The effect of intermolecular interactions on the electro-optical properties of porphyrin dendrimers with conjugated dendrons

Article abstract of DOI:10.1039/b209910c

We have synthesised a new family of dendrimers with stilbene dendrons attached to a porphyrin core via a stilbene unit and compared their properties with a family of dendrimers with the same core and dendrons but with the dendrons attached via a phenyl un

Full text of DOI:10.1039/b209910c

View Article Online / Journal Homepage / Table of Contents for this issue
The effect of intermolecular interactions on the electro-optical
properties of porphyrin dendrimers with conjugated dendrons
Michael J. Frampton,a Steven W. Magennis,b Jonathan N. G. Pillow,a Paul L. Burn*a
and Ifor D. W. Samuel*b
aThe Dyson Perrins Laboratory, Oxford University, South Parks Road, Oxford, UK
OX1 3QY
bOrganic Semiconductor Centre, School of Physics and Astronomy, University of St Andrews,
North Haugh, St Andrews, Fife, UK KY16 9SS
Received 9th October 2002, Accepted 20th November 2002
First published as an Advance Article on the web 20th December 2002
We have synthesised a new family of dendrimers with stilbene dendrons attached to a porphyrin core via a
stilbene unit and compared their properties with a family of dendrimers with the same core and dendrons but
with the dendrons attached via a phenyl unit. The oxidation and reduction half potentials of the two dendrimer
families were found to be the same and independent of generation indicating that the dendrons were not
creating a micro-environment for the core. However, the rate of heterogeneous electron transfer was found to
be strongly dependent on link type and generation. The photoluminescence quantum yield (PLQY) of the
dendrimers was also found to be strongly dependent on the method of attachment of the core. In solution the
dendrimers with the stilbene link between core and dendrons had PLQYs 1.5 times higher than their phenyl
counterparts but in the solid state the trend was reversed with the phenyl linked dendrimers generally having a
higher PLQY. The difference in properties has been assigned to the comparative openness of the dendrimer
architectures and the effect of the dendrons on the shape of the porphyrin core.
the porphyrin core. However, these studies have tended to be
Introduction
carried out in solution and little is known about the effect of the
dendrimer structure on the porphyrin cores in the solid state. In
contrast to the studies on porphyrin cored dendrimers with
saturated links there have been fewer publications of porphyrin
cored dendrimers with conjugated dendrons, with stilbene6 and
phenylene7,8 based dendrons having been reported. Our
interest in porphyrin cored dendrimers has arisen from the
reports that porphyrins can be used as red light-emitters in
organic light-emitting diodes (OLEDs).9–11 Porphyrins have a
propensity to p-stack and be non-luminescent in the solid state.
In OLED devices this problem is generally overcome by
incorporating the porphyrins as a guest in a blend at low
concentration. Our alternative approach to controlling the
porphyrin–porphyrin interactions in the solid state is to simply
incorporate the porphyrin as the core of a dendrimer.
Functional dendrimers that contain the active component at
the core of the dendrimer are becoming increasingly studied.
The dendritic architecture provides an elegant method of
controlling the intermolecular interactions of the active core.
The control is achieved at three levels; first, the nature of the
dendron attached to the core, second the number of dendrons
attached to the core, and finally the generation of the dendron.
Dendrons can be divided into two main types, those that
contain conjugated links between the branching points and
those that have saturated links. It is not surprising that
dendrimers with the same core but different dendron types,
saturated or unsaturated, would give rise to dendrimers with
different properties.1 One of the more widely studied families of
dendrimers are those in which the core is comprised of a
porphyrin. Most of these studies have utilised flexible dendrons
with saturated links between the branching points.2–5 In such
dendrimers the nature of the dendron and generation have had
a strong influence on the photophysical and redox properties of
We have previously reported that porphyrin cored dendri-
mers with stilbene based dendrons attached via phenyl moieties
to the meso-positions can be used as neat layers in OLEDs.12,13
We found that the efficiency of the OLEDs was generation
Fig. 1 Porphyrin cored dendrimers with stilbene dendrons linked to the core via phenyl units.
DOI: 10.1039/b209910c
J. Mater. Chem., 2003, 13, 235–242
This journal is # The Royal Society of Chemistry 2003
235

View Article Online
dependent with the second generation being twice as efficient as
the first. However, the origin of this dependence of efficiency on
generation was not clear, that is whether it was due to a
difference in luminescence efficiency or charge injection and
transport. In this paper we probe the role of the stilbene
dendrons in controlling the intermolecular interactions of the
porphyrin cores both in solution and the solid state using a
combination of electrochemistry, photoluminescence and
photoluminescence quantum yield measurements. We discuss
the effect of generation on the opto-electronic properties of two
families of dendrimers that only differ in the way the stilbene
dendrons are attached to the porphyrin ring. The first family of
dendrimers (1, 2, and 3) have the stilbene dendrons attached to
the phenyl ring that is directly connected to the meso-positions
of the porphyrin core whilst the second family (15, 16, and 17)
has the dendrons connected via the more extended stilbene unit.
Finally, we propose an explanation for the observation that
OLED efficiency increases with generation.
Results and discussion
Syntheses
The synthesis of the first, 2 ([G-1]PP), and second, 3 ([G-2]PP),
generation tetra-phenylporphyrin dendrimers Fig. 1 has been
reported previously.6 The strategy for the formation of the
tetra-stilbene analogues 15, 16, and 17 is shown in Scheme 1.
The basic method involves extending the stilbene dendrons
with aldehyde foci, 6, 7, and 8, by the attachment of a
4-styrylbenzaldehyde unit before subsequent condensation
with pyrrole to give the porphyrin. The extension of the foci
was achieved in two steps. The first step involved the reaction
of dimethyl 4-cyanobenzylphosphonate 5 with the aldehyde
focused dendrons (6–8) to give dendrons with a 4-cyanostyryl
focus (9–11). The cyano groups of the 4-cyanostyryl focused
dendrons were then reduced with diisobutylaluminium hydride
(DIBAL-H) and a subsequent aqueous acid hydrolysis gave the
aldehyde moiety required for the porphyrin formation. The
phosphonate for the Horner–Wadsworth–Emmons reaction
was obtained from 4-cyanobenzyl bromide 4 in a yield of 90%
after purification by reaction with an excess of trimethyl
phosphite at 100 uC.
Following this general procedure the zeroeth generation
aldehyde 6 was coupled with phosphonate 5 at room tempera-
ture using potassium tert-butoxide as base and tetrahydrofuran
as solvent. After purification by column chromatography
zeroeth generation 4-cyanostyryl focused dendron 9 was
isolated in an 89% yield. Reduction of 9 with DIBAL-H
followed by an aqueous acid work-up gave the 4-styrylbenz-
aldehyde focused dendron 12 as a mixture of cis- and trans-
isomers. This suggests that 9 also contained some cis-isomer.
The cis-isomer was equilibrated to the trans-isomer by heating
with catalytic iodine in toluene at 100uC. After purification 12
was isolated in a 95% yield. The first and second generation
4-cyanostyryl focused dendrons 10 and 11 were prepared in an
analogous manner to the zeroeth generation and were isolated
in yields of 92% and 95% respectively. Reduction of nitrile
moieties was also carried out under the same conditions as the
zeroeth generation, but no isomerisation step was required, and
after purification by column chromatography the aldehyde 13
was obtained in a yield of 80% and 14 in a yield of 73%.
The final step in the synthesis of the tetra-stilbeneporphyrins
15–17 was the trifluoroacetic acid catalysed condensation of the
stilbene aldehyde focused dendrons with pyrrole. A solution of
the zeroeth generation stilbene aldehyde 12 with one equivalent
of pyrrole and one equivalent of trifluoroacetic acid in
dichloromethane was stirred in the dark under argon for one
week to give the zeroeth generation tetra-stilbeneporphyrin 15
([G-0]StP) in a yield of 25% after 2,3-dichloro-5,6-dicyano-1,4-
benzoquinone oxidation and purification. The reactions to
Scheme 1 Reagents and conditions: i, P(OMe)3, 100 uC; ii, t-BuOK,
THF, rt; iii, DIBAL-H, THF, 278 uC, then HCl(aq), rt; iv, pyrrole,
CF3CO2H, CH2Cl2, Ar, rt then DDQ.
prepare the first ([G-1]StP) and second ([G-2]StP) generation
tetra-stilbeneporphyrin dendrimers 16 and 17 were analogous
and after purification the porphyrins were collected in yields of
15% and 22% respectively.
In our previous study of the tetra-phenylporphyrins we
found that there was a strong correlation between the hydro-
dynamic radii and the electrochemical and device properties of
the dendrimers. With the extra styryl unit it was expected that
the hydrodynamic radii of the tetra-stilbeneporphyrins would
be greater than for the tetra-phenylporphyrins. Gel permeation
chromatography (GPC) showed that the dendrimers were
236
J. Mater. Chem., 2003, 13, 235–242

View Article Online
¯
mono-disperse and we used Mv along with the Hester–Mitchell
equation and the Mark–Houwink relationship to estimate their
hydrodynamic radii (Rh).13 The Mv of the [G-0]StP, [G-1]StP,
¯
and [G-2]StP generation tetra-stilbeneporphyrins were 2050,
3858, and 6330 respectively corresponding to hydrodynamic
˚
radii of 10, 14, and 19 A respectively. These radii are larger
than those of the tetra-phenylporphyrin derivatives which had
˚
radii of 7, 12, and 16 A for the zeroeth to second generation
respectively.13 This corresponds to 43%, 17%, and 19%
increases in radii for the equivalent generations. For the
tetra-phenylporphyrins it was observed that in moving from
the zeroeth to second generation there was a dramatic change
in the porphyrins’ electrochemical properties. Hence the
corresponding increase in radii in going to the tetra-
stilbeneporphyrins might also be expected to have a strong
effect on the properties of the porphyrin cores.
Electrochemistry
We first studied the effect of the introduction of the styryl
spacer on the electrochemical properties. The electrochemical
properties of the tetra-stilbeneporphyrin dendrimers were
studied by cyclic voltammetry and the experiments were
carried out under the same conditions used for the tetra-
phenylporphyrins so that direct comparison between the two
families could be made. The measurements were taken at room
temperature as 1.0 mM solutions of the dendrimers in
dichloromethane and at a scan rate of 35 mV s21. For all
traces the chemical reversibility associated with each electro-
chemical process was checked by performing repetitious scans.
The cyclic voltammograms of [G-0]StP, [G-1]StP, and [G-2]StP
are shown in Figs. 2 to 4 and a summary of the results for the
tetra-stilbeneporphyrins and tetra-phenylporphyrins are in
Table 1. For [G-0]PP three porphyrin based redox processes
were observed, two chemically reversible reductions (21.76
and 22.05 V) and one chemically reversible oxidation (0.44 V).
For the [G-0]StP we again observed two porphyrin based
chemically reversible reductions but in contrast to [G-0]PP we
also saw two chemically reversible oxidations (Fig. 2). The
potentials at which the reductions and first oxidations take
place are similar for both porphyrin types. This is consistent
with the stilbene units being orthogonal to the porphyrin plane
and hence not increasing the conjugation length of the
porphyrin core. However, the observation of the chemical
reversibility of the porphyrin dication for [G-0]StP indicates
that the stilbene unit does have an effect on the oxidative
processes. The separation between the anodic (Epa) and
Fig. 3 Cyclic voltammogram of [G-1]StP 16, scan rate ~ 35 mV s21
potentials are quoted against the ferrocenium/ferrocene couple.
;
Fig. 4 Cyclic voltammogram of [G-2]StP 17, scan rate ~ 35 mV s21
potentials are quoted against the ferrocenium/ferrocene couple.
;
cathodic (Epc) peaks of the first oxidation and reduction of
[G-0]StP 15 and [G-0]PP 1 were around 90 mV.
For the first generation tetra-phenylporphyrin [G-1]PP 2 two
chemically reversible reductions and one chemically reversible
oxidation were also observed. The potentials at which the
redox processes occurred were close to those for the zeroeth
generation. For the [G-1]StP 16 (Fig. 3) the reduction and first
oxidation potentials were the same as those for [G-0]StP 15.
Table 1 Electrochemical data for [G-n]P and [G-n]StP
Dendrimer
E(o1)/V E(o2)/V
E(r1)/V E(r2)/V DEp(r1)a/V
1
2
3
15
16
17
10.44
10.56
—
10.45
10.47
10.54
—
—
—
10.72
—
—
21.76
21.72
21.75
21.71
21.68
21.70
22.08
22.04
22.17
22.02
21.98
22.03
0.09
0.24
0.58
0.09
0.09
0.14
ar1
~ first reduction. Conditions: solvent ~ dichloromethane;
[Dendrimer] ~ 1.0 mM; [(n-Bu)4NPF6] ~ 0.1 M; glassy carbon
working electrode; platinum wire counter electrode; Ag/3 M NaCl/
AgCl(sat) reference electrode; ferrocenium/ferrocene couple as
Fig. 2 Cyclic voltammogram of [G-0]StP 15, scan rate ~ 35 mV s21
potentials are quoted against the ferrocenium/ferrocene couple.
;
standard; scan rate ~ 35 mV s21
.
J. Mater. Chem., 2003, 13, 235–242
237

View Article Online
However, unlike the zeroeth generation only one chemically
reversible oxidation was seen for the first generation dendrimer.
The second porphyrin based oxidation of the first generation
was determined to be close to the potential of a chemically
irreversible oxidation, which we believe is due to oxidation of
the dendrons. Although the half potentials for the reductions of
2 and 16 are the same there was a distinct difference in the Epa
2 Epc for the reductions. For the first reduction in going from
[G-0]PP 1 to [G-1]PP 2 Epa 2 Epc increases from 90 mV to
240 mV indicating slowed heterogeneous electron transfer from
the electrode to the porphyrin core. This is consistent with an
increase in the hydrodynamic radii of the porphyrins. In
contrast, in going from the zeroeth to first generation in the
tetra-stilbeneporphyrin there is no change in Epa 2 Epc which is
#90 mV in both cases. This is in spite of the hydrodynamic
radii of [G-1]StP 16 being 40% and 17% larger than that of
[G-0]StP 15 and [G-1]PP 2, respectively. This suggests that
moving the dendron attachment point further from the core
gives a more open flexible structure allowing the porphyrin to
be more easily accessed. Interestingly we do not see any
hysteresis in the oxidation of [G-1]PP 2 although we do not
have an explanation for this at this stage.
Electrochemical analysis of [G-2]PP 3 showed that it had
two chemically reversible reductions but no chemically rever-
sible oxidation was observed. For the second generation tetra-
stilbeneporphyrin dendrimer [G-2]StP 17 we again observed
two chemically reversible reductions and in contrast to [G-2]PP
3 one chemically reversible oxidation (Fig. 4). For [G-2]PP 3
there is a large hysteresis observed for the reduction processes
with Epa 2 Epc ~ 580 mV for the first reduction indicating that
the rate of heterogeneous electron transfer is significantly
slowed. For [G-2]StP some hysteresis is also observed for the
reduction and oxidation processes with the Epa 2 Epc being
#50% greater, at 140–160 mV, than for the first generation.
This clearly shows that with the larger dendron the porphyrin
core is more shielded although less so than for the tetra-
phenylporphyrin dendrimers. This is again consistent with the
more open structure caused by the styrene link of the dendron
to the porphyrin leaving the porphyrin core more exposed.
It is important to note that the redox potentials at which
these processes occurred were the same as those for the lower
generations in each of the families indicating that these rigid
stilbene dendrons do not create a microenvironment that
changes the redox properties of the porphyrins.
Fig. 5 Solution UV-vis and PL spectra for 15, 16, and 17.
two families. For the tetra-phenylporphyrin family the weight-
ing of the Q(0,0) and Q(0,1) bands was seen to be generation
dependent with the ratio of the Q(0,0) to the Q(0,1) changing
from 1 : 0.7 to 1 : 0.9 in going from the zeroeth to second
generation. The relative intensities of the Q(0,0) and Q(0,1)
emission bands is dependent on the relative orbital energies of
the excited states of the porphyrins.15,16 Therefore, different
ratios of the intensities of the two peaks could imply different
porphyrin excited state energies. Such a change has not been
reported for porphyrin cored dendrimers with flexible den-
drons. We believe that the change in Q(0,0) and Q(0,1)
intensities arises from the steric demands of rigid dendrons
causing the shape of the porphyrin core to deviate, to different
degrees, from planarity. This causes a change to the excited
state energy levels which in turn alters the electronic spectra.17
In contrast, no large change is seen in the weighting for the
tetra-stilbeneporphyrins and it is only for the second genera-
tion that a slight increase in the Q(0,1) compared to the Q(0,0)
is observed. In fact it is interesting to note that the Q(0,1)
transition for the tetra-stilbeneporphyrins is weaker than for
the tetra-phenylporphyrins which would also be consistent with
the lower steric demand of the dendrons. That is, the porphyrin
ring does not have to change its shape significantly with
generation.
In going from solution to the film there is a change in the PL
spectra for the tetra-stilbeneporphyrin dendrimers (Fig. 6). For
all three generations there is a modest red shift in the peaks,
#23 nm for the Q(0,0) and 12 nm for the Q(0,1) transitions, a
broadening of the emission, and a pronounced red tail. The
broadening of the spectra and red tail suggests the possibility of
a contribution to the emission from excimers or aggregates.
The contribution to the red end of the spectrum is greatest for
the zeroeth generation which is consistent with it having the
most open structure. The PL spectrum of [G-2]StP is closest to
that of the solution spectrum which is also consistent with it
Therefore, the electrochemistry indicates that extension of
the attachment point of the dendron from a phenyl to a stilbene
unit does not change the redox potentials or HOMO–LUMO
energy gap of the porphyrin cores. In addition, although the
tetra-stilbeneporphyrins have larger hydrodynamic radii than
the tetra-phenylporphyrins, the stilbene moiety gives a more
open structure causing the porphyrin core to be more
accessible.
UV-visible spectra and photoluminescence
The solution UV-visible absorption and photoluminescence
(PL) spectra of the tetra-stilbeneporphyrins are shown in
Fig. 5. The absorption spectra consist of dendron absorption
with a peak around 310 nm, which increases in intensity with
increasing generation number,the porphyrin Sore´tbandat430nm
and four porphyrin Q bands between 520 and 655 nm. The
PL spectra of the three generations of tetra-stilbeneporphyrins
show two peaks at 661 and 727 nm corresponding to the Q(0,0)
and the Q(0,1) porphyrin core transitions.14 The PL is only
slightly red-shifted when compared to the tetra-phenylpor-
phyrin dendrimers which had peaks at 655 and 720 nm. This is
consistent with the electrochemical measurements that showed
the HOMO–LUMO energy gaps of the two families of
materials being essentially the same (Table 1). However,
there is a significant difference in the PL spectra between the
Fig. 6 Film PL spectra for 15, 16, and 17.
238
J. Mater. Chem., 2003, 13, 235–242

View Article Online
Table 2 Dendrimer PLQYs
l
ex/nm
Dendrimer (solution) (%)a
PLQYsol PLQYfilm Film
(%)
Film abs.
thickness/nm @ lex
1
2
3
15
16
17
419
429
429
427
427
428
12
12
10
15
18
17
2.2b
1.7b
1.8b
0.81c
0.98c
1.9c
130
110
130
130
150
160
0.4
0.9
0.4
0.65
1.0
0.5
aAll solution measurements were taken in tetrahydrofuran (HPLC
grade). bFilm spun from tetrahydrofuran. cFilm spun from chloro-
form. Films excited at 442 nm.
were found to be similar for all generations. That is, the
attachment of the dendrons via stilbene units enhances the
PLQY of the porphyrins. However, in going from solution to
the solid state we observe the most interesting PLQY results.
It might be expected that for the zeroeth generation dendrimers
in each family we might see a decrease in the film PLQY due to
p-stacking but observe a significant increase in going to higher
generations as the core becomes more protected by the
dendrons.18 However, this is not the case for either family.
For the tetra-phenylporphyrin dendrimer family there is a five-
fold decrease in going from solution to the solid state with all
generations essentially having the same PLQY. For the tetra-
stilbeneporphyrins the decrease in PLQY is approximately
eighteen-fold for the zeroeth and first generation and ten-fold
for the second generation. The decrease in PLQY of almost an
order of magnitude seen for both families of dendrimers in
going from solution to the solid state indicates that there is
enhanced non-radiative decay in the solid state. Enhanced non-
radiative decay can arise from the migration of excited states to
quenching sites including excimers and aggregates. This larger
relative decrease for the tetra-stilbeneporphyrins is consistent
with them having a more open structure allowing increased
porphyrin core–porphyrin core interactions.
A final aspect of this paper is an understanding of the
improvement in OLED performance in going from the first to
second generation in the tetra-phenylporphyrin series. The
electroluminescence external quantum efficiency for the first
generation was 0.02% and for the second 0.04% in a device
configuration of ITO/porphyrin/Ca.12 Although these EL
efficiencies are not very high they actually do represent
almost optimised efficiency for these fluorescent materials.
Given that OLEDs based on molecular fluorescent materials
can only have a maximum internal efficiency of 25% of the
PLQY the maximum internal EL efficiency of [G-1]PP and
[G-2]PP based devices can only be of the order 0.45%. How-
ever, the outcoupling of light in these simple devices is usually
around a fifth due to the refractive indices of the materials. This
means that the maximum external efficiency of devices based
on these two porphyrins would be around 0.09%. Therefore,
given that the PLQY of the two dendrimers is the same we
believe that the difference in efficiency is likely to be due to the
difference in charge mobility, which is governed by generation.
Fig. 7 Film PL spectra for 1, 2, and 3.
being the most sterically hindered in the series. Another
difference between the solution PL and film PL spectra of the
tetra-stilbeneporphyrins is that the Q(0,1) transition becomes
more pronounced in the solid state. In addition, the ratio of the
Q(0,1) to Q(0,0) transitions was found to decrease on going
from the zeroeth to first and second generation. We believe one
factor causing the change in the ratio of the Q(0,0) to Q(0,1)
transitions is that the porphyrin core adopts a different
conformation in the solid state with the conformation being
partially controlled by the dendrons and solid state inter-
molecular interactions.
For the more sterically encumbered tetra-phenylporphyrin
the changes in the PL spectra are less (Fig. 7). Although there is
a red shift in the PL spectra on moving from solution to the
solid state it is less than that observed for the tetra-
stilbeneporphyrin series with the Q(0,0) and Q(0,1) peaks
moving to the red by only #14 nm and #8 nm respectively.
The PL spectra are only slightly broader in the solid state in
marked contrast to the tetra-stilbeneporphyrin family. How-
ever, the largest difference between the solution and solid state
PL spectra is in the weighting of the Q(0,0) and Q(0,1)
transitions. In solution we observed that the ratio of these
transitions increased with increasing generation. However, in
the solid state this trend is reversed with the Q(0,1) being the
main transition for the zeroeth generation dendrimer with it
decreasing with increasing generation. For the second genera-
tion the ratio of the Q(0,0) to Q(0,1) transitions in the solid
state was seen to begin to reach parity with the solution
spectrum. These PL spectra again clearly demonstrate that the
porphyrin emission is very sensitive to the porphyrin shape,
which can be governed by the local environment which in turn
can be controlled by the method of dendron attachment and
generation. It is interesting to note that even the zeroeth
generation of the tetra-phenylporphyrin series has less of a red
tail. We believe this is due to the bulky tert-butyl groups being
attached close enough to the porphyrin core to decrease the
level of porphyrin p-stacking. This view is consistent with the
physical properties of [G-0]PP which is far more soluble than
the simple meso-tetraphenylporphyrin which lacks the bulky
groups. Therefore, the branching of the stilbene dendrons
closer to the core gives better control of the core–core
interactions.
Conclusion
We have found in a new family of porphyrin cored conjugated
dendrimers with stilbene based dendrons that the proximity of
the first branching point is important in controlling the core–
core interactions. We have shown that when the dendrons are
attached via a stilbene unit the dendrimer is more open and the
cores more susceptible to intermolecular interactions than
when the dendrons are connected via a phenyl ring. Whilst the
dendrimers with the stilbene links between the dendrons and
the porphyrin core show superior electrochemical and photo-
luminescence properties in solution, in the solid state the
The different abilities of the phenyl and stilbenyl attachment
to the porphyrin ring to control the core–core interactions was
confirmed by PL quantum yield (PLQY) measurements. We
carried out the measurements in both solution and the solid
state and the results are summarised in Table 2. For the tetra-
phenylporphyrin dendrimers the solution PLQYs were found
to be independent of generation and around 10–12%. With the
attachment of the dendrons via the stilbene rather than a
phenyl unit the PLQY was seen to increase by around 50% and
J. Mater. Chem., 2003, 13, 235–242
239

View Article Online
phenyl linked dendrimers were more luminescent. Therefore,
when designing conjugated dendrimers it is necessary not only
to choose the correct type of dendron for the core but also the
correct linking unit between dendron and core.
(4-Cyanobenzyl)phosphonic acid dimethyl ester 5
A suspension of 4-(bromomethyl)benzonitrile (27.9 g, 0.142
mmol) in trimethyl phosphite (84.0 cm3, 0.711 mmol) was
stirred at 100 uC for 66 h. Unreacted trimethyl phosphite was
removed by distillation under reduced pressure. The residue
was purified by column chromatography over silica using a
dichloromethane–ethyl acetate mixture (1 : 0 to 0 : 1) as
eluent, to give a white solid of 5 (28.9 g, 90%). Mp 80-81 uC
(Found: C, 53.3; H, 5.4; N, 6.25. C10H12NO3P requires C, 53.3;
H, 5.4; N, 6.2%); nmax(KBr)/cm21 2225 (CMN), 1254 (PLO) and
1054 and 1035 (P–O–C); lmax(MeOH)/nm 233 [log e/dm3
mol21 cm21 (4.27)], 239sh (4.21), 263sh (2.75), 271 (2.85) and
281 (2.80); dH(400 MHz, CDCl3) 3.19 (2 H, d, J 22, CH2P), 3.68
(6 H, d, J 11, OCH3) and 7.39 and 7.59 (4 H, AA’BB’); dC(100.6
MHz, CDCl3) 33.1 (d, J 138, CH2P), 53.0 (d, J 7, CH3), 111.0
(d, J 4), 118.6 (d, J 2), 130.5 (d, J 6), 132.3 (d, J 3) and 137.1 (d,
J 10); m/z (APCI1) 226.1 (MH1, 100%).
Experimental
Measurements
NMR spectra were recorded on a Bruker DPX 400 MHz or an
AMX 500 MHz spectrometer: sp ~ surface phenyl; cp ~ core
phenyl; cv ~ core vinyl; bp ~ branch phenyl. All J values are
in Hertz. IR spectra were recorded on a Perkin-Elmer 1000
infrared spectrometer. UV-visible spectra were recorded on a
Perkin-Elmer UV-visible Lambda 14P spectrometer and were
recorded as a solution in spectroscopic grade chloroform,
dichloromethane, or methanol. Mass spectra were recorded
on a Hewlett-Packard 1050 atmospheric pressure chemical
ionisation mass spectrometer (APCI) (1ve mode) or a
Micromass TofSpec 2E for matrix-assisted laser desorption/
ionisation–time-of-flight (MALDI-TOF) from dithranol (1,8,9-
trihydroxyanthracene) in reflectron mode. Melting points
were recorded on a Gallenkamp melting point apparatus
and are uncorrected. Microanalyses were carried out in the
Inorganic Chemistry Laboratory, Oxford, UK. Gel permeation
chromatography was carried out using PLgel mm Mixed-A
columns (600 mm 1 300 mm lengths, 7.5 mm diameter) from
Polymer Laboratories calibrated with polystyrene narrow
standards (Mp ~ 1300 to 11.2 6 106) in tetrahydrofuran
with toluene as flow marker. The tetrahydrofuran was degassed
with helium and pumped at a rate of 1 mL min21 at 30.0 uC.
Light petroleum refers to the fraction of boiling point 60–
80 uC, unless otherwise stated, and ether refers to diethyl ether.
When solvent mixtures are used for chromatography over silica
the proportions are given by volume.
(E)-4-[2-(3,5-Di-tert-butylphenyl)vinyl]benzonitrile:
[G-0]StCN 9
Potassium tert-butoxide (5.92 g, 52.7 mmol) was added to a
stirred solution of 3,5-di-tert-butylbenzaldehyde 6 (9.599 g,
43.96 mmol) and 5 (10.0 g, 52.7 mmol) in tetrahydrofuran
(100 cm3). The reaction was initially exothermic and the
solution was stirred at room temperature for 19 h. Water
(250 cm3) was added and the aqueous layer was separated and
extracted with ether (100 cm3). The organic layers were
combined and washed with brine (100 cm3), dried over
anhydrous magnesium sulfate, filtered and the solvent
removed. The residue was purified by column chromatography
over silica using a dichloromethane–light petroleum mixture
(2 : 3) as eluent to give 9 as a white solid (12.5 g, 89%), mp
145 uC (lit. 146.5 uC23).
Electrochemistry was performed using an EG&G Princeton
Applied Research potentiostat/galvanostat model 263A. All
measurements were made at room temperature on samples
dissolved in freshly distilled dichloromethane with 0.1 M tetra-
n-butylammonium hexafluorophosphate as the electrolyte.
Dichloromethane was purified by the literature procedure
and then freshly distilled from calcium hydride.19 The electro-
lyte was purified by recrystallisation from ethanol. The solu-
tions were deoxygenated with argon. A glassy carbon working
electrode, platinum wire counter electrode, and Ag/3 M NaCl/
AgCl(sat) reference electrode were used. The ferrocenium/
ferrocene couple was used as standard,20 and the ferrocene was
purified by sublimation. Figs. 2–4 show a small chemically
reversible process near 0 V due to residual ferrocene. In all
cases several scans were carried out to confirm the chemical
reversibility of the redox processes.
Solution photoluminescence quantum yields (PLQYs)
were measured by a relative method using rhodamine 101 in
ethanol at room temperature as a standard.21 The dendrimers
were dissolved in tetrahydrofuran (THF) and freeze–thaw
degassed. Photoluminescence spectra were recorded in a JY
Horiba Fluoromax 2 fluorimeter, with the dendrimer solutions
excited at the Sore´t maxima. The absorbances of the standard
and samples were similar and small (j0.1). The accuracy of
these measurements is estimated to be ¡15% of the stated
value.
(E)-4-[2-(3,5-Di-tert-butylphenyl)vinyl]benzaldehyde:
[G-0]StCHO 12
A solution of 9 (10.0 g, 31.5 mmol) in toluene (50 cm3) and
dichloromethane (50 cm3) was cooled to 278 uC. Diiso-
butylaluminium hydride (1.0 M in hexanes, 45 cm3, 45 mmol)
was added and the solution stirred at 278uC for 30 min
then at room temperature for 19 h. Aqueous hydrochloric
acid (3 M, 100 cm3) was added and the mixture stirred for
3 min. The solution was extracted with ether (250 cm3) and
the extract was washed with aqueous hydrochloric acid
(3 M, 100 cm3), water (100 cm3), aqueous sodium bicarbonate
(5% w/v, 100 cm3), brine (100 cm3), dried over anhydrous
magnesium sulfate, filtered and the solvent removed. The
residue was dissolved in toluene (100 cm3) and heated with
catalytic iodine (400 mg, 1.58 mmol) under argon at 100 uC
for 24 h and allowed to cool. The solution was washed
with saturated sodium metabisulfite (2 6 100 cm3) and brine
(100 cm3), dried over anhydrous magnesium sulfate, filtered
and the solvent removed. The residue was purified by passing
through a plug of silica with a dichloromethane–light petro-
leum mixture (1 : 1) as eluent to give a pale yellow solid
of 12 (4.95 g, 95%). Mp 93–94 uC (Found: C, 86.2; H, 8.8.
C23H28O requires C, 86.2; H, 8.8%); nmax(KBr)/cm21 1695
(CLO), 1599 (CLC) and 969 (CLC–H trans); lmax(CHCl3)/nm
342 [log e/dm3 mol21 cm21 (4.53)]; dH(400 MHz, CDCl3) 1.39
(18 H, s, t-butyl H), 7.16 and 7.33 (2 H, d, J 16, vinylic H),
7.43 (3 H, m, sp H), 7.69 and 7.88 (4 H, AA’BB’, cp H) and
10.01(1 H, s, CHO); dc(100.6 MHz, CDCl3) 31.4, 34.9, 121.2,
123.0, 126.6, 126. 8, 130.2, 133.4, 135.1, 135.7, 143.8, 151.2
and 191.6; m/z (APCI1) 321.3 (MH1, 100%).
Films were spin-coated from a THF or chloroform solution
with a dendrimer concentration of 10 mg ml21 at 1000 rpm for
1 min to give a thickness of about 150 nm. Their PLQYs were
measured using an integrating sphere in accordance with
Greenham et al.22 using a Helium Cadmium laser (Kimmon) as
the excitation source. The excitation power was 0.05 mW at
442 nm and the sphere was purged with nitrogen.
240
J. Mater. Chem., 2003, 13, 235–242

View Article Online
[G-0]StP 15
7.68 (1 H, s, G1-bp H), 7.72 and 7.91 (4 H, AA’BB’, cp H) and
10.03 (1 H, s, CHO); m/z (APCI1) 637.5 (MH1, 100%).
A solution of 12 (2.00 g, 6.24 mmol), pyrrole (0.43 g, 6.2 mmol)
and trifluoroacetic acid (0.480 ml, 6.24 mmol) in dichloro-
methane (460 cm3) was stirred in the dark under argon for
7 days. 2,3-Dichloro-5,6-dicyano-1,4-benzoquinone (1.42 g,
6.24 mmol) was added and the mixture stirred for 10 min. The
solution was neutralised by the addition of an excess of sodium
hydrogen carbonate then filtered through a plug of silica using
a dichloromethane–light petroleum mixture (1 : 2 to 1 : 0) as
eluent. The main fraction was collected, the solvent removed,
[G-1]StP 16
A solution of 13 (1.00 g, 1.57 mmol), pyrrole (157 ml,
2.26 mmol) and trifluoroacetic acid (121 ml, 1.57 mmol) in
dichloromethane (114 cm3) was stirred at room temperature
under argon for 9 days. 2,3-Dichloro-5,6-dicyano-1,4-benzo-
quinone (357 mg, 1.57 mmol) was added and the mixture
stirred for 30 min. The solution was neutralised by the addition
of an excess of sodium hydrogen carbonate then filtered
through a short plug of silica using dichloromethane as eluent.
The main fraction was collected and the solvent removed. The
residue was purified by column chromatography over silica
using a dichloromethane–light petroleum mixture (1 : 3 to
2 : 3) as eluent. The main fraction was collected, the solvent
removed and the residue recrystallised from a dichloro-
methane–methanol mixture to give a red–purple solid of 16
(153 mg, 15%). Mp w 279 uC (decomp.) (Found: C, 89.1; H,
8.2; N, 2.2. C204H230N4 requires C, 89.5; H, 8.5; N, 2.1%);
and the residue recrystallised from
a dichloromethane–
methanol mixture to give 15 (563 mg, 25%) as a red–purple
solid. Mp w 280 uC (decomp.) (Found: C, 87.45; H, 7.9; N,
3.85. C108H118N4 requires C, 88.1; H, 8.1; N, 3.8%); nmax(KBr)/
cm21 3320 (N–H) and 964 (CLC–H trans); lmax(CH2Cl2)/nm
305 [log e/dm3 mol21 cm21 (5.01)], 429 (5.85), 521 (4.31),
559 (4.28), 594 (3.84) ad 652 (3.93); dH(500 MHz, CDCl3)
22.68 (2 H, s, NH), 1.45 (72 H, s, t-butyl), 7.45 and 7.51 (8 H,
d, J 16, cv H), 7.46 (4 H, dd, J 1.5 and 1.5, sp H), 7.55 (8 H, d, J
1.5, sp H), 7.95 and 8.25 (16 H, AA’BB’, cp H) and 8.97 (8 H, s,
b-pyrrolic H); m/z (MALDI) calc. 1471.9; found 1471.9 (M1,
100%).
n
max(KBr)/cm21 3317 (N–H), 1594 (CLC) and 961 (CLC–H
trans); lmax(CH2Cl2)/nm 318 [log e/dm3 mol21 cm21 (4.60)],
335sh (4.45), 429 (4.85), 444 (2.88), 521 (3.41), 559 (3.38), 595
(2.94) and 652 (2.99); dH(500 MHz, CDCl3) 22.65 (2 H, s,
NH), 1.41 (144 H, s, t-butyl), 7.25 and 7.36 (16 H, d, J 16, G1-
vinyl H), 7.41 (8 H, dd, J 1.5 and 1.5, sp H), 7.48 (16 H, d, J 1.5,
sp H), 7.51 and 7.60 (4 H, d, J 16, cv H), 7.72 (4 H, s, G1-bp H),
7.76 (8 H, s, G1-bp H), 7.99 and 8.29 (16 H, AA’BB’, cp H) and
8.99 (8 H, s, b-pyrrolic H); m/z (MALDI) 2738.4 (M1, 100%).
[G-1]StCN 10
Potassium tert-butoxide (1.01 g, 9.00 mmol) was added to
a stirred solution of 7 (4.00 g, 7.38 mmol) and 5 (2.02 g,
8.97 mmol) in tetrahydrofuran (80 cm3), and the mixture
stirred at room temperature for 16.5 h. Water (100 cm3) was
added and the resultant slurry extracted with dichloromethane
(1 6 200 cm3, 16 100 cm3). The organic layers were combined
and washed with brine (100 cm3), dried over anhydrous
magnesium sulfate, filtered and the solvent removed. The
residue was purified by column chromatography over silica
using a dichloromethane–light petroleum mixture (2 : 3) as
eluent to give a white solid of 10 (4.36 g, 92%). Mp 254–255 uC
(Found: C, 89.0; H, 8.8; N, 2.2. C47H55N requires C, 89.05; H,
8.7; N, 2.2%); nmax(KBr)/cm21 2223 (CMN) and 959 (CLC–H
trans); lmax(CH2Cl2)/nm 322 [log e/dm3 mol21 cm21 (4.93)],
332sh (4.89); dH(400 MHz, CDCl3) 1.39 (36 H, s, t-butyl H),
7.17 and 7.28 (4 H, d, J 16, G1-vinyl H), 7.22 and 7.29 (2 H, d, J
16, cv H), 7.40 (2 H, dd, J 1.5 and 1.5, sp H), 7.43 (4 H, d, J 1.5,
sp H), 7.60 (2 H, d, J 1, G1-bp H), 7.64 (2 H, 1/2AA’BB’, cp H)
and 7.67–7.69 (3 H, m, cp H and G1-bp H); m/z (APCI1) 634.5
(MH1, 15%).
[G-2]StCN 11
Potassium tert-butoxide (0.50 g, 4.45 mmol) was added to
a stirred suspension of 8 (4.00 g, 3.42 mmol) and 5 (1.00 g,
4.45 mmol) in tetrahydrofuran (100 cm3) and the reaction
mixture was stirred at room temperature for 67 h under argon.
Water (100 cm3) was added and the mixture extracted with
dichloromethane (1 6 200 cm3, 1 6 100 cm3). The organic
layers were combined, washed with water (100 cm3) and brine
(100 cm3), dried over anhydrous magnesium sulfate, filtered
and the solvent removed. The residue was purified by column
chromatography over silica using a dichloromethane–light
petroleum mixture (1 : 1) as eluent followed by recrystallisa-
tion from a dichloromethane–light petroleum mixture to give
a white solid of 11 (4.11 g, 95%). Mp w 290 uC (decomp.)
(Found: C, 90.1; H, 8.8; N, 1.1. C95H111N requires C, 90.1; H,
8.8; N, 1.1%); nmax(KBr)/cm21 2226 (CLN), 1595 (CLC) and
960 (CLC–H trans); lmax(CH2Cl2)/nm 323 ([log(e/dm3 mol21
cm21) 5.27] and 332sh (5.23); dH(400 MHz, CDCl3) 1.41 (72 H,
s, t-butyl), 7.20 and 7.31 (8 H, d, J 16, G2-vinyl H), 7.24 and
7.31 (2 H, m, cv H), 7.32 (4 H, s, G1-vinyl H), 7.41 (4 H, dd, J
1.5 and 1.5, sp H), 7.46 (8 H, d, J 1.5, sp H), 7.66–7.72 (10 H, m,
G1-bp H, G2-bp H, and cp H), 7.70 (2 H, 1/2AA’BB’, cp H)
and 7.76 (1 H, br m, G1-bp H); m/z (MALDI) 1266.9 (M1,
100%).
[G-1]StCHO 13
A solution of 10 (4.00 g, 6.31 mmol) in dichloromethane
(25 cm3) and toluene (25 cm3) was cooled to 278 uC under
argon. Diisobutylaluminium hydride (1.0 M in hexanes,
8.8 cm3, 8.8 mmol) was added and the mixture was stirred at
278 uC for 30 min. The reaction mixture was then allowed to
warm to room temperature and stirred for 19.5 h. Aqueous
hydrochloric acid (3 M, 80 cm3) was added carefully and the
mixture stirred for 30 min, then extracted with dichloro-
methane (2 6 80 cm3). The organic layers were combined,
washed with aqueous hydrochloric acid (3 M, 100 cm3) and
brine (100 cm3), dried over anhydrous magnesium sulfate,
filtered and the solvent removed. The residue was purified by
column chromatography over silica using a dichloromethane–
light petroleum mixture (1 : 1) as eluent to give a white solid of
13 (3.23 g, 80%). Mp 235 uC (Found: C, 88.6; H, 8.8. C47H56O
requires C, 88.6; H, 8.9%); nmax(KBr)/cm21 1688 (CLO) and 954
(CLC–H trans); lmax(CH2Cl2)/nm 323 [log e/dm3 mol21 cm21
(5.32)], 331sh (5.29) and 358sh (5.03); dH(400 MHz, CDCl3)
1.40 (36 H, s, t-butyl H), 7.18 and 7.29 (4 H, d, J 16, G1-
vinyl H), 7.28 and 7.35 (2 H, d, J 16, cv H), 7.41 (2 H, dd, J 1.5
and 1.5, sp H), 7.44 (4 H, d, J 1.5, sp H), 7.63 (2 H, s, G1-bp H),
[G-2]StCHO 14
A suspension of 11 (3.17 g, 2.50 mmol) in dichloromethane
(25 cm3) and toluene (23 cm3) was stirred at 278 uC under
argon. Diisobutylaluminium hydride (1.5 M in toluene,
2.20 cm3, 3.25 mmol) was added and the mixture stirred at
278 uC for 10 min and allowed to warm to room temperature,
with stirring, over 20.5 h. Aqueous hydrochloric acid (3 M,
20 cm3) was carefully added and the mixture stirred for 20 min,
then extracted with dichloromethane (2 6 100 cm3). The
organic layers were combined, washed with water (100 cm3)
and brine (100 cm3), dried over anhydrous magnesium sulfate,
filtered and the solvent removed. The residue was purified by
column chromatography over silica with dichloromethane as
J. Mater. Chem., 2003, 13, 235–242
241

View Article Online
eluent to give a white solid of 14 (2.31 g, 73%). Mp w 250 uC
(Found: C, 89.8; H, 8.9. C95H112O requires C, 90.0; H, 8.9%);
n
References
max(KBr)/cm21 1697 (CLO), 1597 (CLC) and 960 (CLC–H
1
C. S. Cameron and C. B. Gorman, Adv. Funct. Mater., 2002, 12,
17.
trans); lmax(CH2Cl2)/nm 323 [log e/dm3 mol21 cm21 (5.36)] and
333sh (5.32); dH(400 MHz, CDCl3) 1.41 (72 H, s, t-butyl), 7.21
and 7.32 (8 H, d, J 16, G2-vinyl H), 7.30 and 7.37 (2 H, d, J 16,
cv H), 7.33 (4 H, s, G1 vinyl H), 7.38 (4 H, dd, J 2 and 2, sp H),
7.46 (8 H, d, J 2, sp H), 7.68 (8 H, m, G1-bp H and G2-bp H),
7.75 and 7.94 (4 H, AA’BB’, cp H), 7.76 (1 H, br m, G1-bp H)
and 10.05 (1 H, s, CHO); m/z (MALDI) 1269.9 (M1, 100%).
2
M. S. Matos, J. Hofkens, W. Verheijen, F. C. De Schryver,
S. Hecht, K. W. Pollak, J. M. J. Fre´chet, B. Forier and W. Dehaen,
Macromolecules, 2000, 33, 2967.
3
4
D.-L. Jiang and T. Aida, J. Am. Chem. Soc., 1998, 120, 10895.
P. Weyermann, J.-P. Gisselbrecht, C. Boudon, F. Diederich and
M. Gross, Angew. Chem., Int. Ed., 1999, 38, 3215.
K. W. Pollak, J. W .Leon, J. M. J. Fre´chet, M. Maskus and
H. D. Abrun˜a, Chem. Mater., 1998, 10, 30.
J. N. G. Pillow, M. Halim, J. M. Lupton, P. L. Burn and
I. D. W. Samuel, Macromolecules, 1999, 32, 5985.
M. Kimura, T. Shiba, T. Muto, K. Hanabusa and H. Shirai,
Macromolecules, 1999, 32, 8237.
M. Kimura, T. Shiba, M. Yamazaki, K. Hanabusa, H. Shirai and
N. Kobayashi, J. Am. Chem. Soc., 2001, 123, 5636.
R. C. Kwong, S. Sibley, T. Dubovoy, M. Baldo, S. R. Forrest and
M. E. Thompson, Chem. Mater., 1999, 11, 3709.
5
6
7
8
9
[G-2]StP 17
A mixture of 14 (1.000 g, 0.787 mmol), pyrrole (55.0 ml,
0.787 mmol), and trifluoroacetic acid (90.0 mg, 0.787 mmol) in
dichloromethane (60.0 cm3) was stirred at room temperature
under argon for 12 days. 2,3-Dichloro-5,6-dicyano-1,4-benzo-
quinone (393 mg, 1.96 mmol) was added and the mixture
stirred for 30 min. The solution was neutralised by the addition
of an excess of sodium hydrogen carbonate then filtered
through a short plug of silica, eluting with dichloromethane.
The solvent was removed and the residue was purified by
column chromatography over silica using a dichloromethane–
light petroleum mixture (1 : 4 to 2 : 3) as eluent. The main frac-
tion was collected, the solvent removed and the residue recrys-
tallised from a dichloromethane–methanol mixture to give 17
(232 mg, 22%) as a red–purple solid. Mp 286 uC (decomp.)
(Found: C, 89.7; H, 8.7; N, 1.3. C396H454N4 requires C, 90.25;
H, 8.7; N, 1.1%); nmax(KBr)/cm21 1594 (CLC) and 960 (CLC–H
10 V. Cleave, G. Yahioglu, P. Le Barny, R. H. Friend and N. Tessler,
Adv. Mater., 1999, 11, 285.
11 D. F. O’Brien, C. Giebler, R. B. Fletcher, A. J. Cadby, L. C. Palilis,
D. G. Lidzey, P. A. Lane, D. D. C. Bradley and W. Blau, Synth.
Met., 2001, 116, 379.
12 P. L. Burn, M. Halim, J. N. G. Pillow and I. D. W. Samuel, SPIE
Symposium on organic light-emitting materials and devices III,
Denver, SPIE—The International Society for Optical Engineering,
Washington, 1999, vol. 3797, p. 66.
13 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.
14 M. Goutermann and G-E. Khalil, J. Mol. Spectrosc., 1974, 53, 88.
15 K. Kalyanasundaram, Photochemistry of Polypyridine and Por-
phyrin Complexes, Academic Press Limited, London, 1992.
16 P. J. Spellane, M. Gouterman, A. Antipas, S. Kim and Y. C. Lui,
Inorg. Chem., 1980, 19, 386.
17 P. Charlesworth, T. G. Truscott, D. Kessel, C. J. Medforth and
K. M. Smith, J. Chem. Soc., Faraday Trans., 1994, 90, 1073.
18 L.-O. Pa˚lsson, R. Beavington, M. J. Frampton, S. W. Magennis,
J. P. J. Markham, J. N. G. Pillow, P. L. Burn and I. D. W. Samuel,
Macromolecules, 2002, 35, 7891.
trans); lmax(CH2Cl2)/nm 314sh (log (e/dm3 mol21 cm21
)
5.92), 322 (5.94), 331sh (5.86), 430 (5.85), 488sh (3.91), 521
(4.38), 559 (4.35), 595 (3.88) and 651 (3.94); dH(500 MHz,
CDCl3) 22.62 (2 H, s, NH), 1.40 (288 H, s, t-butyl), 7.22 and
7.33 (32 H, d, J 16, G2 vinyl H), 7.39–7.40 (32 H, G1 vinyl H
and sp H), 7.46 (32 H, d, J 1.5, sp H), 7.56 and 7.63 (8 H, d, J
16, cv H), 7.68 (8 H, s, G2-bp H), 7.71 (16 H, s, G2-bp H), 7.80
(4 H, s, G1-bp H), 7.82 (8 H, s, G1-bp H), 8.03 and 8.32 (16 H,
AA’BB’, cp H) and 9.02 (8 H, s, b-pyrrolic H); m/z (MALDI)
5272.6 (M1, 100%).
19 D. D. Perrin and W. L. F. Amarego, Purification of Laboratory
Compounds, Pergamon Press, Oxford, 1988.
20 G. Gritzner and J. Kuta, Electrochim. Acta, 1984, 29, 869.
21 T. Karstens and K. Kobs, J. Phys. Chem., 1980, 84, 1871.
22 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.
Acknowledgements
We are grateful to CDT Oxford Ltd, and the EPSRC and
SHEFC for financial support. IDWS is a Royal Society
University Research Fellow.
23 R. Schenk, H. Gregorius, J. Heinze and K. Mu¨llen, J. Am. Chem.
Soc., 1991, 113, 2634.
242
J. Mater. Chem., 2003, 13, 235–242