An Organic Light‐Emitting Diode with Field‐Effect Electron Transport

Full text

Page 1

DOI: 10.1002/adfm.200700769
An Organic Light-Emitting Diode with Field-Effect Electron
Transport**
By Sarah Schols,* Stijn Verlaak, Cédric Rolin, David Cheyns, Jan Genoe, and Paul Heremans
1. Introduction
Organic light-emitting devices have attracted much attention
since the first report of electroluminescence in a device based
on an organic thin film.[1]Over the last decade, remarkable
progress has been achieved in the development of organic
light-emitting diodes (OLEDs) based on polymers as well as
small-molecule semiconductors.[2]It has been shown that the
current in OLEDs is typically space charge limited,[3,4]i.e., lim-
ited by the bulk of the semiconductor. In other words, the mea-
sured current is a drift current determined by the mobility of
the charge carriers. The carrier mobility in the organic semi-
conductor element of OLEDs is typically low.[4,5]To obtain ef-
ficient charge transport at reasonable driving voltages, the total
thickness of the organic layers in OLEDs is therefore generally
limited to 80 to 100 nm.[6]Consequently, light is generated very
close (within ca. 50 nm) to the metallic cathode. The proximity
–
[*] S. Schols, Dr. S. Verlaak, C. Rolin, D. Cheyns, Dr. J. Genoe,
Prof. P. Heremans
IMEC v.z.w. – SOLO/PME
Kapeldreef 75, 3001 Leuven (Belgium)
E-mail: sarah.schols@imec.be
S. Schols, D. Cheyns, Prof. P. Heremans
Electrical Engineering Department, Katholieke Universiteit Leuven
3001 Leuven (Belgium)
S. Schols
FWO Vlaanderen
Egmontstraat 5, 1000 Brussels (Belgium)
[**] This work is partially supported by the EU-funded Project OLAS (con-
tract No. 015034). The authors thank Merck for PTAA that has been
provided through the EU-funded Integrated Project POLYAPPLY (con-
tract No. 507143). S. Schols thanks the FWO Vlaanderen for financial
support.
of the metal electrode induces severe absorption losses if the
OLED is used as a waveguide. This phenomenon has been
demonstrated by Andrew et al., who have reported a substan-
tial increase of the lasing threshold of a poly(2-methoxy-5-(2′-
ethylhexoxy)-1,4-phenylenevinylene) (MEH-PPV)-based or-
ganic laser upon the insertion of a thin silver layer.[7]Reufer et
al. have proved that this detrimental waveguide loss caused by
metallic layers may be reduced by increased optical confine-
ment.[8]However, this is only possible when the polymer layer
is sufficiently thick, in the order of several hundreds of nano-
meters. The low carrier mobilities of thick organic semiconduc-
tor layers give rise to a large electrical resistance, thus preclud-
ing the use of thick films in the fabrication of efficient OLEDs.
To circumvent this problem, transparent electrodes fabricated
from materials such as indium tin oxide (ITO)[9]or aluminum-
doped zinc oxide (AZO)[10]have been proposed as alternatives
to inject electrons and holes in the organic semiconductor.
Yamamoto et al.[11]and Görrn et al.[12]have demonstrated that
thin ITO and AZO layers, respectively, can indeed be used as
low-loss contacts in waveguide structures. This is mainly be-
cause of the much lower optical losses in the visible spectral
range for these materials as compared to metallic layers.[13]In
addition, Reufer et al. have reported that thick, more conduc-
tive ITO layers can also be used for this purpose, provided that
a crosslinked hole-transport layer is inserted between the
highly conductive ITO and the emission layer.[14]
Recently,light-emittingorganic
(LEOFETs) have been proposed as lateral light-emitting de-
vices, complementary to traditional vertical light-emitting di-
odes. Since these structures combine the optical output of an
OLED and the gate control of an organic field-effect transistor
in one single device, they may become interesting constructs in
the field of organic displays. Displays based on LEOFETs may
field-effecttransistors
136
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2008, 18, 136–144
We describe an organic light-emitting diode (OLED) using field-effect to transport electrons. The device is a hybrid between a
diode and a field-effect transistor. Compared to conventional OLEDs, the metallic cathode is displaced by one to several mi-
crometers from the light-emitting zone. This micrometer-sized distance can be bridged by electrons with enhanced field-effect
mobility. The device is fabricated using poly(triarylamine) (PTAA) as the hole-transport material, tris(8-hydroxyquinoline) alu-
minum (Alq3) doped with 4-(dicyanomethylene)-2-methyl-6-(julolindin-4-yl-vinyl)-4H-pyran (DCM2) as the active light-emit-
ting layer, and N,N′-ditridecylperylene-3,4,9,10-tetracarboxylic diimide (PTCDI-C13H27), as the electron-transport material.
The obtained external quantum efficiencies are as high as for conventional OLEDs comprising the same materials. The quan-
tum efficiencies of the new devices are remarkably independent of the current, up to current densities of more than 10 A cm–2.
In addition, the absence of a metallic cathode covering the light-emission zone permits top-emission and could reduce optical
absorption losses in waveguide structures. These properties may be useful in the future for the fabrication of solid-state high-
brightness organic light sources.
FULL PAPER

Page 2

eliminate the difficult integration of an organic light-emitting
structure and the organic driving backplane. LEOFETs may
also be used to study the optoelectronic performance of organ-
ic semiconductors. In contrast to OLEDs where charge trans-
port occurs perpendicularly to the organic layers, charge trans-
port in LEOFETs occurs in the plane and the carriers are
transported by field-effect.[15]In field-effect transistors the cur-
rent flowing between the source and the drain is modulated by
applying a bias to a third contact, the gate electrode. In this
way, charge carriers can be accumulated or depleted in the
semiconductor close to the semiconductor/insulator interface.
The gate field provides an additional degree of control over
the amount of charge present in the semiconductor. In addi-
tion, the accumulated charges can flood deep traps, giving rise
to a higher effective mobility for the remaining carriers. The
field-effect mobility can be several orders of magnitude higher
than the charge carrier mobility in a conventional OLED.[16]
Although the basic concept of a LEOFET dates back from
1996,[17]the development of LEOFETs is still in a relatively
embryonic stage. The working principle of a LEOFET is based
on the simultaneous injection of electrons and holes into a dou-
ble or ambipolar layer by tuning of the gate–source and drain–
source voltages. The accumulated charge is zero when the bias-
ing conditions are such that the potential at some point in the
channel equals the gate potential. Consequently, at this point
the electron and hole accumulation layers vanish. Exciton for-
mation occurs near this point and radiative relaxation of these
excitons to the ground state leads to light emission. The first
LEOFET has been demonstrated by Hepp et al.[18]and is based
on vacuum-evaporated tetracene as the organic semiconductor.
Since this first demonstration, LEOFETs have been fabricated
using polymers,[19,20]small molecules,[21–23]and a heterostruc-
ture of p- and n-type organic semiconductors.[24–29]Recent re-
ports in the literature have focused on LEOFETs based on
an ambipolar polymeric semiconductor.[30,31]An overview of
light-emitting organic transistors has also been recently pub-
lished.[32]
In LEOFET devices, the charge carriers have a field-effect
mobility, and depending on the type of LEOFET, the light-
emission zone can be located at some distance from the metal-
lic source and drain electrodes. However, to obtain these prop-
erties, three different electrodes need to be used: the source,
the drain, and the gate. In addition, most organic field-effect
materials demonstrate rather weak photoluminescence; analo-
gously, most organic light-emitting materials show poor field-
effect performance.[33]Indeed, a high carrier mobility, which is
a prerequisite for transistors, is favored by structural features
such as tight intermolecular p-stacking. Examples of suitable
materials include pentacene[34]and regioregular poly(3-hexyl-
thiophene) (P3HT).[35]However, the photoluminescence yield
is usually low in such materials with strong intermolecular cou-
pling. Conversely, materials exhibiting high photoluminescence
quantum yields, for example, poly(p-phenylene vinylene)s
(PPVs)[36]are usually characterized by low intermolecular cou-
pling, and therefore limited hopping transport between mole-
cules.[37]Therefore, it would be desirable to form a more elabo-
rate heterojunction consisting of charge transport layers and a
light-emitting layer, which can be each optimized separately.
Such heterojunction concepts have been used in high-perfor-
mance OLEDs.[38,39]However, this concept has been difficult
to apply to the above-mentioned LEOFET device architec-
tures.
In this paper, a novel two-electrode light-emitting device
structure is proposed. The device is a hybrid structure incorpo-
rating aspects of both a diode and a field-effect transistor.
Compared to conventional OLEDs, the cathode is displaced by
one to several micrometers from the light-emission zone. Since
the light-emission zone is not covered by metal, the device can
be used for top emission or even as a waveguide. The microme-
ter-sized distance between the cathode and the active region
can be bridged by electrons with an enhanced field-effect mo-
bility. Owing to this high charge carrier mobility, large current
densities are achievable. The external quantum efficiency at
these high current densities is as high as that of conventional
OLEDs comprising the same materials. In contrast to
LEOFETs, only two electrodes are used in our architecture.
Moreover, light emission in our novel device structure always
occurs at a fixed position, irrespective of the applied bias, in
contrast to the situation in LEOFETs where the emission zone
can be moved within the channel by varying the bias condi-
tions.[30,31]
2. Results and Discussion
2.1. Device Fabrication
The schematic architecture of the device is illustrated in Fig-
ure 1. The device has been fabricated on top of an ITO-coated
glass substrate and comprises an organic hole-transport layer
(HTL), an organic light-emitting layer, and an organic elec-
tron-transport layer (ETL). Prior to the deposition of these or-
ganic layers, an insulating layer of SiO2has been deposited by
sputtering, covering only part of the ITO surface. ITO serves
as the hole-injecting electrode in the device. The cathode has
been formed by the deposition of a thin layer of 0.6 nm LiF fol-
lowed by the deposition of 100 nm Al. This cathode has been
Adv. Funct. Mater. 2008, 18, 136–144© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.afm-journal.de137
Figure 1. Schematic architecture of the OLED exhibiting field-effect elec-
tron transport. The device comprises an organic hole-transport layer, an
organic electron-transport layer, and an organic light-emitting layer. The
dashed lines AA′, BB′, and CC′C″C? indicate three different cross-sctions.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 3

not positioned vertically above the ITO anode in the trench of
the SiO2insulator. Instead, it has been located directly above
the insulator, as shown in Figure 1. The distance between the
metallic top electrode and the insulator edge ranges from one
to several micrometers. Accurate alignment of the metallic
cathode has been achieved by using an integrated shadow mask
technique[40]and the angled deposition of the LiF/Al layer
(Fig. 2).
The integrated shadow mask has been obtained by patterning
a 20 lm thick negative photoresistSU8-25. The different organ-
ic layershave beendepositedafter application andpatterningof
the SU8-25 resist. After deposition of the organic layers, the
sample is mounted on a triangular sample holder and loaded in
an ultra high vacuum system to evaporate the cathode. During
this deposition process the flux has been maintained at a 45° an-
glewithrespecttothesubstrate.TheSU8-25 profilethuscreates
a shadowed region with a span similar to the thickness of the
SU8-25 such that the substrateis only partially covered with the
metal. Figure 3 shows a scanning electron microscopy (SEM)
image of the device structure. The shadowed region where no
metal has been deposited is clearly discernible. It is also clearly
apparent that the walls of the SU8-25 layer are slightly re-en-
trant (i.e., negatively sloped), which is typical for negative
photoresists.
The molecular structures of the organic materials used to
fabricate the device are shown in Figure 4. Poly(triarylamine)
(PTAA)[41]has been selected as the hole-transport material.
PTAA can be deposited by spin-coating, which yields a film
with a smooth top surface that is favorable for the deposition
of an additional layer. The bulk hole mobility of
PTAA is approximately 10–2cm2V–1s–1.[41]The ac-
tive light-emitting layer is formed by tris(8-hydroxy-
quinoline) aluminum (Alq3) doped with 2% of
4-(dicyanomethylene)-2-methyl-6-(julolindin-4-yl-vi-
nyl)-4H-pyran (DCM2). DCM2is a well-known red-
light-emitting fluorescent dye used in OLEDs.[42]
The photoluminescence quantum yield of this dye
family is about 40%.[43]The electron-transport ma-
terial has been chosen based on the two criteria
that are important in our device architecture.
First, the organic electron-transport material should
have a high electron field-effect mobility since
this mobility determines the performance of the de-
vice. Secondly, since the electron transport is in-
tended to occur at the heterojunction between the
ETL and the light-emitting layer, the lowest unoc-
cupied molecular orbital (LUMO) of the electron-
transport material should be slightly lower than
the LUMO of Alq3 and DCM2. N,N′-ditridecyl-
perylene-3,4,9,10-tetracarboxylic diimide (PTCDI-
C13H27) satisfies these two conditions and has there-
fore been used as the electron-transport material.
Gundlach et al. have reported a field-effect mobility
of about 0.6 cm2V–1s–1for PTCDI-C13H27using Cr
top-contacts.[44]A mobility of 0.28 cm2V–1s–1has
been obtained using LiF/Al top-contacts. Even
higher field-effect mobilities can be achieved by an-
nealing after fabrication.[45]The energy levels of
the highest occupied molecular orbital (HOMO)
and LUMO of PTAA,[41]Alq3,[46]DCM2,[47]and
PTCDI-C13H27
are schematically
Figure 4e.
[48]
depicted in
138www.afm-journal.de
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2008, 18, 136–144
SU8-25
Figure 2. Schematic depiction of the processing sequence for the fabrication of the
light-emitting field-effect device structure: a) situation before creation of the inte-
grated shadow mask, b) deposition of the integrated shadow mask, c) deposition of
the organic layers, and d) deposition of the metallic cathode with the substrate
mounted on a triangular sample holder. During this deposition process, the atomic
flux is at a 45° angle with respect to the substrate.
Figure 3. Scanning electron microscopy image of the device structure.
a)Metalliccathode,b)shadowedregionwherenometalisdeposited,c)the
20 lmthickSU8-25 layer,andd)arrowindicatingtheinsulatoredge.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 4

2.2. Device Operation
To explain the device operation, in Figure 5 we show the
band diagrams through sections AA′, BB′, and CC′C″C? of
Figure 1 for the device in forward bias, i.e., for a positive an-
ode-to-cathode bias. Under these conditions, electrons are in-
jected from the cathode into PTCDI-C13H27at the cross-sec-
tion AA′. The semiconductor PTAA and the light-emitting
layer Alq3:DCM2on top of the SiO2(cross-section AA′) are
depleted of holes. As further discussed below, the LUMO off-
set at the interface between Alq3and PTCDI-C13H27prohibits
electron injection from PTCDI-C13H27into Alq3in the AA′
cross-section. As a result, the depleted PTAA and Alq3:DCM2
layers on top of the SiO2insulator behave as a combined di-
electric layer in series with the SiO2insulator, and an electron
accumulation layer is formed in the PTCDI-C13H27layer at the
interface with Alq3:DCM2. These electrons are transported lat-
erally by the electric field from the cathode towards the SiO2
edge (from C′ to C″). The field-effect mobility of these charge
carriers is larger than the charge carrier mobility of a conven-
tional OLED, making transport over several micrometers pos-
sible. Near the insulator edge, at position C″, electrons are in-
jected into the Alq3:DCM2layer, where they recombine with
holes that are injected from the ITO and transported vertically
through the PTAA (see cross section BB′ in Fig. 5). The elec-
tron injection at position C″ into Alq3is enabled by an en-
hanced vertical electric field at that position. The enhancement
of the vertical field at C″ has two different origins: first, the ver-
tical and lateral electric fields in the channel (section C′C″) col-
lapse into only a vertical field at position C″; furthermore, the
presence of a space charge of holes in PTAA increases the ver-
tical electric field that attracts electrons to be injected in the
Alq3:DCM2layer. This latter behavior is similar to the en-
hanced hole injection as a result of the space charge of elec-
trons at the anode of OLEDs reported by Van Woudenbergh
et al.[49]Exciton formation occurs and radiative relaxation of
these excitons to the ground state results in light emission near
the SiO2edge, at a micrometer-sized distance from the metallic
contact. As a result, this heterojunction device allows the
minimization of optical losses at the metal cathode. The vicini-
ty of the ITO bottom-contact is of minor importance with re-
Adv. Funct. Mater. 2008, 18, 136–144 © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.afm-journal.de 139
Figure 4. Molecular structure of the organic materials used to fabricate
the light-emitting device: a) PTAA, b) PTCDI-C13H27, c) Alq3, and d) DCM2.
e) Energy level diagram for these materials. The dotted lines indicate the
highest occupied molecular orbital (HOMO) and the lowest unoccupied
molecular orbital (LUMO) positions of DCM2.
Figure 5. Band diagrams for the device under forward bias conditions
through different sections of Fig. 1: a) AA′, b) BB′, and c) CC′C″C? cross
sections.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 5

spect to absorption losses because it has been shown that thin
transparent ITO layers have low optical losses in the
550–750 nm spectral range.[13]
We have confirmed this view of the device operation by mea-
suring the characteristics of several control devices. We have
fabricated control transistors comprising the stacks PTAA/
PTCDI-C13H27 and PTAA/Alq3:DCM2/PTCDI-C13H27 on a
100 nm thick SiO2dielectric using LiF/Al source and drain
top-electrodes. For the PTAA/PTCDI-C13H27 transistor, we
have extracted a field-effect mobility of 0.2 cm2V–1s–1for elec-
trons in PTCDI-C13H27at the heterojunction interface with
PTAA. It is striking that the hole-transporting organic semi-
conductor PTAA provides a good growth surface for the elec-
tron-conducting PTCDI-C13H27, and that the hetero-interface
is appropriate for electron transport. This can be attributed to
the smooth top surface of PTAA and the formation of a high-
quality interface free of electron traps. We infer the latter from
the observation that the electron mobility in PTCDI-C13H27
transistors on PTAA-coated SiO2is the same as that measured
in control transistors that have a gate dielectric consisting of
SiO2 coated with poly-a-methylstyrene (PaMS). PaMS is
known to provide a high-quality, electron-trap-free surface[50]
that allows good electron transport.[51]
Upon the insertion of Alq3:DCM2 between PTAA and
PTCDI-C13H27, the effective saturation field-effect electron
mobility drops to 0.08 cm2V–1s–1. The output and transfer
characteristics of this transistor are shown in Figure 6. We have
correlated the reduction in mobility to the different growth be-
havior of PTCDI-C13H27 on Alq3:DCM2, as compared to
growth on a PTAA surface. Figure 7a and b show the morphol-
ogy of the PTCDI-C13H27layer upon deposition onto PTAA
and PTAA/Alq3:DCM2 surfaces. PTCDI-C13H27 deposited
on PTAA/Alq3:DCM2 reveals a much rougher topography
than PTCDI-C13H27grown on top of PTAA; this correlates
well with the lower field-effect mobility. The root mean
square (rms) roughness of the PTCDI-C13H27layer is increased
from 6.7 to 10.7 nm upon the insertion of an Alq3:DCM2layer.
Since both substrates (PTAA and PTAA/Alq3:DCM2) are
morphologically similar (both are amorphous flat surfaces
with the same rms roughness of 0.4 nm), the observed differ-
ences in the growth behavior of PTCDI-C13H27likely originate
from interfacial energy differences between the PTCDI-
C13H27and PTAA and PTCDI-C13H27and Alq3:DCM2combi-
nations.
Additionally, a control transistor has been fabricated com-
prising PTAA/Alq3:DCM2 as the organic layers, to verify
whether Alq3:DCM2 can be used as the ETL instead of
PTCDI-C13H27. No field-effect transport of electrons has been
observed in this structure. The non-planar molecular structure
of Alq3apparently prohibits efficient lateral electron transport.
This control experiment also proves that electron transport in
the transistor comprising the organic stack PTAA/Alq3:DCM2/
PTCDI-C13H27 and in our novel device structure (between
positions C′ and C″ in Fig. 1) indeed occurs through PTCDI-
C13H27and not via Alq3:DCM2.
2.3. Optical and Electrical Characterization
Figure 8a displays the experimentally measured electrical
characteristics of a light-emitting device constructed using the
architecture depicted in Figure 1. The device has a width of
1 mm and has been measured under an inert N2atmosphere
140www.afm-journal.de
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2008, 18, 136–144
Figure 6. Electrical characteristics of a top-contact transistor comprising
PTAA/Alq3:DCM2/PTCDI-C13H27as the organic layers (W/L = 2000/50).
a) Output characteristics for various gate voltages, and b) transfer charac-
teristics of the same transistor. A saturation mobility of 0.08 cm2V–1s–1
and a threshold voltage of ca. 0 Vare extracted for this device.
Figure 7. 2 lm × 2 lm atomic force microscopy surface scans of 50 nm
thick PTCDI-C13H27layers on top of a) PTAA and b) PTAA/Alq3:DCM2. A
higher surface roughness is measured for PTAA/Alq3:DCM2. The rms
roughness is 10.7 nm for PTCDI-C13H27 on PTAA/Alq3:DCM2 versus
6.7 nm for PTCDI-C13H27grown on PTAA.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 6

immediately after evaporation. The metallic cathode of the de-
vice is displaced by 3.5 lm with respect to the insulator edge.
Under forward bias, the current increases with increasing volt-
age. The electrical characteristics can be explained by the
equivalent circuit shown in the inset of Figure 8b. This equiva-
lent circuit consists of an n-type transistor with a diode be-
tween the gate and the drain.
The device emits red light upon the radiative decay of the ex-
citons. The optical output intensity as a function of the applied
voltage bias is shown in Figure 8b. It can be seen that the opti-
cal output tracks the current characteristics. The normalized
electroluminescence spectrum of the device is shown in Fig-
ure 9. The peak emission wavelength is located at 636 nm and
corresponds to emission from DCM2. Two control experiments
have enabled us to verify that the observed emission indeed
originates from DCM2. In the first control experiment, we have
changed the dye from DCM2to Btp2Ir(acac),[52]a phosphores-
cent dye with characteristic spectral features that can be easily
be recognized. In another control experiment, we have used
undoped Alq3; this device emits green light, peaking at
540 nm, as is characteristic for emission from Alq3.
In our device structure, the emission zone is well defined and
light emission always occurs at a fixed position, namely near
the edge of the insulator. This position is independent of the
applied bias. The left panel of Figure 10 shows a photograph in
reflection of a device without biasing. The metal electrode
(white reflecting area) and the insulator edge, indicated by an
arrow, are easily discernible. For this device, the distance be-
tween the cathode and the insulator edge is 19 lm. The right
panel of Figure 10 shows an image of the optical output in the
dark when the device is in forward bias. A narrow line of red
light appears alongside the edge of the insulator. The light in-
tensity increases with the bias voltage. Despite the narrow
emission zone, the red light can easily be observed by the
naked eye. The measured width of the emission zone is about
2 lm.
Adv. Funct. Mater. 2008, 18, 136–144© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.afm-journal.de 141
Figure 8. Experimentally measured characteristics of the light-emitting
diode with field-effect electron transport. This device has a width of 1 mm
and a distance between top electrode and insulator edge of 3.5 lm.
a) Current–voltage characteristics, b) corresponding light output. The in-
set shows the equivalent circuit of this device. c) External quantum effi-
ciency of the device as a function of the current.
Figure 9. Normalized electroluminescence spectrum of the light-emitting
diode with field-effect electron transport. The spectrum corresponds to
DCM2emission with a maximum positioned at wavelength k = 636 nm.
Figure 10. Left panel: Optical microscopy reflection image of a device
without bias. The white area is the reflective metal cathode, the arrow indi-
cates the insulator edge. Right panel: Optical microscopy image under for-
ward bias. A narrow line of light appears along the insulator edge. The
width of the line is estimated to be about 2 lm.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 7

2.4. Analysis
The experimentally observed behavior has been qualitatively
verified by numerical simulations. A simplified device, com-
prising a HTL and a light-emitting ETL, has been simulated
using the Silvaco Atlas 2D device simulator. The material pa-
rameters used in the simulations are summarized in Table 1.
For the ETL, an electron mobility of 8 × 10–2cm2V–1s–1has
been used, which corresponds to the experimentally measured
electron field-effect mobility of PTCDI-C13H27in our device.
The hole mobility of the ETL is taken to be 5 × 10–4cm2V–1
s–1. On the other hand, for the HTL, a hole mobility of 5 × 10–3
cm2V–1s–1and a negligible electron mobility are assumed. The
HOMO and LUMO of the ETL and HTL are assumed to be
similar to the HOMO and LUMO of PTCDI-C13H27 and
PTAA, respectively. Typical values for the singlet lifetime (s)
and exciton diffusion length (Ldiff) have been obtained from
the literature.[53]The top contact is assumed to be located 6 lm
from the insulator edge. Figure 11a illustrates the simulated re-
combination zone as a result of applying a positive bias to the
anode. It can be clearly seen that recombination occurs near
the insulator edge, several micrometers away from the metallic
contact. The simulated recombination zone has a width of
about 2 lm, which is consistent with the experimental observa-
tions. Figure 11b confirms the presence of an electron accumu-
lation layer at the interface between the HTL and the light-
emitting ETL when a positive bias is applied to the anode with
respect to the cathode. This electron accumulation region
vanishes beyond the insulator edge, since electrons and holes
recombine there.
The external quantum efficiency of the fabricated devices
has been estimated based on the luminance, electrolumines-
cence spectra, and current. The measured maximum external
quantum efficiency of the device is 0.02%, which is similar to
that of a light-emitting diode comprising the same materials.
Reference OLEDs with Alq3:DCM2active layers have been
demonstrated with up to 0.5% efficiency.[42]In these reference
OLEDs, Alq3is used as the ETL instead of PTCDI-C13H27.
However, we have not been able to use Alq3in our structure
because it does not conduct electrons in the thin-film transistor
configuration. Further optimization of material combinations
is needed to increase the external quantum efficiency.
The light intensity can be tuned by changing the distance be-
tween the metallic contact and the insulator edge. We have
measured different devices for which the top-contact displace-
ment with respect to the insulator edge is systematically varied
from 0.8 to 9 lm. In Figure 12a the measured maximum exter-
nal quantum efficiency of each of these devices is plotted as a
function of the distance between the cathode and the insulator
edge. As expected, the maximum external quantum efficiency
remains constant. On the other hand, at the maximum break-
down voltage of the SiO2insulator (about 30 V), higher cur-
rents are possible in devices with shorter cathode displace-
ments with respect to the insulator edge. As a result, the
maximum achievable optical output power and brightness can
be higher for smaller distances. This is illustrated in Figure 12b,
which shows the light intensity of three different devices where
the distances between the metallic cathode and the insulator
edge are 7.3, 5.1, and 3.5 lm.
142www.afm-journal.de
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2008, 18, 136–144
Table 1. Summary of the parameters used in the 2D device simulator.
HTLLight-emitting ETL
HOMO [eV]
LUMO [eV]
le[cm2V–1s–1]
lh[cm2V–1s–1]
Ldiff[nm]
s [s]
5.1
1.8
5.4
3.4
1 × 10–7
5 × 10–3
8 × 10–2
5 × 10–4
10
16 × 10–9
Figure 11. a) The simulated recombination rate and b) the simulated electron accumulation layer under forward bias.
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 8

Taking into account a 2 lm wide emission zone, the maxi-
mum hole current density in the device at the breakdown volt-
age of 30 V is estimated to be 13 A cm–2. This current density
is much higher than the current density obtained for conven-
tional OLEDs (typically in the order of 10–2A cm–2at the point
of maximum external quantum efficiency). The maximum elec-
tron current density in the accumulation layer of the light-emit-
ting device is even higher. Assuming that the current flow is
confined to a 1.5 nm thick electron accumulation layer, current
densities of 1800 A cm–2are achieved. This is significantly
higher than the current densities achieved in ambipolar organic
light-emitting transistors reported up till now;[30,31]the im-
proved performance arises from the higher electron mobility of
our device. Figure 8c shows the measured external quantum
efficiency as a function of the current. No significant roll-off is
observed within experimental error up to the maximum cur-
rent.
3. Conclusions
We have realized an OLED with field-effect electron trans-
port. In our device configuration, the metallic top-contact is re-
mote from the light-emission zone. The micrometer-sized dis-
tance between the cathode and the light-emission zone is
bridged by electrons with an enhanced field-effect mobility.
The light-emission occurs at a fixed position irrespective of the
applied bias. We have fabricated this device using Alq3:DCM2
as a host–guest light-emitting material system, PTCDI-C13H27
as the electron-transport material, and PTAA as the hole-
transport material. Light-emission is correlated with the cur-
rent, and can be modulated by the anode-to-cathode voltage.
The measured device operation has been verified by 2D nu-
merical simulations and by the characteristics of a number of
control devices. Devices with smaller distances between the
cathode and the insulator edge allow for larger currents at the
maximum operating voltage, and therefore also yield higher
brightness. The external quantum efficiency has been con-
firmed to be as high as in a conventional OLED comprising
the same materials. The quantum efficiency is remarkably con-
stant up to the maximum current, which corresponds to a hole
current density on the order of 10 A cm–2. This high current
density, in combination with reduced optical absorption losses
arising from the remoteness of the metal cathode, may lead to
interesting applications as waveguide OLEDs and possibly a
laser structure.
4. Experimental
Materials: PTCDI-C13H27and Alq3were purchased from Aldrich
and purified once by sublimation before loading into an ultra high vac-
uum system (p = 10–8torr (1 torr=133.3Pa)). PTAA, received from
Merck, was used as received. DCM2was purchased from H. W. Sands
and also used without further purification. The integrated shadow mask
was realized using SU8-25, which was purchased from MicroChem and
used as received.
Device Processing: The devices were fabricated on ITO-coated glass
substrates with a sheet resistance of ca. 16 X square–1. Prior to the de-
position of the organic layers, a ca. 100 nm thick insulating SiO2layer
was deposited on top of ITO by sputtering in a Pfeiffer Spider 630 in-
strument. The trenches were selectively wet-etched through the SiO2to
contact the ITO layer, which serves as the hole-injecting electrode.
Next, a single layer of negative, epoxy-type, near-UV photoresist SU8-
25 was patterned by photolithography to obtain a ca. 20 lm thick layer.
This layer was resistant to solvents, acids, and bases and was character-
ized by excellent thermal stability. Subsequently, the substrate was
cleaned with solvents (acetone and isopropanol) and exposed to an
UV-ozone ambient for 15 min. Then, PTAA was spin-coated at
1000 rpm on top of the patterned structure. The sample was baked on a
hotplate at 110°C for 20 min after spin-coating to evaporate the sol-
vent. In the next step, a 20 nm thick light-emitting layer was deposited
under ultra high vacuum (p = 10–8
100:2 (mass ratio) Alq3and DCM2. Subsequently, 50 nm of PTCDI-
C13H27was evaporated as the ETL. The deposition rate during evapo-
ration was maintained at 0.5 Å s–1, and the substrate was kept at room
temperature. The sample was then mounted on a triangular sample
holder followed by the evaporation of 0.8 nm LiF and 100 nm Al. Dur-
ing this deposition process, the flux was maintained at an angle of 45°
with respect to the substrate. In this way, the SU8-25 profile created a
shadowed region where no LiF and Al were deposited. The shadowed
region had a span of about 20 lm, which was similar to the thickness of
the SU8-25 layer. All processing steps after UV-ozone treatment were
carried out in a dry nitrogen glovebox (<1 ppm O2, <5 ppm H2O) or in
ultra high vacuum. The glovebox and ultra high vacuum systems were
attached to each other. Therefore, samples could be transported from
high vacuum to the glovebox and vice versa without exposure to ambi-
ent atmosphere.
Device Characterization: The devices were characterized immedi-
ately after evaporation in an inert N2atmosphere. The electrical char-
acteristics were measured using an Agilent 4156C parameter analyzer.
A calibrated integrated sphere (SphereOptics Hoffman GmbH) was
torr) by co-evaporation of
Adv. Funct. Mater. 2008, 18, 136–144© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.afm-journal.de143
Figure 12. a) The maximum external quantum efficiency as a function of
the distance from the insulator edge. The maximum external quantum effi-
ciency is constant irrespective of the distance. b) Light intensity versus ap-
plied voltage of three different devices for which the top-contact displace-
ment with respect to the insulator edge is 7.3 lm (dotted lines), 5.1 lm
(dashed lines), and 3.5 lm (solid lines).
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport

Page 9

used for light-intensity measurements and external quantum efficiency
calculations. To determine the spectral characteristics, the emitted light
was detected by means of an optical multichannel analyzer (OMA) in
conjunction with a charge coupled device (CCD). These measurements
were performed at room temperature in a cryostat to prevent photo-
oxidation.
Received: July 13, 2007
Revised: August 31, 2007
Published online: December 18, 2007
–
[1] C. W. Tang, S. A. VanSlyke, Appl. Phys. Lett. 1987, 51, 913.
[2] L. S. Hung, C. H. Chen, Mater. Sci. Eng. R 2002, 39, 143.
[3] M. Pope, C. E. Swenberg, Electronic Processes in Organic Crystals
and Polymers, Vol.1, 2nd ed., Oxford University Press, New York
1999, Ch.2.
[4] P. W. M. Blom, M. J. M. de Jong, IEEE J. Quantum Electron. 1998, 4,
105.
[5] C. Hosokawa, H. Tokailin, H. Higashi, T. Kusumoto, Appl. Phys. Lett.
1992, 60, 1220.
[6] I. D. Parker, J. Appl. Phys. 1994, 75, 1656.
[7] P. Andrew, G. A. Turnbull, I. D. W. Samuel, W. L. Barnes, Appl. Phys.
Lett. 2002, 81, 954.
[8] M. Reufer, S. Riechel, J. M. Lupton, J. Feldmann, U. Lemmer,
D. Schneider, T. Benstem, T. Dobbertin, W. Kowalsky, A. Gombert,
K. Forberich, K. Wittwer, U. Scherf, Appl. Phys. Lett. 2004, 84, 3262.
[9] G. Parthasarathy, P. E. Burrows, V. Khalfin, V. G. Kozlov, S. R. For-
rest, Appl. Phys. Lett. 1998, 72, 2138.
[10] H. Kim, C. M. Gilmore, J. S. Horwitz, A. Piqué, H. Murata, G. P.
Kushto, R. Schlaf, Z. H. Kafafi, D. B. Chrisey, Appl. Phys. Lett. 2000,
76, 259.
[11] H. Yamamoto, T. Oyamada, H. Sasabe, C. Adachi, Appl. Phys. Lett.
2004, 84, 1401.
[12] P. Görrn, T. Rabe, T. Riedl, W. Kowalsky, G. Galbrecht, U. Scherf,
Appl. Phys. Lett. 2006, 89, 161 113.
[13] V. G. Kozlov, G. Parthasarathy, P. E. Burrows, V. B. Khalfin, J. Wang,
S. Y. Chou, S. R. Forrest, IEEE J. Quantum Electron. 2000, 36, 18.
[14] M. Reufer, J. Feldmann, P. Rudati, A. Ruhi, D. Müller, K. Meerholz,
C. Karnusch, M. Gerken, U. Lemmer, Appl. Phys. Lett. 2005, 86, 221
102.
[15] F. Dinelli, M. Murgia, P. Levy, M. Cavallini, F. Biscarini, D. M. de
Leeuw, Phys. Rev. Lett. 2004, 92, 116 802.
[16] C. Tanase, E. J. Meijer, P. W. M. Blom, D. M. de Leeuw, Phys. Rev.
Lett. 2003, 91, 216 601.
[17] A. Dodabalapur, H. E. Katz, L. Torsi, Adv. Mater. 1996, 8, 853.
[18] A. Hepp, H. Heil, W. Weise, M. Ahles, R. Schmechel, H. von Seggern,
Phys. Rev. Lett. 2003, 91, 157 406.
[19] M. Ahles, A. Hepp, R. Schmechel, H. von Seggern, Appl. Phys. Lett.
2004, 84, 428.
[20] J. Swensen, D. Moses, A. J. Heeger, Synth. Met. 2005, 153, 53.
[21] T. Oyamada, H. Uchiuzou, S. Akiyama, Y. Oku, N. Shimoji, K. Mat-
sushige, H. Sasabe, C. Adachi, J. Appl. Phys. 2005, 98, 074 506.
[22] J. Reynaert, D. Cheyns, D. Janssen, R. Müller, V. I. Arkhipov, J. Gen-
oe, G. Borghs, P. Heremans, J. Appl. Phys. 2005, 97, 114 501.
[23] C. Santato, R. Capelli, M. A. Loi, M. Murgia, F. Cicoira, V. A. L.
Roy, P. Stallinga, R. Zamboni, C. Rost, S. F. Karg, M. Muccini, Synth.
Met. 2004, 146, 329.
[24] C. Rost , S. Karg, W. Rieß, M. A. Loi, M. Murgia, M. Muccini, Appl.
Phys. Lett. 2004, 85, 1613.
[25] C. Rost , S. Karg, W. Rieß, M. A. Loi, M. Murgia, M. Muccini, Synth.
Met. 2004, 146, 237.
[26] M. A. Loi, C. Rost, M. Murgia, S. Karg, W. Rieß, M. Muccini, Adv.
Funct. Mater. 2006, 16, 41.
[27] F. Dinelli, R. Capelli, M. A. Loi, M. Muccini, A. Fachetti, T. J. Marks,
Adv. Mater. 2006, 18, 1416.
[28] R. Capelli, F. Dinelli, M. A. Loi, M. Murgia, R. Zamboni, M. Muccini,
J. Phys.: Condens. Matter 2006, 18, S2127.
[29] S. De Vusser, S. Schols, S. Steudel, S. Verlaak, J. Genoe, W. D. Ooster-
baan, L. Lutsen, D. Vandezande, P. Heremans, Appl. Phys. Lett. 2006,
89, 223 504.
[30] J. Zaumseil, R. H. Friend, H. Sirringhaus, Nat. Mater. 2006, 5, 69.
[31] J. Zaumseil, C. L. Donley, J. Kim, H. Friend, H. Sirringhaus, Adv.
Mater. 2006, 18, 2708.
[32] M. Muccini, Nat. Mater. 2006, 5, 605.
[33] T. Oyamada, H. Uchiuzou, S. Akiyama, Y. Oku, N. Shimoji, K. Mat-
sushige, H. Sasabe, C. Adachi, J. Appl. Phys. 2005, 98, 074 506.
[34] Y. Y. Lin, D. J. Gundlach, S. F. Nelson, T. N. Jackson, IEEE Electron
Device Lett. 1997, 18, 606.
[35] H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K. Bech-
gaard, B. M. W. Langeveld-Voss, A. J. H. Spiering, R. A. J. Janssen,
E. W. Meijer, P. Herwig, D. M. de Leeuw, Nature 1999, 401, 685.
[36] I. D. W. Samuel, G. Rumbles, C. J. Collison, R. H. Friend, S. C. Mor-
atti, A. B. Holmes, Synth. Met. 1997, 84, 497.
[37] M. Muratsubaki, Y. Furukawa, T. Noguchi, T. Ohnishi, E. Fujiwara,
H. Tada, Chem. Lett. 2004, 33, 1480.
[38] J. Kido, Y. Lizumi, Appl. Phys. Lett. 1998, 73, 2721.
[39] C. Adachi, M. A. Baldo, M. E. Thompson, S. R. Forrest, J. Appl.
Phys. 2001, 90, 5048.
[40] S. De Vusser, S. Steudel, K. Mymy, D. Janssen, S. De Jonge, J. Genoe,
P. Heremans, Appl. Phys. Lett. 2006, 88, 103 501.
[41] J. Veres, S. Ogier, S. Leeming, B. Brown, D. Cupertino, Mater. Res.
Soc. Symp. Proc. 2002, 708, BB8.7.1.
[42] V. Bulovic, A. Shoustikov, M. A. Baldo, E. Bose, V. G. Kozlov, M. E.
Thompson, S. R. Forrest, Chem. Phys. Lett. 1998, 287, 455.
[43] C. W. Tang, S. A. VanSlyke, C. H. Chen, J. Appl. Phys. 1989, 65, 3610.
[44] D. J. Gundlach, K. P. Pernstich, G. Wilckens, M. Grüter, S. Haas, B.
Batlogg, J. Appl. Phys. 2005, 98, 064 502.
[45] S. Tatemichi, M. Ichikawa, T. Koyama, Y. Taniguchi, Appl. Phys. Lett.
2006, 89, 112 108.
[46] M. A. Baldo, M. E. Thompson, S. R. Forrest, Pure Appl. Chem. 1999,
71, 2095.
[47] Y. Hamada, H. Kanno, T. Tsujioka, H. Takahashi, T. Usuki, Appl.
Phys. Lett. 1999, 75, 1682.
[48] C. Rost, D. J. Gundlach, S. Karg, W. Rieß, J. Appl. Phys. 2004, 95,
5782.
[49] T. van Woudenbergh, P. W. M. Blom, J. N. Huiberts, Appl. Phys. Lett.
2003, 82, 985.
[50] T. W. Kelley, D. V. Muyres, P. F. Baude, T. P. Smith, T. D. Jones,
Mater. Res. Soc. Symp. Proc. 2003, 771, L6.5.1.
[51] L. Chua, J. Zaumseil, J. Chang, E. C. Ou, P. K. Ho, H. Sirringhaus,
R. H. Friend, Nature 2005, 434, 194.
[52] S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razzaq, H. Lee, C.
Adachi, P. E. Burrows, S. R. Forrest, M. E. Thompson, J. Am. Chem.
Soc. 2001, 123, 4304.
[53] I. Sokolik, R. Priestley, A. D. Walser, R. Dorsinville, C. W. Tang,
Appl. Phys. Lett. 1996, 69, 4168.
______________________
144www.afm-journal.de
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2008, 18, 136–144
FULL PAPER
S. Schols et al./An OLED with Field-Effect Transport