Anomalous Stress-Induced Hump Effects in Amorphous Indium Gallium Zinc Oxide TFTs
Anomalous Stress-Induced Hump Effects in Amorphous Indium Gallium Zinc Oxide TFTs
Transactions on Electrical and Electronic Materials. 2012. Feb, 13(1): 47-49
Copyright ©2012, The Korean Institute of Electrical and Electronic Material Engineers
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
  • Received : November 29, 2011
  • Accepted : January 21, 2012
  • Published : February 25, 2012
Export by style
Cited by
About the Authors
Yu-Mi Kim
Kwang-Seok Jeong
Ho-Jin Yun
Seung-Dong Yang
Sang-Youl Lee
Hi-Deok Lee

In this paper, we investigated the anomalous hump in the bottom gate staggered a-IGZO TFTs. During the positive bias stress, a positive threshold voltage shift was observed in the transfer curve and an anomalous hump occurred as the stress time increased. The hump became more serious in higher gate bias stress while it was not observed under the negative bias stress. The analysis of constant gate bias stress indicated that the anomalous hump was influenced by the migration of positively charged mobile interstitial zinc ion towards the top side of the a-IGZO channel layer.
Oxide-based thin-film transistors (oxide TFTs) such as amorphous indium- gallium zinc oxide (a-IGZO), zinc indium oxide (ZIO), zinc tin oxide (ZTO) have shown excellent electri-cal properties, including a high mobility and an excellent on/off current ratio. Especially, a-IGZO TFTs have been recognized as a very promising alternative to display backplanes of active matrix organic light emitting diodes (AMOLEDs) and thin-film transistor liquid crystal displays (TFT-LCDs) [1 , 2] . For practical applications, it was important to have the bias independent reliability and a number of groups have studied the stability of a-IGZO TFTs [3 - 7] . Particularly, the hump characteristics were considered as one of the critical issues during the operation in TFT displays affecting pixel brightness [8] . It has been reported that the stability of oxide TFTs are influenced by the absorption/desorption of gas molecules from ambient atmosphere which can be prevented by forming the passivation/capping layer on the exposed back-channel surface.
According to Ref. [9] of Huang et al, the absorption of H 2 O molecules in the IGZO back channel can cause a hump in the transfer curve, which can be suppressed by the passivation layer.Even in some devices with passivation layer, however, the hump had been observed. In Ref. [10] of Tsai et al, the positive bias stress-induced hump was reported in N 2 O plasma treated IGZO TFTs with SiO x capping layer where the hump was not observed in untreated device with the same capping layer. Although several studies have previously introduced the hump characteristics in ZnO-based TFTs, the detailed investigation on the origin of the hump has not been clarified yet.
In this work, we analyzed the anomalous hump after the elec-trical stress in a-IGZO TFTs with the bottom gate structure.
The bottom gate staggered a-IGZO TFTs were produced on glass substrate. Devices were fabricated as follows: First, sputter-deposited Molybdenum (Mo) was patterned on glass substrates to form the gate. Then, 400 nm thick Si 3 N 4 gate insulator film was deposited by plasma-enhanced chemical vapor deposition (PECVD). Next, 70 nm thick amorphous indium gallium zinc oxide (a-IGZO) layer was grown by radio-frequency sputtering. The Mo was formed by dc-sputtering as the source and drain electrodes. After forming the source and drain on device, 200 nm
Lager Image
The cross-sectional (a) schematic and (b) TEM image of a-IGZO TTFs.
Lager Image
Transfer characteristics a-IGZO TFTs.
thick SiO x was deposited as a passivation layer by PECVD. Finally, the sample was annealed in O 2 atmosphere at 300℃. The cross-sectional schematic and the transmission electron microscopy (TEM) image of the device are shown in Fig. 1 (a) and (b), respec-tively.
The TFT dimension used for this study was 90 by 8 μm 2 ( W×L ). The electric characteristic analysis of the device was carried out by using an Agilent 4155B semiconductor parameter analyzer at room temperature, in ambient air and in the dark.
The transfer characteristics of a-IGZO TFTs are shown in Fig. 2 and the inset shows the schematic cross section of the device.The threshold voltage ( Vth ) of 0.47 V was determined by using constant drain current method (0.1 nA × W/L ). The sub-threshold swing (SS) and on/off current ratio were 1.13 V/decade and 9.62 ×10 7 , respectively. The field-effect mobility (μ FE ) was extracted from the linear-region as a function of gate voltage, and the device ex-hibited effective mobility ranging between 0.18-1.34 cm 2 /V s.
Figure 3 (a) shows the transfer characteristics of a-IGZO TFTs in accordance with the gate-bias stress time at a fixed drain volt-age of 0.05 V. The constant gate bias stress of 20 V was applied while the source and drain electrode was grounded. A positive shift was observed with unchanged sub-threshold slope in the beginning. The threshold voltage shift could be explained by using an electron charge trapping mechanism [4] , [11] . The trapped
Lager Image
Transfer characteristics of a-IGZO TFTs under positive bias stress at VD=0.05 V (a) The bias stress conditions are VG=20 V and VSD=0 V. (b) The positive bias conditions are VG=10 V/ 20 V/ 40 V and VSD=0 V. The anomalous hump is observed in the sub-threshold region while the duration of positive bias stress over 300 s and more serious as the bias stress voltage increases.
Lager Image
Transfer characteristics of a-IGZO TFTs under negative bias stress at VD=0.05 V. The bias stress conditions are VG=-20 V and VSD=0 V.
electrons in the interface between the active layer and the gate oxide by the positive gate bias reduced the effective applied gate voltage, which caused a positive shift in the threshold voltage. In addition, an anomalous hump also came out for the duration of bias stress over 300 s and such a transfer curve seemed to be the combination of a dominant current path and a parasitic current path. Transfer characteristics of a-IGZO TFTs under different gate bias stress are shown in Fig. 3 (b). The applied constant gate bias were 10 V, 20 V and 40 V, respectively, while the source and drain electrode was grounded. The hump appeared more clearly as gate bias stress voltage increased.
However, during negative gate bias stress, the hump was not observed. Figure 4 shows the transfer characteristics according to the constant gate bias stress of -20 V as the function of stress time. A negative shift of threshold voltage was observed with no hump and an unchanged sub-threshold slope in the transfer curve.
In poly-Si TFTs with top gate structure, the stress-induced hump was reported by the generation of the parasitic transistor as a result of the high electric field due to thinner insulator thick-ness at the active edge along the channel width direction [12] . In this works, however, the device has bottom gate structure where the gate insulator was formed after the gate patterning and so the parasitic transistor formation in the active edge was difficult. Moreover, the insulator thickness thinning at the edge was not exhibited as shown in the cross-sectional TEM image of Fig. 1 (b).
It was generally accepted that free carriers in a-IGZO mainly originated from point defects (vacancies, interstitials, and antisites) in the active layer like as other ZnO-based film. There-fore, the electrical characteristics of a-IGZO TFTs were affected significantly by the density of these point de-fects, which will be determined by the process and ambient. Among these point defects, it was well known that the free electrons mainly originated from oxygen vacancies, the thermally excited oxygen atoms leaving free carriers at sites [13 - 15] . While these intrinsic oxygen vacancies were stable, there were also metastable point defects consisting of mobile positively charged zinc interstitial ions. The migration barrier of the zinc interstitial was so low that the zinc interstitials were mobile below room temperature [16] . The zinc interstitials were often referred to in the atomic defect model for Schottky barrier at the grain boundary in polycrystalline ZnO-base material due to their charge and mobile characteristics [17 - 19] . In the bottom gate a-IGZO, it was possible during the posi-tive gate bias stress, that the positively charged zinc interstitials would be accumulated at the top side of the channel layer by the field-induced migration and lower the body potential accelerat-ing the sub-channel formation. This means that the a-IGZO TFT with the bottom-gate structure had two parallel current paths under the constant gate bias stress. One was the main path through the accumulation layer induced by the gate voltage and the other was the parasitic path by the mobile zinc interstitials at the top side of the channel layer. Moreover, considering the electron trapping occurred under the gate in the positive gate bias stress as mentioned before, the sub-channel can turn on earlier. During the negative bias stress, the mobile zinc intersti-tial moved toward the gate and could not affect the hump.
Therefore, the transfer characteristics of a-IGZO TFT might be improved without humps through the process adjustment by which the formation of mobile zinc interstitial could be sup-pressed.
In this study, the anomalous hump in a-IGZO TFTs with the bottom-gate structure was investigated under constant positive bias stress and negative bias stress. During the constant positive bias stress, the anomalous hump was observed in the trans-fer curve, which seemed to be the combination of a dominant current path and a parasitic current path. However, the hump was not observed under the negative bias stress. The positively charged mobile zinc interstitial ions were accumulated near the top side of the active layer by the field-induced migration under the positive gate bias stress. Then, those lower the body potential and accelerate sub-channel formation at the top side of the a-IGZO active layer.
This research was financially supported by the Ministry of Ed-ucation,Science Technology (MEST) and the National Research Foundation of Korea (NRF) through the Human Resource Train-ing Project for Regional Innovation, and by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2010-0028160).
Jeong J. K , Jeong J. H , Choi J. H , Im J. S , Kim S. H , Yang H. W , Kang K.N , Kim K. S , Ahn T. K , Chung H. J , Kim M. K , Gu B. S , Park J. S , Mo Y. G , Kim H. D , Chung H. K (2008) SID Int. Symp. Digest Tech. Papers 39 1 -    DOI : 10.1889/1.3069591
Lee J. H , Kim D. H , Yang D. J , Hong S. Y , Yoon K. S , Hong P. S , Jeong C. O , Park H. S , Kim S. Y , Lim S. K , Kim S. S , Son K. S , Kim T. S , Kwon J. Y , Lee S. Y (2008) SID Int. Symp. Digest Tech. Papers 39 625 -    DOI : 10.1889/1.3069740
Lee J. M , Cho I. T , Lee J. H , Kwon H. I (2008) Appl. Phys. Lett. 93 093504 -    DOI : 10.1063/1.2977865
Suresh A , Mutha J. F (2008) Appl. Phys. Lett. 92 033502 -    DOI : 10.1063/1.2824758
Triska J , Conley J. F , Presley R , Wager J. F (2010) J. Vac. Sci. Technol. B 28 C5I1 -    DOI : 10.1116/1.3455494
Moon Y. K , Lee Sih , Kim W. S , Kang B. W , Jeong C. O , Lee D. H , Park J. W (2009) Appl. Phys. Lett. 95 013507 -    DOI : 10.1063/1.3167816
Lee J. S , Park J. S , Pyo Y. S , Lee D. B , Kim E. H , Stryakhilev D , Kim T. W , Jin D. U , Mo Y. G (2009) Appl. Phys. Lett. 95 123502 -    DOI : 10.1063/1.3232179
Lee J. H , Park S. G , Han S. M , Han M. K , Park K. C (2008) Solid-State Electronics 52 462 -    DOI : 10.1016/j.sse.2007.10.030
Huang S. Y , Chang T. C , Chen M. C , Liau W. L (2011) Electrochemical and Solid-State Letters 14 H177 -    DOI : 10.1149/1.3534828
Tsai C. T , Chang T. C , Chen S. C , Lo Ikai , Tsao S. W , Hung M. C , Chang J. J , Wu C. Y , Huang C. Y (2010) Appl. Phys. Lett. 96 242105 -    DOI : 10.1063/1.3453870
Libsch F. R , Kanicki J (1993) Appl. Phys. Lett. 62 1286 -    DOI : 10.1063/1.108709
Huang C. F , Peng C. Y , Yang Y. J , Sun H. C , Chang H. C , Kuo P. S , Chang H. L , Liu C. Z , Liu C. W (2008) IEEE Electron Device Letters 29 1332 -    DOI : 10.1109/LED.2008.2007306
Kofstad P (1962) J. Phys. chem. Solids 23 1571 -    DOI : 10.1016/0022-3697(62)90239-1
Bonasewicz P , hirschwald W , Neumann G (1986) Phys. Status solidi A 97 593 -    DOI : 10.1002/pssa.2210970234
Gavryushin V , Raciukaitis G , Juodzbalis D , Kazlauskas A , Kubertavicius V (1994) J. Cryst. Growth 138 924 -    DOI : 10.1016/0022-0248(94)90933-4
Janotti Anderson , Van de Walle Chris G (2007) Phys. Rev. B 76 165202 -    DOI : 10.1103/Phys-RevB.76.165202
Gupta T. K (1990) J. Am. Ceram. Soc. 73 1817 -    DOI : 10.1111/j.1151-2916.1990.tb05232.x
Gupta T. K , Carlson W. G , Hower P. L (1981) J. Appl. Phys. 52 4104 -    DOI : 10.1063/1.329262
Jiang S. L , Xie T. T , Yu H , Huang Y. Q , Zhang H. B (2008) Microelectronic Engineering 85 371 -    DOI : 10.1016/j.mee.2007.07.011