JSS F
OCUSI
SSUE ONO
XIDET
HINF
ILMT
RANSISTORSThe Time Response of the On-Current for the Amorphous
In-Ga-Zn-O Thin Film Transistor to the Illumination Pulse
Ya-Hsiang Tai,a,zChun-Yi Chang,bYa-Wei Chen,band Yi-Jung ChenbaDepartment of Photonics & Institute of Display, National Chiao Tung University, Hsinchu 30010, Taiwan bDepartment of Photonics & Institute of Electro-Optical Engineering, National Chiao Tung University, Hsinchu 30010, Taiwan
In this study, the time response behavior of the amorphous indium gallium zinc oxide (a-IGZO) thin film transistors (TFTs) to the illumination pulse is analyzed. The mechanism is proposed to correlate the oxygen vacancy reacting with the light-induced electron-hole pairs. The temperature effect on the time response to the illumination pulse is also studied. The higher excitation level, either from light or temperature, results in the similar excited and recovering behaviors. The formulas for the time response are proposed to be possibly used in the simulation for the circuit performance in real situation of illumination, which is important in the development of transparent electronics using a-IGZO TFT.
© The Author(s) 2014. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND,http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is not changed in any way and is properly cited. For permission for commercial reuse, please email:oa@electrochem.org. [DOI:10.1149/2.012409jss] All rights reserved.
Manuscript submitted May 19, 2014; revised manuscript received August 4, 2014. Published August 15, 2014.This paper is part of the JSS Focus Issue on Oxide Thin Film Transistors.
Nowadays, amorphous-silicon (a-Si) thin film transistor (TFTs) have been used widely in flat-panel displays (FPD) in production, with the growing need for large area displays for home entertainment and the full adoption of digital broadcasting. Even with the success, there are still some drawbacks for the TFTs. For example, the low mo-bility leads to low open area in a pixel because of designing the large size to achieve the high current. For the state-of-art process rule with definite gate length, adopting TFTs with higher mobility is one possi-ble solution to achieve high aperture ratio in a small pixel area without sacrificing the display performance. Therefore, new TFTs with high mobility are needed to replace a-Si TFTs. Among those newly pro-posed TFTs with high mobility, poly-crystalline silicon (poly-Si) TFTs and amorphous indium-gallium-zinc-oxide (a-IGZO) TFTs attract the most attention.
Although poly-Si TFT hasμeffclose to 100 cm2/V s, it requires
ad-ditional re-crystallization steps such as excimer-laser annealing, metal seeding or solid phase crystallization. These add more complexity and costs to the process. The substrate size used by poly-Si TFT technol-ogy is about 4 generations behind what a-Si TFTs can achieve today, while a-IGZO device encounters no special issue when it comes to the glass size. A-IGZO TFTs have many advantages like low temperature process (below 300◦C) and high on/off ratio (∼106).1In addition, it is highly transparent in visible light with transmittance over 90% as illustrated due to wide bandgap (∼3 eV).
These properties open up to new applications such as transparent electronics, flexible electronics, and photo sensor.2,3Even though a-IGZO TFTs exhibit good electrical characteristics and stability in the dark state, the significant electrical instability is observed when they are illuminated. Many article reported threshold voltage (VTH) shift and mobility change after illumination at different light intensities, and those changes can recover in time.4,5In the applications of trans-parent electronics and photo sensor, the a-IGZO TFTs are expected to operate in the transmission of ambient light with frequently varying intensity. In addition, most of the reports studied the change in the devices under for long-time illumination,6–8 but the response to the illumination in short time may also induce the false operation in trans-parent electronics and should be carefully studied. Therefore, the time response of a-IGZO TFTs to the ambience and the induced instability needs to be investigated. In this paper, we analyze the response of the a-IGZO TFTs to the light changing in time, propose the descriptive model, and explain the mechanism.
zE-mail:yhtai@mail.nctu.edu.tw
Experimental
The a-IGZO TFTs in this work are based on the bottom-gate TFT devices with symmetrical source/drain (S/D) fabricated on the glass substrate. The shaped Ti/Al/Ti (50/200/50 nm) gate electrodes were capped with 400-nm-thick SiNx gate dielectric, which was deposited by plasma enhanced chemical vapor deposition (PECVD) at 370◦C. The active layer of 60-nm-thick a-IGZO film was deposited by DC magnetron sputtering system using a target of In:Ga:Zn= 1:1:1 in atomic ratio with the O2/Ar ratio about 6%. For the S/D metals,
Ti/Al/Ti of 50/200/50 nm in thickness was prepared by DC sputtering at room temperature. Then, the devices are capped with passivation at 280◦C as protection layer to avoid the disturbance of outside surround-ing. We used the very same TFT sample with width of 20μm and length of 5μm for each light-stress condition to avoid the difference from device to device.
The TFT was biased at the fixed gate voltage (VG) of 10 V and drain voltage (VD) of 10 V during the illumination. After the experiment of illumination, the TFT was put in dark environment for at least 1 hr for the better recovery to the relatively stable performance. The illumination is done by white light of LED in the intensities of light of 7616 lux, 12648 lux, and 17272 lux. Furthermore, the measurements are conducted at different temperatures of 318 K, 333 K and 348 K with fixed illumination. The electrical properties of transistor were measured in sweeping voltage or sampling mode by Keithley 4200 semiconductor parametric analyzer.
Response to the Pulse Illumination
Monitoring threshold voltage shift.— Before the study of time
re-sponse, we need to verify that the change in the drain current (ID) can reflect the change in the threshold voltage (VTH), which is pre-viously reported.9Fig.1shows the I
D-VGcurves under steady light illuminations. The inset of Fig.1shows the parallel shift of curves in the on region. This parallel shift fairly indicates that the slope of the
ID-VG curve does not change by the illumination. We can correlate VTHandIDby the transformation equation:
VT H = ID/(d ID/dVG) [1] where the slope dID/dVGof the ID-VGcurves keeps constant during the experiment of response time. Accordingly, by monitoring the change of IDbefore, during and after the illumination pulse, the time response ofVTHcan be calculated by Eq.1.
Figure 1. Drain current versus gate voltage under illumination of different
light intensities.
Positive bias stress induced instability.— Fig.2ashows the time responses ofIDshined with the light pulse at different intensities and duration, whereIDis the difference between the sampled ID(t) and its initial value ID(t= 0 s). It was initially dark and then the light was turned on for about 20 seconds and then turned off to be dark again.
It is firstly noticed that theIDdecays even before the illumination. Many articles reported the slow decrease of the drain current in the dark,10,11which is attributed to the mechanism of positive bias stress
Figure 2. Time responses ofIDto the light pulse at different intensities (a) before and (b) after deducting the PBS components.
(PBS) induced instability. It is further observed that the behavior of decay is like exponential decay with time. Therefore, as a first order approximation, we propose an equation to formulate the PBS instability, and the PBS instability can be expressed by
ID,P BS(t)= I∞,P BS− [I∞,P BS− I0,P BS]e−t/T auP B S, [2] whereI0, PBSandI∞, PBSare the initial and the expected saturation value of the current change owing to PBS, respectively, and TauPBSis the fitting parameter for the time constant.12,13In this study, we assume that this effect of PBS instability follows Eq.2even under the illumi-nation. Thus, we fit the data in the region before illumination to Eq.2
and extrapolate the curves to the light responded region, as shown by in the dotted curves in Fig.2a. The fitting parameters of these back-ground PBS curves are listed in inset table in Fig.2a. In such a way, we take the extrapolated components ofIDas the PBS background and we simply deduct them from to total response in this study.
After deducting the fitted extrapolation part of the PBS compo-nents, the curves after deduction are redrawn in Fig.2b. The trend for the excited response ofIDis induced by light, and that for the re-covering behavior corresponds to the period after the light is shut off. According to the previous report,14the excitation and recovery behav-iors are well described with a unified stretched exponential function based on the photo-induced charge trapping model, which supports the mechanism of photo induced carrier trapping.
Excited behavior.—In Fig.3, we replot the rising parts in Fig.2b, which correspond to the current increaseID excited by the illumination. The excited response ofIDcan be fitted by
ID,Excited= I∞,Excited[1− e−t/T auE xci t e], [3]
Figure 3. (a)IDand (b) log(1− ID/I∞, Excite) versus time for the excited response in the illumination period at different light intensities.
whereI∞, Exciteis the expected saturation value ofID, Excite, TauExcite is the fitting parameter, and their values are subject to change with the illumination intensity. The inset table in Fig.3lists the parameters of fitting formula for the different illumination.
The mechanism of excited behavior is correlated to the electrons generated in the conduction band via the ionization of neutral oxygen vacancy,15,16which is expressed by
VO → VO2++ 2e−, [4]
where VO and VO2+are the numbers of neutral and ionized oxygen vacancy, respectively. The continuous illumination at fixed light in-tensity keeps the process of ionizing the neutral oxygen vacancy. Here we assume VO2+is generated at a constant rate RGen.
On the other hand, in the reverse reaction, the created ionized oxygen vacancy VO2+can recapture the electrons and diminish. From the view point of chemical reaction, the rate of the reverse reaction should be proportional to the number of VO2+and it can be written as
RVO2+→VO = VO
2+/τ, [5]
whereτ reflects the reaction rate. Therefore, the net increase rate RNet of VO2+can be expressed as:
RN et= d(VO2+)/dt = RGen−RV o2+→V o= RGen−(VO2+/τ). [6] We further assume that VO2+is negligible before the illumination, i.e., VO2+ (t= 0) = 0. According to the previous report,17VO2+ is proportional to the released trap charge, which leads to the decrease ofVTH and the increase ofID according to Eq.1, as shown in Eq.7:
VO2+∝ CO X(VT H/q) ∝ ID, [7] where COXis the gate insulator capacitance and q is elementary elec-tron charge. Since VO2+andIDare in proportion as shown in Eq.7, we can rewrite the differential Eq.6into:
d[ID(t)]
dt = C −
ID(t) T auE xci t e
, [8]
where constant C and TauExciteare in the same proportion to RGenand τ in Eq.5accordingly. The solution of Eq.7is consistent with Eq.3, which is the fitting formula we used, if only
ID(∞) = CT auE xci t e. [9]
This consistency suggests that the mechanism of the excited be-havior of drain current is strongly related to the ionization of neutral oxygen vacancy induced by the illumination. Therefore, we propose thatIDincrease shown in Fig.3can reflect the VO2+increase.
Moreover, we discuss the parameters for fitting formula of excited behavior with respect to the light intensity. Since the initial value of
I0, Exciteis set to 0 A, only TauExciteandID, Exciteneed to be discussed. Firstly, as shows in Fig.4, it is observed that the TauExciteis almost independent of the light intensity. As mentioned in Eq.5, the recapture rate RVo2+→Vorepresents the reverse reaction rate that VO2+recaptures the electrons back to become VO again. This is consistent with the result that the recapture of the electrons is not affected by the light illumination.
As forI∞, Excite, it can be seen in Fig.4thatI∞, Exciteis roughly in proportion to the light intensity. Since TauExciteis constant to the light intensity, according to Eq.9,I∞, Exciteis linear to C, which is in ratio to the generation rate RGen. The higher light intensity means more electrons in VO can be released by the more incident photons. Therefore, the linear light dependence of RGenis reflected inI∞, Excite, too.
Fitting Formula for Recovering Behavior.— In Fig.5, we enlarge the falling parts in Fig.2b, which correspond to the recovering behav-ior after the light is shut off. For the recovering behavbehav-ior, since the light is shut off, only the reverse reaction in Eq.3retains. It implies the exponential decay of VO2+is in the trend of
VO2+(t)= VO2+(t= ∞) − VO2+(t= ∞) − VO2+(t= 0) e−t/τ, [10]
Figure 4. The parameter TauExciteandI∞, Exciteversus light intensity.
whereτreflects the rate of reverse reaction. The curves in Fig.6can be fitted in the form of
ID,Rec(t)= I∞,Rec− [I∞,Rec− I0,Rec]e−t/T auRec [11]
whereI0, RecandI∞, Recis the initial and expected saturation value ofID, Rec, respectively, and TauRecis the fitting parameter. The inset table in Fig.5lists the parameters of fitting formula for the different illumination.
Figure 5. IDversus time for the recovering behavior after the light of dif-ferent intensity is shut off with y-axis in (a) linear and (b) logarithmic scale.
Figure 6. The TauRecandI∞, Recversus light intensity.
The consistency of the form forIDand VO2+in Eqs.8and9as well as that in Eqs.2and6suggest that the mechanism of both the excited and recovering behaviors of drain current is strongly related to oxygen vacancy. More specifically, it is related to the ionization of neutral oxygen vacancy VO induced by the illumination and the recapture of the electrons by the ionized oxygen vacancy VO2+.
In the analysis of recovering behavior, we do not need to discuss parameter for fitting formula with the light intensity, because the light is turned off during this time. Considering only the reverse reaction of Eq.4, we instead discuss the recovering parameters with respect to the initial value ofI0, Rec.
Fig.6plots TauRecandI∞, Recversus the light intensity. It can be
seen that TauRecis nearly a constant for different cases, like TauExcite is. This also exhibits the independence of the electron reception on the light illumination. In Fig.6, it is also observed thatI∞, Recincreases with the light intensity. The non-zero value ofI∞, Recdepicts that not all but only a portion of VO2+can recapture the electrons. Some ionized oxygen vacancies change in states after the illumination, which can be long term compared to the measurement period. It is suspected that those vacancies become slow trapping states during the illumination, and thus theI∞, Recgets higher.
Temperature effect.— The experiment of light pulse response is
conducted at different temperatures. In other words, the device is firstly raised to the higher temperature. At the raised temperature, the pulse illumination with a light intensity of 17272 lux is shined on the device. In the meantime, the drain current is measured in sampling mode with fixed gate and drain voltages. Fig.7ashows the drain current change and the fitted PBS background under different temperatures at 318 K, 333 K, and 348 K, while the intensity of light is fixed at 17272 lux. At different temperatures, the slope dID/dVGof the ID− VGcurves is no more constant. Thus, we also convertIDto VTHusing Eq.1and plot them together in Fig.7aand Fig.7b.
At the first look, the trend ofVTH corresponding to the illumi-nation pulse at raised temperature acts like the wayID does to the higher light intensity. We can use the same formula to fit the behavior of time response at different temperatures. By the same procedures of fitting and deduction for the PBS background, we extract the fitting parameters and plot them in Figs.8and9.
For the excited behavior, the dependences of TauRec,ID∞, Excite,
andVTH∞, Excite on the temperature are plotted in Fig.8. It is
ob-served that theVTH ∞, Exciteincreases with temperature, which can reflect the increase in VO2+. Many papers reported that the instability mechanism in a-IGZO TFTs under PBITS conditions can be explained by the models of charge injection and defect creation. The temperature causes the VO2+increase in the IGZO.18,19The more VO2+created by temperature and light result in the increase ofVTH∞, Excite. On the
other hand, TauExciteis almost independent of the temperature, which is just like the case of illumination.
Figure 7. Time responses of the changes in drain current and threshold voltage
to the light pulse at different temperatures (a) before and (b) after deducting the PBS components.
As for the recovering behavior, Fig. 9 shows that the TauRec, ID∞, Rec, andVTH ∞, Recversus temperature. It can be seen that TauRec does not change much by the temperature. It is also noticed thatVTH∞, Recincreases with the temperature. The behaviors of the
both parameters at higher temperature exhibit similar responses to those with higher illumination intensity.
In summary, the excitations from temperature have the similar effects on the change in the drain current or threshold voltage to the illumination with higher intensity. After removing the excitation, the trend of restoring to its initial state is alike, too. The role of oxygen
Figure 8. The parametersID∞, Excite,VTH∞, Excite, and TauExciteas a func-tion of temperature.
Figure 9. The parametersID∞, Rec,VTH∞, Rec, and TauRecas a function of temperature.
vacancy is believed to be important in the behavior of the time response to the light pulse.
Conclusions
In this paper, the time response of a-IGZO TFTs to the illumina-tion pulse and the temperature effect are studied. We characterize the time response by sampling the drain current change, which reflects the voltage shift. By deducting the component of PBS instability, we de-velop two fitting formulas to depict the behavior of rise and fall of the drain current change in time upon the illumination pulse, respectively. The mechanisms are strongly related to the ionization of neutral oxy-gen vacancy and the recapture of the electrons by the ionized oxyoxy-gen vacancy. The developed formulas can be helpful to describe the drain current change in time responded to the more complex illumination or temperature conditions.
Acknowledgments
This work was supported by the Ministry of science and Technol-ogy under MOST 103-2221-E-009-061.
References
1. S. Tomita, T. Nakamura, T. Morita, T. Imai, T. Okada, H. Hayashi, Y. Saruhashi, M. Fuchi, M. Hashimoto, M. Tada, T. Endo, K. Saito, and H. Nakamura, “An In-Cell Capacitive Touch-Sensor Integrated in an LTPS WSVGA TFT-LCD,”SID Symp. Dig. Technical paper, 42(1), 629 (2011).
2. B. Kim, H. N. Cho, W. S. Choi, S. H. Kuk, Y. H. Jang, J. S. Yoo, S. Y. Yoon, M. Jun, Y. K. Hwang, and M. K. Han, “Highly Reliable Depletion-Mode a-IGZO TFT Gate
Driver Circuits for High-Frequency Display Applications Under Light Illumination,”
IEEE Electron Devices Letters, 377(4), 757 (1995).
3. T. C. Fung, C. S. Chuang, K. Nomura, H. P. D. Shieh, H. Hosono, and J. Kanick, “Photofield-effect in amorphous In-Ga-Zn-O (a-IGZO) thin-film transistors,”Journal of Information Display, 9(4), 21 (2008).
4. T. Y. Hsieh, T. C. Chang, T. C. Chen, M. Y. Tsai, Y. T. Chen, F. Y. Jian, Y. C. Chung, H. C. Ting, and C. Y. Chen, “Investigating the Drain-Bias-Induced Degradation Behavior Under Light Illumination for InGaZnO Thin-Film Transistors,”Applied Physics Letters, 99(2), 022104-022104-3 (2011).
5. C. Chen, K. Abe, T.C. Fung, H. Kumomi, and J. Kanicki, “Amorphous In–Ga–Zn– O Thin Film Transistor Current-Scaling Pixel Electrode Circuit for Active-Matrix Organic Light-Emitting Displays,”Journal of the Society for Information Display,
17(6), 525 (2009).
6. T. C. Chen, T. C. Chang, T. Y. Hsieh, M. Y. Tsai, C. T. Tsai, S. C. Chen, C. S. Lin, and F. Y. Jian, “Analyzing the effects of ambient dependence for InGaZnO TFTs under illuminated bias stress,” Surface and Coatings Technology, 231(25), 465 (2013).
7. H. X. Ming, W. C. Fei, L. Hai, X. Q. Yu, Z. Rong, and Z. Y. Dou, “Impact of Interfacial Trap Density of States on the Stability of Amorphous InGaZnO-Based Thin-Film Transistors,”Chinese Physics Letters29(6), 0673021 (2012).
8. S. Park, E. N. Cho, and I. Yun, “Instability of Light Illumination Stress on Amorphous InGaZnO Thin Film Transistors,”Journal of the Society for Information Display,
21(8), 333 (2013).
9. J. Yao, N. Xu, S. Deng, J. Chen, J. She, H. D. Shieh, P. T. Liu, and Y. P. Huang, “Electrical and Photosensitive Characteristics of a-IGZO TFTs Related to Oxygen Vacancy,”IEEE Electron Devices Society58(4), 1121 (2011).
10. T. C. Chen, T. C. Chang, T. Y. Hsieh, C. T. Tsai, S. C. Chen, C. S. Lin, F. Y. Jian, and M. Y. Tsai, “Investigation of the gate-bias induced instability for InGaZnO TFTs under dark and light illumination,”Thin Solid Films, 520(5), 1422 (2011). 11. T. C. Chang, T. Y. Hsieh, W. S. Lu, F. Y. Jian, C. T. Tsai, S. Y. Huang, and C. S. Lin,
“Investigating the degradation behavior caused by charge trapping effect under DC and AC gate-bias stress for InGaZnO thin film transistor,”Applied Physics Letters,
99(2), 022104-022104-3 (2011).
12. B. Kim, H. N. Cho, W. S. Choi, S. H. Kuk, Y. H. Jang, J. S. Yoo, S. Y. Yoon, M. Jun, Y. K. Hwang, and M. K. Han, “Highly Reliable Depletion-Mode a-IGZO TFT Gate Driver Circuits for High-Frequency Display Applications Under Light Illumination,”
IEEE Electron Device letters, 33(4), 528 (2012).
13. J. K. Jeong, “The status and perspectives of metal oxide thin-film transistors for active matrix flexible displays,”Semiconductor Science and Technology, 26(3), 034008 (2011).
14. K. Hoshino, D. Hong, H. Q. Chiang, and J. F. Wager, “Constant Voltage Bias Stress Testing of a-IGZO Thin-Film Transistors,”IEEE Transactions on Electron Devices,
56(7), 1365 (2009).
15. J. Yao, N. Xu, S. Deng, J. Chen, J. She, and H. P. D. Shieh, “Electrical and Photosensi-tive Characteristics of a-IGZO TFTs Related to Oxygen Vacancy,”IEEE Transactions on Electron Devices, 58(4), 1121 (2011).
16. K. W. Lee, H. S. Shin, K. Y. Heo, K. M. Kim, and H. J. Kim “Light effects of the amorphous indium gallium zinc oxide thin-film transistor,”Journal of Information Display, 10(4), 1598 (2009).
17. S. Y. Lee, S. H. Kuk, M. K. Song, and M. K. Han, “The hysteresis and off-current of amorphous indium-gallium-zinc oxide thin film transistors with various active layer thicknesses under the light illumination,” Active-Matrix Flatpanel Displays
and Devices, 2012 19th International Workshop on, pp. 151–154.
18. J. M. Lee, I. T. Cho, J. H. Lee, and H. I. Kwon, “Bias-stress-induced stretched-exponential time dependence of threshold voltage shift in InGaZnO thin film transis-tors,”Applied Physics Letters, 93(9), 093504-093504-3 (2008).
19. H.S. Shin, Y. S. Rim, Y. G. Mo, C. G. Choi, and H. J. Kim, “Effects of high-pressure H2O annealing on amorphous IGZO thin-film transistors,”Phys. Status Solidi, 208(9),
