Persistent photoconductivity (PPC) effect and its light-intensity dependence of both enhancement and depletion (E-/D-) mode amorphous InGaZnO (a-IGZO) thin-film transistors (TFTs) are systematically investigated. Density of oxygen vacancy (VO) defects of E-mode TFTs is relatively small, in which formation of the photo-induced metastable defects is thermally activated, and the activation energy (Ea) decreases continuously with increasing light-intensity. Density of VO defects of D-mode TFTs is much larger, in which the formation of photo-induced metastable defects is found to be spontaneous instead of thermally activated. Furthermore, for the first time it is found that a threshold dose of light-exposure is required to form fully developed photo-induced metastable defects. Under low light-exposure below the threshold, only a low PPC barrier is formed and the PPC recovery is fast. With increasing the light-exposure to the threshold, the lattice relaxation of metal cations adjacent to the doubly ionized oxygen vacancies () is fully developed, and the PPC barrier increases to ∼ 0.25 eV, which remains basically unchanged under higher light-exposure. Based on the density of VO defects in the channel and the condition of light illumination, a unified model of formation of photo-induced metastable defects in a-IGZO TFTs is proposed to explain the experimental observations.
Received: 09 July 2020
Revised: 18 August 2020
Accepted manuscript online: 25 August 2020
Fund: Project supported in part by the National Natural Science Foundation of China (Grant Nos. 61974101 and 61971299), the State Key Laboratory of ASIC and System, Fudan University (Grant No. 2019KF007), the Natural Science Foundation of Jiangsu Province, China (Grant No. SBK2020021406), the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (Grant No. 19KJB510058), and the Suzhou Science and Technology Bureau (Grant No. SYG201933).
Yalan Wang(王雅兰), Mingxiang Wang(王明湘), Dongli Zhang(张冬利), and Huaisheng Wang(王槐生) A systematic study of light dependency of persistent photoconductivity in a-InGaZnO thin-film transistors 2020 Chin. Phys. B 29 118101
Fig. 1.
Schematic cross-sectional diagram of an inverted staggered back gate a-IGZO TFT.
Fig. 2.
Transfer curves of two kinds of a-IGZO TFTs by simulation, and measured in dark state and under different wavelength illumination under VDS = 1 V in logarithmic scale (left, bottom axis) and linear scale (right, up axis).
E-mode a-IGZO TFTs
Parameters
Value
Description
NC/cm−3
2.0×1019
effective DOS in CBM
NV/cm−3
5×1018
effective DOS in VBM
Eg/eV
3.2
band-gap
Tail states acceptor-like DOS
gTA(E) = NTA exp[(E − EC)/WTA]
NTA/eV−1⋅cm−3
1.4×1018
density of tail states at E = EC
WTA/eV
0.36
conduction-band-tail slope
Tail states donor-like DOS
gTD(E) = NTDexp[(EV − E)/WTD]
NTD/eV−1⋅cm−3
4.5×1019
density of tail states at E = EV
WTD/eV
0.1
Valence-band-tail slope
Deep donor-like DOS
gGD(E) = NGD exp{−[(E − EGD)/WGD]2}
NGD/eV−1⋅cm−3
1×1017
density of donor-like at E = EV
EGD/eV
1
mean energy of donor-like
WGD/eV
0.1
standard deviation of donor-like
D-mode a-IGZO TFTs
Deep donor-like DOS
gGD(E) = NGD exp{−[(E − EGD)/WGD]2}
NGD/eV−1⋅cm−3
5×1018
density of donor-like at E = EV
EGD/eV
1.3
mean energy of donor-like
WGD/eV
0.1
standard deviation of donor-like
Shallow donor-like DOS
gOV(E) = NOV exp{−[(E − EOV)/WOV]2}
NOV/eV−1⋅cm−3
1×1018
peak of OV states
EOV/eV
2.9
mean energy of OV states
WOV/eV
0.23
standard deviation of OV states
Table 1.
Fitting parameters of the two a-IGZO TFTs.
Fig. 3.
The time dependent photo-induced current ΔI (difference between ID under light and dark) are fitted with stretch-exponential model of an E-mode a-IGZO TFT for different light-intensities respectively in the (a) photo-excitation and (b) recovery stages. Hollow circles are experimental data, and solid lines are fitting curves.
Fig. 4.
Light-intensity dependence of (a) IS; (b) βex, βre; and (c) τex, τre.
Fig. 5.
The T dependence of (a) IS; (b) βex, βre; and (c) τexτre under three light-intensities of 0.3 mW/cm2, 1 mW/cm2, and 5 mW/cm2.
Light-intensity
Ea_ex/eV
Ea_re/eV
0.3 mW/cm2
0.25
0.22
1 mW/cm2
0.13
0.19
5 mW/cm2
0.11
0.15
Table 2.
Comparison of Ea_ex and Ea_re under three light-intensities.
Fig. 6.
The time dependent photo-induced current Δ I are fitted with the stretch-exponential model of a D-mode a-IGZO TFT for different light-intensities respectively in the (a) photo-excitation and (b) recovery stages. Hollow circles are experimental data, and solid lines are fitting curves.
Fig. 7.
Light-intensity dependence of (a) IS; (b) βex, βre; and (c) τex, τre.
Fig. 8.
Light-exposure dependence of (a) IS; (b) βex, βre; and (c) τex, τre in double logarithmic scales.
Fig. 9.
The change of ΔI with illumination time in photo-excitation stage at five different temperatures for light-exposures of (a) 6 mJ/cm2, (b) 30 mJ/cm2, (c) 60 mJ/cm2, (d) 150 mJ/cm2, and (e) 3000 mJ/cm2.
Fig. 10.
The time dependent normalized current ΔI are fitted with stretch-exponential model for different temperatures in the recovery stage at (a) 30 mJ/cm2 and (b) 3000 mJ/cm2; (c) 1/kT dependence of ; (d) light-exposure dependence of Ea_re extracted from (c).
Samples
Photo-excitation
PPC recovery
IS
βex
τex
Ea_ex
βre
τre
Ea_re
E-mode TFTs
1
↓
−1
↓
↓
−1
↓
D-mode
< DT
1.1
−0.25
−0.28
−
−0.12
1.7
↑
TFTs
> DT
0.16
−0.11
−0.28
−
−
slightly ↓
slightly ↓
Table 3.
Comparison of PPC light dependency of E-mode and D-mode a-IGZO TFTs. The values represent the power exponent n in the power exponent relationship between the relevant parameters and light-intensity or exposure, and the arrows (↑ ↓) represent the increase or decrease with the increase of the light-intensity or exposure.
[1]
Nomura K, Ohta H, Takagi A, Kamiya T, Hirano M, Hosono H 2004 Nature432 488 DOI: 10.1038/nature03090
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
