A distinctive example of the cooperative interplay of structure and environment in tuning of intramolecular charge transfer in second-order nonlinear optical chromophores

Article abstract of DOI:10.1002/chem.200204356

The strongly enhanced cooperative influence of medium polarity and organic structural design on the first hyperpolarizability β of a novel family of highly polarizable azinium-(CH=CH-thienyl)-dicyanomethanido chromophores 1-3 is described. The dyes can be

Full text of DOI:10.1002/chem.200204356

FULL PAPER
A Distinctive Example of the Cooperative Interplay of Structure and
Environment in Tuning of Intramolecular Charge Transfer in Second-Order
Nonlinear Optical Chromophores
[
a]
[a]
[b]
[a]
Alessandro Abbotto,* Luca Beverina, Silvia Bradamante, Antonio Facchetti,
[
c]
[a]
[c]
Christopher Klein, Giorgio A. Pagani,* Mesfin Redi-Abshiro, and
[
c]
R¸diger Wortmann
Abstract: The strongly enhanced coop-
erative influence of medium polarity
and organic structural design on the first
hyperpolarizability b of a novel family of
highly polarizable azinium-(CHˆCH-
the ground and excited states present b
values in a broad range, eventually
switching from negative to positive.
Our investigation is based on a combi-
nation of experimental (UV/Vis spec-
troscopy, multinuclear NMR spectrosco-
py, and electrooptical absorption meas-
urements) and computational (ab initio)
approaches. It is shown that: 1) b and mb
are dramatically influenced, even by
orders of magnitude, by a complex,
non-monotonic interplay of structure
and medium action, which in turn affects
molecular ICT and bond length alterna-
tion (BLA), 2) the computations, vali-
dated by different experimental data,
are to be recommended as an extremely
useful tool in the search for a greatly
improved set of molecular nonlinear
optical (NLO) responses (in the case of
1 ± 3 they show that such conditions may
be attained only in a narrow and limited
range of dielectric constants in which the
annelation effect operates most effi-
ciently), and 3) the search for the most
favorable molecular NLO response of a
highly polarizable chromophore both in
solution and in solid matrices should
simultaneously take into account not
only the molecular design supplemented
by annelation effects but also the polar-
ity of the medium.
thienyl)-dicyanomethanido
chromo-
phores 1 ± 3 is described. The dyes can
be efficiently synthesized by regioselec-
tive protonation/alkylation of the corre-
sponding bidentate anion precursors.
Consecutive annelation of the pyridyl
ring of 1 (pyridine ! quinoline ! acri-
dine) and medium polarity effects are
responsible for an extraordinarily varia-
ble range of intramolecular charge
transfer (ICT), leading to a large set of
p-electron distribution patterns. Ac-
cordingly, systems with remarkably dif-
ferent zwitterionic/quinoid character in
Keywords: ab initio calculations ¥
annelation ¥ heterocyclic zwitterions
¥
NMR spectroscopy ¥ nonlinear
optics
Introduction
[
a] Prof. Dr. A. Abbotto, Prof. Dr. G. A. Pagani, Dr. L. Beverina,
Dr. A. Facchetti
Department of Materials Science and INSTM
University of Milano-Bicocca, Via Cozzi 53
The electronic and structural properties of organic push ± pull
conjugated compounds are of considerable interest for non-
linear optical (NLO) applications in photonic and electrooptic
devices such as high-speed photonic switching and electro-
optic modulators.[ To generate highly efficient NLO-phores
with a large quadratic hyperpolarizability b, extensive molec-
ular polarization of the p-electron structure is usually
provided by means of a strong interaction between end-
capped donor and acceptor groups. This situation gives rise to
a dominant intramolecular charge-transfer (ICT) transition
from the ground state (g) to their first excited state (e). The
sign and value of b have been correlated with the bond length
alternation (BLA)[ structural parameter, whereas other
authors have correlated b, through the two-state model, to
2
0125 Milano (Italy)
Fax : (‡39)02-64485403
1]
E-mail: alessandro.abbotto@unimib.it, giorgio.pagani@unimib.it
[
b] Dr. S. Bradamante
CNR-Institute of Molecular Sciences and Technologies
Via Golgi 19, I-20133, Milano (Italy)
[
c] C. Klein, M. Redi-Abshiro, Prof. Dr. R. Wortmann
Department of Physical Chemistry
University of Kaiserslautern
Erwin-Schroedinger-Strasse, 67663, Kaiserslautern (Germany)
Supporting information for this article is available on the WWW under
http://www.chemeuj.org or from the author. UV/Vis spectra of 1a,1b,
2]
2
a,2b and 3a in different solvents (Figures S1 ± S5); optical and
electrooptical absorption spectra of 1b and 2b in dioxane (Figures S6
and S7); computed structures of 1c ± 3c in dioxane and chloroform
2[3]
[4]
other expressions such as c and MIX. All three descriptors
are correlated between themselves.[ In the plot of b against
BLA it has been shown that in push ± pull polyenes b moves
4]
(
Figures S8 and S9); table of computed energies (Table S10); plot of b
0
against e (Figure S11); x,y,z coordinates of all structures.
Chem. Eur. J. 2003, 9, 1991 ± 2007
DOI: 10.1002/chem.200204356
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1991

A. Abbotto, G. A. Pagani et al.
FULL PAPER
from positive (BLA < 0, neutral ground state) to negative
because of the extremely low solubility of the zwitterions of
(
BLA > 0, dipolar ground state), with values passing through
type A–that in all cases retain highly polar natures–in any
[5]
zero (BLA ˆ 0, cyanine limit).
solvent of low polarity. The EOAM (electrooptical absorption
measurements) approach[
16, 17]
seemed practicable since the
We have reported the synthesis of a novel series of
chromophores[ A endowed with solvent-dependent, very
large, and negative mb values (Scheme 1). Indeed, the
solvent polarity has been shown both theoretically and
experimentally[ to have a great influence on ICT, and thus on
b. Changing the strength of the donor and the polarity of the
solvent b can even change the sign.
6]
measurements can be obtained at concentrations typical of
UV/Vis absorption spectroscopy. In addition, we felt it might
be profitable to carry out ab initio computations that, by
taking into account of the solvent effects, could mimic the
contributions of the dipolar and zwitterionic structures in the
different environments.
[7]
[8]
9]
Since, on the whole, the com-
bination of BLA, which strictly
depends upon the p-electron
structure, and of the solvent
influences the change in the rel-
ative contributions of the two
neutral and dipolar limit forms
of the chromophore ground and
excited states, we decided, taking
the zwitterion 1a (see Scheme 2)
as a starting point, to extend the
investigation to very similar sys-
tems presenting in their struc-
tures other acceptor groups, but
ones with small or negligible
variation in their electron-with-
Scheme 1. Resonance between zwitterionic/aromatic ± neutral/quinoid (system A) and neutral/aromatic ± di-
polar/quinoid (DANS-like systems) limit forms.
drawing capacities and that
would in turn be able to influence
the BLA on the basis of bond
Our molecular design approach was based on the following
main strategic choices:
localization only. We took the decision, unprecedented in the
literature, to use benzofusion (annelation effect)[ of the
acceptor pyridinium ring as a regulator of p-delocalization.
Accordingly, we prepared and characterized chromophores
1 ± 3 and, to address solubility problems, also planned the
synthesis of the type b N-decyl-substituted chromophores
(Scheme 2). Indeed, Scheme 3 shows that the bond lengths of
the bonds from the pyridyl-nitrogen atom to the position at
which the double bond in 1 ± 3 is joined show considerable
alternation. Thus, consecutive annelation of pyridine with one
and then with two benzene rings in 1 ± 3 can cause significant
bond alternation and would be expected to influence the
relative importance of the aromatic/quinoid character of the
azinium ring, thus affecting the entire p framework.
Scheme 2 clearly shows that, whereas the pyridine ring loses
aromaticity in the right-hand side formula, the acridinium ring
can better sustain a quinoid character thanks to the restored
full aromaticity of the two benzene side rings. The quinoli-
nium ring is somewhere in the middle. It is therefore to be
expected that in the ground state the importance of the ZW
form should sequentially decrease, and the Q form simulta-
neously increase, on going from pyridine to quinoline to
acridine. To the best of our knowledge the annelation effect
has never been used as a structural tool for affecting the
relative contribution of the two limit forms, but the acridine
ring was also never reported as an acceptor moiety in NLO
18]
1
) Exploitation of p-excessive and p-deficient heteroaromat-
ic compounds[ as key constituent donors and acceptors of
the push ± pull NLO-phore.[ Theoretical and experimen-
tal studies have demonstrated that p-excessive and p-
10]
11]
[11]
deficient heterocycles can efficiently function as primary
and/or auxiliary[ donor and acceptor groups in push ±
pull dyes; furthermore, the reduced ring aromaticity
relative to benzene of the five-membered heterocycles
leads both to an increased transition moment and to
change in dipole moment.[
12]
13]
2
) In contrast to the most common approach in the litera-
ture (e.g., for 4-dimethylamino-4'-nitrostilbene DANS,
Scheme 1), the dipolar (zwitterionic) limit formula retains
aromaticity, whereas the neutral limit formula is qui-
noid.[
14]
We realized, however, that for the above zwitterions, both
the large solvatochromic response and the highly solvent-
dependent mb values, which increase as the solvent polarity
decreases, needed a rationale. This prompted us to investigate
the solvent response of the zwitterionic systems more closely,
and in this we were strongly supported by our recent
[15]
experimental findings that the solvent modulates the extent
of ICT within a series of pyridine-based systems. We thus
decided to obtain mb values in a solvent possessing a dielectric
constant e definitely lower than those of the previously
dyes. Indeed, the quinolinium ring has hardly appeared in the
NLO literature[
14d, 19]
and in none of these cases did the
considered solvents (DMF and CHCl ). Preliminary experi-
3
ments convinced us that EFISH (electric-field-induced sec-
ond harmonic) generation[ experiments were not viable
authors mention the annelation effect as a precise tool for
designing new dyes with optimized values of b.
1]
1992
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
Scheme 2. Effect of annelation of the pyridine ring on the contribution of the zwitterionic/aromatic (ZW) and neutral/quinoid (Q) resonance structures in
the description of the ground state.
phore 3 is reported in Scheme 5. Bromo derivatives 7 and 8
are new compounds. Compound 7 was prepared by a
Horner± Wittig reaction between the diethyl[(5-bromothien-
2
-yl)methyl]phosphonate anion and 4-formylquinoline in
THF. Condensation of 4-methylacridine (9) with 5-bromo-2-
formylthiophene in AcONa/AcOH solution afforded 8 as a
mixture of the E and Z geometric isomers. This result is
surprising in view of the large steric hindrance experienced by
the cis diastereoisomer. Isomers (E)-8 and (Z)-8 can conven-
iently be separated by column chromatography.
In a previous paper[ we described the preparation of
sodium salt 4. Similarly, 5 and 6 were synthesized from the
parent bromothiophenes 7 and 8, respectively, by treatment
with malonodinitrile sodium salt in dimethoxyethane, with
tetrakis(triphenylphosphine)palladium(0) as catalyst. Note
that both (E)-8 and (Z)-8 diastereoisomers react with
malonodinitrile sodium salt to afford the same product (E)-
6 (Scheme 6). It is likely that each bromo isomer gives the
Scheme 3. Bond lengths [ä] in pyridine, quinoline, and acridine (averages
of different X-ray data reported in the literature for alkyl derivatives).
In short, for the detailed investigation of systems 1 ± 3 we
will use a combination of a number of approaches, very
different in nature and targets: organic design and synthesis,
multinuclear NMR investigation to provide insights into
charge-mapping and bond length alternation, linear absorp-
tion, EOA measurements in 1,4-dioxane for comparison with
and discussion of previous measurements in solvents of higher
polarity, and computational (ab initio) investigation, with
effects of the medium always included. We will show that the
successive steps we propose herein–molecular design includ-
ing annelation effects and choice of the
6]
medium–can provide an extremely varia-
ble (to the extent of orders of magnitude)
set of b values. We believe that the proce-
dure outlined herein provides an unprece-
dented novel methodology to access opti-
mized NLO properties.
Results
Synthesis of chromophores
Scheme 4. Synthesis of pyridine chromophores 1a and 1b and quinoline chromophores 2a and 2b:
The synthesis of chromophores 1 and 2 are
reported in Scheme 4 and that of chromo-
i) HCl, DMSO; ii) CF
3
SO
3
C
10
H
21, acetone, RT; iii) NaH, THF, reflux; iv) CH
CO , CH CN, D.
2
(CN)
2
, NaH, DME,
[Pd(PPh ], RT; v) CF
3
)
4
3
SO
3
C
10
H
21, K
2
3
3
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1993

A. Abbotto, G. A. Pagani et al.
ZW) and quinoid (Q) limit
FULL PAPER
(
resonance formulas, and the
solvatochromic effect–defined
as the dependence of the dyes×
UV/Vis absorption spectra on
solvent polarity–is a useful
method for investigating their
nature. Depending on chromo-
phore structure and solvent po-
larity,[ one of the two limit
formulas will be more suitable
to describe the chemical bond-
ing of the real molecule in the
two states (Scheme 7).
20]
Scheme 5. Synthesis of the acridine derivative 3a and failure to prepare 3b: i) AcOH, ZnCl
2
, 2008C; ii) AcOH,
C H
3 10
AcONa, 1308C; iii) CH
2
(CN)
2
, NaH, DME, [Pd(PPh
3
)
4
], RT; iv) HCl, DMSO; v) CF
3
SO 21, different
Each of the chromophores
conditions.
1
± 3 presents a strong charge-
Scheme 7. Zwitterionic/aromatic (ZW) and neutral/quinoid (Q) resonance
structures of chromophores 1 ± 3.
Scheme 6. Preparation of the E diastereoisomer of the acridine-based
sodium salt 6 from both the E and the Z forms of the precursor bromo
2 2 3 4
derivative 8: i) CH (CN) , NaH, DME, [Pd(PPh ) ], RT.
transfer (CT) band in the visible region of the spectrum.
Table 1 lists absorption data (lmax) relative to the CT band in
selected solvents (the corresponding spectra can be found in
the Supporting Information; Figures S1 ± S5). Maximum
wavelength values listed in Table 1 are taken at the local
maximum of the lowest energy band, on the assumption that
corresponding sodium salt and that (Z)-6 only subsequently
isomerizes to (E)-6.
Finally, the last synthetic step to prepare dyes 1 ± 3 was
designed on the basis of our previous experience on bidentate
anions. All chromophores 1a ± 3a were conveniently prepared
in high yields (> 80%) by regioselective nitrogen protonation
of anions 4 ± 6 with 10% hydrochloric acid. Among the N-
alkyl substituents we made use of a medium-length chain to
improve the solubility of the final zwitterions 1 ± 3. The
sodium salts 4 ± 6 exhibit different behavior toward alkylation
with alkyl triflates. In fact, while the reaction between 4 and
CF SO C H took place efficiently and in high yield,
[21]
this band should refer to the monomeric species. A recent
investigation into the aggregation of merocyanine dyes has
shown that the band of lowest energy can be ascribed to the
monomers, with subsidiary vibronic maxima or shoulders at
shorter wavelengths; dye aggregates are hypsochromically
[22]
shifted with respect to the monomer.
If the molecular ground state is better described by ZW, the
charge-transfer direction is from the negatively charged
dicyanomethanido moiety to the N-alkylazinium group. The
other possibility, envisioned when Q is the better ground state
descriptor, is from the aminoquinoid group to the dicyano-
vinyl acceptor. The first situation gives rise to a negative
solvatochromic response, which becomes positive in the latter
case.
3
3
10 21
compound 5 reacted less easily and 6 did not at all. Compound
b was prepared in acetone, but sodium salt 5 afforded 2a
1
under the same conditions. Probably the quinoline nitrogen
atom is less reactive and so the alkyl triflate has time to react
with acetone to form triflic acid, which protonates 5 to give
2
a. To avoid this inconvenience, the alkylation reaction was
carried out in acetonitrile and in the presence of K CO .
2
3
Compound 2b was isolated in pure form after column
chromatography. All attempts to prepare 3b with different
alkylation reagents and/or under different experimental
conditions failed. However, the fact that 3b was not available
does not constitute a problem: compound 3a is fortunately
soluble enough and the optical properties of the series 1/2 (a)
are indistinguishable from those of series 1/2 (b).
Table 1. Solvatochromic data[a] (lmax [nm] of the charge-transfer band) for
the chromophores 1a ± 3a, 1b, and 2b in selected solvents.
Compound DMSO Acetone THF CHCl [
b]
Dioxane[b] Dl
3
(
e)
(46.7)
(20.7)
(7.6) (4.8)
(2.2)
1a
608
619
697
714
787
635
645
723
738
678
676
686
739
758
672
684
702
746
765
658
709
708
748
763
630
À 101
1
2
2
3
b
a
b
a
À 89
À 51
À 49
Solvatochromic data
‡ 157
Both the ground and the excited states of molecules 1 ± 3 can
be represented as linear combinations of the zwitterionic
[
a] Dl ˆ lmax(DMSO) À lmax(dioxane); positive Dl values mean positive
solvatochromism. [b] The lowest energy transition was considered.
1994
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
3
Multinuclear NMR investigation
coupling constants of dyes 1 ± 3 in general, and J of the p-
bridge in particular, would be expected to vary on going from
a ZW to a Q limit structure (Scheme 8).
13C and 15
N NMR data: Theoretical studies have shown the
strict correlation between molecular NLO activity and charge
distribution on the donor, spacer, and acceptor chromophore
groups.[
8b, 23]
We had previously proposed, and confirmed,
[24]
[25]
the validity of empirical p-electron/shift relationships [Eq. (1)
and Eq. (2)] to correlate the variation of the p electron
2
density on sp carbon and nitrogen atoms with variation of the
corresponding chemical shift.
1
3
p
Dd( C) ˆ À160Dq
C
(1)
(2)
3
Scheme 8. Use of the J(H,H) coupling constant of the central ethenylic
unit of chromophores 1 ± 3 to monitor the different contributions of the
zwitterionic/aromatic (ZW) and neutral/quinoid (Q) resonance structures
to the description of the ground state.
1
5
p
Dd( N) ˆ À366.34Dq
N
The 13C and 15N NMR shifts (Figure 1) therefore allow a
map of charge distribution on the chromophore structure in
solution to be drawn and could, for the first time, provide a
useful experimental correlation between p-charge and NLO
activity.
Proton NMR spectra of compounds 1b,2b , and 3a were
recorded in solvents of different polarities. We chose deu-
terated dimethyl sulfoxide, chloroform, and dioxane: their
corresponding protonated analogues have dielectric constant
e values of 46.7, 4.80, and 2.21, respectively. By this approach it
is possible to analyze: 1) whether the ground state chemical
representation of dyes 1 ± 3 is closer either to the ZW or to the
Q structure in that particular environment, and 2) the
sensitivity of each structure to solvent polarity.
1
3
H NMR data: J(H,H) coupling constants are particularly
sensitive to H-C-C-H bond lengths, angles, and carbon
hybridization: in short, to molecular structure.[4b, 26] Combined
theoretical and NMR investigations on cyanine, merocyanine,
and polyenes dyes have shown a clear correlation between the
coupling constant of trans vicinal protons in the conjugated
1
Figure 2 shows H NMR spectra of 1b in these three
[27]
3
chain and the corresponding CÀC bond order. In addition,
solvents. Table 2 lists J(H,H) values for the chromophores
solvent-dependent NMR studies conducted on push ± pull
under investigation in selected solvents. Considering the same
chromophore system, and therefore looking at Table 2 along
the rows, it can be noted that a decrease in solvent polarity is
concomitant with a decrease in
polyenes suggested that the ground-state structure can be
significantly affected by solvent polarity.[
4c, 28]
Therefore, the
3
the J(H,H) coupling constant.
This result suggests the conclu-
sion that the double-bond char-
acter of the p bridge is reduced,
indicative of an increased con-
tribution of the quinoid limit
formula. This change is larger
for the pyridine-based system
3
(
D J(H,H) ˆ 1.51 Hz) than for
3
the quinoline- (D J(H,H) ˆ
1
(
.31 Hz) and acridine-based
3
D J(H,H) ˆ 1.07 Hz) systems.
Alternatively, when the same
3
solvent is considered, J(H,H)
1) > J(H,H) (2) > J(H,H) (3),
3
3
(
showing that the pyridine-
based system 1 is always more
zwitterionic than 2, while 3 has
the most quinoid character in
all solvents.
A further remarkable result
is that a drastic change in the
chemical shifts is recorded on
going from DMSO to less polar
solvents. Pyridine protons shift
upfield by about 1 ppm on go-
Figure 1. 13C and 15N (italics) NMR data in DMSO at 278C for compounds 1a ± 3a, 1b and 2b; nd means that the
ing from DMSO to CDCl /di-
3
signal was not detected.
oxane, meaning that they move
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1995

A. Abbotto, G. A. Pagani et al.
FULL PAPER
and 3a and to analyze the
resonance structures of the
chromophores in terms of the
Q and ZW limit contributions.
The EOA technique meas-
ures the influence of the square
of an external electric field on
the molar decadic absorption
coefficient according to Equa-
E
tion (3), where e is the absorp-
tion coefficient in the presence
of the electric field E.
E
2
e (f, nÄ ) ˆ e( nÄ )[1‡L(f, nÄ )E ‡...]
(3)
The relative change induced
by the field L is a function of
the wavenumber nÄ ˆ 1/l and the
angle f between the polariza-
tion vector of the incident light
and the applied field. Typically
the EOA spectrum is recorded
for two polarizations f ˆ 08 and
9
08, and multilinear regression
analysis in terms of the optical
absorption spectrum and its
first derivative yields a set of
regression coefficients D, E, F
Figure 2. 1H NMR spectra of 1b in [D
6
3 8
]DMSO, CDCl , and [D ]dioxane at 258C (aromatic region).
1
Table 2. Solvent-dependent H NMR coupling constants [Hz] of the
and G, from which the ground state dipole moment m and the
g
central ethenylic unit in systems 1 ± 3.[
a]
dipole difference Dm between the ground state (g) and the
excited state (e) may be calculated.[ An important addi-
tional quantity is the magnitude of the transition dipole
30]
Compound
Dioxane
CHCl
3
DMSO
1
2
3
b
b
a
13.61
12.99
12.69
14.02
13.42
12.78
15.12
14.33
13.76
moment m , which may be determined from the integrated
absorption spectrum. It has been demonstrated in a series of
eg
papers[
3, 31]
that knowledge of Dm, m and of the wavelength of
eg
[
a] Measured at 298 K.
optical excitation (l ) of the CT transition obtained from UV/
eg
Vis and EOA spectra allows the static second order polar-
from a typical pyridinium salt environment to a more quinoid-
like structure. In addition, while the DMSO spectrum exhibits
very sharp signals, the dioxane and CDCl3 spectra are
somewhat broader. This result can be explained by assuming
completely free rotation about the CÀC bonds adjacent to the
CˆC in a zwitterionic-like structure (DMSO), whereas an
izability b [Eq. (4) ] and the anisotropy of the first order
0
polarizability da of the dye to be estimated in a two-level
0
model [Eq. (5)].
6
m
2e gDml 2e g
b
0
ˆ
(4)
(5)
2
ꢀ
hc†
equilibrium between different rotamers can be envisioned in a
2
quinoid-like structure (CDCl , dioxane). Another possibility
2m leg
3
eg
da
0
ˆ
could be due to the formation of molecular aggregates, a
known phenomenon[ more likely in nonpolar or weakly
polar solvents. A similar trend, but more attenuated, can be
observed for all of the other chromophores.
ꢀhc†
29]
Furthermore, EOA results allow us to analyze the reso-
nance structure of the chromophores in a simple CT model,
which describes the ground and excited states as linear
combinations of the Q and ZW limit structures on the basis of
Electrooptical and nonlinear-optical properties
2
the resonance parameter c . According to this simple model,
donor–acceptor-substituted p-conjugated chain molecules
2
Optical absorption spectra of donor–acceptor-substituted p
systems typically show a single intense charge-transfer (CT)
transition at low energy. A powerful experimental technique
may be classified from polyene-type (c ꢁ 0) through neutro-
2
cyanines (™cyanine limit∫, c ꢁ 0.5) to zwitterionic systems
2
(c ꢁ 1). The implications of this classification for the opti-
with which to characterize these CT transitions is electro-
mization of chromophores for nonlinear-optical and photo-
refractive applications have been discussed in a number of
publications.[ A further parameter of this model is the
maximal hypothetical dipole change Dmmax: the difference of
optical absorption (EOA) spectroscopy.[
16, 17]
In this work we
32]
have utilized this technique to determine the ground and
excited state dipole moments of the chromophores 1b, 2b,
1996
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
the dipole moments of the Q and the ZW structures. The
model parameters c and Dmmax can be estimated from
Equation (6) and Equation (7)
further evidence of its reduced dipolar nature relative to 1b
and 2b. The spectra of 3a are therefore displayed in conven-
tional absolute units (Figure 3) and the regression could be
2
extended over the complete CT band. The ground state dipole
2
2
2
À1/2
[30]
c ˆ 1/2[1 À Dm(4meg‡Dm )
]
(6)
(7)
moments were calculated as usual
from E À 6D (cf.
Table 3) in the case of 3a and from E À 2D for 1b and 2b,
taking contributions from electric field-induced dimer disso-
ciation[ into account. Dipole differences were calculated
2
Dmmax ˆ Dm/(1 À 2c )
22]
Figures S6 and S7 in the Supporting Information and
Figure 3 display the EOA spectra of 1b, 2b, and 3a,
respectively, and Table 3 summarizes all EOA results. Be-
cause of partial aggregation (dimer formation) of 1b and 2b in
[30]
from the average values of regression coefficients F and G.
Generally, the EOA spectra of the chromophores are
governed by two effects. The first is the electrodichroic effect
caused by orientation of the chromophores in the externally
applied electric field. This effect is measured by the regression
coefficient E. In all cases, strongly positive electrodichroism is
observed: that is, the absorption for parallel polarization (f ˆ
0
8) of the incident light field relative to the externally applied
field increases, while decreasing for the orthogonal polar-
ization (f ˆ 908). This allows the conclusion that the ground
state dipole moment m and the transition dipole moment m
g
eg
are essentially parallel and oriented along the conjugated
chain of the chromophores. The second effect is the band shift
effect (Stark effect) of the CT band in the electric field due to
the change in the dipole moment Dm upon electronic
excitation. The largest shift of the EOA spectrum relative to
the optical absorption spectrum is clearly observed for 3a
(
Figure 3), indicating a large value of Dm, while the shifts in
the spectra of 1b and 2b (see Figures S6 and S7 in the
Supporting Information) are much smaller.
Figure 3. Optical (e/ nÄ ) and electro-optical absorption spectra (Le/ nÄ ) of 3a
in dioxane, Tˆ 298 K. Data points for parallel (*: f ˆ 08) and perpendic-
ular polarisation (*: f ˆ 908) of the incident light relative to applied
electric field and multilinear regression curves.
Quantitative evaluation of the regression coefficients shows
that the ground state dipole moment increases strongly in the
sequence 3a, 2b, and 1b (Table 3), while the dipole difference
decreases and even becomes negative in the case of 1b. This
may be interpreted in terms of larger ZW contributions to the
electronic ground states in 2b and 1b. The calculated
Table 3. Results of electrooptical absorption measurements in dioxane
(
Tˆ 298 K) and derived nonlinear-optical properties of 1b, 2b, and 3a.
2
Parameter units
Compound
resonance parameter c shows that 1b has the largest ZW
1
b
2b
3a
character (56%) and at the same time the largest ground state
dipole moment. The ZW character of 2b is considerably
reduced (46%), and 3b is already largely Q-like (only 20%
À20 À2
2
D
E
F
[10
[10
V
m ]
7820 Æ 530
3660 Æ 430
80400 Æ 2700
1170 Æ 580
1780 Æ 580
763
±
36.8
56.2 Æ 1.1
5.5 Æ 1.5
61.7 Æ 1.8
104
300 Æ 90
26500 Æ 540
7090 Æ 390
7430 Æ 390
630
À20 À2
2
V
m ]
118500 Æ 3100
À40
À1
2
[10 CV m ] À 2780 Æ 620
2
ZW character). The resonance parameter c correlates
À40
À1
2
G
[10 CV m ] À 2860 Æ 620
linearly with the ground state dipole moment, with slope
l
eg
[nm]
708
±
37.0
À30
À30
2
À1
94 Â 10 Cm and intercept at 14 Â 10 Cm. The slope
e
[m mol
]
3410
31.6
À30
m
m
eg
g
[10 Cm]
agrees reasonably with the value of Dm and the intercept
max
À30
[10 Cm]
66.7 Æ 1.1
À 8.9 Æ 1.1
57.8 Æ 1.8
98
À 92 Æ 15
0.56 Æ 0.01
74.5 Æ 0.2
32.7 Æ 0.5
46.7 Æ 1.9
79.3 Æ 1.8
69
338 Æ 14
0.20 Æ 0.01
78.6 Æ 1.1
with the dipole moment of two cyano groups attached to a
weakly polarized p system. Thanks to the close similarity of
the topologies of the three chromophores, the dipole differ-
ence Dmmax calculated in the CT model should be similar for all
of them. In fact, the estimated values of Dmmax are found to be
almost identical for all chromophores, a result that nicely
corroborates the model assumptions.
À30
Dm [10 Cm]
À30
m
e
[10 Cm]
À40
À1
2
Da
0
[10 CV m ]
À50
À2
3
b
0
[10 CV m ]
66 Æ 18
c2
0.46 Æ 0.01
73.8 Æ 0.2
À30
Dmmax [10 Cm]
dioxane at the concentration of the EOA measurement, only
the lowest energy vibronic band of the spectra were taken into
account in the regression analysis. These bands originate from
the nonaggregated monomer forms of the chromophores,
while the higher energy bands strongly overlap with the dimer
absorption. For the same reason, the absorption and EOA
spectra of 1b and 2b are displayed in Figures S6 and S7 in the
Supporting Information in relative units only. Compound 3a
did not show dimer formation in dioxane solution, providing
Ab initio computations
Need for computational approaches to provide a rationale for
the experimental results and to assess predictability: The
inversion of sign in the solvatochromic response on going
from the pyridine and quinoline derivatives 1 and 2 to the
acridine derivative 3 (Table 1), the large solvent-dependent
1
H NMR chemical shift variations of the pyridine-based
system 1b (Figure 2), and the dependence of the coupling
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1997

A. Abbotto, G. A. Pagani et al.
FULL PAPER
constant value of the central ethenylic unit on structural
program packages.[34] Because of the considerable size of the
molecules under investigation and the high computational
requirements due to the inclusion of the solvent effect (see
below), we carried out full geometry optimizations and
frequency computations at the Restricted Hartree ± Fock
(RHF) level of theory, using 6-31G* as a basis set and
(
pyridine ± quinoline ± acridine) and medium (dioxane ±
CHCl ± DMSO) effects (Table 2) represent unequivocal ex-
3
perimental evidence of the high tuneability of the p-electron
structure of the class of azinium-[-CˆC-thienyl]dicyanome-
thanido dyes 1 ± 3. Evidence that this result should be
successfully translatable into structure- and medium-based
adopting C symmetry for all the species. Test density func-
s
fine control is suggested by the set of the b experimental
values obtained from EOA measurements in dioxane (Ta-
ble 3) and by the very different NLO EFISH data obtained for
tional theory (DFT) calculations did not produce significantly
different conclusions.[ The higher computational cost and
the belief that the accuracy of the HF approach would not be
much improved by higher levels of theory for the purposes of
our work discouraged us from pursuing different approaches.
Conversely, for the reasons outlined in the previous section
pertaining to the experimentally inaccessible full range of
solvent polarities, we paid considerable attention to obtaining
clues on the interaction of polarizable systems, such as 1 ± 3,
with different media by means of calculations. Thus, struc-
tures, electronic parameters and b values were computed,
taking the solvent effect into account by adoption of
Onsager×s self-consistent reaction field (SCRF) approach,[
as available in the Gaussian package. In this model the
chromophores are inserted in a spherical cavity in a contin-
uum of solvent molecules: the solute radius and the dielectric
constant e of the solvent are the input parameters for Onsager
SCRF computations.[
0
35]
system 1 when the solvent is shifted to CHCl and DMF
3
7]
(
Table 4).[ Unfortunately, lack of NLO EFISH measure-
ments for systems 2 and 3, due to low solubility, and
experimental difficulties in use of the EOAM technique in
polar solvents, prevented us from obtaining a full set of
experimental NLO values as a function of structure and
solvent.
All these results are functions of the complex interplay of
structural (BLA) and solvent effects. Experimentally based
discrimination between the two contributions appears com-
plex. We thus reverted to computational approaches and
decided to perform a full and detailed ab initio investigation
of the structural, electronic and NLO properties of systems
36]
1
± 3, considering both structural and solvent effects. Our goal
37]
is: a) to rationalize and complement available experimental
data with a set of additional important parameters, b) to
extend the investigation to media in which experimental
measurements are not viable or reliable, including several
solvents of different polarity and the gas phase. Agreement of
experimental and computed results should, when both are
available, allow us to produce a reliable picture of the effect of
structure and medium on molecular properties, and conclu-
sions based on the combination of different experimental
techniques and ab initio study would be legitimate.
Ab initio ground state geometric and electronic computed
parameters: To investigate the dependence of molecular
parameters on solvent effects we have performed geometry
optimization and charge computations with several solvent
dielectric constants, ranging from 1.0 (gas phase) to 46.7
(DMSO). In this way, we were able to compute all the
relevant molecular parameters in solvents of very different
natures, going from apolar to highly polar media. However,
for direct comparison with experimentally determined (NMR
and b) data, most of the values presented refer to the gas
phase and to the following solvents: dioxane (e ˆ 2.21), CHCl3
(e ˆ 4.80), DMF (e ˆ 36.70), and DMSO (e ˆ 46.70).
Computational details: Computational investigation was
carried out for molecules 1c ± 3c, taken as representative
models of the compounds studied in this work, and on the
assumption that the different substitution at the pyridine-like
nitrogen (H versus Me, alkyl chain) would not affect the p-
electron structure and NLO properties of the molecules.
Results are therefore directly correlated with experimental
data obtained for the species presented in the previous
section. All ab initio computations used the GAUSSIAN 98[
Optimized structures of compounds 1c ± 3c in the gas phase
and in DMSO, dioxane and CHCl , including relevant bond
3
length values, are presented in Figure 4, Figure 5, and
Figures S8 and S9 (see the Supporting Information), respec-
tively. Table S10 (in the Supporting Information) gives
absolute energy values (atomic units) and frequency analysis
33]
Table 4. Experimentally measured and computed[a] dipole moments and first hyperpolarizabilities of compounds 1 ± 3.
System
e)
Gas phase
Dioxane[b]
CHCl
(4.8)
3
DMF
(
(2.2)
(36.7)
[
c]
[d]
[e]
[c]
[d]
[e]
[c]
[d]
[e]
[c]
[d]
[e]
m
g
b
0
mb
0
m
g
b
0
mb
0
m
g
b
0
mb
0
m
g
b
0
mb
0
[
f]
[f]
1
exptl
calcd
20.0
27.1
(28.2)
16.9
24.2
9.8
À 62
À 30
À 1240
À 810
À 6990
À 9600
À 2000
À 4200
[
g]
18.4
20.3)
17
310
45.8
(37.5)
À 210
54.1
(50.8)
À 78
(
2
3
exptl
calcd
exptl
calcd
44
31
230
89
750
750
2230
1760
16.6
14.3
22
29
360
420
47.6
48.1
À 230
À 280
À 11000
À 13500
56.9
58.5
À 69
À 62
À 3900
À 3600
19.8
[
a] RHF/6 ± 31G*//RHF/6 ± 31G*; C
s
symmetry point group; compounds 1c ± 3c. [b] Experimentally determined values from electrooptical absorption
À30
measurements on 1b, 2b, and 3a. [c] in Debye. [d] Values according to the phenomenological convention (ref. [44]); 10 esu. [e] Values according to the
phenomenological convention (ref. [44]); 10 esu. [f] EFISH measurements for compound 1c (ref. [7]). [g] B3LYP/6 ± 31G*//B3LYP/6 ± 31G* computed
À48
values in parentheses.
1998
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
methyl rotation. In the case of
2
c and 3c, frequency analysis
therefore showed that the Cs
geometry does not represent a
local minimum. However, we
have also assumed for these
compounds that computation-
al data relative to the C ge-
s
ometry would be sufficiently
accurate for our purposes.[
38]
The results clearly reveal
not only that the structures of
the three compounds are dra-
matically affected by the po-
larity of the surrounding me-
dia, but also that the predom-
inant character (ZW versus Q)
is not maintained throughout
the range of polarity consid-
ered. In particular, the natures
of the three systems are pre-
dominantly quinoid in the gas
phase, whereas in the highly
polar DMSO the zwitterionic
character is strongly predom-
inant. An intermediate situa-
tion is found for less polar
solvents, such as dioxane and
CHCl , in which, however, the
3
contribution of the zwitterion-
ic form is always predominant.
Because of the benzoconden-
sation of the azine ring on
going from the pyridine to
the quinoline and acridine de-
rivative, the contribution of
the quinoid limit form increas-
es in all of the solvents on
going from 1c to 3c.
A detailed quantitative in-
vestigation on the solvent-de-
pendent ICT was performed
by computation of natural
charges (natural population
analysis (NPA))[
39, 40]
in the
gas phase and in different
solvents, site by site (Figure 6).
To focus attention better on
acceptor± donor properties of
the terminal moieties and on
the extent of ICT, Figure 6
shows natural charges grouped
Figure 4. Optimized structures of chromophores 1c ± 3c in the gas phase.
by acceptor (azinium), donor
results (number of imaginary frequencies) for the computed
structures. Frequency analysis found one imaginary frequency
for the pyridine derivative 1c, corresponding to the N-methyl
rotation. Two and three imaginary frequencies were found for
the quinoline and acridine derivatives 2c and 3c, respectively.
In both cases the lowest imaginary frequency is due to the
(thienyldicyanomethanido), and bridging ethene units. In this
way, we wanted quantitatively to probe the contribution of the
two limit formulas to the description of the ground state in
terms of natural charges residing on the two end groups.
Namely, it was found that the negative total charge on the
donor group (dicyanomethanido moiety) increases on going
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1999

A. Abbotto, G. A. Pagani et al.
FULL PAPER
shared between the two side
units, with no significant in-
volvement of the central dou-
ble bond, which remains sub-
stantially neutral.
The annelation effect, expe-
rienced on going from the
pyridine 1c to the quinoline
2
3
c and the acridine derivative
c, operates, in terms of
charge distribution, in the di-
rection of increasing quinoid
character. Indeed, annelation
promotes charge transfer from
the donor to the acceptor site,
as can be seen from the de-
crease in negative and positive
total charge on the donor and
acceptor moiety, respectively.
A levelling off of annelation-
induced charge variation is
observed at higher polarities,
as a consequence of a predom-
inant frozen (less flexible)
zwitterionic structure. These
results are discussed, in con-
junction with all other studies,
in the following section.
Ab initio computed solvent-
dependent first hyperpolariza-
bilities: Computed static (zero
frequency) first hyperpolariz-
ability b values and mb values
0
0
are collected in Table 4 for a
number of different media,
together with experimentally
ascertained values. Computed
b0 values were obtained at the
HF level by the coupled per-
turbed Hartree ± Fock (CPHF)
method,[ by use of the rou-
tine available in the Gaussian
program. From the calculated
values of the ten independent
components bijk (i,j,k ˆ x,y,z)
of the first hyperpolarizability
tensor, we obtained bvec and
btot values from Equa-
41]
[42]
tions (8)–(10)
,
through
Figure 5. Optimized structures of chromophores 1c ± 3c in DMSO.
the adoption of Kleinman
symmetry relations.[
43]
from the poorly polar dioxane to the highly polar DMF and
DMSO, consistently with an increased contribution of the ZW
form in the latter two solvents. Total charges correspondingly
became more positive on the acceptor group (azine moiety).
It should be noted that the charge is almost symmetrically
b
b
i
ˆ biii‡(1/3)S i= k(bikk‡bkik‡bkki
)
i,k ˆ x,y,z
(8)
(9)
vec ˆ (m
x
b
x
‡m
y
b
y
‡m
z
b
z
)/ j m j
2
2
2
1/2
btot ˆ (b ‡b ‡b †
x
y
z
(10)
2000
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
Figure 6. Computed solvent-dependent natural charge distributions of compounds 1c ± 3c. Acceptor (left bar), ethenyl unit (mid bar), and donor (right bar)
group charges are shown for each compound in each medium.
The bvec values are identical to the btot values when the
intramolecular charge transfer is collinear to the molecular
ground-state dipole moment. We have checked that the two b
values are very similar to each other and consequently only
the bvec values are presented. According to the ™phenomeno-
logical∫ convention used in EFISH experimental measure-
ments,[ calculated b values were multiplied by a factor of 0.5
to make them directly comparable with experimentally
determined values.
state dipole moments, affecting optical[18] absorption in
solution.
15N and 13C NMR experiments were performed in DMSO.
Table 5 lists the sum of carbon chemical shifts on the ethenyl
(CˆC) and thienyldicyanomethanido (T(CN) ) fragments and
2
in the whole donor-spacer unit (Tot.). Chemical shifts can be
44]
p
converted into p-electron densities (Dq ) on the carbon
framework by use of Equation (1). Values are reported as
variations with respect to the system 1, taken as a reference.
Note that only the donor-spacer fragment was considered in
this computation, since it remains unchanged along the whole
series of chromophores. Therefore, on moving from the
pyridine-based systems 1a/b to the corresponding quinoline
Discussion
p
p
2
a/b and acridine 3a derivatives, total q variations (Dq Tot)
Structural control of ICT: The extent of the solvatochromic
response of these dyes in solution (Table 1) is the first
indication of their high molecular polarizability. The Dlmax
can be ascribed only to intramolecular charge transfer
Table 5. Sum (Sd) of the 13C shifts [ppm] of 1 ± 3 on the ethenyl (CˆC) and
(DMSO ! dioxane) sign inversion on going from the pyr-
2
thienyldicyanomethanido (T(CN) ) molecular fragments and variation of
p
idine- to the acridine-based systems confirms the influence of
molecular design at the acceptor level on fine control over the
p-electron distribution on the whole chromophore structure.
In this case we manipulate the aromaticity of the pyridinium/
annelated pyridinium rings to control the p-electron-with-
drawal strength of the end-heterocycle ring acceptor and
therefore the extent of the charge transfer from the donor
group. In fact, in the azine series (pyridine, quinoline and
acridine)–as in the benzenoid one (benzene, naphthalene
and anthracene)–the aromaticity of the central ring decreas-
es through annelation. The result is a modulation of ground
the local p-electron densities Dq on going from 1a/b to the corresponding
a/b and 3a systems.
2
1
3
Dqp[a]
Compound Sd( C NMR)
CˆC
T(CN)
Tot
CˆC T(CN)
Tot
2
2
1
1
2
2b
3
a
b
a
247.9 814.7
247.3 816.4
244.4 834.3
244.0 835.1
252.9 834.3
1062.6
1078.7
1063.7
1079.1
1087.1
0.00
0.00
0.02
0.02
À 0.03
0.00
0.00
À 0.12
À 0.12
À 0.12
0.00
0.00
À 0.10
À 0.10
À 0.15
a
[
a] Calculated according to Equation (1); negative values correspond to a
decrease of p-electron density
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2001

A. Abbotto, G. A. Pagani et al.
FULL PAPER
promoted by pyridine annelation. From this study two
important conclusions can be drawn. First, N-protonated (a
derivatives) and N-alkylated (b derivatives) systems exhibit
identical Dq values (1a ! 2a and 1b ! 2b). Thus, p-molec-
those presented here. Both 13C NMR and computation have
shown the rigidity of the p-structure in highly polar solvents
such as DMSO. However, things change dramatically when
solvent polarity starts to decrease. A first hint of this behavior
p
1
ular polarization is substantially independent of H/alkyl s
substitution at the azine nitrogen atom and makes it
legitimate to compare the NLO properties of differently s-
functionalized derivatives at the azine nitrogen. Second, on
going from the pyridine to the quinoline and then to the
acridine dye there is a slight decrease (À0.10 and À0.15,
respectively) in the p-electron density on the donor site. This
experimental evidence is nicely in
is the drastic change in the H NMR chemical shifts and
H-CˆC-H coupling constants in 1 ± 3 as a function of solvent
polarity.
An interesting complementary way to visualize the strongly
solvent-dependent molecular properties is to plot the com-
puted ground state dipole moment m values against the
solvent×s dielectric constant e (Figure 7a). Two regions of e
accord with the greater accepting
properties of acridinium and qui-
nolinium relative to the pyridini-
um group. In addition, this result
is even more important in view of
the fact that DMSO is certainly
not the best solvent in which to
evaluate this trend, since its high
polarity pushes all the three chro-
mophores toward their zwitter-
ionic limit. Under such conditions,
the p-framework along the ICT
direction is almost identical in all
of the chromophores, and so the
charge distribution on the donor
site is similar in the three species.
Unfortunately, poor chromo-
phore solubility in solvents of
low dielectric constant, in which
higher overall sensitivities would
be expected, prevented us from
performing multinuclear NMR
experiments in CHCl and diox-
3
ane.
The above behavior is in full
agreement with the predications
of ab initio computation. From
the computed absolute p-electron
densities in DMSO, it is possible
p
to derive the Dq values for the -
(
CˆC)-thienyldicyanomethanido
fragment, if molecule 1c is taken
as the reference system. The com-
p
puted Dq values are indeed very
p
small (Dq < 0.004 electrons) and
differences in the three systems
are negligible. In short, the high
solvent polarity of DMSO freezes
the chromophore p-electron dis-
tribution in an almost completely
charge-separated (ZW) structure.
Environmental control of ICT:
Medium polarity proves to be a
very efficient tool with which to
steer the molecular response for
highly polarizable systems such as
Figure 7. Plots of computed m (a) and static mb
0
(b) against dielectric constant e of the surrounding medium,
and mb
0
against (e À 1)/(2e‡1) (c) for 1c (^), 2c (*), and 3c (~).
2002
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
values with different effects are then identified: 1) a large
range (7 < e < 47) of dielectric constants in which m values of
chromophores 1 ± 3 are very large and do not vary signifi-
cantly either with structure or with solvent polarity; this is in
agreement with predominantly zwitterionic character in the
systems, and 2) a much smaller e range (1 < e < 7) in which the
dipole moments decrease dramatically; here the contribution
of the quinoid structure increases for all the systems, in
particular for the acridine-based derivative 3a.
ble contribution of the Q and ZW limit formulas to dyes×
descriptions, with an increased/decreased contribution of the
ZW/Q structure in higher polar solvents. This result is fully
supported experimentally by the J behavior of this bond in
such solvents.
3
Nonlinear optical molecular response: Table 4 collects both
experimentally determined (EFISH and EOAM) and com-
puted m, b and mb values for 1 ± 3 in different media. The
0
0
This very large sensitivity of the p-structure in the 1 < e < 7
range is confirmed by the solvatochromic response of 1b in
bulk. Host ± guest films of 1b were prepared with matrices of
different dielectric constants. Loading of 1b was kept very low
excellent agreement found between the computed and the
available experimental mb results, also taking account of the
0
implicit errors arising from measurements and computational
approximations, is remarkable and allow us to view the
computed values in solvents for which NLO experimental
data are not accessible as realistic.
The overall picture that emerges from the combination of
experimentally determined and computational investigation
leads to important and, in some ways, amazing and unexpect-
ed conclusions.
(
<0.5 wt%) to prevent formation of aggregates and micro-
domains and to avoid substantial changes in e on going from
the pure matrix to the blend. In particular, we chose host
polymer matrices with low dielectric constants, to match the
region in which the molecular polarization of dyes 1 ± 3 is
significantly affected. Some of these templates–such as
PMMA, polyimides, and sol ± gel glasses–have been exten-
sively used in the preparation of active electrooptic materi-
First, chromophore NLO b responses are low and positive
0
in media of very low dielectric constant (gas phase) while
being low and negative in solvents of high dielectric constant
[
45]
als. The solid-state solvatochromic data for 1b are collected
in Table 6. The solvatochromic shift of dye 1b from the low-
polar polystyrene (e ˆ 2.5 ± 2.6; Table 6, entry 8) to the high-
(DMF). Note that mb (DMF) is indeed much higher than mb0
0
(gas) but that most of the contribution comes from the higher
m values. Indeed, these results are to be expected because both
limit situations are characterized by a molecular p-structure
frozen in either the quinoid or the zwitterionic limit formula.
In accordance with the two-state model, these situations lead
to the lowest second-order NLO activity. In this context (at
these extreme situations) annelation of the pyridine acceptor
has very little influence on the extent of the charge transfer
and the molecular properties are environmentally controlled.
These data are in line with the solvatochromic and NMR
results.
polar siliceous matrix (e > 6; entry 1) is extremely large (Dl
max
as high as À132 nm). This surprisingly high solvatochromic
shift, even larger than that found in solution (Dlmax ˆ À
89 nm), is probably due to the peculiar nature of the siliceous
matrix, in which multiple and strong intermolecular inter-
actions occur between the highly polar dye and the pore
surface, exposing hydroxy functionalities.[
46]
Table 6. Solvatochromic data (lmax [nm] of the charge-transfer band) for
the chromophore 1b in selected film matrices.
Second, completely different behavior is observed when
media possessing suitable dielectric constants (2 < e < 7), such
as chloroform and dioxane solvents, are considered. We have
already shown how they greatly affect the molecular dipole
moments. In this case the description of the dyes× ground or
excited states is a ™loose∫ combination of both Q and ZW
limit formulas, which can easily be perturbed. In this situation,
slight structural modifications and/or minute e variations
greatly affect the nonlinear optical response. In dioxane,
Entry
Matrix
e
lmax [nm]
1
2
3
4
5
6
7
8
siliceous
> 6
592
616
628
648
676
680
702
724
[
a]
poly(p-hydroxystyrene)
poly(ethyleneglycol)
3.6 ± 4.0
[
b]
[a]
1b
PMMA
3.2 ± 3.5
3.1 ± 3.3
2.7 ± 2.9
2.5 ± 2.6
polymaleimide
poly(vinylbenzyl chloride)
polystyrene
[
a] Unknown. [b] As free standing neat film obtained by spin coating.
simply through annelation of the pyridine ring, mb changes
0
from negative (ca. À1200 for 1b) to positive (ca. ‡800 for 2b)
Finally, Table 7 shows computed ethenyl bond lengths of
to highly positive (ca. ‡2200 for 3a). Moving from dioxane to
the central spacer unit in dioxane, chloroform, and DMSO for
compounds 1 ± 3. A progressive elongation of this bond in the
series of chromophores is observed when the solvent polarity
is concomitantly increased. This data demonstrate the varia-
chloroform (De ꢁ 2.6) the nonlinear optical figure of merit mb
0
increases by a factor of ten. However, to stress the overall
sensitivity of 1 ± 3 second-order NLO responses with medium
polarity, Figure 7b and Figure S11 (in the Supporting Infor-
mation) show plots of mb and b , respectively, against e.
0
0
Table 7. Solvent-dependent computed CÀC bond lengths [ä] of the
central ethenylic unit in systems 1 ± 3.[
Alternatively, the permittivity dependence of mb in terms of
0
a]
the function (e À 1)/(2e‡1), which controls the magnitude of
the reaction field,[ is illustrated in Figure 7c.
8d]
Compound
Dioxane
CHCl
3
DMSO
These graphs clearly show the highly strategic role of the
polarity of the surrounding media in determining the NLO
molecular response. It should be noted how, upon changing
1
2
3
c
c
c
1.404
1.417
1.435
1.348
1.347
1.349
1.337
1.337
1.340
the value of the dielectric constant e, mb values can not only
[
a] Calculated at the RHF/6 ± 31G*//RHF/6 ± 31G* level; C
s
symmetry
0
point group.
change sign, but more importantly, vary by orders of
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2003

A. Abbotto, G. A. Pagani et al.
FULL PAPER
magnitude. For instance, jmb j of acridine derivatives is
Experimental Section
0
À44
predicted to change from zero to 2.2 Â 10 esu (a value that
General: 1H NMR spectra were recorded at 300 and 500 MHz. UV/Vis
spectra were recorded with a Varian Cary 1E spectrophotometer. Mass
spectra were determined at an ionizing voltage of 70 eV. Anhydrous
solvents were prepared by continuous distillation over sodium sand, in the
presence of benzophenone and under nitrogen or argon, until the blue
color of sodium ketyl was permanent. Anhydrous N,N-dimethylformamide
would be one of the largest ever reported in the literature)
within a range of only 0.4 e units (from 3.4 to 3.8)!
(
DMF) was supplied by Fluka. Acetone was dried over Na
days. Diisopropylamine was heated at reflux over CaH for 4 h and distilled
SO (4 h). Melting
2 4
SO for a few
2
Conclusion
under nitrogen prior to use. Extracts were dried over Na
points are uncorrected.
2
4
In this study a combination of organic synthesis, multinuclear
Electrooptical absorption measurements (EOAM): The EOAM character-
1
13
15
(
H, C, N) NMR study, optical absorption spectroscopy,
ization of the NLO-phores was carried out by previously described
methods.[
16, 31]
Electrooptical absorption spectra were recorded in dioxane,
nonlinear optical measurements, and ab initio computations
have been applied to the design, preparation, and character-
ization of a novel family of push ± pull chromophores 1–3, in
which fine control over molecular properties is based on
structural and media effects.
The chromophore design of the electron-poor central
pyridine ring as a p-acceptor unit conjugated with the
dicyanomethanido donor, based on the different aromaticity
achieved through annelation, provides a set of structures that
exhibits either similar or differentiated properties, depending
on the surrounding media. In fact, the ground and excited
states of these systems can be described as linear combina-
tions of two limit formulas, ZW and Q, the relative weights of
which significantly affect the molecular performance in terms
of linear and nonlinear optical responses.
This study demonstrates that dyes 1 ± 3 are undoubtedly
highly polarizable structures. On the basis of multinuclear
NMR analysis and solvatochromic data, all of the systems 1 ± 3
are primarily zwitterionic over a large range of high dielectric
constants, while the quinoid contribution increases as e
decreases. Eventually, they are completely quinoid for e < 2.
These results are confirmed both by nonlinear optical
measurements in dioxane, chloroform and DMF and by
computations. Bond-localized quinoid arrangements must be
a realistic structural prerequisite: their evolution into charge-
separated structures by ICT must be made possible by
favorable interaction with polar media. Computations are
invaluable in suggesting proper environments for adjusting
the ICT and should allow improvement, even of orders of
magnitude, of the first hyperpolarizability performance of
highly polarizable NLO-phores.
In short, dyes 1 ± 3 give access to large second-order
molecular NLO activities, which can only be reached, how-
ever, if medium polarity effects are carefully taken into
consideration. An excellent NLO response in solution might
vanish when the active chromophore is dispersed in a matrix
with unsuitable dielectric properties. Vice versa, a mediocre
NLO response for a newly designed chromophore could be
the result of an unfortunate choice of solvent, as dictated by
instrumental availability or solubility properties; the same
chromophore might prove to perform very well under
appropriately chosen environmental (solvent, matrix) condi-
tions. The commonly established procedure for NLO com-
pounds to report b values in one solvent only, may in certain
cases be insufficient to draw definite conclusions on the
overall chromophore performance and the prospect of differ-
ent design strategies.
which was purified and carefully dried prior to use by distillation from Na/
K under argon. Supplementary optical absorption spectra required for
evaluation of the EOAM spectra were determined with a Perkin-Elmer
Lambda 900 spectrophotometer.
Materials: 4-Pyridinecarbaldehyde, 4-quinolinecarbaldehyde, and 2-bro-
mothiophene were commercially available. 1-(4-Pyridyl)-2-{2-[5-(dicyano-
methanido)thienyl]}ethene sodium salt and decyl triflate were prepared by
known procedures.[
6]
Film preparation: A solution of 1b (0.1 mg) in THF (0.1 mL) was added to
a solution obtained from polymeric samples (entries 2, 3, 5-±8, Table 6,
ꢁ
15 mg) dissolved by sonication in either THF or CHCl
mixture was sonicated for 2 min and then spin-coated (1500 ± 2000 rpm,
0 s) onto sodium lime glass substrates (Fisher). A 1b film (entry 4,
Table 6) was prepared by spin-coating a solution of 1b (5 mg) in CHCl
(1 mL). Siliceous film (entry 1, Table 6) was prepared from a mixture of 1b
0.2 mg) and Si Cl (0.01 mL) in THF (1 mL). After spin-coating, the
3
(1 mL). The
1
3
(
3
O
3
8
film was dried under vacuum for 10 min and rinsed twice with diluted
ammonium hydroxide solution (2%) and finally with water. UV/Vis
spectra were immediately recorded after film preparation.
1
-(4-1H-Pyridinium)-2-{2-[5-(dicyanomethanido)thienyl]}ethene
(1a):
Aqueous HCl (10%) was added dropwise to a suspension of 1-(4-
pyridyl)-2-{2-[5-(dicyanomethanido)thienyl]}ethene sodium salt (0.185 g,
0.68 mmol) in H O (5 mL) until pH 1 was reached. After the mixture had
2
been stirred at room temperature for 30 min, the blue precipitate was
collected, washed with water, and dried under vacuum at 808C to afford the
pure product (0.140 g, 0.56 mmol, 82.4%): m.p. 234 ± 2358C(DMF);
1
H NMR ([D
6
]DMSO, 258C, TMS): d ˆ 14.2 (broad, 1H), 8.40 (d,
3
3
J(H,H) ˆ 7.0 Hz, 2H), 7.97 (d, J(H,H) ˆ 15.1 Hz, 1H), 7.75 (d, 2H), 7.19
3
(
d, J(H,H) ˆ 4.2 Hz, 1H), 6.34 (d, 1H), 6.25 (d, 1H); elemental analysis
calcd (%) for C14
H
9
N
3
2
S ¥ 0.5H O (260.3): C 64.58, H 3.87, N 16.14; found: C
6
5.01, H 4.02, N 15.61.
1
-(4-1-Decylpyridinium)-2-[2-[5-(dicyanomethanido)thienyl]ethene (1b):
Decyl triflate (0.21 g, 0.74 mmol) in dry acetone (1 mL) was added
dropwise to a suspension of 1-(4-pyridyl)-2-{2-[5-(dicyanomethanido)thie-
nyl]}ethene sodium salt (0.19 g, 0.70 mmol) in the same solvent (4 mL). The
color of the solution changed immediately from orange to blue. After the
mixture had been stirred overnight at room temperature, the precipitate
was collected and washed with EtOH and water to give the practically pure
product (0.141 g, 0.36 mmol, 51.4%) as a blue solid: m.p. 225 ± 2268C
1
3
(
EtOH); H NMR ([D
6
]DMSO, 258C, TMS): d ˆ 8.45 (d, J(H,H) ˆ 7.1 Hz,
3
3
2H), 7.99 (d, J(H,H) ˆ 15.1 Hz, 1H), 7.75 (d, 2H), 7.20 (d, J(H,H) ˆ 4.2,
3
1
2
H), 6.35 (d, 1H), 6.25 (d, 1H), 4.25 (t, J(H,H) ˆ 7.2 Hz, 2H), 1.81 (m,
3
H), 1.18 ± 1.35 (m, 14H), 0.84 ppm (t, J(H,H) ˆ 6.9 Hz, 3H); elemental
analysis calcd (%) for C24
29 3 2
H N S ¥ 0.5H O (400.6): C 71.95, H 7.56, N 10.49;
found: C 71.95, H 7.16, N 10.07.
1
-(4-1H-Quinolinium)-2-[2-[5-(dicyanomethanido)thienyl]ethene
(2a):
Aqueous HCl (10%) was added dropwise to a suspension of 1-(4-
quinolyl)-2-{2-[5-(dicyanomethanido)thienyl]}ethene sodium salt (0.21 g,
0.645 mmol) in H O (7 mL) until pH 1 was reached. After the mixture had
2
been stirred for 30 min at room temperature, the blue precipitate was
collected, washed with water and dried under vacuum to afford the pure
1
product (0.16 g, 0.52 mmol, 82.3%): m.p. >2408C; H NMR ([D
6
]DMSO,
3
2
58C, TMS): d ˆ 14.10 (s, 1H), 8.62 (d, J(H,H) ˆ 8.6 Hz, 1H), 8.51 (d,
3
3
3
J(H,H) ˆ 6.5 Hz, 1H), 8.26 (d, J(H,H) ˆ 14.3 Hz, 1H), 7.93 (t, J(H,H) ˆ
2004
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
3
7
(
.8 Hz, 1H), 7.86 (d, J(H,H) ˆ 8.4 Hz, 1H), 7.81 (d, 1H), 7.68 (t, 1H), 7.46
5.10 mmol) in H
2
O (50 mL) until pH 1 was reached. After the mixture
d, 3J(H,H) ˆ 4.2 Hz, 1H), 6.99 (d, 1H), 6.45 ppm (d, 1H); elemental
had been stirred at room temperature for 30 min, the blue precipitate
was collected, washed with water, and dried under vacuum at 808C to
analysis calcd (%) for C18
H
11
N
3
S ¥ 2H
2
O (337.4): C 64.07, H 4.49, N 12.45;
found: C 64.41, H 4.27, N 12.88.
afford the pure product (1.49 g, 4.22 mmol, 82.9%). m.p. >3008C(DMF);
1
H NMR ([D
6
]DMSO, 258C, TMS): d ˆ 12.2 (broad, 1H), 8.27 (d,
1
-(4-1-Decylquinolinium)-2-{2-[5-(dicyanomethanido)thienyl]}ethene
3
3
3
J(H,H) ˆ 8.3 Hz, 2H), 8.01 (d, J(H,H) ˆ 13.5 Hz, 1H), 7.84 (d,
(
2b): Decyl triflate (0.49 g, 1.71 mmol) in dry acetonitrile (30 mL) was
added dropwise to a suspension of 1-(4-quinolyl)-2-{2-[5-(dicyanometha-
nido)thienyl]}ethene sodium salt (0.50 g, 1.55 mmol) and K CO (0.64 g,
.65 mmol) in the same solvent (30 mL). The color of the solution changed
3
3
J(H,H) ˆ 4.9 Hz, 1H), 7.71 (t, J(H,H) ˆ 7.6 Hz, 2H), 7.52 (d, J(H,H) ˆ
3
3
8
6
.2 Hz, 2H), 7.35 (t, J(H,H) ˆ 7.6 Hz, 2H), 7.16 (d, J(H,H) ˆ 13.7 Hz, 1H),
2
3
3
.88 ppm (d, J(H,H) ˆ 4.9 Hz, 1H); elemental analysis calcd (%) for
4
22 13 3
C H N S (351.4): C 75.19, H 3.73, N 11.96, S 9.12; found: C 75.03, H 3.81,
immediately from orange to green. After the mixture had been heated at
reflux for 2 h, the solvent was evaporated and the residue was chromato-
graphed on alumina (neutral, acetone/CHCl
N 11.51, S 8.60.
1
-(9-Acridyl)-2-{2-[5-(dicyanomethanido)thienyl]}ethene sodium salt (6):
3
1:1) to afford the pure
1
A solution of malononitrile (0.15 g, 2.2 mmol) in 1,2-dimethoxyethane
(4 mL) was added to an ice-cooled suspension of sodium hydride (0,18 g,
60% in mineral oil, 4.5 mmol) in the same solvent (10 mL), and the mixture
was stirred at room temperature under nitrogen for 1 h. 1-(9-Acridyl)-2-(5-
bromothien-2-yl)ethene (0.41 g, 1.1 mmol) and tetrakis(triphenylphosphi-
ne)palladium(0) (0.14 g, 0.1 mmol) were added, and the reaction mixture
was stirred overnight at room temperature. The solvent was remove under
reduced pressure, and the solid was taken up with dry benzene (25 mL) and
product (0.09 g, 0.20 mmol, 13.1%): m.p. 156 ± 1588C;
H NMR
3
(
[D
6
]DMSO, 258C, TMS): d ˆ 8.69 (d, J(H,H) ˆ 8.8 Hz, 1H), 8.55 (d,
3
3
J(H,H) ˆ 7.05 Hz, 1H), 8.29 (d, J(H,H) ˆ 14.2 Hz, 1H), 8.11 (d, 1H), 7.98
3
3
(
4
2
t, J(H,H) ˆ 7.4 Hz, 1H), 7.77 (d, 1H), 7.70 (t, 1H), 7.49 (d, J(H,H) ˆ
3
.3 Hz, 1H), 7.01 (d, 1H), 6.50 (d, 1H), 4.57 (t, J(H,H) ˆ 7.22, 1H), 1.82 (m,
H), 1.40 ± 1.10 (m, 14H), 0.84 ppm (t, J(H,H) ˆ 6.5, 3H); elemental
analysis calcd (%) for C28
C 76.18, H 6.98, N 9.38.
31 3
H N S (441.7): C 76.27, H 7.50, N 9.20; found:
washed with water (5 mL) to give the product as a purple solid (0.41 g;
1
-(4-Quinolyl)-2-[5-(dicyanomethanido)thien-2-yl]ethene sodium salt
1
1
8
.1 mmol; 97%). m.p. >2408C; H NMR ([D
6
]DMSO, 258C, TMS): d ˆ
(
5): Malononitrile (0.35 g, 5.30 mmol) was added portionwise to an ice-
3
3
.39 (d, J(H,H) ˆ 8.7 Hz, 2H), 8.09 (d, J(H,H) ˆ 8.6 Hz, 2H), 7.81 (t,
cooled suspension of sodium hydride (0.25 g, 60% in oil, 6.25 mmol) in
anhydrous 1,2-dimethoxyethane (27 mL), and the mixture was stirred
under nitrogen atmosphere at room temperature for 30 min. 1-(4-Quinol-
yl)-2-(5-bromothien-2-yl)ethene (0.64 g, 2.22 mmol) and tetrakis(triphe-
nylphosphine)palladium(0) (0.25 g, 0.22 mmol) were added to the above
solution, and the mixture was stirred overnight at room temperature. The
solvent was removed under reduced pressure and the resulting gummy
solid was taken up with benzene (3 Â 10 mL) and hexane (30 mL), and
3
3
J(H,H) ˆ 7.6 Hz, 2H), 7.58 (t,
3
J(H,H) ˆ 7.7 Hz, 2H), 7.26 (s, 2H), 7.03 (d,
3
J(H,H) ˆ 3.9 Hz, 1H), 6.10 ppm (d, J(H,H) ˆ 3.9 Hz, 1H).
1
-(9-Acridyl)-2-(5-bromothien-2-yl)ethene (8): A mixture of sodium ace-
tate (8.17 g, 99.6 mmol), acetic acid (98%, 28 mL), and 9-methylacridine
1.63 g, 8.5 mmol) was heated at 808C for few minutes. 5-Bromo-2-
(
thiophenecarbaldehyde (3.22 g, 17.0 mmol) was then added, and the
reaction mixture was heated at reflux for 8 h. After cooling, the reaction
mixture was poured into iced water (140 mL), and ammonium hydroxide
finally washed with H
solid (0.67 g, 2.07 mmol, 93.2%), which was used in the next step without
2
O (2 Â 8 mL) to give the product as a red-brown
(
25%) was added until pH 14 was reached. The basic solution was
extracted with ethyl acetate (5 Â 65 mL) and the organic phase was washed
twice with water and dried over Na SO , and the solvent was evaporated to
1
further purification: m.p. >2408C; H NMR ([D
6
]DMSO, 258C, TMS): d ˆ
.72 (d, 3J(H,H) ˆ 4.8 Hz, 1H), 8.33 (d, J(H,H) ˆ 8.3 Hz, 1H), 7.95 (d,
3
2
4
8
give the crude product (4.05 g), which was purified by column chromatog-
raphy on silica gel (dichloromethane/diethyl ether 20:1). Two isomers were
3
3
J(H,H) ˆ 8.5 Hz, 1H), 7.72 (t, 1H), 7.68 (d, 1H), 7.66 (d, J(H,H) ˆ 11.8 Hz,
3
1
1
H), 7.57 (t, 1H), 7.04 (d, 1H), 7.03 (d, 1H), 6.09 ppm (d, J(H,H) ˆ 3.8 Hz,
1
isolated: cis isomer (0.21 g, 0.6 mmol, 6.6%): m.p. 134 ± 1358C; H NMR
H).
3
(
CDCl
3
,
258C, TMS): d ˆ 8.28 (d, J(H,H) ˆ 8.8 Hz, 2H), 8.15 (d,
3
1
-(4-Quinolyl)-2-(5-bromothien-2-yl)ethene (7): A mixture of 5-bromo-2-
J(H,H) ˆ 8.6 Hz, 2H), 7.76 ± 7.82 (m, 2H), 7.45 ± 7.53 (m, 2H), 7.28 (d,
[
47]
3
3
3
chloromethylthiophene
(4.60 g, 21.76 mmol) and triethylphosphite
J(H,H) ˆ 12.1 Hz, 1H), 6.99 (d, J(H,H) ˆ 12.1 Hz, 1H), 6.70 (d,
3
(
3.60 g, 21.76 mmol) was heated at reflux at 1108C for 6 h. The resulting
J(H,H) ˆ 3.9 Hz, 1H), 6.64 ppm (d, J(H,H) ˆ 3.9 Hz, 1H); MS (EI,
‡
oil was heated at 2008C and at 1 mmHg in a Kugelrohr apparatus to distil
low-boiling components, leaving the diethyl 5-bromothien-2-ylmethane-
phosphonate (6.02 g, 19.22 mmol, 88%) as a light red, oily residue; this was
used for subsequent steps without further purification. H NMR (CDCl
70 eV): m/z (%): 367 (52) [M] ; trans isomer (1.10 g, 3.0 mmol, 35.5%):
1
m.p. 197 ± 2008C (AcOEt); H NMR (CDCl
3
, 258C, TMS) d ˆ 8.27 (d,
3
3
J(H,H) ˆ 8.8 Hz, 2H), 8.24 (d, J(H,H) ˆ 8.7 Hz, 2H), 7.74 ± 7.82 (m, 2H),
1
3
3
3
,
7.63 (d, J(H,H) ˆ 16.3 Hz, 1H), 7.51 ± 7.58 (m, 2H), 7.07 (d, J(H,H) ˆ
3 3
3
4
2
58C, TMS): d ˆ 6.88 (d, J(H,H) ˆ 3.7 Hz, 1H), 6.71 (tt, J(P,H) ˆ 1.0 Hz,
16.3 Hz, 1H), 7.05 (d, J(H,H) ˆ 3.9 Hz, 1H), 6.94 ppm (d, J(H,H) ˆ
3
3
‡
J(H,H)
ˆ
3.7 Hz, 1H), 4.07 (dq, J(P,H) ˆ 8.22 Hz, 6H), 3.26 (dd,
3.9 Hz, 1H); MS (EI, 70 eV): m/z (%): 367 (100) [M] ; elemental analysis
2
3
J(P,H) ˆ 20.7 Hz, 2H), 1.28 ppm (t, J(H,H) ˆ 7.1 Hz, 9H).
calcd (%) for C19
H 3.26, N 3.69.
H12BrNS (366.3): C 62.30, H 3.30, N 3.82; found: C 61.80,
A suspension of sodium hydride in oil (60% by weight; 0.16 g, correspond-
ing to 0.27 g, 6.96 mmol) was thoroughly washed with anhydrous hexane
and then suspended in anhydrous THF (15 mL). A solution of diethyl
9
-Methylacridine (9): A mixture of diphenylamine (13.25 g, 78.3 mmol),
acetic acid (60%, 8 mL), and ZnCl (22.91 g, 168.1 mmol) was heated at
708C for 17 h. After cooling, the solid was taken up with warm water (2 Â
0 mL), which was discarded, and then extracted with hot water (4 Â
0 mL). The aqueous phases were combined, NH OH (30%) was added,
2
1
5
6
5
-bromothien-2-ylmethanephosphonate (1.99 g, 6.36 mmol) in THF (5 mL)
was added under nitrogen to this suspension, followed by the addition of a
solution of 4-quinolylcarbaldehyde (1.00 g, 6.36 mmol) in the same solvent
4
and the basic solution was extracted with ethyl acetate (3 Â 200 mL). The
(
5 mL). The mixture was cautiously heated on an oil-bath at 508C until the
organic phase was washed with water (100 mL) and dried over Na SO , and
2
4
evolution of hydrogen had ceased, and then at reflux for 2 h. The mixture
was poured onto ice (100 mL) and the aqueous phase was extracted with
ether (4 Â 50 mL). The ethereal phases were combined, washed with water
the solvent was evaporated to give the product as a yellow solid (4.01 g,
[
48]
1
2
2
2
0.8 mmol, 26.5%). m.p. 1148C (lit.
114 ± 1158C); H NMR (CDCl
3
,
3
3
58C, TMS): d ˆ 8.22 (d, J(H,H) ˆ 8.8 Hz, 2H), 8.19 (d, J(H,H) ˆ 8.6 Hz,
2 4
and dried over Na SO , and the solvent was finally evaporated to afford a
3
3
H), 7.71 (t, J(H,H) ˆ 7.4 Hz, 2H), 7.50 (t, J(H,H) ˆ 7.5 Hz, 2H), 3.06 ppm
dark oil (1.97 g). The pure product was obtained after column chromatog-
raphy (silica gel, ether) as a light yellow solid (1.041 g, 3.29 mmol, 51.7%):
(
s, 3H).
1
m.p. 113 ± 1148C (after sublimation); H NMR ([D
6
]DMSO, 258C, TMS):
3
3
d ˆ 8.87 (d, J(H,H) ˆ 4.7 Hz, 1H), 8.42 (d, J(H,H) ˆ 7.9 Hz, 1H), 8.03 (d,
3
3
J(H,H) ˆ 7.9 Hz, 1H), 7.81 (d, 1H), 7.79 (t, J(H,H) ˆ 6.9 Hz, 1H), 7.73 (s,
3
2
H), 7.66 (t, 1H), 7.32 (d, J(H,H) ˆ 3.8 Hz, 1H), 7.28 (d, 1H); elemental
Acknowledgements
analysis calcd (%) for C15
H10BrNS (316.2): C 56.98, H 3.19, N 4.43; found:
C 57.31, H 3.42, N.4.67
This work was supported in part by grants from the CNR-Progetto
Finalizzato MSTA II and the MURST National Project (Grant No.
9803191626). We thank Dr. G. Archetti for part of the computational
investigations and UV/Vis measurements.
1
-(4-10H-Acridinium)-2-{2-[5-(dicyanomethanido)thienyl]}ethene (3a):
Aqueous HCl (10%) was added dropwise to a suspension of 1-(9-
acridyl)-2-{2-[5-(dicyanomethanido)thienyl]}ethene sodium salt (1.91 g,
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2005

A. Abbotto, G. A. Pagani et al.
FULL PAPER
[
1] a) Polymers for Second-Order Nonlinear Optics (Eds.: G. A. Lindsay,
K. D. Singer), ACS Symposium Series 601, American Chemical
Society, Washington, DC, 1995; b) Molecular Nonlinear Optics (Ed.:
J. Zyss), Academic Press, New York, 1994; c) Optical Nonlinearities in
Chemistry, (Ed.: D. M. Burland), Chem. Rev. 1994, 94, 1 ± 278; d) P. N.
Prasad, D. J. Williams, Introduction to Nonlinear Optical Effects in
Molecules and Polymers, Wiley, New York, 1991.
Barzoukas, C.-T. Chen, M. Blanchard-Desce, S. R. Marder, J. W.
Perry, Inorg. Chim. Acta 1996, 242, 43 ± 49.
[20] C. Reichardt, Solvent Effects in Organic Chemistry, Verlag Chemie,
New York, 1979, Chapter 6.
[21] A preliminary investigation into the concentration-dependent UV/Vis
absorption of 1b and 2b in dioxane has shown that the normalized
absorption of the lowest energy band increases with decreasing
concentration, while that of the highest energy band decreases
correspondingly, supporting the hypothesis that the former band
could be attributable to the monomeric species. A vibronic structure
of the spectra at higher dilution was also observed. A detailed UV/Vis,
NMR and computational investigation into aggregates is under way
and will be reported in due course.
[
[
2] S. R. Marder, J. W. Perry, G. Bourhill, C. B. Gorman, B. G. Tiemann,
K. Mansour, Science 1993, 261, 186 ± 189.
3] a) R. Wortmann, C. Poga, R. J. Twieg, C. Geletneky, J. Chem. Phys.
1
996, 105, 10637 ± 10647; b) R. Wortmann, C. Glania, P. Kramer, K.
Lukaszuk, R. Matschiner, R. J. Twieg, F. You, Chem. Phys. 1999, 245,
07 ± 120.
1
[
4] a) M. Barzoukas, C. Runser, A. Fort, M. Blanchard-Desce, Chem.
Phys. Lett. 1996, 257, 531 ± 537; b) W. H. Thompson, M. Blanchard-
Desce, V. Alain, J. Muller, A. Fort, M. Barzoukas, J. T. Hynes, J. Phys.
Chem. 1999, 103, 3766 ± 3771; c) G. Bourhill, J. L. Bredas, L. T. Cheng,
S. R. Marder, F. Meyers, J. W. Perry, B. G. Tiemann, J. Am. Chem. Soc.
[22] F. W¸rthner, S. Yao, T. Debaerdemaeker, R. Wortmann, J. Am. Chem.
Soc. 2002, 124, 9431 ± 9447.
[23] a) S. Di Bella, G. Lanza, I. Fragal a¡ , S. Yitzchaik, M. A. Ratner, T. J.
Marks, J. Am. Chem. Soc. 1997, 119, 3003 ± 3006; b) J. Zheng, C. H.
Huang, Y. Y. Huang, D. G. Wu, T. X. Wei, A. C. Yu, X. S. Zhao, New J.
Chem. 2000, 24, 317 ± 321; c) J. L. Bredas, F. Meyers, Nonlinear Opt.
1991, 1, 119 ± 123.
1
994, 116, 2619 ± 2620; d) C. Maertens, C. Detrembleur, P. Dubois, R.
Jerome, C. Boutton, A. Persoons, T. Kogej, J. L. Bredas, Chem. Eur. J.
999, 5, 369 ± 380.
1
X
[24] We have previously defined the charge demand c of a substituent
[
5] a) S. R. Marder, C. B. Gorman, B. G. Tiemann, L. T. Cheng, J. Am.
Chem. Soc. 1993, 115, 3006 ± 3007; b) G. U. Bublitz, R. Ortiz, C.
Runser, A. Fort, M. Barzoukas, S. R. Marder, S. G. Boxer, J. Am.
Chem. Soc. 1997, 119, 2311 ± 2312.
group X as the fraction of p-charge withdrawn from an adjacent
charged sp -carbon atom, and have subsequently used charge
2
demands as a measure of the capacity of group X to delocalize
positive or negative charge:a) S. Bradamante, G. A. Pagani, J. Chem.
Soc. Perkin Trans. 2 1986, 1035 ± 1045; b) S. Bradamante, G. A. Pagani,
Pure Appl. Chem. 1989, 61, 709 ± 716; c) E. Barchiesi, S. Bradamante,
R. Ferraccioli, G. A. Pagani, J. Chem. Soc. Perkin Trans. 2 1990, 375 ±
383.
[
6] A. Abbotto, S. Bradamante, A. Facchetti, G. A. Pagani, J. Org. Chem.
1
997, 62, 5755 ± 5765.
[
7] a) A. Abbotto, S. Bradamante, A. Facchetti, G. A. Pagani, I. Ledoux,
J. Zyss, Mater. Res. Soc. Symp. Proc. 1998, 488, 819 ± 822; b) A.
Abbotto, S. Bradamante, A. Facchetti, G. A. Pagani, L. Yuan, P. N.
Prasad, Gazz. Chim. Ital. 1997, 127, 165 ± 166.
[25] a) A. Abbotto, S. Bradamante, G. A. Pagani, J. Org. Chem. 1993, 58,
449 ± 455; b) A. Abbotto, V. Alanzo, S. Bradamante, G. A. Pagani, J.
[
[
8] a) A. Painelli, Chem. Phys. 1999, 245, 185 ± 197; b) R. Cammi, B.
Mennucci, J. Tomasi, J. Am. Chem. Soc. 1998, 120, 8834 ± 8847;
c) I. D. L. Albert, T. J. Marks, M. A. Ratner, J. Phys. Chem. 1996, 100,
Chem. Soc. Perkin Trans. 2 1991, 481 ± 488; c) A. Abbotto, S.
Bradamante, G. A. Pagani, J. Org. Chem. 1993, 58, 444 ± 448; d) C.
Gatti, A. Ponti, A. Gamba, G. A. Pagani, J. Am. Chem. Soc. 1992 114,
8634 ± 8644.
9
714 ± 9725; d) R. Wortmann, D. Bishop, J. Chem. Phys. 1998, 108,
1
001 ± 1007.
[26] G. Scheibe, W. Seiffert, G. Hohlneicher, C. Jutz, H. J. Springer,
Tetrahedron Lett. 1966, 5053 ± 5059.
[27] R. Radeglia, S. D‰hne, J. Mol. Struct. 1970, 5, 399 ± 411.
[28] M. Blanchard-Desce, V. Alain, P. V. Bedworth, S. R. Marder, A. Fort,
C. Runser, M. Barzoukas, S. Lebus, R. Wortmann, Chem. Eur. J. 1997,
3, 1091 ± 1104.
9] a) C. Dehu, F. Meyers, E. Hendrickx, K. Clays, A. Persoons, S. R.
Marder, J. L. Bredas, J. Am. Chem. Soc. 1995, 117, 10127 ± 10128; b) C.
Runser, A. Fort, M. Barzoukas, C. Combellas, C. Suba, A. Thiebault,
R. Graff, J. P. Kintzinger, Chem. Phys. 1995, 193, 309 ± 319.
[
[
[
[
10] A. R. Katritzky in Handbook of Heterocyclic Chemistry, Pergamon
Press, Oxford, 1983.
11] S. Bradamante, A. Facchetti, G. A. Pagani, J. Phys. Org. Chem. 1997,
[29] A. Mishra, R. K. Behera, P. K. Behera, B. K. Mishra, G. B. Behera,
Chem. Rev. 2000, 100, 1973 ± 2011.
[30] F. Steybe, F. Effenberger, P. Kr‰mer, C. Glania, R. Wortmann, Chem.
Phys. 1997, 219, 317 ± 331.
1
0, 514 ± 524.
12] I. D. L. Albert, T. J. Marks, M. A. Ratner, J. Am. Chem. Soc. 1997, 119,
575 ± 6582.
13] a) C. W. Dirk, H. E. Katz, M. L. Schilling, L. A. King, Chem. Mater.
6
[31] a) S. Beckmann, K.-H. Etzbach, P. Kramer, K. Lucaszuk, A.
Matschiner, A. J. Schmidt, P. Schumacher, R. Sens, G. Seibold, R.
Wortmann, F. W¸rthner, Adv. Mater. 1999, 11, 536 ± 541; b) F.
W¸rthner, C. Thalacker, R. Matschiner, K. Lukaszuk, R. Wortmann,
Chem. Commun. 1998, 1739 ± 1740.
[32] a) F. W¸rthner, R. Wortmann, R. Matschiner, K. Lukaszuk, K.
Meerholz, Y. DeNardin, R. Bittner, C. Br‰uchle, R. Sens, Angew.
Chem. 1997, 109, 2933 ± 2936; Angew. Chem. Int. Ed. Engl. 1997, 36,
2765 ± 2768; b) K. Meerholz, Y. DeNardin, R. Bittner, R. Wortmann,
F. W¸rthner, Appl. Phys. Lett. 1998, 73, 4 ± 6; c) F. W¸rthner, C.
Legrand, E. Mecher, K. Meerholz, R. Wortmann, Polym. Mat. Sci.
Eng. 1999, 80, 252 ± 254; d) R. Wortmann, F. W¸rthner, A. Sautter, K.
Lukaszuk, R. Matschiner, K. Meerholz, SPIE Proceed. 1998, 3471,
41 ± 49.
[33] GAUSSIAN 98, Revision A.9, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakr-
zewski, J. A. Montgomery Jr., R. E. Stratmann, J. C. Burant, S.
Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O.
Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S.
Replogle, J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.
1
1
990, 2, 700 ± 705; b) K. Y. Wong, A. K.-Y. Jen, V. P. Rao, Phys. Rev. A
994, 49, 3077 ± 3080.
[
14] a) J. O. Morley, R. M. Morley, R. Docherty, M. H. Charlton, J. Am.
Chem. Soc. 1997, 119, 10192 ± 10202; b) J. Abe, Y. Shirai, N. Nemoto,
F. Miyata, Y. Nagase, J. Phys. Chem. B 1997, 101, 576 ± 582; c) M.
Szablewski, P. R. Thomas, A. Thornton, D. Bloor, G. H. Cross, J. M.
Cole, J. A. K. Howard, M. Malagoli, F. Meyers, J. L. Bredas, W.
Wenseleers, E. Goovaerts, J. Am. Chem. Soc. 1997, 119, 3144 ± 3154;
d) G. J. Ashwell, R. C. Hargreaves, C. E. Baldwin, G. S. Bahra, C. R.
Brown, Nature 1992, 357, 393 ± 395; e) A. Fort, C. Runser, M.
Barzoukas, C. Combellas, C. Suba, A. Thiebault, SPIE Proceed.
1
994, 2285, 2 ± 10.
15] A. Abbotto, S. Bradamante, G. A. Pagani, J. Org. Chem. 2001, 66,
883 ± 8892.
16] J. J. Wolff, R. Wortmann, Adv. Phys. Org. Chem. 1999, 32, 121 ±
17.
17] a) W. Liptay in Excited States, Vol. 1. Dipole Moments and Polar-
izabilities of Molecules in Excited Electronic States (Ed.: E. C. Lim),
Academic Press, New York, 1974, pp. 129 ± 229; b) R. Wortmann, K.
Elich, S. Lebus, W. Liptay, P. Borowicz, A. Grabowska, J. Phys. Chem.
[
[
[
8
2
1
992, 96, 9724 ± 9730.
[
[
18] J. March, Advanced Organic Chemistry, Wiley, New York, 1985, p. 41.
19] a) V. Alain, M. Blanchard-Desce, I. Ledoux-Rak, J. Zyss, Chem.
Commun. 2000, 5, 353 ± 354; b) F. W¸rthner, S. Yao, R. Wortmann,
Polym. Mater. Sci. Eng. 2000, 83, 186 ± 187; c) V. Alain, A. Fort, M.
2006
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 1991 ± 2007

Tuning Intramolecular Charge Transfer
1991±2007
[
34] Computations were partially carried out at the Cineca Supercomputer
Center in Bologna (Italy) on SGI Origin 3800 and IBM SP RS/6000
Power3 machines.
[38] Such a decision was based on semiempirical (PM3) computations
showing that the C
1
and C
s
geometries have very similar energies
À1
(difference is below 1 kcalmol ) and that deviation from planarity in
[
35] Full-geometry optimizations and energy calculations at the B3LYP/
1
the more stable C structure is not significant.
6
± 31G*//B3LYP/6 ± 31G* level of theory were carried out for
[39] A. E. Reed, R. B. Weinstock, F. Weinhold, J. Chem. Phys. 1985, 83,
735 ± 746.
compound 1c (C symmetry) in the gas phase and in dioxane, CHCl
s
3
and DMF. DFT-computed dipole moment values are included in
Table 4.
[40] Gaussian NBO. Version 3.1 (as incorporated in the GAUSSIAN 98
package).
[
[
36] a) L. Onsager, J. Am. Chem. Soc. 1936, 58, 1486 ± 1493; b) M. W.
Wong, M. J. Frisch, K. B. Wiberg, J. Am. Chem. Soc. 1991, 113, 4776 ±
[41] a) J. Gerratt, I. M. Mills, J. Chem. Phys. 1968, 49, 1730 ± 1739;
b) G. J. B. Hurst, M. Dupuis, E. Clementi, J. Chem. Phys. 1988, 89,
385 ± 395; c) H. Sekino, R. J. Bartlett, J. Chem. Phys. 1991, 94, 3665 ±
3669; d) S. P. Karna, M. Dupuis, J. Comput. Chem. 1991, 12, 487 ± 504.
[42] D. R. Kanis, M. A. Ratner, T. J. Marks, Chem. Rev. 1994, 94, 195 ± 242.
[43] a) D. A. Kleinman, Phys. Rev. 1962, 126, 1977 ± 1979; b) C. Bosshard,
K. Sutter, P. Pr e√ tre, J. Hulliger, M. Flˆrsheimer, P. Kaatz, P. G¸nter,
Organic Nonlinear Optical Materials - Advances in Nonlinear Optics,
Vol. 1 (Eds.: A. F. Garito, F. Kajzar), Gordon and Breach, Basel, 1995.
[44] A. Willets, J. E. Rice, D. M. Burland, D. P. Shelton, J. Chem. Phys.
1992, 97, 7590 ± 7599.
4
782.
37] We are aware that the Onsager reaction field model should be used
only under the strict assumption that all chromophores are spherical in
shape, a case clearly not applicable here. Recent work by Dalton and
Robinson on a Monte Carlo treatment of the electrooptic activity of
dipolar chromophores has indeed pointed out the effect of spherical
versus ellipsoidal shapes, the latter fitting experimental data more
closely (B. H. Robinson, L. R. Dalton, J. Phys. Chem. 2000, 104,
4
785 ± 4795). Unfortunately, other theoretical models available in the
Gaussian package, such as the polarized continuum model (PCM), in
which the cavity is defined as the sum of a series of interlocking
spheres, or the alternative self-consistent isodensity polarized con-
tinuum model (SCI-PCM), did not satisfy convergence criteria during
geometry optimizations even of the smallest system 1c. It became
evident that the Onsager model was the only method able to satisfy
convergence criteria completely for the full set of compounds and
solvents to be considered in this study. Such a simplification could be
responsible for a certain degree of deviation from experimental data.
However, we believe that the overall good, and sometimes very good,
agreement with different experimental data do validate this computa-
tional approach.
[45] a) Photonic Polymer Systems (Eds.: D. L. Wise, G. E. Wnek, D. J.
Trantolo, T. M. Cooper, J. D. Gresser), Marcel Dekker, New York,
1998; b) L. R. Dalton, A. W. Harper, R. Ghosn, W. H. Steir, M. Ziari,
H. Fetterman, Y. Shi, R. V. Mustacich, A. K.-Y. Jen, K. J. Shea, Chem.
Mater. 1995, 7, 1060 ± 1081.
[46] P. Innocenzi, E. Miorin, G. Brusatin, A. Abbotto, L. Beverina, G. A.
Pagani, M. Casalboni, F. Sarcinelli, R. Pizzoferrato, Chem. Mater.
2002, 14, 3758 ± 3766.
[47] S. Gronowitz, S. Liljefors, Chem. Scr. 1979, 13, 39 ± 45.
[48] U. Mˆller, D. Cech, F. Schubert, Liebigs Ann. Chem. 1990, 1221 ± 1225.
Received: August 19, 2002
Revised: December 23, 2002 [F4356]
Chem. Eur. J. 2003, 9, 1991 ± 2007
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2007