Reliability improvement of InGaZnO thin film transistors encapsulated under nitrogen
ambient
Chun-Yu Wu, Huang-Chung Cheng, Chao-Lung Wang, Ta-Chuan Liao, Po-Chun Chiu, Chih-Hung Tsai, Chun-Hsiang Fang, and Chung-Chun Lee
Citation: Applied Physics Letters 100, 152108 (2012); doi: 10.1063/1.3702794
View online: http://dx.doi.org/10.1063/1.3702794
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/100/15?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Electrical stability enhancement of the amorphous In-Ga-Zn-O thin film transistor by formation of Au nanoparticles on the back-channel surface
Appl. Phys. Lett. 102, 102108 (2013); 10.1063/1.4795536
Highly stable amorphous In-Ga-Zn-O thin-film transistors produced by eliminating deep subgap defects Appl. Phys. Lett. 99, 053505 (2011); 10.1063/1.3622121
Investigating the degradation behavior caused by charge trapping effect under DC and AC gate-bias stress for InGaZnO thin film transistor
Appl. Phys. Lett. 99, 022104 (2011); 10.1063/1.3609873
Effect of channel thickness on density of states in amorphous InGaZnO thin film transistor Appl. Phys. Lett. 98, 122105 (2011); 10.1063/1.3570641
Transition of dominant instability mechanism depending on negative gate bias under illumination in amorphous In-Ga-Zn-O thin film transistor
Appl. Phys. Lett. 98, 033504 (2011); 10.1063/1.3540500
For the NAE devices, the threshold voltage (Vth) shift is significantly decreased from 1.88 to
0.09 V and the reduction of saturation drain current is improved from 15.75 to 5.61 lA as compared to the bare a-IGZO counterparts after PGBS. These improvements are attributed to the suppression of negatively charged oxygen adsorption on the a-IGZO backsurface and thereby well maintain the channel potential of NAE devices, which in turn sustain the Vthduring PGBS.VC 2012
American Institute of Physics. [http://dx.doi.org/10.1063/1.3702794] Thin film transistors (TFTs) are widely used as pixel
switch in the active-matrix liquid crystal display (AMLCD). The carrier mobility and device uniformity of TFTs are well known to strongly affect the properties of AMLCD. Due to the low carrier mobility of hydrogenated amorphous silicon (a-Si:H) TFTs, a larger dimensional device is required to obtain the higher drive current and makes the a-Si:H TFTs actualize high brightness and high aperture ratio in TFT array difficultly. To overcome this problem, various crystal-lization methods such as excimer laser annealing have been proposed to enhance the carrier mobility, i.e., polycrystalline silicon TFTs (poly-Si TFTs).1 However, the uniformity of electrical characteristics in poly-Si TFTs is still a critical issue, which seriously restricted its applications. Recently, the amorphous InGaZnO TFTs (a-IGZO TFTs) have been considered as a good alternative for the AMLCD and active-matrix organic light-emitting diode (AMOLED) applications due to their higher carrier mobility than the traditional a-Si:H TFTs and superior device uniformity than the poly-Si TFTs.2–4Although the high-performance a-IGZO TFTs have been proposed,2–4 the poor device reliability is one of the main limitations for the commercial product exploitation.
It is already known that the ambient effects such as oxy-gen molecules and light illumination play important roles in the electrical degradation of metal oxide transistors.5–9The study by Kanget al. reported that the threshold voltage (Vth)
shift of a-IGZO TFTs was about 47 V as the oxygen pressure increased from 8.5 106to 760 Torr.10Therefore, for prac-tical application, it is necessary to fabricate the robust a-IGZO TFTs which can immunize against the unfavorable environment effects. In this work, we propose a simple method to construct a highly stable a-IGZO TFTs. In order to isolate the interference of environment atmosphere, the fabricated a-IGZO TFTs were finally encapsulated under nitrogen ambient. As compared with the bare TFT devices, the nitrogen ambient encapsulation (NAE) ones exhibit neg-ligible Vth shift and minor saturation drain current (IDsat)
reduction after DC positive gate bias stress (PGBS). The
pos-sible mechanism of the instability phenomenon is also investigated.
Bottom-gate inverted-staggered a-IGZO TFTs were fab-ricated on the glass substrate in this work. First, the Ti/Al/Ti (50/180/150 nm) trilayer metal film was deposited through dc-sputter system and the gate pattern was transferred by li-thography process. Next, a 300-nm-thick SiNxfilm was
de-posited by plasma-enhanced chemical vapor deposition (PECVD) at 200C as the gate insulator. A 30-nm-thick a-IGZO active layer was then deposited by the dc-sputtering system at 300 W of plasma discharge power and 5 mTorr of process pressure in a gas mixture ratio of O2/Ar¼ 20%.
Then, the thermal annealing process was carried out at 220C for 2 h in clean dry air ambient. After defining the active region, the 200-nm-thick SiOxas etch stop layer was
deposited by PECVD at 170C and patterned by dry etching to open the source/drain (S/D) contact hole, followed by the Ti/Al/Ti (50/180/150 nm) metal S/D formation. Some of these samples attached with driers, named as NAE TFTs, were then simultaneously encapsulated under the nitrogen ambient by using a covered glass. A thin adhesive layer of UV glue was applied to fix the covered glass for the protec-tion of the devices from the outer ambience. Eventually, the adhesive layer was solidified by means of UV light exposure for 10 min in N2 ambient. The photographic image of the
NAE TFTs devices was shown in Fig.1. For comparison, the conventional bare TFTs without encapsulation were also fab-ricated with the same process run.
The a-IGZO TFTs with L¼ 4 lm and W ¼ 30 lm are employed in this work. The Vthis determined at normalized
drain current (IDS L/W) ¼ 1 nA for VDS¼ 0.1 V. The
sub-threshold swing (SS) is extracted from the inverse maximum slope of the semi-logarithmic IDS versus VGS curve at
VDS¼ 0.1 V. The initial Vth, SS, and field-effect mobility
(lFE) are 0.44 V, 291 mV/decade, and 12.8 cm2/Vs for the
NAE a-IGZO TFTs, respectively, while those values are 0.81 V, 290 mV/decade, and 12.1 cm2/Vs for the bare devi-ces, as shown in Figs. 2(a) and 2(b). The corresponding
cross-sectional diagrams of the NAE and bare TFTs are also plotted in Figs. 2(a) and 2(b), accordingly. Although both devices appear the similar SS and lFE, however, the bare
devices exhibit larger initial Vth. The oxygen adsorption on
the a-IGZO surface is relatively easy for the bare TFTs, thus its initial Vthis significantly increased.
10
On the contrary, the lower Vthof NAE devices is attributed to the reduced
nega-tively charged oxygen adsorption on the a-IGZO backsurface.
Prior to the PGBS experiment, all the samples were cured on a hot plate at 100C for 2 h to decrease the influ-ence of ambient moisture. The electrical instabilities of NAE and bare devices are performed with the DC PGBS of VG¼ 30 V and VDS¼ 0 V for 1500 s at room temperature as
shown in Figs.2(a)and2(b), respectively. The Vthshift as a
function of bias stress duration is plotted in Fig.2(c)and the inset shows the SS variation with the stress time. During PGBS, the parallel transfer characteristics shift to the posi-tive direction with no significant degradation of SS for both devices are observed, which indicates that the interface state creations are negligible in a-IGZO TFTs after PGBS. How-ever, a much greater Vth shift of 1.88 V is observed for the
bare devices as compared to 0.09 V for the NAE ones under the same stress time. Previous investigations have reported that a combination of the charge trapping at the dielectric/ channel layer and the oxygen adsorption at the a-IGZO back-surface is responsible for the electrical-stress degradation of a-IGZO TFTs.11–13 Considering that both devices have an identical vertical electric field during PGBS test, the charge trapping at their dielectric/channel layers will be similar. Therefore, the remarkable Vth shift of the bare TFTs is
mainly ascribed to the influence of the ambient environment. Although the SiOxetch stop layer coated on the back channel
region can be used as a passivation layer, the outer ambience such as oxygen molecules can still penetrate through the SiOx layer due to the less dense thin film deposited by
PECVD at relatively low temperature.14,15As mentioned in previous oxide-semiconductor device literatures,11,16–18 adsorption oxygen molecules on their surface would cause a charge transfer phenomenon and the reaction could be expressed in the form of O2þ e ! O2. As the applied
gate bias is raised beyond the Vthof a-IGZO TFTs, the
con-duction electrons [e] in the a-IGZO thin film will be
consid-erably increased and then extracted by the surrounded oxygen molecules, resulting in a significant increase in nega-tively charged oxygen [O2] adsorption on the device
back-surface. For the bare TFTs, the depletion layer formed underneath the a-IGZO backsurface will therefore be enlarged and simultaneously raise the channel potential, leading to the increase of Vthas the oxygen molecules are
absorbed. The schematic plot and band diagram for the bare TFTs after PGBS are shown in Figs.3(a)and3(b)to explain the observed experimental results, respectively. In contrast, for the NAE devices, the undesirable oxygen molecules are obstructed outside the covered glass and effectively elimi-nate the oxygen penetration through the SiOxetch stop layer,
as shown in Fig. 3(c). The corresponding band diagram of NAE devices after PGBS is also shown in Fig.3(d). Due to the suppression of the oxygen adsorption on the a-IGZO backsurface for the NAE devices, the channel potential will not be elevated, thereby sustaining the Vthduring PGBS. FIG. 1. The photograph of the NAE a-IGZO TFTs. The dotted lines indicate
the covered glass substrate that was used to isolate the interference of envi-ronment atmosphere.
FIG. 2. The transfer characteristics of (a) NAE a-IGZO TFTs and (b) bare ones during PGBS of VG¼ 30 V and VDS¼ 0 V for 1500 s. The inset showed
the corresponding cross-sectional diagrams. (c) Stress time dependence of Vthshift for the NAE TFTs and bare ones. The inset showed the SS variation
with the stress time for both devices.
152108-2 Wu et al. Appl. Phys. Lett. 100, 152108 (2012)
Figures4(a)and4(b)show the output characteristics of the NAE and bare devices after PGBS of VG¼ 30 V and
VDS¼ 0 V for 1500 s, respectively. The IDsatreduction of the
bare devices (DIDsat¼ 15.75 lA) after PGBS is more severe
than the NAE ones (DIDsat¼ 5.61 lA) at VDS¼ 10 V and
VGS¼ 6 V. Owing to the large amount of oxygen adsorption
during PGBS for the bare devices, the Vth will be
signifi-cantly increased to degrade the drain current. Nevertheless, the slight IDsatreduction is still observed in the NAE devices,
which can be ascribed to the minor electron trapping at or near the dielectric/channel interface.
In summary, the effect of nitrogen encapsulation on the electrical instabilities of a-IGZO TFTs under the DC PGBS is investigated. The NAE devices exhibit superior Vth
stabil-ity (DVth¼ 0.09 V) than the bare ones (DVth¼ 1.88 V) after
PGBS at VG¼ 30 V for 1500 s. Furthermore, a small IDsat
reduction of 5.61 lA after PGBS test is also observed for the NAE TFTs as compared to 15.75 lA for the bare ones at VDS¼ 10 V and VGS¼ 6 V. The improvement can be
attrib-uted to the elimination of the negatively charged oxygen adsorption on the backsurface for the NAE devices. Practi-cally, such simple method to construct a highly stable a-IGZO TFTs is promising for the applications in AMLCD and AMOLED.
This work was supported by the National Science Coun-cil of the Republic of China under the Grant Number NSC 99-2221-E-009-168-MY3, and in part by the AU Optronics, Advanced Display Technology Research Center, for process equipment support. Also, the authors would like to thank the Nano Facility Center (NFC) of National Chiao Tung Univer-sity and the National Nano Device Laboratory (NDL) for fa-cility utilization.
1J. S. Im, R. S. Sposili, and M. A. Crowder,Appl. Phys. Lett.
70, 3434 (1997).
2
K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono,
Nature (London)432, 488 (2004).
3J. B. Kim, C. Fuentes-Hernandez, and B. Kippelen,Appl. Phys. Lett.93,
242111 (2008).
4
J. H. Na, M. Kitamura, and Y. Arakawa,Appl. Phys. Lett.93, 063501 (2008).
5
Q. H. Li, Q. Wan, Y. X. Liang, and T. H. Wang,Appl. Phys. Lett.84, 4556 (2004).
6D. Zhang, C. Li, S. Han, X. Liu, T. Tang, W. Jin, and C. Zhou, Appl. Phys. Lett.82, 112 (2006).
7
T.-C. Chen, T.-C. Chang, T.-Y. Hsieh, C.-T. Tsai, S.-C. Chen, C.-S. Lin, M.-C. Hung, C.-H. Tu, J.-J. Chang, and P.-L. Chen,Appl. Phys. Lett.97, 192103 (2010).
8
T.-C. Chen, T.-C. Chang, C.-T. Tsai, T.-Y. Hsieh, S.-C. Chen, C.-S. Lin, M.-C. Hung, C.-H. Tu, J.-J. Chang, and P.-L. Chen,Appl. Phys. Lett.97, 112104 (2010).
9J.-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).
10D. Kang, H. Lim, C. Kim, I. Song, J. Park, Y. Park, and J. G. Chung, Appl. Phys. Lett.90, 192101 (2007).
11
J. K. Jeong, H. W. Yang, J. H. Jeong, Y.-G. Mo, and H. D. Kim,Appl. Phys. Lett.93, 123508 (2008).
FIG. 4. The output characteristics of (a) NAE a-IGZO TFTs and (b) bare ones after PGBS of VG¼ 30 V and VDS¼ 0 V for 1500 s.
12Y.-C. Chen, T.-C. Chang, H.-W. Li, S.-C. Chen, J. Lu, W.-F. Chung,
Y.-H. Tai, and T.-Y. Tseng,Appl. Phys. Lett.96, 262104 (2010).
13
J.-M. Lee, I.-T. Cho, J.-H. Lee, W.-S. Cheong, C.-S. Hwang, and H.-I. Kwon,Appl. Phys. Lett.94, 222112 (2009).
14A. G. Erlat, R. J. Spontak, R. P. Clarke, T. C. Robinson, P. D. Haaland, Y.
Tropsha, N. G. Harvey, and E. A. Vogler,J. Phys. Chem. B, 103, 6047 (1999).
15K. 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).
16S. A. Hoenig and J. R. Lane,Surf. Sci.11, 163 (1968). 17S. Strassler and A. Reis,Sens. Actuators
4, 465 (1983).
18
J. M. Wu,Nanotechnology21, 235501 (2010).
152108-4 Wu et al. Appl. Phys. Lett. 100, 152108 (2012)
