Rachana Acharya1,2, Boyu Peng3, Paddy K L Chan3, Guido Schmitz2, Hagen Klauk1. 1. Max Planck Institute for Solid State Research , Stuttgart 70569 , Germany. 2. Institute of Materials Science , University of Stuttgart , Stuttgart 70569 , Germany. 3. Department of Mechanical Engineering , University of Hong Kong , Hong Kong , Hong Kong.
Abstract
The properties of organic thin-film transistors (TFTs) and thus their ability to address specific circuit design requirements depend greatly on the choice of the materials, particularly the organic semiconductor and the gate dielectric. For a particular organic semiconductor, the TFT performance must be reviewed for different combinations of substrates, fabrication conditions, and the choice of the gate dielectric in order to achieve the optimum TFT and circuit characteristics. We have fabricated and characterized organic TFTs based on the small-molecule organic semiconductor 2,7-diphenyl[1]benzothieno[3,2-b][1]benzothiophene in combination with an ultrathin hybrid gate dielectric consisting of aluminum oxide and a self-assembled monolayer. Fluoroalkylphosphonic acids with chain lengths ranging from 6 to 14 carbon atoms have been used to form the self-assembled monolayer in the gate dielectric, and their influence on the TFT characteristics has been studied. By optimizing the fabrication conditions, a turn-on voltage of 0 V with an on/off current ratio above 106 has been achieved, in combination with charge-carrier mobilities up to 0.4 cm2/V s on flexible plastic substrates and 1 cm2/V s on silicon substrates.
The properties of organic thin-film transistors (TFTs) and thus their ability to address specific circuit design requirements depend greatly on the choice of the materials, particularly the organic semiconductor and the gate dielectric. For a particular organic semiconductor, the TFT performance must be reviewed for different combinations of substrates, fabrication conditions, and the choice of the gate dielectric in order to achieve the optimum TFT and circuit characteristics. We have fabricated and characterized organic TFTs based on the small-molecule organic semiconductor 2,7-diphenyl[1]benzothieno[3,2-b][1]benzothiophene in combination with an ultrathin hybrid gate dielectric consisting of aluminum oxide and a self-assembled monolayer. Fluoroalkylphosphonic acids with chain lengths ranging from 6 to 14 carbon atoms have been used to form the self-assembled monolayer in the gate dielectric, and their influence on the TFT characteristics has been studied. By optimizing the fabrication conditions, a turn-on voltage of 0 V with an on/off current ratio above 106 has been achieved, in combination with charge-carrier mobilities up to 0.4 cm2/V s on flexible plastic substrates and 1 cm2/V s on silicon substrates.
The development of organic
thin-film transistors (TFTs) is important
for potential portable and wearable electronic applications.[1,2] In order to ensure portability and safe handling of the systems,
it is necessary that their power consumption is sufficiently low so
they can be powered by solar cells or small batteries. This requires
that the TFTs employed in such devices have low operating voltages.
Moreover, for the realization of low-power complementary circuits,
it is beneficial to have TFTs with turn-on voltages as close to 0
V as possible in order to minimize the off-state drain current while
maximizing the on/off current ratio obtained with a certain maximum
gate–source voltage. Low-voltage operation of organic TFTs
can be achieved by employing ultrathin hybrid gate dielectrics composed
of a thin metal oxide layer and an organic self-assembled monolayer
(SAM), which provide a large dielectric capacitance while minimizing
gate leakage and improving the overall TFT performance.[3−6] A suitable choice of molecule for the SAM is essential in achieving
the desired semiconductor morphology and the optimum TFT characteristics.
Previously, silane-based SAMs that bind to silicon dioxide (SiO2) as the gate oxide have been chemically and electrically
engineered by using molecules with different functional groups to
achieve an optimum crystalline microstructure of the organic semiconductor
and thus obtain high-performance TIPS-pentacene TFTs in terms of stability
and charge-carrier mobility.[7] Alkyl silanes
and fluoroalkyl silanes have also been studied in pentacene TFTs to
evaluate their impact on turn-on voltages and transport mechanisms
in the organic semiconductor.[8] Another
attractive combination of materials for the hybrid gate dielectric
is aluminum oxide (AlO) together with
a SAM based on phosphonic acids, especially those with a saturated
aliphatic chain of carbon and hydrogen atoms attached to the phosphonic
acid head group, which align by mutual cohesive forces to form self-assembled
monolayers.[9] The turn-on voltage of organic
TFTs can be controlled by employing gate dielectrics based on mixed
SAMs with different combinations of alkylphosphonic acids and fluoroalkylphosphonic
acids.[10−14] Kraft et al. used fluoroalkylphosphonic acids to shift the threshold
voltage of both p-channel and n-channel organic TFTs by about 1 V
due to their strong electron-withdrawing character as compared to
the alkylphosphonic acid-based monolayers.[15] Another approach is to vary the chain length of the phosphonic acid
molecules in order to tune the threshold voltage and other electrical
properties of the TFTs.[16−18] In this work, we explore the
use of hybrid AlO/SAM gate dielectrics
in combination with a relatively new organic semiconductor, 2,7-diphenyl[1]benzothieno[3,2-b][1]benzothiophene (DPh-BTBT)[19−21] and different
fabrication conditions to study their influence on the performance
of p-channel organic TFTs. In addition, we investigate the use of
fluoroalkylphosphonic acids with a wide range of chain lengths in
combination with DPh-BTBT to explore the effect of the thickness of
the SAM on the morphology of the organic semiconductor layer and the
TFT performance. This combination of materials and fabrication conditions
could be instrumental in designing improved low-voltage organic complementary
circuits.
Experimental Section
All TFTs were fabricated in the inverted staggered (bottom-gate,
top-contact) configuration with either DPh-BTBT or, for comparison,
dinaphtho[2,3-b:2′,3′-f]thieno[3,2-b]thiophene (DNTT)[22−24] (both purchased
from Sigma Aldrich; Figure ) as the organic semiconductor and fluoroalkylphosphonic acids
(Specific Polymers) of various chain lengths (Figure and Table ), which form the SAM in the hybrid gate dielectric.
Figure 1
Chemical
structures of organic semiconductors DNTT and DPh-BTBT,
schematic cross sections of the TFTs fabricated on flexible polyethylene
naphthalate (PEN) substrates with patterned gate electrodes and of
the TFTs fabricated on doped silicon substrates with a common gate
electrode, and chemical structures of fluoroalkylphosphonic acids
with five different chain lengths used to form the self-assembled
monolayer in the hybrid dielectric.
Table 1
Chemical Names and Formulas of the
Fluoroalkylphosphonic Acids Employed for the Self-Assembled Monolayers
as Part of the Hybrid Gate Dielectric
chemical name
chemical formula
1H,1H,2H,2H-perfluorohexylphosphonic acid
FC6PA
1H,1H,2H,2H-perfluorooctylphosphonic
acid
FC8PA
1H,1H,2H,2H-perfluorodecylphosphonic acid
FC10PA
1H,1H,2H,2H-perfluorododecylphosphonic acid
FC12PA
1H,1H,2H,2H-perfluorotetradecylphosphonic
acid
FC14PA
Chemical
structures of organic semiconductors DNTT and DPh-BTBT,
schematic cross sections of the TFTs fabricated on flexible polyethylene
naphthalate (PEN) substrates with patterned gate electrodes and of
the TFTs fabricated on doped silicon substrates with a common gate
electrode, and chemical structures of fluoroalkylphosphonic acids
with five different chain lengths used to form the self-assembled
monolayer in the hybrid dielectric.TFTs were fabricated on heavily doped silicon
substrates and on
flexible, 125 μm-thick polyethylene naphthalate (PEN) substrates
(Teonex Q65; kindly provided by William A. MacDonald, DuPont Teijin
Films, Wilton, U.K.). For the gate electrodes, a 30 nm-thick layer
of aluminum was deposited by thermal evaporation in vacuum with a
rate of about 2 nm/s. For the TFTs on the silicon substrates, the
aluminum was deposited without patterning and served as a common gate
electrode for all TFTs on the substrate, whereas on the PEN substrates,
the aluminum was deposited through a polyimide shadow mask (CADiLAC
Laser, Hilpoltstein, Germany) to define a patterned gate electrode
for each transistor. The substrates were then exposed to an RF oxygen
plasma (oxygen flow rate: 30 sccm; oxygen partial pressure: 10 mTorr;
RF power: 200 W; duration: 30 s) to increase the thickness of the
native aluminum oxide to about 3.6 nm. The substrates were then immersed
into a 1 mM solution of a particular fluoroalkylphosphonic acid in
2-propanol to allow a monolayer to self-assemble on the aluminum oxide
surface. Each substrate was kept immersed in the solution for a minimum
of 12 h, after which it was rinsed with 2-propanol to remove any physisorbed
molecules, dried with nitrogen, and subsequently heated to a temperature
of 80 °C for 10 min to stabilize the monolayer. After formation
of the SAM, a nominally 25 nm-thick layer of the organic semiconductor
(DPh-BTBT or DNTT) was deposited onto the hybrid AlO/SAM gate dielectric by sublimation in vacuum with a rate
of about 2 nm/min. During the DPh-BTBT semiconductor deposition, the
substrate was held either at room temperature (25 °C) or at an
elevated temperature (100 °C). For the DNTT TFTs, the substrate
was heated to 60 °C during the semiconductor deposition. On the
PEN substrates, the semiconductor was deposited through a shadow mask,
whereas on the silicon substrates, the semiconductor was deposited
without a mask and patterned later by scratching under a microscope.
In the last step, 30 nm-thick gold was deposited by thermal evaporation
in vacuum with a rate of 1.2 nm/s and patterned using a shadow mask
to define the source and drain contacts. The static electrical characteristics
of the TFTs were measured using a Micromanipulator probe station and
an Agilent 4156C Semiconductor Parameter Analyzer. All substrates
were stored and measured in ambient air at room temperature in a yellow
light cleanroom environment. The transfer characteristics were measured
by applying a drain–source voltage (VDS) of −2 V (on silicon substrates) or −3 V (on
PEN substrates) and sweeping the gate–source voltage (VGS) from 1 to −2 V (in steps of −30
mV) or from 1 to −3 V (in steps of −40 mV) and back
to 1 V. The effective charge-carrier mobility (hereafter referred
to as the carrier mobility) was extracted from the following equation: , where L and W are the channel length and channel width,
respectively. The TFTs
on the silicon substrates have a channel length of 100 μm and
a channel width of 200 μm, while the TFTs on PEN have a channel
length of 20 μm and a channel width of 100 μm. Cdiel is the capacitance of the gate dielectric
per unit area, which was determined using a sine wave generator, a
transimpedance amplifier (to convert the displacement current into
a voltage), and an oscilloscope. Apart from the carrier mobility,
other TFT characteristics were also extracted from the transfer curves
for comparison. The turn-on voltage was considered as the gate–source
voltage (VGS) at which the drain current
(ID) is at its minimum,[25] signifying the transition between the off-state and the
on-state. The threshold voltage (Vth)
was obtained by fitting the measured transfer curve to the following
equation: . As one indicator of the quality of the
gate dielectric in reducing leakage currents, the gate current (IG) was measured as a function of the gate–source
voltage, and its values at VGS = −2
V (on silicon substrates) and VGS = −3
V (on PEN substrates) were considered for comparison. The on/off current
ratio (ION/IOFF) was calculated as the ratio between the drain current (ID) measured at the most negative applied gate–source
voltage (either VGS = −2 V or VGS = −3 V) and the drain current (ID) measured at a gate–source voltage
of zero (VGS = 0 V). The thin-film morphology
of the semiconductor surface was imaged by a Zeiss Merlin scanning
electron microscope with an acceleration voltage of 3 kV. Grazing
incidence X-ray diffraction (XRD) was performed using a Rigaku SmartLab
system equipped with a 9 kW copper X-ray source (λ = 1.5406
Å). The X-ray incident angle was constant at 0.2°, and the
diffraction signals were captured with a 0D detector in the out-of-plane
direction. The horizontal width of the incident X-ray beam was 5 mm,
which is smaller than the width of the investigated substrates.
Results and Discussion
The organic semiconductor DPh-BTBT
was deposited by thermal sublimation
in vacuum with the provision of heating the substrate during the semiconductor
deposition process. Conventionally, substrates are heated to an optimum
temperature to facilitate the diffusion of semiconductor molecules
on the surface and form polycrystalline thin films for maximum carrier
mobility and overall TFT performance.[26,27] However, increasing
the substrate temperature beyond the optimum may result in island-mode
growth or desorption of semiconductor molecules from the surface.[28] In Section , we investigate how the substrate temperature during
the DPh-BTBT deposition affects the DPh-BTBT film growth and the characteristics
of TFTs fabricated on silicon and on flexible PEN substrates, and
based on this information, DPh-BTBT TFTs were fabricated on the respective
substrates with different fluoroalkylphosphonic acids forming the
gate dielectric SAM. Section examines the influence of the chain length of the
fluoroalkylphosphonic acid on the semiconductor morphology and the
TFT performance.
Optimum Substrate Temperature
during Semiconductor
Deposition
A striking difference is observed between TFTs
fabricated on silicon and on PEN in the way the morphology of the
DPh-BTBT layer and the TFT performance are affected by the substrate
temperature during the semiconductor deposition. As seen in Figure a–c, the TFTs
fabricated on silicon have carrier mobilities of 0.56 cm2/V s when the substrate is held at a temperature of 25 °C during
the DPh-BTBT deposition and 1.1 cm2/V s when the DPh-BTBT
is deposited at a substrate temperature of 100 °C. This is the
expected behavior for organic semiconductors deposited by thermal
sublimation where a higher substrate temperature promotes larger grain
size and enhanced crystallinity in the thin film.[29] In contrast, for DPh-BTBT TFTs fabricated on PEN, Figure d–f shows
that heating the substrate during the DPh-BTBT deposition to a temperature
of 100 °C yields a carrier mobility that is 3 orders of magnitude
lower than that obtained without substrate heating.
Figure 2
Transfer curves of DPh-BTBT
TFTs fabricated on silicon substrates
with the substrate held at temperatures of (a) 25 °C and (b)
100 °C during the semiconductor deposition and (c) carrier mobilities
extracted from the transfer curves. Transfer curves of DPh-BTBT TFTs
fabricated on flexible PEN substrates with the substrate held at temperatures
(d) 25 °C and (e) 100 °C during the semiconductor deposition
and (f) carrier mobilities extracted from the transfer curves.
Transfer curves of DPh-BTBT
TFTs fabricated on silicon substrates
with the substrate held at temperatures of (a) 25 °C and (b)
100 °C during the semiconductor deposition and (c) carrier mobilities
extracted from the transfer curves. Transfer curves of DPh-BTBT TFTs
fabricated on flexible PEN substrates with the substrate held at temperatures
(d) 25 °C and (e) 100 °C during the semiconductor deposition
and (f) carrier mobilities extracted from the transfer curves.The thin-film morphologies as
seen in the scanning electron microscopy
(SEM) images show that when the PEN substrate is heated during the
semiconductor deposition, the DPh-BTBT molecules form isolated islands
(Figure c, inset)
instead of a continuous film formed when the PEN substrate is held
at 25 °C (Figure d, inset). Since a percolation path for the molecules is harder to
form through isolated islands, the carrier mobility is significantly
lower. For silicon substrates held at 100 °C (Figure a, inset) and 25 °C (Figure b, inset), the thin-film
morphologies indicate thin films with almost complete coverage in
both cases, thereby establishing that the substrate temperature has
little influence on the thin-film morphology of DPh-BTBT films on
silicon substrates.
Figure 3
Out-of-plane XRD spectra and corresponding scanning electron
microscopy
images of DPh-BTBT thin films deposited onto hybrid AlO/FC10PA SAM dielectrics on silicon substrates
held at temperatures of (a) 100 °C (red line) and (b) 25 °C
(black line) and on PEN substrates held at temperatures of (c) 100
°C (red line) and (d) 25 °C (black line) during the semiconductor
deposition.
Out-of-plane XRD spectra and corresponding scanning electron
microscopy
images of DPh-BTBT thin films deposited onto hybrid AlO/FC10PASAM dielectrics on silicon substrates
held at temperatures of (a) 100 °C (red line) and (b) 25 °C
(black line) and on PEN substrates held at temperatures of (c) 100
°C (red line) and (d) 25 °C (black line) during the semiconductor
deposition.Out-of-plane XRD measurements
(Figure ) performed
on the DPh-BTBT films on PEN
and silicon substrates indicate the possible orientation of the DPh-BTBT
molecules, establishing the influence of the substrate temperature
on the thin-film morphology for the two different substrates. On the
PEN substrate for the films deposited at a substrate temperature of
25 °C , the DPh-BTBT molecules are mostly in an upright-standing
orientation, with a dominant diffraction peak at 2θ = 4.65°
that corresponds to a (001) interlayer spacing of 19 Å. The appearance
of the peaks corresponding to (002) and (003) further indicates a
good layer-by-layer structure of the film. On the other hand, the
DPh-BTBT films deposited on PEN at a substrate temperature of 100
°C are characterized by a mix of upright-standing and lying-down
orientation of the molecules, with peaks appearing at 2θ = 4.65°,
23.8°, and 27.1° that correspond to the (001), (020), and
(120) orientations, respectively. The intensity of the (001) signal
is much smaller than that measured on the film deposited at a substrate
temperature of 100 °C. Only the semiconductor molecules that
are standing upright contribute to the lateral carrier transport,
which explains why the carrier mobility is significantly larger when
the semiconductor deposition is carried out at a substrate temperature
of 25 °C, rather than 100 °C. In contrast, it is observed
that the XRD spectra of the DPh-BTBT films on the silicon substrates
are very similar for the two substrate temperatures, with a single
dominant peak at 2θ = 4.65°, corresponding to molecules
in an upright-standing configuration. Since this orientation is favorable
to lateral carrier transport, this also explains the carrier mobilities
of 0.56 and 1.1 cm2/V s for DPh-BTBT TFTs on silicon substrates
for substrate temperatures of 25 °C and 100 °C, respectively,
and the fact that the transistor characteristics depend only weakly
on the substrate temperature during the semiconductor deposition.
Additional numerical data regarding the XRD spectra can be found in
the Supporting Information (Table S2).
This work aims to study the effect of substrate heating during the
semiconductor deposition. A more systematic investigation of the influence
of varying the substrate temperature would be required to deduce the
optimum substrate temperature for the DPh-BTBT deposition. To the
best of our knowledge, this anomalous behavior is unique to the semiconductor
DPh-BTBT and remains unexplained; nevertheless, it helps to identify
the conditions required for fabricating devices and circuits on these
two types of substrate. These appropriate substrate temperatures (100
°C for the silicon substrates and 25 °C for the PEN substrates)
were used to fabricate the TFTs described in the following section.
Effect of Fluoroalkylphosphonic Acid Chain
Length
By using fluoroalkylphosphonic acids with different
chain lengths, SAMs of varying thicknesses are formed, and the capacitance
of the hybrid AlO/SAM dielectric can
be tuned from 1.1 μF/cm2 (with the FC6PASAM) to 0.61 μF/cm2 (with the FC14PASAM) (Table ).
Along with this, the different SAMs also influence the thin-film morphology
of the semiconductor layer deposited onto them, which is crucial since
the gate field-induced carrier channel is confined to the first one
or two monolayers of the semiconductor layer close to the interface
with the SAM.[30]Figure shows the thin-film morphology of DPh-BTBT
films with a nominal thickness of 25 nm deposited onto different SAMs.
It is evident that, with increasing chain length of the fluoroalkylphosphonic
acid, the film coverage decreases and the film appears to be more
sparsely connected. However, as will be shown below, this has little
effect on the charge-carrier mobilities of the TFTs, which suggests
that all films are characterized by a sufficient degree of percolation.
Table 2
Measured Unit-Area
Capacitance of
Hybrid AlO/SAM Gate Dielectrics Employing
Fluoroalkylphosphonic Acids with Five Different Chain Lengths To Form
the SAM
fluoroalkylphosphonic acid
gate dielectric capacitance Cdiel (μF/cm2)
FC6PA
1.1
FC8PA
0.88
FC10PA
0.75
FC12PA
0.58
FC14PA
0.61
Figure 4
Scanning
electron microscopy (SEM) images showing the thin-film
morphology of DPh-BTBT films deposited onto hybrid AlO/SAM gate dielectrics based on fluoroalkylphosphonic
acids with chain lengths of 6, 8, 10, 12, and 14 carbon atoms (from
left to right) on silicon substrates, with the substrate held at a
temperature of 100 °C during the semiconductor deposition.
Scanning
electron microscopy (SEM) images showing the thin-film
morphology of DPh-BTBT films deposited onto hybrid AlO/SAM gate dielectrics based on fluoroalkylphosphonic
acids with chain lengths of 6, 8, 10, 12, and 14 carbon atoms (from
left to right) on silicon substrates, with the substrate held at a
temperature of 100 °C during the semiconductor deposition.Figure shows the
carrier mobility (Figure a) and turn-on voltage (Figure b) of the DPh-BTBT TFTs fabricated on silicon substrates
as a function of the chain length of the fluoroalkylphosphonic acid.
Results for DNTT TFTs are shown for comparison. For both semiconductors,
the carrier mobility is relatively small when the chain length is
either short (FC6PA, FC8PA) or long (FC14PA), and it has a maximum value in the case of a medium chain
length (FC10PA, FC12PA). Upon observing the
same trend for both semiconductors, the influence of the chain length
of the fluoroalkylphosphonic acid on the carrier mobility in
the organic semiconductor layer is evident.
Figure 5
(a) Carrier mobilities and (b) turn-on
voltages of DPh-BTBT and
DNTT TFTs with hybrid AlO/SAM gate dielectrics
based on fluoroalkylphosphonic acids with chain lengths of 6, 8, 10,
12, and 14 carbon atoms fabricated on silicon substrates. (c) Transfer
curves of five DPh-BTBT TFTs with a hybrid AlO/FC10PA-SAM gate dielectric fabricated on a silicon
substrate, showing a turn-on voltage of exactly 0 V. The substrate
was held at a temperature of 100 °C during the DPh-BTBT deposition
and at a temperature of 60 °C for the DNTT deposition.
An interesting aspect
of the combination of fluoroalkylphosphonic
acid SAMs with the organic semiconductor DPh-BTBT is that the turn-on
voltage is extremely close to 0 V, as compared to the significantly
more positive turn-on voltages of the DNTT TFTs (Figure b). The near-zero turn-on voltage
of the DPh-BTBT TFTs is highly beneficial for the design of low-voltage,
low-power circuits, especially complementary inverters and complementary
ring oscillators.[31]Figure c shows the transfer curves of five DPh-BTBT
TFTs with an AlO/FC10PA-SAM
gate dielectric fabricated on the same silicon substrate; all TFTs
have a turn-on voltage of exactly 0 V. Similarly, transfer curves
of five DPh-BTBT TFTs with a hybrid AlO/SAM dielectric based on all fluoroalkylphosphonic acid molecules
available as well as TFTs with a bare AlO dielectric have been included in the Supporting Information (Figure S2).(a) Carrier mobilities and (b) turn-on
voltages of DPh-BTBT and
DNTT TFTs with hybrid AlO/SAM gate dielectrics
based on fluoroalkylphosphonic acids with chain lengths of 6, 8, 10,
12, and 14 carbon atoms fabricated on silicon substrates. (c) Transfer
curves of five DPh-BTBT TFTs with a hybrid AlO/FC10PA-SAM gate dielectric fabricated on a silicon
substrate, showing a turn-on voltage of exactly 0 V. The substrate
was held at a temperature of 100 °C during the DPh-BTBT deposition
and at a temperature of 60 °C for the DNTT deposition.DPh-BTBT TFTs with AlO/SAM gate dielectrics
based on fluoroalkylphosphonic acids with five different chain lengths
were also fabricated on flexible PEN substrates without heating the
PEN during the semiconductor deposition. Thirty TFTs were measured
on each substrate to determine the influence of the chain length of
the fluoroalkylphosphic acid on the TFT characteristics, such
as the threshold voltage, the gate leakage current, the carrier mobility,
and the on/off current ratio. This could be helpful in identifying
the appropriate choice of the fluoroalkylphosphonic acid to meet the
specific device or circuit design requirements. Figure a shows the transfer curves
of these TFTs and the threshold voltage plotted as a function of the
chain length of the fluoroalkylphosphonic acid. Compared to the TFTs
with an AlO gate dielectric, that is,
without a SAM (as indicated by a dashed red line), the fluoroalkylphosphonic
acid SAM leads to a significant shift in threshold voltage toward
more positive values (Figure b), producing a threshold voltage very close to 0 V. This
phenomenon has also been observed for pentacene p-channel and F16CuPc n-channel TFTs[15] as well
as for TFTs employing DNTT and its derivatives as the semiconductor.[32] As seen in Figure b, in spite of the hybrid AlO/SAM gate dielectric being thicker than the AlO gate dielectric, the threshold voltage of all
the TFTs with the hybrid gate dielectrics is much closer to 0 V, indicating
the influence of the SAM in controlling the threshold voltage of the
TFTs. Figure c also
shows the gate leakage current measured at the maximum gate–source
voltage (VGS = −3 V) plotted as
a function of the chain length. The gate current of the TFTs with
the hybrid AlO/SAM gate dielectrics are
1 to 2 orders of magnitude smaller compared to the TFT with the AlO gate dielectric, confirming the beneficial
effect of the SAM in the hybrid gate dielectrics.
Figure 6
(a) Transfer curves of
DPh-BTBT TFTs with hybrid AlO/SAM gate
dielectrics based on fluoroalkylphosphonic
acids with five different chain lengths fabricated on flexible PEN
substrates. The transfer curve of a DPh-BTBT TFT with an AlO gate dielectric without SAM is shown for comparison.
(b) Threshold voltage, (c) gate leakage current, (d) carrier mobility,
and (e) on/off current ratios of the TFTs from panel (a) plotted as
a function of the chain length of the fluoroalkylphosphonic acid.
The red dashed lines indicate the parameter values of the TFT with
an AlO gate dielectric without SAM.
(a) Transfer curves of
DPh-BTBT TFTs with hybrid AlO/SAM gate
dielectrics based on fluoroalkylphosphonic
acids with five different chain lengths fabricated on flexible PEN
substrates. The transfer curve of a DPh-BTBT TFT with an AlO gate dielectric without SAM is shown for comparison.
(b) Threshold voltage, (c) gate leakage current, (d) carrier mobility,
and (e) on/off current ratios of the TFTs from panel (a) plotted as
a function of the chain length of the fluoroalkylphosphonic acid.
The red dashed lines indicate the parameter values of the TFT with
an AlO gate dielectric without SAM.Both the carrier mobility (Figure d) and the on/off
current ratio (Figure e) of the TFTs fabricated on PEN substrates
have maximum values for a medium chain length (0.4 cm2/V
s and 5 × 105 for TFTs with the FC10PASAM), similar to the TFTs fabricated on silicon substrates (Figure a). When alkylphosphonic
acids of varying chain lengths (from 6 to 18 carbon atoms) were used
for the hybrid AlO/SAM gate dielectrics
in pentacene TFTs, the alkyl SAMs with a medium chain length (10 to
14 carbon atoms) were also reported to produce the best TFT performance.[16−18] This was attributed to the observation that SAMs based on molecules
with medium chain lengths tend to form the most well-ordered and most
densely packed monolayers.[33−35] Similar dynamics between the
fluoroalkylphosphonic acids might be at play here, but a spectroscopic
investigation would be required to gain more information on the packing
density and the ordering of the monolayers.Static water contact
angles on different fluoroalkylphosphonic
acid monolayer surfaces were measured on PEN substrates (Figure ) to ensure the quality
of the monolayers and estimate the ordering and packing density in
the SAMs. Apart from verifying the hydrophobic nature of the different
SAMs, the variation of the water contact angle with the chain length
provides an insight into the relative surface energies and underlying
molecular interactions in the SAM. A smaller water contact angle observed
for the shorter and longer chain lengths indicates more wettability
on these surfaces and a higher surface energy. In comparison, the
medium-chain-length FC10PASAM surface has the largest
water contact angle, indicating a low-wettability surface with low
surface energy. This could be correlated to the most well-ordered
and densely packed monolayer with maximum cohesive forces among the
individual fluoroalkylphosphonic acid molecules, thus lowering the
energy of the surface.
Figure 7
Water contact angles measured on the surfaces of fluoroalkylphosphonic
acid SAMs with different chain lengths on PEN substrates.
Water contact angles measured on the surfaces of fluoroalkylphosphonic
acid SAMs with different chain lengths on PEN substrates.Out-of-plane XRD measurements on DPh-BTBT thin
films on different
fluoroalkylphosphonic acidSAMs also give insight into the dependence
of the carrier mobility on the chain length of the phosphonic acids
(Figure ). On the
medium-chain-length FC10PASAM, the DPh-BTBT molecules
are predominantly standing upright, with a dominant diffraction peak
at 2θ = 4.65°. On the shorter-chain-length FC6PA and on the longer-chain-length FC14PASAMs, the XRD
spectra indicate additional peaks at 2θ = 23.9° and 26.95°.
This indicates that a fraction of the DPh-BTBT molecules on these
SAMs are lying face-down, which is unsuitable for efficient charge
transport in the lateral direction. This could explain the observed
dependence of the carrier mobility on the chain length of the gate
dielectric SAM. Additional numerical data regarding the XRD spectra
can be found in the Supporting Information (Table S3).
Figure 8
Out-of-plane XRD spectra measured on DPh-BTBT thin films deposited
onto three different hybrid AlO/SAM dielectrics
on PEN substrates.
Out-of-plane XRD spectra measured on DPh-BTBT thin films deposited
onto three different hybrid AlO/SAM dielectrics
on PEN substrates.
Conclusions
In this work, we have investigated the relatively new organic semiconductor
DPh-BTBT, in combination with hybrid gate dielectrics based on fluoroalkylphosphonic
acid SAMs. We observed a peculiar dependence of the semiconductor
morphology and the resulting TFT performance on the choice of the
substrate (silicon or PEN) and the substrate temperature during the
semiconductor deposition: for DPh-BTBT molecules to form a closed
layer with optimum carrier mobility, the semiconductor deposition
should be carried out with substrate heating on silicon substrates
and without substrate heating on flexible PEN substrates. Furthermore,
the chain length of the fluoroalkylphosphonic acid molecules forming
the SAM has a strong influence on the thin-film morphology of the
semiconductor and the corresponding TFT performance. We find that
the TFTs with the medium-chain-length FC10PA-SAM have the
highest carrier mobility and the largest on/off current ratio. By
employing DPh-BTBT in combination with fluoroalkylphosphonic acid
gate dielectric SAMs, organic TFTs with a turn-on voltage of 0 V can
be fabricated, which may be beneficial for low-voltage, low-power
integrated circuits.
Authors: Michael Geiger; Marion Hagel; Thomas Reindl; Jürgen Weis; R Thomas Weitz; Helena Solodenko; Guido Schmitz; Ute Zschieschang; Hagen Klauk; Rachana Acharya Journal: Sci Rep Date: 2021-03-18 Impact factor: 4.379