FULL PAPER

Polymer Light-emitting Diodes Based on End-capped

Poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl]
Zhang, Qiushu*()
School of Mechatronics Engineering, University of Electronic Science and Technology of China,
Chengdu, Sichuan 611731, China
We demonstrate polymer light-emitting diodes (LEDs) based on poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl]
with end capper dimethylphenyl or N,N-bis(4-methylphenyl)-N-phenylamine. The introduction of end-capper
groups increased the device luminance and efficiency, while greatly depressing the green emission. For the devices
constructed of poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl] end capped with dimethylphenyl, the maximum luminance reached 381 cd/m2 at 122 mA/cm2. The maximum external quantum efficiency was 0.16% at 117 mA/cm2,
which is more than five times higher than that of the non-end-capped polymer LEDs. The electroluminescence (EL)
maximum was at 485 nm, blue shifted by 52 nm with respect to that of the non-end-capped polyfluorene devices. It
is proposed that efficient hole trapping at end capper and increased resistance of polyfluorene to oxidation are responsible for the improved device performance and color stability.
Keywords

polymers, light-emitting diodes, polyfluorene, thin films, end-capping

Introduction
End-capping is one of the strategies in macromolecular engineering which can be exploited to modify
polymer properties to satisfy specific application requirements. For polymer light-emitting diode (LED)
applications, end-capping has been directed toward improving device efficiency,1-4 stabilizing blue emission,1,4-6 tuning the emission color,2,3,7 as well as realizing white electroluminescence.8 As a means for increasing polymer LED device performance, end-capping will not result in phase separation (a common
phenomenon in polymer blend systems) with time and
does not alter the electronic properties of the polymer
backbone.1 Furthermore, when end-capped with some
hole transport moieties, polyfluorenes can become more
resistant to oxidation.9
A few mechanisms have been proposed to explain
the increase in polymer LED device efficiency arising
from end-capping. Miteva et al. supposed that for their
devices based on end-capped polymer, most holes in the
emissive layer might be pointed to the end-capper
groups instead of sites with less efficient emission,
which subsequently recombined with the electrons on
the polyfluorene main chain.1 However, Nakazawa et al.
believed that the position of the exciton recombination
zone has an effect on light emission and device efficiency.4 They suggested that end-capping gives rise to
improved hole injection so that the recombination zone
is moved away from the polymer/anode interface. In
addition, energy transfer from polymer backbone to

end-capper could also lead to the enhanced efficiency.3

It was reported that end-capping could significantly
suppress troublesome long wavelength EL band in
polyfluorenes, which has been assigned to aggregates/
excimers,10,11 an emissive keto defect generated by virtue of thermo-, photo-, or electro-oxidative degradation,12 or a chemical defect located close to the cathode.13 Many end-cappers have been found to play an
important role in reducing green emission, among which
are mono-functional fluorene derivatives,14 anthracene,15 crosslinkable moieties,16,17 hole-trapping
groups,1 sterically hindered groups,18,19 polyhedral oligomeric silsesquioxanes,20 and so on. A possible reason
for this improvement of polyfluorene blue emission is
that the chain ends, rather than aggregates and excimer
forming sites, preferentially become the places where
the electron-hole recombinations occur.1 In some cases,
suppression of low energy emission is ascribed to the
shift of the recombination zone away from the polymer/
anode interface, since the polyfluorene molecules only
have strong interchain interactions near the anode interface but not in the bulk.4 Kadashchuk et al. found that
polyfluorenes end capped with certain hole-transporting
molecules, like triphenylamine derivatives, demonstrate
high stability against oxidation, which leads to considerably decreased concentration of keto defects.9
In this work, we fabricated and characterized polymer LED devices based on poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl] (PF2/6),21 end capped with dimethylphenyl (DMP) or N,N-bis(4-methylphenyl)-N-phenylamine (TPA).4 Compared to PF2/6 based devices,

* E-mail: qiushuzhang@uestc.edu.cn; Tel.: 0086-028-61830598; Fax: 0086-028-61830598

Received November 20, 2009; revised and accepted April 1, 2010.

1482

2010 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Chin. J. Chem. 2010, 28, 14821486

Polymer Light-emitting Diodes Based on Poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl]

end-capped PF2/6 devices displayed improved properties in terms of luminance and efficiency. Predominant
long wavelength emission observed in PF2/6 was significantly suppressed in end-capped PF2/6s. The effect
of end-capping on the enhancement of device performance and color stability is discussed.

Experimental
The chemical structures of PF2/6 and end-capped
PF2/6s are shown in Figure 1. PF2/6 was purchased
from Sigma-Aldrich Corporation, United States. It is
comprised of yellow-green crystal-like particulates and
soluble in common organic solvents such as toluene,
xylene, CHCl3 and tetrahydrofuran. Both end-capped
PF2/6s were synthesized by American Dye Source, Inc.,
Canada. They are light yellow powder, and highly soluble in toluene and tetrahydrofuran. The weight average
molecular weight Mw and polydispersity of DMP- and
TPA-end-capped PF2/6 are 69 000, 54 000 and 2.4, 2.5,
respectively, as determined by gel permeation chromatography in tetrahydrofuran using polystyrene standards.

fonate) (PEDOT:PSS) was spin coated over the substrate from 1.3 wt% water dispersion. The emissive layers consisted of PF2/6, or end-capped PF2/6. They were
formed at the top of PEDOT:PSS films by spin casting
from polymer toluene solutions (10 mg/mL) at a speed
of 2000 r/min. The film thicknesses are ca. 40 nm for
PF2/6, ca. 70 nm for DMP-end-capped PF2/6, and ca.
80 nm for TPA-end-capped PF2/6. Devices were dried
in a vacuum chamber at room temperature for a minimum of 24 h before the deposition of the Al film. The
manufacturing was completed with the thermal evaporation of the aluminum cathode (ca. 200 nm) at 2.67

10 4 Pa through a shadow mask. The overlap between

the two electrodes gave device active areas of 9 mm2.
The optical absorption (UV-Vis) spectra were measured with an Agilent-8453 UV-visible spectrophotometer from polymer films spin cast onto a quartz plate.
Photoluminescence (PL) spectra were obtained with a
PSI (Photon Technology International, Canada) fluorescence spectrometer. All device testing was implemented in air at room temperature. Current-voltage
characteristics were measured on a Keithley 236
source-measure unit. The power of EL emission was
measured through using an ILX Lightwave
OMM-6810B optical multimeter equipped with a silicon
power/wavehead (OMH-6722B). By assuming Lambertian distribution of the EL emission, luminance (cd/m2)
was calculated via utilizing the forward output light
power and the EL spectra of the devices.

Results and discussion

The optical absorption (UV-Vis) spectra of thin
films of PF2/6 and end-capped PF2/6s are shown in
Figure 2. The absorption of PF2/6 exhibited an onset at
415 nm and a peak at 362 nm. The UV-Vis spectra of
DMP- and TPA-terminated PF2/6 are very similar to
each other. The absorption of DMP-end-capped PF2/6
had an onset at 425 nm, the same as that of
TPA-end-capped PF2/6. And both spectra exhibited a
peak at nearly the same wavelength. However, the opti-

Figure 1 Chemical structures of light emitting polymers: (a)

PF2/6, (b) DMP-end-capped PF2/6, (c) PF2/6 end-capped with
TPA.

The devices with a configuration of ITO/PEDOT:

PSS/emissive layer/aluminum were fabricated at ambient conditions on glass substrates covered by patterned
indium-tin-oxide (ITO) electrodes. The ITO substrates
were precleaned by two successive ultrasonic rinses in
acetone and isopropyl alcohol. After drying them with a
nitrogen gun, a 45-nm thick film of poly(3,4-oxyethyleneoxythiophene) doped with poly(styrene sulChin. J. Chem. 2010, 28, 14821486

Figure 2 Optical absorption (UV-Vis) spectra of thin films of

PF2/6, DMP-end-capped PF2/6 and TPA-end-capped PF2/6.

2010 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cjc.wiley-vch.de

1483

Zhang

FULL PAPER
cal absorption spectra of PF2/6 with end-capper groups
were clearly red-shifted compared to that of PF2/6,
which indicates that the conjugation length in endcapped PF2/6s might be longer than that in PF2/6. From
the UV-Vis spectra, the band-gap energies were determined to be 2.99 eV for PF2/6 and 2.92 eV for
end-capped PF2/6s, respectively.
PL spectra of thin films of PF2/6, DMP-end-capped
PF2/6 and TPA-end-capped PF2/6 are presented in Figure 3. The PL emissions from PF2/6 and end-capped
PF2/6 thin films showed well-defined vibronic features.
The PL of PF2/6 had a maximum at ca. 440 nm. A
weaker long-wavelength emission (ca. 530 nm) is pronounced in the spectrum. The two end-capped PF2/6s
displayed very similar PL spectra with the maximum at
ca. 410 nm. The low energy emission bands were completely suppressed in the case of end-capped PF2/6s.
End-capping brings about red-shift in absorption and
blue-shift in PL emission. As a consequence, the Stokes
shift was reduced, which might be due to the increased
stiffness of the polymer upon end-capping.

Figure 3 Photoluminescence spectra of thin films of PF2/6,

DMP-end-capped PF2/6 and TPA-end-capped PF2/6.

Figure 4 shows the current density-electric field

characteristics of polymer LEDs based on PF2/6 and
end-capped PF2/6s. The turn-on voltage was around 2 V
(5105 V/cm) for PF2/6 based devices, around 5 V
(7.14105 V/cm) for DMP-end-capped PF2/6 devices,
and around 5 V (6.25105 V/cm) for TPA-end-capped
PF2/6 devices, respectively. A crossover is observed
between the current density-electric field characteristics
of the devices based on PF2/6 and end-capped PF2/6s,
which arises at ca. 2.5106 V/cm for DMP-end-capped
polymer, and at ca. 2.1106 V/cm for TPA-end-capped
polyfluorene. As the charge carrier mobility in conjugated polymers is field-dependent,22 it is reasonable to
assume that the charge carrier mobility in end-capped
PF2/6s is more sensitive to the electric field than that in
PF2/6 as evidenced by the greater currents in the
end-capped PF2/6 devices than PF2/6 devices at the
higher fields.
The current dependence of light power for the poly1484

www.cjc.wiley-vch.de

Figure 4 Current density-electric field characteristics of the

LEDs based on PF2/6 (), DMP-end-capped PF2/6 (), and
TPA-end-capped PF2/6 ().

mer LED devices was recorded to calculate device efficiencies and brightness. Figure 5 displays luminance as
a function of the current density for the three types of
devices under study. End-capping decreases the threshold current for light emission from 96.9 mA/cm2 to 2.4
(DMP-end-capped PF2/6) and 1.2 mA/cm2 (TPA-endcapped PF2/6), which suggests that there is better balance between holes and electrons within the end-capped
PF2/6 emission layer. The charge balance between holes
and electrons significantly affects the device efficiency
because a surplus of either of the charge carriers results
in a current increase that does not enhance the emission,
but raises Joule heating which causes more rapid polymer degradation.23,24 For PF2/6 active layer, the maximum brightness was 179 cd/m2 at 214 mA/cm2. The
maximum luminance efficiency and maximum external
quantum efficiency were calculated to be 0.084 cd/A
and 0.03% at 214 mA/cm2, respectively. In comparison
with the devices based on non-end-capped polymer,
end-capping improves EL properties in light of luminance and efficiency. In the case of TPA-terminated
PF2/6, the luminance reached 327 cd/m2 at 211 mA/cm2.
The maximum luminance efficiency and maximum ex-

Figure 5 Luminance-current density characteristics of the

LEDs based on PF2/6 (), DMP-end-capped PF2/6 (), and
TPA-end-capped PF2/6 ().

2010 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Chin. J. Chem. 2010, 28, 14821486

Polymer Light-emitting Diodes Based on Poly[9,9-di-(2'-ethylhexyl)fluorenyl-2,7-diyl]

ternal quantum efficiency were 0.156 cd/A and 0.069%

at 205 mA/cm2. The best performance was observed for
DMP-end-capped PF2/6 devices with a maximum luminance of 381 cd/m2 at 122 mA/cm2, a maximum luminance efficiency of 0.319 cd/A at 117 mA/cm2, and a
maximum external quantum efficiency of 0.16% at 117
mA/cm2, which are higher than the PF2/6 devices by a
factor of 2.1, 3.8 and 5.3, respectively.
As shown in Figure 6, end-capping significantly
suppresses the long wavelength emission that is overwhelming in the PF2/6 EL spectrum. The blue excitonic
emission bands that display the intrinsic characteristics
of polyfluorenes become predominant in EL emission
from end-capped PF2/6s. It is especially pronounced for
DMP-end-capped PF2/6. Its EL peaks at 420, 445 and
485 nm, which all fall into violet-blue region. The
maximum intensity is at 485 nm, blue shifted by 52 nm
relative to that of PF2/6.

Figure 6 Electroluminescence spectra of the LEDs based on

PF2/6, DMP-end-capped PF2/6, and TPA-end-capped PF2/6.

For polyfluorene devices having a calcium cathode,

electron-hole recombination is expected to happen close
to the anode since electron injection from calcium to
polyfluorene is considered to be easier than hole injection from PEDOT:PSS to polyfluorene.25,26 With calcium cathode in the device structure, the enhancement
of device performance originating from end-capping is
therefore believed to correlate with the improved hole
injection caused by end-capping, which moves the exciton recombination zone from the polymer/anode interface into the bulk of the light emitting polymer.4
However, in all of our devices, electron-hole recombination could occur near the cathode because of the poor
electron injection from aluminium cathode to polymer.
This indicates that the end-capped PF2/6 device properties could have nothing to do with the position of the
exciton recombination zone. The increased efficiency
and color stability achieved with the end-capped PF2/6
devices might be due to efficient hole trapping at the
end-capper groups.1 In the PF2/6 emissive layer, a certain portion of the electron-hole recombinations might
occur at sites with less efficient emission, like aggreChin. J. Chem. 2010, 28, 14821486

gates or excimer forming sites. However, for the active

layer of end-capped PF2/6, the end-cappers may have
preference over aggregates or excimer forming sites
when they compete with each other with respect to hole
trapping. Consequently, most holes could be directed to
the end-capper groups, subsequently recombining with
the electrons on the polymer main chain. Additionally, it
is also likely that end-capping enhances the resistance of
polyfluorene to oxidation, hence considerably decreasing concentration of keto defects.9 DMP-end-capping
could give rise to more efficient hole trapping at the
end-capper groups and/or higher resistance of polyfluorene to oxidation than TPA-end-capping, which is perhaps responsible for the better device performance observed from DMP-end-capped PF2/6 based LEDs.
Besides the end-capper moieties, the end-capper
concentration can also affect device properties. Endcapping could have an influence upon charge transport
and injection. At low end-capper concentration, the
current could drop sharply compared to non-end-capped
polymer because of severe charge trapping on endcapper group. However, current could increase gradually at higher concentrations since charge transport
could be enabled by hopping via end-capper. Moreover,
charge injection could be facilitated at higher
end-capper concentrations. As shown in Figure 4, the
devices based on end-capped PF2/6 exhibited the current comparable to that of the non-end-capped polyfluorene devices, which suggests the high end-capper
concentration in the end-capped PF2/6s. The same conclusion can be reached from the fact that the long
wavelength emission is nearly completely depressed in
the case of end-capped PF2/6 (Figures 3 and 6) as the
green emission bands observed from PF2/6 will decrease continuously with increasing end-capper concentration.1 On the other hand, exciton quenching at
charged traps might occur.27 Energy of excitons could
be transferred to the trapped charges via Forster mechanism, which is very efficient at a sufficiently large
charge concentration and eventually results in the release of charges from the traps. With the high
end-capper concentration, exciton quenching might
harm EL emission from the end-capped PF2/6 devices
under study. However, the process of carrier recombination and light emission is still dominated by the EL enhancement resulting from efficient hole trapping at the
end-capper groups rather than sites with less efficient
emission. Consequently, end-capped PF2/6 devices
demonstrated better device properties (Figure 5). Further investigation is needed concerning the influence of
end-capper concentration on polymer LED device performance.

Conclusion
End-capped PF2/6 was used as active layer to fabricate polymer LED devices. Investigation results demonstrate that end-capping PF2/6 with DMP or TPA clearly

2010 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cjc.wiley-vch.de

1485

Zhang

FULL PAPER
improves device properties, while significantly suppressing green emission bands. The best performance is
obtained from devices with a DMP-end-capped PF2/6
emissive layer that exhibited a maximum luminance of
381 cd/m2, a maximum luminance efficiency of 0.319
cd/A, and a maximum external quantum efficiency of
0.16 %, which are 25 times higher than those of the
PF2/6 LEDs. All three EL peaks belong to the violet-blue zone with a maximum emission at 485 nm,
which has a blue shift of 52 nm with respect to that of
the PF2/6 based devices. Efficient hole trapping at the
end-capper groups as well as increased resistance of
polymer to oxidation might lead to the enhancement of
device properties and blue emission.

10
11
12
13
14
15
16

Acknowledgment
17

The author would like to acknowledge David Keith

Chambers and Joseph Cannon at Louisiana Tech University for help in device testing.

18

References

19

5
6
7
8
9

Miteva, T.; Meisel, A.; Knoll, W.; Nothofer, H. G.; Scherf,

U.; Muller, D. C.; Meerholz, K.; Yasuda, A.; Neher, D. Adv.
Mater. 2001, 13, 565.
Huang, Y.; Lu, Z.; Peng, Q.; Jiang, Q.; Xie, R.; Han, S.;
Dong, L.; Peng, J.; Cao, Y.; Xie, M. Mater. Chem. Phys.
2005, 93, 95.
Cao, J.; Zhou, Q.; Cheng, Y.; Geng, Y.; Wang, L.; Ma, D.;
Jing, X.; Wang, F.; Kane-Maguire, L. A. P.; Officer, D. L.
Synth. Met. 2005, 152, 237.
Nakazawa, Y. K.; Carter, S. A.; Nothofer, H.-G.; Scherf, U.;
Lee, V. Y.; Miller, R. D.; Scott, J. C. Appl. Phys. Lett. 2002,
80, 3832.
Lee, J.-I.; Lee, V. Y.; Miller, R. D. ETRI J. 2002, 24, 409.
Neher, D. Macromol. Rapid Commun. 2001, 22, 1365.
Lee, J.-F.; Hsu, S. L.-C. Polymer 2009, 50, 2558.
Kuo, C.-H.; Cheng, W.-K.; Lin, K.-R.; Leung, M.-K.; Hsieh,
K.-H. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 4504.
Kadashchuk, A.; Schmechel, R.; von Seggern, H.; Scherf,
U.; Vakhnin, A. J. Appl. Phys. 2005, 98, 024101.

20
21
22
23

24
25
26
27

Yang, N. C.; Lee, S. M.; Yoo, Y. M.; Kim, J. K.; Suh, D. H.

J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 1058.
Kwag, G.; Park, E.; Lee, S. N. J. Appl. Polym. Sci. 2005, 96,
1335.
Tokito, S.; Weinfurtner, K.-H.; Fujikawa, H.; Tsutsui, T.;
Taga, Y. Proc. SPIE 2001, 4105, 69.
Park, J. H.; Ko, H. C.; Kim, J. H.; Lee, H. Synth. Met. 2004,
144, 193.
Lee, J.-I.; Klaerner, G.; Miller, R. D. Chem. Mater. 1999, 11,
1083.
Lee, J. I.; Hwang, D. H.; Park, H.; Do, L. M.; Chu, H. Y.;
Zyung, T.; Miller, R. D. Synth. Met. 2000, 111112, 195.
Klrner, G.; Lee, J. I.; Lee, V. Y.; Chan, E.; Chen, J. P.;
Nelson, A.; Markiewicz, D.; Siemens, R.; Scott, J. C.;
Miller, R. D. Chem. Mater. 1999, 11, 1800.
Chen, J. P.; Klaerner, G.; Lee, J.-I.; Markiewicz, D.; Lee, V.
Y.; Miller, R. D.; Scott, J. C. Synth. Met. 1999, 107, 129.
Klarner, G.; Davey, M. H.; Chen, W.-D.; Scott, J. C.; Miller,
R. D. Adv. Mater. 1998, 10, 993.
Lee, J.-I.; Klrner, G.; Chen, J. P.; Scott, J. C.; Miller, R. D.
Proc. SPIE 1999, 3623, 2.
Fenenko, L.; Nakanishi, Y.; Tokito, S.; Konno, A. Jpn. J.
Appl. Phys., Part 1 2006, 45, 550.
Gong, X.; Iyer, P. K.; Moses, D.; Bazan, G. C.; Heeger, A.
J.; Xiao, S. S. Adv. Func. Mater. 2003, 13, 325.
Bozano, L.; Carter, S. A.; Scott, J. C.; Malliaras, G. G.;
Brock, P. J. Appl. Phys. Lett. 1999, 74, 1132.
Sainova, D.; Miteva, T.; Nothofer, H. G.; Scherf, U.;
Glowacki, I.; Ulanski, J.; Fujikawa, H.; Neher, D. Appl.
Phys. Lett. 2000, 76, 1810.
Tessler, N.; Harrison, N. T.; Thomas, D. S.; Friend, R. H.
Appl. Phys. Lett. 1998, 73, 732.
Janietz, S.; Bradley, D. D. C.; Grell, M.; Giebeler, C.; Inbasekaran, M.; Woo, E. P. Appl. Phys. Lett. 1998, 73, 2453.
Brown, T. M.; Kim, J. S.; Friend, R. H.; Cacialli, F.; Daik,
R.; Feast, W. J. Synth. Met. 2000, 111112, 285.
Romanovskii, Y. V.; Arkhipov, V. I.; Bssler, H. Phys. Rev.
B 2001, 64, 033104.
(E0911201 Pan, B.)

1486

www.cjc.wiley-vch.de

2010 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Chin. J. Chem. 2010, 28, 14821486