Organic Electronics 15 (2014) 1672–1677

Contents lists available at ScienceDirect

Organic Electronics
journal homepage: www.elsevier.com/locate/orgel

Fabrication of organic thin film transistors on Polyethylene

Terephthalate (PET) fabric substrates
Gi Sung Ryu a, Seung Hyeon Jeong a, Byoung Chul Park b, Beob Park b, Chung Kun Song a,⇑
a
Department of Electronics Engineering, Dong-A University, South Korea
b
Kolon Glotech, INC., South Korea

a r t i c l e i n f o a b s t r a c t

Article history: In this paper organic thin film transistors (OTFTs) are directly fabricated on fabric sub-
Received 15 November 2013 strates consisting of Polyethylene Terephthalate (PET) fibers. A key process is coating the
Received in revised form 17 March 2014 polymer layers on the fabric in order to reduce the large surface roughness of the fabric
Accepted 18 March 2014
substrate. Two polymers, i.e. polyurethane (PU) and photo-acryl (PA), are used to reduce
Available online 4 April 2014
the large surface roughness and simultaneously improve the process compatibility of the
layers with the subsequent OTFTs processes while also retaining the original flexibility
Keywords:
of the fabric. The surface roughness of the PU/PA-coated fabric is significantly reduced to
Organic thin film transistors
e-Textile
0.3 lm. Furthermore, the original flexibility of the PET fabric remained after coating of
Fabric substrates the PU/PA polymer layers. The mobility of the OTFTs fabricated on the PU-PA coated fabric
Polyurethane substrate is 0.05 ± 0.02 cm2/V s when three PA layers and 90 nm thick pentacene layer
Photo-acryl were used. The performance does not vary even after 30,000 bending test.
Pentacene Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction diodes [14], and organic photovoltaic [15] are imple-

mented in a single fiber and then the fibers are woven
Electronic textiles (e-textiles) where electronic func- together to realize a function [13]. In order to realize the
tions are integrated into the textiles are garnering signifi- dream of in-cloth e-textiles, a unique creativeness is
cant attention because they provide greater convenience required for weaving the functional fibers.
and functionality than the conventional hand-carried Recently, on-cloth e-textile applications has primarily
devices. It is expected that e-textiles will be applied in very focused on fabricating conductors on fabric substrates
diverse areas, including healthcare, sports, fashion, and using thermal evaporation with metals as well as solution
military [1,2]. processes with carbon nanotubes (CNTs) and conducting
E-textile research began with the wearable electronics polymers [16–18]. Active devices such as sensors and bat-
where conventional electronic circuit boards were simply teries have been successfully fabricated on fabrics [19–23].
attached to the textile surface [3–7]. Nowadays, e-textiles However, OTFTs have not yet been reported to have been
are being developed for on-cloth applications where the fabricated on fabric substrates due to the large surface
electronic devices are fabricated directly onto the textiles roughness of the fabric substrates; this roughness is a sig-
[1]. E-textiles are expected to evolve into textiles with in- nificant problem for OTFTs whose currents flow along the
cloth applications where active devices such as organic surface of fabric substrates.
thin film transistors (OTFTs) [8–13], organic light emitting In this paper the direct fabrication of OTFTs on fabric
substrates, which are key devices for on-cloth e-textiles,
are investigated. In particular, a surface smoothing process
⇑ Corresponding author. Tel.: +82 51 200 7711; fax: +82 51 200 6965.
was developed in order to reduce the surface roughness of
E-mail address: [email protected] (C.K. Song).

http://dx.doi.org/10.1016/j.orgel.2014.03.019
1566-1199/Ó 2014 Elsevier B.V. All rights reserved.
 G.S. Ryu et al. / Organic Electronics 15 (2014) 1672–1677 1673

the fabric substrate. Based on the surface smoothing pro- In the first step, two polymers polyurethane (PU) and a
cess, OTFTs were fabricated using the conventional process Si-based polymer were coated separately and their feasi-
used for to plastic substrates [24]. The second purpose of bility was investigated. The contact angle and surface
this study is preservation of the original flexibility, i.e. roughness were measured and evaluated in order to iden-
drape properties, of the fabric substrates even after the tify the suitability for the subsequent OTFT processes. In
smoothing process and OTFT fabrication. Furthermore, the second step, an ozone (O3) treatment and photo-acryl
the performance of the OTFTs including their bending (PA) were combined in order to further reduce the surface
characteristics and reliability in air was examined. roughness and to enhance the wettability of the first layer.
The wettability and surface roughness were examined
2. Experiments using the contact angle and 3D profiling in order to deter-
mine their compatibility with the subsequent OTFT pro-
In general, a fabric substrate consisting of fibers pro- cesses. Furthermore, the drape properties were also
vides two types of roughnesses; a nano-scale roughness tested using cantilever method [27]; then, they were com-
associated with a single fiber surface and a micro-scale pared with a PET film and a cotton fabric in order to deter-
roughness related to the periodic bulges generated by the mine whether the initial flexibility of the PET fabric was
woven fibers whose size is approximately equal to the maintained.
diameter of the fiber. Normally, the nano-scale roughness The OTFT processes were performed on the fabric sub-
interrupts carrier transportation because charge carriers strate coated with these two polymer layers. First, alumi-
are trapped and released at the valleys on the dielectric num was thermally evaporated with a thickness of 80 nm
in channel [25,26]. However, it can be cured by coating a using a shadow mask for gate electrodes. Then, polyvinyl-
polymer film such as polyurethane used in this paper. By phenol (PVP) was spin-coated and annealed for 10 min at
the way, the micro scale roughness is not a simple problem 110 °C and then continuously for 20 min at 200 °C for the
to be resolved by coating a thick film because the height of gate dielectric. The detailed process can be found in [28].
bulges is too large to be buried in a thick polymer film and The thickness was approximately 300 nm. Pentacene was
the flexibility of textile should be maintained with the subsequently evaporated at a rate of 0.3 Å/s using a sha-
thickly coated film simultaneously. Otherwise, the large dow mask for the semiconductor layer; thickness was
roughness may be a cause of discontinuity of metal and 90 nm. During the evaporation, the temperature of the fab-
semiconductor thin films. Therefore, the surface roughness ric substrate was maintained at 80 °C for high quality crys-
is a significant difficulty that must be addressed in order to tal [29]. Finally, the source and drain electrodes were
effectively realize electronics on fabrics. formed via evaporating gold on the pentacene layer using
In this study, fabrics woven with Polyethylene Tere- a shadow mask. The OTFTs with a bottom gate and top
phthalate (PET) fibers whose diameter was about 100 lm contact structure were completed on the fabric substrate
were used as the substrate. As shown in Fig. 1, a two-step as shown in Fig. 1. This process has been applied to the
process was applied to address the difficulty of surface OTFTs fabricated on glass and plastics [30,31], and thus
roughness. The first step is reducing the large surface the difference of results can be estimated to result from
roughness to be as small as possible through coating a the substrate.
polymer layer on the fabric substrate without considering The OTFT performance on the PET fabric substrate was
the process compatibility with the subsequent OTFT pro- investigated through measuring the electrical characteris-
cesses. The second step is rendering the first layer compat- tics according to the surface roughness and thickness of
ible with the OTFT processes, in particular rendering the pentacene layer. Furthermore, bending tests were also
making the surface adhesive to the gate dielectric of OTFTs, conducted by laminating the fabric substrate containing
using surface treatments and an additional coating of a OTFTs on a plastic substrate and bending it repeatedly
polymer layer coating. Furthermore, the original flexibility and then examining the variations in performance. The
must be maintained even with the two step processes. reliability in air was also examined after the bending test.

Fig. 1. The structure of polymer coating layers and OTFTs fabricated on fabric substrates; the first layer to reduce the large surface roughness of the fabric
substrates and the second layer to enhance compatibility of the first layer with OTFTs processes.
1674 G.S. Ryu et al. / Organic Electronics 15 (2014) 1672–1677

3. Results and discussion of the difference in the thermal expansions between the
PU layer and PET fabric.
The surface roughness of the fabric substrate before and It is important to maintain the original tactile flexibility
after the first polymer coating is depicted in Fig. 2a, and is of the fabric even after PU-coating. The drape properties of
also compared with that of polycarbonate (PC) plastic and the fabric substrate were measured using a cantilever
a glass substrate. The bare PET fabric exhibited a roughness experiment. This involved a bending length, (‘/2), when
of 10.05 lm; this was reduced to 1.92 lm after being the fabric moved by the length of ‘ until the front line of
coated with PU. The Si-based polymer further decreased the fabric touched the slope of 41.5°, as shown in Fig. 3.
the surface roughness to 0.59 lm. However, the contact It was found that the smaller the bending length was, the
angle of the Si-based polymer layer increased to 102°, better the drape property was. As shown in Fig. 3, the
which resulted in the second coating being difficult. Thus, bending length of the PU-coated PET fabric was 4.3 cm,
the PU-coated fabric substrate was selected for the second which is similar to the 4.9 cm of the original PET fabric,
step process. However, the roughness should be reduced and it was significantly smaller than the 13.9 cm of the
even further. For comparison, the surface roughnesses of PET film. Thus, the drape properties of the PET fabric were
the PC and glass substrate were 0.028 lm and maintained with the PU-coating layer. The drape proper-
0.0026 lm, respectively. ties of the PU-coated fabric were identical to those of cot-
The thermal endurance of the PU-coated fabric sub- ton and nylon fabrics; thus, the tactile flexibility of the PU-
strate was tested through heating for 20 min at various coated PET fabric was identified to be similar to natural
temperatures and through comparing the roughness varia- clothes.
tions as shown in Fig. 2b. The roughness did not vary until The second step was performed on the PU-coated fabric
180 °C. However, at 200 °C it had increased 2.5-fold. Fur- substrate in order to enhance the adhesion of the PU layer
thermore, it appeared that the fabric wrinkled as a result on the PVP, which is the gate dielectric in the OTFTs, and
also to further decrease the surface roughness. The contact
angle of the PU layer was 95°; thus, PVP was not uniformly
coated over the layer. Moreover, the surface roughness of
1.92 lm was too large to fabricate OTFTs. Therefore, photo-
sensitive acryl (PA) was coated over the PU layer. The PA
was spin-coated and exposed to UV for 45 s and then
annealed at 130° for 30 min. The UV-sensitive PA could
reduce the thermal annealing temperature, and thus the
exposure time to heat, in order to minimize the variation
of the PU layer. The low annealing temperature of PA
avoided the increases in the PU surface roughness that
resulted from high temperatures, as shown in Fig. 2b. The
PA layer surface was treated with O3 to enhance its adhe-
sion. The roughness was reduced to 0.64 lm, but this
remains large for OTFTs. The PA/O3 processes were
repeated until a sufficiently small roughness was obtained.
With three PA/O3 layers, the roughness was improved to
0.32 lm, which is sufficiently small for OTFTs, as shown
in Fig. 2a. However, using a single PA layer with the same
thickness as that of the three layers could not be produced
simply by reducing the spin velocity because the high

Fig. 2. (a) The surface roughnesses of PET fabric substrates and polyure-
thane/photo-acryl coated PET fabrics; those of a plastic and glass
substrate are also included for comparison. (b) The surface roughness of
polyurethane-coated PET fabric substrates with respect to the annealing Fig. 3. The results of cantilever bending test for the various fabric
temperatures. substrates; the inset is the picture of cantilever measurement system.
 G.S. Ryu et al. / Organic Electronics 15 (2014) 1672–1677 1675

viscosity of PA increased the difficulty of obtaining a uni- different thicknesses were prepared. One device had a sin-
form PA layer and it generated clusters of PA instead. gle PA layer with a surface roughness of 0.64 lm, and the
In Fig. 4 the transfer and output characteristics of the other had three PA layers with a surface roughness of
OTFTs on the PET fabric substrates with three PA layer 0.32 lm. The surface of PVP/3-PA with 0.3 lm thick PVP
and 90 nm thick pentacene are presented. The OTFTs layer coated over three PA layers is estimated to become
exhibited the clear transistor characteristics producing a much smoother than that of PVP/1-PA because 0.3 lm
mobility of 0.05 ± 0.02 cm2/V s, a threshold voltage of thick PVP layer may sufficiently cover 3 PA layer with
8.06 ± 2.50 V, a sub-threshold slope of 1.20 ± 1.02 V/dec, 0.32 lm roughness. As shown in Fig. 5a, the three PA layers
and an on/off current ratio of 8.67 ± 4.52  105, as with 90 nm pentacene (3PA/pen-90 nm) exhibited better
described in Table 1. However, in general, only eight performance than that of the single PA layer with 90 nm
devices from the twenty fabricated exhibited this transis- pentacene (1PA/pen-90 nm). The mobility of
tor behavior, which is a yield of 20%. It appears that the 3PA/pen-90 nm was 0.05 ± 0.02 cm2/V s, which is larger
low performance compared with those on the plastic sub- than that of 1PA/pen-90 nm exhibiting 0.03 ± 0.01 cm2/
strate [32] as well as the low yield result from the high sur- V s. The sub-threshold slope of 1.20 ± 1.02 V/dec in 3PA/
face roughness of the PVP gate dielectric that has been pen-90 nm was smaller than that of 1PA/pen-90 nm
generated during the high temperature PVP dielectric pro- (2.27 ± 2.53 V/dec); and the off-state current of
cess as shown in Fig. 2b. Therefore, a low temperature pro- 3PA/pen-90 nm was 1.17 ± 2.30  10 3 pA/lm, which
cess for the gate dielectric is essential to obtain high was also smaller than that of 1PA/pen-90 nm
performances and yields for OTFTs on fabric substrates. (2.65 ± 3.43  10 2 pA/lm). These results are summarized
In order to examine the effects of the surface roughness in Table 1. The enhancement of the three PA layers is
on the performance, second coating layers of PA with attributed to the smaller surface roughness of 0.32 lm,

Fig. 4. (a) The transfer and (b) the output characteristics of OTFTs with three PA layers and 90 nm thick pentacene fabricated on PET fabric substrate.

Table 1
Summary of performance parameters of OTFTs on fabric substrates.

Device Mobility(cm2/V s) Vth (V) SS (V/dec) Ion/Ioff Ioff (pA/lm)

3PA/pen-90 nm (PA:3layer/pen:90 nm) 0.05 ± 0.02 8.06 ± 2.50 1.20 ± 1.02 8.67 ± 4.52  105 1.17 ± 2.30  10 3

1PA/pen-90 nm (PA:1layer/pen:90 nm) 0.03 ± 0.01 7.88 ± 4.02 2.27 ± 2.53 1.10 ± 1.34  104 2.65 ± 3.43  10 2

31A/pen-90 nm (PA:1layer/pen:45 nm) 0.003 ± 0.002 7.99 ± 5.22 6.45 ± 3.42 4.59 ± 3.23  102 7.39 ± 6.74  10 2
1676 G.S. Ryu et al. / Organic Electronics 15 (2014) 1672–1677

decreases because the bulk current path decreases. How-

ever, in this case 1PA/pen-45 nm with 45 nm pentacene
exhibited the larger Ioff than 3PA/pen-90 nm with 90 nm
thick pentacene. The surface roughness of PA layers may
be a cause of the large Ioff. The rougher surface of 1PA/
pen-45 nm provides the larger electric field at the bulge
region between the gate electrode and the S/D electrodes,
resulting in the large gate leakage current and thus the
large Ioff. This interpretation can be supported by the large
Ioff of 1PA/pen-90 nm similar to that of 1PA/pen-45 nm as
shown in Fig. 5 even though it has the thicker pentacene
layer.
The OTFTs with a thin pentacene layer of 45 nm were
also compared with the thick pentacene OTFTs in Fig. 5a.
Both OTFTs had only one PA layer. The performance of
the thick pentacene OTFTs was 10-fold stronger than that
of the thin pentacene OTFTs. In 45 nm thick pentacene film
there exist voids, which are the area unfilled by pentacene
molecules as shown in Fig. 5b. Meanwhile, in 90 nm thick
pentacene the film was completely filled as shown in
Fig. 5c. The voids may interrupt the carrier flow, resulting
in the smaller Ion for the 45 nm than that for 90 nm. There-
fore, the origin of the inferior performance of OTFT with
45 nm thick pentacene to 90 nm thick OTFT can be found
in the existence of the voids. Generally, on the flat sub-
strate such as glass the 45 nm thickness was thick enough
to completely fill the voids. However, on the textile sub-
strate it is found out that the 45 nm thickness is not
enough to fill the voids because of the large surface rough-
ness. In conclusion, it appears that the thick pentacene
layer compensated the channel degradation caused by
the voids in pentacene related to the rough surface of the
PVP dielectric induced from the rough surface of the single
PA layer. The OTFTs with thick pentacene and the three PA
layers produced the best results among those investigated.
Thus, it is clear that the small surface roughness of the gate
dielectric is the most important factor in obtaining good
performance in OTFTs on fabric substrates. Furthermore,
a thick semiconductor layer at least thicker than 90 nm
can mitigate the degradation caused by rough gate dielec-
tric surfaces.
The bending test was conducted by laminating the fab-
ric substrate containing OTFTs onto a plastic substrate. The
radius of the bending curve was 3 cm. After bending it
30,000 times, the characteristics of the OTFTS were
unchanged. Generally, the performance variations under
bending, which exerts compression or tension force on
Fig. 5. (a) Comparison of transfer characteristics of OTFTs with the
pentacene molecules, is related to modifying bonding
various thickness of PA and pentacene layer; the solid line for OTFTS with
the three PA layers and pentacene with a thickness of 90 nm (3PA/pen- overlap between the molecules, which results in variations
90 nm), the dark circle symbolized line for OTFTs with the one PA layer of the potential barrier for hopping, i.e. the Poisson effect
and pentacene with a thickness of 90 nm (1PA/pen-90 nm), and the dark [33]. In this case, the reason for the strong bending toler-
triangle symbolized line for OTFTs with the one PA layer and pentacene
ance is found in the fabric characteristics that absorb com-
with a thickness of 45 nm (1PA/pen-45 nm), (b) SEM picture of pentacene
layer with 45 nm thickness; there exist voids which were not filled by
pression and tension forces by moving the individual fibers
pentacene molecules, and (c) SEM picture for 90 nm thick pentacene under bending and thus prevents the bending force from
without voids. reaching the pentacene layer.
Furthermore, the reliability in air was stronger than
as shown in Fig. 2a) since the smoother surface alleviates expected. The OTFTs were exposed to air for 10 hrs without
the interruption of carrier transport related to the rough passivation and the mobility was measured in every hour,
surface. their mobility was only reduced by 2% from the initial
The surface effects of PA layer also can be found in Ioff. mobility. The empirical formula extracted from the mea-
Normally, Ioff decreases as the thickness of the active layer sured mobility according to time was to be
 G.S. Ryu et al. / Organic Electronics 15 (2014) 1672–1677 1677

the forces induced by the bending and thus maintained

the bonding orbital between the molecules of pentacene.

Acknowledgement

This work was supported by the IT R&D Program of

MKE/KEIT [10041957, Design and Development of fiber-
based flexible display] and partly by the Human Resources
Development of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP) grant funded by the
Korea government Ministry of Knowledge Economy (No.
20114030200030).

References
Fig. 6. Picture of a cotton cloth attached with PET fabric containing
OTFTs; the fabric was naturally bent and folded along the winkles of the [1] S. Park, S. Jayaraman, IEEE Eng. Med. Biol. Mag. 22 (2003) 41–48.
[2] F. Carpi, D. De Rossi, IEEE Trans. Inf. Technol. Biomed. 9 (2005) 295–
cloth.
318.
[3] E.R. Post, M. Orth (Eds.), Presented at IEEE International Symposium
on Wearable Computers, Cambridge, Massachusetts, 1997, pp. 13–
14.
l(t) = 0.05248 cm2/V s – (1.11  10 4 cm2/V s2)  t, and [4] E.R. Post, M. Orth, P.R. Russo, N. Gershenfeld, IBM Syst. J. 39 (2000)
thus the time for the mobility half reduction was about 840–860.
230 h. [5] R. Parker, R. Riley, M. Jones, D. Leo, L. Beex, T. Milson (Eds.),
Presented at International Interactive Textiles for the Warrior
In Fig. 6 the PET fabric substrate containing OTFTs was Conference, Cambridge, Massachusetts, 2002, pp. 9–11.
attached to a cotton cloth. The fabric naturally bent and [6] R.B. Katragadda, Y. Xu, Sens. Actuators, A 143 (2008) 169–174.
folded along the wrinkles of the cloth. This demonstrates [7] K. Cherenack, C. Zysset, T. Kinkeldei, N. Münzenrieder, G. Tröster,
Adv. Mater. 22 (2010) 5178–5182.
the possibility of using the OTFTs for ID tags and flexible [8] J. W Han, M. Meyyappan, Appl. Phys. Lett. 98 (2011) 192102.
displays on clothes. [9] M. Maccioni, E. Orgiu, P. Cosseddu, S. Locci, A. Bonfiglio, Appl. Phys.
Lett. 89 (2006) 143515.
[10] J.B. Lee, V. Subramanian, IEEE Trans. Electron. Dev. 52 (2005) 6509–
6516.
4. Conclusion [11] R. Gonzalez, N.J. Pinto, Synth. Met. 151 (2005) 275–278.
[12] G. Mattana, P. Cosseddu, B. Fraboni, G.G. Malliaras, J.P. Hinestroza, A.
Bonfiglio, Org. Electron. 12 (2011) 2033–2039.
Surface roughness is an essential factor in fabricating [13] M. Hamedi, R. Forchheimer, O. Inganas, Nat. Mater. 6 (2007) 357–
OTFTs on fabric substrates. In this paper, two polymer lay- 362.
ers were used to reduce the surface roughness of the PET [14] B. O’Connor, K.H. An, Y. Zhao, K.P. Pipe, M. Shtein, Adv. Mater. 19
(2007) 3897–3900.
fabric substrates. The surface roughness of the PET fabric [15] B. O’Connor, K.P. Pipe, M. Shtein, Appl. Phys. Lett. 92 (2008) 193306.
substrates, which were coated by polyurethane (PU) and [16] S. Khumpuang, K. Miyake, T. Itoh, Sens. Actuators, A 169 (2011) 378–
photo-acryl (PA) polymer layers, was reduced from 382.
[17] D.S. Hecht, L. Hu, G. Gruner, Curr. Appl. Phys. 7 (2007) 60–63.
10.05 lm to 0.32 lm. Furthermore, with the coating layers, [18] T. Yamashita, K. Miyake, T. Itoh (Eds.), Symposium on Design, Test,
the drape properties of the fabric were not altered from the Integration and Packaging of MEMS/MOEMS, Cannes, 2012, pp. 25–
original properties without the coating layers. 27.
[19] D. Chen, S. Lei, Y. Chen, Sensors 11 (2011) 6509–6516.
The mobility of the OTFTs fabricated on the PET fabric [20] T. Itoh, S. Yamashita, T. Kobayashi, N. Shibayama, K. Miyake (Eds.),
substrates was 0.05 ± 0.02 cm2/V s, with a threshold volt- Symposium on Design, Test, Integration and Packaging of MEMS/
age of 8.06 ± 2.50 V, a the sub-threshold slope of MOEMS, Cannes, Massachusetts, 2012, pp. 25–27.
[21] J.F. Gu, S. Gorgutsa, M. Skorobogatiy, Smart Mater. Struc. 19 (2010)
1.2 ± 1.02 V/dec, and an on/off current ratio of
115006.
8.67 ± 4.52  105. Generally, the OTFT performance was [22] Z. Wang, L. Zhang, J.L. Volakis, Textiles Light Ind. Sci. Technol. (TLIST)
worse than that on plastic and glass substrates, and this 2 (2013) 105–112.
[23] R. Bhattacharya, M.M. de Kok, J. Zhou, Appl. Phys. Lett. 95 (2009)
resulted from the surface roughness of the PU/PA-coated
223305.
substrates, which remained larger than the plastic and [24] Y.H. Kim, S.K. Park, D.G. Moon, W.K. Kim, J.I. Han, Displays 25 (2004)
glass substrates. Furthermore, the high temperature pro- 167–170.
cess of the PVP dielectric enhanced the roughness further [25] S. Steudel, S.D. Vusser, S.D. Jonge, D. Janssen, S. Verlaak, J. Genoe, P.
Heremans, Appl. Phys. Lett. 85 (2004) 4400–4402.
and thus degraded the performance. In order to reduce [26] S.E. Fritz, T.W. Kelley, C.D. Frisbie, J. Phys. Chem. B 109 (2005)
the roughness and improve the performance, a low tem- 10574–10577.
perature process is required. [27] Chua et al., Adv. Mater. 16 (2004) 1609.
[28] Steudel et al., Appl. Phys. Lett. 85 (2004) 4400.
The performance degradation caused by the surface [29] E. de Bilbao, D. Soulat, G. Hivet, J. Launay, A. Gasser, Int. J. Mater.
roughness was partly compensated by a thick pentacene Form. 1 (2008) 835–838.
layer at least larger than 90 nm. The mobility was [30] H.S. Byun, Y.X. Xu, C.K. Song, Thin Solid Films 493 (2005) 278–281.
[31] C.K. Song, M.K. Jung, B.W. Koo, J. Korean Phys. Soc. 39 (2001) 271–
increased 10-fold by increasing the thickness of the penta- 274.
cene layer from 45 nm to 90 nm. Furthermore, after bend- [32] M.W. Lee, C.K. Song, J. Korean Phys. Soc. 54 (2009) 514–517.
ing it 30,000 times, the performance of the OTFTs remained [33] T. Sekitani, Y. Kato, S. Iba, H. Shinaoka, T. Someya, T. Sakurai, S.
Takagi, Appl. Phys. Lett. 86 (2005) 073511.
unchanged. The fabric substrate itself appeared to absorb