Fluorescent carbazole dendrimers for the detection of nitroaliphatic taggants and accelerants

Article abstract of DOI:10.1039/c2jm32072j

The detection of explosive taggant 2,3-dimethyl-2,3-dinitrobutane (DMNB) and accelerant nitromethane (NM) by oxidative quenching of dendrimer fluorescence is examined. Two fluorescent dendrimers incorporating a 3,6-disubstituted-9-n-hexylcarbazole-based core and first-generation biphenyl-based dendrons linked directly or with acetylene bridges are reported. The dendrimers display good photoluminescence (PL) quantum yields in both solution and thin films and appropriate excited state energies for oxidation by the nitroaliphatic analytes. The dendrimer natural excited state lifetimes in solution of 6.8 and 7.4 ns were found to be significantly longer than previously reported fluorescent conjugated polymers and dendrimers for sensing applications. Steady-state PL quenching measurements in solution revealed the highest quenching efficiencies for the detection of nitroaliphatics reported to date of 59 ± 1 M^-1 for DMNB and 78 ± 1 M^-1 for NM. Furthermore, PL lifetime quenching measurements confirmed that the dendrimers were quenched by a predominantly collisional quenching mechanism. As such, the unprecedented quenching efficiencies with nitroaliphatics in solution are due to the combination of the long excited state lifetimes of the dendrimers and efficient collisional quenching. The fluorescence of dendrimer thin films was also reversibly quenched by exposure to pulses of sub-saturation concentrations of analyte vapours. However, in the thin film case the sensitivity towards DMNB was found to be greater than NM, highlighting the disparity between solution and thin film fluorescence quenching measurements.

Full text of DOI:10.1039/c2jm32072j

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 12507
www.rsc.org/materials
PAPER
Fluorescent carbazole dendrimers for the detection of nitroaliphatic taggants
and accelerants
*
Andrew J. Clulow, Paul L. Burn, Paul Meredith and Paul E. Shaw
*
Received 3rd April 2012, Accepted 16th May 2012
DOI: 10.1039/c2jm32072j
The detection of explosive taggant 2,3-dimethyl-2,3-dinitrobutane (DMNB) and accelerant
nitromethane (NM) by oxidative quenching of dendrimer fluorescence is examined. Two fluorescent
dendrimers incorporating a 3,6-disubstituted-9-n-hexylcarbazole-based core and first-generation
biphenyl-based dendrons linked directly or with acetylene bridges are reported. The dendrimers display
good photoluminescence (PL) quantum yields in both solution and thin films and appropriate excited
state energies for oxidation by the nitroaliphatic analytes. The dendrimer natural excited state lifetimes
in solution of 6.8 and 7.4 ns were found to be significantly longer than previously reported fluorescent
conjugated polymers and dendrimers for sensing applications. Steady-state PL quenching
measurements in solution revealed the highest quenching efficiencies for the detection of nitroaliphatics
reported to date of 59 ꢀ 1 Mꢁ1 for DMNB and 78 ꢀ 1 Mꢁ1 for NM. Furthermore, PL lifetime
quenching measurements confirmed that the dendrimers were quenched by a predominantly collisional
quenching mechanism. As such, the unprecedented quenching efficiencies with nitroaliphatics in
solution are due to the combination of the long excited state lifetimes of the dendrimers and efficient
collisional quenching. The fluorescence of dendrimer thin films was also reversibly quenched by
exposure to pulses of sub-saturation concentrations of analyte vapours. However, in the thin film case
the sensitivity towards DMNB was found to be greater than NM, highlighting the disparity between
solution and thin film fluorescence quenching measurements.
explosives since it is safer to ship and store when compared with
nascent high-explosives.15
Introduction
The prevailing global security situation means that the detection
of explosives is of paramount importance in domestic environ-
ments and in current and past conflict zones. Much attention
has been given to the detection of nitroaromatic explosive
residues such as 2,4,6-trinitrotoluene (TNT) and 2,4-
dinitrotoluene (DNT) due to their common usage in explosive
compositions and landmines.1–6 Plastic explosives based on
organic nitramines and nitrate esters have also received
considerable attention.7,8 In contrast, there have been relatively
few reports on the efficient fluorescence quenching-based
sensing of nitroaliphatics such as 2,3-dimethyl-2,3-dinitrobutane
(DMNB) and nitromethane (NM).9–13 DMNB is a highly
volatile tagging agent in commercially manufactured plastic
explosives and hence is an important target analyte for detec-
tion.14 NM is highly volatile and while it is has low detonation
sensitivity in its neat form, when mixed with a variety of
chemicals such as acids, bases, fuels and inorganic oxidants, it
yields explosive compositions. NM is therefore used in
Detection of explosive analytes by oxidative fluorescence
quenching is effective for materials that have a high electron
affinity such as nitroaromatics. As a generic technique, fluores-
cent quenching has the potential to deliver rapid and sensitive
detection with simple, low cost instrumentation suitable for
portable or in-field use. In the detection process, the fluorescence
sensing material is optically excited. If the analyte of interest is in
the environment under scrutiny and has a suitable electron
affinity the sensor is quenched by oxidation of the exciton thus
providing a signal change. While in principle DMNB and NM
can be detected by this method, in practise the response from the
fluorescent sensing element is often very weak,11,16 a fact that has
severely hindered the development of suitable sensing fluo-
rophores. Herein we exploit the primarily collisional nature of
fluorescence quenching by DMNB and NM in solution by
creating sensing materials with long natural radiative lifetimes
and appropriate energy levels. These molecules enable the
detection of DMNB and NM in solution at levels comparable to
those of higher electron affinity analytes17,18 and with the best
sensitivity for nitroaliphatics reported to date.6,13 The sensing
capabilities of thin dendrimer films are then evaluated and
comparisons drawn with the solution based quenching assay.
Centre for Organic Photonics & Electronics, The University of Queensland,
Brisbane, Queensland, 4072, Australia. E-mail: p.burn2@uq.edu.au;
meredith@physics.uq.edu.au
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12507–12516 | 12507

View Online
was itself formed by a Sonogashira reaction from the corre-
sponding bromide focussed dendron 6.19 Gel permeation chro-
Results and discussion
Dendrimer synthesis and physical properties
matography showed that both dendrimers were mono-disperse
20
and their hydrodynamic radii calculated from the Mvs were
ꢀ
The structures and syntheses of the two dendrimers developed
are shown in Scheme 1. The dendrimers are comprised of first
generation biphenyl dendrons with 2-ethylhexyloxy surface
groups and a carbazole-based core. Recent reports have sug-
gested that materials containing carbazole units can display high
affinities for DMNB in solution,6,13 however the origin of this
affinity has not been explored. The dendrimers of this work were
therefore designed to determine the influence of the carbazole
unit on the sensitivity towards nitroaliphatic analytes and
whether the affinity for the analytes in solution could be trans-
lated to the solid state. Dendrimer 4 has the dendrons directly
attached to the carbazole unit while dendrimer 5 has the den-
drons linked to the carbazole unit via acetylene moieties. As
illustrated in Scheme 1 a convergent synthetic strategy was
chosen in which functionalised dendrons 2 and 3 were prepared
before attachment to the dibrominated carbazole 1. The key step
in the synthesis of each dendrimer was a metal-catalysed cross
coupling. Dendrimer 4, in which the dendrons are directly
attached to the carbazole unit, was formed by a standard Suzuki
reaction with the boronate ester focussed dendron 219 in a 52%
yield. In the case of dendrimer 5 a biphasic Sonogashira
coupling, in which the aqueous base deprotects acetylene 3 prior
to coupling, gave the required material in a 47% yield. Dendron 3
ꢁ
ꢁ
8.4 A and 9.1 A for 4 and 5, respectively. Differential scanning
calorimetry gave glass transition temperatures of 37 ꢂC for 4 and
ꢂ
25 C for 5.
Photophysical and electrochemical properties
For dendrimers 4 and 5 to be able to detect DMNB and NM by
fluorescence quenching there are two minimum requirements:
first, they must have detectable (preferably strong) fluorescence
under excitation with UV or visible radiation; and second, their
energetic structure must enable photoinduced electron transfer
from the excited fluorophore to the ground state analyte (i.e.,
oxidation of the dendrimer singlet exciton). The second
requirement demands a negative energy difference between
the excited state ionisation potential of the fluorophore and the
ground state analyte electron affinity. This difference (and the
general energetic landscape) was estimated via electrochemistry
and optical spectroscopy by taking the energy difference between
the calculated or measured reduction potentials of the analyte
and the fluorescent dendrimers. The optical absorption and
photoluminescence (PL) spectra of dendrimers 4 and 5 dissolved
in dichloromethane are shown in Fig. 1. The onset of absorption
Scheme 1 Synthesis of dendrimers 4 and 5. Conditions and reagents: (i) Pd(PPh3)4, PhMe, tert-BuOH, aq. K2CO3, 100 ꢂC, N2, 52%. (ii) Pd(PPh3)4, CuI,
PhMe, aq. NaOH, n-Bu4NBr, 80 ꢂC, N2, 47%. (iii) Pd(PPh3)4, CuI, 2-methylbut-3-yn-2-ol, piperidine, 80 ꢂC, N2, 91%.
12508 | J. Mater. Chem., 2012, 22, 12507–12516
This journal is ª The Royal Society of Chemistry 2012

View Online
Fig. 1 Absorption and PL spectra of 4 and 5 in dichloromethane solu-
tion. The PL spectrum of unsubstituted 9-n-hexylcarbazole (HxCbz) is
included for comparison.
Fig. 2 Thin film absorption and PL spectra of dendrimers 4 and 5.
aggregation. The PLQYs of the dendrimer thin films were found
to be similar to their solution PLQYs being 0.27 ꢀ 0.06 and
0.17 ꢀ 0.02 for 4 and 5 respectively. The decrease in the PLQY of
5 in the solid state is consistent with the proposed fluorophore
aggregation increasing non-radiative decay rates in the thin films.
The increase in PLQY of biphenyl-substituted carbazoles such as
4 in thin films has been observed previously and is attributed to
the restriction of intramolecular bond rotations eliminating non-
radiative decay pathways in the solid state.22
for both dendrimers is around 375 nm although dendrimer 5 has
much stronger absorption at the longer wavelengths due to the
linking acetylene moieties. Both dendrimers have identical PL
spectra with emission peaks at 375 nm and 395 nm and shoulders
at 412 nm and 438 nm. The emission is red-shifted compared to
the bare 9-n-hexylcarbazole core (Fig. 1) that displays emission
peaks at 350 nm and 366 nm with shoulders at 384 nm and 405
nm. This observation is consistent with an extension of the
conjugation of the emissive chromophore in 4 and 5 relative to
the unsubstituted core. The photoluminescence excitation (PLE)
spectra correlate well with the absorption of each dendrimer at
detection wavelengths between 335 nm and 475 nm, suggesting
that emission is from a single excited state of comparable energy
for both 4 and 5. No evidence of emission from the dendrons was
observed showing that energy is transferred from the dendrons to
the emissive chromophore. Solution photoluminescence
quantum yields (PLQYs) were recorded in dichloromethane by
the relative method using quinine sulfate as the standard21 and
were found to be 0.19 ꢀ 0.01 and 0.23 ꢀ 0.01 for 4 and 5
respectively. The time-resolved PL decays at 400 nm in toluene
solutions were measured to determine the natural excited state
lifetime for each of the dendrimers. For dendrimer 4 the natural
radiative lifetime was 7.4 ns and that of 5 was of similar
magnitude at 6.8 ns. The combination of long fluorophore life-
time with lower PLQY implies a slow radiative decay rate for the
active chromophore allowing a longer sampling time for
quenching interactions.
We used Cyclic (CV) and Differential Pulse (DPV) Voltam-
metry in combination with the optical energy gap to estimate the
reduction potentials for 4 and 5. CV showed that both the den-
drimers underwent chemically reversible oxidations and for 5
DPV was used to determine the E1/2. The E1/2s for 4 and 5 were
0.71 V and 0.75 V respectively versus the ferrocenium–ferrocene
(Fc+/Fc) couple. The E1/2s for the reductions lay outside the
electrochemical window and so we estimated the reduction
potentials by adding the energy of the optical gap. To calculate the
optical gap we used a standard method: first, the absorption and
PL spectra were rescaled by dividing the absorbance by photon
energy (in eV) and PL by the photon energy cubed;23 second, the
spectra were normalised to the lowest energy absorption feature
and the highest energy emission feature; and finally, the crossover
point of the rescaled absorption and emission plots was taken as
the optical energy gap. Using this method the optical gap was to
determined to be 3.4 eV for both dendrimers giving estimated
reduction potentials of approximately ꢁ2.7 V for both 4 and 5
versus the Fc+/Fc couple. The first reduction potential of DMNB
has been reported as ꢁ2.2 V versus the Fc+–Fc couple.9 The first
reduction potential of NM has not been measured directly by
cyclic voltammetry as it decomposes on electrode surfaces in
The absorption and PL spectra of thin films of the dendrimers
are shown in Fig. 2. The absorption spectra are similar to their
solution counterparts with only slight red shifts in the peaks and
shoulders observed. The thin film PL spectra of 4 and 5 are also
comparable to that in solution except in the latter case there was
a tail in the PL at the red end of the spectrum consistent with
fluorophore aggregation in the solid state. This suggests that the
first-generation biphenyl dendrons are sufficient to prevent
significant aggregation of the carbazole chromophores when
bound directly to the core. However, separation of the dendrons
from the carbazole by addition of the bridging acetylene moieties
results in a less sterically encumbered structure and greater
a
four-electron process liberating methylhydroxyamine.24
However, based on the fact that DMNB and NM are simple
nitroaliphatic molecules a similar first reduction potential would
be expected. The difference of 0.5 V between the reduction
potentials of the dendrimers and the analytes is greater than the
reported exciton binding energies in organic semiconductors,
which are often around 0.4 eV.25 Therefore 4 and 5 both fulfil the
two fundamental criteria for detection of nitroaliphatic analytes
by oxidative fluorescence quenching.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12507–12516 | 12509

View Online
Solution quenching measurements
The affinity (quenching efficiency) of an analyte for a particular
fluorophore is generally quantified by steady-state Stern–
Volmer measurements in solution. In such an experiment the
quenching efficiency (loss of fluorophore emission) as a function
of the analyte concentration is determined under steady-state
illumination. Changes in absorption and fluorescence were
monitored as known concentrations of analytes were added to
a
toluene solution of constant dendrimer concentration.
Attenuation of the excitation beam and inner-filter effects
caused by analyte absorption were corrected by our previously
reported method.26 The corrected data was then fitted to the
Stern–Volmer equation for low quencher concentrations,17
given by
ꢀ
ꢁ
F0
F
kq
¼ 1 þ Ks
þ
½Qꢃ ¼ 1 þ fKs þ Kcg½Qꢃ
kr þ knr
¼ 1 þ Ksv½Qꢃ
(1)
where F0 is the integrated PL intensity in the absence of the
analyte, F is the integrated PL intensity after analyte addition,
Ks is the quenching efficiency due to the static binding of
quencher molecules, Kc is the quenching efficiency due to
collisional interactions with quencher molecules, kq is the
bimolecular quenching rate constant, kr is the radiative decay
rate constant, knr is the non-radiative decay rate constant, [Q] is
the quencher concentration and Ksv is the Stern–Volmer
quenching constant for the analyte. The Stern–Volmer
quenching constant thus determined is a combination of static
binding efficiency (Ks) and the branching ratio of the rate of
bimolecular quenching to all other intramolecular decay
processes occurring (Kc). The steady-state Stern–Volmer plots
obtained for the dendrimers with the analytes NM and DMNB
are shown in Fig. 3. All runs were repeated in triplicate with
separate analyte solutions and the data from each run combined
to give the overall Stern–Volmer plots. Linear plots were
obtained for both analytes with intercepts close to the theoret-
ical value of 1.0.
Fig. 3 Steady-state Stern–Volmer plots for 4 and 5 with NM and
DMNB in toluene solution.
ꢀ
ꢁ
ꢀ
ꢁ
s0
s
kr þ knr þ kq½Qꢃ
kr þ knr
kq
¼
¼ 1 þ
½Qꢃ
kr þ knr
¼ 1 þ s0kq½Qꢃ ¼ 1 þ Kc½Qꢃ
(2)
where s0 is the radiative lifetime in the absence of quencher, s is
the radiative lifetime after analyte addition and Kc is the colli-
sional portion of the Stern–Volmer constant as seen in eqn (1).
The change in fluorescence lifetime observed upon the addition
of analyte is due to collisional quenching by the analyte and
hence the Kc provides a measure of the quenching caused by
collisional interactions alone. The difference between the exper-
imentally determined Ksv and Kc is due to static interactions (Ks)
that quench the fluorescence by reducing the effective concen-
tration of fluorophore but do not change the observed fluores-
cence lifetime.
The Stern–Volmer results are summarised in Table 1 and we
believe that the quenching constants observed with the nitro-
aliphatic reagents NM and DMNB are the largest reported to
date. The Stern–Volmer analyses gave Ksv values of 78 ꢀ 1 Mꢁ1
and 59 ꢀ 1 Mꢁ1 for 4 and 67 ꢀ 1 Mꢁ1 and 55 ꢀ 1 Mꢁ1 for 5 with
NM and DMNB respectively. The overall magnitude of
quenching was observed to follow the trend NM > DMNB for
both dendrimers and it was found on average that the fluores-
cence of 4 was quenched more effectively than 5 by both of the
analytes.
The steady-state Stern–Volmer measurement does not give
any information as to whether the observed quenching is due to
a static process (the analyte forms a ground state ‘dark
complex’ with the fluorophore prior to photoexcitation) or
a collisional process (the analyte interacts briefly with the flu-
orophore in its photoexcited state). Therefore to further probe
the underlying mechanism, time-resolved fluorescence quench-
ing measurements were performed. By equating the observed
fluorescence lifetime with the reciprocal of the sum of all decay
processes occurring, the lifetime Stern–Volmer equation can be
written as23
Table 1 Summary of solution and thin film PL quantum yields
(PLQYs), PL lifetime, steady-state Stern–Volmer constant (Ksv) and
collisional quenching constant (Kc) data
Dendrimer/ Solution Thin film Solution
PLQY
analyte
PLQY
lifetime (ns) Ksv (Mꢁ1
)
Kc (Mꢁ1
)
4/NM
0.19
0.27
7.4
6.8
78 ꢀ 1
59 ꢀ 1
67 ꢀ 1
55 ꢀ 1
75 ꢀ 1
59 ꢀ 1
62 ꢀ 1
49 ꢀ 1
4/DMNB
5/NM
5/DMNB
0.23
0.17
12510 | J. Mater. Chem., 2012, 22, 12507–12516
This journal is ª The Royal Society of Chemistry 2012

View Online
The time-resolved Stern–Volmer measurements were per-
formed using similar analyte concentrations as the steady-state
measurements, and changes in the fluorescence lifetime were
monitored at 400 nm with each analyte addition. In these
experiments corrections for analyte absorption were not required
as attenuation of the excitation beam only affects the PL inten-
sity and not the PL decay rate. The normalised lifetime data were
fitted with single exponential decays convoluted with the recor-
ded instrument response function (IRF). The lifetime quenching
data collected for 4 and 5 is displayed in Fig. 4 and this shows
a clear decrease in the fluorescence lifetime upon addition of the
analytes. The corresponding lifetime Stern–Volmer plots are
displayed in Fig. 5.
The measured collisional Stern–Volmer constants, Kc, were
essentially the same as the steady-state Stern–Volmer constants
(Table 1) revealing that the fluorescence of the dendrimers is
predominantly quenched in solution by a collisional mechanism.
High levels of collisional quenching by nitroaliphatics have been
observed previously and is attributed to their inability to engage
in p–p interactions with the conjugated dendrimers.17,18 What is
remarkable however, is the magnitude of the Ksv and Kc
constants with 4 and 5, which are much larger than those
previously published. By way of example, an earlier reported
dendrimer possessing the same first-generation biphenyl
Fig. 5 Lifetime Stern–Volmer plots for dendrimers 4 and 5 with NM
and DMNB in toluene solution.
dendrons and a bifluorene core18 had a steady-state Ksv of only
5 ꢀ 1 Mꢁ1 for DMNB in spite of it having a higher PLQY and the
interactions also being primarily collisional. Comparison of the
performance of the dendrimers of this work with conjugated
polymer systems is more difficult as the collisional Stern–Volmer
constant determined by photoluminescence lifetime analysis
possesses a contribution from delayed static quenching by
diffusion of the exciton formed along the polymer chain.18 There
is only a single report of the same lifetime Stern–Volmer analysis
undertaken with DMNB and a conjugated polymer, poly(9,9-di-
n-octylfluoren-2,7-diyl).18 In that work 30% of interactions of the
polymer and analyte in solution were static, and the Ksv was also
18
an order of magnitude lower (5 ꢀ 1 Mꢁ1
)
than the carbazole
containing dendrimers of this current work. Indeed, in many
cases static binding of analytes has been observed to be the
dominant quenching mechanism for conjugated polymers.27,28
We also determined the bimolecular collisional quenching rate
constants (kq) for each analyte using the observed natural life-
times and collisional Stern–Volmer constants (eqn (2)). The
bimolecular quenching rate constants were all found to be on the
order of 109–1010 Mꢁ1 sꢁ1. These values are in the range typical of
diffusion-controlled rate constants29 in solution indicating that
quenching of the fluorophore by the analytes is highly efficient
with most collisions resulting in quenching. Based on these
results it would be reasonable to assign the observed trend in
relative quenching efficiencies, NM > DMNB, to the smaller
Fig. 4 Fluorescence lifetime quenching of 4 and 5 by NM and DMNB in
toluene solution; dashed line indicates instrument response function
(IRF).
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12507–12516 | 12511

View Online
molecular volume of NM that allows it to diffuse to a fluo-
rophore more rapidly than DMNB. The key to the high
quenching affinities with these new dendrimers is their relatively
long singlet excited state lifetimes. While 4 and 5 have excited
state lifetimes of 7.4 ns and 6.8 ns respectively, the previously
reported first-generation dendrimers with bifluorene containing
chromophores have lifetimes an order of magnitude shorter at
0.8 ns,18 and conjugated polymers typically have lifetimes of
around a nanosecond.9,28 The observed dendrimer PL lifetimes of
this work are about half those reported for unconjugated
carbazole chromophores.30 Hence, we have exploited the direct
correlation between the natural fluorescence lifetime and the
dominance of collisional quenching interactions to maximise the
quenching response towards the nitroaliphatic analytes in solu-
tion. These results provide important insight for the design of
more effective sensing materials: for analytes that display
predominantly collisional quenching, such as nitroaliphatics, the
highest quenching affinities are achieved by maximising the
natural fluorescence lifetime of the sensing molecule.
Thin film quenching measurements
The fluorescence quenching effects that were observed in the
solution-based measurements were translated into a detectable
quenching response in the solid state. Thin films of dendrimers 4
and 5 of ꢄ50 nm thickness were spin-cast onto fused silica
substrates from toluene solutions and the PL of the films at 400
nm (lexc ¼ 320 nm) was monitored over time. Natural photo-
degradation of the dendrimer films was observed with 35–50% of
the initial fluorescence intensity lost within the first 30 minutes of
continuous exposure to 320 nm radiation.
Fig. 6 Thin film quenching of 4 and 5 by DMNB and NM. Traces
indicate an 800 s period from a total sampling time of 1800 s. The
differences in baseline slope arise from the different times at which
sampling was performed during the 1800 s experiments.
The films were exposed to a flow of nitrogen (16.7 cm3 sꢁ1
)
throughout the measurements and this served as the carrier gas
flow for the introduction of analyte vapours. With the nitrogen
flow held constant by a mass flow controller the analytes were
introduced by injecting known volumes of air presaturated with
analyte into the carrier gas stream over 1 second. Such transient
exposure measurements allow for a pulse of analyte of sub-
saturation concentration to be introduced and the reversibility of
the quenching interactions evaluated. The saturated vapour
concentrations of DMNB and NM have been reported to be 2.7
ppm and 41300 ppm respectively.9,31
intensity in Fig. 6 is due to the relative volatility of the two
analytes, with the volumes injected resulting in greater concen-
trations of NM.
The natural rate of photodegradation of thin films of 5 was
greater that of 4 and made the measurements more difficult to
perform. Nevertheless the film measurements were in agreement
with the solution-based measurements, with the quenching of the
fluorescence of 5 being observed to be weaker than 4 with both
analytes. The response of 5 to repeated 1.5 ppm injections of
DMNB was very weak (<1% quench) and injections early in the
quenching measurements were obscured by the photo-
degradation of the film. The responses to injections of NM
vapour are also weaker than the corresponding interactions with
4 giving an average quenching response of approximately 0.002%
ppmꢁ1 in the concentration range below 4400 ppm. Once again
the quenching interaction with NM was reversible with the
baseline fluorescence recovering within 10 seconds of exposure.
The ability to recover the baseline signal under ambient flow
conditions is by no means a ubiquitous quality of films of fluo-
rescent sensing materials with some requiring chemical or
thermal treatment to be reactivated for sensing.6,12,13 The
reversible nature of the quenching interactions of 4 and 5 thin
films with the nitroaliphatics is advantageous for a reusable
sensing array.
Reversible fluorescence quenching responses were observed
upon exposure of dendrimer 4 films to both DMNB and NM
(Fig. 6). A 3% quenching response was detected on exposure to
1.5 ppm of DMNB vapour and the fluorescence recovered to the
baseline within 20 seconds under the flow of nitrogen. Injection
of smaller quantities of DMNB gave a corresponding decrease in
the quenching response of the film with the response to 0.6 ppm
DMNB being beyond the limit of reliable detection. Repeated
injections of 2000 ppm of NM vapour into the carrier gas flow
gave reversible quenching responses from 15–25% and quenching
down to 500 ppm NM was observed. In all cases the baseline
level of fluorescence was almost completely recovered within 10
seconds of NM exposure. In contrast to the results from the
solution measurements the films display a greater quenching
efficiency for DMNB in the solid state. The average initial
quenching for DMNB was approximately 2% ppmꢁ1 compared
to around 0.01% ppmꢁ1 for NM. The difference in signal
12512 | J. Mater. Chem., 2012, 22, 12507–12516
This journal is ª The Royal Society of Chemistry 2012

View Online
at The University of Queensland, Australia. Melting points
€
were measured using a Buchi Melting Point B-545 and are
Conclusions
In summary, we have developed two carbazole-containing den-
drimers capable of detecting the explosive component analytes
DMNB and NM with unparalleled sensitivity in solution. The
levels of detection observed are comparable to those of higher
electron affinity nitroaromatic analytes with dendrimer systems
bearing the same dendrons but different cores.17,18 Sensing in
solution proceeds predominantly via collisional quenching and is
facilitated by long excited state lifetimes. In designing more
effective fluorescent sensing molecules for high electron affinity
targets (whether they be nitroaliphatic or otherwise) one must
consider and understand whether static or collisional quenching
is the dominant mechanism. Nitroaliphatic detection has been an
issue for sensing by oxidative fluorescence quenching and
understanding the collisional nature of the quenching interac-
tions plays an important role in improving sensitivity towards
DMNB and NM in solution. Traditionally solution-based
steady-state Stern–Volmer measurements have been used to
screen the sensing potential of a new fluorophore.3,4,8–10,13
However, it has been observed that a high Stern–Volmer
quenching efficiency whilst useful in a laboratory setting is not
necessarily a promising attribute for thin film sensing arrays. For
deployment in a portable fluorescence sensor device thin film
properties are of the utmost importance. The ability of our
system to detect sub-saturation levels of both DMNB and NM
reversibly in the solid state is a critical step on the path to
developing practical sensing materials.
corrected.
UV-Visible absorption spectra were recorded on a Varian
Cary 5000 UV-Vis-NIR spectrophotometer in spectrophoto-
metric grade solvent. Absorption lmax values are quoted to the
nearest nm and shoulders are denoted by sh. Infrared absorp-
tion spectra were measured with a Perkin Elmer Spectrum 100
FT-IR spectrophotometer as neat samples using an ATR
interface.
Glass transition temperatures (Tg) were recorded on a Perkin
Elmer Diamond DSC Differential Scanning Calorimeter cali-
brated to an indium standard. Decomposition temperatures
(Tdec) are quoted for 5% weight decomposition and were recor-
ded using a Perkin Elmer STA 6000 Simultaneous Thermal
Analyser under a nitrogen atmosphere.
Cyclic and differential pulse voltammetry was undertaken on
a BASi EpsilonEC and values are quoted relative to the fer-
rocenium/ferrocene couple.32 A silver/silver nitrate in acetoni-
trile reference electrode was used in combination with platinum
or glassy carbon working electrodes and a platinum counter
electrode. Acetonitrile for the reference electrode was stored
ꢁ
over
3
A
molecular sieves. Solvents for electrochemical
measurements were dried and distilled on the day of use.
Dichloromethane was dried over calcium hydride and tetrahy-
drofuran was dried over sodium/benzophenone and distilled
before it was dried over and distilled from lithium aluminium
hydride.
Gel permeation chromatography was performed on a Polymer
Laboratories PL GPC 50 using PLgel 3 mm mixed-E columns
(2 ꢅ 300 mm lengths, 7.5 mm diameter) from Polymer Labora-
Experimental details
General methods and synthesis of organic materials
tories calibrated against polystyrene narrow standards [Mp
¼
Unless otherwise indicated all chemicals were obtained from
commercial suppliers and used as received. Solvents were
distilled by rotary evaporation prior to use. Anhydrous tetra-
hydrofuran for reactions was dried over sodium/benzophenone
and distilled immediately prior to use. Thin layer chromatog-
raphy was performed on Merck aluminium plates coated with
silica gel 60 P254. Column chromatography was performed with
Merck silica gel (0.063–0.200 mm). When solvent mixtures are
used the proportions are given by volume.
(162 ꢁ 3.8) ꢅ 104] in tetrahydrofuran containing 0.01% toluene
as a flow marker. The eluent was degassed with helium and
pumped at a rate of 0.5 cm3 minꢁ1 at 39.3 ꢂC. UV absorption at
256 nm was used to detect the eluting species.
Thin films were spin-cast using a Speciality Coating Systems
(SCS) G3-8 Spincoater and surface profilometry was performed
on a Veeco Dektak 150 to determine the film thicknesses.
Spectroscopy
1H and 13C NMR spectra were recorded on either a Bruker
AV400 (400 MHz) or AV500 (500 MHz) spectrometers. Chemical
shifts are reported in parts per million (ppm) and are referenced to
the residual solvent peak (dH(CDCl3) ¼ 7.26, dC(CDCl3) ¼ 77.0).
Coupling constants, J, are reported in Hertz (Hz) to nearest 0.5
Hz. Peak multiplicities are labelled in the following manner:
singlet (s), doublet (d), triplet (t), or multiplet (m). Peak identities
are abbreviated as phenyl (Ph) and carbazolyl (Cbz). Peak
assignments were made with the aid of 13C DEPT, 1H–1H COSY,
1H–13C HSQC and 1H–13C HMBC spectra.
Photoluminescence (PL) and photoluminescence excitation
(PLE) spectra were recorded on a Jobin-Yvon Horiba Fluo-
romax 4 in spectrophotometric grade solvents. Solution photo-
luminescence quantum yield (PLQY) measurements were
measured relative to quinine sulfate (PLQY ¼ 0.55) in 0.5 M
sulfuric acid.21 PLQY values were determined from plots of
absorbance versus integrated PL for the dendrimers relative to
that of a quinine sulfate standard (0.02 < A < 0.10). The standard
errors in the gradients of these plots calculated by least squares
regression were propagated using the chain rule to determine the
error in PLQY. The error in solution PLQY values was found to
be ꢀ0.01.
Matrix-assisted Laser Desorption Ionisation Time of Flight
(MALDI-TOF) mass spectra were recorded on an Applied
Biosystems Voyager MALDI-TOF mass spectrometer using
dithranol (1,8,9-anthracenetriol) as the matrix. Electrospray
ionisation (ESI) mass spectra were recorded on a Bruker HCT
3D Ion Trap with methanol/dichloromethane as solvent.
Elemental analysis was performed by the Microanalytical
Service of the School of Chemistry and Molecular Biosciences
Thin film PLQY measurements were performed using the
procedure described by Greenham et al.33 Thin film samples were
photoexcited using the 325 nm output of a HeCd laser under
nitrogen. Unless otherwise stated the excitation wavelength used
for PL measurements was 320 nm.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12507–12516 | 12513

View Online
Solution quenching measurements
substrates were clamped in a custom built sample holder in
a Jobin Yvon Horiba Fluorolog Tau fluorometer. The dendrimer
film was placed on the opposite side of the substrate to the
excitation beam (320 nm, angle of incidence ¼ 60ꢂ) and the PL at
400 nm was detected at right angles to the excitation pulse facing
the film. A flow of nitrogen (16.7 cm3 sꢁ1) was established into the
chamber from an aperture facing the film (angle of incidence ¼
60ꢂ) and held constant throughout the measurement. The flow of
nitrogen was controlled by a Bronkhorst EL-FLOW Select mass
flow controller calibrated for nitrogen within the flow range 2.5–
125 cm3 sꢁ1. The apparatus was purged with nitrogen for 20 min.
The PL intensity was then monitored for 30 min during which
time analytes were injected. Analytes were introduced by
manually injecting a known volume (between 0.2 and 20 cm3) of
saturated analyte vapour into the nitrogen carrier stream over
1 s. Cotton wool was placed inside the syringes at the outlet to
prevent the injection of particles of analyte into the carrier gas
flow. The stoppered syringes were allowed to saturate with
analyte vapours at 22 ꢂC overnight prior to injection. The
approximate concentration of analyte introduced was estimated
from the reported saturated vapour concentrations9,31 and the
dilution ratio of the analyte in the nitrogen carrier stream.
An optically dilute (absorbance approximately 0.1 at peak)
dendrimer solution was prepared and used as a stock solution.
Analyte solutions of known concentration were prepared from
the dendrimer stock such that dendrimer concentrations
remained constant throughout all measurements. In all experi-
ments the initial dendrimer solution volume was 2.5 cm3 and
analyte solutions were added in 0.05 cm3 aliquots.
For steady-state Stern–Volmer measurements the initial
absorption and PL spectra of the dendrimer solutions were
measured before analyte solutions were added. After each
addition of analyte the absorption and PL spectra were again
recorded. Addition of the analyte to the solution caused atten-
uation of the excitation beam and absorption of lower wave-
length dendrimer fluorescence due to UV absorption by the
analyte. These two effects can be reliably corrected for providing
the measurements are done within the Beer–Lambert regime
(absorbance # 0.6 at excitation wavelength) and hence the
analyte concentrations were selected to ensure that this criterion
was fulfilled. All measurements were completed in triplicate with
separately prepared analyte solutions. The largest error in the
calculation of steady-state Stern–Volmer constants was found to
arise from the corrections applied for analyte absorption,
specifically in the accurate measurement of wavelengths on the
spectrophotometer and fluorometer. We have assumed a toler-
ance of ꢀ1 nm in wavelength recognition between the instru-
ments and the error quoted is the deviation in the observed Ksv
value when varying the excitation wavelength within this range.
Time-resolved Stern–Volmer measurements were performed
on a Jobin-Yvon Horiba Fluorolog FL3-11 with a time-corre-
lated single photon counting module. The materials were excited
at 372 nm with 1.2 ns pulses and a repetition rate of 1 MHz using
the output of a Horiba Scientific NanoLED pulsed diode light
source. Fluorescence decays were measured at 400 nm with an
instrument response function (IRF) with a FWHM of approxi-
mately 1.5 ns. The data was fitted to a single exponential decay
following convolution with the IRF. Due to the proximity of the
detection and excitation wavelengths scatter of the excitation
beam was observed in the decays. The background signal was
recorded using spectrophotometric grade toluene as a scatterer,
which was subtracted from the observed data prior to fitting. The
errors associated with the calculation of the collisional Stern–
Volmer constants are generally small due to the lack of correc-
tions needed for analyte absorption. The error quoted is the
standard error in the gradient of the time-resolved Stern–Volmer
plots calculated by least squares regression.
4-{3,5-Bis[4-(2-ethylhexyloxy)phenyl]phenyl}-2-methylbut-3-
yn-2-ol (3). 619 (0.100 g, 0.177 mmol) was dissolved in piperidine
(1.4 cm3). Copper(I) iodide (0.002 g, 8.8 mmol) was added to the
solution and the mixture was sparged with nitrogen for 10 min.
Tetrakis(triphenylphosphine)palladium(0) (0.010 g, 8.8 mmol)
was added to the mixture, which was then sparged with nitrogen
for 10 min. 2-Methylbut-3-yn-2-ol (0.10 cm3, 1.03 mmol) was
added and the mixture was stirred at 80 oC for 16 h. The mixture
was allowed to cool to room temperature before the volatiles
were removed in vacuo. The residue was purified by column
chromatography over silica using dichloromethane : n-hexane
(7 : 3) as eluent to yield 3 as an orange oil (0.092 g, 91%). Found
C, 82.2; H, 9.4%; C39H52O3 requires C, 82.4; H, 9.2%. lmax
(CH2Cl2/nm): 264 (log 3/dm3 molꢁ1 cmꢁ1 ¼ 4.69), 273 sh (4.67).
ꢀnmax (film, ATR/cmꢁ1): 3349 (br, O–H Stretch), no prominent
C^C stretch was observed. dH (500 MHz, CDCl3): 0.92–0.98
(12H, m, ethylhexyl CH3), 1.32–1.59 (16H, m, ethylhexyl CH2),
1.67
(6H,
s,
HOC(CH3)2ꢁ),
1.74–1.79
(2H,
m,
OCH2CH(C4H9)(C2H5)), 2.17 (1H, s, HOC(CH3)2ꢁ), 3.90 (4H,
m, OCH2CH(C4H9)(C2H5)), 6.99 (4H, ½ AA0BB0, surface PhH),
7.54–7.56 (6H, m, branching and surface PhH), 7.66 (1H, dd, J ¼
2.0 Hz and 2.0 Hz, branching PhH). dC (126 MHz, CDCl3): 11.1,
14.1, 23.0, 23.8, 29.1, 30.5, 31.5, 39.4, 65.6, 70.6, 82.3, 93.6, 114.8,
123.4, 125.2, 128.1 (two overlapping peaks), 132.6, 141.4, 159.2.
m/z [ESI+]: 551.4 (M ꢁ OH+), 569.5 (M + H+).
Thin film quenching measurements
3,6-Bis{3,5-bis[4-(2-ethylhexyloxy)phenyl]phenyl}-9-n-hexylcarbazole
(4). 134 (0.076 g, 0.186 mmol) and 219 (0.284 g, 0.464 mmol) were
dissolved in toluene (3 cm3). Aqueous potassium carbonate (2 M,
1 cm3) and tert-butanol (1 cm3) were added and the mixture was
deoxygenated by sparging with nitrogen for 10 min. Tetrakis-
(triphenylphosphine)palladium(0) (0.011 g, 9.26 mmol) was
added and the mixture was deoxygenated by sparging with nitrogen
for 10 min. The mixture was stirred at 100 ꢂC under nitrogen for 4 h
before it was allowed to cool to room temperature. The mixture
was diluted with toluene (25 cm3) and water (25 cm3) and the
Fused silica substrates (12 mm diameter) were cleaned by rinsing
with acetone, 2-propanol (under ultrasonication) and toluene
prior to dendrimer deposition. Dendrimer films were spin-cast
from 10 mg cmꢁ3 solutions in toluene. Dendrimer solutions were
pipetted onto the surface of the cleaned substrates until the
surface was covered by the solution and the films were immedi-
ately spun at 2000 rpm for 2 min (acceleration ¼ 2000 rpm sꢁ1).
Under these spinning conditions the film thicknesses were found
to be 48 ꢀ 5 nm and 52 ꢀ 6 nm for 4 and 5 respectively. The
12514 | J. Mater. Chem., 2012, 22, 12507–12516
This journal is ª The Royal Society of Chemistry 2012

View Online
layers were separated. The aqueous layer was extracted with
toluene (2 ꢅ 25 cm3) and the combined organics were washed
with water (2 ꢅ 25 cm3) and brine (25 cm3). The organics were
dried over anhydrous magnesium sulfate and the organics dec-
anted. The magnesium sulfate was washed with additional
solvent and the decanted organics combined before the volatiles
were removed in vacuo. The residue was purified by column
chromatography over silica using dichloromethane : n-hexane
(1 : 4–3 : 7) as eluent to yield 4 as an off-white glassy solid (0.119
g, 52%). Found C, 84.6; H, 9.0; N, 1.2%; C86H109NO4 requires C,
84.6; H, 9.0; N, 1.1%. lmax (CH2Cl2/nm): 259 sh (log 3/dm3 molꢁ1
cmꢁ1 ¼ 5.19), 263 (5.20), 281 sh (5.10), 296 sh (4.98), 316 sh
(4.54), 336 sh (3.94), 354 sh (3.57). dH (400 MHz, CDCl3): 0.92–
1.00 (27H, m, hexyl and ethylhexyl CH3), 1.36–1.57 (38H, m,
PhOCH2CH(C4H9)(C2H5)), 1.86–1.92 (2H, m, NCH2CH-
2(C4H9)), 3.91 (8H, m, PhOCH2CH(C4H9)(C2H5)), 4.31 (2H, t,
J ¼ 7.0 Hz, NCH2CH2(C4H9)), 7.01 and 7.61 (16H, AA0BB0,
surface PhH), 7.40 (2H, d, J ¼ 8.5 Hz, CbzH), 7.69–7.72 (8H, m,
branching PhH and CbzH), 8.34 (2H, d, J ¼ 1.5 Hz, CbzH). dC
(126 MHz, CDCl3): 11.1, 14.0, 14.1, 22.5, 23.1, 23.9, 26.9, 28.9,
29.1, 30.5, 31.5, 39.4, 43.4, 70.6, 88.1, 90.5, 109.0, 113.9, 114.9,
122.5, 124.2, 124.5, 124.9, 128.0, 128.2, 129.8, 132.8, 140.5, 141.5,
159.3. m/z [MALDI-TOF] anal. calcd 1267.8 (97%), 1268.8
(100%), 1269.8 (51%), 1270.8 (17%), 1271.8 (5%). Found 1268.0
(87%), 1269.0 (100%), 1269.9, (56%), 1270.9 (15%). Tg ¼ 25.1 ꢂC.
T5% dec ¼ 399.8 ꢂC. E1/2(OX, DCM, DPV) ¼ 0.75 V. GPC, Mw
¼
ꢀ
ꢀ
ꢀ
1863, Mn ¼ 1856, Mv ¼ 1862.
9-n-Hexylcarbazole was prepared by a literature procedure35
hexyl
and
ethylhexyl
CH2),
1.76–1.83
(4H,
m,
for optical studies.
PhOCH2CH(C4H9)(C2H5)), 1.93–2.00 (2H, m, NCH2CH2-
(C4H9)), 3.94 (8H, m, PhOCH2CH(C4H9)(C2H5)), 4.39 (2H, t,
J ¼ 7.0 Hz, NCH2CH2(C4H9)), 7.05 and 7.70 (16H, AA0BB0,
surface PhH), 7.53 (2H, d, J ¼ 8.5 Hz, CbzH), 7.72 (2H, dd, J ¼
1.5 Hz and 1.5 Hz, branching PhH), 7.85–7.87 (6H, m, over-
lapping branching PhH and CbzH), 8.50 (2H, d, J ¼ 1.5 Hz,
CbzH). dC (126 MHz, CDCl3): 11.3, 14.2, 14.3, 22.7, 23.2, 24.0,
27.2, 29.2, 29.3, 30.7, 31.8, 39.6, 43.5, 70.7, 109.3, 115.0, 119.3,
123.7, 123.8, 124.5, 125.7, 128.5, 132.7, 133.9, 140.7, 142.1, 143.2,
159.3. m/z [MALDI-TOF] anal. calcd 1219.8 (100%), 1220.8
(97%), 1221.8 (47%), 1222.8 (15%), 1223.8 (4%). Found 1219.9
Acknowledgements
The Authors acknowledge funding from the Australian Research
Council (DP0986838) and The University of Queensland (Stra-
tegic Initiative – Centre for Organic Photonics & Electronics).
AJC is supported by an Endeavour International Postgraduate
Research Scholarship and a University of Queensland Research
Scholarship. PLB is supported by an Australian Research
Council Federation Fellowship (Project no. FF0668728). PM is
supported by a University of Queensland Vice Chancellor’s
Senior Research Fellowship.
(100%), 1220.9 (95%), 1221.9 (55%), 1222.9 (15%). Tg ¼ 36.8 ꢂC
ꢂ
(no melting transitions were observed below 200 C). T5% dec
¼
399.0 ꢂC. E1/2(OX, THF, CV) ¼ 0.71 V. GPC, Mw ¼ 1613, Mn ¼
ꢀ
ꢀ
ꢀ
1606, Mv ¼ 1612.
References
3,6-Bis({3,5-bis[4-(2-ethylhexyloxy)phenyl]phenyl}ethynyl)-9-n-
hexylcarbazole (5). 134 (0.050 g, 0.120 mmol), 3 (0.150 g, 0.264
mmol) and tetra-n-butylammonium bromide (0.002 g, 5.66 mmol)
were dissolved in toluene (3 cm3). Aqueous sodium hydroxide
(20% w/w, 3 cm3) was added and the mixture was sparged with
nitrogen for 10 min. Copper(I) iodide (0.001 g, 5.99 mmol) was
added and the mixture was deoxygenated by sparging with
nitrogen for 10 min. Tetrakis(triphenylphosphine)palladium(0)
(0.007 g, 5.99 mmol) was added and the mixture was deoxygen-
ated by sparging with nitrogen for a further 10 min. The mixture
1 A. W. Czarnik, Nature, 1998, 394, 417.
2 J.-S. Yang and T. M. Swager, J. Am. Chem. Soc., 1998, 120, 11864.
3 S. J. Toal and W. C. Trogler, J. Mater. Chem., 2006, 16, 2871.
4 M. Guo, O. Varnavski, A. Narayanan, O. Mongin, J.-P. Majoral,
M. Blanchard-Desce and T. Goodson III, J. Phys. Chem. A, 2009,
113, 4763.
5 H. Cavaye, A. R. G. Smith, M. James, A. Nelson, P. L. Burn,
I. R. Gentle, S.-C. Lo and P. Meredith, Langmuir, 2009, 25, 12800.
6 G. Tang, S. S. Y. Chen, P. E. Shaw, K. Hegedus, X. Wang, P. L. Burn
and P. Meredith, Polym. Chem., 2011, 2, 2360.
7 J. C. Sanchez, S. J. Toal, Z. Wang, R. E. Dugan and W. C. Trogler, J.
Forensic Sci., 2007, 52, 1308.
ꢂ
8 J. C. Sanchez, A. G. DiPasquale, A. L. Rheingold and W. C. Trogler,
Chem. Mater., 2007, 19, 6459.
was stirred at 80 C under nitrogen for 4 h and was allowed to
cool to room temperature. The mixture was diluted with toluene
(25 cm3) and water (25 cm3) and the layers were separated. The
aqueous layer was extracted with toluene (2 ꢅ 25 cm3) and the
combined organics were washed with water (2 ꢅ 25 cm3) and
brine (25 cm3). The organics were dried over anhydrous
magnesium sulfate and the organics were decanted. The
magnesium sulfate was washed with additional solvent and the
decanted organics combined before the volatiles were removed in
vacuo. The residue was purified by column chromatography over
silica using dichloromethane : n-hexane (1 : 4) as eluent to yield 5
as an orange glassy solid (0.071 g, 47%). Found C, 84.9; H,
8.8; N, 1.1%; C89H107NO6 requires C, 85.2; H, 8.7; N, 1.1%. lmax
(CH2Cl2/nm): 269 sh (log 3/dm3 molꢁ1 cmꢁ1 ¼ 4.91), 284 (5.01),
314 sh (4.82), 345 (4.67). ꢀnmax (film, ATR/cmꢁ1): 2207 (w, C^C
stretch). dH (500 MHz, CDCl3): 0.88 (3H, t, J ¼ 7 Hz, hexyl
CH3), 0.91–0.98 (24H, m, ethylhexyl CH3), 1.29–1.58 (38H, m,
9 S. W. Thomas III, J. P. Amara, R. E. Bjork and T. M. Swager, Chem.
Commun., 2005, 4572.
10 M. E. Germain and M. J. Knapp, Chem. Soc. Rev., 2009, 38, 2543.
11 K. J. Balkus Jr, T. J. Pisklak, G. Hundt, J. Sibert and Y. Zhang,
Microporous Mesoporous Mater., 2008, 112, 1.
12 A. Lan, K. Li, H. Wu, D. H. Olson, T. J. Emge, W. Ki, M. Hong and
J. Li, Angew. Chem., Int. Ed., 2009, 48, 2334.
13 T. Naddo, X. Yang, J. S. Moore and L. Zang, Sens. Actuators, B,
2008, 134, 287.
14 D. Jones, R. Augsten and K. Feng, J. Therm. Anal. Calorim., 1995, 44,
533.
15 S. B. Markofsky, Ullman’s Encyclopedia of Industrial Chemistry,
Wiley-VCH Verlag GmbH & Co, 2000.
16 M. E. Germain, T. R. Vargo, P. G. Khalifah and M. J. Knapp, Inorg.
Chem., 2007, 46, 4422.
17 D. A. Olley, E. J. Wren, G. Vamvounis, M. J. Fernee, X. Wang,
P. L. Burn, P. Meredith and P. E. Shaw, Chem. Mater., 2010, 23, 789.
18 H. Cavaye, P. E. Shaw, X. Wang, P. L. Burn, S.-C. Lo and
P. Meredith, Macromolecules, 2010, 43, 10253.
19 S.-C. Lo, E. B. Namdas, P. L. Burn and I. D. W. Samuel,
Macromolecules, 2003, 36, 9721.
hexyl
and
ethylhexyl
CH2),
1.75–1.80
(4H,
m,
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12507–12516 | 12515

View Online
20 P. L. Burn, R. Beavington, M. J. Frampton, J. N. G. Pillow,
M. Halim, J. M. Lupton and I. D. W. Samuel, Mater. Sci. Eng., B,
2001, 85, 190.
21 J. N. Demas and G. A. Crosby, J. Phys. Chem., 1971, 75, 991.
22 S. Dong, Z. Li and J. Qin, J. Phys. Chem. B, 2008, 113, 434.
23 J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer,
New York, 3rd edn, 2006.
24 W. J. Aixill, P. B. Mills, R. G. Compton, F. Prieto and M. Rueda, J.
Phys. Chem. B, 1998, 102, 9187.
25 S. F. Alvarado, P. F. Seidler, D. G. Lidzey and D. D. C. Bradley,
Phys. Rev. Lett., 1998, 81, 1082.
28 S. W. Thomas III, G. D. Joly and T. M. Swager, Chem. Rev., 2007,
107, 1339.
29 P. Atkins and J. de Paula, Atkins’ Physical Chemistry, Oxford
University Press, Oxford, 8th edn, 2006.
30 G. E. Johnson, J. Phys. Chem., 1980, 84, 2940.
31 W. M. Jones and W. F. Giauque, J. Am. Chem. Soc., 1947, 69, 983.
32 B. W. D’Andrade, S. Datta, S. R. Forrest, P. Djurovich,
E. Polikarpov and M. E. Thompson, Org. Electron., 2005, 6, 11.
33 N. C. Greenham, I. D. W. Samuel, G. R. Hayes, R. T. Phillips,
Y. A. R. R. Kessener, S. C. Moratti, A. B. Holmes and
R. H. Friend, Chem. Phys. Lett., 1995, 241, 89–96.
ꢂ
ꢂ
26 H. Cavaye, H. Barcena, P. E. Shaw, P. L. Burn, S.-C. Lo and
P. Meredith, Organic semiconductors in sensors and bioelectronics
II, Proc. SPIE, 2009, 7418, 741803.
34 C. Beginn, J. V. Grazulevicius, P. Strohriegl, J. Simmerer and
D. Haarer, Macromol. Chem. Phys., 1994, 195, 2353.
35 H. Wang, G. Chen, X. Xu, H. Chen and S. Ji, Dyes Pigm., 2010, 86,
238.
27 D. Zhao and T. M. Swager, Macromolecules, 2005, 38, 9377.
12516 | J. Mater. Chem., 2012, 22, 12507–12516
This journal is ª The Royal Society of Chemistry 2012