nmat978PRINT
10/9/03
3:27 pm
Page 678
ARTICLES
Solution-processed ambipolar organic
field-effect transistors and inverters
E. J. MEIJER1,2, D. M. DE LEEUW*1, S. SETAYESH1, E. VAN VEENENDAAL1, B.-H. HUISMAN1,
P. W. M. BLOM3, J. C. HUMMELEN3, U. SCHERF4 AND T. M. KLAPWIJK2
1
Philips Research Laboratories,Professor Holstlaan 4,5656 AA Eindhoven,The Netherlands
Delft University of Technology,Faculty of Applied sciences,Department of NanoScience,Lorentzweg 1,2628 CJ Delft,The Netherlands
3
Materials Science Center,University of Groningen,Nijenborgh 4,9747 AG Groningen,The Netherlands
4
Bergische Universitat Wuppertal,Department of Chemistry,Gauss-strasse 20,D-42097 Wuppertal,Germany
*e-mail: dago.de.leeuw@philips.com
2
Published online: 21 September 2003; doi:10.1038/nmat978
There is ample evidence that organic field-effect
transistors have reached a stage where they can be
industrialized, analogous to standard metal oxide
semiconductor (MOS) transistors. Monocrystalline silicon
technology is largely based on complementary MOS
(CMOS) structures that use both n-type and p-type
transistor channels. This complementary technology has
enabled the construction of digital circuits, which operate
with a high robustness, low power dissipation and a good
noise margin. For the design of efficient organic integrated
circuits, there is an urgent need for complementary
technology, where both n-type and p-type transistor
operation is realized in a single layer, while maintaining the
attractiveness of easy solution processing. We demonstrate,
by using solution-processed field-effect transistors, that
hole transport and electron transport are both generic
properties of organic semiconductors. This ambipolar
transport
is
observed
in
polymers
based
on
interpenetrating networks as well as in narrow bandgap
organic semiconductors. We combine the organic ambipolar
transistors into functional CMOS-like inverters.
T
he main difficulty in achieving ambipolar transistor operation is
the injection of holes and electrons into a single semiconductor
from the same electrode. This electrode needs to have a
workfunction that allows injection of holes in the highest occupied
molecular orbital (HOMO) of the semiconductor, and the injection of
electrons in the lowest unoccupied molecular orbital (LUMO).
Consequently,this will result in an injection barrier of at least half of the
bandgap energy for one of the carriers. To circumvent injection
problems due to this barrier, in double-carrier devices, such as lightemitting diodes1 and photovoltaic cells2, two different electrodes are
used to allow injection or collection of holes and electrons.Alternatively,
one-electrode material can be used in combination with two different
semiconductors,where one has its HOMO level and the other its LUMO
level aligned with the metal work function. The semiconductors can be
spatially separated by evaporation techniques to form discrete n-type
and p-type transistors. A summary of the state-of-the-art mobilities of
discrete n-type and p-type organic transistors is presented in ref. 3.
Discrete n-type and p-type transistors have been integrated into CMOS
logic circuits4. Moreover, the two semiconductors can also be
evaporated on top of each other, resulting in a heterostructure that
exhibits ambipolar transistor behaviour5,6. However, to fully exploit the
advantageous properties of organic molecules, such as ease of
processing leading to low-cost, high-volume, large-area applications7–9,
solution processing is required. Here we demonstrate ambipolar
transistors using heterogeneous blends, consisting of interpenetrating
networks of p-type and n-type semiconductors, as suggested by
Tada et al.10. The distinction between n-type and p-type organic
semiconductors, however, is artificial because it depends on the
position of the work function of the electrode relative to the HOMO
and LUMO energies of the organic semiconductor. Ambipolar charge
transport is an intrinsic property of pure organic semiconductors.
We show that ambipolar transport is observed experimentally on
reduction of the injection barriers by using low bandgap
semiconductors. The transistors have been combined into CMOS-like
inverters. An analytical model is presented that describes the details of
the inverter transfer characteristics.
nature materials | VOL 2 | OCTOBER 2003 | www.nature.com/naturematerials
678
©2003 Nature Publishing Group
©2003 Nature Publishing Group
nmat978PRINT
10/9/03
3:28 pm
Page 679
ARTICLES
2.8 eV LUMO OC1C10-PPV
3.7 eV LUMO PCBM
φB
PCBM
5.1 eV
Au
OC1C10-PPV
OC1C10-PPV : PCBM
Source (Au)
Drain (Au)
Insulator (SiO2)
n
Gate (N++-Si)
Figure 1 Schematic cross-section of the FET geometry used in this study,the
molecular structures of PCBM and OC1C10-PPV,and an artist’s impression of
the interpenetrating networks of the two semiconductors.
Device fabrication is described in the Methods section.
A representation of the interpenetrating network and cross-section of
the field-effect transistor are presented in Fig. 1.The energy levels for the
blend of [6,6]-phenyl C61-butyric acid methyl ester (PCBM) and
poly[2-methoxy-5-(3′,7′-dimethyloctyloxy)]-p-phenylene vinylene
(OC1C10-PPV) are represented in Fig. 2.For simplicity,the energy levels
are drawn as straight lines, but band bending occurs at the
semiconductor/insulator interface upon an applied gate bias.This shifts
the Fermi level in the semiconductor, which in turn can result in band
bending at the electrode/semiconductor interface. The HOMO level of
OC1C10-PPV, at 5.0 eV, is aligned with the work function of gold, at
5.1 eV,which will result in an ohmic contact for hole injection from gold
into the OC1C10-PPV network. Owing to the large bandgap of OC1C10PPV, gold is a blocking contact for electrons11 into OC1C10-PPV. The
alignment of the gold work function with the LUMO level of the PCBM
is not as good, and the mismatch in energy levels results in an injection
barrier φB of 1.4 eV for electron injection into the PCBM network.
However, this injection barrier can be significantly reduced to 0.76 eV,
due to the formation of a strong interface dipole layer at the Au/PCBM
interface (J. K. J. van Duren et al., unpublished work). A similar
reduction of injection barrier has been observed for the Au/C60 interface
by ultraviolet photoemission spectroscopy12.We note that the injection
barrier can in general be further reduced in a field-effect transistor by
applying both a lateral source-drain field and a perpendicular gate
field13.A narrow injection barrier will allow tunnelling of charge carriers
from the electrode to the semiconductor.
The transfer characteristics of a transistor with gold electrodes and
PCBM as the single semiconductor show good electrical performance
and an electron field-effect mobility of 10–2 cm2 V–1 s–1 at gate voltage
Vg = 20 V, demonstrating that the energy level mismatch between gold
and PCBM can be overcome with the field-effect. In the PCBM
transistor, no hole current was observed. Typical output characteristics
of a field-effect transistor based on the OC1C10-PPV:PCBM blend in
combination with Au electrodes (see Fig. 3a, b) demonstrate operation
in both the hole-enhancement and electron-enhancement mode.
5.0 eV HOMO OC1C10-PPV
6.1 eV HOMO PCBM
Au
Figure 2 Device band diagram of interpenetrating networks of OC1C10-PPV and
PCBM in contact with Au electrodes, when no biases are applied to the transistor.
For simplicity the energy levels are drawn as straight lines. Note that band bending due to
an applied gate voltage can reduce the barrier for electron injection into the PCBM network.
For high negative Vg, the transistor is in the hole-enhancement mode
and its performance is identical to a unipolar transistor based on
OC1C10-PPV, with a field-effect mobility of 7 × 10–4 cm2 V–1 s–1 at
Vg = –20 V. For low Vg and high drain voltages, Vds, the current shows a
pronounced increase with increasingly negative Vds (see Fig. 3a), which
is typical of an ambipolar transistor, and which is not present in a
unipolar device based on OC1C10-PPV. At positive Vg, the transistor
operates in the electron-enhancement mode (Fig.3b),with a field-effect
mobility of 3 × 10–5 cm2 V–1 s–1 at Vg = 30 V, two orders of magnitude
lower than the electron mobility in a PCBM transistor. At low drain
voltages we observe a nonlinear current increase,indicating that there is
a barrier for electron injection from gold into PCBM, which is much
more pronounced than in a unipolar PCBM transistor, where the
output characteristics look qualitatively similar to the hole currents in
the blend transistor shown in Fig. 3a. (We remark that the super-linear
output characteristics at low Vds observed in the blend transistor,
indicate an injection problem from the Au into the PCBM, which is
probably due to the selective wetting of PCBM and OC1C10-PPV on
gold, which can also account for the low electron mobility in the blend
transistor as compared with the PCBM-only transistor.) At low gate
voltages and high drain voltages, we again observe a pronounced
increase in current, typical of an ambipolar transistor, and which is not
observed in a unipolar PCBM transistor. We also measured ambipolar
transport in blends of poly-3-hexylthiophene (P3HT) with PCBM (in a
1:4 weight ratio),with electron and hole mobilities comparable to those
obtained for the OC1C10-PPV:PCBM blend. This shows that ambipolar
transport is a generic property of interpenetrating networks,with workfunction-matched LUMO and HOMO levels. Optimization of
processing conditions and matching with the metal work function may
lead to increases of the mobilities.
The importance of energy level matching was investigated using fieldeffect transistors of single semiconductor compounds. In combination
with gold electrodes, semiconductors such as OC1C10-PPV and related
PPVs, poly(2,5-thienylene vinylene) (PTV), and P3HT showed only ptype transistor behaviour. For PCBM and N,N′-bis[3-[2-[2-(1butoxy)ethoxy]ethoxy]propyl]perylene-3,4,9,10-tetracarboxyldiimide
(PPEEB) with gold, we measured only n-type transistor action. Of the
wide bandgap semiconductors investigated with gold electrodes, only in
pentacene transistors evidence of both p-channel and n-channel
operation was obtained. N-type conduction was only observed in
vacuum,not in air,indicating that ionic motion is unlikely to be the origin
nature materials | VOL 2 | OCTOBER 2003 | www.nature.com/naturematerials
679
©2003 Nature Publishing Group
©2003 Nature Publishing Group
nmat978PRINT
10/9/03
3:28 pm
Page 680
ARTICLES
a
c
–70
–60
–0.3
Vg = –30 V
Blend p-channel
PIF p-channel
–50
0V
Ids (µA)
Ids (nA)
–0.2
–40
–25 V
–30
Vg = –40 V
–0.1
–20
–35 V
–20 V
–10
0V
–15 V
–5 V
–30 V
0
0.0
0
–5
–10
–15
–20
–25
–30
0
–10
Vds (V)
–20
–30
Vds (V)
b
d
0.2
10
PIF n-channel
Vg = 40 V
Blend n-channel
Ids (µA)
Ids (nA)
Vg = 30 V
35 V
5
0.1
25 V
15 V
30 V
20 V
25 V
0
0
5
10
15
20
25
0V
0.0
20 V
30
0
10
20
30
Vds (V)
Vds (V)
Figure 3 The output characteristics of the blend transistor as well as the output characteristics of the PIF transistor demonstrate ambipolar operation. The OC1C10PPV:PCBM ambipolar transistor operating in a, hole-enhancement, and b, electron-enhancement mode.The PIF ambipolar transistor operating in c, hole-enhancement and
d, electron-enhancement mode.
of the observed current. Transfer characteristics are presented in Fig. 4.
A hole and electron mobility of respectively 10–2 cm2 V–1 s–1 (at
Vg = –20V,Vds = –2V) and 10–6 cm2 V–1 s–1 (at Vg = 100V,Vds = 20V) was
derived.We note that the ring geometry with enclosed drain contact did
enable the measurement of the low electron current.The drain currents
measured at Vg = 100 V and Vds = 50 V were more than a factor of ten
higher than the measured gate currents, emphasizing the evidence for
real n-channel operation in pentacene. Figure 4 shows that a large
positive voltage is needed to detect the electron current, indicating the
presence of a large injection barrier for electrons. Pentacene transistors
using silver electrodes gave an improved electron current,but still a high
injection barrier.
The small electron current in pentacene, and the lack of ambipolar
transistor action in the other wide-bandgap semiconductors presumably
is due to the presence of large injection barriers,although effects of charge
trapping and low mobilities may also play a role.We can reduce the barrier
by applying small bandgap semiconductors for which we used poly(3,9di-t-butylindeno[1,2-b] fluorene) (PIF), which has a bandgap energy of
1.55 eV (ref. 14). Typical output characteristics of a field-effect transistor
based on PIF in combination with gold electrodes (Fig. 3c,d) demonstrate
operation both in the hole-enhancement and electron-enhancement
mode.At low drain voltages we observe a nonlinear current increase that
indicates a small barrier for both hole and electron injection from gold
into PIF.In the PIF transistor we find the field-effect mobility for the holes
(at Vg = –30V) to be 4 × 10–5 cm2 V–1 s–1 and for the electrons (at Vg = 30V)
5×10–5 cm2 V–1 s–1,but the onset for the field-effect differs for the holes and
electrons: for the hole-accumulation mode, current flow is observed at
Vg = –15 V, whereas for the electron-enhancement mode current flow is
observed at Vg = +10 V. The reason for this is unclear at present.
The measurements of Fig. 3c,d. show that the use of small bandgap
semiconductors is a promising way to reduce injection barriers and to
obtain ambipolar transistors.We note that the semiconductor purity is of
the next importance to the requirements for charge injection in order to
achieve ambipolar operation. For unintentionally doped systems, the
transport of minority carriers is unfavourably influenced by the presence
of the dopants.Furthermore,the purity is important in order to minimize
trapping effects. A detailed investigation on the influence of purity is
however beyond the scope of this paper.
The ambipolar output characteristics of Fig. 3 can be described
using a simple analytical expression. The current increase at high drain
voltages and low gate voltage can readily be understood when
considering that under certain biasing conditions both holes and
electrons are accumulated in the transistor channel, forming a pnjunction15.For a large negative Vg and a small negative Vds,the transistor
nature materials | VOL 2 | OCTOBER 2003 | www.nature.com/naturematerials
680
©2003 Nature Publishing Group
©2003 Nature Publishing Group
nmat978PRINT
10/9/03
3:28 pm
Page 681
ARTICLES
10–4
a
10–6
VIN
Electron enhancement
Ambipolar
FET
10
Vds = +20 V
Vds = +50 V
Gate current
VDD = +30 V
V0UT (V)
Ids(A)
10–9
Ambipolar
FET
VOUT
20
10–7
10–8
VDD
30
10–5
0
O
O
10–10
–10
4:1
O
10–11
–20
10–12
–30
10–13
–20
0
20
40
60
80
O
VDD = –30 V
n
100
–30
Vg (V)
–20
–10
0
10
20
30
VIN (V)
Figure 4 Transfer characteristics of a precursor pentacene field-effect
transistor with gold electrodes, demonstrating electron currents at high
gate voltages. The inset shows the molecular structure of pentacene.
b
VDD
40
Ambipolar
FET
VOUT
30
VIN
20
(1)
[[
[[[
T
[[
– –Vg + Vso + Vds
[[
L
–Vg + Vso
2To,h
2To,h
T
[
Ids, h (L) =
AhWCi
Ambipolar
FET
10
VDD = +40 V
V0UT (V)
behaves as a unipolar hole-accumulation transistor. If in a unipolar
device, Vds is increased beyond Vds = Vg – Vso (where Vso is the switch-on
voltage of the field-effect, defined as the gate voltage that needs to be
applied to reach the flatband condition16),a depletion region around the
drain will develop and the drain current saturates. However, for an
ambipolar transistor, electrons will start to accumulate at the drain
electrode.This electron accumulation region forms a pn-junction in the
channel with the hole-accumulation region at the source electrode.
This electron accumulation region is responsible for the observed
current increase at high Vds.
When the transistor is biased in the hole-accumulation mode, the
hole current can be described by a percolation model based on variablerange hopping of charge carriers in an exponential density of states
(DOS)16,17:
0
–10
–20
n
–30
VDD = –40 V
–40
–40
–30
–20
–10
0
10
20
30
40
VIN (V)
Figure 5 Transfer characteristics of CMOS-like inverters based on two identical
ambipolar transistors. Depending on the polarity of the supply voltage,VDD,the inverter
works in the first or the third quadrant.A schematic representation of the electrical
connections in the inverter is given in the insets.a,Inverter characteristics of an inverter
based on two identical OC1C10-PPV:PCBM field-effect transistors.The solid lines are
modelled on the basis of equations (1) and (2).b,Inverter characteristics of an inverter
based on two identical PIF field-effect transistors.
where Ah is a prefactor for the hole current16,18, Ci the insulator
capacitance per unit area, T0,h is the width of the exponential DOS for
holes17, and
barrier for electrons of the OC1C10-PPV:PCBM transistor here, assume
that we can use the same Vso for both hole accumulation and electron
accumulation, and do not model the PIF transistor characteristics.
Then, in the electron accumulation mode of the OC1C10-PPV:PCBM
transistor, the current can be described analogously to equation (1):
(2)
[[[
[[
L
Vg – Vso
2To,e
T
–
[[
Vg – Vso – Vds
[[
Ids, e(L) =
– AeWCi
2To,e
T
[
[[
= 1 x+ 1 x
2
2
.
It is assumed that most of the charge carriers are located in the
exponential tail of the DOS,that is,equation (1) breaks down for T>T0,h.
In the model it is also assumed that the transistor operates in the linear
regime (|Vds|<<|Vg|), but the resulting expression of equation (1) turns
out to be a reasonable description of the hole current close to saturation
as well. For a complete description of the electron-accumulation mode
of the OC1C10-PPV:PCBM transistor and for the description of both
electron and hole accumulation in the PIF transistor, we should
consider the injection-limited current.This has recently been modelled
by thermally assisted hopping from the electrode into the localized states
of the organic semiconductor, which are broadened due to disorder19.
For the sake of simplicity, however, we neglect the observed injection
[[x
where Ae is a prefactor for the electron current,and T0,e is the width of the
exponential DOS for electrons.Under bias conditions where both holes
and electrons accumulate in the channel, the transistor is composed of
nature materials | VOL 2 | OCTOBER 2003 | www.nature.com/naturematerials
681
©2003 Nature Publishing Group
©2003 Nature Publishing Group
nmat978PRINT
10/9/03
3:28 pm
Page 682
ARTICLES
two transistors in series, both in saturation, one of length Lh where only
holes accumulate and one of length Le where only electrons accumulate.
We assume that the charge transfer across the pn-junction, separating
the two transistors, is not a limiting factor in the device performance.
Then, the source-drain current can readily be calculated from the
condition of current continuity across the pn-junction,
Ids,h(Lh) = Ids,e(Le) and the relation Le + Lh = L. The result is that the total
current is simply the sum of the source–drain currents in equations (1)
and (2), that is, with L in the denominator. This is true for all bias
conditions, for example, for bias conditions where there is no
accumulation of electrons both terms in equation (2) vanish.
Summarizing, the ambipolar transistor can be represented as a p-type
and n-type transistor, both with length L, connected in parallel.
Complementary logic is possible because both polarities of charge
can be induced in the transistor. Inverters based on two identical
ambipolar transistors were constructed,with a common gate as the input
voltage, VIN. These devices, both for an OC1C10-PPV:PCBM blend and
for PIF, demonstrate CMOS-like inverter operation (Fig. 5).A high gain
of 10 for the OC1C10-PPV:PCBM blend inverter and of 11 for the PIF
inverter is easily achieved (the steepness of the inverter characteristic),in
combination with a good noise margin (the position of the voltage switch
in the inverter characteristic). For unipolar logic, this combination is
difficult to achieve,and typically requires a levelshifter.Depending on the
polarity of the supply voltage, VDD, the inverter works in the first or
the third quadrant of Fig. 5,which is a particular feature of the ambipolar
transistor-based inverter, as unipolar logic works only in one quadrant.
Furthermore,we see a small dependence of the output voltage20,VOUT,at
low and high values of the input voltage, VIN, which is normally not
observed in CMOS-based inverters.This change of VOUT is a direct result
of the fact that both transistors that make up the inverter can be
considered as a parallel circuit of an n-type and a p-type transistor.
Therefore,there is always a leakage current flowing through the inverter,
as neither transistor is ever fully switched off. The switching speed of the
ambipolar inverter will typically be higher than for unipolar logic, but
not as high as for CMOS-based logic, due to this leakage current. Using
the sum of equations (1) and (2) for the ambipolar current, the inverter
characteristics can already be modelled qualitatively by using the various
biasing conditions in the inverter, and the ratio of the extracted fieldeffect mobilities for electrons and holes. From the modelling of the
p-channel behaviour with equation (1),we find T0,h = 530 K.We assume
T0,e = T0,h, and we neglect the injection-limited current observed for the
n-channel. The inverter characteristics calculated in this way are plotted
as solid lines in Fig. 5a.We note that,although we have taken a very simple
model for the ambipolar transistor behaviour, we already get a good
description for the inverter characteristics,including the change in VOUT.
In conclusion, we have shown that ambipolar transport is a generic
property of organic semiconductors.Crucial in field-effect transistors is
matching of the work function of the source and drain electrodes with
the energy levels of the semiconductor. This is experimentally realized
by using both interpenetrating networks and small bandgap
semiconductors. Discrete ambipolar transistors could be combined
into CMOS-like inverters that exhibit high gain and good noise margin.
The ambipolar transport has been modelled and a good description for
the CMOS-like inverters was obtained.Optimization of processing and
energy-level matching will reduce the injection barriers, enhance the
field-effect mobilities, and pave the way for an organic CMOS
technology based on single solution processed semiconductors.
precursor route22,23, PPEEB24, and PIF14. Transistors based on a single semiconductor material were made
by spin-coating the following solutions: PCBM 1 wt% in chlorobenzene, OC1C10-PPV 0.4 wt% in toluene
or chlorobenzene, precursor pentacene, PPEEB, and P3HT 0.5 wt% in CHCl3, PIF 1 wt% in CHCl3.
Transistors based on semiconducting blends were made by spinning solutions of PCBM mixed with
either OC1C10-PPV or P3HT. The interpenetrating networks in these blend transistors are typically used
in organic photovoltaic cells research2. PCBM and OC1C10-PPV (typically 4:1 by weight) were spun from
a 0.5 wt% solution in chlorobenzene. Prior to spin-coating, the mixed solution was stirred for one hour at
80 °C. PCBM and P3HT were mixed in various ratios and spin-coated from a 1 wt% chloroform solution.
All completed devices were annealed in vacuum of 10–4 mbar for 15 hours at 90 °C. The electrical
transport measurements were performed in vacuum, because the devices slowly deteriorated when
measured in air25. The measurements were performed at room temperature with an HP 4156B
semiconductor parameter analyser. For all single semiconductors smooth films were obtained, apart from
PCBM that exhibits a poor wetting behaviour on the substrates used, and precursor pentacene films that
are continuous but microcrystalline and rough. The OC1C10-PPV:PCBM films were investigated with
atomic force microscopy, and showed a similar surface morphology as reportedpreviously26 for the same
mixture. This indicates that the constituents are uniformly mixed.
Received 28 May 2003; accepted 14 August 2003; published 21 September 2003.
References
1. Burroughes, J. H. et al. Light-emitting diodes based on conjugated polymers. Nature 347, 539−541
(1990).
2. Yu, G., Gao, J., Hummelen, J. C., Wudl, F. & Heeger, A. J. Polymer photovoltaic cells: enhanced
efficiencies via a network of internal donor-acceptor heterojunctions. Science 270, 1789−1791 (1995).
3. Dimitrakopoulos, C. D. & Malenfant, P. R. L. Organic thin film transistors for large area electronics.
Adv. Mater. 14, 99−117 (2002).
4. Crone, B. et al. Large-scale complementary integrated circuits based on organic transistors. Nature
403, 521−523 (2000).
5. Dodabalapur, A., Katz, H. E., Torsi, L. & Haddon, R. C. Organic heterostructure field-effect
transistors. Science 296, 1560−1562 (1995).
6. Dodabalapur, A., Katz, H. E., Torsi, L. & Haddon, R. C. Organic field-effect bipolar transistors. Appl.
Phys. Lett. 68, 1108−1110 (1996).
7. Huitema, H. E. A. et al. Plastic transistor in active-matrix displays. Nature 414, 599 (2001).
8. Sirringhaus, H., Tessler, N. & Friend, R. H. Integrated optoelectronic devices based on conjugated
polymers. Science 280, 1741−1744 (1998).
9. Gelinck, G. H., Geuns, T. C. T. & de Leeuw, D. M. High-performance all-polymer integrated circuits.
Appl. Phys. Lett. 77, 1487−1489 (2000).
10. Tada, K., Harada, H. & Yoshino, K. Polymeric bipolar thin-film transistor utilizing conducting
polymer containing electron transport dye. Jpn J. Appl. Phys. 35 L944−L946 (1996).
11. Blom, P. W. M., de Jong, M. J. M. & van Munster, M. G. Electric-field and temperature dependence of
the hole mobility in poly(p-phenylene vinylene). Phys. Rev. B. 55, R656−R659 (1997).
12. Veenstra, S. C., Heeres, A., Hadziioannou, G., Sawatzky, G. A. & Jonkman, H. T. On interface dipole
layers between C60 and Ag or Au. Appl. Phys. A, 75, 661−666 (2002).
13. Sze, S. M. Physics of Semiconductor Devices (Wiley, New York, 1981).
14. Reisch, H., Wiesler, W., Scherf, U. & Tuytuylkov, N. Poly(indenofluorene) (PIF), a novel low band gap
polyhydrocarbon. Macromolecules 29, 8204−8210 (1996).
15. Neudeck, G. W., Bare, H. F. & Chung, K. Y. Modelling of ambipolar a-Si:H thin-film transistors. IEEE
Trans. Electron Dev. 34, 344−350 (1987).
16. Meijer, E. J. et al. Switch-on voltage in disordered organic field-effect transistors. Appl. Phys. Lett. 80,
3838−3840 (2002).
17. Vissenberg, M. C. J. M. & Matters, M. Theory of the field-effect mobility in amorphous organic
transistors. Phys. Rev. B 57, 12964−12967 (1998).
18. Detcheverry, C. & Matters, M. Device simulation of all-polymer thin-film transistors. Proc. ESSDERC
328−331 (2000).
19. van Woudenbergh, T., Blom, P. W. M., Vissenberg, M. C. J. M. & Huiberts, J. N. Temperature
dependence of the charge injection in poly-dialkoxy-p-phenylene vinylene. Appl. Phys. Lett. 79, 1697−
1699 (2001).
20. Chung, K. Y., Neudeck, G. W. & Bare, H. F. Analytical modelling of the CMOS-like a-Si:H TFT
inverter circuit. IEEE J. Solid-State Circ. 23, 566−572 (1988).
21. Meijer, E. J. et al. Dopant density determination in disordered organic field-effect transistors. J. Appl.
Phys. 93, 4831−4835 (2003).
22. Herwig, P. T. & Müllen, K. A soluble pentacene precursor: synthesis, solid-state conversion into
pentacene and application in a field-effect transistor. Adv. Mater. 11, 480−483 (1999).
23. Brown, A. R., Jarrett, C. P., de Leeuw, D. M. & Matters, M. Field-effect transistors made from solutionprocessed organic semiconductors. Synth. Met. 88, 37−55 (1997).
24. Cormier, R. A. & Gregg, B. A. Synthesis and characterization of liquid crystalline perylene diimides.
Chem. Mater. 10, 1309−1319 (1998).
25. de Leeuw, D. M., Simenon, M. M. J., Brown, A. R. & Einerhand, R. E. F. Stability of n-type doped
conducting polymers and consequences for polymeric microelectronic devices. Synth. Met. 87, 53−59
(1997).
26. Shaheen, S. E. et al. 2.5% efficient organic plastic solar cells. Appl. Phys. Lett. 78, 841−843 (2001).
Acknowledgements
METHODS
Thin-film transistor test devices were made, using heavily doped silicon wafers as gate electrode, with a
200 nm thermally grown SiO2 layer as gate dielectric. Using conventional lithography, gold or silver
source and drain contacts were defined with channel widths of 1 to 10 mm and lengths of 10 to 40 µm.
The SiO2 layer was treated with the primer hexamethyldisilazane. The drain electrode was contained
within a circular source electrode to eliminate parasitic leakage currents21. Organic semiconductors
investigated were PCBM, regio-regular head-to-tail coupled P3HT, pentacene and PTV both by a
The authors acknowledge Brian Gregg (National Renewable Energy Laboratory, Golden, Colerado, USA)
for providing a sample of PPEEB, Jitendra Jadam for the synthesis of the PIF, Eugenio Cantatore (Philips
Research) for useful discussions, Henny Herps (Philips Research) for the design of Fig. 1, and also
gratefully acknowledge The Dutch science foundation NWO/FOM through the ‘Laboratorium zonder
muren’ project.
Correspondence should be addressed to D.M.deL.
Competing financial interests
The authors declare that they have no competing financial interests.
nature materials | VOL 2 | OCTOBER 2003 | www.nature.com/naturematerials
682
©2003 Nature Publishing Group
©2003 Nature Publishing Group