Q72 ECS Solid State Letters, 2 (9) Q72-Q74 (2013) 2162-8742/2013/2(9)/Q72/3/$31.00©The Electrochemical Society
Analysis of Electrical Characteristics and Reliability Change
of Zinc-Tin-Oxide Thin-Film Transistors by Photo-Thermal
Treatment
Yu-Chun Chen,aTing-Chang Chang,a,b,zHung-Wei Li,cWan-Fang Chung,d
and Tien-Yu Hsieha
aDepartment of Physics, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan
bAdvanced Optoelectronics Technology Center, National Cheng Kung University, Tainan, Taiwan cDepartment of Photonics & Institute of Electro-Optical Engineering, National Chiao Tung University, Hsinchu 30010, Taiwan
dDepartment of Electronics Engineering & Institute of Electronics, National Chiao Tung University, Hsinchu 30010, Taiwan
This study investigates the influence of a photo-thermal treatment on the electrical characteristics and bias-induced instability of amorphous Zn-Sn-O thin film transistors. Sequences of measurements made in both vacuum and oxygen ambient reveal the most pronounced threshold voltage (VT) shift, i.e., the highest sensitivity of oxygen, occurs in a device with photo-thermal-treatment.
After gate-bias stress in vacuum, the treated device shows less VT shift than that untreated, which can be attributed to residual
adsorbed gas molecules on the backchannel. Thus, the proposed photo-thermal-treatment should be conducted before depositing a passivation layer to increase the reliability of devices after stress.
© 2013 The Electrochemical Society. [DOI:10.1149/2.010309ssl] All rights reserved.
Manuscript submitted March 20, 2013; revised manuscript received June 11, 2013. Published July 12, 2013.
Amorphous metal oxide-based semiconductors (AOSs) have demonstrated their benefits in applications such as thin film transis-tors (TFTs) in next generation displays due to their superior electrical performance, transparency to visible light and tunable carrier concen-trations even when deposited at room temperature.1,2Recently, zinc
oxide (ZnO)-based materials, such as zinc tin oxide (ZTO), and indium gallium zinc oxide, have been intensively studied for adoption as the channel material in TFTs to replace amorphous silicon, especially for flexible and transparent display applications.3,4The ZnO-based TFTs
can offer enough field effect mobility and low off-state current, which fit the high frame rate and low power consumption requirements for the future displays.
Although ZnO-based TFTs have demonstrated excellent perfor-mance, there are still some reliability issues in these devices.5–8The
electrical degradation of devices under gate bias stress has been mainly attributed to the charge trapping model and the influence of the ambi-ent atmosphere due to electric field-induced oxygen adsorption.9Even
though suitable passivation layers to eliminate the influence of ambi-ent gas on the devices have been studied,10,11certain pre-treatments
can be conducted on the active layer in order to obtain stable electrical characteristics. The mechanisms and effects of these pre-treatments on bias-induced electrical instability require further examination.
In this study, a pre-treatment based on chemical reaction theory is proposed, and the principal mechanism of electrical instability is presented to demonstrate the pre-treatment’s effect on the subsequent passivation process.
Experimental
Bottom-gate bottom-contact configured sol-gel derived a-ZTO TFTs are used in this study, and shown in the inset of Figure1. The detailed fabrication procedure of a-ZTO TFTs has been previously reported.9The TFTs were run through a sequence of stages, each of
which consists of alternating from vacuum to oxygen ambient atmo-spheres. Stages 1 and 2 were performed identically, and measurements were made under vacuum and different oxygen partial pressures of 1 torr, 10 torr, 100 torr and 760 torr oxygen ambient. The treatment is performed after Stage 2, where TFTs were subjected to either a photo-only, thermal-photo-only, or photo-thermal treatment. After treatment, sub-sequent measurements were preformed under vacuum and different oxygen partial pressures in Stage 3. All current-voltage characteristics were measured at room temperature and in a vacuum chamber with
zE-mail:[email protected]
gas-flow system and probe station using an Agilent B1500 precision semiconductor parameter analyzer. The threshold voltage (VT) was determined as the gate-to-source voltage (VG) which induced a drain current (ID) of 1 nA and subthreshold swing (SS) is determined by the equation SS= dVGS/d(log IDS) (V/dec) while the current is between 10−10and 10−9A.8,12The light illumination of 10000 lux in intensity
in this work was obtained by a halogen lamp whose spectrum is shown in the inset of Figure2.
Experimental Results and Discussion
Figure 1a shows the ID-VG electrical characteristics of a-ZTO TFTs at VD of 1 V in vacuum (1× 10−4 torr, 1.3× 10−2 pascal), and increasing oxygen ambient pressures of 1 torr (133 pascal), 10 torr (1333 pascal), 100 torr (1.3× 104 pascal) and 760 torr (1.013 × 105pascal) (Stage 1). Clearly, the V
Tvariation of a-ZTO TFTs is
Figure 1. Transfer ID -VG characteristics of a-ZTO TFTs in vacuum
(1× 10−4 torr, 1.3× 10−2 pascal), and increasing oxygen ambient pres-sures of 1 torr (133 pascal), 10 torr (1333 pascal), 100 torr (1.3×104pascal)
and 760 torr (1.013×105pascal) during (a) Stage 1 and (b) Stage 2, both before
photo-thermal treatment. (c) Stage 3, after photo-thermal treatment in vacuum (1× 10−4torr, 1.3× 10−2pascal). (d) Delta VTof device at Stages 1-3. The
inset shows the schematic cross-sectional view of a fabricated bottom-gate a-ZTO TFT.
) unless CC License in place (see abstract). ecsdl.org/site/terms_use
address. Redistribution subject to ECS terms of use (see 140.113.38.11
ECS Solid State Letters, 2 (9) Q72-Q74 (2013) Q73
Figure 2. (a) Comparison between ID-VGcharacteristics of devices after
dif-ferent treatments in vacuum ambient and without treatment in oxygen ambient. The inset shows the illumination spectrum of the halogen lamp. Delta VTof
device for (b) with thermal-only treatment and (c) with photo-only treatment for Stages 1-3.
closely correlated with the amount of surrounding oxygen molecules. The SS of the ID-VG in different environment of vacuum and oxy-gen ambient are similar, exhibiting no defect creation at the gate insulator/semiconductor interface under oxygen ambient. In general, literature has described the behavior of surrounding oxygen molecules as capturing electrons from the conduction band and causing the ad-sorption of oxygen ions (O−, O2−) on the active layer, resulting in a depletion layer in the backchannel and an increase of VTin the ZTO TFTs.13,14Subsequently, in our experiment, the chamber was pumped
down to 1× 10−4torr without breaking the vacuum and oxygen was fed into the chamber again (Stage 2). Figure 1b shows the ID-VG of the device at the same oxygen partial pressures as in Stage 1. A comparison of Figures1band1ashows that the positive VTshift of the device in Stage 2 decreases, demonstrating that the degree of VT shift, i.e., the sensitivity to oxygen, decreases. Based on the results in Figure1b, this implies that the adsorbed oxygen is not in fact removed by the vacuum process, and the oxygen tends to form chemisorption bonds. This paper suggests a photo-thermal treatment process to ef-fectively remove the chemisorbed gases.
In the photo-thermal treatment process, light illumination is ap-plied at 110◦C in a vacuum environment for 30 minutes. After apply-ing the treatment to the device, the electrical characteristic is again measured in different oxygen partial pressures (Stage 3), as shown in Figure1c, which indicates that the sensitivity to oxygen is notably increased. It exhibits a VT shift about 30 V with rare variations in SS between the ID-VG after treatment and in oxygen ambient. This proposed method is based on the phenomenon that photo-generated holes can easily migrate to the chemisorbed gas sites due to the built-in electric field and transform the chemisorbed state into a physisorbed state. It is known that the adsorbing force of the physical adsorption is mainly by Van der Waals force. The physisorbed gases can then be desorbed by substrate heating (baking), a process that can also be used to remove the residual moisture existing in the ZTO film. In addition, during the treatment process, the environment should be kept in vac-uum to ensure effective evacuation of oxygen from the chamber. In
Figure 3. Time evolution of theVTof devices with/without photo-thermal
treatment after dynamic stress in vacuum ambient for 4000 seconds.
order to prevent electron re-capture by the physically adsorbed gases on the channel layer, the illumination, heating and vacuum system should be applied at the same time.
To further verify the assumption above, illumination-only (photo-only) and baking-only (thermal-(photo-only) treatments are applied on the devices in vacuum ambient. Figure2ashows the electrical character-istics of the device before and after different treatments. It indicates obviously that the device after photo-thermal treatment exhibits the most negative VT shift when compared to that in oxygen ambient. In addition, Figures2band2cshow that the variations of VT in the devices with photo-only and thermal-only treatment during Stages 1-3, with Stages 1 and 2 representing devices that have not yet been treated. From the results of Figures1d,2band2c, the variation of VT during the first to second run shows a similar behavior. However, the degree of VT shift, i.e., the sensitivity of oxygen, is highest for the device with photo-thermal treatment. Thus, both the illumination and baking in vacuum ambient are essential to desorb the chemisorbed gases on the backchannel.
To examine the influence of the treatment on the electrical sta-bility of ZTO TFTs, gate bias stress and recovery phases (dynamic stress) were applied on the devices both with and without photo-thermal treatment in vacuum ambient.15 The gate bias stress was
carried out with the gate bias kept at VT+ 10 V and grounded source and drain (stress phase). After the stress, the gate bias was sequentially switched to ground to observe the recovery behavior (recovery phase). Figure3shows the progressive shift of VTfor devices with and without photo-thermal treatment during dynamic stresses in vacuum ambient. The device without photo-thermal treatment exhibited a more severe threshold voltage shift during the stress phase, but less recovery during the recovery phase. In general, the positive VTshift during stress has been suggested to be due to electron trapping in the preexisting traps located at the interface or in the gate dielectric. However, this result indicates that the charge trapping model is not sufficient to account for the dynamic stress behavior. Thus, the degradation of the device without treatment can be associated with residual chemisorption of oxygen from the result in Figure1.
To address this issue, the influence of the residual chemisorption of oxygen onVTin the stress and recovery phases is examined by fitting the stretched-exponential equation from the charge trapping mechanism.9,15The stretched-exponential equation is defined asV
T = VTO{1 − exp[−(t/τ)β]} where VT0is theVTat infinite time, β is the stretched-exponential exponent, and τ is the constant charac-teristic trapping time for stress phase or de-trapping time for recovery phase. Figures4aand4bshow the experimental data and fitting curves for the stress and recovery phases. Clearly, the experimental results are in good agreement with the stretched-exponential equation.
Compared to the treated device, the characteristic trapping time constant of the device without treatment decreases from 2677 to 790 seconds during stress, whileβ remains approximately constant (∼0.58), as shown in Figure4a. Even in vacuum, the electrical field-induced chemisorption of oxygen on the active backchannel may still occur during the stress phase. This result indicates that electrical field-induced chemisorption of oxygen on the backchannel of active layer
) unless CC License in place (see abstract). ecsdl.org/site/terms_use
address. Redistribution subject to ECS terms of use (see 140.113.38.11
Q74 ECS Solid State Letters, 2 (9) Q72-Q74 (2013)
Figure 4. Time dependence ofVTof devices with/without treatment during
(a) stress phase and (b) recovery phase in vacuum ambient. The measured data are well fit to the stretched-exponential equation.
as well as trapping of electrons in the interface or bulk dielectric both occur in the device without pre-treatment during the stress phase. The effects of the electrical field which can induce chemisorption of oxygen on the ZnO-based thin film has been described in a previ-ous study.16 Nevertheless, as shown in Figures4aand4b, the more
residual oxygen during dynamic stress, the less trapping time and the more de-trapping time exhibited. Accordingly, the photo-thermal treatments for removing the adsorbed gas species of the ZnO-based TFTs are indeed helpful to improve the device bias stability.
Although the effect of oxygen on ZnO-based TFTs has previously been studied, this work investigates the influence of a gas-desorption treatment process on the electrical characteristic and electrical stability
of devices. The device with treatment shows better electrical stability than the device without treatment, even in vacuum ambient. Thus, in order to obtain more stable electrical characteristics, the photo-thermal treatment proposed in this work should be conducted to improve device electrical stability.
Acknowledgment
This work was performed at National Science Council Core Facilities Laboratory for Nano-Science and Nano-Technology in Kaohsiung-Pingtung area, NSYSU Center for Nanoscience and Nanotechnology, and was supported by the National Science Council of the Republic of China under Contract Nos. NSC-102-2120-M-110-001.
References
1. Toshio Kamiya, Kenji Nomura, and Hideo Hosono,Sci. Technol. Adv. Mater., 11, 044305 (2010).
2. T-C Chang, F-Y Jian, S-C Chen, and Y-T Tsai,Mater. Today, 14(12), 608 (2011). 3. H. Q. Chiang, J. F. Wager, R. L. Hoffman, J. Jeong, and D. A. Keszler,Applied
Physics Letters, 86, 013503 (2005).
4. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono,Nature, 432, 488 (2004).
5. T-Y Hsieh, T-C Chang, T-C Chen, M-Y Tsai, Y-T Chen, and F-Y Jian,IEEE Electron Device Letters, 33(7), 1000 (2012).
6. T-C Chen, T-C Chang, T-Y Hsieh, W-S Lu, and F-Y Jian,Applied Physics Letters,
99, 022104 (2011).
7. T-Y Hsieh, T-C Chang, T-C Chen, Y-T Chen, M-Y Tsai, and A-K Chu,IEEE Trans-actions on Electron Devices, 59(12), 3389 (2012).
8. C.-K. Lee, H. Y. Jung, S. Y. Park, B. G. Son, C.-K. Lee, H. J. Kim, R. Choi, D.-H. Kim, J.-U. Bae, W.-S. Shin, and J. K. Jeong,IEEE Electron Device Letters,
34(2), 253 (2013).
9. Y-C Chen, T-C Chang, H-W Li, S-C Chen, Jin Lu, and W-F Chung,Applied Physics Letters, 96, 262104 (2010).
10. Shinhyuk Yang, Doo-Hee Cho, Min Ki Ryu, Sang-Hee Ko Park, Chi-Sun Hwang, Jin Jang, and Jae Kyeong Jeong,Applied Physics Letters, 96, 213511 (2010). 11. T-C Chen, T-C Chang, T-Y Hsieh, S-C Chen, C-S Lin, and M-C Hung,Applied
Physics Letters, 97, 192103 (2010).
12. J. S. Park, T. S. Kim, K. S. Son, J. S. Jung, K.-H. Lee, J.-Y. Kwon, B. Koo, and S. Lee,
IEEE Electron Device Letters, 31(5), 440 (2010).
13. J. Lagowski, E. S. Sproles, and H. C. Gatos,Journal of Applied Physics, 48, 3566 (1977).
14. W-F Chung, T-C Chang, H-W Li, S-C Chen, and Y-C Chen,Applied Physics Letters,
98, 152109 (2011).
15. T. Cho, J. M. Lee, J. H. Lee, and H. I. Kwon,Semiconductor Science And Technology, 24, 015013 (2009).
16. Stuart A. Hoenig and John R. Lane,Surface Science, 11, 163 (1968).
) unless CC License in place (see abstract). ecsdl.org/site/terms_use
address. Redistribution subject to ECS terms of use (see 140.113.38.11
