of series II, dipole moments have not been
measured.
[54] A. Painelli, F. Terenziani, J. Am. Chem. Soc. 2003, 125, 5624–5625.
[55] F. Terenziani, A. Painelli, J. Lumin. 2004, 35, 474–478.
[56] F. Terenziani, A. Painelli, A. Girlando, R. M. Metzger, J. Phys.
Chem. B 2004, 108, 10743–10750.
[20] J. Knoester, in Proceedings of the International School of Physics
“Enrico Fermi” 2001, Course CXLIX, Organic Nanostructures: Sci-
ence and Applications (Eds.: V. M. Agranovich, G. C. La Rocca),
IOS Press, The Netherlands, 2002, pp. 149–186.
[21] C. Lambert, W. Gaschler, E. Schmälzlin, K. Meerholz, C. Bräuchle,
J. Chem. Soc. Perkin Trans. 2 1999, 2, 577–587.
[57] R. S. Mulliken, J. Am. Chem. Soc. 1952, 74, 811–824.
[58] A. Painelli, F. Terenziani, J. Phys. Chem. A 2000, 104, 11041–11048.
[59] L. Del Freo, F. Terenziani, A. Painelli, J. Chem. Phys. 2002, 116,
755–761.
[60] For calculations presented here, eight vibrational states are enough
to converge results in series II, while at least 18 vibrational states
are needed for calculating spectra of series I; this is due to the much
[22] W. Leng, F. Würthner, A. Myers Kelley, J. Phys. Chem. B 2004, 108,
10284–10294.
[23] D. Beljonne, W. Wenseleers, E. Zojer, Z. Shuai, H. Vogel, S. J. K.
Pond, J. W. Perry, S. R. Marder, J.-L. BrØdas, Adv. Funct. Mater.
2002, 12, 631–641.
[24] C. Lambert, E. Schmälzlin, K. Meerholz, C. Bräuchle, Chem. Eur. J.
1998, 4, 512–521.
Chem. Eur. J. 2006, 12, 3089 – 3102
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3101

M. Blanchard-Desce et al.
greater vibrational relaxation energy for the chromophore of series
I than for the chromophore of series II.
[64] M. Mladenova, L. Ventelon, M. Blanchard-Desce, Tetrahedron Lett.
1999, 40, 6923–6926.
[61] G. D. Scholes, Annu. Rev. Phys. Chem. 2003, 54, 57–87.
[62] D. A. Kleinman, Phys. Rev. 1962, 126, 1977–1979.
[63] While for spectra of products of series I, eight vibrational states are
not enough, they lead to converged results for ground-state proper-
ties.
Received: July 28, 2005
Published online: January 23, 2006
3102
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3089 – 3102