Nature of the fluorescent state of N-arylcarbazole derivatives as derived from directly measured values of the excited state dipole moment

Article abstract of DOI:10.1021/jp0035481

Laser-induced changes in microwave dielectric loss of benzene solutions of two electron donor-acceptor systems, N-(4-cyanophenyl)carbazole and N-(1-naphthyl)carbazole, have been quantitatively measured with a view to determining the dipole moments and hence the nature of the fluorescent states of these systems. The dipole moments so determined are found to be much lower than that expected for a twisted intramolecular charge transfer state. On the basis of the measured values and the results of AM1 calculations, it is concluded that, contrary to what is commonly believed, these donor-acceptor systems emit from a locally excited state.

Full text of DOI:10.1021/jp0035481

5438
J. Phys. Chem. A 2001, 105, 5438-5441
Nature of the Fluorescent State of N-Arylcarbazole Derivatives as Derived from Directly
Measured Values of the Excited State Dipole Moment
A. Samanta,*,†,‡ S. Saha,† and R. W. Fessenden*,‡
School of Chemistry, UniVersity of Hyderabad, Hyderabad 500 046, India and Radiation Laboratory,
UniVersity of Notre Dame, Notre Dame, Indiana 46556-0579, USA
ReceiVed: September 28, 2000; In Final Form: March 30, 2001
Laser-induced changes in microwave dielectric loss of benzene solutions of two electron donor-acceptor
systems, N-(4-cyanophenyl)carbazole and N-(1-naphthyl)carbazole, have been quantitatively measured with
a view to determining the dipole moments and hence the nature of the fluorescent states of these systems.
The dipole moments so determined are found to be much lower than that expected for a twisted intramolecular
charge transfer state. On the basis of the measured values and the results of AM1 calculations, it is concluded
that, contrary to what is commonly believed, these donor-acceptor systems emit from a locally excited state.
Introduction
One of the most commonly employed methods for the
determination of the change in the dipole moment on electronic
excitation is based on the correlation between the difference in
the wavenumber of the absorption and fluorescence maxima
and a solvent polarity function, ususally defined in terms of
the dielectric constant and the refractive index of the medium.17-22
Among the numerous correlations available,21 the one suggested
by Lippert and Mataga is the most popular.17-19 However, this
method suffers from the disadvantage that the change in the
dipole moment of the molecule depends on the third power of
the “Onsager cavity radius”, the radius of the spherical cavity
in which the molecule (assumed to be a point dipole) is
embedded. This quantity is most often chosen somewhat
arbitrarily. Quite obviously, the estimated values of the excited-
state dipole moments are therefore subject to uncertainty. Hence,
the conclusions derived from these values, such as the nature
of the fluorescent state, may not be correct.
Taking into consideration the limitation of the solvatochromic
methods, we have recently embarked upon measurements of
the excited-state dipole moments of various electron donor-
acceptor molecules23,24 employing a technique that involves
measurement of the time-resolved changes in the dielectric loss
of a solution of the system of interest in a nonpolar solvent.25-28
This method is more direct and is known to provide more
reliable values of the excited-state dipole moment.26,28 In this
particular work, we have focused our attention on two N-aryl
carbazoles (Chart 1). The results show that the measured values
of the excited state dipole moment of these molecules are far
less than those obtained previously from the solvatochromic
method.29
Intramolecular charge transfer in electron donor-acceptor
molecules has been a topic of extensive investigation in recent
years. The current activity on this topic can be ascribed to a
number of reasons ranging from the importance of the charge-
transfer process in photosynthesis, solar energy conversion,
chemical transformation, nonlinear optical properties of materi-
als, to testing of contemporary theories of electron transfer. Of
late, a number of studies have been directed toward a special
type of charge transfer process, commonly termed the twisted
intramolecular charge transfer (TICT) phenomenon.1-16 The
TICT phenomenon was introduced by Grabowski and co-
workers1,2 to account for the anomalous fluorescence band of
4-(N,N-dimethylamino)-benzonitrile (DMABN).10,11 There oc-
curs a transfer of an electron from the dimethylamino to the
cyanophenyl moiety and a decoupling of the donor and the
acceptor orbitals through internal twisting of the dimethylamino
group in the excited state of the molecule. It was proposed that
the normal and anomalous fluorescence bands of DMABN
originate respectively from the locally excited (LE) and TICT
states of the molecule. Subsequent studies have shown that the
photophysical behavior (in particular, the fluorescence proper-
ties) of a wide variety of other donor-acceptor systems could
also be accounted for in terms of the TICT mechanism.3-8 It
should be noted in this context that the TICT process may not
always lead to dual fluorescence as is observed in the case of
DMABN and a number of analogous systems. The TICT state
has even been proposed for donor-acceptor systems which
exhibit a single fluorescence band.12-16 In these systems, the
emission may originate from either an LE state or a TICT state.
The nature of the emitting state (whether LE or TICT) in these
molecules is usually determined based on the estimated dipole
moment of the fluorescent state of the molecule. Because a TICT
state is expected to be much more polar than the LE state, the
two states can easily be distinguished based on the magnitude
of the excited-state dipole moment.
Experimental Section
Materials. The two systems were prepared according to the
published procedure.30 Compound I was prepared as follows
and a similar procedure was adopted for II. An equimolar (3
mmol) mixture of carbazole and NaH was stirred in dry DMF
under nitrogen for 2 h. The sodium carbazole so formed was
then heated at 120 °C with 4-fluorobenzonitrile (3 mmol) and
NaI (6 mmol) in dry DMF for about 20 h. The product and
unreacted starting materials were precipitated by adding water
to the reaction mixture. The precipitate was dried and the product
* To whom correspondence should be addressed. E-mail: assc@
uohyd.ernet.in and fessenden.1@nd.edu, respectively.
† School of Chemistry, University of Hyderabad.
‡ Radiation Laboratory, University of Notre Dame.
10.1021/jp0035481 CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/12/2001

Fluorescent State of N-Arylcarbazole Derivatives
J. Phys. Chem. A, Vol. 105, No. 22, 2001 5439
CHART 1
signal amplitudes of the sample were compared with those from
the reference compound, diphenylcyclopropenone (DPCP), and
necessary corrections were made for adjustments of cavity
coupling. Because the response time of the apparatus (36 ns),
as determined by the cavity Q, was significantly higher than
the fluorescence lifetime of the systems, the cavity Q was
lowered by introducing a resistive plate in the cavity. The
measured Q factor under these conditions gave a time constant
of about 9 ns. The shapes of the decay curves were fitted by
integrating a suitable differential equation taking into account
the pulse shape of the excitation laser (12 ns full-width, half-
maximum Gaussian), the response time of the apparatus from
the Q factor, an additional 1 ns time constant for the subsequent
amplifier, and the excited-state lifetime of the compound. The
equation allowed for a fraction of the singlet to decay into a
long-lived (on this time scale) triplet. This calculation was also
used to model the rising portion of the data curves observed
for DPCP, which gives a step response as a result of the
transformation of this compound into nonpolar photoproducts.26
The response times so measured agreed with the value from
the Q factor. Photolysis of the samples was achieved with 355
nm pulses (approximately 10 mJ/pulse over 1 cm2) from a
Quanta-Ray PRO 230-10 Nd:YAG laser. The laser dose was
monitored while photolyzing the samples to allow correction
for any fluctuation in laser intensity. Considerable care was
given to verify, by experiments at several laser pulse energies,
that there were no effects such as ground-state bleaching.
The fluorescence lifetime of the systems was measured on a
single photon counting spectrofluorimeter equipped with a N2
laser (Photon Technology International, GL 3300) as the
excitation source. Spectrophotometric measurements were made
using a Shimadzu spectrophotometer (UV 3101PC).
was separated by column chromatography using a silica gel
column. A mixture of ethyl acetate and n-hexane (50:50) was
used as eluent. Colorless crystals were obtained from absolute
ethanol upon slow evaporation of the solvent. The structure of
the compound was confirmed by NMR spectroscopy.
The reference compound, diphenylcyclopropenone (Aldrich),
was recrystallized several times from an ethanol-water mixture
before photolysis. Benzene (Fischer Scientific, spectranalyzed)
was used without any further purification. The solutions were
deoxygenated by bubbling with argon prior to the time-resolved
experiments.
Methodology. The present method is based on the principle
that addition of a dipolar solute to a nonpolar solvent introduces
a dielectric loss (ꢀ") that is proportional to the square of the
dipole moment (µ) of the solute and is given by
ꢀ′′ ) A[S]µ2g(ωτ)
where [S] is the molar concentration of solute. A is a constant
that involves the dielectric constant of the media, the absolute
temperature and fundamental constants such as Avogadro’s
number and the Boltzmann constant. The term, g(ωτ), takes
into account the nature and rate of molecular reorientation (ω
is the angular frequency of the microwaves, and τ is the
molecular rotational relaxation time). In an ideal case, g(ωτ) is
given by ωτ/[1 + (ωτ)2] but usually g(ω) is different than this
form in view of the fact that there exists a distribution of
relaxation times even for a rigid molecule. The unknown values
of g(ωτ) were determined by measuring the loss at the desired
frequency for known concentrations of the ground-state mol-
ecules in static experiments as done previously.26,28
Theoretical Calculation. The ground state dipole moments
of the two derivatives were obtained by theoretical calculations
for the AM1 optimized geometries of the molecules. The
calculations were performed on a Personal Computer using the
commercially available Hyperchem package. Unrestricted ge-
ometry optimization at the semiempirical level was performed
following initial optimization of the geometry by MMX
molecular mechanics program. The gradient norms were
monitored to test for successful convergence. Moreover, because
the crystal structures of the two derivatives are available,31 the
dipole moment values were also calculated for the structures
obtained through X-ray crystallography.
The transient experiments use a comparison with the change
in dielectric loss of a reference compound to eliminate all factors
of the experimental setup. The equation employed for the
measurement of the excited-state dipole moment is given by26
Results and Discussion
∆(µ2)s ) (Vsgr/Vrgs)∆(µ2)r
Time dependent variation of the dielectric loss signals
observed upon photolysis of benzene solutions of I, II, and the
reference compound, DPCP, are shown in Figure 1. It should
be noted that the curve for DPCP, which represents a decrease
in dielectric loss, has been inverted for comparison with the
signals for the carbazoles, which show an increase in loss on
excitation. That the signals do not arise from any spurious effect,
such as heating of the solution or the cavity, has been verified
by performing blank experiments on benzene. Further, to be
sure that only the dipole moment change of the molecule gives
rise to the dielectric loss, we have verified that no signal appears
from a benzene solution of anthracene (with a similar absor-
bance), a molecule for which ∆µ is known to be negligible.
The best fit to the decay curves, calculated by taking into
account the factors described in the Experimental section, are
shown as solid curves in Figure 1. The curve for DPCP provides
not only the amplitude but also shows that the response time
matches the data. The experimental curves in Figure 1, parts a
where the subscripts s and r refer to the unknown sample and
reference compounds, respectively, the g terms are g(ωτ) as
defined above, and V represents the amplitude of the transient
signal. The use of the ground state to measure the loss factor g
assumes that the direction of the dipole in the ground and excited
states is the same. This assumption is strictly valid for compound
I because of its symmetry. For compound II, the small dipole
moment makes the direction unclear so the dipole in the excited
state may be in a different direction. The effect of the change
in axis of the molecular reorientation on the g(ωτ) value should
not be large however. Thus, the value of the excited state dipole
moment determined for this compound should not be greatly
in error.
The details of microwave measurements were described in
earlier publications.25,26,28 The microwave signal source was a
40 mW klystron with frequency in the range 9.0-9.2 GHz. The

5440 J. Phys. Chem. A, Vol. 105, No. 22, 2001
Samanta et al.
TABLE 1: Excited State Dipole Moments of
N-arylcarbazoles along with the Values of Parameters Used
in the Calculations
DPCP
I
II
amplitudea
g(ωt)
µg/D
37.07
0.288
5.1b
88.59
0.1774
2.45c
5.6
102.16
10.4
57.75
0.1843
1.57c
5.4
71.50
8.6
τf/ns
∆(µ2)
µe
(5.1)2
0
-5.1
∆µ
7.95
0.075
7.03
0.01
d
φT
a Value from curve fitting corrected for difference in dose, cavity
Q, and absorbances. b From ref 26. c Represents AM1 calculated dipole
moment for the structures obtained from X-ray crystallography. For
fully optimized geometries µg values were 2.86 and 1.64 D for I and
II, respectively. d The triplet yield based on the assumption that the
triplet state has the same dipole moment as the excited singlet.
Figure 1. Time-resolved changes in the dielectric loss of argon-bubbled
benzene solutions of (a) 10 mM of I and (b) 8 mM of II. The inset
shows the behavior for the reference compound, DPCP, at 3 mM
concentration. The latter represents a reduction in dielectric loss so
has been inverted for presentation. The solid curves were calculated
by the method described in the text. The lack of a return to the baseline
(in a and b, particularly in a) is the result of formation of triplet state.
the electron donor moiety is different between the two systems.
According to the TICT mechanism, an electron is transferred
from the amino nitrogen to the cyanophenyl moiety, and the
distance between the center of the positive and the center of
the negative charges is almost the same for the two systems.
Therefore, one should expect the dipole moment of the TICT
state of I to be similar in magnitude to that for DMABN.
However, the measured excited-state dipole moment of I is far
lower than that for the TICT state of DMABN (ca. 17 D).33
Interestingly, the dipole moment of the LE state of DMABN
(ca. 10 D)33 is almost identical to the measured dipole moment
for the excited state of I. In view of this, we conclude that it is
the LE state from which the fluorescence of the systems
originates.
Because the dielectric loss measurements have been made
in the relatively nonpolar solvent, benzene, one might argue
that in polar solvents such as acetonitrile, the emitting state could
be a more polar TICT state. However, the following observations
clearly argue against this idea. First, had there been a change
in the nature of the emitting state while going from the relatively
nonpolar solvent benzene to the more polar solvent acetonitrile,
one would have expected the plot of the emission energy against
the microscopic solvent polarity parameter, ET(30)34 to be
nonlinear with a change of slope in between. However, such
measurements (in benzene, 1,4-dioxane, tetrahydrofuran, and
acetonitrile) have been found to be linear (see Figure 2) with
correlation coefficients of 0.994 for I and 0.999 for II. Quite
obviously, it is the same state that is the origin of fluorescence
in benzene and acetonitrile.
Second, the fluorescence yields of the present systems are
quite high as shown in Table 2 (φf of 0.38-0.52 for I and 0.19-
0.29 for II) despite the fact that the fluorescence from a TICT
state is expected to be forbidden.2,3 In addition, the TICT
fluorescence yield is expected to be quite sensitive to the polarity
of the medium. Bhattacharyya et al.8 and others have demon-
strated that the TICT fluorescence yield of DMABN increases
with an increase in the polarity of the medium until it reaches
a maximum. Subsequent increases in the polarity of the medium
have been found to decrease the TICT fluorescence yield. The
fluorescence yield of the two systems studied here is reported
to be very much the same in nonpolar n-hexane (0.34 for I and
0.12 for II) and polar acetonitrile (0.38 for I and 0.19 for II).29
The near constancy of the fluorescence yield of the present
systems also goes against the TICT hypothesis.
and b, show (on this time scale) an approach to a plateau rather
than to the baseline. The plateau represents the formation of
some longer-lived triplet state. If the dipole moment of the triplet
is equal to that of the excited singlet, then the quantum yields
are 0.075 and 0.01 for I and II, respectively.
The fluorescence lifetimes of the carbazoles were measured
because those values are required for the calculation of the true
amplitudes (those corresponding to the absence of distortion
by a finite pulse width and restricted response time). The values
obtained for I and II are 5.6 and 5.4 ns, respectively. The
calculation also requires the knowledge of the ground state
dipole moments of the systems. These values have been
evaluated by two different procedures. First, the structures of
the molecules have been obtained through X-ray crystallography,
and then the dipole moments corresponding to these structures
have been obtained through AM1 calculations. Second, the
geometries of the molecules were fully optimized theoretically
and the ground-state dipole moments evaluated by AM1 method.
The values so obtained are 2.45 and 1.57 D for I and II by the
first method and 2.86 and 1.64 D by the second method. Even
though the dipole moment values obtained for a given system
by the two methods are very close, the former set using the
experimental geometries has been used in the calculation of the
excited-state dipole moments.
The excited-state dipole moments of I and II, as obtained by
the analysis of the dielectric loss data in benzene (Table 1), are
10.4 ((0.5) and 8.6 ((0.5) D, respectively. The changes in the
dipole moment of the molecules on electronic excitation (∆µ)
lie between 7 and 8.0 D. (The much smaller amplitude of the
curve for II in Figure 1b is mainly because of a smaller
absorbance for that compound.) These directly measured values
of the change in the dipole moment of the molecules are almost
one-half of the previously estimated values based on a range
of polar solvents.29,32 Therefore, a reinterpretation of the nature
of the fluorescent state of the two systems may be required. It
must also be noted that our calculation of the dipole moments
has assumed that only a single excited-state species is present.
The discussion below addresses these points together.
Third, a more reliable indicator of the nature of the emitting
state may be29 the radiative rate constant which is also shown
in Table 2. These values are seen to decrease only moderately
Of the two molecules studied here, I resembles the prototype
TICT system, DMABN, rather closely in the sense that only

Fluorescent State of N-Arylcarbazole Derivatives
J. Phys. Chem. A, Vol. 105, No. 22, 2001 5441
dioxane are significantly lower than previously determined from
the solvatochromic fluorescence data in a range of solvents.
From the present measurement, it is concluded that the present
systems fluoresce from a state that has predominantly an LE
character. On the basis of the available photophysical data of
the systems, it is argued that no TICT state could be seen to be
contributing to the fluorescence even in polar media.
Acknowledgment. Financial assistance for this work has
been obtained from Council of Scientific and Industrial Research
(CSIR), India and the office of Basic Energy Sciences of the
U.S. Department of Energy. This is contribution No NDRL-
4263 from the Notre Dame Radiation Laboratory.
References and Notes
(1) Rotkiewicz, K.; Grellman, K. H.; Grabowski, Z. R. Chem. Phys.
Lett. 1973, 19, 315.
(2) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D.
J.; Baumann, W. NouV. J. Chim 1979, 3, 443.
(3) Rettig, W. Angew. Chem. 1986, 25, 969.
(4) Lippert, E. W. R.; Koutecky, V. B.; Heisel, F.; Miehe, J. A. AdV.
Chem. Phys. 1987, 68, 1.
Figure 2. Plot of the energy of the fluorescence maximum against
the solvent parameter ET(30) for the two compounds studied here (data
in Table 2). The solvents are (left to right) toluene, 1,4-dioxane,
tetrahydrofuran, and acetonitrile. The straight lines are least-squares
fits.
(5) Topics in Current Chemistry; Rettig, W., Ed.; Springer-Verlag:
Berlin, 1994; p 253.
TABLE 2: Fluorescence Decay Properties of Compounds I
and II in Some Solvents
(6) Rettig, W. Fluorescence Spectroscopy, New Methods and Applica-
tions; Wolbeis, O. S., Ed.; Springer-Verlag: Berlin, 1993; p 31.
(7) Progress in Photochemistry and Photophysics; Rettig, W., Bau-
mann, W., Eds.; CRC Press: Boca Raton, 1992; Vol. 6, p 79.
(8) Bhattacharyya, K.; Chowdhury, M. Chem. ReV. 1993, 93, 507.
(9) Zachariasse, K. A.; Groby, M.; von der Haar, T.; Hebecker, A.;
Il’chev, Y. V.; Morawski, O.; Ru¨ckert, I.; Ku¨hnle, W. J. Photochem.
Photobiol. A: Chemistry 1997, 105, 373.
(10) Lippert, E.; Lu¨der, W.; Moll, F.; Nagele, H.; Boos, H.; Prigge, H.;
Blankenstein, I. S. Angew. Chem. 1961, 73, 695.
(11) AdVances in Molecular Spectroscopy; Lippert, E.; Luder, W.; Boss,
H., Eds.; Pergamon: New York, 1962.
(12) Jones, G., II; Jackson, W. R.; Choi, C. Y.; Bergmark, W. R. J.
Phys. Chem. 1985, 89, 294.
(13) Gompbel, J. A. V.; Schuster, G. B. J. Phys. Chem. 1989, 93, 1292.
(14) Rechthaler, K.; Ko¨hler, G. Chem. Phys. 1994, 189, 99.
(15) Soujanya, T.; Fessenden, R. W.; Samanta, A. J. Phys. Chem. 1996,
100, 3507.
Cyanophenyl carbazole (I)
solvent
toluene
1,4-dioxane
tetrahydrofuran
acetonitrile
ET(30) peak energy (cm-1
)
φfa
τfb
kfc
33.9
36.0
37.4
45.6
26 385
25 707
25 063
23 095
0.48 3.60 13.3
0.52 5.29
0.48 6.20
0.38 8.00
9.8
6.8
4.8
Naphthyl carbazole (II)
ET(30) peak energy (cm-1
solvent
)
φfa
τfb
kfc
toluene
33.9
36.0
37.4
45.6
26 385
25 974
25 707
24 390
0.27 4.47 6.0
0.29 5.87 4.9
0.23 5.93 3.9
0.19 7.04 2.7
1,4-dioxane
tetrahydrofuran
acetonitrile
a Fluorescence quantum yield, measured relative to the value in
acetonitrile which has been reported previouly29 and corrected for
refractive index. b Fluorescence lifetime in ns. c Radiative rate constant
(16) Fery-Forgues, S.; Fayet, J. P.; Lopez, A. J. Photochem. Photobiol.
A: Chem. 1993, 70, 229.
(17) Lippert, V. E. Z. Elektrochem. 1957, 61, 962.
(18) Mataga, N.; Kaifu, Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1956,
29, 465.
(s-1), k × 10-7
.
from toluene (presumed to be similar to benzene) to acetonitrile
solvent, and the value is very similar in toluene and dioxane.
The very big decrease of 5-10-fold in the case of 4-(N,N-
dialkylamino)-benzonitriles1,35 is not seen for the present
compounds. This, again, suggests that there is no reversal of
the LE and TICT states while the solvent is changed in this
series. The presence of a single excited state seems evident for
the carbazoles in these solvents.
(19) Mataga, N. Bull. Chem. Soc. Jpn. 1963, 36, 654.
(20) Ravi, M.; Samanta, A.; Radhakrishnan, T. P. J. Phys. Chem. 1994,
98, 9133.
(21) Koutek, B. Collect. Czech. Chem. Commun. 1978, 43, 2368.
(22) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 1st ed.;
Plenum Press: New York, 1983.
(23) Samanta, A.; Fessenden, R. W. J. Phys. Chem. 2000, 104, 8577-
82.
(24) Samanta, A.; Fessenden, R. W. J. Phys. Chem. 2000, 104, 8972-
5.
On the basis of the results of the time-resolved microwave
experiments and the above arguments, we are led to conclude
that the emitting state of the present systems is not as polar as
previously thought. The fluorescence of I and II originates from
a state in which the molecular geometry is not very different
from that in the ground state. The reason for a smaller dipole
moment than that found in the previous work29 is not clear, but
may involve the selection of the radius parameter in the
Lippert-Mataga treatment.
(25) Fessenden, R. W.; Carton, P. M.; Paul, H.; Shimmamori, H. J. Phys.
Chem. 1979, 83, 1676.
(26) Fessenden, R. W.; Carton, P. M.; Shimmamori, H.; Scaiano, J. C.
J. Phys. Chem. 1982, 86, 3803.
(27) De Haas, M. P.; Warman, J. M. Chem. Phys. 1982, 73, 35.
(28) Fessenden, R. W.; Hitachi, A. J. Phys. Chem. 1987, 91, 3456.
(29) Rettig, W.; Zander, M. Chem. Phys. Lett. 1982, 87, 229.
(30) Bridger, R. F.; Law, D. A.; Bowman, D. F.; Middleton, B. S.;
Ingold, K. U. J. Org. Chem. 1968, 33, 4329.
(31) Saha, S.; Samanta, A., to be published.
(32) Schneider, F.; Zander, M. Ber. Bunsen-Ges. Physik. Chem. 1971,
75, 887.
(33) Schuddeboom, W.; Jonker, S. A.; Warman, J. M.; Leinhos, U.;
Kuhnle, W.; Zachariasse, K. A. J. Phys. Chem. 1992, 96, 10 809.
(34) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry,
2nd ed.; VCH: Weinheim, 1988.
Conclusion
Time-resolved microwave dielectric loss measurements of two
N-arylcarbazole derivatives suggests that the dipole moments
of the fluorescent states of the two systems in benzene and 1,4-
(35) Rettig, W. J. Lumin. 1980, 26, 21.