Impact of bias stability for crystalline InZnO thin-film transistorsHojoong Kim, Daehwan Choi, Solah Park, Kyung Park, Hyun-Woo Park, Kwun-Bum Chung, and Jang-YeonKwon

Citation: Appl. Phys. Lett. 110, 232104 (2017); doi: 10.1063/1.4985295View online: http://dx.doi.org/10.1063/1.4985295View Table of Contents: http://aip.scitation.org/toc/apl/110/23Published by the American Institute of Physics

Articles you may be interested in Hydrogen anion and subgap states in amorphous In–Ga–Zn–O thin films for TFT applicationsApplied Physics Letters 110, 232105 (2017); 10.1063/1.4985627

An oxidation-last annealing for enhancing the reliability of indium-gallium-zinc oxide thin-film transistorsApplied Physics Letters 110, 142102 (2017); 10.1063/1.4979649

H and Au diffusion in high mobility a-InGaZnO thin-film transistors via low temperature KrF excimer laserannealingApplied Physics Letters 110, 133503 (2017); 10.1063/1.4979319

BaTiO3/SrTiO3 heterostructures for ferroelectric field effect transistorsApplied Physics Letters 110, 232902 (2017); 10.1063/1.4985014

Solution-processed gadolinium doped indium-oxide thin-film transistors with oxide passivationApplied Physics Letters 110, 122102 (2017); 10.1063/1.4978932

The effect of device electrode geometry on performance after hot-carrier stress in amorphous In-Ga-Zn-O thinfilm transistors with different via-contact structuresApplied Physics Letters 110, 202103 (2017); 10.1063/1.4983713

Impact of bias stability for crystalline InZnO thin-film transistors

Hojoong Kim,1,2 Daehwan Choi,1,2 Solah Park,1,2 Kyung Park,2 Hyun-Woo Park,3

Kwun-Bum Chung,3 and Jang-Yeon Kwon1,2,a)

1School of Integrated Technology, Yonsei University, Incheon 406-840, South Korea2Yonsei Institute of Convergence Technology, Yonsei University, Incheon 406-840, South Korea3Division of Physics and Semiconductor Science, Dongguk University, Seoul 100-715, South Korea

(Received 8 March 2017; accepted 29 May 2017; published online 8 June 2017)

Crystallized InZnO thin-film transistors (IZO TFTs) are investigated to identify a potential for the

maintenance of high electrical performances with a consistent stability. The transition from an

amorphous to a crystallization structure appeared at an annealing temperature around 800 �C, and

it was observed using transmission electron microscopy and time-of-flight secondary ion mass

spectrometry analysis. The field-effect mobility of the crystallized IZO TFTs was boosted up to

53.58 cm2/V s compared with the 11.79 cm2/V s of the amorphous devices, and the bias stability

under the negative stress was greatly enhanced even under illumination. The defect states related to

the oxygen vacancy near the conduction band edge decreased after the crystallization, which is a

form of electrical structure evidence for the reliability impact regarding the crystallized IZO TFTs.

Published by AIP Publishing. [http://dx.doi.org/10.1063/1.4985295]

The oxide semiconductor thin-film transistors (TFTs)

form a huge branch in the field of active matrix displays and

diodes due to its attractive properties of a reasonable electri-

cal mobility, transparency for visible light, and low fabrica-

tion costs.1–3 However, an instability issue regarding the

voltage-bias stress and which accelerated under illumination

has persisted in terms of the general use of the oxide material

in technology. The general origins of the bias instability are

known as the defect creation, charge trapping, vacancy ioni-

zation, and environmental conditions.4,5 The bias stress gen-

erates or traps carriers inside the channel or at the interface

between the channel and the gate insulator, resulting in the

threshold voltage (VTH) shift. The ionization of the oxygen

vacancy (VO) under illumination creates electron carriers,

thereby accelerating the VTH shift. Last, the moisture and

oxygen in air can also affect the carrier level so that it makes

passivation treatment or structural modification necessary. A

high electrical mobility can be expected for the In contained

ZnO semiconductor due to the high carrier density in the In,

while simultaneously generated defects also degrade the

device stability.6,7 The doping of a carrier suppressor such as

Zr, Hf, and Ga into the semiconductor can stabilize the defect

level for the development of device reliability, but most car-

rier suppressors also reduce the electrical mobility.8,9

A previous study reported the c-axis aligned crystal

InGaZnO (CAAC-IGZO) TFTs as an effective source of the

negative bias illumination.10,11 The CAAC-IGZO decreased

the defect level at the deep bandgap states more than the

amorphous IGZO, even maintaining its electrical characteris-

tics. In spite of the IGZO crystallization without the axis ori-

entation, a high temperature annealed crystallization can also

reduce the defect level and enhance the device stability.12 It

is understood that the defect level in the sub-conduction

band is effectively diminished by the crystallization; there-

fore, the simple annealing process for crystallizing the oxide

semiconductor can be an alternative to the preservation of the

reliability without a deterioration of its initial characteristics.

Even though the IGZO is a credible material for many appli-

cations, the limitation on electrical mobility at around 10 cm2/

V s shows that a higher conductive oxide is still required.

Consequently, crystallizing InZnO (IZO) TFTs have a poten-

tial in terms of the maintenance of a high electrical mobility

with reliability among the other oxide materials.

In this study, amorphous IZO TFTs are transformed into

the crystal phased devices by high temperature annealing.

The phase variation of the IZO thin film is identified as a

function of the annealing temperature, and the electrical

characteristics and stability of both devices are compared

through an electrical structure based analysis.

The amorphous IZO TFTs were fabricated as the bottom-

gate structure on a highly doped P-type Si wafer through the

use of a 100 nm thermally grown SiO2 as a gate insulator. The

IZO thin film (In:Zn¼ 7:3) was deposited as an active chan-

nel layer using radio frequency (RF) sputtering with a sputter

power of 100 W and an oxygen partial pressure (O2/O2 þAr)

of approximately 9.1% under 5 mTorr of working pressure. A

lift-off technique was exploited for patterning the channel

layer with a width/length area of 200 lm/50 lm through

photolithography. An alpha step measurement identified the

channel thickness as 50 nm. To crystallize the IZO thin film,

a post-channel annealing was progressed for 1 h in a tube

furnace with a varying temperature of 600 �C, 800 �C, and

1000 �C. Mo electrode of 150 nm was deposited as a source

and drain contact using sputtering, and last, all devices were

heated at 200 �C for 1 h to reduce the contact resistance.

To identify the crystal structure of the IZO thin films

depending on the annealing temperature, X-ray diffraction

(XRD) patterns were measured as a function of post-channel

annealing temperature (TP), which is shown in Fig. S1 (sup-

plementary material). Up to 600 �C, XRD peak of the IZO

does not show any particular peak intensity that is related to

the IZO. While at a further TP increment at 800 �C, the dif-

fraction peaks of (211), (222), and (400) as the In2O3 crystal

a)Author to whom correspondence should be addressed: jangyeon@

yonsei.ac.kr

0003-6951/2017/110(23)/232104/5/$30.00 Published by AIP Publishing.110, 232104-1

APPLIED PHYSICS LETTERS 110, 232104 (2017)

peaks (JCPDS 06-0416) shows the film was changed from

amorphous to crystalline structure. The clear peak at

2h¼ 31� is attributed to the (222) plane of the bixbyite In2O3

structure,13 and the phase change oriented from the ZnO is

not detectable on the diffraction pattern. The peak intensity

at TP¼ 1000 �C is not significantly increased compared to

that at the TP¼ 800 �C, indicating the crystallinity of the

IZO did not grow any further with raising temperature.

Figs. 1(a)–1(d) display the cross-sectional transmission

electron microscopy (TEM) images of the IZO thin films

with varying TP. The images at the downside of each bright-

field image show the high magnification focus of the IZO

film, and the insets indicate the selected area electron diffrac-

tion (SAED) patterns of the IZO. In the bright-field images

of as-deposition (as-dep) and TP¼ 600 �C annealed samples,

the morphology variation is not appeared as increasing TP.

Alternatively, a clear crystalline shape is identifiable in the

enlarged images of the TP¼ 800 �C and 1000 �C. The SAED

patterns also show the definite evidence of the IZO phase

transformation from the amorphous to the polycrystalline

state after the TP¼ 800 �C. In addition, the creation of a sur-

face undulation on the IZO and a shallow interlayer between

the IZO and the SiO2 are discovered in the images over

800 �C. The surface undulation is considered as a formation

of the crystalline grain,14 and the interlayer creation could be

originated from a mixture of the two layers caused by the

high annealing temperature.

Following the crystallization confirmation, the IZO film

composition variation was investigated owing to the mor-

phology degradation after the 800 �C as shown earlier. The

graphs in Figs. 2(a)–2(d) are the depth profiles from the IZO

to the SiO2 measured by the time-of-flight secondary ion

mass spectrometry (TOF-SIMS). The interface between the

IZO and SiO2 is relatively distinguishable at the as-dep (a)

and TP¼ 600 �C states (b), while the Zn ion in the IZO film

is diffused into the SiO2 layer after the TP¼ 800 �C (c). The

diffusion caused a reduction of Zn composition ratio in the

IZO, and created a layer with higher amount of In. The com-

positional variation was advanced as the Zn diffusion deeply

penetrated into the SiO2 layer at the TP¼ 1000 �C (d).

After the identification of the structural changes in the

IZO thin film, the electrical characteristics of the IZO TFTs

were investigated. Figure 3(a) shows the transfer curves of

the IZO TFTs as a function of the TP with a 10 V of drain-

source voltage (VDS). The transfer characteristic of the as-

dep IZO device demonstrated a typical semiconducting

curve shape with the field-effect mobility (lFE) of 11.79 cm2/

V s, the VTH of 18.68 V, and the subthreshold swing (SS) of

0.41 V/dec. The lFE defined by transconductance (gm) was

calculated by the following equation:

gm ¼W

LCox lFE VDS; (1)

where the gm is the gradient of the transfer curve, W/L is the

ratio of the channel width and length, and Cox is the capaci-

tance per unit area of the gate insulator. The lFE at the maxi-

mum value of the gm was represented as the electrical

mobility. The transfer curve of the as-dep IZO TFTs was

changed to a constant line shape after 600 �C due to the

higher electrical conductivity of the channel by the electron

carrier generation.15 As the TP approached to 800 �C, the

conductive channel was retransformed to the semiconductor,

predicted as a consequence of the channel crystallization.

The variation of the electrical conductivity could be the

result of the diminution of the VO in the amorphous channel

by the IZO crystalline arrangement. The device parameters

of the lFE, VTH, and SS at 800 �C were calculated to

53.58 cm2/V s, 8.31 V, and 0.58 V/dec, respectively. The sig-

nificant lFE boost could be owing to the variation of the

channel composition ratio, which is caused by the high In

amount in the semiconductor.6,16 Lastly at the TP¼ 1000 �C,

the transfer curve was negatively shifted by the creation of

the electron carrier, and the values of the lFE, VTH, and SS

were 46.34 cm2/V s, �1.86 V, and 0.96 V/dec, respectively.

The degradation of the lFE and SS could be originated from

the severe deformation of the channel and interfacial layers

by the extreme annealing temperature. The device parame-

ters at each TP are summarized in Table I.

The electrical reliability as the variation of the film phase

structure was investigated. The as-dep and TP¼ 800 �C IZO

TFTs were selected as the amorphous and crystalline devices

for evaluating under the bias stress. The test was implemented

inside a dark chamber under the vacuum condition to mini-

mize the environmental effects from the air. The stress test

was carried out until 7200 s and the measurement interval

was set to 600 s. A gate-source voltage (VGS) of �40 V and

FIG. 1. Cross-sectional TEM images of

the IZO for (a) as-dep, (b) 600 �C, (c)

800 �C, and (d) 1000 �C. The high mag-

nification focus images, which are

located at the downside, allow for the

identification of the phase transforma-

tion. The insets are the diffraction pat-

terns of the IZO for each post-annealing

temperature.

232104-2 Kim et al. Appl. Phys. Lett. 110, 232104 (2017)

the VDS of 10 V were applied for the negative bias stress

(NBS), and VGS¼ 40 V and VDS¼ 10 V for the positive bias

stress (PBS). The transfer curve of the as-dep IZO TFTs under

the NBS was gradually shifted to the negative side. Whereas

which of the 800 �C annealed device was slightly moved to

the positive side within relatively early test time. The parallel

shift of the transfer curve under the NBS was attributed to the

donor-like trap states by the gate electric field.17,18 The trap

states at the IZO/SiO2 interface were shifted by the NBS upon

the Fermi level so that the traps donate charges to the conduc-

tion band. On the other hand, the trap sites in IZO TFTs were

eliminated by the crystallization, which reduced the negative

VTH shift under NBS. The sudden positive shift of VTH

revealed that new charge trapping sites could be created in the

crystallized IZO TFTs and capture electrons during the mea-

surement. The electron trapping at the interface between the

channel and the gate insulator or at the grain boundary inside

the channel could be attributed to the positive VTH shift even

under the NBS. A similar phenomenon was reported on the

poly-Si TFTs.19 The electrons trapped and fixed into the

defect sites can also erase the defect density, and as a conse-

quence, the SS characteristic was significantly improved in

the first 1200 s of the NBS. The gate voltage movement at the

1 nA drain current (DV1nA) in the as-dep device was approxi-

mately �16.0 V. In the case of the PBS, the evolution at the

as-dep IZO TFTs shows a gradual movement to the positive

side with a degradation of the SS value, while they remained

stationary for most of the stress time at the TP¼ 800 �C. In

the next phase, the NBS and PBS under illumination (NBIS

and PBIS) were performed with a white LED light of 290 lux.

The negative shifting of the as-dep curves was drastically

increased under the NBIS compared to the NBS by the

FIG. 2. TOF-SIMS depth profiles of

the IZO thin films at each TP: (a)

as-dep, (b) 600 �C, (c) 800 �C, and (d)

1000 �C.

FIG. 3. Electrical characteristics of the IZO TFT devices. (a) Transfer

curves of the TFTs as a function of the post-annealing temperature at

VDS¼ 10 V. (b) The evolution of the transfer curves under the bias-stress

tests summarized as the gate voltage movement at the 1 nA drain current.

TABLE I. Electrical characteristics of IZO TFTs as a function of post-

annealing temperature.

Post-annealing temperature (TP)

as-dep 800 �C 1000 �C

lFE (cm2/V s) 11.79 53.58 46.34

VTH (V) 18.68 8.31 �1.86

S.S (V/dec) 0.41 0.58 0.96

232104-3 Kim et al. Appl. Phys. Lett. 110, 232104 (2017)

assistance of the illumination. After crystallization, however,

the DV1nA was decreased from �34.6 V to �5.4 V, and this

reduction ratio was much larger than that under the NBS test

(even though the positive shift at the TP¼ 800 �C was

included). It means that the VO ionization accelerated by the

illumination was also decreased. Therefore, it leads to the pos-

sibility that the defect states which corresponding to the dimi-

nution of the VO was decreased after the crystallization of the

IZO TFTs. It is also noticeable that the TP¼ 800 �C under

NBIS showed SS reduction. Decreasing SS could explain the

elimination of trap site on the interfaces or the channel grain

boundary during the measurement as similar to the NBS test.

The movement of the transfer curves under the PBIS was rela-

tively comparable to that under the PBS for both devices. The

DV1nA under all of the stress tests is summarized in Fig. 3(b),

and Fig. S3 (supplementary material) shows the evolution of

the transfer curves under the bias stress tests.

In addition, a similar report has been published regard-

ing the stability of the IZO crystallization. Oh et al. fabri-

cated IZO TFTs by varying the composition ratio of In and

Zn.20 An increasing of the In content in the polycrystalline

ZnO transformed the IZO film into the amorphous phase.

As the phase variation, the device stability was simulta-

neously degraded whereas the electrical characteristics were

improved. They mentioned that the In weakened the crystal

structure of the ZnO and created the VO, which showed a

lower stability in the IZO TFTs with the high In amount. In

our study, the bias stability of the crystalline IZO TFTs was

enhanced with the high In concentration, which is regarded

for the possibility of high performance devices with an

excellent stability in terms of the crystallized oxide TFTs.

For a further investigation in a perspective of the elec-

tronic structures, a systematic spectroscopic analysis was

performed in the beamline 2A of Pohang Acceleration

Lightsource. Figure 4(a) shows the normalized O K1 edge X-

ray absorption spectroscopy (XAS) spectra of the as-dep and

TP¼ 800 �C IZO films as the amorphous and crystalline phased

devices, respectively. The relative peak intensities in the spec-

tra from 530 eV to 550 eV reflect the hybridized orbitals of the

oxygen p-projected density of states for the In 5sp, Zn 4sp, and

O 2p.21 The post-annealing process increased the relative inten-

sities of the K1 edge spectra, which can be attributed to the

enlarged conduction band area and the dominant feature of

orbital hybridization of In2O3.22 With the increased number of

the unoccupied hybridized orbital states, the carrier transport

on the conduction band can be enhanced. The defect states

below the conduction band edge around 531 eV were detected

from the second-derivative spectra. Figure 4(b) depicts the

enlargement of the band edge states below the conduction

band. The amount of the defect state for the TP¼ 800 �C was

clearly decreased compared to that of the as-dep IZO, which

results from the reduction of VO below the conduction

band.23–27 Regarding the decreased amount of the defect states,

the NBS could not create the donor-like defect states in the

IZO after the crystallization. In addition, the reduction of VO

density suppressed the VO ionization under the illumination

stress. Therefore, the crystalline IZO TFTs showed the highly

reliable and stable performance under NBS and NBIS.

In conclusion, crystallized IZO TFTs were revealed,

and their electrical characteristics and bias stability were

investigated in a comparison with the amorphous IZO TFTs.

The electrical mobility was greatly improved in the crystal-

lized devices due to the compositional variation and the

phase transformation, which were mainly oriented by the

In2O3, respectively. The reduction of the vacancy defects

that are located near the conduction band edge supports the

origin of the stability in the crystalline phase. Crystallization

is potentially an effective reliability solution, even in the

highly defective semiconductors, for the increasing of the

electrical performances.

See supplementary material for the characterization of

the IZO films by XRD and XPS. Output curves of IZO TFTs

and the evolution of transfer curves under the stress tests are

also included.

This research was supported by the MSIP (Ministry of

Science, ICT and Future Planning), Korea, under the “ICT

Consilience Creative Program” (IITP-2017-2017-0-01015)

supervised by the IITP (Institute for Information and

communications Technology Promotion). This work was

also supported by Samsung Display Co., Ltd.

1E. Fortunato, P. Barquinha, and R. Martins, Adv. Mater. 24, 2945 (2012).2J. S. Park, W.-J. Maeng, H.-S. Kim, and J.-S. Park, Thin Solid Films 520,

1679 (2012).3J. Y. Kwon and J. K. Jeong, Semicond. Sci. Technol. 30, 024002 (2015).4T.-C. Fung, K. Abe, H. Kumomi, and J. Kanicki, J. Disp. Technol. 5, 452

(2009).

FIG. 4. (a) The normalized XAS O K1 edge spectra of the as-dep and

800 �C post-annealed IZO and the second derivative O K1 edge spectra are

respectively at the bottom of the graph. (b) The band edge states over a nar-

row energy region below the conduction.

232104-4 Kim et al. Appl. Phys. Lett. 110, 232104 (2017)

5J.-Y. Kwon, D.-J. Lee, and K.-B. Kim, Electron. Mater. Lett. 7, 1 (2011).6H. Hosono, J. Non-Cryst. Solids 352, 851 (2006).7R. Martins, P. Barquinha, I. Ferreira, L. Pereira, G. Goncalves, and E.

Fortunato, J. Appl. Phys. 101, 044505 (2007).8S. J. Kim, S. Yoon, and H. J. Kim, Jpn. J. Appl. Phys., Part 1 53, 02BA02

(2014).9S. Parthiban and J.-Y. Kwon, J. Mater. Res. 29, 1585 (2014).

10S. Yamazaki, J. Koyama, Y. Yamamoto, and K. Okamoto, SID Dig. 43,

183 (2012).11S. Yamazaki, ECS Trans. 64, 155 (2014).12K. Park, H.-W. Park, H. S. Shin, J. Bae, K.-S. Park, I. Kang, K.-B. Chung,

and J.-Y. Kwon, IEEE Trans. Electron Devices 62, 2900 (2015).13J. Jia, S.-i. Nakamura, and Y. Shigesato, Phys. Status Solidi A 213, 2291

(2016).14S. Yamazaki, H. Suzawa, K. Inoue, K. Kato, T. Hirohashi, K. Okazaki,

and N. Kimizuka, Jpn. J. Appl. Phys., Part 1 53, 04ED18 (2014).15P. Barquinha, G. Goncalves, L. Pereira, R. Martins, and E. Fortunato, Thin

Solid Films 515, 8450 (2007).16G. H. Kim, B. D. Ahn, H. S. Shin, W. H. Jeong, H. J. Kim, and H. J. Kim,

Appl. Phys. Lett. 94, 233501 (2009).17Y. Ueoka, Y. Ishikawa, J. P. Bermundo, H. Yamazaki, S. Urakawa, M. Fujii,

M. Horita, and Y. Uraoka, ECS J. Solid State Sci. Technol. 3, Q3001 (2014).

18K. Ghaffarzadeh, A. Nathan, J. Robertson, S. Kim, S. Jeon, C. Kim, U.-I.

Chung, and J.-H. Lee, Appl. Phys. Lett. 97, 113504 (2010).19C.-Y. Chen, J.-W. Lee, S.-D. Wang, M.-S. Shieh, P.-H. Lee, W.-C. Chen,

H.-Y. Lin, K.-L. Yeh, and T.-F. Lei, IEEE Trans. Electron Devices 53,

2993 (2006).20S. Oh, B. S. Yang, Y. J. Kim, M. S. Oh, M. Jang, H. Yang, J. K. Jeong,

C. S. Hwang, and H. J. Kim, Appl. Phys. Lett. 101, 092107 (2012).21J. G. Chen, Surf. Sci. Rep. 30, 1 (1997).22D.-Y. Cho, J. Song, K. D. Na, C. S. Hwang, J. H. Jeong, J. K. Jeong, and

Y.-G. Mo, Appl. Phys. Lett. 94, 112112 (2009).23K. B. Chung, J. P. Long, H. Seo, G. Lucovsky, and D. Nordlund, J. Appl.

Phys. 106, 074102 (2009).24B. D. Ahn, J.-S. Park, and K. B. Chung, Appl. Phys. Lett. 105, 163505

(2014).25T. S. Herng, D.-C. Qi, T. Berlijn, J. B. Yi, K. S. Yang, Y. Dai, Y. P. Feng,

I. Santoso, C. S�anchez-Hanke, X. Y. Gao, A. T. S. Wee, W. Ku, J. Ding,

and A. Rusydi, Phys. Rev. Lett. 105, 207201 (2010).26G. S. Chang, E. Z. Kurmaev, D. W. Boukhvalov, L. D. Finkelstein, S.

Colis, T. M. Pedersen, A. Moewes, and A. Dinia, Phys. Rev. B 75, 195215

(2007).27H.-W. Park, J.-S. Park, J. H. Lee, and K.-B. Chung, Electrochem. Solid

State Lett. 15, H133 (2012).

232104-5 Kim et al. Appl. Phys. Lett. 110, 232104 (2017)