Content uploaded by Dae Hwan Kim
Author content
All content in this area was uploaded by Dae Hwan Kim on Jan 11, 2017
Content may be subject to copyright.
Cation composition effects on electronic structures of In-Sn-Zn-O
amorphous semiconductors
Ji-Young Noh, Hanchul Kim, Ho-Hyun Nahm, Yong-Sung Kim, Dae Hwan Kim et al.
Citation: J. Appl. Phys. 113, 183706 (2013); doi: 10.1063/1.4803706
View online: http://dx.doi.org/10.1063/1.4803706
View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v113/i18
Published by the American Institute of Physics.
Additional information on J. Appl. Phys.
Journal Homepage: http://jap.aip.org/
Journal Information: http://jap.aip.org/about/about_the_journal
Top downloads: http://jap.aip.org/features/most_downloaded
Information for Authors: http://jap.aip.org/authors
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
Cation composition effects on electronic structures of In-Sn-Zn-O
amorphous semiconductors
Ji-Young Noh,
1,2
Hanchul Kim,
1
Ho-Hyun Nahm,
2
Yong-Sung Kim,
2,3,a)
Dae Hwan Kim,
4
Byung-Du Ahn,
5
Jun-Hyung Lim,
5
Gun Hee Kim,
5
Je-Hun Lee,
5
and Junho Song
5
1
Department of Physics, Sookmyung Women’s University, Seoul 140-742, South Korea
2
Korea Research Institute of Standards and Science, Daejeon 305-340, South Korea
3
Department of Nano Science, University of Science and Technology, Daejeon 305-350, South Korea
4
School of Electrical Engineering, Kookmin University, Seoul 136-702, South Korea
5
Samsung Display Co., Ltd., Yongin 446-711, South Korea
(Received 11 January 2013; accepted 16 April 2013; published online 9 May 2013)
Based on density-functional theory calculations, the effects of cation compositions on electronic
structures of In-Sn-Zn-O amorphous semiconductors were investigated. We considered various
composition ratios of In, Sn, and Zn in O stoichiometric condition, and found that the conduction
band minimum (CBM) energy level decreases and the valence band tail (VBT) energy level
extent increases as the sum of In and Sn ratios (RInþRSn) increases. The CBM lowering is
attributed to the increased overlap of the In-5sand Sn-5sorbitals as the RInþRSn increases, and
correspondingly the electron effective masses (m
e) are found to be reduced. The VBT increase is
found to be due to the increased density of the In and Sn atoms, near which the O-2pinter-site
pprcoupling is larger than that near the Zn atoms. The acute O-(In,Sn)-O angles are suggested
to be structurally important, giving the stronger O-O pprcoupling. V
C2013 AIP Publishing LLC.
[http://dx.doi.org/10.1063/1.4803706]
I. INTRODUCTION
Amorphous oxide semiconductors (AOS) have high
electron mobility (>10 cm
2
V
1
s
1
) even in the amorphous
phase.
1
The amorphous InGaZnO
4
is a prototype AOS, and
various kinds of AOS with alternative cations (Sn, Hf, Zr,
Al, Si, and so on) have been suggested.
2–6
Basically, the
high electron mobility originates from the cation s-like orbi-
tals of the conduction band (CB) states. Since the sorbitals
are spherical, their inter-site overlap is robust to the orienta-
tional disorder of the amorphous structure.
7
It is in stark con-
trast to the hydrogenated amorphous Si (a-Si:H), which has
the sp
3
-hybridized conduction bands and exhibits severe
localization of the sp
3
orbitals in amorphous phase (giving
the conduction band tail states). Thus, the electron mobility
of a-Si:H is degraded to about 1cm
2
V
1
s
1
from that of
the crystalline Si of about 200 cm
2
V
1
s
1
.
With the high mobility, the AOS has process compati-
bility. The low (room) temperature thin-film deposition pro-
cess (usually sputtering) makes it possible to manufacture
large-area thin-films on a variety choice of substrates, such
as soda-lime glasses and flexible organic plastics, with low-
cost and a high production (deposition) rate.
8,9
Electronics
on a large-area transparent substrate for commercial flat-
panel displays have used a-Si:H thin-film transistors (TFTs),
and large-area low-cost solar cells have been made by using
a-Si:H pn junctions. The requirement of high speed TFTs for
the next-generation high-resolution larger-area flat-panel dis-
plays makes the use of a-Si:H TFTs doubted due to its funda-
mental limit of electron mobility, and a new way by utilizing
the AOS-based TFTs is nowadays intensively pursued.
10
However, the facing bottleneck of the AOS-based TFTs
for display applications is their instabilities under bias and/or
illumination stress (IS). By positive gate-bias stress (PBS) in
AOS-based TFTs, the threshold voltage (Vth) of the TFTs is
known to be positive-shifted,
11–17
and by negative bias and
illumination stress (NBIS)
4,5,17–24
or only by IS,
17,25–29
the
Vth is largely negative-shifted. The NBIS is known to be the
most serious among the stresses,
18
and it is important in
active matrix liquid crystal displays (AM-LCD), since it is
applied to the TFTs in AM-LCD steadily in the normal off-
state stand-by condition. On the other hand, the PBS is stead-
ily applied to the TFTs in active matrix organic light
emitting diode (AM-OLED) displays during the normal state.
Understanding the microscopic mechanisms of the insta-
bilities in AOS-based TFTs has been recently advanced. For
the PBS instability, (i) it has been suggested that the elec-
trons attracted to the gate electrode are trapped at the inter-
face between the AOS channel and the gate insulator and/or
injected into the gate insulator.
11–13,17
(ii) The electron traps
at H
2
O-related trap sites have been also suggested.
14
The
electron traps can make a positive-shift of the Vth electro-
statically, but their meta-stability is not clear yet. (iii)
Surface doping induced by O
2
and H
2
O molecules has been
suggested as another mechanism.
15
(iv) More importantly, it
has been proposed that acceptor-like deep states is generated
in AOS by the PBS.
16
For the NBIS and IS instabilities, a
few plausible models have been suggested. (i) Holes are gen-
erated by the light illumination and trapped at the interface
and/or injected into the gate insulator.
4,22,23
However, the IS
instability has been observed not only in AOS TFTs but also
in AOS thin films without a gate insulator through persistent
photo-conductivity (PPC),
26,29
and thus there has been con-
sidered to be an intrinsic mechanism in AOS materials.
22,26
a)
yongsung.kim@kriss.re.kr
0021-8979/2013/113(18)/183706/7/$30.00 V
C2013 AIP Publishing LLC113, 183706-1
JOURNAL OF APPLIED PHYSICS 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
(ii) NBIS or IS desorbs surface O atoms (reduction) and
accordingly dope electrons,
5
and an overlying passivation
layer on AOS TFTs has been adopted to avoid it.
5
Although
somewhat improved, the NBIS and IS instabilities have been
still serious even with avoiding the surface effect.
4,19,20,22
(iii) Photo-ionized O-vacancies (VO) persist with slow elec-
tron recombination.
19–22,24,30,31
The VOcan give photo-
carriers and be a hole-trap, but the meta-stability of the
ionized VOremains unclear. The recovery time of the
negative-shifted Vth in AOS TFTs
19,26–28
and the persistency
of the PPC in AOS thin films
26,29
are an order of days at
room temperature, and the energy barrier from the meta-
stable to the stable state should be very high. The activation
energy has been reported to be in the range of 0.9–1.0 eV
experimentally.
29
(iv) Meta-stable O2
2peroxide defect for-
mation in AOS by NBIS or IS has been recently proposed.
32
With excited holes in the valence band tail (VBT) states in
AOS, the empty ppr-like VBT states induce a driving force
to form an O-O bond through the pprbonding interaction.
Along with the O-O bond formation, the pprstate is height-
ened up into the CBs, and the electrons excited are doped
with occupying the conduction band minimum (CBM) state.
The energy barrier from the meta-stable O2
2peroxide to the
normal amorphous oxide configuration is 0.97 eV in the
hybrid density-functional theory (DFT) calculations.
32
In order to improve the stability of AOS-based TFTs
against the bias and illumination stresses, a variety of
attempts, such as a novel TFT architecture (including adopt-
ing a passivation layer),
5,23,33,34
changing the gate insulator
material,
4,24
process design (for examples, O
2
or water
annealing),
21,24,35
impurity doping in AOS,
2
and optimizing
the cation species and compositions of the AOS materi-
als,
26,36
have been made. The improvement of stability
should not accompany (or at least minimize) any degradation
of the electron mobility in the AOS. In this study, we investi-
gate the cation composition effects on the electronic struc-
tures of In-Sn-Zn-O amorphous semiconductors to optimize
both the stability and the mobility. We find that the CBM
and VBT show nearly monotonic changes with respect to the
In and Sn cation composition ratios. The In-5sand Sn-5sor-
bital overlaps and the stronger O-O pprorbital coupling
near the In and Sn atoms are found to be responsible for the
CBM and VBT changes, respectively. The results indicate
that the electron mobility and the stability are in a trade-off
relation, when they are controlled by the In and Sn contents
in amorphous In-Sn-Zn-O.
II. METHODS
We performed DFT calculations as implemented in the
Vienna ab initio simulation package (VASP) code.
37,38
The
projector-augmented wave pseudopotentials
39
and the plane-
wave basis set with a kinetic energy cutoff of 400 eV were
used. The local-density approximation
40
plus U(LDAþU)
41
was used for the exchange correlation energy with the on-
site Coulomb energy Uof 6 eV for the In, Ga, and Zn d
states. The Sn 4dwere treated as core states. For the
Brillouin zone integration in DFT molecular dynamics simu-
lations, a single k-point at (0.25, 0.25, 0.25) was used. The
total energy difference between the single k-point and the
222 Monkhorst-Pack mesh is only a few meV. In elec-
tronic density-of-states calculations, we used the 3 33
mesh including the Ck-point. The LDAþUmethod with
U¼6 eV was tested for crystalline ZnO. The major role of U
for the Zn-3dstates in ZnO is correction of the energy level
position of the Zn-3dbands.
42
The Zn-3dbands are located
near 7 eV below the valence band maximum (VBM) in our
LDA þUcalculations, and is close to the X-ray emission
spectroscopy measurements.
43
With this correction, the band
gap underestimation problem of the LDA is partially
improved by suppressing the spurious Zn-3dand O-2p
hybridization in the LDA. The calculated LDA þUband
gap of ZnO was 1.74 eV, while the LDA band gap was
0.80 eV. The calculated LDA and LDA þUgaps are consist-
ent with the previous LDA þUcalculations,
44
but still lower
than the experimental,
45
hybrid functional,
46,47
and GW
48–53
results. Thus, here we focus only on the changes of the band
gaps with respect to the cation compositions.
The considered metal compositions of the In-Sn-Zn-O
amorphous semiconductors are listed in Table I. We used
cubic supercells. The number of atoms for each atom species
(NIn,NSn ,NZn, and NO) and the total number of atoms (N
t
)in
the supercell are shown in Table I. We only considered O
stoichiometric condition as in the binary oxides of In
2
O
3
,
SnO
2
, and ZnO. The supercell volumes (V
t
) are derived from
the experimental mass densities of In
2
O
3
(7180 kg/m
3
), SnO
2
(6900 kg/m
3
), and ZnO (5600 kg/m
3
). We also considered
amorphous InGaZnO
4
as a reference AOS. The number of
atoms in the InGaZnO
4
cubic supercell is N
t
¼168 with
NIn ¼24, NGa ¼24, NZn ¼24, and NO¼96. The InGaZnO
4
supercell volume is also derived from the experimental mass
densities [Ga
2
O
3
(6000 kg/m
3
)] (V
t
¼1972.9 A
˚
3
). The aver-
age volume per unit formula obtained by this a priori
approach for the amorphous phase InGaZnO
4
is 9.6% larger
than the theoretically optimized volume per unit formula for
the crystalline phase InGaZnO
4
. The amorphous structures
were generated by melt-and-quenching molecular dynamics
simulations within the Nos
e canonical ensemble. A random
amorphous oxide structure was melted at 3000 K for 6 ps
and then quenched down to 300 K at a quenching rate of
900 K/ps. The amorphous structure was then optimized with
the static DFT calculations until the Hellmann-Feynman
forces were less than 0.02 eV/A
˚.
TABLE I. Seven considered amorphous In-Sn-Zn-O semiconductors in this
work are listed. NIn,NSn ,NZn, and NOare the numbers of In, Sn, Zn, and O
atoms, respectively, in the supercell. N
t
is the total number of atoms in the
supercell, and V
t
is the total supercell volume. RInþRSn is the sum of the In
and Sn composition ratios.
Sample NIn NSn NZn NON
t
V
t
(A
˚
3
)RIn þRSn
ISZO#1 24 24 24 108 180 2221.0 0.667
ISZO#2 12 24 24 90 150 1835.6 0.600
ISZO#3 6 27 27 90 150 1824.2 0.550
ISZO#4 10 16 24 71 121 1481.1 0.520
ISZO#5 6 8 16 41 71 869.2 0.467
ISZO#6 10 12 28 67 117 1432.5 0.440
ISZO#7 6 18 36 81 151 1714.9 0.400
183706-2 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
The limited number of atoms with the supercell struc-
tures and the melt-and-quenching simulation method to gen-
erate the amorphous structures can restrict accurate
description of the experimentally obtained amorphous struc-
tures (typically by sputtering and annealing). Nevertheless,
such a structural modeling has been widely accepted in pre-
vious DFT studies and explained the experimental results
such as extended X-ray absorption fine structures
(EXAFS).
54
We could also reproduce the well known trade-
off relation between the mobility and stability of AOS, and
its chemical origins are focused on here.
III. RESULTS
A. Trends
The calculated electronic densities of states (DOS) in
LDAþUfor the considered In-Sn-Zn-O amorphous semicon-
ductors are shown in Fig. 1. We aligned the energy levels by
comparing the In-4dlevels of the In-Sn-Zn-O with those of
the crystalline InGaZnO
4
. Minimizing the absolute value of
the difference of the In-4dpartial DOS between the amor-
phous In-Sn-Zn-O and the crystalline InGaZnO
4
is found to
yield a good alignment of the energy levels, as shown in Fig.
1(a). We can see the overall agreements of the deep O-2s,
In-4d, Zn-3d, and O-2pstates in level positions between the
difference compositions. The zero energy level is set to the
top-most valence band state level of the crystalline
InGaZnO
4
. Near the valence and conduction band edges, the
changes of the VBT extents in energy level and the CBM
levels are found, as shown in Fig. 1(b).
We plot the top-most VBT levels, the CBM levels, and
their differences (the fundamental LDAþUenergy gaps) as
a function of the sum of the In and Sn ratios (RInþRSn )in
Fig. 2(a). We define here the composition ratio of a metal
cation as R
M
¼N
M
/(NIn þNSn þNZn), where Mis one of the
In, Sn, and Zn. It is found that the levels show almost mono-
tonic changes with respect to the RInþRSn ratio. We could
not find any such correlations of the energy levels with the
other parameters, for example, with the In ratio (RIn ) nor
with the Sn ratio (RSn), while necessarily there are correla-
tions with the Zn ratio [RZn ¼1ðRIn þRSn)]. The CBM
level is found to be lowered, the top-most VBT level is
raised, and the band gap is correspondingly narrowed, as the
RInþRSn ratio increases. Although the LDAþUexchange
correlation energy in DFT is well known to underestimate
the band gaps severely, here we are only interested in the
changes of the band edge levels with respect to the cation
compositions and their chemical origins, qualitatively.
Figure 2(b) shows the calculated electron effective masses
(m
e) of the In-Sn-Zn-O’s. The m
eshows a monotonic
decrease as the RInþRSn ratio increases.
B. Conduction band minimum
The CBM level lowering as the RInþRSn ratio increases
can be understood from the cation metal s-orbital overlaps.
In Fig. 3(a), the calculated portions of the s-projected CBM
density of states for the In-5s, Sn-5s, Zn-4s, and O-2sare
shown. The CBM states of the In-Sn-Zn-O’s are found to
FIG. 1. Calculated total electronic densities of states for the considered In-
Sn-Zn-O samples (a) in a wide range of the energy level and (b) near the va-
lence and conduction band edges. With decreasing the sum of the In and Sn
ratios from the ISZO#1, the density of states is shifted by þ0.02 eV
1
A
˚
3
.
FIG. 2. (a) Calculated top-most VBT levels (filled circles), CBM levels
(filled rectangles), and energy gaps (solid line) in LDAþUare plotted as a
function of the sum of the In and Sn ratios (RInþRSn). (b) Calculated elec-
tron effective masses (filled triangles) are plotted as a function of the
RInþRSn.
FIG. 3. Calculated (a) atomic-orbital portions (s-orbital projected DOS) of
the CBM state and (b) average inter-atomic distances between the In and Sn
atoms, as a function of the RInþRSn ratio in amorphous In-Sn-Zn-O.
183706-3 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
mainly consist of the In-5sand Sn-5satomic orbital states.
The Sn-5scharacter is found to be slightly stronger than the
In-5sfor the CBM in some metal compositions, and the sum
of the In-5sand Sn-5sportions is approximately as high
as 80%, irrespective of the metal compositions. Both the
In-5sand Sn-5sorbitals with similar contributions form
the main channel of the electron conduction in amorphous
In-Sn-Zn-O.
On the other hand, the effective inter-atomic distances
between the In and Sn atoms in In-Sn-Zn-O are changed
depending on the cation compositions. As shown in Fig.
3(b), the average inter-atomic distances between the In and
Sn atoms are closer as the RInþRSn ratio increases. The aver-
age inter-atomic distance between the In atoms in InGaZnO
4
is 4.35 A
˚, of which the RIn ratio is 33%. The average inter-
atomic distance between the In and Sn atoms becomes closer
down to 3.59 A
˚, when the RInþRSn is 67% in In-Sn-Zn-O, as
shown in Fig. 3(b).
It can be easily expected that the closer inter-atomic dis-
tances between In and Sn atoms should make the overlaps
between the In-5sand Sn-5satomic orbitals more effective
in amorphous In-Sn-Zn-O. As the In and Sn density
increases, the increased overlaps of the In-5sand Sn-5sorbi-
tals extend the conduction band dispersion, as schematically
shown in Fig. 4. Accordingly, the CBM level is lowered
[Fig. 2(a)], and the electron effective mass is reduced [Fig.
2(b)], as the RInþRSn ratio increases.
C. Valence band tail
The understanding of the CBM changes with respect to
the RInþRSn ratio in In-Sn-Zn-O is rather straightforward
through the conducting channels made of the In and Sn 5s
orbitals. However, the change of the VBT extent depending
on the RInþRSn ratio [Fig. 2(a)] is not simply as expected,
because the VBT states are thought to be mostly O-related
and barely related to the cation compositions.
The orbital character of the VBT states is mostly O-2p
in the ionic-bonded material. The upper part of the valence
O-2pbands consists of the O-2panti-bonding states, while
the lower part of the bands consists of the O-2pbonding
states. In Fig. 5, we schematically draw the valence O-2p
bands comprising the various O-2porbital states. In crystal-
line oxides [Fig. 5(a)], the crystal field split levels of the O-
2patomic orbital states are well degenerated in inter-sites by
the structural order, and by the inter-site p-p orbital coupling,
the valence O-2pbands are delocalized. The lower part
of the bands is the pprand pppmixed bonding states, and
the upper part is the pprand pppmixed anti-bonding
states. We find that the top-most valence band state of the
crystalline InGaZnO
4
is mostly the delocalized pprand
pppmixed anti-bonding state [see Fig. 6(a)].
One of the main differences in valence O-2pbands
between the amorphous and crystalline oxides is the exis-
tence of the VBT states, which is the highest among the va-
lence O-2pstates [see Fig. 5(b)]. Among the inter-site p-p
orbital coupling, the ppris the highest in energy level, and
then the VBT states in amorphous oxides should have at least
partly the character of the O-O pprorbitals. In disordered
amorphous structure, the crystal field splitting of the atomic
O-2porbitals is complex, and the high energy O-2porbitals
will contribute to the VBT states more strongly [indicated by
the high energy porbital levels in Fig. 5(b)]. We plot the cal-
culated charge densities of the top-most VBT states for the
considered seven amorphous In-Sn-Zn-O compositions
[Figs. 6(b)–6(h)] and the amorphous InGaZnO
4
[Fig. 6(i)].
For the amorphous oxides, all show the O-O pprorbitals in
the VBT states clearly. For amorphous InGaZnO
4
, the ppr
character of the VBT state is also seen in other works.
55,56
We find a slight overlap of the Zn-3dand O-2pbands in
DOS of the amorphous oxides [see Fig. 1(a)]. By the p-d
FIG. 5. Schematic diagram of the va-
lence O-2pbands in (a) crystalline and
(b) amorphous oxides. The lower part of
the bands is pprand pppbonding states,
while the upper part is pprand ppp
anti-bonding. In amorphous oxides, the
highest level is the ppranti-bonding
state resulting from the high energy p
orbitals in the disordered structure.
FIG. 4. Schematic diagram showing the In and Sn 5sorbital overlaps and
the conduction band dispersions, when (a) the cations are isolated, (b) the
inter-distances between the In and Sn atoms are closer, and (c) the inter-
distances are furthermore closer.
183706-4 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
hybridization, a small contribution of the Zn-3dstates in
VBT can be induced, which can be seen in Fig. 6(e), as pre-
viously reported.
55
However, the Zn-3dcontribution to the
VBT states is found to be weaker than the O-O ppr, as seen
in the charge densities shown in Fig. 6.
Then, why is the O inter-site ppr-related VBT
level extent affected by the RInþRSn ratio in amorphous
In-Sn-Zn-O? We first investigate the O-O inter-distances in
the amorphous In-Sn-Zn-O, because, when the O-O inter-
distance is closer, the O-O inter-site coupling (ppr-ppr)
can be stronger. Figure 7shows the calculated radial pair-
distribution functions (rPDF) of O-O for the amorphous In-
Sn-Zn-O and InGaZnO
4
. The nearest neighbor O-O distances
are found in the range of 2.5-3.5 A
˚, and we could not find a
significant reduction of the O-O distances for the amorphous
In-Sn-Zn-O having a high RInþRSn ratio. All the calculated
amorphous oxides have a similar O-O distance distributions,
as shown in Fig. 7. Therefore, the VBT extent increase
with the RInþRSn ratio cannot be attributed to the O-O inter-
distance shortening.
The O-2porbital directions are various in amorphous
oxides by the orientational disorder of the amorphous struc-
ture, which is, in fact, the origin of the porbital localization
and the VBT states. Among the various O-2porbitals, the
O-2pdirected along the M-O bond is electrostatically stable,
since the Mcation atoms attract electrons, and their elec-
tronic energy levels are relatively deep in the valence O-2p
bands [low energy porbitals, indicated in Fig. 5(b)]. On the
other hand, the O-2porbital directed off the M-O bonds is
high in electrostatic energy, and forms a higher energy level
in the valence O-2pbands [high energy porbitals, indicated
in Fig. 5(b)]. By coupling of the off-bond O-2porbital with a
nearby O-2porbital, the probability of forming the VBT
states becomes high. Therefore, the inter-site coupling
between the off-bond O-2porbitals can be important for the
VBT characterization.
In the case of the inter-site coupling between two off-
bond O-2porbitals, the O-M-O bond angles can be important
in amorphous oxides. As the O-M-O bond angle is smaller,
the inter-site pprcoupling of the off-bond O-2porbitals can
be stronger (Fig. 8), and especially when the O-M-O bond
angle is acute smaller than 90, the pprcoupling can be sig-
nificantly enhanced, as shown in Fig. 8(c). It is worthwhile
to mention the preferred coordination numbers of In, Sn, and
Zn cations in oxides. The In and Sn atoms preferentially
form octahedral configuration with bonding to nearby 6 O
atoms (6-fold coordination) [Figs. 9(c) and 9(d)], while the
Zn atoms preferentially form tetrahedral configuration with 4
nearby O atoms (4-fold coordination) [Fig. 9(a)]. The Ga
atoms are known to have a preferentially 5-fold coordination
number [Fig. 9(b)]. Especially, for the octahedral configura-
tion with the nearby 6 O atoms, the O-M-O angles are 90in
a crystal and can be acute more probably in amorphous
structure.
FIG. 6. Calculated charge densities of (a) the top-most valence state of the
crystalline InGaZnO
4
, [(b)-(i)] the top-most VBT state of the amorphous In-
Sn-Zn-O [(b)–(h)], and amorphous InGaZnO
4
(i). The dashed lines connect
two O atoms that have the pprorbital in the VBT state. The solid line in
(a) indicates the ppporbital ordering.
FIG. 7. Calculated rPDF between O atoms for amorphous In-Sn-Zn-O and
InGaZnO
4
. In the parenthesis, the RInþRSn ratios are indicated. With
decreasing the RInþRSn ratio from the ISZO#1 (67%), the rPDF is shifted by
þ1 ea.
FIG. 8. Schematic diagrams showing the inter-site coupling between two
high-energy (off-bond) O-2porbitals, which depends on the O-M-O angles:
(a) 180, (b) 120, and (c) 90.
183706-5 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
We investigated the O-M-O bond angle distributions for
the amorphous In-Sn-Zn-O and InGaZnO
4
, as shown in Figs.
10(a)–10(h). For In and Sn atoms, the number of O-M-O
angles [M¼(In, Sn)] are abundant in the range of 70-90,
and there is a small peak near 140-180, which indicates the
octahedral configuration of InO
6
and SnO
6
[see Figs. 9(c) and
9(d)]. On the other hand, near Zn atoms, the largest amount of
O-M-O angles (M¼Zn) are found in the range of 80–120.
The O-Zn-O bond angles are larger than the small O-(In, Sn)-
O bond angles. For amorphous InGaZnO
4
, the number of
O-Ga-O bond angles is the most in the range of 80-110[Fig.
10(h)]. Although the O-O distances in the amorphous oxides
do not show a significant difference between the different cat-
ion compositions (as shown in Fig. 7), the O-M-O bond angles
show a clear difference depending on the metal cation species,
M. That is because the In-O (2.1 A
˚) and Sn-O (2.1 A
˚)
bond lengths are larger than the Zn-O (1.9 A
˚) bond length,
and thus even with the smaller O-M-O bond angles, the O-O
distances keep a value in the range of 2.5-3.5 A
˚(see Fig. 7),
irrespective of the nearby Mcations in the amorphous oxides.
The densities of acute O-M-O bond angles are plotted as
a function of the RInþRSn ratio in Fig. 10(i). They are found
to increase as the RInþRSn ratio increases in the amorphous
oxides. When In and Sn contents are increased in amorphous
In-Sn-Zn-O, a large number of acute O-M-O angles will
appear, and they make the O-O ppr-pprcoupling between
two off-bond O-2porbitals stronger. That is, although the
VBT states are mostly characterized by the O-2porbitals,
they are affected by the metal cation compositions. The O
atoms near the In and Sn cations are more strongly interact-
ing and have a higher probability of forming the pprhybri-
dized orbital states in the VBT in amorphous In-Sn-Zn-O.
As the RInþRSn ratio increases, the RZn ratio decreases
correspondingly. In our results, as the RZn ratio decreases,
the VBT level extent increases. If we considered the partial
contribution of the Zn-3dstates in the VBT states, the trend
should be opposite. However, it is not the case, because as
the RZn ratio increases, only the Zn-3dpeak intensity in the
total DOS is heightened [as shown in Fig. 1(a)], whereas the
Zn-3d band dispersion is little changed. For the p-d hybrid-
ization strength, the band overlap between the Zn-3dand O-
2pstates is more influential than the density of the Zn-3d
states. Thus, the VBT level extent changes with the cation
compositions are not considered to be related to the Zn-3d
states.
IV. DISCUSSION
Recently, it has been proposed that the NBIS and IS
instabilities of AOS-based TFTs are partially attributed to
the VBT holes generated by the negative gate bias and/or
illumination stress.
32
We showed that the VBT states in
AOS are characterized by the pprelectronic orbitals, in
Sec. III C. The empty ppr-like VBT states by the excited
holes can generate a driving force to form an O-O bond by
the pprbonding interaction, resulting in the O2
2peroxide
defect formation. The O2
2peroxide defect is previously
found to be a meta-stable donor defect.
32
Therefore, the
increased VBT extents in the In- and Sn-rich AOS can wor-
sen the hole-induced (NBIS and IS) instabilities of the
AOS, even though the electron mobility can be enhanced by
the increased densities of In and Sn. On the other hand, the
VO’s in AOS have been also proposed as an origin to give
the NBIS instability in AOS-based TFTs.
19–22,24,30,31
Since
the In-O and Sn-O bonds are weaker than the Ga-O and
Zn-O bonds, the VOformation can be enhanced with the
increased densities of In and Sn, which can worsen again but
as a different mechanism the NBIS and IS instabilities of the
AOS. Therefore, we can expect that the electron mobility
and stability of AOS can be in a trade-off relation, when they
are controlled by the In and Sn compositions in AOS.
V. CONCLUSION
Based on DFT calculations, the effects of cation compo-
sitions on the electronic structures of amorphous In-Sn-Zn-O
semiconductors were investigated. We found that the CBM
energy level and the electron effective mass (m
e) decrease
with the sum of In and Sn composition ratios (RInþRSn ).
The electron carrier mobility can be increased with increas-
ing the In and Sn contents in amorphous In-Sn-Zn-O by
reducing the effective mass. However, as the sum of In and
FIG. 9. Local atomic configurations near a (a) Zn, (b) Ga, (c) Sn, and (d) In
in amorphous ISZO#1 and amorphous InGaZnO
4
.
FIG. 10. (a)–(h) O-M-O bond angle distributions for the amorphous In-Sn-
Zn-O and amorphous InGaZnO
4
. The Mis for In (circle), Sn (square), Zn
(triangle), and Ga (diamond). (i) Volumetric densities of the acute O-M-O
bond angles (90) in the amorphous oxides, as a function of the RIn þRSn
ratio.
183706-6 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
Sn ratios increases, the VBT is found to be extended, which
can worsen the stability of the AOS-based TFTs.
32
The VBT
extent is found to be due to the increased density of In and
Sn cations, near which the O-2pinter-site pprcoupling is
stronger than that near the Zn atoms. The acute O-(In,Sn)-O
angle is suggested to be an important local structural parame-
ter in amorphous oxide semiconductors, which renders the
O-O pprcoupling more effective. The electron mobility
and the stability are thus in a trade-off relation, when they
are controlled by the In and Sn contents in amorphous In-Sn-
Zn-O.
ACKNOWLEDGMENTS
This research was supported by Nano R&D program
through the National Research Foundation (NRF) of Korea
funded by the Ministry of Education, Science, and Technology
(MEST) (No. 2012-0006199).
1
K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono,
Nature 432, 488 (2004).
2
B. S. Yang, M. S. Huh, S. Oh, U. S. Lee, Y. J. Kim, M. S. Oh, J. K. Jeong,
C. S. Hwang, and H. J. Kim, Appl. Phys. Lett. 98, 122110 (2011).
3
J.-S. Park, K. Kim, Y.-G. Park, Y.-G. Mo, H. D. Kim, and J. K. Jeong,
Adv. Mater. 21, 329 (2009).
4
J.-Y. Kwon, J. S. Jung, K. S. Son, K.-H. Lee, J. S. Park, T. S. Kim, J.-S.
Park, R. Choi, J. K. Jeong, B. Koo, and S. Y. Lee, Appl. Phys. Lett. 97,
183503 (2010).
5
S. Yang, D.-H. Cho, M. K. Ryu, S.-H. K. Park, C.-S. Hwang, J. Jang, and
J. K. Jeong, Appl. Phys. Lett. 96, 213511 (2010).
6
I. J. Kang, C. H. Park, E. Chong, and S. Y. Lee, Curr. Appl. Phys. 12, S12
(2012).
7
J. Robertson, J. Non-Cryst. Solids 354, 2791 (2008).
8
K. Nomura, A. Takagi, T. Kamiya, H. Ohta, M. Hirano, and H. Hosono,
Jpn. J. Appl. Phys., Part 1 45, 4303 (2006).
9
Y. Sun and J. A. Rogers, Adv. Mater. 19, 1897 (2007).
10
H. Yabuta, M. Sano, K. Abe, T. Aiba, T. Den, H. Kumomi, K. Nomura, T.
Kamiya, and H. Hosono, Appl. Phys. Lett. 89, 112123 (2006).
11
P. G€
orrn, H. H€
olzer, W. Kowalsky, J. Wang, T. Weimann, P. Hinze, and
S. Kipp, Appl. Phys. Lett. 90, 063502 (2007).
12
A. Suresh and J. F. Muth, Appl. Phys. Lett. 92, 033502 (2008).
13
J.-M. Lee, I.-T. Cho, J.-H. Lee, and H.-I. Kwon, Appl. Phys. Lett. 93,
093504 (2008).
14
M. E. Lopes, H. L. Gomes, M. C. R. Medeiros, P. Barquinha, L. Pereira,
E. Fortunato, R. Martins, and I. Ferreira, Appl. Phys. Lett. 95, 063502
(2009).
15
J. K. Jeong, H. W. Yang, J. H. Jeong, Y.-G. Mo, and H. D. Kim, Appl.
Phys. Lett. 93, 123508 (2008).
16
K. Nomura, T. Kamiya, M. Hirano, and H. Hosono, Appl. Phys. Lett. 95,
013502 (2009).
17
W.-T. Chen, H.-W. Hsueh, H.-W. Zan, and C.-C. Tasi, Electrochem.
Solid-State Lett. 14, H297 (2011).
18
J.-H. Shin, J.-S. Lee, C.-S. Hwang, S.-H. K. Park, W. S. Cheong, M. Ryu,
C.-W. Byun, J.-I. Lee, and H. Y. Chu, ETRI J. 31, 62 (2009).
19
H. Oh, S.-M. Yoon, M. K. Ryu, C.-S. Hwang, S. Yang, and S.-H. K. Park,
Appl. Phys. Lett. 97, 183502 (2010).
20
M. D. H. Chowdhury, P. Migliorato, and J. Jang, Appl. Phys. Lett. 97,
173506 (2010).
21
K. H. Ji, J.-I. Kim, H. Y. Jung, S. Y. Park, R. Choi, U. K. Kim, C. S.
Hwang, D. Lee, H. Hwang, and J. K. Jeong, Appl. Phys. Lett. 98, 103509
(2011).
22
H. Oh, S.-M. Yoon, M. K. Ryu, C.-S. Hwang, S. Yang, and S.-H. K. Park,
Appl. Phys. Lett. 98, 033504 (2011).
23
H. Oh, S.-H. K. Park, C.-S. Hwang S. Yang, and M. K. Ryu, Appl. Phys.
Lett. 99, 022105 (2011).
24
K. H. Ji, J.-I. Kim, H. Y. Jung, S. Y. Park, R. Choi, Y. G. Mo, and J. K.
Jeong, Microelectron. Eng. 88, 1412 (2011).
25
P. Barquinha, A. Pimentel, A. Marques, L. Pereira, R. Martins, and E.
Fortunato, J. Non-Cryst. Solids 352, 1756 (2006).
26
P. G€
orrn, M. Lehnhardt, T. Riedl, and W. Kowalsky, Appl. Phys. Lett. 91,
193504 (2007).
27
C.-S. Chuang, T.-C. Fung, B. G. Mullins, K. Nomura, T. Kamiya, H.-P.
D. Shieh, H. Hosono, and J. Kanicki, Proc. Soc. Inf. Display 39,1215
(2008).
28
D. P. Gosain and T. Tanaka, Jpn. J. Appl. Phys., Part 1 48,03B018
(2009).
29
D. H. Lee, K. Kawamura, K. Nomura, T. Kamiya, and H. Hosono,
Electrochem. Solid-State Lett. 13, H324 (2010).
30
B. Ryu, H.-K. Noh, E.-A. Choi, and K. J. Chang, Appl. Phys. Lett. 97,
022108 (2010).
31
H.-K. Noh, K. J. Chang, B. Ryu, and W.-J. Lee, Phys. Rev. B 84, 115205
(2011).
32
H.-H. Nahm, Y.-S. Kim, and D. H. Kim, Phys. Status Solidi B 249, 1277
(2012).
33
M. Kim, J. H. Jeong, H. J. Lee, T. K. Ahn, H. S. Shin, J.-S. Park, J. K.
Jeong, Y.-G. Mo, and H. D. Kim, Appl. Phys. Lett. 90, 212114 (2007).
34
J. S. Park, T. S. Kim, K. S. Son, K.-H. Lee, W.-J. Maeng, H.-S. Kim, E. S.
Kim, K.-B. Park, J.-B. Seon, W. Choi, M. K. Ryu, and S. Y. Lee, Appl.
Phys. Lett. 96, 262109 (2010).
35
J. S. Park, T. S. Kim, K. S. Son, W.-J. Maeng, H.-S. Kim, M. Ryu, and
S. Y. Lee, Appl. Phys. Lett. 98, 012107 (2011).
36
T. Riedl, P. G€
orrn, H. H€
olzer, and W. Kowalsky, Phys. Status Solidi
(RRL) 1, 175 (2007).
37
G. Kresse and J. Furthm€
uller, Phys. Rev. B 54, 11169 (1996).
38
G. Kresse and D. Joubert, Phys. Rev. B 59, 1758 (1999).
39
P. E. Bl€
ochl, Phys. Rev. B 50, 17953 (1994).
40
D. M. Ceperley and B. J. Alder, Phys. Rev. Lett. 45, 566 (1980).
41
V. I. Anisimov, I. V. Solovyev, M. A. Korotin, M. T. Czy_
zyk, and G. A.
Sawatzky, Phys. Rev. B 48, 16929 (1993).
42
W.-J. Lee and Y.-S. Kim, J. Korean Phys. Soc. 60, 781 (2012).
43
C. L. Dong, C. Persson, L. Vayssieres, A. Augustsson, T. Schmitt, M.
Mattesini, R. Ahuja, C. L. Chang, and J.-H. Guo, Phys. Rev. B 70, 195325
(2004).
44
A. Janotti, D. Segev, and C. G. Van de Walle, Phys. Rev. B 74, 045202
(2006).
45
J. E. Jaffe and A. C. Hess, Phys. Rev. B 48, 7903 (1993).
46
F. Oba, A. Togo, I. Tanaka, J. Paier, and G. Kresse, Phys. Rev. B 77,
245202 (2008).
47
S. J. Clark, J. Robertson, S. Lany, and A. Zunger, Phys. Rev. B 81, 115311
(2010).
48
S. Massidda, R. Resta, M. Posternak, and A. Baldereschi, Phys. Rev. B 52,
R16977 (1995).
49
M. Shishkin and G. Kresse, Phys. Rev. B 75, 235102 (2007).
50
M. Shishkin, M. Marsman, and G. Kresse, Phys. Rev. Lett. 99, 246403
(2007).
51
F. Fuchs, J. Furthm€
uller, F. Bechstedt, M. Shishkin, and G. Kresse, Phys.
Rev. B 76, 115109 (2007).
52
B.-C. Shih, Y. Xue, P. Zhang, M. L. Cohen, and S. G. Louie, Phys. Rev.
Lett. 105, 146401 (2010).
53
M. Stankovski, G. Antonius, D. Waroquiers, A. Miglio, H. Dixit, K.
Sankaran, M. Giantomassi, X. Gonze, M. C
ote, and G.-M. Rignanese,
Phys. Rev. B 84, 241201(R) (2011).
54
K. Nomura, T. Kamiya, H. Ohta, T. Uruga, M. Hirano, and H. Hosono,
Phys. Rev. B 75, 035212 (2007).
55
A. Walsh, J. L. F. Da Silva, and S.-H. Wei, Chem. Mater. 21,5119
(2009).
56
T. Kamiya, K. Nomura, and H. Hosono, J. Disp. Technol. 5, 273 (2009).
183706-7 Noh et al. J. Appl. Phys. 113, 183706 (2013)
Downloaded 10 May 2013 to 222.111.165.128. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jap.aip.org/about/rights_and_permissions
