SS10198 81 Total pages 4

Japanese Journal of Applied Physics 50 (2011) 04DF11 REGULAR PAPER

DOI: 10.1143/JJAP.50.04DF11

Transparent Oxide Thin-Film Transistors Using n-(In2 O3 )0:9 (SnO2 )0:1 /InGaZnO4
Modulation-Doped Heterostructures
Satoshi Taniguchi
, Mikihiro Yokozeki, Masao Ikeda1 , and Toshi-kazu Suzuki2
Core Device Development Group, Sony Corporation, 4-14-1 Asahicho, Atsugi, Kanagawa 243-0014, Japan
1
Advanced Material Laboratory, Sony Corporation, 4-14-1 Asahicho, Atsugi, Kanagawa 243-0014, Japan
2
Center for Nano Materials and Technology, Japan Advanced Institute of Science and Technology (JAIST),
1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan
Received September 22, 2010; revised November 2, 2010; accepted November 8, 2010; published online April 20, 2011

We investigated transparent oxide thin-film transistors (TFTs) using n-(In2 O3 )0:9 (SnO2 )0:1 /InGaZnO4 (n-ITO/IGZO) modulation-doped
heterostructures, which are effective in achieving high carrier mobilities. From transmittance measurements and UV photoemission spectroscopy,
n-ITO/IGZO modulation-doped heterostructures are expected to realize the type-II energy band lineup, in which both the conduction band
minimum and the valence band maximum of n-ITO are higher in energy than those of IGZO. Van der Pauw Hall measurements revealed Hall
mobility enhancement and two-dimensional behavior of electrons at the n-ITO/IGZO interface. Using the n-ITO/IGZO modulation-doped
heterostructures, we obtained TFTs with higher electron mobility than that of IGZO TFTs. We consider that modulation doping is a promising
method for performance improvements of TFTs using transparent oxide semiconductors. # 2011 The Japan Society of Applied Physics

of this work is to investigate transparent oxide TFTs using

1. Introduction
Since the first demonstration of thin-film transistors (TFTs)
using amorphous InGaZnO4 (a-IGZO),1) which is a trans-
parent oxide semiconductor (TOS) with a high electron
mobility of 10 cm2 /(Vs) and high reliability compared
with amorphous Si, TOSs have been widely investigated as a
TFT channel material to meet the requirements of driver
circuits for flat panel displays, such as active-matrix organic
light emitting displays or liquid crystal displays. Although
instability under current bias stress has been a problem of
TOS TFTs, various protective materials, such as Al2 O3 ,2,3)
have been proposed to obtain TOS TFTs with improved
stability, which are suitable for mass production. Moreover,
owing to the high transparency4) or low fabrication temper-
ature5,6) of TOS TFTs, their application to new displays, such
as transparent displays or flexible displays is expected.
To continuously satisfy the requirements of next-genera-
tion displays, the realization of TOS carrier mobilities
comparable to those of polycrystalline Si is an important
issue.7) As a method of achieving higher carrier mobilities,
modulation doping, which has been used for III–V semi-
conductors,8,9) is effective. In this method, reduced impurity
scattering owing to the spatial separation of carriers and
impurities leads to the enhancement of mobility. For TOSs,
the theoretical analysis of modulation doping10) and natural
modulation doping11) have been reported. In the former work,
high electron mobilities were calculated for ZnO:Al/InGaO3
modulation-doped heterostructures, in which ZnO:Al has a
higher conduction band energy minimum than that of InGaO3
and electrons transfer into InGaO3 . In the latter work, high
hole mobilities were obtained for LaO:Mg/CuSe:Mg hetero-
structures with uniform impurity doping. Since acceptors
were activated only in the LaO:Mg layer, the valence band
energy maximum of which is lower than that of CuSe:Mg,
modulation doping were naturally realized. These high
carrier mobilities were attributed to the spatial separation of
carriers and ionized impurities.
However, there are few experimental studies on designed
modulation-doped heterostructures using TOSs. The purpose


E-mail address: Satoshi.Taniguchi@jp.sony.com
04DF11-1
n-(In2 O3 )0:9 (SnO2 )0:1 /a-IGZO (n-ITO/IGZO) modulation-
doped heterostructures. On the basis of energy band lineup
estimation using transmittance measurements and UV
photoemission spectroscopy, we suppose that ITO and
IGZO form a type-II heterostructure, in which both the
conduction band minimum and the valence band maximum
of ITO are higher in energy than those of IGZO. From
van der Pauw Hall measurements, compared with ITO or
IGZO single films, we observed enhancements in the Hall
mobility for n-ITO/IGZO modulation-doped heterostruc-
tures and a two-dimensional (2D) behavior of electrons at
the n-ITO/IGZO interface. Moreover, by the application of
the n-ITO/IGZO modulation-doped heterostructure to TFTs,
higher electron mobilities than those of IGZO TFTs were
realized. We consider modulation-doping to be a promising
method for improving the performance of TFTs using TOSs.
2. Experiments and Results
2.1 Energy band lineup of ITO/IGZO
Throughout this work, ITO and IGZO films were deposited
at 80  C by Ar-plasma DC sputtering with facing targets in
O2 gas ambient. Doping concentrations of all the films were
controlled by varying the O2 partial pressure during
deposition. In order to investigate the energy band lineup
of ITO/IGZO, we carried out transmittance measurements
and UV photoemission spectroscopy (UPS) using ITO and
IGZO single films on c-sapphire substrates. Figure 1 shows
the Tauc plots obtained from the transmittance measure-
ments of ITO and IGZO. The optical bandgap energy for
ITO is estimated to be 3:6 eV using a square Tauc plot,
which is suitable for polycrystalline materials. In fact, as
shown later, our ITO is polycrystalline. On the other hand,
the optical bandgap energy for a-IGZO is estimated to be
3:0 eV using a square-root Tauc plot, which is suitable for
amorphous materials. Also as shown later, our IGZO is
amorphous. In addition, from the UPS of each film, it is
found that the valence band energy maximum of ITO is
higher than that of IGZO with a valence band discontinuity
EV  0:5{0:6 eV. It is reported that, for In2 O3 , the
fundamental bandgap energy is 0:8 eV smaller than the
optical bandgap energy since the valence band energy
# 2011 The Japan Society of Applied Physics
Jpn. J. Appl. Phys. 50 (2011) 04DF11
Fig. 1. (Color online) Tauc plots of ITO and IGZO. Optical bandgap
energy for ITO is estimated to be 3:6 eV using square Tauc plot. Optical
bandgap energy for a-IGZO is estimated to be 3:0 eV using square-root
Tauc plot.
Fig. 2. (Color online) (a) Supposed type-II energy band lineup of ITO
and IGZO, and (b) n-ITO/IGZO modulation-doped heterostructure.
maximum lies inside the optical bandgap.12) From these
values, we suppose the fundamental bandgap energy of ITO
to be 2:8 eV for our films and the energy band lineup of
ITO and IGZO with the conduction band discontinuity
EC  0:3{0:4 eV, as shown in Fig. 2(a). Therefore, we
expect to realize type-II n-ITO/IGZO modulation-doped
heterostructures, as shown in Fig. 2(b), in which electrons
transfer from n-ITO to IGZO and accumulate on the IGZO
side at the heterointerface.
2.2 n-ITO/IGZO modulation-doped heterostructure
We deposited n-ITO (10 nm)/undoped-IGZO (340 nm)
modulation-doped heterostructures on c-sapphire substrates,
as shown in Fig. 3(a), where ITO was deposited on IGZO
sequentially. Figure 4 shows cross-sectional images of the
heterostructure obtained by transmission electron micro-
scope. As seen in this figure, n-ITO exhibits a crystal grain
indicating the polycrystalline phase, but IGZO is amorphous.
We confirmed a smooth heterointerface between n-ITO and
IGZO.
We investigated electron transport properties of the
modulation-doped heterostructures using van der Pauw Hall
measurements. For comparison with modulation-doped
heterostructures, n-ITO (10 nm)/undoped-ITO (340 nm)
[Fig. 3(b)] and n-IGZO (340 nm) [Fig. 3(c)] on c-sapphire
04DF11-2
S. Taniguchi et al.
Fig. 3. (Color online) Sample structures: (a) n-ITO/IGZO, (b) n-ITO/
ITO, and (c) n-IGZO.
Fig. 4. (Color online) Cross-sectional transmission electron microscope
images of the n-ITO/IGZO modulation-doped heterostructure.
Fig. 5. (Color online) Relationship between sheet electron concentration
and Hall mobility at room temperature.
substrates were also prepared and measured. For all
structures, relatively thick films were employed to avoid
the influence of the film/substrate interface on electron
transport. For a systematic comparison, we employed sheet
electron concentrations in the order of 1013 –1014 /cm2 for all
samples. Figure 5 shows the relationship between sheet
electron concentration and Hall mobility. We find that the
modulation-doped heterostructures show high Hall mobili-
ties of 40 cm2 /(Vs) in comparison with other structures.
This Hall mobility enhancement can be attributed to weaker
electron scattering for the modulation-doped heterostruc-
tures, in which electrons are spatially separated from ionized
donors in n-ITO. Moreover, we observed a logarithmic
temperature dependence of conductance correction for a
wide range of sheet electron concentrations. This is a
characteristic behavior for two-dimensional systems,13) and
thus indicates 2D electron gas formation at the heterointer-
face.
# 2011 The Japan Society of Applied Physics
SS10198
Jpn. J. Appl. Phys. 50 (2011) 04DF11 S. Taniguchi et al.

2.3 Fabrication and characteristics of TFTs

We applied the n-ITO/IGZO modulation-doped heterostruc-
ture to TFTs. Bottom-gate n-ITO/IGZO TFTs were
fabricated as follows. After n-ITO (10 nm)/IGZO (30 nm)
deposition on a SiO2 (100 nm)/n-Si substrate, device mesas
were isolated by dilute HCl etchant, followed by ohmic
metallization for source and drain electrodes through Au
(200 nm)/Ni (10 nm) deposition by resistive heater evapora-
tion and lift-off. To obtain a reasonable threshold voltage of
n-ITO/IGZO TFTs, we employed modulation doping giving
Fig. 6. (Color online) Schematics of n-ITO/IGZO and IGZO TFTs, with
the sheet electron concentrations of the middle of 1012 /cm2 , source (S), drain (D), and gate (G) electrodes.
which is much lower than those of Hall measurement
samples. For comparison, bottom-gate TFTs using IGZO
(40 nm)/SiO2 (100 nm)/n-Si substrate were also fabricated. ohmic electrodes and the n-ITO/IGZO interface is a reason
We carried out postannealing only for the IGZO TFTs at for the low contact resistance. Figures 8(a) and 8(b) show
350  C for 30 min under O2 ambient, which is optimal for square-root-current and logarithm-current transfer charac-
obtaining better device performance of IGZO TFTs.14,15) teristics of TFTs at a drain–source voltage of 10 V,
Figure 6 shows the schematics of both TFTs. The channel respectively. From Fig. 8(a), for the n-ITO/IGZO TFT and
length and width of TFTs are 10 and 100 m, respectively. the IGZO TFT, we obtain threshold voltages of 5:1 and
Output characteristics of both TFTs are shown in Fig. 7. 2.2 V, and electron mobilities of 20 and 7.5 cm2 /(Vs),
In comparison with the IGZO TFT, a much higher drain respectively. Owing to the modulation doping giving a
current of 20 mA/mm is obtained for the n-ITO/IGZO sheet electron concentration in the mid-1012 /cm2 range, a
TFT. From Fig. 7, we find that on resistances of the n-ITO/ negative threshold voltage is observed for the n-ITO/IGZO
IGZO TFT estimated from the slope in the linear region are TFT. To obtain the electron mobilities, we assume a
lower than those of the IGZO TFT. This can be attributed dielectric constant of 11.516) for IGZO. For the n-ITO/
to not only higher electron mobility but also lower contact IGZO TFT, a much higher electron mobility is realized.
resistance. We consider that the short distance between the From Fig. 8(b), we obtain the on-off ratio of 109 for the n-

Fig. 7. (Color online) Output characteristics of n-ITO/IGZO and IGZO TFTs.

Fig. 8. (Color online) (a) Square-root-current and (b) logarithm-current transfer characteristics of n-ITO/IGZO and IGZO TFTs.

04DF11-3 # 2011 The Japan Society of Applied Physics

SS10198
Jpn. J. Appl. Phys. 50 (2011) 04DF11
Table I. Parameters of n-ITO/IGZO and IGZO TFTs.
Threshold Mobility Subthreshold
voltage [cm2 /(Vs)] swing
(V) (V/decade)
n-ITO/IGZO TFT 5:1 20 1 >109
IGZO TFT 2.2 7.5 0.5
Fig. 9. (Color online) Relationship between transconductance and gate-
source voltage of n-ITO/IGZO and IGZO TFTs.
ITO/IGZO TFT and 108 for the IGZO TFT, both of which
are sufficient for practical use. On the other hand, Fig. 8(b)
exhibits a subthreshold swing of 1 V/decade for the n-ITO/
IGZO TFT and 0.5 V/decade for the IGZO TFT; the
subthreshold swing for the n-ITO/IGZO TFT is poor in
comparison with that for the IGZO TFT. Since an
improvement in the subthreshold swing is often realized
by postannealing for IGZO TFTs14) as in the case of our
samples, we expect the postannealing of n-ITO/IGZO TFTs
to be effective in improving their subthreshold swing. The
obtained TFT parameters are summarized in Table I.
Figure 9 shows the relationship between gate-source voltage
and transconductance gm ¼ @ID =@VGS , where ID and VGS are
drain current and gate-source voltage, respectively. For the
IGZO TFT, transconductance increases monotonically with
increasing gate-source voltage. On the other hand, for the n-
ITO/IGZO TFT, we observe an ‘‘N-shaped’’ behavior of the
transconductance; after the region of decrease from 9 to
16 V, the transconductance again increases. We consider
that this behavior can be attributed to the influence of the
IGZO/SiO2 interface channel in addition to the n-ITO/
IGZO interface channel. With increasing gate-source
voltage, a turn-on of the additional IGZO/SiO2 interface
channel results in the re-increasing transconductance.
Similar behavior of the additional channel has been observed
for III–V compound semiconductor heterostructure field
effect transistors.17)
S. Taniguchi et al.
3. Summary
On–off
ratio We investigated TFTs using n-ITO/IGZO modulation-
doped heterostructures. On the basis of the results of
transmittance measurements and UV photoemission spec-
>108 troscopy, we suppose that n-ITO/IGZO modulation-doped
heterostructures have the type-II energy band lineup, in
which both the conduction band minimum and the valence
band maximum of n-ITO are higher in energy than those of
IGZO. With this structure, it is possible to realize electron
transfer from n-ITO to IGZO and electron accumulation on
the IGZO side at the heterointerface. Van der Pauw Hall
measurements revealed Hall mobility enhancement and two-
dimensional behavior of electrons at the n-ITO/IGZO
interface. Using the n-ITO/IGZO modulation-doped hetero-
structures, we obtained TFTs with higher electron mobility
than that of IGZO TFTs, although further improvement of
subthreshold swing is needed. We consider modulation
doping to be a promising method for improving the
performance of TFTs using TOSs.
Acknowledgements
We would like to thank to H. Koga and H. Yoshida for
technical support and fruitful discussion.
1) K. Nomura, H. Ohata, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono:
Nature 432 (2004) 488.
2) T. Arai, N. Morosawa, K. Tokunaga, Y. Terai, E. Fukumoto, T. Fujimori,
T. Nakayama, T. Yamaguchi, and T. Sasaoka: SID Int. Symp. Dig. Tech.
Pap. 41 (2010) 1033.
3) J. M. Lee, I. T. Cho, J. H. Lee, W. S. Cheong, C. S. Hwang, and H. I.
Kwon: Appl. Phys. Lett. 94 (2009) 222112.
4) K. Nomura, A. Takagi, T. Kamiya, H. Ohta, M. Hirano, and H. Hosono:
Jpn. J. Appl. Phys. 45 (2006) 4303.
5) D. H. Kim, N. G. Cho, H. G. Kim, H. S. Kim, J. M. Hong, and I. D. Kim:
Appl. Phys. Lett. 93 (2008) 032901.
6) W. Lim, E. A. Douglas, S. H. Kim, D. P. Norton, S. J. Pearton, F. Ren, H.
Shen, and W. H. Chang: Appl. Phys. Lett. 93 (2008) 252103.
7) T. Kamiya, K. Nomura, and H. Hosono: J. Disp. Technol. 5 (2009) 273.
8) R. Dingle, H. L. Stormer, A. C. Gossard, and W. Wiegmann: Appl. Phys.
Lett. 33 (1978) 665.
9) T. Mimura, S. Hiyamizu, T. Fujii, and K. Nanbu: Jpn. J. Appl. Phys. 19
(1980) L225.
10) J. J. Robbins and C. A. Wolden: Appl. Phys. Lett. 83 (2003) 3933.
11) T. Kamiya and H. Hosono: Int. J. Appl. Ceram. Technol. 2 (2005) 285.
12) A. Walsh, J. L. F. Da Silva, S.-H. Wei, C. Körber, A. Klein, L. F. J. Piper,
A. DeMasi, K. E. Smith, G. Panaccione, P. Torelli, D. J. Payne, A.
Bourlange, and R. G. Egdell: Phys. Rev. Lett. 100 (2008) 167402.
13) T. Ando, A. B. Fowler, and F. Stern: Rev. Mod. Phys. 54 (1982) 437.
14) H. Hosono, K. Nomura, Y. Ogo, T. Uruga, and T. Kamiya: J. Non-Cryst.
Soilds 354 (2008) 2796.
15) H. Q. Chiang, B. R. McFarlane, D. Hong, R. E. Presley, and J. F. Wager:
J. Non-Cryst. Soilds 354 (2008) 2826.
16) H.-J. Chung, J. H. Jeong, T. K. Ahn, H. J. Lee, M. Kim, K. Jun, J.-S. Park,
J. K. Jeong, Y.-G. Mo, and H. D. Kim: Electrochem. Solid-State Lett. 11
(2008) H51.
17) C. C. Eugster, T. P. E. Broekaert, J. A. D. Alamo, and C. G. Fonstad: IEEE
Electron Device Lett. 12 (1991) 707.
04DF11-4 # 2011 The Japan Society of Applied Physics
SS10198