Content uploaded by Franco Cacialli
Author content
All content in this area was uploaded by Franco Cacialli
Content may be subject to copyright.
Cross-linking of a poly(3,4-ethylene dioxythiophene):(polystyrene sulfonic
acid) hole injection layer with a bis-azide salt and the effect of atmospheric
processing conditions on device properties
Oliver Fenwick, Kate Oliver, and Franco Cacialli
Citation: Appl. Phys. Lett. 100, 053309 (2012); doi: 10.1063/1.3680606
View online: http://dx.doi.org/10.1063/1.3680606
View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v100/i5
Published by the AIP Publishing LLC.
Additional information on Appl. Phys. Lett.
Journal Homepage: http://apl.aip.org/
Journal Information: http://apl.aip.org/about/about_the_journal
Top downloads: http://apl.aip.org/features/most_downloaded
Information for Authors: http://apl.aip.org/authors
Downloaded 22 Sep 2013 to 187.95.35.18. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
Cross-linking of a poly(3,4-ethylene dioxythiophene):(polystyrene sulfonic
acid) hole injection layer with a bis-azide salt and the effect of atmospheric
processing conditions on device properties
Oliver Fenwick, Kate Oliver, and Franco Cacialli
a)
Department of Physics and Astronomy and London Centre for Nanotechnology, University College London,
Gower Street, London WC1E 6BT, United Kingdom
(Received 4 October 2011; accepted 9 January 2012; publish ed online 3 February 2012)
We investigate the role of atmospheric conditions during processing when cross-linking a
poly(3,4-ethylene dioxythiophene):(polystyrene sulfonic acid) hole injection layer with a bis-azide
salt. In particular, we investigate nitrogen atmosphere and air, since there is a competing reaction
of the cross-linker with oxygen. We show enhanced work function when processing under inert
conditions, with device properties otherwise unaffected. When processing is done in air, higher
irradiation dosages are needed to cross-link and the resulting films show lower work functions.
Surprisingly, the finished devices display an unexpected two-fold increase in efficiency which we
attribute to increased electron trapping in these films.
V
C
2012 American Institute of Physics.
[doi:10.1063/1.3680606]
Organic materials have been making multiple inroads
into the electronics market over the past 20 years, finding
applications from light-emitting diodes (LEDs) in displays,
lighting,
1,2
and night-vision readable displays
3
to photovol-
taics,
4,5
upconversion devices,
6
and radio frequency identifi-
cation tags (RFID). Their attractiveness is attributed to their
highly tunable optoelectronic properties combined with low
cost processing that is suitable for large area devices. One
persistent material in this story has been the conjugated poly-
mer poly(ethylene dioxythiophene), PEDOT, doped with
poly(styrene sulphonic acid), PSS. This blend (PEDOT:PSS)
exhibits conductivities in the region of 1–10
5
Scm
1
, tuna-
ble by the blend ratio and by additives,
7,8
and has found
application in plastic circuitry, conductive coatings, and as
both a hole injection layer in polymer LEDs and an anode its
own right. Its use as a hole injection layer (HIL) in LEDs,
which is the focus of this paper, is made possible both by its
high work function,
9
which aids hole injection from a lower
work function transparent conductive electrode such as in-
dium tin oxide (ITO) and by its transparency in the visible
region. Furthermore, PEDOT:PSS is processable from dis-
persion in water allowing it to be readily deposited as part of
multilayer structures with organic soluble light-emitting
polymers. Because of its importance to the plastic electronics
industry, developing methods of processing and patterning
PEDOT:PSS whilst preserving its electrical properties is
paramount.
Despite the wide variety of deposition techniques for or-
ganic semiconductors such as spray-coating, roll-to-roll
printing, and inkjet printing, photolithography remains
highly important due to its high throughput and its ability to
combine large area patterning with (sub-)micron sized
feature widths. One way photolithography can be made
compatible with PEDOT:PSS by blending a water soluble
photo-activated cross-linking agent into the PEDOT:PSS
suspension before deposition. Direct patterning by cross-
linking of PEDOT:PSS has the added benefit that it opens up
routes to the development of multi-layer devices where
another water-soluble conjugated polymeric material, such
as a perylene
10
or a polyrotaxane,
11
is deposited on top of
the PEDOT:PSS.
Accordingly, there have been reports of methods for
cross-linking PEDOT:PSS and these concentrate on the use
of water soluble azides, including bis(fluorinated phenyl az-
ide) (AAATf),
12
and 4,4
0
-diazido-2,2
0
-disulfonic acid benza-
lacetone disodium salt (DAB).
13
In this publication, we
focus on the latter of these two cross-linking agents for the
fact that it can be photo-activated at longer wavelengths than
the fo rmer (400 nm for DAB, Fig. 1, compared to 250 nm
for AAATf), avoiding the use of high energy radiation that
has the potential to damage other active components in
multilayer devices, and also because the slightly longer
wavelengths are readily compatible with unconventional
photolithography techniques such as scanning near-field op-
tical lithography
14,15
that commonly use tapered optical
fibers. A previous investigation
13
of PEDOT:PSS cross-
linked with DAB has demonstrated a photolithography pro-
cess with minimum line width of 3.5 lm and line spacing of
1 lm applied to the patterning of electrodes for the fabrica-
tion of transistors in all-polymer integrated circuits. We aim
to build on this work by considering its application as a hole
injection layer in polymer LEDs which depend critically on
the wo rk function of the PEDOT:PSS layer.
PEDOT:PSS-DAB films were prepared from a dispersion
in water of 1.3 wt. % PEDOT:PSS (Sigma Aldrich, conductive
grade 483095) with 0.5 wt. % DAB (Toyo Gosei Kogyo Co.
Ltd.). ITO-coated glass substrates were oxygen-plasma treated
no more than 5 min prior to spin-coating the PEDOT:PSS on
top to ensure freezing-in of the work function increase.
16
Resulting PEDOT:PSS films were 100 nm thick. Cross-
linking was achieved by illuminating the film through the
glass/ITO substrate from a UV bromograph with emission (3.9
mW/cm
2
) overlapping the broad absorption peak of DAB at
a)
Author to whom correspondence should be addressed. Electronic mail:
f.cacialli@ucl.ac.uk.
0003-6951/2012/100(5)/053309/4/$30.00
V
C
2012 American Institute of Physics100, 053309-1
APPLIED PHYSICS LETTERS 100, 053309 (2012)
Downloaded 22 Sep 2013 to 187.95.35.18. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
around 400 nm (Fig. 1) until the minimum dose required for
insolubilisation was achieved—E
N
2
¼ 195 mJ=cm
2
in nitrogen
atmosphere or E
air
¼ 780 mJ=cm
2
in air. This dose is larger
than previously reported, probably due to both absorption by
the ITO/glass substrate and the broad spectrum of the bromo-
graph which extends beyond the absorption band of the DAB.
Nevertheless, the larger dose required for insolubilisation in air
compared to nitrogen atmosphere (E
air
=E
N
2
¼ 4) is consistent
with reported values and is due to competition between the
intended reaction of the triplet nitrenes with C-H bonds that is
responsible for cross-linking and unintended reaction of the
same triplet nitrenes with oxygen.
13
The broad absorption peak
of DAB around 400 nm (Fig. 1) was observed to diminish
upon UV irradiation indicating reaction of the azide group and
yielding a resulting film that is transparent in the visible part of
the spectrum. The irradiated films were then rinsed in de-
ionized water to remove non-cross-linked PEDOT:PSS and
unreacted DAB. Retention of the PEDOT:PSS during rinsing
was clear by eye, and thickness measurements made with a
Dektak surface profilometer confirmed complete retention of
the layer for films irradiated above the minimum dose. It is
worth noting that UV-irradiation may also improve bonding
between the PEDOT:PSS layer and the substrate, although we
do not currently have any direct evidence for such a process.
Devices were then constructed by spin-coating a 100 nm
thick layer of the prototypical conjugated polymer poly(9,9
0
dioctylfluorene-alt-benzothiadiazol) (F8BT) from 1.3 wt. % in
xylene on top of the PEDOT:PSS layer. The choice of an or-
ganic soluble active layer ensures that we can compare our
cross-linked devices with analogues incorporating a HIL of
untreated PEDOT:PSS without introducing solven t compatibil-
ity issues. The device was finished with a top electrode of cal-
cium (50 nm) capped with aluminium (150 nm).
To investigate the suitability of cross-linked
PEDOT:PSS hole injection layers for LEDs, we first meas-
ured the work functions of cross-linked PEDOT:PSS layers
on ITO by the macroscopic Kelvin probe technique. The
measured work function of pristine PEDOT:PSS was
4.93 6 0.03 eV, but was 0.1 eV lower than this when
cross-linked in air (4.85 6 0.04 eV) or 0.1 eV higher
(5.01 6 0.03 eV) than for pristine PEDOT:PSS when cross-
linked in nitrogen (Table I). These values were calculated
against a reference of freshly cleaved highly oriented pyro-
lytic graphite (work function 4.475 eV (Ref. 17)). We note
that this formulation of PEDOT:PSS has a slightly lower
work function than some reported in the literature, but our
results are consistent with our previous report on the same
formulation.
12
Even the modest increase in work function for
nitrogen cross-linked films is important for polymer devices
since their highest occupied molecular orbital (HOMO) typi-
cally sits below the Fermi level of the anode (5.9 eV for
F8BT (Ref. 18)). The reason for the opposite effect when
cross-linking in air may be caused (a) by reduced cross-
linking in the PSS-rich surface layer where oxygen could be
expected to be more abundant, thereby favoring the compet-
ing nitrene-oxygen reaction and resulting in a certain amount
of re-solubilisation of the PSS that will tend to reduce the
dipole-effect that this surface-layer has on the work function
of the film or (b) by introduction of electronegative oxygen
species into the positively charged PSS-rich surface layer by
nitrene-oxygen reaction that will also tend to reduce the
dipole-effect.
We also observed a modest two-fold increase in conduc-
tivity of PEDOT:PSS after cross-linking by four-point probe
measurement. This could be in part due to the chemical dop-
ing caused by the introduction of DAB into PEDOT chains
but could equally be due to direct photochemical doping of
the PEDOT:PSS under UV illumination. Indeed previous
studies of azide cross-linking showed a 70-fold
12
or 100-
fold
13
increase in conductivity. It may be that the higher
energy photons (from a 250 nm LED (Ref. 12) or a mercury
lamp
13
) that we aim to avoid in this study are responsible for
the large increase. In our case, we illuminate through the
ITO-coated glass substrate and, as a result, high energy
photons from the 253 nm line of the bromograph may be
absorbed before reaching the PEDOT:PSS.
Representative current-light-voltage (JVL) characteris-
tics measured under vacuum of finished LEDs incorporating
the cross-linked and analogue PEDOT:PSS layers are plotted
in Fig. 2. As expected from our work-function measurements
on the pristine films, the light turns on at higher voltages for
the air-cross-linked device (2.78 V) than both the non-cross-
linked devices (2.41 V) and the N
2
-cross-linked ones
(2.30 V). Indeed this can also be seen in the current
FIG. 1. (a) The chemical structures of PEDOT, PSS, and DAB. (b) The
absorption spectra of both a neat PEDOT:PSS film and a PEDOT:PSS film
incorporating DAB (film thicknesses 100 nm), plotted alongside the emis-
sion spectrum of the UV bromograph used to activate the cross-linking. (c)
Atomic force micrographs of PEDOT:PSS films on ITO for both a neat
PEDOT:PSS films and a film that has been cross-linked with DAB (50 s in
N
2
) and rinsed to remove residual non-cross-linked material.
053309-2 Fenwick, Oliver, and Cacialli Appl. Phys. Lett. 100, 053309 (2012)
Downloaded 22 Sep 2013 to 187.95.35.18. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
characteristics where there is a sharp transition from unipolar
charge injection at low voltages to bipolar characteristics
above the light turn-on voltage. Perhaps surprisingly and de-
spite the unfavorable hole injection properties of the
PEDOT:PSS, we observe much higher quantum efficiencies
in the air cross-linked devices—0.43% when cross-linked in
air compared to 0.17% and 0.14%, respectively, for those
cross-linked in N
2
and those not cross-linked at all. These
higher quantum efficiencies for air cross-linked devices are
at the expense of slightly higher operating voltages, but lu-
minous efficacy in terms of the electrical driving power is
nonetheless trebled to 0.13 lm/W when compared to devices
incorporating non-cross-linked PEDOT:PSS.
We can investigate more deeply by looking in detail at
the current voltage characteristics. It is known that the
energy levels of F8BT (HOMO 5.9 eV, LUMO 3.5 to
3.3 eV (Refs. 18 and 19)) result in energy level line-up
between the LUMO and the cathode (Ca work function
2.8 eV (Ref. 20)), and thus that currents below light turn-
on are electron-dominated and that the onset of bipolar char-
acteristics is therefore expected to be controlled by the work
function of the ITO/PEDOT:PSS anode. Both the low and
high voltage current characteristics of Fig. 2 can be fitted by
a power law (J / V
x
). The switch between unipolar and
bipolar transport is seen in all devices by a sharp change in
the exponent, x. The bipolar current turn-on voltages, as the
light turn-on voltages, mirror the work functions measured
by Kelvin probe as we would expect. However, a closer look
at the sub-threshold currents shows some unexpected differ-
ences between the devices. The sub-threshold current in
nitrogen cross-linked devices is about twice that in a non-
cross-linked device (Table I), whereas it is reduced by a fac-
tor of three for the air cross-linked devices. In the sub-
threshold region, charge injection is unaffected by the anode
work function so we can infer that there must be either (a) a
change in either the bulk electron transport, (b) a change in
pinhole density in the PEDOT:PSS layer, or (c) modification
in electron trapping properties of the F8BT-PEDOT inter-
face. Atomic force microscope images of the cross-linked
PEDOT:PSS films showed no difference in morphology
compared to the non-cross-linked films (1.80 nm and
1.71 nm RMS roughness, respectively, Fig. 1) and no evi-
dence of pinholes in either case, ruling out explanation (b).
We find explanation (c) more likely than (a) since accumula-
tion of electrons at the F8BT-PEDOT interface has been
documented previously
21
and is also consistent with the
observed increase in electroluminescence quantum efficiency
for the air-cross-linked devices, which under this hypothesis
show the greatest degree of electron accumulation by means
of their lower sub-threshold currents. Indeed if we look at
the sub-threshold power law exponent, x , at 1 V below light
turn-on in the three cases, we see much lower values (1.3)
for the non- and N
2
-cross-linked cases than for the air-cross-
linked case (1.7). This indicates current characteristics
closer to space charge limited (x ¼ 2) than ohmi c (x ¼ 1) for
the air cross-linked case. In fact, it is likely that the F8BT-
PEDOT:PSS interface is altered in cross-linked devices for
two reasons: (a) introduction of Na
þ
ions from the DAB salt
and (b) the use of a post-deposition rinsing step that is made
possible by cross-linking and may remove low molecular
weight impurities otherwise present in PEDOT:PSS (such as
sulphuric acid or other PSS hydrolysis products).
In conclusion, we have cross-linked PEDOT:PSS hole
injection layers with a bis-azide salt, DAB, for use in light-
emitting diodes. We investigated the effect of the atmos-
pheric conditions during activation of the cross-linker which
TABLE I. Properties of the treated PEDOT:PSS films and operational characteristics of their devices (errors determined from the standard deviation of meas-
urements from many devices/samples).
Device properties Film properties
Film
Luminance
turn-on
a
(V)
EQE
max
(%)
Luminous
efficacy
b
(lm/W)
Sub-threshold
current
c
(mA/cm
2
)
Conductivity
(S/cm)
Work
function (eV)
PEDOT:PSS 2.41 6 0.17 0.14 6 0.06 0.05 6 0.01 0.4 6 0.2 0.08 4.93 6 0.03
PEDOT:PSS-DAB (N
2,
50 s) 2.30 6 0.08 0.17 6 0.03 0.05 6 0.02 0.8 6 0.2 0.19 4.85 6 0.04
PEDOT:PSS-DAB (air, 200 s) 2.78 6 0.05 0.43 6 0.06 0.13 6 0.02 0.13 6 0.04 0.18 5.01 6 0.03
a
Defined for light emission above 0.2 cd/m
2
.
b
Defined as the ratio of luminous flux to electrical power at 100 cd/m
2
.
c
Defined at a bias of 1 V.
FIG. 2. Current-voltage-light characteristics of LEDs fabricated with a
PEDOT:PSS hole injection layer compared with devices where the hole
injection layer has been cross-linked using DAB. (Inset) External quantum
efficiency of the same devices plotted against current density.
053309-3 Fenwick, Oliver, and Cacialli Appl. Phys. Lett. 100, 053309 (2012)
Downloaded 22 Sep 2013 to 187.95.35.18. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
is by UV exposure, finding that films cross-linked in air
show unfavorable wo k functions (4.85 eV) compared to
those cross-linked in N
2
atmosphere (5.01 eV) or non-cross-
linked (4.93 eV). This is reflected in the light turn-on voltage
in the respective devices. However, rather surprisingly, we
find more than a two-fold increase in efficiency for the air
cross-linked devices over the others. Lower sub-threshold
electron currents with more space charge limited character in
these devices suggest that this is due to a greater degree of
electron trapping at the F8BT-PEDOT:PSS interface.
We thank the EC (contracts: MRTN-CT-2006-036040
(THREADMILL), PITN-GA-2009–238177 (SUPERIOR)
and grant agreement no. 212311 FP7/2007-2013 (ONE-P))
and the EPSRC.
1
S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Luessem,
and K. Leo, Nature 459(7244), 234 (2009).
2
S. Brovelli, F. Meinardi, G. Winroth, O. Fenwick, G. Sforazzini, M. J.
Frampton, L. Zalewski, J. A. Levitt, F. Marinello, P. Schiavuta et al., Adv.
Funct. Mater. 20(2), 272 (2010).
3
O. Fenwick, J. K. Sprafke, J. Binas, D. V. Kondratuk, F. Di Stasio, H. L.
Anderson, and F. Cacialli, Nano Lett. 11(6), 2451 (2011).
4
P. Li, O. Fenwick, S. Yilmaz, D. Breusov, D. J. Caruana, S. Allard, U.
Scherf, and F. Cacialli, Chem. Commun. 47(31), 8820 (2011).
5
J. Peet, J. Y. Kim, N. E. Coates, W. L. Ma, D. Moses, A. J. Heeger, and G.
C. Bazan, Nature Mater. 6(7), 497 (2007).
6
D. Y. Kim, D. W. Song, N. Chopra, P. De Somer, and F. So, Adv. Mater.
22(20), 2260 (2010).
7
C. Ko, Appl. Phys. Lett. 90(6), 063509 (2007).
8
Sigma Aldrich, Materials Matters (2007), Vol. 2, p. 15.
9
T. M. Brown, J. S. Kim, R. H. Friend, F. Cacialli, R. Daik, and W. J. Feast,
Appl. Phys. Lett. 75(12), 1679 (1999).
10
T. Tang, J. Qu, K. Mu¨llen, and S. E. Webber, Langmuir 22(1), 26 (2005).
11
S. Brovelli, G. Latini, M. J. Frampton, S. O. McDonnell, F. E. Oddy, O.
Fenwick, H. L. Anderson, and F. Cacialli, Nano Lett. 8(12), 4546 (2008).
12
G. Winroth, G. Latini, D. Credgington, L.-Y. Wong, L.-L. Chua, P. K. H.
Ho, and F. Cacialli, Appl. Phys. Lett. 92(10), 103308 (2008).
13
F. J. Touwslager, N. P. Willard, and D. M. de Leeuw, Appl. Phys. Lett.
81(24), 4556 (2002).
14
D. Credgington, O. Fenwick, A. Charas, J. Morgado, K. Suhling, and F.
Cacialli, Adv. Funct. Mater. 20(17), 2842 (2010).
15
A. Charas, H. Alves, J. M. G. Martinho, L. Alcacer, O. Fenwick, F.
Cacialli, and J. Morgado, Synth. Met. 158(16), 643 (2008).
16
T. M. Brown, G. M. Lazzerini, L. J. Parrott, V. Bodrozic, L. Buergi, and
F. Cacialli, Org. Electron. 12(4), 623 (2011).
17
W. N. Hansen and G. J. Hansen, Surf. Sci. 481(1–3), 172 (2001).
18
E. Moons, J. Phys.: Condens. Matter 14(47), 12235 (2002).
19
L. L. Chua, J. Zaumseil, J. F. Chang, E. C. W. Ou, P. K. H. Ho, H. Sirring-
haus, and R. H. Friend, Nature 434(7030), 194 (2005).
20
Y. Park, V. Choong, E. Ettedgui, Y. Gao, B. R. Hsieh, T. Wehrmeister,
and K. Mullen, Appl. Phys. Lett. 69(8), 1080 (1996).
21
K. Murata, S. Cina, and N. C. Greenham, Appl. Phys. Lett. 79(8), 1193
(2001).
053309-4 Fenwick, Oliver, and Cacialli Appl. Phys. Lett. 100, 053309 (2012)
Downloaded 22 Sep 2013 to 187.95.35.18. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
