Azo-azulene derivatives as second-order nonlinear optical chromophores

Article abstract of DOI:10.1002/1521-3765(20000717)6:14<2599::AID-CHEM2599>3.0.CO;2-G

The molecular and solid state nonlinear optical (NLO) properties of several (phenylazo)-azulenes are investigated. In particular, (4-nitrophenylazo)-azulene (2b) exhibits a quadratic hyperpolarizability (β(vec)) of 80 x 10^-30 cm⁵ esu recorded at 1.907 μm by the electric field-induced second-harmonic (EFISH) technique. This molecular material, which crystallizes in the monoclinic noncentrosymmetric space group Pc, exhibits an efficiency 420 times that of urea in second-harmonic generation. The origin of the optical nonlinearity in azo-azulene is discussed in relation with crystal structures and semiempirical calculations within the INDO/SOS formalism, and compared with that of the well known disperse red one (DR1) organic dye.

Full text of DOI:10.1002/1521-3765(20000717)6:14<2599::AID-CHEM2599>3.0.CO;2-G

FULL PAPER
Azo-Azulene Derivatives as Second-Order Nonlinear Optical Chromophores
Pascal G. Lacroix,*[a] Isabelle Malfant,[a] Gabriel Iftime,*[b] Alexandru C. Razus,[b]
Keitaro Nakatani,[c] and Jacques A. Delaire[c]
Abstract: The molecular and solid state nonlinear optical (NLO) properties of
several (phenylazo)-azulenes are investigated. In particular, (4-nitrophenylazo)-
azulene (2b) exhibits a quadratic hyperpolarizability (bvec) of 80 Â 10À30 cm5 esu
recorded at 1.907 mm by the electric field-induced second-harmonic (EFISH)
technique. This molecular material, which crystallizes in the monoclinic noncentro-
symmetric space group Pc, exhibits an efficiency 420 times that of urea in second-
harmonic generation. The origin of the optical nonlinearity in azo-azulene is
discussed in relation with crystal structures and semiempirical calculations within the
INDO/SOS formalism, and compared with that of the well known disperse red one
(DR1) organic dye.
Keywords: chromophores ´ donor±
acceptor systems ´ nonlinear optics
´ semiempirical calculations ´ solva-
tochromism
examples illustrate the fact that the search for new NLO
chromophores is still a very active research area, which now
encompasses newer generations of molecules, with electronic
properties of greater complexity than those of the substituted
benzene previously investigated.
For instance, azulene (1) is a non-benzenoid aromatic
molecule with a polar resonance structure (Scheme 1). One of
its characteristic features is the electron drift from the seven-
Introduction
The elaboration of nonlinear optical (NLO) materials has
attracted considerable interest for the past two decades
because of their potential applications in optoelectronics,
telecommunications, and optical information processing. In
particular, donor± acceptor substituted p-conjugated organics
with low-lying charge transfer excited states have received
much attention for their large second-order NLO re-
sponse.[1, 2] Many of them have been investigated both
theoretically,[3, 4] and experimentally,[5] in the substituted
stilbene series, which has been the most promicing family
until the mid nineties. However, several new organic deriv-
atives have recently appeared as very promicing alternative
structures. Among recent examples, thiophene-based stilbene
analogues have been reported to possess significantly en-
hanced optical nonlinearities,[2a, 6] and push ± pull arylethynyl
porphyrins,[7] and polyenes[8] leading to very large hyper-
polarizabilities have also been under investigations. These
Scheme 1.
membered ring toward the five-membered ring. This is in
accordance with an experimental dipole moment of about 0.8
D.[9] The five- and seven-membered rings of azulene carry
negative and positive charges, respectively. Therefore, the
five-membered ring can act as electron donor, and the seven-
membered ring as electron acceptor. Owing to this property,
azulene appears to be a versatile fragment which might be
used for the design of various NLO materials.
[a] Dr. P. G. Lacroix, Dr. I. Malfant
Laboratoire de Chimie de Coordination (CNRS)
205 route de Narbonne
Several types of push ± pull compounds can be imagined, in
which azulene will act as donor, as acceptor or as conjugated
bridge. The electron-acceptor character of azulene is support-
ed by the reaction with nucleophiles (at the seven-membered
ring),[10] or more recently by the use of an azulene derivative
as acceptor in charge-transfer complexes.[11] The electron-
donor character of azulene has been proven by the reaction
with electrophiles (at the five-membered ring),[10] by the
synthesis of charge-transfer complexes in which the azulene
31077 Toulouse cedex(France)
E-mail: pascal@lcc-toulouse.fr
[b] Dr. G. Iftime, Dr. A. C. Razus
C.D. Nentzescu Institute of Organic Chemistry
POX 258, 71141
Bucharest 15 (Romania)
[c] Dr. K. Nakatani, Prof. Dr. J. A. Delaire
Â
P.P.S.M. Ecole Normale Superieure de Cachan (UMR 8531)
Â
Avenue du President Wilson
94235 Cachan cedex(France)
Chem. Eur. J. 2000, 6, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2599 $ 17.50+.50/0
2599

P. G. Lacroix, G. Iftime et al.
FULL PAPER
moiety acts as donor,[12] or by the stabilization of methylca-
tions.[13] Semiempirical and ab initio calculations have shown
that azulene derivatives should exhibit NLO response com-
parable to those of paranitroaniline, and therefore should
deserve more investigations in materials chemistry.[14]
Surprisingly, only few reports on azulene-based materials
with NLO properties have been published to date. The first
one in 1996 by Asato et al.
phores. In particular, following the earlier report that azulene
could have the same donating strength than N,N-diethylani-
line,[15] 2b may exhibit electronic properties related to those of
2-[N-ethyl-4-(4-nitrophenylazo)phenyl-amino]ethanol (DR1,
3), which has already been widely used as NLO chromo-
phore.[19] This possibility is supported by the resonance
structures illustrated in Scheme 3. We present here an
described compounds which
contain azulene or guiazulene
as electron donor groups and
dicyanovinyl or 1,3-diethyl-2-
thiobarbituric acid as electron-
acceptor groups.[15] In the sec-
ond one, we have recently re-
ported on the first observation
of second harmonic generation
(SHG) in microcrystalline sam-
ples of azulene-based chromo-
phores.[16] In particular, we
found that 6-methoxy-1-[2-(4-
Scheme 3.
nitrophenyl)ethenyl]-azulene
exhibits an efficiency 58 times that of urea in second-harmonic
generation. Recently, a ferrocene derivative connected by an
ethenyl bond to the seven-membered ring of azulene was
described.[17] This compound exhibits a first-order hyper-
polarizability of 26 Â 10À30 cm5 esuÀ1. No measurements of its
efficiency in solid state have been reported.
In the present contribution, we wish to report on second-
order NLO properties of azo-azulenes derivatives (2), a family
of compounds that can readily be obtained by reaction of
azulene (1) with the appropriate diazonium salt (Scheme 2).[18]
extensive study of 2b, in comparison with 3 and with the
related (phenylazo)-azulene (2a), and (4-methoxyphenyl-
azo)-azulene (2c). The molecular hyperpolarizabilities are
investigated experimentally by spectroscopy and by the
electric field-induced second-harmonic (EFISH) technique,[20]
in combination with a quantum chemical analysis, within the
proven INDO/S-SOS formalism[4] to describe the electronic
structure NLO property relationships. The relations between
molecular and bulk nonlinearity are discussed in relation with
the crystal structure for 2b. Finally, the possibility to use
azulene-based chromophores as radical anions is discussed.
Results and Discussion
Design and general features of azo-azulenes: Contrary to 1-[2-
(phenyl)ethenyl]-azulenes (4) derivatives, which require a
2-step synthesis in dry atmosphere and are usually obtained as
a mixture of E:Z isomers,[16] 1-(4-Y-phenyl-azo)-azulenes (2)
are readily obtained in one step, by the reaction of azulene (1)
with the appropriate diazonium salt in ethanol, in presence of
sodium acetate, at room temperature, as previously repor-
ted[18c] (Scheme 2). This route, which turns out to be suitable
for linking azulene to phenyls bearing various donors and
acceptors substituents, allows a modulation of the electronic
properties of the chromophores. On the other hand, it is well
known that the nature of the conjugation path also contrib-
utes to the extend of the NLO response.[21, 22] In particular,
theoretical studies have predicted that azobenzenes should
exhibit larger hyperpolarizabilities than stilbenes analogues in
some cases,[21d] even if this fact is usually not observed
experimentally.[5a] Therefore replacing the ethenylene frag-
ment in 4 by azo groups might offer attracting opportunities
not only from the synthetic point of view, but also for NLO
purpose. It is interesting to note that, although these readily
available compounds have long been known,[18a,b] before
second-order NLO properties were first reported in 1961,[23]
Scheme 2.
This route provides the opportunity to expand the p-
conjugated system to various donor and acceptor substituents,
and hence to tune the push ± pull character of the chromo-
2600
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2600 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 14

Azo-Azulene Derivatives
2599±2608
our study is the first NLO investigation of azo-azulene. By
contrast, the related compound 3, has been widely used for
about fifteen years, especially in poled polymer matrix.[24]
Azulenic derivatives possess additional features that can
attract some interest depending on the NLO applications. It is
well known that second-order NLO chromophores have to be
embodied into noncentrosymmetric environment if the mo-
lecular hyperpolarizability (b) has to give rise to an observ-
able bulk susceptibility (c(2)).[25] The most traditional approach
which guarantees stable acentric organizations of NLO
chromophores in the solid state is based on the possibility of
growing large noncentrosymmetric single crystals.[26] Azulenic
derivatives frequently possess low dipole moments,[14b] and it
is usually assumed that they are more likely to crystallise in a
non-centrosymmetric fashion, although this idea is somewhat
controversial.[27] The dipole moments of the compounds
investigated in the present study are gathered in Table 1. As
Figure 1. Atom labeling scheme, and thermal vibration ellipsoids for 2b.
H atoms are omitted for clarity.
Table 1. Decomposition temperature (Td in 8C), redoxpotentials ( E in V
versus SCE), and dipole moments (m in D) of azo-azulenes.
[a]
Compound (-Y)
Td
Ered
Eox
m[b]
2a (-H)
2b (-NO2)
2c (-OCH3)
245
235
275
À 1.26
À 0.86
À 1.29
1.07
1.16
0.94
2.63
7.20
2.27
[a] Peak potentials. [b] Data from ref. [28b].
anticipated, 2b exhibits a reduced dipole than 3 (7.2 D vs
8.7 D),[28] which is indicative of possible uses of the crystal
growth methodology for the design of SHG efficient material
containing azulene moieties (vide infra). However, in the case
of 2b versus 3, a difference of 1.5 D is probably not large
enough to be the only reason to account for such different
crystal environment.
Figure 2. Atom labeling scheme, and thermal vibration ellipsoids for 2c
molecule 1 (top) and molecule 2 (bottom). H atoms are omitted for clarity.
Among other interesting physico-chemical features, azu-
lenic derivatives frequently exhibit an excellent thermal
stability,[15] which is an important prerequisite for many
applications. Following this idea, the use of azulene as a
donor group could give rise to thermally stable NLO
chromophores. In addition, enhanced thermal stabilities of
the azobenzene derivatives versus the stilbene analogues has
also be reported.[29] The values given in Table 1 indicate a
good range of stability for 2 (235 ± 2758C).
Description of the structures: Figures 1, 2, and 3 illustrate the
atomic numbering scheme employed for molecules 2b, 2c,
and 3, respectively. The X-ray structure analyses of 2b reveals
two molecules present in a non centrosymmetric crystal cell
(monoclinic space group Pc), which refer to one another by a
glide mirror. The molecule is quasi planar, with a largest
deviation of 0.131 Š observed at O(2). The structure of 2c
consists of four discrete molecules embodied in a centrosym-
metric environement. Two molecules are present in the
asymmetric unit cell (molecules 1 and 2). The angle between
the mean plan of molecules 1 (largest deviation of 0.174 Š at
C(15)) and 2 (largest deviation of 0.043 Š at C(33)) is equal to
80.08. The asymmetric unit cell of 3 is made of two molecules
of dye with an additionnal molecule of water. The largest
deviation from planarity is 0.158 Š at O(12) for molecule 1,
Figure 3. Atom labeling scheme, and thermal vibration ellipsoids for 3
molecule 1 (top) and molecule 2 (bottom). H atoms are omitted for clarity.
and 0.113 Š at O(21) for molecule 2. The angle between the
dyes is equal to 46.58. The majority of the azobenzene
containing structures reported in the litterature show signifi-
cant twist of at least one of the phenyl rings within the
azobenzene frame.[30] This situation is not observed here, with
angles between the phenyl and the azulene mean plans equal
to 4.58, 8.88 and 1.88 for 2b and 2c (molecule 1 and 2),
respectively. These values are in the same range of magnitude
Chem. Eur. J. 2000, 6, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2601 $ 17.50+.50/0
2601

P. G. Lacroix, G. Iftime et al.
FULL PAPER
Table 3. Absorption maxima (lmax in nm) of the low energy charge transfer
transitions for 2a ± c and 3 in solvents of different polarities.
of the 3.18 and 5.78 measured between the two aromatic rings
of 3 (molecule 1 and molecule 2, respectively). Selected bond
À
2a
2b
2c
3
lengths are given in Table 2. A shortening of the two C N
bond, with averages of 1.405(3), 1.407(4), 1.404(4) Š for 2b
cyclohexane
CCl4
dioxane
ethyl acetate
THF
toluene
CH2Cl2
DMF
chloroform
MeOH
417
422
422
420
423
424
424
426
424
421
422
421
421
428
450
458
460
462
464
468
472
474
478
480
482
486
487
506
425
429
428
424
428
430
428
430
429
425
425
426
423
434
454
458
474
480
488
464
482
502
472
465
465
466
466
483
Table 2. Selected bond lengths (Š) for 2b, 2c, and 3.
2b
C(4) N(2)
N(2) N(3)
N(3) C(7)
À
1.424(3)
1.277(3)
1.387(3)
À
À
2c
À
C(4) N(2)
N(2) N(3)
N(3) C(7)
C(24) N(22)
N(22) N(23)
N(23) C(27)
1.424(4)
1.280(3)
1.390(4)
1.419(4)
1.288(3)
1.389(4)
EtOH
À
CH3CN
acetone
DMSO
À
À
À
À
3
À
C(14) N(12)
N(12) N(13)
N(13) C(131)
C(24) N(22)
N(22) N(23)
N(23) C(231)
1.425(3)
1.262(3)
1.400(3)
1.460(4)
1.241(3)
1.416(4)
mophores is well described in term of a ground and an excited
state having charge transfer character, and is related to the
energy of the optical transition (E), its oscillator strength (f),
and the difference between ground and excited state dipole
moment (Dm) through the relation:
À
À
À
À
À
3 e2 ꢀh f Dm
2 m E3
E4
b ˆ
Â
(1)
2
2
ꢀE2 À ꢀ2 ꢀhw† †ꢀE2 À ꢀꢀhw† †
and 2c (molecule 1 and molecule 2), respectively, are
indicative of significant conjugation of the whole p system.
These bond lengths are shorter than the 1.412(3) Š and
1.438(4) Š found for 3 (molecule 1 and 2, respectively). In
in which ꢀhw is the energy of the incident laser beam.
Equation (1) shows that large solvatochromism is strongly
indicative of a large hyperpolarizability. Solvatochromism
analysis (see Experimental Section) gives Dm values of 3.1 and
5.2 D for 2b and 3, respectively. Therefore, assuming a two-
level description of the NLO response, comparable intensi-
ties, lmax and solvatochromic shifts evidenced for 2b and 3
suggest the same b magnitude for both chromophores.
À
addition, the Table reveals an increase of the N N bond
lengths, in the azulenic derivatives versus those of 3. All these
features allow the possibility of long range electron delocal-
ization.
Optical spectroscopy: The absorption spectra of 2b and 3
recorded in dioxane are compared in Figure 4. Both exhibit an
intense band at 460 nm (e ˆ 31700mÀ1 cmÀ1) and 474 nm (e ˆ
31000mÀ1 cmÀ1) for 2b and 3, respectively. The absorption
These observations can be compared with the INDO
derived spectroscopic properties gathered in Table 4. In the
case of 3, the calculation indicates a single intense absorption
Table 4. Comparison of experimental (in dioxane) and INDO derived
spectroscopic data (lmax in nm, e in MÀ1 cmÀ1, oscillator strength (f) and Dm
in D) for 2b and 3.
Experimental
INDO
lmax
e
Dm
lmax
f
Dm
4.7
1.7
13.7
14.7
2b
3
460
31700
31000
3.1
1 ! 3
445
440
402
379
0.77
0.45
1.24
1.12
1 ! 4
474
5.2
molecule 1
molecule 2
maxima located at 402 nm (f ˆ 1.24) and 379 nm (f ˆ 1.12) for
molecule 1 and 2, respectively, a difference which is due to
different molecular geometries. Molecule 2 is slightly twisted
(5.78) versus molecule 1 (3.18), a situation which can reduce
the overlap and hence the intensity (f) and the wavelength of
the electronic transition. In the case of 2b, two intense and
closely located transitions at 445 nm (1 ! 3) and 440 nm (1 !
4) can account for the experimental band observed at 460 nm.
The sum of their intensity lies in the range of magnitude of
that calculated for 3, in agreement with the experimental
extinction coefficients. We make the assumption that the
calculated spectra (lmax and intensities) roughly fit the
Figure 4. Electronic spectrum for 2b recorded in dioxane, 3 (dotted line) is
given as a reference.
maxima (lmax) recorded in solvent of different polarities are
given in Table 3. The data clearly indicate that 2b and 3
exhibit a strong solvatochromism, with lmax values lying in the
same range of magnitude (450 ± 500 nm) depending on the
solvent. Solvatochromism is usually indicative of a change in
dipole moments (Dm) upon electronic excitations. According
to the well known and widely used ªtwo-level modelº,[20a, 31]
the hyperpolarizability (b) of ªpush ± pullº p-organic chro-
2602
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2602 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 14

Azo-Azulene Derivatives
2599±2608
experimental data, in order to relate the NLO response and
the INDO-based electronic properties in the next Section.
However, it must be noticed that the agreement between
calculated and experimental Dm value is not excellent for
compound 3. Although the origin of this difference has not
been carefully analysed, it has to be reminded here that
experimental estimations of Dm by mean of solvatochromic
shifts may not be fully reliable in some cases.[32] In particular,
this method implies the knowledge of the Onsager radius, a
parameter hard to define for these compounds (see Exper-
imental Section).
when the second harmonic is close to the absorption maxima,
in agreement with Equation (1). It must be observed that
INDO-SOS calculation indicates the same magnitude for the
NLO response of both compounds, the hyperpolarizability of
2b (43.6 Â 10À30 cm5 esuÀ1) being slightly higher than the
averaged value calculated for 3 (40.6 Â 10À30 cm5 esuÀ1), in
agreement with the experimental data. This qualitative
agreement makes the analysis of the calculations reliable.
Within the framework of the SOS perturbation theory, the
molecular hyperpolarizability can be related to all excited
states of the molecule and can be partitioned into two
contributions, so called two-level (b2level) and three-level
(b3level) terms.[33] Analysis of term contributions to the
molecular hyperpolarizability of 2b and 3 indicates that
two-level terms dominate the nonlinearity. Table 6 indicates
the main electronic transitions involved in b2level for com-
pounds 2 and 3. In the case of 2b, it can mainly be related to
the dominant 1 ! 3 transition, which contribute 45% to the
nonlinearity, while the single 1 ! 5 transition account for
about 90% of the b2level of 3. Therefore, understanding these
transitions provides qualitative understanding of b. The
charge transfer associated with these transitions are compared
in Figure 5. It clearly appear that the five-membered ring of
the azulene moiety acts as the donor in 2 b.
Molecular hyperpolarizabilities: The hyperpolarizabilities of
2a ± c were measured by the EFISH technique. In this
technique, the noncentrosymmetry of the solution is induced
by dipolar orientation of the chromophores by a pulse of high
voltage, yielding a signal proportional to the square of G, the
third-order nonlinear coefficient. G and b are related by the
relation:
G ˆ Nf2w fw f0(g‡mbvec/5kT)
(2)
N is the number of chromophore per unit volume, f2w, fw and f0
are Lorentz local field factors. If the cubic hyperpolarizability
(g) is neglected, EFISH, combined with the measure of the
dipole moment (m) leads to bvec, the vector component of the
bijk tensor along the dipole moment direction. The EFISH
data are gathered in Table 5. It must be reminded here from
Equation (1), that b (and hence bvec) are dependent on the
energy of the incident laser beam. On the other hand,
(3e2 ꢀh fDm)/2mE3 is the static hyperpolarizability b(0), inde-
pendent on the laser. This is the parameter of interest to take
into account for the comparison of the intrinsic optical
nonlinearity of the molecules. b(0)vec is equal to 57 Â
10À30 cmÀ5 esuÀ1 for 2b,
a
value close to the 47 Â
10À30 cmÀ5 esuÀ1 measured for 3. It must be observed that
the agreement between experimental and INDO data is
satisfactory for 2b and c.
Figure 5. Differences in electronic populations between the ground state
and the excited state for the main transitions involved in the NLO response
of 2b (1 ! 3 transition) and 3 (1 ! 5 transition). The white (black)
contributions are indicative of an increase (decrease) of electron density in
the charge transfer process.
A chemical quantum analysis can provide a rationale for
the understanding of the origin of the NLO response of 2b
versus that of 3. The INDO calculated hyperpolarizabilities
are reported in Table 5. The data indicates the b enhancement
Table 5. INDO-SOS NLO response (btotal, b2level, and bvec in 10À30 cm5 esuÀ1) at various wavelengths (incident laser beam), compared with experimental bvec
and SHG powder efficiences (I2w) versus that of urea for 2a ± c and 3.
,
Calculated hyperpolarizabilities
b(1.907 mm)
Experimental data[a]
b(0)vec
b(1.064 mm)
b(0)
bvec
I2w
total
2 level
vec
2.3
total
2 level
vec
total
2 level
vec
1.907 mm
1.356 mm
1.907 mm
2a
2b
2c
36.0
252.2
26.7
328.3
17.4
68.9
16.5
104.0
13.2
62.8
16.0
50.0
14.2
80.8
! 12.7
! 43.6
7.1
80
< 5.4[c]
! 5.4
! 57
6.5
420
0
242.2
!< 4.1[c]
molecule 1
molecule 2
32.9
33.3
24.2
29.2
4.5
4.4
18.9
21.0
19.0
26.5
11.3
14.2
17.2
18.7
18.1
22.6
! 10.3
! 12.6
3
125[b]
molecule 1
molecule 2
153.6
96.2
243.7
162.7
151.0
92.9
63.7
46.1
122.0
93.6
62.4
44.7
47.7
35.8
98.3
77.9
! 46.6
! 34.7
[a] Experimental b(0) obtaines using 2-level description, with b(0) ˆ b[1 À (2lmax/lw)2][1 À (lmax/lw)2]. [b] Data from ref. [27]. [c] Limited accuracy due to weak
EFISH signal.
Chem. Eur. J. 2000, 6, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2603 $ 17.50+.50/0
2603

P. G. Lacroix, G. Iftime et al.
FULL PAPER
Table 6. Energies (lmax in nm), oscillator strengths (f), dipole moment changes between ground and excited state (Dm in D), contribution in the b2level, and
composition of the main two transitions involved in the NLO responses of 2 and 3.
Compound
Transition
lmax
f
Dm
State (%)[a]
Composition[b] of CI expansion
2a
1 ! 8
1 ! 9
1 ! 3
1 ! 4
267
250
445
440
0.62
0.50
0.77
0.45
7.1
30
21
45
9
À 0.489
‡0.479c41!44 À 0.413c43!46
7.4
4.7
1.7
À 0.411c42!45 À 0.408c42!44 À 0.393c42!45
2b
À 0.714c41!44 À 0.438
‡0.371c45!54
c51!52 À 0.540c51!52
0.579c45!52
c45!52 À 0.485c45!54
2c
molecule 1
1 ! 4
1 ! 9
1 ! 9
1 ! 4
423
255
257
430
0.88
0.71
0.69
0.82
1.86
10.0
11.35
1.55
27
26
29
22
À 0.795
À 0.359
0.523c48!51
c49!50 À 0.482c48!50
c49!51 À 0.379c49!52
molecule 2
À 0.626
‡0.400c48!50‡0.375c49!52
c48!51 À 0.396c45!50
À 0.806c49!50
3
molecule 1
1 ! 5
1 ! 9
1 ! 5
1 ! 9
403
267
352
262
1.24
0.18
1.12
0.22
13.74
17.1
14.7
22.5
92
5
87
8
0.818
‡0.477c60!62
À 0.801c60!61 À 0.389c60!61
molecule 2
À 0.762c59!61 À 0.562c60!62
0.718c59!60
c60!61 À 0.474c60!61
[a] State % ˆ bg!e(i)/Si bg!e(i). [b] Orbital 43 is the HOMO and 44 the LUMO for 2a, orbital 51 is the HOMO and 52 is the LUMO for 2b, orbital 49 is the
HOMO and 50 the LUMO for 2c, orbital 60 is the HOMO and 61 the LUMO for 3.
By contrast, 2a and 2c are modest chromophores. Interest-
ingly, the experimental and calculated magnitudes of bvec
(Table 5) significantly disagree, although it clearly appears
that the NLO responses are deeply reduced versus that of 2b
and 3. We may tentatively relate this discrepancy to the fact
that the dipole moments of 2a and 2c are deeply reduced
(Table 1) and no longer parallel to b. Large experimental
uncertainties can arise from small EFISH signals (propor-
tional to m). This effect becomes dramatic at 1.064 mm where b
and m are almost orthogonal, which leads to almost vanishing
bvec (Table 5). Moreover, the modest INDO accuracy on the
involving the whole p-electron system, 2a and 2c exhibit
electronic properties dominated by the azulene fragment.
Consequently, such chromophores should exhibit similar and
rather modest lmax, solvatochromic shift (Table 3) and b
(Table 5), whatever the donating strength of the phenyl
substituent would be.
Bulk efficiencies in second-harmonic generation: In molec-
ular materials, the susceptibility tensor (c(2)) which determines
the second-order response is related to the underlying
molecular hyperpolarizability tensor (b). The above section
has shown that 2b possesses a large b, in the range of
magnitude of that of the well known molecular structure 3.
However, the main bottleneck to the development of
molecular second-order NLO material is the compulsory
noncentrosymmetric environment of the chromophores if the
molecular hyperpolarizability is to contribute to an observ-
able bulk nonlinear susceptibility.[25] It is generally assumed
that traditionnal push ± pull chromophores, such as 1,4-
disubstituted benzene and 4,4'-disubstituted stilbene, have
large dipole moments and approximately linear shapes, which
cause most of them to crystallize in centrosymmetric space
groups. One strategy employed to encourage noncentrosym-
metric crystallization of neutral organic chromophores is to
increase their geometrical asymmetry by introducing a sub-
stituent in the ortho or meta position of one of the aromatic
rings. Thus, whereas crystals of 4-nitroaniline are centrosym-
metric, 2-methyl-4-nitroaniline crystallize in the noncentro-
symmetric space group Cc.[34, 35] Another strategy is to lower
the ground state dipole moment, without cancelling the
intramolecular charge transfer. 3-Methyl-4-nitropyridine-1-
oxide was the first organic material that exemplifies with
success this strategy.[36] Along these lines, it seems that the
substituted azulenes investigated in the present study should
offer more chance for noncentrosymmetric crystallization, by
virtue of their bent shape and reduced dipole moment.
However, the development of single-crystal NLO materials is
hamped by the absence of general rules for predicting crystal
structure from molecular structure, and the zero efficiency of
2c (Table 5) proves the limitations of this methodology.
angle value between b and m can affect the magnitude of bvec
,
especially as the angle value is large. The fact that calculations
are based on solid state molecular structures, whereas
molecules in solution can adopt another geometry may also
explain the discrepancy between theory and experiments.
However, it clearly appears either from experimental or
calculated data that the NLO response of 2a and 2c is far
reduced versus that of 2b and 3. Although no transition
clearly dominate their NLO response, as indicated in Table 6,
the origin of the charge transfers (and hence b) can
unambigously be found in the azulene moiety itself. This is
evidenced in Figure 6, in which the directions and magnitudes
of b are compared for the four chromophores. While 2b and 3
exhibit the same features with a strong charge transfer
Figure 6. Orientation of b for 2a ± c and 3.
2604
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2604 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 14

Azo-Azulene Derivatives
2599±2608
Nevertheless, the noncentrosymmetric crystal structure of 2b
allows an investigation of the relations between microscopic
and macroscopic second-order optical nonlinearities for this
material.
anions, which could be used in the design of organic metals.[11]
In order to further understand the electrochemical properties
of these molecular materials, we have drawn the frontier
orbitals of 2b and 2c in Figure 8. Interestingly, it can be
Following Zyss approach,[26] the hyperpolarizability tensor
(components bijk in the molecular frame) is related to the
corresponding crystalline first-order nonlinearity c(2) (compo-
nents dIJK in crystalline frame) through the following relation:
dIJK(À2w;w,w) ˆ NfI2w fJw fKw Scos(I,i)cos(J,j)cos(K,k)bijk
N is the number of chromophores per unit volume, fI2w, fJw, and
fKw are Lorentz local-field factors. The summation is per-
formed over all molecules in the unit cell, and the cosine
product terms represent the rotation from the molecular
reference frame into the crystal frame. 2b crystallizes in space
group Pc (monoclinic point group 2). Assuming a simplified
one-dimensional description of the molecular nonlinear
tensor of the molecules, b has only one non vanishing
coefficient along the charge transfer axis x of the molecule
(namely bxxx). This model leads to
Figure 8. Comparison of the frontier orbitals for 2b (left) and 2c (right).
observed that most of the electron density of the highest
occupied molecular orbital (HOMO) is carried by the azulene
fragment, in both cases. Therefore, the HOMO of 2b and 2c
exhibit the same features. This is consistent with oxydation
potentials lying in the same range of magnitude (Table 1)
whatever the nature of the substituted phenyl is. By contrast,
the lowest unoccupied molecular orbitals (LUMO) of both
materials are very different. In 2b, most of the electron
density is located on the nitrophenyl substituent, in relation
with the withdrawing effect of NO2. This observation results
in a shift of the reduction potentials to higher values for this
material (Table 1).
dZZZ ˆ Nbxxx cos3q
dZXX ˆ Nbxxx cosqsin2q
where the Lorentz local-field factors have been assumed
equal to 1. All other components of the tensor are negligible
(q is defined as the angle between the main intramolecular
charge transfer axis 0x and the glide mirror of the crystal). The
optimization of dZZZ can be achieved with q ˆ 08, while the
angular factor weighting bxxx in the expression of dZXX is
maximized and equal to 0.385 for q ˆ 54.748. The orientation
of b in the crystal cell is shown in Figure 7. q reaches a value of
The possibility to isolate stable radical anions is illustrated
in Figure 9 for 2b. Two reduction processes are observed at
Figure 7. CAMERON view of 2b showing the angle (q ˆ 15.18) between b
and the glide mirror (dotted line) of the cell.
15.18, which corresponds to angular factors of 0.901 and 0.065
for dZZZ and dZXX, respectively. Assuming the calculated b
value of 68.9 Â 10À30 cm5 esuÀ1 and the crystallographic deter-
mination of N ˆ 3.15  1021 cmÀ3 lead to dZZZ ˆ 19.6 À
10À8 cm2 esuÀ1 and dZXX ˆ 1.41  10À8 cm2 esuÀ1. These data
indicate a large quadratic NLO response for the crystal.
Consequently, the SHG efficiency of 2b has been found to be
420 times that of urea (Table 5), a very large value for this new
class of materials.
Figure 9. Cyclic voltammogram of 2b in acetonitrile.
À0.86 V (E1) and À1.24 V (E2). They exhibit satisfactory
reversibilities (Ered À Eox ˆ 88 mV and 71 mV for E1 and E2,
respectively). The stability of the radical anion depends on the
comproportionation constant (Kcon), calculated for the equi-
librium
Kcon
)
2b‡(2b)2À
2(2b)À
*
Electrochemical properties of azulene-based chromophores:
Beside NLO properties, azulene derivatives possess electronic
features of greater complexity than those of the earliest
benzene derivatives, such as the possibility of forming radical
the redoxpotential being as follows: E1 À E2 ˆ 0.059 log
(Kcon). A potential difference of 0.38 V leads to large
Chem. Eur. J. 2000, 6, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2605 $ 17.50+.50/0
2605

P. G. Lacroix, G. Iftime et al.
FULL PAPER
135.5, 129.2, 129.0, 126.5, 126.4, 125.2, 122.2, 119.9; 1H NMR (CDCl3) and
13C NMR (CDCl3) are in agreement with the ones previously reported.[41]
C16H12N2: calcd C 82.73, H 5.21, N 12.06; found C 82.65, H 5.30, N 12.05.
comproportionation constant (Kcon ˆ 2.75  106), which indi-
cates that radicals such as 2b could be isolated and charac-
terized. Interestingly, it has recently been mentioned that
radicals might offer intriguing candidates for NLO pur-
pose.[37±39] This might stimulate further challenging researches
using azulene-based radicals as NLO materials with extended
electronic properties.
Azulene-1-azo-(4'-methoxybenzene) (2c): Yield: 98 mg, 84%; m.p.: 988C
‡
(lit.: 968C);[18a,b] MS (m/z): 262 [M] ; 1H NMR (CD2Cl2): d ˆ 9.31 (d, J ˆ
9.8 Hz, 1H, H(8)), 8.35 (d, J ˆ 9.3 Hz, 1H, H(4)), 8.28 (d, J ˆ 4.5 Hz, 1H,
H(2)), 7.97 (d, J ˆ 8.9 Hz, 2H, C6H4), 7.75 (t, J ˆ 9.9 Hz, 1H, H(6)), 7.44 (d,
J ˆ 4.9 Hz, 1H, H(3)), 7.43 (t, J ˆ 9.7 Hz, H(5±7)), 7.32 (t, J ˆ 9.7 Hz, 1H,
H(5±7)), 7.05 (d, J ˆ 8.9 Hz, 2H, C6H4), 3.90 (s, 3H, CH3); 13C NMR (CDCl3):
160.8, 148.5, 143.9, 143.5, 139.4, 138.3, 137.9, 135.5, 126.1, 125.9, 125.1, 123.8,
119.7, 114.2, 55.6; C17H14N2O: calcd C 77.82, H 5.38, N 10.68; found C 77.77,
H 5.40, N 10.61.
Conclusion
Azulene-1-azo-(4'-nitrobenze) (2b): 4-Nitroaniline (9 mg, 0.5 mmol) were
dissolved in a mixture of concentrated H2SO4 (1 mL) and a drop of water.
The mixture was heated for 5 min to ensure complete dissolution of the
amine, then cooled at 08C. A solution containing sodium nitrite (35 mg,
0.5 mmol) in water (0.5 mLI was slowly added to the reaction mixture, and
the solution was allowed to stand at room temperature for 5 min. The
solution of the resulting diazonium salt was then slowly added under
stirring to a cooled suspension (08C) containing azulene (64 mg, 0.5 mmol)
and sodium acetate (2 g) in ethanol (5 mL. The mixture was stirred for
10 min at room temperature, then the organic compound was separated
between CH2Cl2 and a 20% aqueous solution of sodium carbonate. The
organic layer was dried (Na2SO4) then the solvent evaporated. The crude
residue was purified by column chromatography with CH2Cl2 on alumina.
After evaporation of the solvent, crystallisation was induced by adding a
small amount of pentane, and gave dark-brown crystals (120 mg, 87%).
M.p.: 2058C (lit. 2048C).[18b] MS (m/z): 277 [M] ; H NMR (CD2Cl2): d ˆ
9.38 (d, J ˆ 9.8 Hz, 1H, H(8)), 8.44 (d, J ˆ 9.3 Hz, 1H, H(4)), 8.35 (d, J ˆ
9.1 Hz, 2H, C6H4), 8.30 (d, J ˆ 4.7 Hz, 1H, H(2)), 8.06 (d, J ˆ 9.1 Hz, 2H,
C6H4), 7.88 (t, J ˆ 9.9 Hz, 1H, H(6)), 7.60 (t, J ˆ 9.7 Hz, 1H, H(5±7)), 7.49 (d,
J ˆ 4.3 Hz, 1H, H(3)), 7.48 (t, J ˆ 9.6 Hz, 1H, H(5±7)); 13C NMR (CDCl3): d ˆ
157.8, 147.2, 145.3, 144.6, 140.7, 140.2, 139.0, 135.7, 128.2, 125.5, 124.8, 122.5,
121.3; C16H11N3O2: calcd C 69.31, H 4.00, N 15.15; found C 69.01, H 4.13, N
15.02.
After having focused on relatively simple NLO systems with
particular emphasis on the paranitroaniline family, chemists
have increasingly extended the field of their investigations
over the last few years. We have presented here the second-
order NLO properties of azo-azulene derivatives. At the
molecular level, 2b exhibits a 57 Â 10À30 cm5 esuÀ1 b(0) value a
little higher than that of the well known ªDisperse Red 1º (3)
chromophore. Interrestingly, 2b crystallizes in the noncen-
trosymmetric space group Pc which results in a SHG
efficiency 420 times that of urea.
Over the last two decades, intense activity has demonstrat-
ed at both molecular and crystalline levels the relevance of
organic media for quadratic nonlinear optics. Azulene-based
materials which have received very limited attention in
nonlinear optics are a family of readily available compounds,
which offer additional electronic features (e.g. reversible
redoxproperties) which could make them attracting candi-
dates as multi-functional molecular devices.
‡
1
X-ray crystallography: The data were collected on a Stoe Imaging Plate
Diffraction System (IPDS) equipped with an Oxford cryosystems cooler
device, with a tube power of 2 kW for 2b and 3, and 1.5 kW for 2c. The
crystal to detector distance was 80 mm. Owing to rather low mxvalues,
0.014, 0.016, and 0.027 for 2b, 2c, and 3, respectively, no absorption
corrections were considered. The structures were solved by direct methods
(Shelxs-86)[42] and refined by least-square procedures. Crystallographic
data for 2b, 2c, and 3 are summarized in Table 7. The weighting Scheme
used in the last refinement cycles was w ˆ w'[1 À (DF/6s(Fo)2]2 where w' ˆ
1/S1nArTr(x) with 5 coefficients Ar for the Chebyshev polynomial ArTr(x),
where x was Fc/Fc(max). The calculations were carried out with the
CRYSTALS package programs[43] running on a PC. The drawings of the
molecular structures were obtained with the help of CAMERON.[43] The
atomic scattering factors were taken from International Tables for X-ray
Crystallography.[44]
Experimental Section
Reagents and equipment: Disperse Red 1 (3), azulene (1), and aromatic
amines for coupling reactions were used as commercial samples. Basic
alumina (activity B II-III, Brockmann) was used for column chromatog-
raphy. TLC analysis was carried out on Merck Aluminoxid 60 F254
aluminum plates. The proton and carbon spectra were recorded on a
Bruker AC-F 200 NMR spectrometer. Mass spectra were recorded on a
SQMD 1000-Carlo Erba 5300 Mega Series mass spectrometer with
capillary column, at 70 eV. Electronic spectra were recorded on
a
Hewlett ± Packard 8452 A spectrophotometer. Thermal measurements
were performed by TG/DTA analysis on a Setaram-TGDTA92 thermo-
analyser. The experiments were conducted under nitrogen on 5 mg of
sample (rate of heating: 108C per min). The decomposition temperature
(Td) was assigned as the intercept of the leading edge of the decomposition
endotherm with the base line of the DTA scan.[29, 40]
Crystallographic data (excluding structure factors) for the structures
reported in this paper have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication no. CCDC-127823 ±
127825. Copies of the data can be obtained free of charge on application to
CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax: (‡44)1223-336-
033; e-mail: deposit@ccdc.cam.ac.uk).
Azulene-1-azobenzene (2a) and azulene-1-azo(4'-methoxybenzene) (2c):
A solution of aryldiazonium chloride was prepared at 08C, by adding
sodium nitrite (37 mg, 0.54 mmol) to
a solution of aromatic amine
hydrochloride (0.54 mmol) and concentrated HCl (0.9 mL) dissolved in
water (2 mL). This solution was slowly added, with stirring, to a mixture of
azulene (72 mg, 0.44 mmol) and sodium acetate (170 mg) in ethanol
(10 mL) at room temperature. After 10 more minutes stirring, the reaction
mixture was extracted with CH2Cl2. The organic solution was washed with
5% aqueous ammonia then with water, and dried (Na2SO4). The solvent
was evaporated and the residue was purified by column chromatography
with n-pentane/CH2Cl2 5:1 on alumina. The product (dark-brown crystals)
was then crystallised from n-hexane.
Spectroscopy: The changes in dipole moment between ground and excited
states (Dm) were obtained from solvatochromism through the Lippert ±
Mataga Equation (3).[45]
nÄCT ˆ nÄgCT‡C1(n2 À 1)/(2n2‡1)‡C2[(D À 1)/(D‡2) À (n2 À 1)/(n2‡2)] (3)
where nÄCT is the frequency (cmÀ1) of the optical transition in a particular
solvent, nÄgCT is the frequency in the gas phase, D and n are the dielectric
constant and refraction indexof the solvent, and C1 and C2 are constants.
The constant C2 can be determined from least-square fit of Equation (2) to
the charge transfer band in different solvent and allows the determination
of Dm through Equation (4):
Azulene-1-azobenzene (2a): Yield: 82 mg, 81%; m.p.: 1208C (lit.: 120 ±
‡
1218C).[18c] MS (m/z): 232 [M] ; 1H NMR (CD2Cl2): d ˆ 9.36 (d, J ˆ
9.8 Hz, 1H, H(8)), 8.39 (d, J ˆ 9.3 Hz, 1H, H(4)), 8.30 (d, J ˆ 4.6 Hz, 1H,
H(2)), 7.97 (m, 2H, C6H5), 7.80 (t, J ˆ 9.8 Hz, 1H, H(6)), 7.54 (m, 2H, C6H5),
7.46 (d, J ˆ 4.3 Hz, 1H, H(3)), 7.45 (m, 2H, H(5±7), C6H5), 7.37 (t, J ˆ 9.7 Hz,
1H, H(5±7)); 13C NMR (CDCl3): d ˆ 154.1, 143.9, 143.8, 139.5, 138.7, 138.5,
C2 ˆ 2mgDm/hca3
(4)
2606
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2606 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 14

Azo-Azulene Derivatives
2599±2608
Table 7. Crystal data collections and refinements.
classical method based on the Gug-
genheim theory.[50] Further details of
the experimental methodology and
data analysis are reported else-
where.[49]
2b
2c
3
Crystal data
chemical formula
molecular weight
crystal system
space group
a [Š]
C16H11N3O2
277.3
monoclinic
Pc
6.4003(6)
3.7733(3)
26.454(3)
C34H28N4O2
524.6
monoclinic
P21/n
15.999(2)
6.318(1)
27.870(3)
C32H38N8O7
646.7
triclinic
Theoretical methods: The all-valence
INDO (intermediate neglect of differ-
encial overlap) method,[51] in connec-
tion with the sum-over-state (SOS)
formalism,[52] was employed. Details
of the computationally efficient IN-
DO/SOS based method for describing
second-order molecular optical non-
linearities have been reported else-
where.[4] Calculation was performed
using the INDO/1 Hamiltonian incor-
porated in the commercially available
MSI software package INSIGHT II
(4.0.0). The monoexcited configura-
tion interaction (MECI) approxima-
tion was employed to describe the
excited states. The 100 energy transi-
tions between the ten highest occupied
molecular orbitals and the ten lowest
unoccupied ones were chosen to un-
dergo CI mixing. Structural parame-
ters used for the INDO calculations
were taken from the present crystal
study for 2b, 2c (molecule 1), and 3.
Å
P1
7.539(1)
11.365(1)
19.358(3)
83.58(2)
83.77(2)
73.81(1)
1578.0
b [Š]
c [Š]
a [deg]
b [deg]
95.79(1)
104.99(1)
g [deg]
V [Š3]
635.6
1.45
0.917
2722.6
1.28
0.75
1(calcd) [g cmÀ3
]
1.36
0.919
m(MoKa) [cmÀ1
]
crystal size [mm]
T [K]
0.32 Â 0.15 Â 0.07
0.15 Â 0.15 Â 0.40
0.37 Â 0.37 Â 0.17
160
160
140
Data collection
radiation
scan mode
MoKa(l ˆ 0.71069 Š)
MoKa
f
4.5 < 2q < 48.3
none
13952
4754
MoKa
f
4 < 2q < 48.4
none
12972
4766
f
2q range
4.6 < 2q < 48.4
none
4131
1891
1616
absorption correction
measured reflections
independant reflections
observed reflections
criteria
2397
I > 2s(I)
3127
I > 3s(I)
I > 3s(I)
Refinement
refinement on
R[a]
Electrochemistry:
Electrochemical
F
F
F
0.046
0.053
425
calculated
Chebychev
1.62
measurements were carried out in an
airtight three-electrodes cell connect-
ed to a vacuum argon/N2 line, using the
interrupt method to minimize the un-
compensated resistance (iR) drop. Cy-
clic voltammetry was recorded in ace-
0.038
0.044
191
calculated
Chebychev
2.21
0.054
0.053
362
calculated
Chebychev
1.72
0.173
1.78
0.014
0.551
0.83
Rw[b]
no of parameters used
H atoms treatment
weighting scheme
coeff. ar
‡
À
tonitrile, with Bu4N BF4 as support-
ing electrolyte (0.1m LÀ1). The working
electrode was platinum, and a saturat-
ed calomel electrode (SCE) was used
À 1.36
À 0.739
2.17
1.24
À 0.663
0.552
0.93
À 0.83
À 0.477
0.242
0.59
À 0.23
À3
as
a
reference. The scan rate was
Electrolysis was per-
D1max [e Š
D1min [e Š
]
100 mVsÀ1
.
À3
]
À 0.74
formed at À1000 mV using platinum
[a] R ˆ S(j j Foj À jFcjj)/S(jFo j ). [b] Rw ˆ Sw(jFoj À jFc j )2/Sw(Fo)2]1/2
.
foil as a working electrode.
where h and c are the Planckꢁs constant and the speed of the light,
respectively, and a is the Onsager radius, estimated to be equal to 4.7 Š and
5.5 Š (for 2b and 3, respectively) from the solute molar volume[46] and the
crystal data available. However, the evaluation of Dm by the Lippert ±
Mataga treatment rises several problems, which have been discussed in
the literature.[47] Therefore, this approach may not be fully reliable.
Acknowledgement
The authors deeply acknowledge Jean-François Delouis, Isabelle Fanton-
Maltey and Olivier Guilbaud for their help in EFISH measurements, and
Dr. de Montauzon for the electrochemical measurement.
NLO properties
Powder efficiencies: The measurements of second harmonic generation
(SHG) intensity were carried out by the Kurtz ± Perry powder technique,[48]
using a picosecond Nd/YAG pulsed (10 Hz) laser operating at l ˆ 1.064 mm.
The outcoming Stokes-shifted radiation at 1.907 mm generated by Raman
effect in a hydrogen cell (1 m long, 50 atm) was used as the fundamental
beam for second harmonic generation. The SHG signal was selected
through a suitable interference filter, detected by a photomultiplier and
recorded on an ultrafast Tektronic TDS 620 B oscilloscope.
[1] a) P. N. Prasad, D. J. Williams, Introduction to Nonlinear Optical
Effects in Molecules and Polymers, Wiley, Chichester, 1991; b) Mo-
lecular Nonlinear Optics: Materials, Physics, and Devices (Ed.: J.
Zyss), Academic Press, New York, 1994; c) Nonlinear Optical Proper-
ties of Organic Materials VII, Vol. 2285 (Ed. G. R. Mohlmann), Proc.
SPIE, The International Society for Optical Engineering: Washington,
DC, 1994.
EFISH measurements: The principle of the electric field induced second
harmonic (EFISH) technique is reported elsewhere.[20, 49] The data were
recorded using the 1.907 mm incident laser beam generated as described
above. The laser delivered pulses of 10 ns. The compounds were dissolved
in dioxane at various concentrations (0 to 5 Â 10À3 mollÀ1). The centro-
symmetry of the solution was broken by dipolar orientation of the
chromophores with a high voltage pulse (around 5 kV) synchronized with
the laser pulse. The dipole moment were measured independently by a
[2] For recent reviews see: a) L. R. Dalton, A. W. Harper, R. Ghosn,
W. H. Steier, M. Ziari, H. Fetterman, Y. Shi, R. V. Mustacich, A. K-Y.
Jen, K. J. Shea, Chem. Mater. 1995, 7, 1060 ± 1081; b) R. G. Benning, J.
Mater. Chem. 1995, 5, 365 ± 378; c) Optical Nonlinearities in Chemistry
(Ed. D. M. Burland), Chem. Rev. 1994, 94, Issue 1; d) T. Verbiest, S.
Houbrechts, M. Kauranen, K. Clays, A. Persoons, J. Mater. Chem.
1997, 7, 2175 ± 2189; e) J. J. Wolff, R. Wortmann, Adv. Phys. Org.
Chem. 1999, 32, 121 ± 217.
Chem. Eur. J. 2000, 6, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2607 $ 17.50+.50/0
2607

P. G. Lacroix, G. Iftime et al.
FULL PAPER
[3] S. R. Marder, D. N. Beratan, L. T. Cheng, Science 1991, 252, 103 ± 106.
[4] a) D. R. Kanis, T. J. Marks, M. A. Ratner, Int. J. Quant. Chem. 1992,
43, 61 ± 82; b) D. R. Kanis, M. A. Ratner, T. J. Marks, Chem. Rev. 1994,
94, 195 ± 242.
[26] J. Zyss, J. L. Oudar, Phys. Rev. A 1982, 26, 2028 ± 2048.
[27] On the correlation of dipole moment with acentric space groups, see:
J. K. Whitesell, R. E. Davis, L. L. Saunders, R. J. Wilson, J. P. Feagins,
J. Am. Chem. Soc. 1991, 113, 3267 ± 3270.
[5] a) L. T. Cheng, W. Tam, S. H. Stevenson, G. R. Meredith, G. Rikken,
S. R. Marder, J. Phys. Chem. 1991, 95, 10631 ± 10643; b) L. T. Cheng,
W. Tam, S. R. Marder, A. E. Stiegman, G. Rikken, C. W. Spangler, J.
Phys. Chem. 1991, 95, 10643 ± 10652.
[6] V. Pushkana Rao, A. K.-Y. Jen, J. Chandrasekhar, I. N. Namboothiri,
A. Rathna, J. Am. Chem. Soc. 1996, 118, 12443 ± 12448.
[7] S. M. LeCours, H.-W. Guan, S. G. DiMagno, C. H. Wang, M. J.
Therien, J. Am. Chem. Soc. 1996, 118, 1497 ± 1503.
[28] a) K. D. Singer, J. E. Sohn, L. A. King, H. M. Gordon, H. E. Katz,
C. W. Dirk, J. Opt. Soc. Am. B 1989, 6, 1339 ± 1350; b) F. Gerson, T.
Gaumann, E. Heilbronner, Helv. Chim. Acta 1958, 41, 1481 ± 1491.
[29] C. R. Moylan, R. J. Twieg, V. Y. Lee, S. A. Swanson, K. M. Betterton,
R. D. Miller, J. Am. Chem. Soc. 1993, 115, 12599 ± 12600.
[30] M. L. Glowka, Z. Olubek, Acta Crystallogr. C 1994, 50, 458 ± 460, and
references herein.
[31] J. L. Oudar, J. Chemla, J. Chem. Phys. 1977, 66, 2664 ± 2668.
Â
[8] 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] H. J. Tobler, A. Bauder, H. H. Gunthard, J. Mol. Spectrosc. 1965, 18,
239.
[32] See for example: a) V. Alain, S. Redoglia, M. Blanchard-Desce, S.
Lebus K. Lukaszuk, R. Wortmann, U. Gubler, C. Bosshard, P. Günter,
Chem. Phys. 1999, 245, 51 ± 71; b) M. Blanchard-Desce, V. Alain, L.
Midrier, R. Wortmann, S. Lebus, C. Glania, P. Kramer, A. Fort, J.
Muller, M. Barzoukas, J. Photochem. Photobiol. A 1997, 195, 115 ± 121.
[33] See for example: a) D. R. Kanis, M. A. Ratner, T. J. Marks, J. Am.
[10] K. P. Zeller, in Houben Weyl, Methoden der Organischen Chemie, Vol.
5/2c, Thieme, Stuttgart, 1985, p. 127.
Á
Chem. Soc. 1992, 114, 10338 ± 10357; b) S. Di Bella, I. Fragala, I.
[11] S. Schmitt, M. Baumgarten, J. Simon, K. Hafner, Angew. Chem. 1998,
110, 1129 ± 1133; Angew. Chem. Int. Ed. 1998, 37, 1078 ± 1081.
[12] J. E. Frey, A. M. Andrews, S. D. Combs, S. P. Edens, J. J. Puckett, R. E.
Seagle, L. A. Torreano, J. Org. Chem. 1992, 57, 6460 ± 6466.
[13] S. Ito, S. Kikuchi, N. Morita, T. Asao, Chem. Lett. 1996, 175 ± 176.
[14] a) J. O. Morley, J. Am. Chem. Soc. 1988, 110, 7660 ± 7663; b) J. O.
Morley, J. Chem. Soc. Perkin Trans. 2 1989, 103 ± 106.
[15] A. E. Asato, R. S. H. Liu, V. Pushkara Rao, Y. M. Cai, Tetrahedron
Lett. 1996, 37, 419 ± 422.
[16] G. Iftime, P. G. Lacroix, K. Nakatani, A. C. Razus, Tetrahedron Lett.
1998, 39, 6853 ± 6856.
[17] R. Hermann, B. Pedersen, G. Wagner, J.-H. Youn, J. Organomet.
Chem. 1998, 571, 261 ± 266.
[18] a) F. Gerson, E. Heilbronner, Helv. Chim. Acta 1958, 41, 1444 ± 1463;
b) F. Gerson, J. Schulze, E. Heilbronner, Helv. Chim. Acta 1958, 41,
1463 ± 1481; c) A. G. Anderson, Jr., J. A. Nelson, J. J. Tazuma, J. Am.
Chem. Soc. 1953, 75, 4980 ± 4989.
[19] Among recent examples for the NLO use of DR1, see: a) D. H. Choi,
J. H. Park, T. H. Rhee, N. Kim, S.-D. Lee, Chem. Mater. 1998, 10, 705 ±
709; b) N. Tirelli, U. W. Suter, A. Altomare, R. Solaro, F. Ciardelli, S.
Follonier, Ch. Bosshard, P. Günter, Macromolecules 1998, 31, 2152 ±
2159; c) H. Jiang, A. K. Kakkar, Macromolecules 1998, 31, 2501 ±
2508.
Ledoux, T. J. Marks, J. Am. Chem. Soc. 1995, 117, 9481 ± 9485.
[34] J. L. Oudar, J. Zyss, Phys. Rev. A 1982, 26, 2016 ± 2027.
[35] B. F. Levine, C. G. Bethea, C. D. Thurmond, R. T. Lynch, J. L.
Bernstein, J. Appl. Phys. 1979, 50, 2523 ± 2527.
[36] J. Zyss, D. S. Chemla, J. F. Nicoud, J. Chem. Phys. 1981, 74, 4800 ± 4811.
Á
[37] S. Di Bella, I. Fragala, T. J. Marks, M. A. Ratner, J. Am. Chem. Soc.
1996, 118, 12747 ± 12751.
[38] a) R. Yam, R. Cohen, G. Berkovic, Nonlinear Opt. 1995, 11, 311;
b) P. M. Lundquist, S. Yitzchaik, T. J. Marks, G. K. Wong, S. Di Bella,
R. Cohen, G. Berkovic, Phys. Rev. B 1997, 55, 14055 ± 14058.
[39] I. Malfant, N. Cordente, P. G. Lacroix, C. Lepetit, Chem. Mater. 1998,
10, 4079 ± 4087.
[40] B. Bosnich, J. Am. Chem. Soc. 1968, 90, 627 ± 632.
[41] A. A. S. Briquet, H.-J. Hansen, Helv. Chim. Acta 1994, 77, 1577 ± 1584.
[42] G. M. Sheldrick, SHELXS86, Program for Crystal Structure Solution,
University of Göttingen, Göttingen, Germany, 1986.
[43] a) D. J. Watkin, C. K. Prout, J. R. Carruthers, P. W. Betteridge,
CRYSTALS Issue 10, Chemical Crystallography Laboratory, Univer-
sity of Oxford, Oxford, 1996; b) D. J. Watkin, C. K. Prout, L. J. Pearce,
CAMERON, Chemical Crystallography Laboratory, University of
Oxford, Oxford, 1996.
[44] D. T. Cromer, J. T. Waber, International Tables for X-ray Crystallog-
raphy, Vol. 4, Kynoch Press, Birmingham, 1974.
[20] For a general discussion of the EFISH technique, see: a) J. L. Oudar, J.
Chem. Phys. 1977, 67, 446 ± 457; b) B. F. Levine, C. G. Betha, J. Chem.
Phys. 1975, 63, 2666 ± 2682, B. F. Levine, C. G. Betha, J. Chem. Phys.
1976, 65, 1989 ± 1993.
[45] N. Mataga, T. Kubota, Molecular Interactions and Electronic Spectra,
Marcel Dekker, New York, 1970, p. 371.
[46] S. Di Bella, T. J. Marks, M. A. Ratner, J. Am. Chem. Soc. 1994, 116,
4440 ± 4445.
[21] a) J. O. Morley, V. J. Dorcherty, D. Pugh, J. Chem. Soc. Perkin Trans. 2
1987, 1351 ± 1355; b) J. O. Morley, J. Chem. Soc. Faraday Trans. 2 1991,
87, 3009 ± 3013; c) J. O. Morley, Int. J. Quantum Chem. 1993, 46, 19 ±
26; d) J. O. Morley, J. Chem. Soc. Perkin Trans. 2 1995, 731 ± 734.
[22] D. Li, M. A. Ratner, T. J. Marks, J. Am. Chem. Soc. 1988, 110, 1707 ±
1715.
[23] P. A. Franken, A. E. Hill, C. W. Peters, G. Weinrich, Phys. Rev. Lett.
1961, 7, 118 ± 119.
[24] Among early attempts to use DR1 in polymers for NLO purpose, see:
K. D. Singer, J. E. Sohn, S. J. Lalama, Appl. Phys. Lett. 1986, 49, 248 ±
250.
[47] W. Liptay, Z. Naturforschg. 1965, 20a, 1441 ± 1471.
[48] a) S. K. Kurtz, T. T. Perry, J. Appl. Phys. 1968, 39, 3798 ± 3813; b) J. P.
Dougherty, S. K. Kurtz, J. Appl. Crystallogr. 1976, 9, 145 ± 158.
[49] I. Maltey, J. A. Delaire, K. Nakatani, P. Wang, X. Shi, S Wu, Adv.
Mater. Opt. Electronics 1996, 6, 233 ± 238.
[50] E. A. Guggenheim, Trans. Faraday Soc. 1949, 45, 714 ± 720.
[51] a) M. C. Zerner, G. Loew, R. Kirchner, U. Mueller-Westerhoff, J. Am.
Chem. Soc. 1980, 102, 589 ± 599; b) W. P. Anderson, D. Edwards, M. C.
Zerner, Inorg. Chem. 1986, 25, 2728 ± 2732.
[52] J. F. Ward, Rev. Mod. Phys. 1965, 37, 1 ± 18.
[25] D. J. Williams, Angew. Chem. 1984, 96, 637 ± 650; Angew. Chem. Int.
Ed. Engl. 1984, 23, 690 ± 703.
Received: August 25, 1999
Revised version: November 26, 1999 [F2002]
2608
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0614-2608 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 14