Matrix dependence of light emission from TCNQ adducts

Article abstract of DOI:10.1039/b104992p

The reactions of primary and secondary amines with 7,7,8,8-tetracyano-p-quinodimethane (TCNQ) lead to mono- and di-substituted adducts. Fluorescence emission has been observed for several of these compounds. The luminescence property of the TCNQ adducts, 2-{4-[(2,6-dimethylmorpholin-4-yl)(4-methylpiperidin-1-yl) methylene]cyclohexa-2,5-dien-1-ylidene}malononitrile (MORPIP) and 2-{4-[cyclohex-1-yltetrahydropyrimidin2(1H)-ylidene]cyclohexa-2,5-dien-1- ylidene}malononitrile (AMINO) were investigated in a variety of environments. These included alcohol solutions and crystals at room temperature and glass forming solvents and polymer films as a function of temperature. The fluorescence quantum yields and Stokes' shifts were found to be very sensitive to the matrix. Crystal structure data show that the molecules are non-planar in the ground state. The matrix effect is discussed in terms of the conformational change during photo-excitation and the constraint imposed on this by the matrix.

Full text of DOI:10.1039/b104992p

View Article Online / Journal Homepage / Table of Contents for this issue
Matrix dependence of light emission from TCNQ adducts
David Bloor,*a Yasuyuki Kagawa,a{ Marek Szablewski,a Mosurkal Ravi,a Stuart J. Clark,a
Graham H. Cross,a Lars-Olof Pa˚lsson,a Andrew Beeby,b Chetin Parmerc and
Garry Rumblesc{
aDepartment of Physics, University of Durham, South Road, Durham, UK DH1 3LE
bDepartment of Chemistry, University of Durham, South Road, Durham, UK DH1 3LE
cDepartment of Chemistry, Imperial College London, Exhibition Road, London, UK
SW7 2AY
Received 5th June 2001, Accepted 18th September 2001
First published as an Advance Article on the web 31st October 2001
The reactions of primary and secondary amines with 7,7,8,8-tetracyano-p-quinodimethane§ (TCNQ) lead to
mono- and di-substituted adducts. Fluorescence emission has been observed for several of these compounds.
The luminescence property of the TCNQ adducts, 2-{4-[(2,6-dimethylmorpholin-4-yl)(4-methylpiperidin-1-
yl)methylene]cyclohexa-2,5-dien-1-ylidene}malononitrile (MORPIP) and 2-{4-[cyclohex-1-yltetrahydropyrimidin-
2(1H)-ylidene]cyclohexa-2,5-dien-1-ylidene}malononitrile (AMINO) were investigated in a variety of
environments. These included alcohol solutions and crystals at room temperature and glass forming solvents
and polymer films as a function of temperature. The fluorescence quantum yields and Stokes’ shifts were found
to be very sensitive to the matrix. Crystal structure data show that the molecules are non-planar in the ground
state. The matrix effect is discussed in terms of the conformational change during photo-excitation and the
constraint imposed on this by the matrix.
fluorescence quantum yields of fluorophores are routinely
1. Introduction
performed in solution.19 Fluorescence in solution is affected by
solvent polarity, the solvatochromism observed for polar
molecules can be very large.20 Fluorescence is also affected
by solvent viscosity.21 Alcohols provide a solvent system where
both polarity and viscosity can be varied. The polarity of the
solvent can be changed for normal alcohols by extending the
molecular length.22,23 The viscosity of the solvent can be
controlled by changing the number of hydroxy groups, this
leads to the enhancement of the hydrogen bonding network
in the medium and increased viscosity, e.g. as in diethylene
glycol and glycerol.24,25 A study of fluorescence lifetimes
and quantum yields was undertaken for two of our chromo-
Since the first synthesis1 and study of the basic chemistry2 of
TCNQ in 1962, the chemistry of adducts of TCNQ has been
extensively explored.3–5 Recently this class of compounds has
attracted interest because of their non-linear optical properties.
While much of this work has focussed on single crystals and the
large molecular hyperpolarisibity (b) of these compounds6–9
they have also been shown, both experimentally and theore-
tically, to possess large ground state dipole moments (m).10,11
The product mb, which is large in these materials, is a useful
molecular figure of merit characterising the second order non-
linearity of poled polymer films containing the chromo-
phores.12 Thus, there is continuing interest in these TCNQ
adducts as components of electro-optical polymers. In contrast,
the fluorescence of non-linear optical chromophores has
attracted little attention. Generally it has been regarded as a
problem since fluorescence excited by either harmonics or
two-photon absorption can affect measurements of non-linear
optical coefficients, even when the fluorescence from the
chromophores is weak. However, we have discovered that
certain of our TCNQ adducts show intense, matrix dependent
fluorescence.11,13
phores,
2-{4-[(2,6-dimethylmorpholin-4-yl)(4-methylpiperi-
din-1-yl)methylene]cyclohexa-2,5-dien-1-ylidene}malononitrile
(MORPIP) and 2-{4-[cyclohex-1-yltetrahydropyrimidin-2(1H)-
ylidene]cyclohexa-2,5-dien-1-ylidene}malononitrile (AMINO),
dissolved in a series of normal alcohols, diethylene glycol and
glycerol to evaluate the influence of the viscosity and polarity of
the medium on their luminescence properties.
The fluorescence was first observed visually from a MORPIP
sample placed adjacent to an ultra-violet lamp used for exciting
fluorescence on TLC plates. A bright yellow emission was
observed from crystalline material deposited inside a test tube
above a weakly fluorescing acetonitrile solution. This observa-
tion indicated that the environment of the chromophore had a
large effect on the emission. Studies were therefore undertaken
of crystalline powders and molecules dispersed in solid, amor-
phous matrices. These included glass-forming solvents at low
temperature and polymers. The latter were studied as the
incorporation of fluorophores into polymer matrices has been
used in the fabrication of electro-luminescent devices. The
quantum efficiency of the fluorescence was found to depend on
both matrix and temperature. Details of the study of the
fluorescence of these TCNQ adducts in solution and solid
matrices are presented here and complement our earlier brief
communication.13
Interest in organic light emitting chromophores has expanded
rapidly since the discovery of efficient electro-luminescence
(EL), its use in light emitting devices14–16 and its potential for
electrically pumped solid state lasers.17,18 The measurement of
photo-luminescence has been used extensively to characterise
fluorescent chromophores and identify potential materials for
use in EL devices. Measurements of radiative lifetimes and
{Present address: Corning International, Communications Products
Div., No.35 Kowa Bld. 3rd Flr., 14-14 Akasaka-chome, Minato-chu,
Tokyo 107-0052, Japan.
{Present address: National Renewable Energy Laboratory, MS3216,
1617 Cole Blvd, Golden, CO 80401, USA.
§The IUPAC name for TCNQ is (cyclohexa-2,5-diene-1,4-diylidene)-
dimalononitrile.
DOI: 10.1039/b104992p
J. Mater. Chem., 2001, 11, 3053–3062
This journal is # The Royal Society of Chemistry 2001
3053

View Article Online
2.2. Physical characterisation
2. Experimental
Room temperature emission and photo-excitation spectra were
recorded with a Perkin–Elmer Luminescence Spectrometer,
LS-50B, and absorption spectra were recorded with a Perkin–
Elmer Lambda 19 spectrophotometer. MORPIP solutions
were prepared in methanol, ethanol, propan-1-ol, butan-1-ol,
pentan-1-ol, hexan-1-ol, ethylene glycol, and glycerol without
degassing. After filtration (0.5 mm filters) solutions were diluted
to give an optical density 0.055¡0.01 at 375 nm. A reference
solution of quinine sulfate dihydrate in 0.5 M sulfuric acid with
the same optical density was also prepared. The quantum yield
of quinine sulfate dihydrate is 0.55 and is relatively temperature
independent.27 The quantum yields of the MORPIP solutions
were calculated by comparing the integrated emission intensity
of the quinine sulfate dihydrate solution to that of samples. All
of the emission spectra were measured with the same spectral
resolution.
Thin films of poly(methylmethacrylate) (PMMA) and bis-
phenol A poly(carbonate) (PC) were prepared as described in
the following example. MORPIP (11 mg) and PMMA pellets
(1.218 g) were dissolved in 3 ml of tetramethylurea (TMU). The
mixture was stirred for several days to ensure complete
dissolution. Glass substrates were cleaned with acetone and a
mixture of acetone and isopropyl alcohol. For films prepared
by spin coating the PMMA solution was dropped on to the
substrate and the spin coater was switched on. A spin rate of
2000 rpm for ten to fifteen seconds gave optically translucent
PMMA films with thickness in the range 3–5 mm. Dip coated
films were prepared by the slow withdrawal of the clean sub-
strate from the solution giving films with thickness w10 mm.
The films were kept in an oven at 80 uC for three to five days
under vacuum to ensure complete removal of solvent prior to
spectral measurements.
TCNQ of 98% purity was obtained from Lancaster Ltd. 2,6-
Dimethylmorpholine, 4-methylpiperidine and N-(3-aminopro-
pyl)cyclohexylamine of 97% purity were obtained from Aldrich
Ltd. All solvents used were HPLC grade. These chemicals were
used without further purification.
2.1 Synthesis
2.1.1 2-{4-[(2,6-Dimethylmorpholin-4-yl)(4-methylpiperidin-
1-yl)methylene]cyclohexa-2,5-dien-1-ylidene}malononitrile
(MORPIP). 2-{4-[(2,6-Dimethylmorpholin-4-yl)(4-methylpiperi-
din-1-yl)methylene]cyclohexa-2,5-dien-1-ylidene}malononitrile
(MORPIP) was prepared in a two-stage reaction. 2,6-Dimethyl-
morpholine (0.527 ml, 4.8 mmol) was added to a solution of
TCNQ (1 g, 4.8 mmol) in 100 ml tetrahydrofuran (THF)
heated at 50 uC. The mixture was stirred at 50 uC for 3 hours,
cooled to room temperature and then stirred overnight. The
solvent was removed under vacuum. The residue was re-
crystallised in acetonitrile twice and dried under vacuum. 0.52 g
of 2-{4-[1-(2,6-dimethylmorpholin-4-yl)-2-nitriloethylidene]-
cyclohexa-2,5-dien-1-ylidene}malononitrile, yield 36%, was
obtained. 4-Methylpiperidine (0.2 ml, 2 mmol) was added to
a solution of 2-{4-[1-(2,6-dimethylmorpholin-4-yl)-2-nitrilo-
ethylidene]cyclohexa-2,5-dien-1-ylidene}malononitrile (0.4 g,
1.36 mmol) in THF (30 ml) heated at 50 uC and stirred for
30 min at 50 uC. The product was observed to precipitate. The
solution was cooled to room temperature and stirred for one
hour. The yellow precipitate was collected by filtration and
dried under vacuum. Re-crystallisation of the solid was
carried out with acetonitrile. Yellow crystals of MORPIP
were obtained (0.24 g, 46%). Analytical data: lmax 417 nm in
acetonitrile. Microanalysis, calculated for C22H28N4O: % C
72.5, H 7.74, N 15.37, found: % C 72.13, H 7.70, N 15.18. Mass
spectrum (EI): 364(Mz) (100%, molecular ion). Decomposi-
tion temperature 260 uC. The molecular structure was con-
firmed by X-ray crystallography,5 detailed accounts of X-ray
crystallographic data for TCNQ adducts will be presented
elsewhere.26
The photoluminescence quantum yield (PLQY) of polymer
thin films prepared as described above, and polycrystalline
samples deposited on to Spectrosil substrates was determined
at room temperature using an integrating sphere (Labsphere)
to collect the light emitted in all directions.28 The excitation was
provided by the 442 nm line of a CW HeCd laser (Kimmon)
and the excitation intensity was about 0.5 mW on an area of
2 mm2. Absorbance of the samples at the excitation wavelength
was in the range 0.3 to 0.5.
2.1.2 2-{4-[Cyclohex-1-yltetrahydropyrimidin-2(1H)-ylid-
ene]cyclohexa-2,5-dien-1-ylidene}malononitrile (AMINO).
2-{4-[Cyclohex-1-yltetrahydropyrimidin-2(1H)-ylidene]cyclohexa-
2,5-dien-1-ylidene}malononitrile (AMINO) was prepared by
adding N-(3-aminopropyl)cyclohexylamine (0.5 ml, 2.8 mmol)
to TCNQ (0.5 g, 2.4 mmol) in acetonitrile (30 ml) heated to
50 uC. The mixture was stirred at 50 uC for three hours and
cooled to room temperature. The product precipitated and
was collected by filtration. The precipitate was recrystallised
from acetone and dried under vacuum. A brown powder
(0.23 g, yield 31.1%) was obtained. Analytical data: 1H NMR:
(d-DMSO), d 7.1 ppm, doublet, benzene protons (2H); d
6.8 ppm, doublet, benzene protons (2H); d 3.6 ppm, quintet,
amine proton (1H); d 3.45 ppm, triplet; -(CH2)- next to NH
(2H); d 3.35 ppm, triplet, -(CH2)- next to N (2H); d 1.95 ppm,
triplet; cyclohexane proton next to N (1H); d 1.7 and 1.0 ppm,
cyclohexane and -(CH2)- protons (12 H). IR: 2179, 2132 cm21
(nitrile stretching). lmax: 368 nm in acetonitrile. Microanalysis:
Calcd for C19H22N4: % C 74.48, H 7.24, N 18.28. Found: % C
74.46, H 7.35, N 18.28. Mass spectrum (EI): 306(Mz) (100%,
molecular ion). Decomposition temperature: 220 uC.
Low temperature emission and photo-excitation spectra
were recorded with an ISA Fluoromax fluorimeter. Low tem-
perature absorption spectra were recorded with a Perkin–
Elmer Lambda 15 spectrophotometer. MORPIP solutions
were prepared in glass forming solvents, either propan-1-ol,
2-methyltetrahydrofuran (2MTHF) or EPA, a 2 : 2 : 5 mixture
of ether, ethanol and isopentane. Typically the optical density
of these solutions at the absorption maximum was below 0.5.
The samples were placed in a special low temperature cuvette in
an Oxford Instruments cryostat (DN1704). Absorption,
emission and photo-excitation spectra were obtained in the
range from room temperature to 80 K as the sample was varied
with an Oxford Instrument temperature controller (ITC-6).
Data were obtained for polymer thin film samples mounted in
the same cryostat.
Fluorescence lifetimes were measured using the time-
correlated single photon counting technique.29 Samples were
excited with a cavity dumped DCM dye laser (Coherent 7210
cavity dumper and 700 Series dye laser) that was synchronously
pumped with the second harmonic of a mode locked Nd : YAG
laser (Coherent Antares 76-s). The resulting 3.8 MHz pulse
train could be tuned over the range 610–680nm. Excitation
pulses in the range 300–340 nm of intensity approximately 10 pJ
were obtained from this pulse train by frequency doubling in a
BBO crystal. The detection system comprised a 0.22 m
subtractive dispersion double monochromator (Spex 1680), a
microchannel plate (Hamamatsu R3809U), a 1GHz amplifier
The NMR spectrum was recorded in d-DMSO, a polar
solvent. Therefore, the aromatic protons sense different electro-
nic environments due to the charge separation between the
donor and acceptors groups. These protons are observed as
doublets at d 7.1 and 6.8 ppm. This splitting is observed in
other TCNQ adducts.10
3054
J. Mater. Chem., 2001, 11, 3053–3062

View Article Online
and timing discriminator (EG&G Ortec 9327), a time-to-
amplitude converter (Tennelec TC864) and a multichannel
pulse to height analyser (Tennelec PCA II). The inherent
response time of the system was ca. 80 ps. Excitation and
detection wavelengths were selected from the steady state data.
Solutions in glass forming solvents and polymer films were
prepared and mounted in the cryostat as described above.
Fluorescence lifetimes were recorded at room temperature,
200 K, 125 K and 80 K.
refractive index (n) of the medium the absorption and emission
energies, relative to the gas phase values, are given by
ꢂꢀ
ꢁ
ꢀ
ꢁꢃ
2ꢀg(ꢀe{ꢀg)
hca3
e{1
n2{1
2n2z1
vambasx~va0bs
{
{
2ez1
ꢁ
(1)
ꢀ
2(ꢀ2e{ꢀ2g)
n2{1
2n2z1
{
{Cða,DÞ
hca3
ꢂꢀ
n2{1
2n2z1
ꢁ
ꢀ
ꢁꢃ
2ꢀe(ꢀe{ꢀg)
hca3
e{1
n2{1
2n2z1
vemmax~ve0m
{
{
2ez1
ꢁ
3. Results
(2)
ꢀ
2(ꢀ2e {ꢀ2g)
Fluorescence emission has been observed from many of the
TCNQ adducts we have synthesised. The emission is more
intense for the asymmetrically substituted adducts than for the
symmetrically substituted compounds we have studied. We
have, therefore, concentrated our studies on two asymmetri-
cally substituted adducts one with two distinct substituents
(MORPIP) and one with an asymmetrically substituted fused
ring (AMINO). The structure of these compounds is, in
general, intermediate between the two limiting forms, the
neutral state with quinoid structure and the fully charge
separated zwitterionic state with benzenoid structure,5,10,26,30
which are shown in Fig. 1. The actual structure will be
determined by the reaction field acting on the molecule,10,31
which will depend on the environment, e.g. on solvent polarity.
X-Ray crystallographic data show that the electron donor
moieties (amino groups) and the p-conjugation unit (the
benzene ring) are not coplanar.6,9,26,30 The twist angle between
the plane of the conjugation unit and the plane containing the
two amino nitrogen atoms and the carbon atom they are
bonded to is approximately 45u. In addition the observed and
calculated ground state dipole moments are large.6,10,11,13,32
Thus, these molecules are expected to show sizeable solvato-
chromism and be affected by the constraints placed on changes
in molecular conformation by rigid matrices. Experimental
data are presented below for the different media used as hosts
for the chromophores.
{
{Cða,DÞ
hca3
where me and mg are the excited and ground state dipole
moments, a is the radius of the spherical Onsager cavity
surrounding the solute molecule and C(a,D) describes the
effects of molecular polarisibility and dispersion, which are
usually treated as small and constant for polar molecules. The
Stokes’ shift is then given by the difference
ems
max
Dv~vambasx{v
ꢀ
ꢁ
2(ꢀe{ꢀg)2 e{1
n2{1
(3)
~
{
zconstant
hca3
2ez1 2n2z1
where (((e–1)/(2ez1))–((n2–1)/(2n2z1)) is the polarity or
solvent density parameter (Df). Different approaches lead to
similar expressions for Dn but with different polarity para-
meters.34,40 All are modified if the effect of solvent molecule
reorientation is included.34,41 These expressions have all been
used to describe the behaviour of chromophores in polar
solvents with some degree of success. This reflects the inaccu-
racies in the models and similarity of the polarity parameters.40
We chose to use the Lippert formalism to provide a consistent
analysis of the experimental data. Deviations from the Lippert
equation can result from specific solvent interactions,19,42 and
relaxation of the approximations involved in its derivation.43,44
The analysis of the solvent induced shift in absorption and
emission maxima is usually based on eqns. (1) and (2) with a
number of simplifying assumptions.35,36,38,40 For example if the
difference in the nature of the ground and excited states is small
eqns. (1) and (2) can be written as:
The combined effect of the solvatochromic shift of the
ground and excited electronic energy levels gives rise to a
solvent dependent Stokes’ shift. There is an extensive literature
on the modelling of solvent–solute interactions.33 However, a
simple approach leading to the Lippert equation, is often
adequate.19,34–39 In terms of the relative permittivity (e) and
ꢂꢀ
ꢁ
ꢀ
ꢁꢃ
2ꢀg(ꢀe{ꢀg)
hca3
e{1
1
2
n2{1
2n2z1
vambasx~va0bs
{
{
(4)
2ez1
and
ꢂꢀ
ꢁ
ꢀ
ꢁꢃ
2ꢀe(ꢀe{ꢀg)
hca3
e{1
2ez1
1
2
n2{1
2n2z1
vemmax~ve0m
{
{
(5)
The case of me&mg has also been considered.35 However, as
noted above the choice of polarity parameter is not crucial.40
3.1 Absorption and emission of alcohol solutions
The intensity of the room temperature emission from alcoholic
solutions of MORPIP and AMINO show a strong solvent
dependence. This is illustrated in Fig. 2, which shows the raw
data for the emission of MORPIP solutions. Both emission and
absorption spectra for MORPIP and AMINO are broad and
featureless, scaled emission spectra and examples of absorption
spectra are shown in Fig. 3. Comparison of absorption and
fluorescence excitation spectra shows that they are identical
within experimental error. The variation in the energy of the
absorption maximum for MORPIP and AMINO dissolved in
normal alcohols, methanol to hexanol, are plotted using
eqn. (4) in Fig. 4. Similar results are obtained for a wider range
of solvents, i.e. THF, dichloromethane, acetone, dimethylfor-
mamide (DMF) and acetonitrile. This trend is not maintained
Fig. 1 Molecular structures of (a) MORPIP, shown in the extreme
quinoid and benzenoid forms, and (b) AMINO, shown in the quinoid
form.
J. Mater. Chem., 2001, 11, 3053–3062
3055

View Article Online
Fig. 2 Unscaled emission spectra of MORPIP in alcoholic solvents.
From least to most intense spectra the solvents are methanol, ethanol,
propanol, butanol, pentanol, hexanol, diethylene glycol and glycerol.
for chloroform, a somewhat less polar solvent. The hypso-
chromic shift for the more polar solvents is a result of the
stabilisation of the benzenoid ‘‘zwitterionic’’ structure so that
mevmg. In low polarity solvents the molecules will be close to
the intermediate, equal bond length structure.
Although all the emission profiles are similar the Stokes’ shift
and quantum yields for MORPIP and AMINO vary con-
siderably, as shown in Table 1. The use of degassed solvent
gave a slightly higher value for the quantum yield, however, the
difference was within the error quoted in Table 1. The
dependence of the Stokes’ shifts on Df for the normal alcohol
solutions is plotted in Fig. 4 together with the absorption data.
Fig. 4 The energy of the absorption maximum (triangles) and Stokes’
shift (diamonds) plotted according to eqns. (3) and (4) for (a) MORPIP
and (b) AMINO dissolved in normal alcohols (methanol to hexanol).
For MORPIP the slopes of the best-fit lines to both sets of data
are the same within the fitting error, while for AMINO the
slopes are similar. These results indicate that for these solutions
the change in dipole moment between the excited and the
ground states must be approximately constant and, from
eqns. (4) and (5), that the excited state dipole must be small.
The emission spectra in Figs. 2 and 3 display little dependence
on solvent bearing out this conclusion. This is shown
quantitatively in Fig. 5, where the emission data are plotted
according to eqn. (5). This shows a negligible dependence of the
position of the emission spectra maxima on solvent polarity for
MORPIP and a weak dependence for AMINO.
Evidence for specific solvent interactions is shown in
Fig. 6(a), where the Stokes’ shifts for MORPIP and AMINO
in diethylene glycol and glycerol do not fit the trend observed
for the normal alcohols. Although diethylene glycol (Df~
0.266) and glycerol (Df~0.265) have almost the same polarity
as butanol (Df~0.264) they have larger Stokes’ shifts. These
solvents have two and three-dimensional hydrogen bonding
networks and high viscosity, 30.2 nP for diethylene glycol and
934 nP for glycerol. Thus, because of the large difference in
viscosity the solvent interaction cannot be completely modelled
by Df. However, ET(30) solvent polarity scale does provide a
good fit for the Stokes’ shifts for all the solutions considered
here, Fig. 6(b). This is because the ET(30) scale is not model
based but is an experimentally determined microscopic polarity
parameter, which takes into account the solvent–solute
interaction.20
Fig. 3 Absorption (to the right) and emission spectra (to the left) for (a)
MORPIP and (b) AMINO in solution in methanol (squares), propanol
(circles), hexanol (triangles), pentanol (star), diethylene glycol (inverted
triangles), and glycerol (diamonds). The fluorescence spectra have been
multiplied in (a) by factors of 15, 7, 4, 4 and 1 and in (b) by factors of
20, 10, 6, 5 and 1 for methanol, propanol, hexanol, diethylene glycol
and glycerol solutions respectively. The emission was excited at the
peak of the photo-excitation spectrum at 375 nm.
Similarly we observe that the fluorescence quantum yields in
normal alcohols are linearly dependent on Df, but the values for
diethylene glycol and glycerol do not follow this trend, as
3056
J. Mater. Chem., 2001, 11, 3053–3062

View Article Online
Table 1 Stokes’ shifts and quantum yields for alcohol solutions of MORPIP and AMINO at room temperature
Solvent
MORPIP
AMINO
Df
Viscosity/nPc
Stokes’ shift/cm21a
Quantum yield (%)b
Stokes’ shift/cm21a
Quantum yield (%)b
Methanol
Ethanol
5160
4620
4240
4150
3970
4010
5040
5400
0.1
0.1
0.4
0.5
0.7
1.0
1.4
11.8
c
7380
6770
6500
6160
6010
5670
7990
6630
0.22
—
0.53
—
1.3
—
2.32
21.8
0.309
0.290
0.275
0.264
0.254
0.243
0.266
0.265
0.544
1.07
1.95
2.54
3.62
4.58
30.2
934
Propan-1-ol
Butan-1-ol
Pentan-1-ol
Hexan-1-ol
Diethylene glycol
Glycerol
aAccuracy¡100 cm21
.
bAccuracy ca. ¡10% of value. The values of viscosity were taken from CRC solvent handbook.
shown in Fig. 7(a). The fluorescence quantum yields of the
glycerol solutions are more than ten times larger than that of
the normal alcohol solutions. In this instance the use of the
ET(30) scale did not give a better fit to the data, Fig. 7(b). This
suggests an additional effect for solutions in viscous solvents.
Previous studies indicate that this is due to the constraint
placed on changes in molecular conformation during excitation
by the hydrogen-bonding networks.24,25
3.2 Glass forming solvents
Solvents that form low temperature glasses have been
extensively employed as matrices for fluorophores.45,46 If the
increase in fluorescence quantum yield in viscous solvents is
due to the constraining effect of the hydrogen bonding network
a similar effect can be anticipated in glassy matrices. Suitable
glass forming solvents are propan-1-ol,24,47 2MTHF48 and
EPA.49 The use of propan-1-ol glasses allows a direct com-
parison to be made with the room temperature solution data.
Spectra were recorded for MORPIP in propan-1-ol, 2MTHF
Fig. 5 Observed shifts in emission maxima, with excitation at the peak
of the photo-excitation spectrum at 375 nm, for alcohol solutions of
MORPIP (squares) and AMINO (circles) plotted according to eqn. (5).
Fig. 7 Fluorescence quantum yield for alcohol solutions of MORPIP
(squares) and AMINO (circles) plotted versus (a) Df and (b) ET(30).
The trend lines are shown for the normal alcohols; MORPIP, full line:
AMINO, dashed line.
Fig. 6 Stokes’ shift for alcohol solutions of MORPIP (squares) and
AMINO (circles) plotted versus (a) Df and (b) ET(30).
J. Mater. Chem., 2001, 11, 3053–3062
3057

View Article Online
Table 2 Observed changes in spectra of MORPIP and AMINO in glass forming solvents at room and low temperature
Absorption
maximuma/cm21
Emission
maximumb/cm21
Absorption shiftc/
cm21 294 to 80 K
Emission shiftd/
Stokes’ shifte/
cm21
Adduct
Solvent/K
cm21 294 to 80 K
MORPIP
Propanol (294)
(80)
2MTHF (294)
(80)
EPA (294)
(80)
EPA (294)
(80)
23600
24900
21400
22700
22500
25100
27030
28520
c
19200
22000
18300
19200
19530
22570
24070
24420
1300
1300
2600
1200
e
3000
900
4400
2900
3100
3500
3000
2600
2960
3830
2990
350
AMINO
aAccuracy¡50 cm21
.
bAccuracy¡100 cm21. Accuracy¡100 cm21
.
dAccuracy¡100 cm21. Accuracy¡70 cm21
.
and EPA glasses and for AMINO in an EPA glass. A summary
of data obtained for MORPIP and AMINO at 294 K and 80 K
in these solvents is given in Table 2.
increase in emission intensity was observed in the glassy phase,
which is formed at ca. 200 K, cf. Fig. 8(b). The integrated
emission intensity rises sharply below 250K becoming approxi-
mately constant below 200 K, Fig. 10. The quantum yield is
estimated to be 13% at 80 K compared with room temperature
value of 0.4%. Non-radiative decay is reduced at low
temperature as thermally activated processes are turned off.44
The increase of emission intensities is influenced both by this
and by changes in the interaction of the MORPIP with the host
medium.
In contrast the data for MORPIP in 2MTHF, Table 2, show
that the maximum of the absorption and emission spectra shifts
by smaller, similar, amounts between 294 K and 80 K.
Consequently, there is a small increase in Stokes’ shift between
294 K and 80 K. If this shift is attributed solely to the solva-
tochromic shift, which results from the density change of the
medium, then on the basis of the trend line in Fig. 4(b) Df must
increase by ca. 0.02 between 294 K and 80 K. While data are
available for the relative permittivity (e) over the range 296 K
to 208 K comparable data are not available for the refractive
index (n); however, on the basis of data for similar solvents Df
should increase by ca. 0.03 over this temperature range. Thus,
the observed change in Stokes’ shift is somewhat smaller than
expected if it is due solely to solvatochromism. Emission
spectra of MORPIP in 2MTHF at various temperatures are
shown in Fig. 9. The increase in emission intensity (i.e.
quantum yield) at low temperatures is similar to that for the
propan-1-ol glass, cf. Fig. 10.
The absorption and emission spectra of MORPIP in propan-
1-ol as a function of temperature are shown in Fig. 8. The
absorption spectra became sharper at low temperature but
with no indication of any vibronic structure, the small feature
just below 26000 cm21 is an instrumental artefact. Thus,
inhomogeneous broadening50 dominates even at 80 K. There is
an increase in optical density, which is commensurate with the
observed narrowing of the absorption band and increase in
sample density. Both the absorption and fluorescence maxima
shift to higher energy as the solvent density, e, n and the pola-
rity parameter (Df) increase at low temperatures as expected for
mevmg, cf. eqns. (4) and (5). However, the emission maximum
shifts by a larger amount than the absorption maximum on
cooling, cf. Fig. 8. This is the opposite of what is expected for
mevmg. The nett effect is that the Stokes’ shift decreases from
ca. 4400 cm21 at room temperature to ca. 2900 cm21 at 80 K
rather than increasing as indicated by eqn. (3). Thus, the
constraint placed on molecular conformation by the glassy
matrix has a significant effect on the emission process. Precise
determinations were not made of the fluorescence quantum
efficiencies below room temperature. However, a dramatic
For MORPIP in EPA the shifts in the maximum in absorp-
tion and emission between 294 and 80 K are similar and there is
a small decrease in the Stokes’ shift, cf. Table 2. Over the same
temperature range there is a distinct difference between the
shifts with temperature for the maximum in absorption and
emission of AMINO in EPA although the values are the
smallest measured, cf. Table 2. This results in a significant
increase in Stokes’ shift for AMINO in comparison to a
decrease for MORPIP.
Fig. 8 (a) Absorption and (b) emission spectra of MORPIP in propan-
1-ol at 294 (triangles), 200 (squares), 150 (circles) and 80 K (diamonds).
Fig. 9 Emission spectra of MORPIP in 2-methyl THF at 294
(triangles), 200 (squares), 150 (circles) and 80 K (diamonds).
3058
J. Mater. Chem., 2001, 11, 3053–3062

View Article Online
yields were determined for thin layers of MORPIP and
AMINO micro-crystals spread on a non-fluorescent silica sub-
strate in the integrating sphere apparatus. Values of 5% and
1.6% were obtained for the photoluminescence quantum yields
of MORPIP and AMINO, respectively. These values are appro-
ximate since there was considerable variation between samples,
which we attribute to variations in the size of microcrystallites
from sample to sample.
3.4 Polymer matrices
Films of PMMA and PC containing MORPIP and AMINO
were prepared by spin and dip coating. PMMA films contain-
ing MORPIP were obtained from solutions in DMF, tetra-
methylurea (TMU) and dichloromethane and films containing
AMINO were obtained from DMF and TMU solutions. PC
films containing MORPIP were obtained from DCM solutions.
These films were used to determine fluorescence quantum yields
at room temperature, in the integrating sphere apparatus,
and study the temperature dependence of the absorption and
emission spectra from room temperature down to 80 K. Absorp-
tion and emission spectra for films containing MORPIP
prepared from DMF and TMU solutions at room temperature,
200 and 80 K are shown in Fig. 11. Room temperature and low
temperature data are listed in Tables 3 and 4, respectively.
Quantitative data were not obtained for the AMINO samples
but qualitatively the emission observed was comparable to that
from the MORPIP in PC sample.
Fig. 10 Integrated emission intensity of MORPIP in propan-1-ol
(squares) and 2-methyl THF (circles) as a function of temperature.
At 80 K the intensities of emission for MORPIP in the
PMMA films prepared from TMU and DMF solutions was
about 2 and 3 larger than that at 294 K, cf. Fig. 11. This is in
contrast with the much larger increase, ca. thirty to one
hundred times, of the emission intensity between room and low
temperature in glass forming solvents. PMMA is below its glass
transition temperature at room temperature and the increase in
density on cooling will be ca. 1–2%. This is much smaller than
that of the glass forming solvents, e.g. EPA contracts by ca.
25% between room and liquid nitrogen temperatures.
3.5 Fluorescence decay times
The fluorescence lifetime of MORPIP in propan-1-ol was
determined at 294 and 80 K by the time resolved single photon
counting technique. The sample was excited at 340 nm and the
Table 4 Low temperature absorption and emission data for polymer
films containing MORPIP obtained from DMF and TMU solutions
Absorption
maximuma/
cm21
Emission
maximuma/
cm21
Stokes’
shiftb/
cm21
Fig. 11 Spectra of the MORPIP doped PMMA films (a) absorption
(right) and emission (left) for films cast from DMF solution at 294
(triangles), 200 (circles) and 80K (diamonds), (b) excitation (right) and
emission (left) for films cast from TMU solution at 294 (triangles), 200
(circles) and 80 K (diamonds).
Casting
solvent
Temperature/
K
DMF
TMU
294
200
80
294
200
80
22400
22400
22700
22400
22400
22600
18200
18500
18700
18900
1900
4200
3900
4000
3500
3400
3500
3.3 Crystalline powders
Fluorescence from a TCNQ adduct was first observed visually
for microcrystalline powder of MORPIP formed above an
acetonitrile solution in a test tube. The fluorescence quantum
19200
aAccuracy¡100 cm21 bAccuracy¡200 cm21
. .
Table 3 Room temperature absorption and emission data for polymer films containing MORPIP and AMINO obtained from different solvents
Absorption
maximuma/cm21
Emission
maximuma/cm21
Stokes’
shiftb/cm21
Quantum yield
(%)c
Adduct
Polymer
PMMA
Casting solvent
MORPIP
DMF
TMU
DCM
DCM
DMF
TMU
c
22620
22570
22420
21930
26950
26980
19100
19250
18760
18330
21210
21340
3520
3320
3660
3600
5740
5640
21
21
19
PC
PMMA
7¡2
d
AMINO
d
aAccuracy¡50 cm21
.
bAccuracy¡100 cm21. Accurate to¡5% of value unless noted. dSee text.
J. Mater. Chem., 2001, 11, 3053–3062
3059

View Article Online
that mg~16¡3 D and me~2¡1 D or using the approximation
me%mg the values are again unchanged within the experimental
error, mg~14¡3 D and me~1.5¡1 D.
The alternative fitting procedure due to Koutek52 has been
shown by Ravi et al.53 and Kumar et al.54 to give reasonable
results for me for molecules with modest mg, i.e. 3–7 D. However,
the agreement with the limited experimental data is not good in
this case. Fitting data for MORPIP and AMINO leads to
values for both molecules for mg and me of ca. 5 and 1 D,
respectively. The former is much smaller than the value of 15 D
determined experimentally for MORPIP and the latter does not
reflect the difference in solvent dependence of the fluorescence
between AMINO and MORPIP, Fig. 5.
Fig. 12 Fluorescence decay of the emission of MORPIP dissolved in
propan-1-ol at (a) 294 and (b) 80 K.
The fluorescence quantum yields of MORPIP and AMINO
dissolved in normal alcohols fall in the range 0.1 to 1%
(Table 1). The presence of oxygen has little effect, as the excited
state lifetime is too short to allow for diffusion to permit
interaction with the dissolved oxygen. The quantum yields are
larger, up to 20%, in the viscous solvents diethylene glycol and
glycerol. The values for AMINO are about twice those found
for MORPIP in the same solvent. When the Stokes’ shifts are
plotted as a function of polarity parameter (Df) the results for
the viscous solvents depart from the trend for the other
alcohols, Fig. 6(a). Although this discrepancy is removed when
the parameter ET(30) is used rather than Df, Fig. 6(b), the
quantum yield data for the viscous solvents remain anomalous,
Fig. 7, showing a strong effect of solvent viscosity on fluores-
cence intensity. The fluorescence properties of chromophores,
which undergo large conformational changes on photo-
excitation, are strongly influenced by solvent viscosity. Tetra-
phenylethylene (TPE) has been extensively studied.25,55,56 The
lifetime of TPE increases from 6 ps in low viscosity solvents to
60 ps in ethylene glycol and 600 ps in glycerol.25 This has been
attributed to the viscous media hindering the rotation of the
phenylene moieties thereby limiting non-radiative decay of
the excited state, which is induced by the torsion motion of the
phenylenes. A decrease in non-radiative decay will lead to an
increase in fluorescence quantum yield as observed for
MORPIP and AMINO.
There is other experimental evidence that there are signi-
ficant conformational changes between ground and excited
state in the TCNQ adducts. As noted earlier X-ray crystal-
lographic data show that the electron donor moieties (amino
groups) and the p-conjugation unit (the benzene ring) in the
TCNQ adducts are not coplanar.6,9,26,30 In MORPIP and
related compounds the twist angle between the molecular
components in the molecular ground state is approximately
45u. Despite this, both theory and experiment show that the
ground state dipole moments of these molecules are
large.6,10,11,13,32 As noted above for MORPIP mg is found to
be approximately constant at 15 D in chloroform, dichoro-
methane and acetone solutions. This is in accord with the result
that for alcohol solutions the change in dipole moment on
excitation is constant and me$0, i.e. mg is approximately
constant. This behaviour is somewhat different from that of
adducts made with tertiary amines, which are rigid planar
molecules. The dipole moments of these molecules increase as
the polarity of the solvent is increased.10,57 Thus, we deduce
that the molecular geometry of MORPIP does not remain fixed
as the molecular environment, e.g. solvent polarity, is changed.
The small value of me is, therefore, likely to be a consequence of
a change in the torsion angle between the component parts of
the molecule in the excited state.
emission was detected at 520 nm at 294K and 450 nm at 80 K
close to the emission maximum at these temperatures, cf.
Table 2. The fluorescence decays observed are shown in
Fig. 12. The decay time at room temperature was too short
to measure the temporal profile of the observed emission being
essentially identical with that of the excitation pulse. Thus the
decay time must be less than ca. 0.1 of the width of the
excitation pulse, i.e. v10 ps. The fluorescence decay from a
sample at 80 K was fitted with a single exponential decay with a
lifetime of 2.5¡0.2 nanoseconds, cf. the linearity of the decay
over two decades shown in Fig. 12. The quality of fit was good
with a raw reduced chi-squared of 1.1 and random weighted
residuals.29
4. Discussion
In principle eqns. (3) to (5) can be used to estimate the mg and me
for MORPIP and AMINO. However, the values obtained
depend crucially on the factor a3, where a is the radius of the
Onsager cavity enclosing the molecule in solution. The model
of a spherical cavity is widely applied although it is clear that
for elongated molecules a more complex model with an
ellipsoidal cavity should be used.10,51 While this approach is
strictly correct the additional complication is not justified given
the approximations involved in the analysis. Furthermore it
has been shown that the equivalent radius of the spherical
cavity deduced from crystal structure data is in reasonable
agreement with that found experimentally for other zwitter-
ionic molecules.10 The crystal structure of MORPIP gives a
3 11,13
˚
volume per molecule of ca. 530 A .
If this is taken as the
˚
volume of the Onsager cavity then a#4 A, which represents a
lower bound for the solution cavity.10 The slopes of the plots of
absorption and emission maxima and the Stokes shift for
MORPIP, Figs. 4 and 5, are 19700¡3300 cm21 (R~0.948),
160¡1700 cm21 (R~20.049) and 18100¡2800 cm21 (R~
0.956), respectively. Using these data the values found from
eqns. (3) to (5) are mg~12¡5 D and me~20.1¡1 D. Modifying
eqns. (4) and (5) to allow for me%mg, i.e. replacing mg (me2mg) by
2
mg in eqn. (4) and me (me2mg) by me mg in eqn. (5), has little
effect on these values giving mg~11¡5 D and me~20.1¡
1 D. The value obtained for mg from the spectral data is similar
to that of 15 D found both theoretically and experimentally, in
chloroform, dichloromethane and acetone, for MORPIP.11,13
This agreement is reasonable given the approximations
involved in the theory and the choice of a, the ellipsoidal
shape of the molecules and the large value of mg.
Crystalstructuredataarenotavailablefor AMINO. However,
although AMINO has only one bulky substituent the
molecular volume will be similar to that of MORPIP so that
MOPAC96-AMI calculations give a larger value for me for
a planar geometry (y10 D) than for a 90u twist (~6 D).7
However, these values are significantly larger than we find
experimentally. Recently similar calculations have been per-
formed for a series of adducts with twist angles between y40u
to y75u and give values for me in the range 9 to 13 D.32 These
values are all much larger than our results for me. However,
˚
a#4 A is a reasonable approximation. The slopes of the plots
of absorption and emission maxima and the Stokes shift for
AMINO, Figs. 4 and 5, are 32400¡2700 cm21 (R~0.984),
3500¡800 cm21 (R~0.913) and 25100¡1200 cm21 (R~
0.995), respectively. In this case from eqns. (3) to (5) we find
3060
J. Mater. Chem., 2001, 11, 3053–3062

View Article Online
theoretical modelling using ab initio methods shows that the
dipole moment increases as the twist angle increases and that
the excited state has a more planar geometry. Hence, we believe
that in low viscosity solvents these molecules undergo signi-
ficant changes in conformation on photo-excitation. Restric-
tion of the change in molecular conformation by viscous
solvents leads to enhanced fluorescence; i.e. the relaxed excited
state has a low fluorescence quantum yield, while that of the
hindered excited state is much higher.
We note that the steric hindrance for AMINO, due to the
bulky cyclohexyl group attached to the cyclic substituent will
be much greater than for MORPIP, see Fig. 1. Molecular
modelling shows that a twist angle of 45u between the mole-
cular sub-units is possible for AMINO, with some slight
deformation of the molecular framework. However, a large
reduction in twist angle on excitation is not possible. This
suggests a smaller change in molecular conformation for
AMINO than for MORPIP both as a function of solvent
polarity and on excitation. It also offers an explanation for the
non-zero value of me for AMINO and the higher fluorescence
intensity for AMINO in all the solvents used. As the viscosity
of diethylene glycol and glycerol is derived from a hydrogen-
bonding network they will interact strongly with AMINO and
produce significant effects on the fluorescence despite the lower
flexibility of this molecule.
than the instrumental resolution, i.e. ¡10 ps. However, at 80 K
in the glassy matrix the decay is well characterised by a single
exponential with a time constant of 2.5 ns, Fig. 12(b). Some
increase in fluorescence and lifetime is to be expected at low
temperatures due to the reduction in thermally induced non-
radiative decay. However, the large changes observed are
indicative of a significant change in the character of the
emission process, from a predominantly non-radiative process
in solution at room temperature to a predominantly radiative
process in the glass at low temperature.
Similar data have been obtained for other adducts with
different substituents so we believe that the behaviour reported
here for MORPIP and AMINO is representative of this class of
compounds. The molecules have large mg and a molecular
conformation in which the plane of the substituents is twisted
with respect to the conjugated ring and CN moieties. When the
molecules are free to relax on excitation me is small and the
molecular geometry of the excited state is less twisted than that
of the ground state and may in some instances be close to
planar, e.g. MORPIP. At room temperature the decay from
this state is predominantly non-radiative. In situations where
the molecular relaxation is inhibited by constraints imposed by
the environment, decay by a radiative process is favoured. This
behaviour may be viewed as an inverse of that where the
ground state is planar and the excited state is twisted with
greater charge transfer in the excited state, i.e. TICT (twisted
intramolecular charge transfer).34–37,39
This hypothesis is also capable of explaining the trends
observed with solid matrices. The molecular packing in a
crystal lattice will leave limited free volume for molecular
deformation on photo-excitation. Hence, larger fluorescence
quantum efficiency is to be expected for crystals than for
solutions with a larger effect for the more flexible MORPIP, as
observed.
Acknowledgements
This work was supported by grants from the Engineering and
Science Research Council (EPSRC). We thank Dr I. D. W.
Samuel and M. Halim for making preliminary measurements
of the fluorescence of the solid samples at room temperature
and N.-A. Hackman for assistance in preparation of polymer
film samples.
The same considerations apply to polymer matrices and glass
forming solvents. The differences in spectra and fluorescence
quantum yield in the polymer matrices reflect both the polarity
and density of the media. PC has a higher polarity than PMMA
and a distinct spectral shift is observed. The differences in
quantum efficiency reflect differences in the free volume avai-
lable in the different polymers. Although fluorescence intensity
for MORPIP in glass forming solvents was not determined as
precisely as that of the room temperature solutions and films a
dramatic increase from an initially small value was observed on
cooling to 80 K. This reflects the change from a fluid to a solid
environment as the glass forms. Initially the adduct is free to
relax with minimal constraint giving a low fluorescence
quantum efficiency. This increases as the solvent viscosity
increases constraining the changes in molecular conformation
on excitation and becomes constant as the glass solidifies into a
phase with low thermal expansion, cf. Fig. 10. The increase in
fluorescence for polymer films over the same temperature range
was much smaller. This is to be expected as the polymer matrix
is in a glassy phase throughout this temperature range and the
reduction in free volume on cooling is small. It is surprising that
the fluorescence quantum yields for the polymer films are
approximately four times the values observed for crystals. A
possible explanation is that the environment in a crystal is
identical for all molecules while in the polymer film there is a
distribution of environments due to variation in the size of the
cavities, which provide the free volume, at the molecular level.
Thus the higher fluorescence intensity observed in the polymer
films will predominantly originate from molecules in the most
constrained environments.
References
1
D. S. Acker, R. J. Harder, W. R. Hertler, W. Mahler, L. R. Melby,
R. E. Benson and W. E. Mochel, J. Am. Chem. Soc., 1960, 82,
6408.
2
D. S. Acker and W. R. Hertler, J. Am. Chem. Soc., 1962, 84, 3370
and following papers (Substituted Quinodimethanes Parts I to V).
B. P. Bespalov and V. V. Titov, Russ. Chem. Rev., 1975, 44, 1091.
G. J. Ashwell, E. J. C. Dawnay, A. P. Kuczyn˜ski, M. Szablewski,
I. M. Sandy, M. R. Bryce, A. M. Grainger and M. Hasan, J. Chem.
Soc., Faraday Trans., 1990, 86, 1117.
3
4
5
6
7
N. Martin, J. L. Segura and C. Seoane, J. Mater. Chem., 1997, 7,
1661.
M. Ravi, S. Cohen, I. Agranay and T. P. Radhakrishnan, Struct.
Chem., 1996, 7, 225.
M. Ravi, D. N. Rao, S. Cohen, I. Agranat and T. P.
Radhakirishnan, Chem. Mater., 1997, 9, 830.
8
9
M. Ravi, Proc. Indian Acad. Sci. (Chem. Sci.), 1998, 110, 133.
M. Ravi, P. Gangopadhyay, D. N. Rao, S. Cohen, I. Agranat and
T. P. Radhakrishnan, Chem. Mater., 1998, 10, 2371.
10 M. Szablewski, P. R. Thomas, A. Thornton, D. Bloor, G. H. Cross,
J. M. Cole, J. A. K. Howard, M. Malagoli, F. Meyers, J. Bre´das,
W. Wenseleers and E. Goovaerts, J. Am. Chem. Soc., 1997, 119,
3144.
11 Y. Kagawa, Ph.D. Thesis, University of Durham, 1998.
12 S. R. Marder and J. W. Perry, Science, 1994, 263, 1706.
13 Y. Kagawa, M. Szablewski, M. Ravi, N.-A. Hackman,
G. H. Cross, D. Bloor, A. S. Batsanov and J. A. K. Howard,
Nonlinear Opt., 1999, 22, 235.
14 J. H. Burroughes, D. C. Bradley, A. R. Brown, R. N. Marks,
K. M. Makay, R. H. Friend, P. L. Burns and A. B. Holmes,
Nature, 1990, 347, 539.
15 D. Braun and A. J. Heeger, Appl. Phys. Lett., 1991, 58, 1982.
16 H. Tanaka, S. Tokito, Y. Taga and A. Okada, J. Mater. Chem.,
1998, 8, 1999.
The fluorescence lifetime measurements provide further
evidence that rapid non-radiative decay occurs when the
adducts are dissolved in low viscosity solvents while viscous
and solid matrices constrain molecular relaxation and open up
a radiative decay channel. The decay of the fluorescence of
MORPIP in propanol at room temperature, shown in Fig. 12(a),
is identical with the laser excitation pulse. Thus, the excited
molecules decay on a time scale either comparable to or shorter
17 F. Hide, M. A. Diaz-Garcia, B. J. Schwartz, M. R. Andersson,
Q. Pei and A. J. Heeger, Science, 1996, 273, 1833.
J. Mater. Chem., 2001, 11, 3053–3062
3061

View Article Online
18 N. Tessler, P. K. H. Ho, V. Cleave, D. J. Pinner, R. H. Friend,
G. Yahioglu, P. L. Barny, J. Gray, M. de Souza and G. Rumbles,
Thin Solid Films, 1999, 363, 64.
19 J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd Edn.,
Kluwer Academic/Plenum Publishers, New York, 1999.
20 C. Reichart, Chem. Rev., 1994, 94, 2319.
21 A. D. Osborne, J. Chem. Soc., Faraday Trans. 2, 1980, 76, 1638.
22 K. G. Casey and E. L. Quitevis, J. Phys. Chem., 1988, 92, 6590.
23 K. M. Keery and G. R. Fleming, Chem. Phys. Lett., 1982, 93, 322.
24 C. J. Tredwell and A. D. Osborne, J. Chem. Soc., Faraday Trans.
2, 1980, 76, 1627.
38 J. Herbich and A. Kapturkiewicz, J. Am. Chem. Soc., 1998, 120,
1014.
39 K. Araki, K. Tada, M. Abe and T. Mutai, J. Chem. Soc., Perkin
Trans. 2, 1998, 1391.
40 A. T. Amos and B. L. Burrows, Adv. Quant. Chem., 1973, 7, 289.
41 C. Streck and R. Reichert, Ber. Bunsen-Ges. Phys. Chem., 1994, 98,
619.
ˇ
42 B. Cunderl´ıkova´ and L. Sikurova´, Chem. Phys., 2001, 263, 415.
43 P. Suppan, J. Photochem. Photobiol. A. Chem., 1990, 50, 293.
44 M. Fall, J.-J. Aaron, M. M. Dieng and C. Pa´rka´nyi, Polymer,
2000, 41, 4047.
25 P. F. Barbara, S. D. Rand and P. M. Rentzepis, J. Am. Chem. Soc.,
1981, 103, 2156.
45 P. F. Barbara, P. M. Rentzpis and L. E. Brus, J. Am. Chem. Soc.,
1980, 102, 2786.
26 Y. Kagawa, M. Szablewski, M. Ravi, N.-A. Hackman,
G. H. Cross, D. Bloor, A. S. Batsanov and J. A. K. Howard, in
preparation.
27 W. H. Melhuish, J. Phys. Chem., 1961, 65, 229.
28 J. C. de Mello, H. F. Wittmann and R. H. Friend, Adv. Mater.,
1997, 9, 230.
46 H. T. Oh, Y. Kanematsu, A. Kurita and T. Kushida, J.
Luminescence, 1996, 66/67, 310.
47 E. Laitinen, K. Salonen and T. Harju, J. Chem. Phys., 1996, 105,
9771.
48 R. Reichart, F. Stickel, R. S. Fee and M. Maroncelli, Chem. Phys.
Lett., 1994, 229, 302.
29 D. J. R. Birch and R. E. Imhof, in Topics in Fluorescence
Spectroscopy, ed. J. R. Lakowicz, Volume 1, Techniques, Plenum
Press, New York, 1991, pp. 1–95.
30 J. C. Cole, J. M. Cole, G. H. Cross, M. Farsari, J. A. K. Howard
and M. Szablewski, Acta Crystallogr., 1997, B53, 812.
31 J. Onsager, J. Am. Chem. Soc., 1936, 58, 1486.
32 P. Gangopadhay, M. Ravi and T. P. Radhakrishnan, Indian
J. Chem., 2000, 39A, 106.
33 D. Grasso and E. Bellio, Chem. Phys. Lett., 1975, 30, 421.
34 H. Lami and N. Glasser, J. Chem. Phys., 1986, 84, 597.
35 J-F. Le´tard, R. Lapouyade and W. Rettig, J. Am. Chem. Soc.,
1993, 115, 2441.
49 S. L. Murov, I. Charmichael and G. L. Hug, The Handbook of
Photochemistry, Marcel Dekker, London, 1993.
50 N. A. Nemkovich, A. N. Rubinov and V. I. Tomin, Topics in
Fluorescence Spectroscopy, ed. J. R. Lakowicz, Volume 2.
Principles, Plenum Press, New York, 1991, pp. 367–428.
51 C. J. F. Bo¨ttcher, Theory of Electric Polarization, Elsevier,
Amsterdam, 1973, Vol. 1; 1978, Vol. 2.
52 B. Koutek, Collect. Czech. Chem. Commun., 1978, 43, 2368.
53 M. Ravi, A. Samanta and T. P. Radhakrishnan, J. Chem. Phys.,
1994, 98, 9133.
54 S. Kumar, V. C. Rao and R. S. Rastogi, Spectrochim. Acta, Part
A, 2001, 57, 41.
36 G. Liu, L. Heisler, L. Li and M. G. Steinmetz, J. Am. Chem. Soc.,
1996, 118, 11412.
55 J. Kordas and M. A. El-Bayoumi, J. Am. Chem. Soc., 1974, 96,
3043.
37 C. F. Zhao, R. Gvishi, U. Narang, G. Ruland and P. N. Prasad,
J. Phys. Chem., 1996, 100, 4526.
56 D. Ben-Amotz and C. B. Harris, Chem. Phys. Lett., 1985, 119, 305.
57 P. R. Thomas, Ph.D. Thesis, University of Durham, 1997.
3062
J. Mater. Chem., 2001, 11, 3053–3062