The third-order NLO properties of D-π-A molecules with changing a primary amino group into pyrrole

Article abstract of DOI:10.1016/S0009-2614(02)01575-0

The nonlinear refractive index, as well as the two-photon absorption (TPA), has been investigated by the femto-second Z-scan method for push-pull molecules, in which a primary donor group is weakened by conversion into a pyrrole or related heterocycle. The results at the fast response time show that, contrary to what had been expected from a previous work, there is a little decrease in the nonlinear refractive index change by introduction of the weaker donor groups. Although the measured TPA spectra showed that the observed nonlinear refractive index change was regarded to be off-resonant; difference in the frequency dispersion among the molecules considerably contributes to the decrease, in addition to difference in the intrinsic static nonlinearity. This shows that there is some future in the development of these non-resonant materials.

Full text of DOI:10.1016/S0009-2614(02)01575-0

Chemical Physics Letters 367 (2003) 62–71
www.elsevier.com/locate/cplett
The third-order NLO properties of D-p-A molecules
with changing a primary amino group into pyrrole
a,
b
b
P. Audebert *, K. Kamada b,1, K. Matsunaga , K. Ohta
a
Laboratoire de Photophysique et Photochimie Supramoleꢀculaires et Macromoleꢀculaires, Ecole Normale Superieure de Cachan,
61 Avenue du President Wilson, 94235 Cachan Cedex, France
b
Photonics Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), AIST Kansai Center,
1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan
Received 19 June 2002; in final form 30 September 2002
Abstract
The nonlinear refractive index, as well as the two-photon absorption (TPA), has been investigated by the femto-
second Z-scan method for push–pull molecules, in which a primary donor group is weakened by conversion into a
pyrrole or related heterocycle. The results at the fast response time show that, contrary to what had been expected from
a previous work, there is a little decrease in the nonlinear refractive index change by introduction of the weaker donor
groups. Although the measured TPA spectra showed that the observed nonlinear refractive index change was regarded
to be off-resonant; difference in the frequency dispersion among the molecules considerably contributes to the decrease,
in addition to difference in the intrinsic static nonlinearity. This shows that there is some future in the development of
these non-resonant materials.
Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction
concerns materials exhibiting nonlinear refractive
index change [6–11]. For such nonlinear optical
molecules and materials, it is desirable to optimize
the c values, while keeping off-resonance for both
one-photon and two-photon absorptions (TPA).
One of us (PA) published some time ago that it
could be possible to increase the c values while
escaping resonance, simply by decreasing the
strength of the donor group, in classical push–pull
chromophores consisting of p-transmitter with
electron withdrawing and donating groups at both
ends [12]. Among such chromophores, two fami-
lies of molecules had been investigated, where a
primary amino group of the molecules in the first
There has been an increasing amount of work
on nonlinear optical (NLO) materials [1–5]. While
the engineering for enhancing second-order NLO
efficiency is relatively well understood and
achieved, the need for efficient third-order mole-
cules and materials still exists, especially for what
* Corresponding author. Fax: +33-1-4740-2454.
E-mail addresses: pierre.audebert@ppsm.ens-cachan.fr (P.
Audebert), k.kamada@aist.go.jp (K. Kamada).
1
Also Corresponding author. Fax: +81-727-51-9637.
0009-2614/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0009-2614(02)01575-0

P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
63
family {p-nitroaniline (PNA), p-cyanoaniline
(PCA) and disperse orange 3 (DO3)} had been
changed into an imide in the second one, while
keeping the rest of the molecule (transmitter and
electron withdrawing group) unchanged. The ex-
planation proposed for the increase in c of the
molecules in the second family as the dipole mo-
ment of the molecule in the ground state was de-
creased, while the dipole moment in the excited
state kept more or less the same value, which
would lead to a neat increase of the c value on the
basis of a simple two-level model [13]. At that
time, the measurements were performed by Z-scan
technique with a 50-ps pulsed laser beam. With
such a duration pulse, however, effects of nuclear
motions including the orientational effect on the
resultant c value could not be completely excluded.
In addition, the measurements had been made at
the sole frequency of 1060 nm, which was of course
off resonance for the molecules concerned, but the
frequency dependence was not investigated.
It was therefore, worthwhile to extend this
study, in order both to find the limits and possibly
ameliorate our former interpretations, as well also
to extend our investigations to new families of
molecules. In our former study we measured and
analyzed PNA, PCA, DO3 and their correspond-
ing imides (designated as IPNA, IPCA and IDO3).
In the present study, we describe new results on
one of the former molecule PNA and its imide
counterpart IPNA, as well as new molecules de-
rived from PNA and DO3 by the conversion of the
amino group into a pyrrole ring in the case of PNA
and DO3 and a (1,5-bisthienyl)-pyrrole ring in the
case of PNA (Scheme 1). As a matter of fact,
conversion of the primary amino group into a
pyrrole ring is another interesting way to diminish
the donor strength since the nitrogen doublet be-
comes a part of the aromatic electrons of the
pyrrole ring.
In this study, we have also investigated the
TPA spectra, as well as the off-resonant nonlinear
Scheme 1. Molecular structures of the molecules studied.

64
P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
refractive index, by the femtosecond Z-scan
method for the molecules above. The reason of
this is that it is essential to know the TPA spectra
for evaluating the resonance enhancement of
the real part of the c value. Additionally the TPA
property of the organic compounds is one of
the current topics, and the present results could
make some contribution to establishing the
structure–property relation for such p-conjugated
systems.
For TERPNA synthesis, the 1,4-dithienyl c-
butyrodicetone [18,19] (1 g, 4 mmoles) and PNA
(550 mg, 4 mmoles) were mixed in a round bot-
tomed flask with p-toluenesulphonic acid (70 mg,
0.4 mmole), and refluxed for 14 h. Upon cooling, a
solid was separated, which is usually pure TER-
PNA. In order to obtain high purity compound,
a purification chromatography is performed in
dichloromethane/petroleum ether 1/1 mixture.
1H NMR (350 MHz):
Furthermore, it should be remarked that not
only the pyrrole (or oligopyrrole) rings act as
electron withdrawing structures, but they can also
serve as polymerizable groups, due to the well-
known electropolymerization of the pyrrole de-
rivatives [14]. Therefore the intrinsic potential of
such molecules is clearly beyond the initial goal,
and they could well be the basis of new performing
NLO polymers.
For PYPNA: 8.32 (dd, 2H, phenyl-NO2); 7.52
(dd, 2H, phenyl-NO2); 7.20 (t, 2H, a-pyrrole); 6.43
(t, 2H, b-pyrrole).
For PYDO3: 8.40 (m, 2H, phenyl-NO2); 8.05 (t,
4H, phenyl-azo); 7.56 (d, 2H, phenylpyrrole); 7.20
(t, 2H, a-pyrrole); 6.40 (t, 2H, b-pyrrole).
For TERPNA: 8.23 (dd, 2H, phenyl-NO2); 7.41
(dd, 2H, phenyl-NO2); 7.15 (dd, 2H, a-thiophene);
6.87 (dd, 2H, b-thiophene); 6.56 (s, 2H, b-pyrrole);
6.55 (dd, 2H, b-thiophene). Mþ ¼ 352 (ioniza-
tion).
2. Experimental part
2.2. Optical measurements
2.1. Synthesis of the molecules
The sample solutions were prepared by dis-
solving the molecules in spectroscopic-grade dim-
ethylsulfoxide (DMSO). The concentrations of the
sample solution were adjusted so that the open
aperture Z-scan signals could be detected (0.50 M
for PNA and IPNA, 0.20 M for PYPNA, 0.32 M
for TERPNA and 0.012 M for DO3 and PYDO3).
The UV–Vis (one-photon) absorption spectra of
the sample solution were recorded by a Shimadzu
UV-3150 spectrometer. The sample solutions in
high concentration were sandwiched between
fused quartz plates and were used for the UV–Vis
measurements. The optical path length of the
solution film was adjusted in order to yield ab-
sorbance below 2. The recorded one-photon ab-
sorption spectrum of the solution film was little
different from those of the diluted solution mea-
sured with a 1-cm quartz cuvette.
PNA and DO3 are commercial compounds and
were purchased (Aldrich Chemicals) and used
without further purification. The synthesis of and
IPNA was described before [12] and was per-
formed according to the method of Toru and co-
workers [15].
PYPNA (a previously known compound
[16,17]) and PYDO3 were performed by the clas-
sical Paal–Knorr reaction, while TERPNA was
prepared by extending a previously published
procedure [18,19]. Typically, for the Paal–Knorr
reaction the conditions were the following:
In a round-bottomed flask were introduced
4 mmol of the amino dye and 5 mmol of di-
methoxytetrahydrofuran, with 20 ml of glacial
acetic acid. The mixture is heated on an oil bath
until it refluxes, after which the heating is imme-
diately removed, and the content of the flask is
poured on a mixture of sodium hydroxide and ice.
After classical work-up with dichloromethane and
chromatography (petroleum ether 1/dichlorome-
thane 1), the desired compound is obtained
(Rf ¼ 0.7) with 60–70% yield.
Third-order nonlinear optical properties of the
sample solutions were measured by the Z-scan
technique [20]. The optical setup is illustrated in
Fig. 1. An optical parametric amplifier (Spectra-
Physics OPA-800) pumped by Ti:sapphire regen-
erative amplifier system (Spectra-Physics Spitfire,

P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
65
Fig. 1. Schematic outline of the femtosecond Z-scan setup. BS, beam sampler; QDP, quad-photodiode; CP, mechanical chopper;
BPC, Berek polarization compensator; GTP, Glan-Laser prism; ND, neutral density filter; L, lens; S, sample; AP, iris aperture;
PD, photodiode.
q0 ¼ að2Þð1 À RÞI0L;
ð1Þ
Merlin, Tsunami and Millennia) operating at the
repetition rate of 1 kHz was used as light source.
The output pulse width was 105–130 fs. The rep-
etition rate used for measurement can be reduced
without changing the beam property by mechani-
cal choppers placed between the optical parametric
amplifier and the measurement optics. A couple of
quad-photodiodes were served to monitor the
stability of beam position during the measure-
ments. The sample solution in a 1-mm quartz cu-
vette was put on a power stage and was scanned
along the optical axis. Three photodiodes con-
nected to box-car averagers were used for signal
detection. The closed and open aperture signals
were corrected for nonlinear absorption and inci-
dent intensity as illustrated in Fig. 1. The closed
aperture measurements were performed at the re-
pletion rate of 10 Hz to suppress the accumulated
thermal effect and the open aperture measurements
were performed at 1 kHz. No spectral change
of UV–Vis absorption was confirmed before and
after the measurements of each sample.
where að2Þ is the TPA coefficient, R is the Fresnel
reflectance at the sample cell wall, I0 is the on-axis
peak intensity, and L is the optical path length of
the sample cell. Eq. (1) holds for the measurements
when the one-photon absorption is absent as in the
present case. The parameter q0 was obtained by
curve fit using the relation [21]
Z
þ1
ð1 À RÞ
TðfÞ ¼
ln½1 þ qðfÞ expðÀx2ފdx;
pffiffiffi
2
pqðfÞ
À1
ð2Þ
for the spatially and temporally Gaussian pulse
along with qðfÞ ¼ q0=ð1 þ f2Þ. Here T is the nor-
malized transmittance, f ¼ z=zR is the normalized
z position, where zR is the Rayleigh range. The zR
values were 3.5–6 mm for the measurements. We
measured at least 4 different incident powers and
confirmed that TPA is the dominant process for
the observed nonlinear absorption by the linear
relation between the obtained q0 and I0 shown as
in Eq. (1). The TPA cross-section rð2Þ is calculated
from að2Þ with the relation rð2Þ ¼ hxað2Þ=2pN,
where hx=2p is the photon energy of the incident
pulse and N is the number density of the solute
molecule. I0 was calculated from the incident pulse
energy, the pulse width, and the beam waist radius
calculated from zR.
2.3. Analysis of the Z-scan data
Both the open and closed aperture data were
recorded in normalized transmittance as a function
of sample position (z). The open aperture data was
analyzed with TPA parameter q0 defined as [21]

66
P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
where f ¼ ðn20 þ 2Þ=3 is the local field correction.
The closed aperture data were analyzed with the
In the following part, the symbols of Re[ ] and the
suffix xxxx are omitted for simplification.
difference between normalized transmittance at the
peak and that at the valley, so-called DTpv, which is
proportional to the nonlinear index of refraction
n2 and the real part of the nonlinear susceptibility
ð3Þ
Re½v Š as
xxxx
3. Results and discussion
I0L
n02
ð3Þ
DTpv / n2I0L /
Re½v Š;
ð3Þ
xxxx
3.1. Results on the nitrophenyl derivatives (PNA,
IPNA, PYPNA and TERPNA)
where n0 is the linear refractive index. We deter-
ð3Þ
mined the Re½v Š of the sample solution with the
xxxx
First, the one- and two-photon absorption
(TPA) spectra were measured because it was es-
sential to know which wavelength region is non-
resonant for further optical measurements. Figs. 2
and 3 display the results of TPA for both couples
PNA/IPNA and PYPNA/TERPNA measured by
the open aperture Z-scan method along with their
one-photon absorption spectra. The TPA proper-
ties are also summarized in Table 1. Although the
TPA peak of IPNA was not clear, the TPA bands
for all the compounds except TERPNA are ob-
served at approximately twice the wavelength of
the lowest energy band of one-photon absorption,
ð3Þ
Re½v Š of a high refractive-index glass (FD2) as
xxxx
reference material measured at the same experi-
mental condition. The value of FD2 was deter-
mined based on that of fused silica [22]. We also
measured a blank solution in the cell under the
identical condition for each measurement and the
contribution from the solvent and the cell walls
ð3Þ
were subtracted before calculating Re½v Š. The
xxxx
real part of the second hyperpolarizability was
calculated based on the relation
ð3Þ
Re½v
Š
;
ð3Þ
xxxx
Re½c Š ¼
ð4Þ
xxxx
f 4N
Fig. 2. One-photon absorption (A) and TPA (B) spectra of
PNA and IPNA. The curves in the panel B are just for eye
guide.
Fig. 3. One-photon absorption (A) and TPA (B) spectra of
PYPNA and TERPNA. The curves in the panel B are just for
eye guide.

P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
67
Table 1
TPA properties of the molecules
b
Molecule
kmax;TPA
(nm)
rð2Þa
ð10À50 cm4 sÞ
Im ½cŠ
(10À36 esu)
PNA
IPNA
ꢀ714
(<590)
ꢀ700
(<570)
932
8:5 Æ 1:7
25 Æ 6:0
ð>4:8 Æ 1:7Þ
14:1 Æ 2:4
ð>9:8 Æ 3:9Þ
40:6 Æ 8:3
ð>40:2 Æ 7:8Þ
810 Æ 160
370 Æ 63
PYPNA
TERPNA
DO3
ð>21:1 Æ 3:3Þ
159:3 Æ 25:9
84:3 Æ 11:4
PYDO3
865
a TPA cross-section measured at 715, 595, 701, and 572 nm for PNA, IPNA, PYPNA and TERPNA, respectively.
b Converted from the rð2Þ values.
exhibiting that TPA transition shares the same
excited state with one-photon absorption and no
hidden TPA state exists below the excited state.
Sharing the excited state with one- and two-pho-
ton transitions is reasonably explained in terms of
molecular symmetry because the parity selection
rule of one- and two-photon transitions breaks for
asymmetric; structures like these molecules.
For IPNA, the TPA peak cannot be clarified
due to the wavelength limitation of our experi-
mental setup; however, the peak is estimated
shorter than 600 nm, which is close to twice the
one-photon absorption maximum (280 nm). The
one-photon absorption and TPA peaks of IPNA
are blue-shifted comparing to PNA and PYPNA.
The blue shift of IPNA arises from rather weak
electron donating ability of imide group than those
of pyrrolyl and amino groups.
On the other hand, the TPA peak of TERPNA
should be located at shorter than 570 nm and is
not observed at twice the wavelength of the one-
photon peak (323 Â 2 ¼ 646 nm). This suggests
that the lowest TPA excited state of TERPNA
must be different from the lowest one-photon
excited state. For other three compounds, the
directions of the transition moments of both one-
and two-photon excited states are expected to be
parallel to the phenylene-NO2 axis. On the other
hand, the TPA transition moment for TERPNA is
not parallel to the phenylene-NO2 axis, but par-
allel to the thienyl-pyrrolyl-thienyl axis. The cor-
responding one-photon excitation may be
forbidden due to the symmetry property for the
direction. The rð2Þ maximum increased from PNA,
to PYPNA and to TERPNA, which may correlate
to the molecular size and the charge transfer na-
ture of the excitation concerned.
Second, the nonlinear refractivity of the com-
pounds was measured by the closed aperture Z-
scan method. The second hyperpolarizability
dominating the nonlinear refractive process c(Àx;
x, x, Àx) is resonantly enhanced by only one-
photon and two-photon absorptions [13]. The re-
sults of the above linear and nonlinear absorption
measurements shows that there is no one- and two-
photon absorptions at the wavelength longer than
900 nm for all the compounds. It is also needed to
pay attention in near IR region for the overtone of
IR absorption of the sample solution. In the case
of DMSO, the overtones are observed at 1100–
1200 nm and 1350–1500 nm. Therefore, the closed
aperture Z-scan measurements at 1007 or 1016 nm
can be regarded as off-resonant condition for the
solvent as well as for all compounds. The obtained
c values are given in the Table 2. We can observe
that weakening the donor group from amino
(PNA) to pyrrolyl (PYPNA), to 2,5-dithienyl-
pyrrolyl (TERPNA) and to maleimide (IPNA)
results into a decrease in c, contrary to what had
been reported previously [12].
The second hyperpolarizability is often ex-
pressed by the following three-level model as
2
2
2
l201ðl0 À l1Þ
l401
x13
l01l1n
x21xn
X
cð0;0;0;0Þ /
À
þ
ð5Þ
x13
n
Here xn is the transition energy between the
ground and nth excited states, l01 and l1n are
transition moments between the ground and low-
est excited states and between the lowest and nth
excited states, respectively, and l0 À l1 denotes the

68
P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
Table 2
Real part of off-resonant second hyperpolarizability
cb
a
Molecule
Concentration
(M)
kmax;OPA
(nm)
(10À36 esu)
PNA
IPNA
0.496
0.499
0.200
388
288
340
323
472
391
þ20:0 Æ 6:01
þ3:79 Æ 2:13
þ12:0 Æ 7:22
þ6:2 Æ 29:2c
þ256 Æ 94c
þ137 Æ 70c
PYPNA
TERPNA
DO3
0.0322
0.0119
0.0123
PYDO3
a Wavelength of one-photon absorption maximum.
b Here c denotes the real part of the second hyperpolarizability, Re[c]. Measured at 1007 nm for PNA and IPNA, at 1016 nm for
PYPNA and TERPNA and at 1280 for DO3 and PYDO3. The signal from the solvent (DMSO) was subtracted. The convention used
for c is based on the power series as lind ¼ aE þ bEE þ EEE, where lind is the induced dipole moment, a and b denote linear po-
larizability and first hyperpolarizability, respectively, E is microscopic electric field.
c Large error due to low solubility.
1
difference in the dipole moments of the two states.
This expression can be considered as the approxi-
mated formula of the following more complicated
equation by ignoring the frequency factor (x ! 0).
F ¼
:
2
ð8Þ
2
½1 À ðx=x1Þ Š½1 À ð2x=x1Þ Š
The estimated frequency factors F for the present
)
experimental optical frequency (x ꢀ 9900 cmÀ1
cðÀx; x; x; ÀxÞ
are 2.89, 1.62 and 2.04, for PNA, IPNA and PY-
PNA, respectively, depending on the one-photon
absorption peak positions (TERPNA is excluded
because it does not seem to be suitable for this
two-level analysis). By using the frequency factor,
the static c can be deduced as 6.9, 2.3 and 5.9 (in
10À36 esu) for PNA, IPNA and PYPNA, respec-
tively. Therefore, it is very probable that the de-
crease in the obtained c values from PNA to IPNA
considerably arises from the frequency dispersion
as well as the intrinsic static c, originally from the
blue shift of the one-photon absorption peak with
weakened donor. Probably the previous two-level
model may be too simplified to explain whole the
observed results; nevertheless one can say that the
transition moment factors in the numerator of
Eq. (7) are estimated to be in the same order for all
compounds.
2
l201ðl0 À l1Þ x1
l401x1ðx21 þ x2=3Þ
/
À
ðx21 À x2Þðx12 À 4x2Þ
3
ðx21 À x2Þ
2
2
2
2
X
l01l1nfx1xn þ x ð4x1 þ xnÞg
þ
:
ð6Þ
2
ðx21 À x2Þ ðxn2 À 4x2Þ
n
However, for the present compounds like PNA,
IPNA and PYPNA, where the TPA transition
shares the same excited state with one-photon
absorption and the molecule has a large dipole
moment at the ground state, only the first term is
expected to be dominant and can be employed for
1
the analysis as in the previous paper ,
2
l201ðl0 À l1Þ x1
cðÀx; x; x; ÀxÞ /
:
ð7Þ
ðx21 À x2Þðx12 À 4x2Þ
Even if the c values were measured at off-resonant
conditions, we can expect enhancement in the c
values due to the frequency factor in the denomi-
nator of Eq. (7). Based on Eq. (7), we can estimate
the enhancement factor compared to the static
values at the limit of zero frequency using the
following equation:
This explanation can be supported by the recent
ab initio molecular orbital calculation of the static
c value for PNA, IPNA and PYPNA [23]. The
calculated results show that the static c values for
the three compounds are in the same order.
Moreover the ab initio calculations reveal that
both IPNA and PYPNA take the nonplanar
structures by tilting the donating groups at the
1
Eq. (7) is more rigorous equation than that in [12] as the
term concerning the virtual transition path of 0-1-1-1-0.

P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
69
optimized geometries. This should cause the par-
tial disruption of the p-conjugated structures,
which contributes to the blue shift of the absorp-
tion peak and also to decrease in the intrinsic static
c values despite the larger size of the molecules
than PNA. Moreover, the calculation shows in-
dications of possible contributions from the
higher-excited states to the c values for IPNA and
PYPNA, which is not taken into account by the
simple two-level model.
Fig. 4 represents both one-photon absorption
and TPA spectra for both DO3 and PYDO3. Each
maximum of TPA spectrum is located at almost
twice the wavelength of the corresponding one-
photon absorption maximum. The maxima of rð2Þ
values of DO3 and PYDO3 were larger than those
of the nitrophenyl derivatives, exhibiting the effect
of longer conjugation length in DO3 family. At
the wavelength longer than 1200 nm, there is no
TPA for both DO3 compounds and no overtones
of IR absorption of the solvent, so the closed
aperture Z-scan measurements were performed at
1280 nm.
The obtained c values at the wavelength are
ðþ256 Æ 94Þ Â 10À36 esu for DO3 and ðþ137 Æ 70Þ
Â10À36 esu for PYDO3 as also shown in Table 2.
These c values are much larger than those of the
nitrophenyl derivatives, reflecting the larger mo-
lecular size and p-conjugation length. This result
shows the trend similar to the previous results for
the nitro series, again corresponding to the one-
photon peak positions.
The previous interpretation that c value in-
creases with the weak donor group was a little
optimistic. The previously observed increase in
the c value from PNA to IPNA [12] can be ex-
plained by taking the molecular orientational ef-
fect into account. The previous c values were
measured with 50-ps pulses, which is close to the
rotational relaxation time of the molecule in so-
lution. It is natural to expect that the rotational
relaxation time of a strongly charge separated
molecule like PNA becomes longer because of the
interaction with the polar solvent molecules. If it
is so, IPNA, which is a weaker charge-separated
molecule, must have shorter rotational relaxation
time than PNA and is quickly reoriented along
the polarization direction of the incident pulse.
Therefore, the anisotropic linear polarizability
contributes to increase apparent c value because
usually the orientational contribution is much
larger than the electronic contribution at room
temperature. The previous c values are probably
the mixture of the orientational and electronic
nonlinearities. This artifact can be neglected for
the present measurements because the incident
pulse width (ꢀ110 fs) is two orders of magnitude
or more shorter than the expected rotational re-
laxation times.
3.2. DO3 family compounds
Beyond DO3, only PYDO3 was accessible in
this family, which is commercially available.
Unfortunately TERDO3 (the analog of PYDO3
but with two thiophene rings substituted on the
positions of the pyrrole ring like in TERPNA)
was not accessible despite repeated attempts using
various reaction conditions, the amino group of
DO3 being definitely too weakly reactive.
Fig. 4. One-photon absorption (A) and TPA (B) spectra of
DO3 and PYDO3. The curves in the panel B are just for eye
guide.

70
P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
Acknowledgements
If we apply the similar discussion concerning
frequency factors for DO3 and PYDO3 as done
for the PNA series, the estimated frequency
factors F (Eq. (8)) for the experimental optical
frequency (x ꢀ 7800 cmÀ1 in this case) are 2.54
and 1.76 for DO3 and PYDO3, respectively.
Then the static c values are deduced as 101 and
78 (in 10À36 esu) for DO3 and PYDO3, respec-
tively. Therefore, DO3 still seems to have a lar-
ger c value than PYDO3 at the static limit,
although the two values become close to each
other. Again this is probably caused by the fact
that PYDO3 contains a much weaker donor
group (pyrrolyl) than DO3 does (amino), and the
presence of the pyrrolyl group may disrupt the
p-conjugation.
P.A. would like to thank K.I. Chane-Ching and
ꢀ
F. Cherioux for many fruitful discussions and their
help in the synthetic work.
References
[1] D.S. Chemla, J. Zyss (Eds.), Nonlinear Optical Properties
of Organic Molecules and Crystals, Academic Press, NY,
1987.
[2] C. Ye, T.J. Marks, J. Yang, G.K. Wong, Macromolecules
20 (1987) 2322.
[3] C. Ye, T.J. Marks, J. Yang, G.K. Wong, Macromolecules
21 (1988) 2899.
[4] M. Alheim, M. Barzoukas, P.V. Bedworth, J.Y. Hu, S.R.
Marder, J.W. Perry, C. Stahelin, B. Zysset, Science 121
(1996) 335.
[5] X.M. Wu, J.Y. Wu, Y.Q. Liu, A.K.Y. Jen, J. Am. Chem.
Soc. 121 (1999) 472.
4. Conclusion
ꢀ
[6] M. Albota, D. Beljonne, J.L. Bredas, J.E. Ehrlich, J.-Y.
Fu, A.A. Heikal, S.E. Hess, T. Kogei, M.D. Levin, S.R.
We have described the preparation and the
third-order NLO properties in the femtosecond
range of two families including new molecules
possessing the same withdrawing group and p-
connecting system and donor groups with various
strength. The results show that the c (real part)
measured at the off-resonant wavelength de-
creases as the decreasing strength of donor group
in spite of the expectation from the previous re-
sults. The measurements of TPA and one-photon
absorption spectra exhibited that both of the
lowest energy TPA transition and the one-photon
transition shared the same initial and final states,
except for TERPNA. The analysis based on the
simplified two-level model considering the ground
and lowest excited states suggested that the vari-
ation of measured c among the molecules is
governed not only by the intrinsic static c but
also by the frequency dispersion. Thus, decreas-
ing the donor strength increases the transition
energy between the two states, and then decreases
the frequency factor at the observation wave-
length even if it is in the off-resonant region. The
present results suggested that not only optimizing
of the static c but also tailoring the transition
energy relating with the frequency factor at the
wavelength in use is important for the future
applications.
€
Marder, D. McCord-Maughon, J.W. Perry, H. Rockel, M.
Rumi, G. Subramaniam, W.W. Webb, X.-L. Wu, C. Xu,
Science 281 (1998) 1653.
[7] C. Andraud, R. Anemian, A. Collet, J.-M. Nunzi, Y.
Morel, P.L. Baldeck, J. Opt. A: Pure Appl. Opt. 2 (2000)
284.
[8] F.Z. Henari, W.J. Blau, L.R. Milgrom, G. Yahoglu, D.
Philips, J.A. Lacey, Chem. Phys. Lett. 267 (1997) 229.
[9] K. Kamada, M. Ueda, T. Sakaguchi, K. Ohta, T. Fukumi,
J. Opt. Soc. Am. B 15 (1998) 838.
[10] K. Kamada, T. Sugino, M. Ueda, K. Tawa, Y. Shimizu, K.
Ohta, Chem. Phys. Lett. 302 (1999) 615.
[11] K. Kamada, M. Ueda, H. Nagao, K. Tawa, T. Sugino, Y.
Shimizu, K. Ohta, J. Phys. Chem. A 104 (2000) 4723.
ꢀ
[12] F. Cherioux, H. Maillotte, J.M. Dudley, P. Audebert,
Chem. Phys. Lett . 319 (2000) 669.
[13] J.F. Ward, Rev. Mod. Phys. 37 (1965) 1.
[14] H.S. Nalwa (Ed.), For a Recent Overview on Conducting
Polymers and Especially Pyrroles, Wiley, New York, 1997,
see Handbook of organic conductive molecules and poly-
mers.
[15] P.Y. Reddy, S. Kondo, T. Toru, Y. Ueno, J. Org. Chem.
62 (1997) 2652.
[16] A.F. Diaz, J. Castillo, K.K. Kanazawa, J.A. Logan, M.
Salmon, O. Fajardo, J. Electroanal. Chem. 133 (1982)
233.
[17] M. Salmon, M. Aguikar, M. Saloma, J. Chem. Soc., Chem.
Commmun. (1983) 570.
[18] K.I. Chane-Ching, J.C. Lacroix, R. Baudry, M. Jouini, S.
Aeiyach, C. Lion, P.C. Lacaze, J. Electroanal. Chem. 453
(1998) 139.
[19] P.E. Just, K.I. Chane-Ching, J.C. Lacroix, P.C. Lacaze,
J. Electroanal. Chem. 479 (1999) 3.

P. Audebert et al. / Chemical Physics Letters 367 (2003) 62–71
71
[20] M. Sheik-Bahae, A.A. Said, T.-H. Wei, D.J. Hagan, E.W.
Van Stryland, IEEE J. Quantum Electron. 26 (1990) 760.
[21] R.L. Sutherland, Handbook of Nonlinear Optics, Mercel
Dekker, New York, 1996.
[22] A. Mito, K. Hagimoto, C. Takahashi, Nonlinear Opt. 13
(1995) 3.
[23] S. Yamada, K. Yamaguchi, K. Kamada, K. Ohta, Mol.
Phys. 100 (2002), in press.