Using Sum-Frequency Generation (SFG) to Probe Electric-Fields
within Organic Field-Effect Transistors
Douglas J. C. Gomes
1
, Silvia G. Motti
2
and Paulo B. Miranda
1
1
São Carlos Institute of Physics, University of São Paulo,
C.P. 369, São Carlos – SP, 13560-970, Brazil
2
Center for Nano Science and Technology @PoliMi, Istituto Italiano di Tecnologia,
Via Pascoli 70/3, 20133 Milano, Italy
Keywords: Sum-Frequency Generation, Organic Transistors, Electric-Field Poling.
Abstract: Organic Field-Effect Transistors (OFETs) have attracted much research interest due to their potential for
unique applications, such as flexible electronics. The operation of OFETs depends on the charge
accumulation at the interface between an organic semiconductor and a dielectric material, induced by the
voltage applied at the gate electrode. Direct measurements of the electric-field distribution in an operating
device are useful for proposing and validating theoretical models for OFET operation. Here we propose
using the second-order nonlinear optical process of Sum-Frequency Generation vibrational spectroscopy
(SFG spectroscopy) to probe the presence of an electric-field in the dielectric layer of OFETs, in a non-
invasive, non-destructive and remote fashion. The OFETs were fabricated with a dielectric layer consisting
of poly(methyl-methacrylate) – PMMA, and an active layer based on poly(3-hexyl thiophene) – P3HT, and
SFG spectra were acquired from the channel region of operating OFETs. It was observed the appearance of
vibrational bands due to carbonyl groups (~ 1720 cm
-1
) of the PMMA layer, whose χ
(2)
were induced by the
electric-field within the dielectric, similarly to a reversible poling of polymers. This phenomenon opens up
the possibility of mapping the spatial charge distribution in the conducting channel using SFG microscopy
in operating devices.
1 INTRODUCTION
Semiconductor devices based on organic materials,
such as transistors or solar cells, have attracted much
research interest due to their potential for unique
applications, such as flexible and/or low-cost
electronics. The operation of organic eld-eect
transistors (OFETs) depends on the modulation of
charge accumulation at the interface between an
organic semiconductor and a dielectric material,
induced by the voltage applied at the gate electrode.
The accumulated charge carriers form a conducting
channel and are driven along this interface by
application of an electric field parallel to the
interface, between drain and source electrodes (Sze
and Kwok, 2007). Although the basic physics of
OFETs is well known, there is currently no single
model that can explain its operation both in the
linear regime and in saturation, after the pinch-off of
the conduction channel (Stallinga and Gomes,
2006). Therefore, direct measurements of the
electric field distribution in an operating device are
useful for proposing and validating theoretical
models.
Several methods have been used to probe the
potential, field or charge distribution in operating
OFETs, including Kelvin probe microscopy (Hallam
et al., 2009), microwave conductivity microscopy
(Babajanyan et al., 2011), charge modulation
spectroscopy (Sciascia et al., 2011), Raman
microscopy (Furukawa et al., 2012), and second-
order nonlinear optical processes, such as second-
harmonic generation (SHG) (Manaka et al., 2007a)
and vibrational spectroscopy by sum-frequency
generation (SFG) (Nakai et al., 2009). The optical
methods have the advantage of being non-invasive,
that is, they do not affect the charge distribution that
is being probed. Among them, SHG and SFG are
very versatile, since they allow probing the field
distribution (Manaka et al., 2007a; Nakai et al.,
2009) and dynamics (Manaka et al., 2007b) within
the organic semiconductor. In this work, we further
194
Gomes, D., Motti, S. and Miranda, P.
Using Sum-Frequency Generation (SFG) to Probe Electric-Fields within Organic Field-Effect Transistors.
DOI: 10.5220/0005812801920196
In Proceedings of the 4th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2016), pages 194-198
ISBN: 978-989-758-174-8
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
extend these methods by showing that SFG
vibrational spectroscopy can be used to selectively
probe the electric-field inside the dielectric layer of
OFETs, which is directly related to the charge
distribution in the OFET channel.
2 MATERIALS AND METHODS
2.1 Materials and Sample Preparation
The bottom-gate/top-electrode OFETs have been
fabricated by sequential spin-coating of a dielectric
layer consisting of poly(methyl-methacrylate) –
PMMA, and an active layer based on poly(3-hexyl
thiphene) – P3HT, using solutions made with
orthogonal solvents. The PMMA layer was ~ 400
nm thick and was spin-coated from a methy-ethyl
ketone solution, while the P3HT film was spin-
coated over the PMMA film from a tetralin solution.
The thickness of this P3HT film was only ~ 30 nm
to minimize the absorption of the SFG light
(wavelength ~ 480 nm) generated at the PMMA
film, while exiting through the P3HT layer.
Figure 1: Scheme of the device structures: (a)
Si/SiO
2
/PMMA/P3HT/Au and (b) Glass/Au/PMMA/P3HT
/Au. (c) Schematic view of the SFG spectroscopy
experiment in the channel of the polarized device.
Two types of substrates were used: for the
measurements of SFG spectra as a function of
applied potential, we used n-doped Si wafers with a
300 nm thermal oxide layer as a gate electrode,
while the temperature dependence was measured on
a device fabricated on glass substrates, with the gate
electrode produced by thermal evaporation of Au.
The top electrodes were also evaporated Au films,
with channel dimmensions L = 100 μm and W = 2
mm. The structure of the devices is shown in Figure
1, together with a scheme of the SFG experiment on
a polarized device.
2.2 SFG Spectroscopy
The second-order nonlinear optical process of sum-
frequency generation (SFG) is forbidden in media
with inversion symmetry, so that it has been widely
used to selectively probe the interfacial region
between two materials, where the inversion
symmetry is usually broken (Shen, 1996). In its
conventional implementation for surface studies,
two high intensity laser pulses (one usually in the
visible and another tunable in the mid-infrared) is
incident at the sample interface, and a sum-
frequency beam is generated in the reflected
direction due to the second-order polarization
generated at the interface (see Equation (1)). The
SFG intensity is proportional to the product of the
input beam intensities and the square of the effective
second-order susceptibility of the interface,

()
,
and it is measured with a photomultiplier, after
spectral and spatial filtering. However, if a strong
DC electric field E
0
is present in the bulk of a
centrosymmetric material, it breaks the inversion
symmetry and leads to a significant increase of the
SFG signal. As shown in Equations (1) and (2), this
increase of

()
has two contributions (Hayes,
2010): (i) a field-induced orientation of the dipoles
in the material that leads to a non-vanishing
()
(
), and (ii) a third-order contribution that
mixes two laser fields E
ω
1
, E
ω
2
and the DC field E
0
.
They can be lumped together as an effective third-
order contribution, which is proportional to the
amplitude of the DC field,

()
. Therefore, if in
the zero-field limit

()
is negligible, the SFG
intensity is proportional to the square of the DC field
in the material.
(1)
Using Sum-Frequency Generation (SFG) to Probe Electric-Fields within Organic Field-Effect Transistors
195
(2)
Our SFG experiments were performed with a
commercial SFG spectrometer (Ekspla, Lithuania),
which is based on a picosecond high energy
Nd
+3
:YAG laser (30 mJ, 20 Hz repetition rate, 1064
nm) and an OPG/OPA/DFG unit pumped by the
third harmonic of the laser (355 nm), to generate
tunable mid-IR beam from 1000 cm
-1
to 4000 cm
-1
with pulse energies at the sample attenuated to ~150
μJ. The visible beam pumping the sample was a
portion (~10 μJ) of the second-harmonic of the laser
(532 nm). The incidence angles were 51° and 60° for
the mid-IR and visible beams, respectively. The SFG
output beam direction was determined by
phasematching along the surface plane, being ~
59.2° for an IR frequency of 1700 cm
-1
. In these
conditions, the coherence length (L
c
= 1/Δk) for the
SFG process in the reflection direction was ~ 76 nm.
This sets the maximum PMMA film thickness from
which the SFG signal is coherently generated (being
proportional to the square of the film thickness).
However, since both types of OFET samples had
reflective gate electrodes, we may also have
collected simultaneously the SFG signal generated
in the PMMA layer in the transmission direction,
which was then reflected by the gate electrode. For
the transmission geometry, the coherence length is
much longer (~ 65 μm), so that in this case the
whole PMMA film is contributing to the detected
SFG signal.
3 RESULTS AND DISCUSSION
Figure 2 displays the electrical characteristics of the
OFET on the Si substrate (Si/SiO
2
/PMMA/
P3HT/Au). These output curves show a nice
saturation and typical p-type OFET performance.
However, it can be seen that they do not cross the
origin, indicating that the gate leakage current is
considerable, since the top contact pads had an area
of ~16 mm
2
, and the whole substrate is conductive.
Figure 3 shows the SFG spectra obtained in the
channel region of the OFET on the Si substrate, for
several gate voltages (keeping V
D
= 0). It can be
seen that upon polarizing the device, a vibrational
resonance arises at ~1720 cm
-1
, which is assigned to
the C=O stretch of the PMMA monomer (its
chemical structure is shown in the inset of Figure 3).
We have checked this assignment by repeating the
experiment on an OFET without the PMMA layer
(only SiO
2
as dielectric), when the SFG peak
disappeared. The solid line is a fit to Equation (3),
which represents the superposition of a vibrational
resonance at
ω
0
with amplitude A and width Γ, plus
a weak nonresonant background,

()
.
(3)
Figure 2: Output curves for the OFET with the structure:
Si/SiO
2
/PMMA/P3HT/Au.
Figure 3: SFG spectra for the OFET on Si substrate, with
several gate voltages, keeping V
D
= 0. The solid lines are
fits to Equation (3). The inset shows the chemical structure
of the PMMA monomer.
From the fits, we can see that the amplitude A of
the resonance increases nearly proportional to the
applied gate voltage, as expected from Equation (2).
Therefore, the amplitude of this peak can be used to
probe the magnitude of the electric field within the
organic dielectric layer (PMMA). The carriers
accumulated at the semiconductor-dielectric
interface could in principle give rise to an increase
of the nonresonant contribution,

()
, but this is not
very significant, as can be seen from the fits and
() ()
()
() ()
22 3 3
000eff eff
EEE
χχ χ χ
=+=
()
()
2
2
2
(2)
0
()
IR
eff IR
SFG NR
IR
A
I
i
ω
χ
χω
ωω
∝= +
−+Γ
PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology
196
from the SFG signal far off resonance (Figure 3).
However, due to interference with the weak
nonresonat background, we can see that the peak
lineshape is a bit asymmetric. It is possible to exploit
this interference to get not only the magnitude, but
also the relative sign of the DC field within the
PMMA layer. For that, we need a stronger

()
,
comparable to the resonant peak nonlinearity (A/Γ).
With this purpose, we then fabricated the OFETs on
glass substrates with Au films as gate electrodes.
Gold has a high

()
at the wavelengths used, so that
it provides a reference signal with which the
resonant peak from PMMA interferes. Figure 4
shows the SFG spectra obtained for the device with
the glass substrate (Glass/Au/PMMA/P3HT/Au),
with both positive and negative gate voltages (and
not polarized, as a reference). With V
G
= 0, we
detect only the nonresonant contribution from the
Au electrode (gate). For negative gate voltages, we
see that the vibrational peak from the C=O stretch of
PMMA interferes destructively with the nonresonant
background from Au, producing a dip in the
spectrum. For positive voltages, the orientation of
the C=O dipoles is inverted, changing the sign of the
peak amplitude A and yielding a constructive
interference with

()
and giving rise to a
pronounced peak at the molecular resonance.
We could estimate the sensitivity of the method
to detect electric fields in the dielectric layer (at least
for these non-optimized conditions) by considering
that the minimum detectable change in the spectra
shown in Figure 3 occurs for V
G
~ - 10V. This
corresponds to an E-field of about 2.510
5
V/cm.
Figure 4: SFG spectra for the OFET with a Gold gate
electrode (Glass/Au/PMMA/P3HT/Au), without and with
both positive and negative gate voltages, keeping V
D
= 0.
Finally, we tried to address the origin of the
effective second-order nonlinearity that we have
been measuring in our devices. Is

()
mostly due to
the field-induced reorientation of the molecular
dipoles (as in a fully reversible electric field poling
of polymers (Bauer, 1996)), or is it a higher order

()
contribution, that does not require any
molecular reorientation in the solid polymer film and
may happen in perfectly isotropic materials? In the
first case, we should expect a strong temperature
dependence of the resonant contribution
()
, as
shown by Equation (4) (μ is the molecular dipole
moment), provided that the DC field is not strong
enough to reach dielectric saturation. If
()
dominates, its temperature dependence should be
rather weak, because it would not depend as strongly
as
()
on the average molecular orientation.
(4)
Figure 5 shows the temperature dependence of
the peak intensity of the spectrum in Figure 4, with
V
G
= +60 V. It can be seen that there is a significant
decrease of the SFG intensity while heating the
sample, which is reversible upon cooling, within the
uncertainty of the measurement. The solid line
represents a signal decrease as predicted by Equation
(4). Therefore, these data support the conclusion that
even for our samples, which are solid and should
have a relatively low molecular mobility, the
()
contribution due to DC field poling of the PMMA
layer is the dominant contribution to the effective
second-order nonlinearity. Any
()
contribution
appears to be negligible.
Figure 5: SFG signal for the OFET with a Gold gate
electrode, biased with V
G
= +60 V and V
D
= 0, upon a
heating and cooling cycle (heating/cooling rate ~ 2
°C/min). The solid line is the temperature dependence
predicted by Equation (4).
()
2
D
C
R
B
E
kT
μ
χ
Using Sum-Frequency Generation (SFG) to Probe Electric-Fields within Organic Field-Effect Transistors
197
4 CONCLUSIONS
Direct measurements of the electric-field distribution
in an operating organic transistor are useful for
proposing and validating theoretical models for their
electrical behavior. Here we showed that Sum-
Frequency Generation vibrational spectroscopy
(SFG spectroscopy) is a valuable tool to probe the
electric field within the (organic) dielectric layer of
the transistor. The SFG spectra from polar groups of
the dielectric (PMMA) show a marked resonance,
whose amplitude is proportional to the applied gate
voltage, which in turn is proportional to the field
within the dielectric, when V
D
= 0. Using the
interference of this resonant contribution with a
large nonresonant background from the gate
electrode, it is possible to determine not only the
relative magnitude of the field within the dielectric,
but also its sign. We further demonstrated that even
for solid samples, where the dipoles are less mobile
than in the case of organic molecules in solution, the
()
contribution due to DC field reorientation of the
dielectric layer is the dominant contribution to the
effective second-order nonlinearity, and any
()
contribution appears to be negligible. This
phenomenon allows probing the electric field in the
dielectric, which is directly related to the charge
distribution in the OFET channel, and opens up the
possibility of mapping the spatial field distribution
in the conducting channel by SFG microscopy of
OFETs (Nakai et al., 2009).
ACKNOWLEDGEMENTS
DJCG gratefully acknowledges a PhD scholarship
from FAPESP and SGM thanks CAPES for a MSc
scholarship. This work has been supported by the
Brazilian agencies FAPESP and CNPq.
REFERENCES
Babajanyan, A., Melikyan, H., Kim, J., Lee, K., Iwamoto,
M., Friedman, B., 2011. Direct imaging of
conductivity in pentacene field-effect transistors by a
near-field scanning microwave microprobe. Org.
Electron. 12, 263–268.
Bauer, S., 1996. Poled polymers for sensors and photonic
applications. J. Appl. Phys. 80, 5531–5558.
Furukawa, Y., Seto, K., Nakajima, K., Itoh, Y., Eguchi, J.,
Sugiyama, T., Fujimura, H., 2012. Infrared and Raman
spectroscopy of organic thin films used for electronic
devices. Vibr. Spectrosc. 60, 5–9.
Hallam, T., Lee, M., Zhao, N., Nandhakumar, I.,
Kemerink, M., Heeney, M., McCulloch, I.,
Sirringhaus,H., 2009. Local charge trapping in
conjugated polymers resolved by scanning Kelvin
probe microscopy. Phys. Rev. Lett. 103, 256803.
Hayes, P. L., Malin, J. N., Jordan, D. S., Geiger F. M.,
2010. Get charged up: Nonlinear optical voltammetry
for quantifying the thermodynamics and electrostatics
of metal cations at aqueous/oxide interfaces. Chem.
Phys. Lett. 499, 183–192.
Manaka, T., Nakao, M., Yamada, D., Lim E., Iwamoto M.,
2007 a. Optical second harmonic generation imaging
for visualizing in-plane electric field distribution.
Optics Express 15, 15964-15971.
Manaka, T., Lim E., Tamura, R., Iwamoto M., 2007 b.
Direct imaging of carrier motion in organic transistors
by optical second harmonic generation. Nat. Photonics
1, 581-584.
Nakai, I. F. et al., 2009. Molecular structure and carrier
distributions at semiconductor/dielectric interfaces in
organic field-effect transistors studied with sum-
frequency generation microscopy. Appl. Phys. Lett. 95,
243304.
Sciascia, C. et al., 2011. Sub-micrometer charge
modulation microscopy of a high mobility polymeric
n-channel field-effect transistor. Adv. Mater. 23,
5086–5090.
Shen, Y. R., 1996. A few selected applications of surface
nonlinear optical spectroscopy. Proc. Natl. Acad. Sci.
USA 93, 12104.
Stallinga, P., Gomes, H., 2006. Modeling electrical
characteristics of thin-lm eld-effect transistors I:
trap-free materials. Synth. Met. 156, 1305–1315.
Sze, S. M., Kwok, K. Ng., 2007. Physics of Semiconductor
Devices, Wiley, New Jersey.
PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology
198