High-efficiency multilayer polymeric blue light-emitting diodes using boronate esters as cross-linking linkages

Article abstract of DOI:10.1039/b514542b

A novel approach for the production of cross-linked and robust hole transport layers for use in multilayer polymeric light-emitting diodes (PLEDs) has been developed. Two alternating triphenylamine-fluorene copolymers (TPAFn, n = 2, 3) with hydroxyl group

Full text of DOI:10.1039/b514542b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
High-efficiency multilayer polymeric blue light-emitting diodes using
boronate esters as cross-linking linkages
a
Jianping Lu,* Yinan Jin, Jianfu Ding, Ye Tao* and Michael Day
b
b
a
b
Received 13th October 2005, Accepted 9th November 2005
First published as an Advance Article on the web 5th December 2005
DOI: 10.1039/b514542b
A novel approach for the production of cross-linked and robust hole transport layers for use in
multilayer polymeric light-emitting diodes (PLEDs) has been developed. Two alternating
triphenylamine–fluorene copolymers (TPAFn, n = 2, 3) with hydroxyl groups on the side chains
and tris(4-dihydroxyboranylphenyl)amine (TBPA) as a cross-linker have been designed and
synthesized. The mixture of TPAFn and TBPA when subjected to mild reaction conditions (2 h
baking at 130 uC under vacuum) undergoes cross-linking reactions to produce cross-linked films
(
X-TPAFn), due to the formation of boronate ester linkages. The resulting X-TPAFn films have
excellent solvent resistance to common organic solvents, such as THF and CHCl , thereby
3
facilitating the fabrication of multilayer PLEDs. Studies have shown that the cross-linking
reaction had no detrimental effects on the photophysical properties of the resulting X-TPAFn
films. In addition, the cross-linked X-TPAFn networks have been shown to have much better
electron-blocking properties than the widely used PEDOT–PSS. Using X-TPAFn as a hole
transport layer, we have investigated the electroluminescent (EL) properties of alternating
fluorene–oxadiazole copolymers OxFn (n = 2, 3) and the poly(9,9-dioctylfluorene) (POF)
homopolymer. The experimental data indicated that an increase in the oxadiazole content of the
polymer lowered the LUMO energy level while decreasing the photoluminescence (PL) quantum
yield. Consequently the best device performance was obtained with OxF3 which was found to have
2
2
21
a maximum luminance of 2010 cd m at 11.5 V and a maximum luminous efficiency of 1.0 cd A
22
at 820 cd m when used with X-TPAF2 as the hole transport layer and calcium as the cathode.
Introduction
applied to very limited polymer pairs. The most convenient
way therefore to construct polymeric multilayer devices is to
cross-link each layer after spin-coating, making the resulting
film insoluble in an organic solvent. In addition, the cross-
linking reaction increases the glass transition temperatures
Organic light-emitting diodes (OLEDs) have been the subject
of intense research over the past two decades, due to their
promising potential for applications in flat-panel displays
1
and general solid state lighting. It is generally accepted that
(
Tg) of the polymers and improves the thermal stability of the
4
device performance can be maximized by the use of a
multilayer structure in which each layer is optimized for
electron transport, hole transport, and luminescence, respec-
resulting device.
Over the past decade, fluorene-based conjugated polymers
(
PFs) have emerged as a promising class of blue light-emitting
2
tively. Through the use of such a multilayer device, the charge
materials because of their high photoluminescence (PL) and
electroluminescence (EL) quantum efficiencies, thermal stabi-
lity, good solubility, and facile functionalization at the C-9
injection and transport can be better balanced and the
formation of excitons can be confined within the emissive
layer. Although the fabrication of multilayer small-molecule
OLEDs is relatively easy via vacuum deposition, the casting
of polymer multilayers via solution processing is more
complicated. This arises because the lower layers may be
partially or even totally dissolved by the solvents used for
casting succeeding layers. Only in rare cases, do two electro-
active polymers have significant solubility differences in
organic solvents to allow the construction of polymeric
5
position of fluorene. However, the widespread application
of polyfluorenes in PLEDs has been seriously hampered by
the troublesome formation of a tailed emission band at long
wavelengths (>500 nm) during device operation, which
leads to both color instability and reduced efficiency. Some
researchers have attributed this instability to interchain
aggregation and/or excimer formation. Consequently they
have introduced bulky groups at the C-9 position of the
fluorene or, alternatively, attached cross-linkable moieties such
as styryl groups to PFs in an attempt to suppress the excimer
3
double-layer devices. However, this approach can only be
6
a
emission and improve the thermal stability of PL spectra.
Institute for Microstructural Sciences (IMS), National Research
Council of Canada (NRC), 1200 Montreal Road, Ottawa, ON K1A 0R6,
Canada. E-mail: Jianping.Lu@nrc-cnrc.gc.ca.;
Ye.Tao@nrc-cnrc.gc.ca.; Fax: (613) 990-0202; Tel: (613) 990-1651
Institute for Chemical Process and Environmental Technology
(ICPET), National Research Council of Canada (NRC), 1200
Montreal Road, Ottawa, ON K1A 0R6, Canada
However, more recent studies suggest that keto defects in the
polymer backbone, originating from photo- and/or electro-
oxidative degradation of 9,9-dialkylated PFs, are mainly
b
7
responsible for this strong low-energy emission. Recently, it
8
has been demonstrated by Heeger et al. and our group that
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 593–601 | 593

View Article Online
the insertion of a thin layer of small molecular electron
transporting materials between the emissive PF and the
cathode could result in improved blue emission stability.
However, the fabrication of such hybridized devices containing
small molecules and polymers compromises the unique merit
of PLEDs, i.e., the use of low cost solution processes. In a
previous study we reported that the formation of cross-linked
networks of oligofluorenes via the thermal dehydration of
boronic acid groups can also stabilize the blue emission while
permeation chromatography (GPC) on a Waters Model 515
3
4
5
HPLC equipped with m-Styragel columns (10 , 10 , and 10 )
using THF as eluant and calibrated with polystyrene
standards. The differential scanning calorimetry (DSC) analy-
2
1
sis was performed under a nitrogen atmosphere (50 mL min
)
2
1
on a TA Instruments DSC 2920 at heating rates of 10 uC min
.
UV–Vis absorption spectra were recorded on a Varian Cary 50
spectrophotometer, while fluorescence measurements were
carried out on a Fluorolog 3 spectrometer with the absorption
maxima of the materials as the excitation wavelengths. Cyclic
voltammetry (CV) measurements were conducted in a 0.1 M
9
allowing multilayer configurations to be achieved. However,
due to the poor electrochemical and photophysical properties
of the materials used, device performance was disappointing.
In this paper, we report on the fabrication of high-efficiency
polymeric multilayer blue light-emitting diodes using boronate
esters as cross-linking linkages. This has involved the design
and synthesis of triphenylamine–fluorene alternating copoly-
mers TPAFn (n = 2, 3), containing hydroxyl groups on the side
chains, as hole transporting materials, and tris(4-dihydroxy-
boranylphenyl)amine (TBPA) as the cross-linker. Heating this
mixture of TPAFn and TBPA at 130 uC under vacuum for 2 h
produced a cross-linked film (X-TPAFn) as a result of the
formation of boronate ester linkages. The resulting X-TPAFn
films have excellent solvent resistance to common organic
tetrabutylammonium hexafluorophosphate (Bu
4 6
NPF ) anhy-
drous CH CN solution using a Solartron SI 1287 potentiostat
3
2
1
at a scan rate of 50 mV s at room temperature under argon.
A silver wire (1 mm diameter), a platinum wire (0.5 mm
diameter), and a platinum disk (1 mm diameter) sealed in a soft
glass rod were used as the quasi-reference electrode, counter-
electrode, and working electrode, respectively. The Ag quasi-
reference electrode was calibrated using a ferrocene–ferrocenium
redox couple as an external standard prior to measurements. The
polymer film was drip-coated onto a Pt disc electrode from
CHCl solution, and then dried at 80 uC or cross-linked at 130 uC
3
under vacuum. Twenty successive cycles were recorded for the
oxidation redox process from 0.0 to 1.2 V and then for the
reduction redox process from 0.0 to 22.5 V, respectively.
3
solvents, such as THF and CHCl . As a result, double-layer
devices have been fabricated by casting electron transport
materials on top of these X-TPAFn hole transport layers
(
HTLs). Using an alternating copolymer of fluorene–oxadia-
Materials
zole (OxF3) as the electron transport and emitting material,
2
2
devices with a maximum luminance of 1430 cd m have been
realized at a driving voltage of 18.5 V even with a Mg–Ag (9 : 1)
alloy as the cathode. In addition, these devices have a peak
Tetrahydrofuran (THF) was dried over a Na–K alloy and
distilled prior to use. The acetonitrile used in the electro-
chemical measurements was refluxed in the presence of
calcium hydride under argon and distilled before use.
Other starting materials were purchased from Aldrich
2
1
luminous efficiency of 0.68 cd A at 12.5 V with a luminance
2
of 125 cd m . When a low-work-function calcium cathode
2
was used, the maximum luminance and luminous efficiency
21
and used as received without further purification.
11
2
2
were further improved to 2010 cd m (11.5 V) and 1.0 cd A
8.5 V), respectively. For comparison purposes we have com-
3 4
Tetrakis(triphenylphosphine) palladium(0) (Pd(PPh ) ),
9
tris(4-dihydroxyboranylphenyl)amine (TBPA), 2-bromo-9,9-
(
1
dioctylfluorene, ,
2
pared our data with a PLED device fabricated with the widely
used poly(3,4-ethylenedioxythiophene)–poly(styrene sulfonate)
2-dihydroxyboranyl-9,9-dioctylfluorene
15
and N,N-bis(4-bromophenyl)aniline (4) were pre-
1
3,14
(1),
(
PEDOT–PSS) as a hole transport layer. The results obtained
pared according to literature procedures. The preparation of
alternating fluorene–oxadiazole copolymers OxFn (n = 2, 3)
and poly(9,9-dioctylfluorene) (POF) homopolymer was
indicate that the X-TPAFn networks are superior to the
PEDOT–PSS networks, leading to much greater brightness
and luminous efficiency. This was felt to be due to the
X-TPAFn being able to function as electron-blocking
materials, thereby greatly enhancing the electron–hole recom-
bination efficiency in the emitting layer. Recently, indepen-
dently of our work, Marks and coworkers have also reported
that the use of a cross-linked hole transporting/electron-
blocking network in PLEDs dramatically enhances the
1
4a
reported previously. The procedures for the preparation of
the triphenylamine–fluorene copolymers (TPAFn, n = 2, 3) are
summarized in Scheme 1, and are essentially the same as the
9
,13,14
widely used known synthetic procedures.
are not reported here in detail.
Therefore, they
TPAF2 (10). To a 50 mL one-necked flask were added
1
1
0
performance of PF-based devices.
tricaprylylmethylammonium chloride (Aliquot 336) (y20%
wt based on monomers), N,N-bis(4-(4,4,5,5-tetramethyl-1,3,2-
Experimental
dioxaborolan-2-yl)phenyl)aniline (5) (1 equiv.), dibromodimer
7
(1 equiv.), and toluene. Once all the monomers were
Characterization
dissolved, 2 M Na CO aqueous solution (3.3 equiv.) was
2
3
1
H NMR spectra were obtained in CDCl3 on a 400 MHz
Varian Unity Inova spectrometer. Mass spectrometry mea-
surements and elemental analysis were carried out by the
University of Ottawa Mass Spectrometry Center and
Canadian Microanalytical Service Ltd., respectively. The
molecular weights of the polymers were determined using gel
added. The flask, equipped with a condenser, was then
evacuated and flushed with argon 3 times to remove air.
3 4
Pd(PPh ) (0.5% equiv.) was then added under a nitrogen
atmosphere. The flask was then once again flushed with argon
3 times. The mixture was then heated to reflux and maintained
for 48 hours under argon. The reaction mixture was then
5
94 | J. Mater. Chem., 2006, 16, 593–601
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Scheme 1 Synthetic route for the preparation of TPAFn and the chemical structures of the materials used in this study.
cooled to room temperature and the organic layer was
separated, concentrated, and precipitated into methanol. The
resultant polymer was collected, dried, and further purified by
the following procedure. First it was dissolved in dichloro-
methane and vigorously stirred with 6 N HCl for 1 hour. The
TPAF3 (11). This polymer was prepared and purified using
a procedure similar to that used for the preparation of 10,
except that in this case equimolar dibromotrimer 9 and
diboronate 5 were used. EA: found: C, 87.39; H, 9.21. Calc.
1
for C101H125NO : C, 87.58; H, 9.10%. H NMR: d (400 MHz;
2 H
organic layer was then separated, dried over MgSO
4
, and
CDCl ; Me Si) 7.76–7.84 (8 H, m), 7.54–7.70 (14 H, m), 7.29–
3 4
passed through a short silica-gel column to remove the catalyst
residues. The resulting copolymer solution was collected,
concentrated, and precipitated into methanol. The pale yellow
solid (85% yield) obtained was dried overnight under vacuum
7.35 (2 H, m), 7.20–7.28 (6 H, m), 7.05–7.11 (1 H, m), 3.52 (4 H,
t), 2.09 (12 H, m), 1.39 (4 H, m), 1.00–1.28 (48 H, m), 0.69–0.88
(24 H, m).
2
at 50 uC. EA: found: C, 86.34; H, 8.51. Calc. for C72H85NO :
Thermal cross-linking of TPAFn using TBPA as a cross-linker
TPAFn (n = 2, 3) (9.0 mg) was dissolved in a solvent mixture of
1
C, 86.79; H, 8.60%. H NMR: dH (400 MHz; CDCl ; Me Si)
3
4
7
7
.74–7.84 (4 H, m), 7.54–7.68 (12 H, m), 7.28–7.35 (2 H, m),
.20–7.28 (6 H, m), 7.05–7.10 (1 H, m), 3.50 (4 H, t), 2.08 (8 H,
CHCl and THF (10 : 1 by volume). TBPA (1.5 equiv.) in THF
3
m), 1.38 (4 H, m), 1.00–1.28 (28 H, m), 0.70–0.88 (14 H, m).
solution was then added and the mixture was filtered through a
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 593–601 | 595

View Article Online
0
.2 mm filter. The resulting mixed solution of TPAFn
and TBPA gave a transparent thin film when spin-coated
onto a quartz substrate. Curing at 130 uC under vacuum (5 6
2
0
2
1
Torr) for 2 h resulted in the production of the cross-
linked film of X-TPAFn. The solvent resistance of the cured
film was tested by rinsing the film with a variety of solvents on
a spin-coater and then evaluating film thickness changes with
UV–vis absorption spectroscopy.
EL device fabrication and testing
ITO-coated glass substrates (15 V/%) were patterned by
conventional photolithography using an acid mixture of HCl
(
6 N) and HNO (0.6 N) as the etchant. After patterning, the
3
substrates were rinsed in deionized water, and then ultra-
sonicated sequentially in acetone and 2-propanol. The active
2
area of each EL device was 5 6 7 mm . Immediately prior to
3
device fabrication, ITO was UV/O -treated for 10 min. A thin
layer of the TPAFn and TBPA mixture, prepared as outlined
above, was then spin-coated onto the ITO substrate at
2
000 rpm for 60 s, and then cured at 130 uC under vacuum
1
for 2 h. The thickness of the resulting film was determined to
be around 40 nm using a Dektak III surface profilometer. A
thin layer (60–70 nm) of the copolymer OxFn was then cast on
top of the X-TPAFn layer to act as the electron transporting
and emitting layer. The device fabrication was completed by
the co-deposition of a Mg–Ag (9 : 1) alloy (100 nm) as a
cathode which was subsequently covered by a 10 nm thick
layer of Ag. The resultant devices were subsequently tested in
air under ambient conditions with no additional protective
encapsulation. EL spectra, and luminance–voltage (L–V) and
current–voltage (I–V) characteristics were recorded using a
combination of a Photo Research PR-650 SpectraScan and a
Keithley 238 Source meter.
3
Fig. 1 H NMR spectra of TPAF2 and TPAF3 in CDCl .
are in excellent agreement with their chemical structures.
Although the NMR signals in the aromatic region (7.00 to
7.85 ppm) are poorly resolved, it still appears that the peaks
associated with the triphenylamine units have relatively higher
intensities in the TPAF2 spectrum than in the TPAF3
spectrum. This is coincident with the fact that TPAF2 has a
higher triphenylamine content than TPAF3.
Cross-linking reaction
9
In our previous paper, we reported that aromatic boronic
acids readily undergo dehydration to produce the six-
membered boroxine rings. It was demonstrated that this
dehydration reaction could be readily applied to carbazole
or oligofluorenes containing diboronic acids which when
heated produce cross-linked networks that facilitated the
fabrication of multilayer devices. However, the boronic
anhydride linkages reported in this earlier paper were found
to be sensitive to hydrolysis when exposed to wet polar
solvents such as THF and acetone. In addition, it was difficult
to get pinhole-free thin films from these small molecular
materials using the spin-coating process. Clearly this imposes
obstacles in obtaining the optimized layer thicknesses. As a
continuation of this earlier work, we explore the feasibility of
using boronate esters as the cross-linking linkaging mechanism
to overcome these shortcomings in this paper. We have
designed and synthesized the alternating triphenylamine–
fluorene copolymers TPAFn (n = 2, 3) bearing hydroxyl
groups on the side chains to act as hole transport materials,
while using tris(4-dihydroxyboranylphenyl)amine (TBPA) as
the cross-linker. With this system the hydroxyl groups can
react with the boronic acids to produce boronate esters in a
condensation reaction which eliminates water. The removal of
the water using heat and vacuum pushes the equilibrium of
the reaction to give the desired cross-linked product. In this
study we found that baking the films at 130 uC under vacuum
Results and discussion
Materials synthesis
The procedures used for the preparation of the TPAFn
polymers (outlined in Scheme 1) are essentially the same as
those used for the preparation of the alternating oxadiazole–
1
4
fluorene and carbazole–fluorene copolymers. However, in
this work, a triphenylamine unit was introduced into each
repeat unit at a molar ratio of amine–fluorene of 1 : 2 for
TPAF2 and 1 : 3 for TPAF3, respectively. In addition, two
6
-hydroxyhexyl groups were attached to the C-9 position of
one of the fluorene units in every repeat unit. The presence
of these hydroxyl groups on the side chains offers reactive
sites with the multi-boronic acids containing cross-linkers
to facilitate cross-linked network formation. Consequently, a
solvent-resistant polymer thin film is obtained. The presence of
these hydroxyl groups on the side chain is clearly evident from
1
the H NMR spectra of TPAF2 and TPAF3 (Fig. 1) in which
the peaks at 3.52 and 1.40 ppm can be assigned to the protons
on the first and second carbons adjacent to the OH group,
respectively. The area ratio between the peak at 3.52 ppm and
the one at 2.08 ppm, ascribed to the protons on the alkyl
carbon adjacent to the C-9 position of the fluorene unit, is
2
2
1
: 2.05 for TPAF2 and 1 : 3.07 for TPAF3, respectively, which
(5 6 10 Torr) for 2 h is sufficient to produce films with
5
96 | J. Mater. Chem., 2006, 16, 593–601
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Fig. 2 UV–Vis absorption spectra of (a) pristine TPAF2, TPAF3 and TBPA films, (b) X-TPAF2 films before and after rinsing with THF and
3 3
CHCl , and (c) X-TPAF3 films before and after rinsing with THF and CHCl . All film thicknesses were about 40 nm.
excellent solvent resistance to solvents such as CHCl
3
, toluene
130 uC in air alone failed to achieve full cross-linking of the
films. This suggests that when heated in air the presence of
water vapor in the air and/or incomplete removal of the water
from the films are possible and can lead to films with poor
solvent resistance.
and even wet THF. As shown in Fig. 2, rinsing the cross-linked
film with organic solvents caused negligible change in the UV–
vis absorption spectra, indicating no change in the film
thickness. The relatively stronger absorbance at 215 nm in
the UV–vis absorption spectra of these cross-linked films when
compared with the pristine TPAFn films clearly indicates the
presence of TBPA in the cross-linked network. It is also worth
noting that the baking of the TPAFn and TBPA mixtures at
Influence of thermal cross-linking on the photophysical and
electrochemical properties of TPAFn
In terms of the photophysical properties of TPAFn, i.e. PL
spectra and absolute quantum efficiencies, it appears that the
cross-linking reaction has no detrimental effect. As can be seen
from Fig. 3, the PL spectra of the cross-linked X-TPAFn
networks are almost identical to those of the pristine TPAFn
films. Even rinsing with organic solvents such as THF and
3
CHCl had very little effect on the PL spectra, indicating the
excellent stability and robustness of these cross-linked films.
This is in marked contrast to the emergence of long-
wavelength emissions observed in the thermal treatment of
6
,7
fluorene homopolymers.
The measurement of the absolute PL quantum yields of
TPAFn thin films before and after cross-linking was performed
using a commercial spectrofluorimeter in combination with an
1
6
integration sphere according to the literature procedure. The
results are summarized in Table 1. From these data it would
appear that the cross-linked X-TPAFn networks have slightly
higher PL quantum efficiencies than the corresponding pristine
TPAFn films.
Table 1 The physical properties of TPAFn, OxFn, and POF
abs
l
max
/
3
a
fl
a
b
W
fl
Polymers
M
n
(610 ) PDI
T
g
/uC nm
l
em/nm
TPAF2
22.0
2.67 135
N. A. N. A. 384
2.9 122 386
N. A. N. A. 384
2.63 127
2.5 114
3.45 nd
382
437, 451
433, 453
434, 452
434, 454
0.16
0.19
0.17
0.18
0.22
c
X-TPAF2 N. A.
TPAF3 c 16.6
X-TPAF N. A.
OxF2
OxF3
POF
a
Obtained from thin films spin-coated on quartz slides. PL quantum
yields of thin films measured in an integration sphere in combination
with a spectrofluorimeter. Cross-linked films obtained by heating the
21.6
25.1
47.2
380, 396 425, 449
381, 393 423, 449, 481 0.28
385 422, 448, 479 0.36
d
b
c
d
mixture of TPAFn and TBPA at 130 uC under vacuum. No
noticeable T
and a melting peak at 159 uC were noted.
g
was observed; instead a crystallization peak at 113 uC
Fig. 3 PL spectra of (a) a X-TPAF2 film and (b) a X-TPAF3 film,
3
before and after rinsing with THF and CHCl .
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 593–601 | 597

View Article Online
Fig. 4 Cyclic voltammograms of TPAF2 and X-TPAF2 thin films on
a Pt electrode.
Fig. 5 The energy-level diagram of the PLED components used in
Fig. 4 shows the cyclic voltammograms of the pure TPAF2
film and the cross-linked X-TPAF2 film measured in a three-
this study.
4 6 3
electrode cell using 0.1 M Bu NPF in anhydrous CH CN as
the supporting electrolyte. From this figure it can be seen that
both films have very similar CV characteristics. Both show
highly reversible redox pairs under both oxidative and
reductive conditions with no apparent changes in the peak
potentials following 20 successive scans. However, a more
detailed investigation of these CV curves reveals that the cross-
linked X-TPAF2 film has slightly more stable CV properties.
For example, the cross-linked film showed a peak current
reduction of less than 10% following 20 oxidation and
reduction scans. This clearly indicates a high redox stability
of the cross-linked film for both hole and electron charging.
However, the non-cross-linked pure TPAF2 films showed a
slightly lower stability during multiple CV scans. About 30%
reduction in the peak current was found both for the oxidation
and reduction processes after 20 scans. In addition, the
oxidation redox pair of the non-cross-linked pure TPAF2 film
showed a certain amount of broadening with each successive
scan, leading to a reduction in the onset potential of the
oxidation process. The CV study on the non-cross-linked
pristine TPAF3 and cross-linked X-TPAF3 films reveals
similar results. These results demonstrate that the cross-
linked films have certain advantages in terms of improved
redox stability.
presented in Fig. 5 along with the data for other PLED
components. From this information it can be seen that the
incorporation of triphenylamine units into the polyfluorene
backbone significantly raises the HOMO energy levels of the
polymers, bringing their values close to the work function of
ITO. This effect obviously facilitates the hole injection from
ITO to the cross-linked X-TPAFn layer.
In order to investigate the performance of the X-TPAFn
films as a hole transport material, a double layer PLED was
fabricated by spin-coating an oxadiazole–fluorene copolymer,
OxF3, as an electron transport and emitting layer onto a
X-TPAFn layer. To facilitate comparisons, a LED device
incorporating the widely used PEDOT–PSS instead of
X-TPAFn as a hole transport layer was also fabricated. The
results of this comparison are presented in Fig. 6. From these
data it can be seen that the X-TPAFn produces a device which
has much superior performance to the PEDOT–PSS-based
22
device. For example, the luminous efficiency at 100 cd m is
increased by a factor of three to four when PEDOT–PSS is
replaced with X-TPAFn. It should be noted that an ideal HTL
not only maximizes hole injection from the anode, but also
blocks electron overflow from the emitting layer, thereby
dramatically increasing the electron–hole recombination effi-
ciency. However, in the case of PEDOT–PSS the lowest
Hole transport and electron-blocking properties of cross-linked
unoccupied molecular orbital (LUMO) level is located at
10
X-TPAFn
3
.4 eV, which is about 0.8 eV below that of OxF3 (Fig. 5).
The charge transport properties of fluorene-based conjugated
polymers can be effectively tuned from electron transporting
to hole transporting by the incorporation of suitable func-
This means that PEDOT–PSS has almost no electron-blocking
capability. Consequently, a significant fraction of the electron
flux just passes through the device to reach the ITO anode
without recombining with holes. However, when PEDDOT-
PSS is replaced by X-TPAFn, the electron-blocking properties
of the hole transport layers are improved due to their higher
LUMO energy levels as indicated in Fig. 5. A similar result has
recently been reported by Marks et al., who observed that
when an organosiloxane interpenetrating network was used
as the hole transport material instead of PEDOT–PSS a
5
,17
tional groups into the polymer backbones.
triphenylamine–fluorene alternating copolymers TPAFn
n = 2, 3) bearing hydroxyl groups on the side chains have
In this study,
(
been synthesized as hole transporting materials. The electron
energy levels of the cross-linked X-TPAFn can be calculated
from the CV results shown in Fig. 4, using the method
1
8
described by Br e´ das et al. and ourselves. These data are
5
98 | J. Mater. Chem., 2006, 16, 593–601
This journal is ß The Royal Society of Chemistry 2006

View Article Online
injection from the Mg–Ag alloy cathode to OxF3 is much
higher than the hole injection barrier from the ITO anode to
the HTL. Consequently, as shown in Fig. 6a, the EL emission
peak for the double-HTL device becomes broader due to large
imbalance in the charge injection and transport in this device,
causing deterioration in performance.
Photophysical, electrochemical and EL properties of OxFn
Although polyfluorenes are widely used as electron transport
blue emitters, the energy barrier for electron injection from air-
stable metals such as aluminium is still high due to their high
LUMO energy levels. Consequently a lot of efforts have been
made to lower the LUMO levels of PFs by introducing
electron-withdrawing groups into the polymers, either in the
1
4,17
backbones or on side chains.
In this study, we have investigated two alternating fluorene–
oxadiazole copolymers OxFn (n = 2, 3) with well-defined
chemical structures as emissive materials and investigated the
effect of oxadiazole content on the photophysical, electro-
chemical, and electroluminescent properties of the resulting
polymers. Cyclic voltammetry studies have shown that both
the HOMO and LUMO energy levels of OxFn are lowered
when the oxadiazole content is increased. Consequently, the
electron injection from the Mg–Ag cathode into OxFn
was greatly improved in comparison with that obtained with
the POF homopolymer. However, we also found that the
incorporation of 2,5-oxadiazole units into the polymer main
chains caused the absolute PL quantum yields of the resulting
polymers to decrease, as shown in Table 1. Although the PL
emission peaks of OxFn and POF appear at similar
wavelengths, Fig. 7 shows that the PL emission tails at long
wavelengths are appreciably greater upon the incorporation of
the oxadiazole units.
Fig. 6 (a) EL spectra and (b) voltage–luminance (filled symbols) and
voltage–efficiency (open symbols) characteristics of the following
fabricated devices: ITO/HTL (40 nm)/OxF3 (65 nm)/Mg–Ag (100 nm)/
Ag (10 nm), where the HTLs are the following materials: 1. PEDOT–
PSS; 2. X-TPAF3; 3. X-TPAF2; 4. PEDOT–PSS/X-TPAF2.
To investigate the EL properties of an OxFn, double-layer,
PLEDs incorporating X-TPAF2 with the structure ITO/X-
TPAF2/OxFn/Mg–Ag/Ag were fabricated. As can be seen
from Fig. 8a, the EL emissions from these devices fall within
the blue region with the CIE 1931 coordinates of (0.201,
0.185), (0.180, 0.176), and (0.188, 0.192) for POF, OxF3, and
OxF2, respectively. In terms of EL performance, the devices
fabricated with OxF2 and OxF3 were much better than those
fabricated from the POF homopolymer, requiring lower turn-
on voltages and achieving 100-fold greater maximum
luminances and 20-fold greater luminous efficiencies. The
improved performance of these materials can be ascribed to
their improved electron injection and transport properties.
When the Mg–Ag alloy was used as the cathode, the energy
barrier for the electron injection into the POF layer was as
high as 1.22 eV. In contrast, the hole injection barrier from
the X-TPAF2 HTL to POF was only about 0.42 eV.
Consequently, charge injection into the POF emitting layer is
greatly unbalanced. However, our experimental results suggest
that the use of the higher oxadiazole content in the OxFn
copolymers cannot ensure better EL performance due to the
reduced PL quantum efficiencies. For example, the maximum
significant improvement in PLED device performance could
1
0
be achieved.
Based on the data contained in Fig. 6, it can also be seen
2
1
that X-TPAF2 gives a higher luminous efficiency (0.69 cd A
)
2
2
and maximum luminance (1430 cd m ) than X-TPAF3
2
0.42 cd A , 950 cd m ) due to its higher triphenylamine
1
22
(
content and hence better hole transport/electron-blocking
properties. In addition, the copolymer TPAF3 contains
trifluorene units, which possess some electron transport
capability. Consequently a fraction of the electron flux
might pass through the X-TPAF3 layer to be unproductively
neutralized at the ITO anode. Once again, this demonstrates
the importance of the electron-blocking capability of the hole
transport layer.
In addition to the above single HTL devices, one with
double HTLs (PEDOT–PSS/X-TPAF2) was also fabricated.
Although the use of the double HTLs resulted in a decrease of
the turn-on voltage from 9 V to 7 V and an increase in the
current density of about 10% at the same electric field strength,
the luminous efficiency was reduced. This result implies that
this double-HTL device was a hole-dominant device. This is
not unusual because the energy barrier for the electron
2
1
luminous efficiency of OxF2 (0.59 cd A ) is lower than that of
2
OxF3 (0.69 cd A ). These EL efficiency data are comparable
1
5
d,8,19
to those reported in the literature,
considering the use of
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 593–601 | 599

View Article Online
Fig. 8 (a) EL spectra and (b) voltage–luminance (filled symbols) and
voltage–efficiency (open symbols) characteristics of the following
fabricated devices: ITO/X-TPAF2 (40 nm)/EML (60–70 nm)/Mg–Ag
Fig. 7 (a) UV–Vis absorption and (b) PL spectra of POF, OxF3,
and OxF2.
(
100 nm)/Ag, where the EMLs are the following materials: 1. POF; 2.
OxF3; 3. OxF2.
the Mg–Ag alloy as the cathode, which provides less ideal
electron-injection into the devices in comparison to the
commonly used calcium electrode. In addition, our polymeric
blue light-emitting diodes were fabricated using the more
economically viable spin-coating process.
injection and transport in OxF3 are better balanced, resulting
in a higher electron–hole recombination efficiency. When
OxF3 was used with Ca as the cathode, a maximum luminance
22
of 2010 cd m at a driving voltage of 11.5 V and a peak
22
2
luminous efficiency of 1.0 cd A at a luminance of 820 cd m
1
were achieved. Furthermore, the turn-on voltage, correspond-
Effects of the work function of the cathode on the device
performance
2
2
ing to 1 cd m , was reduced to 5 V. Although calcium was
reported as a catalyst for the formation of keto defects in
polyfluorenes, leading to significant emission peak broadening
A comparison between the EL data in Fig. 8 and 9 reveals that
the use of the low-work function Ca cathode significantly
increases the luminance and efficiency of the devices fabricated
with POF, while slightly improving the EL performance of
OxF3-based devices. However, the overall performance of the
OxF2-based device was not improved when Ca was used.
Indeed, the maximum luminance and luminous efficiency of
the OxF2 device decreased in comparison with the device
using a Mg–Ag cathode. In many respects this result is not
surprising, considering that OxF2 is an electron-dominant
emitting material. Using the data from Fig. 5, the energy
barrier for hole injection from X-TPAF2 to OxF2 can be easily
calculated. The value obtained is higher than that for electron
injection from Ca, meaning that there is an imbalance between
charge injection and transport in the emitting layer, resulting
in a lower luminous efficiency. On the contrary, the charge
8a
in blue PLEDs, the low-energy emission tail in our devices
using Ca as the cathode was small, as can be seen from Fig. 9a.
Conclusions
We have demonstrated that the formation of boronate esters
between hydroxyl groups and boronic acids under mild
conditions can be employed to fabricate cross-linked and
robust thin films with excellent solvent resistance. This cross-
linking reaction has been shown to have no detrimental
effects on the photophysical properties of the resulting films
and greatly facilitates the fabrication of multilayer PLEDs.
Two triphenylamine–fluorene alternating copolymers (TPAFn,
n = 2, 3) bearing hydroxyl groups on the side chains, and
a cross-linker (tris(4-dihydroxyboranylphenyl)amine, TBPA),
6
00 | J. Mater. Chem., 2006, 16, 593–601
This journal is ß The Royal Society of Chemistry 2006

View Article Online
References
1
(a) C. W. Tang and S. A. Vanslyke, Appl. Phys. Lett., 1987, 51,
13; (b) J. R. Sheats, H. Antoniadis, M. Hueschen, W. Leonard,
J. Miller, R. Moon, D. Roitman and A. Stocking, Science, 1996,
73, 884; (c) R. H. Friend, R. W. Gymer, A. B. Holmes,
9
2
J. H. Burroughes, R. N. Marks, C. Taliani, D. D. C. Bradlay,
D. A. Dos Santo, J. L. Br e´ das, M. L o¨ gdlund and W. R. Salaneck,
Nature, 1999, 397, 121.
2
3
(a) C. Adachi, T. Tsutsui and S. Saito, Appl. Phys. Lett., 1990, 57,
531; (b) J. Morgado, R. H. Friend and F. Cacialli, Appl. Phys.
Lett., 2002, 80, 2436.
(a) M. Strukelj, F. Papadimitrakopoulos, T. M. Miller and
L. J. Rothberg, Science, 1995, 267, 1969; (b) J. P. Lu, A. R. Hlil,
Y. Sun, A. S. Hay, T. Maindron, J. P. Dodelet and M. D’Iorio,
Chem. Mater., 1999, 11, 2501; (c) F. Huang, L. Hou, H. Wu,
X. Wang, H. Shen, W. Cao, W. Yang and Y. Cao, J. Am. Chem.
Soc., 2004, 126, 9845.
4
5
(a) E. Bellmann, S. E. Shaheen, S. Thayumanavan, S. Barlow,
R. H. Grubbs, S. R. Marder, B. Kippeen and N. Peyghambarian,
Chem. Mater., 1998, 10, 1668; (b) G. Kl a¨ rner, J. I. Lee, V. Y. Lee,
E. Chan, J. P. Chen, A. Nelson, D. Markiewwicz, R. Siemens,
J. C. Scott and R. D. Miller, Chem. Mater., 1999, 11, 1800.
(a) Q. Pei and Y. Yang, J. Am. Chem. Soc., 1996, 118, 7416; (b)
M. Leclerc, J. Polym. Sci., Part A: Polym. Chem., 2001, 22, 1365;
(
c) A. W. Grice, D. D. C. Bradeley, M. T. Bernius, M. Inbasekaran,
W. W. Wu and E. P. Woo, Appl. Phys. Lett., 1998, 73, 629; (d)
C. F. Shu, R. Dodda, F. I. Wu, M. S. Liu and A. K. Y. Jen,
Macromolecules, 2003, 36, 6698.
6
(a) J. I. Lee, G. Kl a¨ rner and R. D. Miller, Chem. Mater., 1999, 11,
1
083; (b) S. Setayesh, A. C. Grimsdale, T. Weil, V. Enkelmann,
K. M u¨ llen, F. Meghdadi, E. J. W. List and G. Leising, J. Am.
Chem. Soc., 2001, 123, 946.
7
8
(a) E. J. W. List, R. Guentner, P. S. Freitas and U. Scherf, Adv.
Mater., 2002, 14, 374; (b) J. M. Lupton, M. R. Craig and
E. W. Meijer, Appl. Phys. Lett., 2002, 80, 4489.
(a) X. Gong, P. K. Iyer, D. Moses, G. C. Bazan, A. J. Heeger and
S. S. Xiao, Adv. Funct. Mater., 2003, 13, 325; (b) J. P. Lu, Y. Tao,
M. D’iorio, Y. N. Li, J. F. Ding and M. Day, Macromolecules,
Fig. 9 (a) EL spectra and (b) voltage–luminance (filled symbols) and
voltage–efficiency (open symbols) characteristics of the following
fabricated devices: ITO/X-TPAF2 (40 nm)/EML (60–70 nm)/Ca
2
004, 37, 2442.
9
Y. Li, J. Ding, M. Day, Y. Tao, J. Lu and M. D’iorio, Chem.
Mater., 2003, 15, 4936.
(30 nm)/Al (100 nm), where the EMLs are the following materials: 1.
1
0 (a) H. Yan, B. J. Scott, Q. Huang and T. J. Marks, Adv. Mater.,
004, 16, 1948; (b) H. Yan, P. Lee, N. R. Armstrong, A. Graham,
G. A. Evmenenko, P. Dutta and T. J. Marks, J. Am. Chem. Soc.,
005, 127, 3172.
1 M. Range, D. Rondeau and M. Leclerc, Macromolecules, 1997, 30,
686.
POF; 2. OxF3; 3. OxF2.
2
2
have been designed and synthesized. The resulting cross-linked
X-TPAFn can be used as the hole transport layer in a PLED,
providing a better electron-blocking capability than the widely
used PEDOT–PSS, thereby greatly enhancing the performance
of OxFn-based devices. We have also investigated the
photophysical, electrochemical and EL properties of the
alternating copolymers of fluorene–oxadiazole OxFn (n = 2,
1
7
12 (a) M. Grell, W. Knoll, D. Lupo, A. Meisel, T. Miteva, D. Neher,
H.-G. Nothofer, U. Scherf and A. Yasuda, Adv. Mater., 1999, 11,
671; (b) E. D. Woo, W. R. Shiang, M. Inbasekaran and G. R. Roof,
US Pat., 5 962 631, 1999.
1
3 (a) M. Rehahn, A.-D. Schluter, G. Wegner and W. J. Feast,
Polymer, 1989, 30, 1060; (b) T. Vahlenkamp and G. Wegner,
Macromol. Chem. Phys., 1994, 195, 1933.
4 (a) J. Ding, M. Day, G. Robertson and J. Roovers,
Macromolecules, 2002, 35, 3474; (b) Y. Li, J. Ding, M. Day,
Y. Tao, J. Lu and M. D’Iorio, Chem. Mater., 2004, 16, 2165.
3
) and compared them with the POF homopolymer. The
1
resulting data clearly indicate that the incorporation of
oxadiazole units into the PF backbone lowers the LUMO
energy levels but decreases the PL quantum yields. The best
device performance was obtained with OxF3, which had a
15 H. R. Goodbrand and N. X. Hu, J. Org. Chem., 1999, 64, 670.
1
6 (a) L.-O. P a˚ lsson and A. P. Monkman, Adv. Mater., 2002, 14, 757;
(b) J. C. Mello, H. F. Wittmann and R. H. Friend, Adv. Mater.,
2
2
maximum luminance of 1430 cd m
at 18.5 V and a
22
1
997, 9, 230.
2
maximum luminous efficiency of 0.69 cd A at 500 cd m
1
1
7 (a) Z. Peng and J. Zhang, Chem. Mater., 1999, 11, 1138; (b)
A. C. Morteani, A. S. Dhoot, J. S. Kim, C. Silva, N. C. Greenham,
C. Murphy, E. Moons, S. Cina, H. Burrouges and R. H. Friend,
Adv. Mater., 2003, 15, 1708; (c) W. Ma, P. K. Iyer, X. Gong, B. Liu,
D. Moses, G. C. Bazan and A. J. Heeger, Adv. Mater., 2005, 17,
274.
when used with X-TPAF2 as the hole transport layer and a
Mg–Ag alloy as the cathode. The luminous efficiency of this
2
1
device could be further improved to 1.0 cd A when calcium
instead of Mg–Ag was used as the cathode.
1
8 (a) J. L. Br e´ das, R. Silbey, D. S. Boudreaux and R. R. Chance,
J. Am. Chem. Soc., 1983, 105, 6555; (b) J. Ding, Y. Tao, M. Day,
J. Roovers and M. D’Iorio, J. Opt. A: Pure Appl. Opt., 2002, 4,
S267.
Acknowledgements
We thank Mr G. Robertson at ICPET for the fruitful
discussion on the NMR spectra.
1
9 J. Jacob, S. Sax, T. Piok, E. J. W. List, A. C. Grimsdale and
K. M u¨ llen, J. Am. Chem. Soc., 2004, 126, 6987.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 593–601 | 601