Nitrogen-doped indium tin oxide (ITO) has been applied in the thin-film strain gauges (TFSGs) due to their high stability, excellent piezoresistivity and antioxidation at elevated temperatures. However, the mechanism on the sensing and stability of the nitrogen-doped ITO TFSGs at high temperatures was not comprehensively clarified. In this work, various ITO TFSGs were fabricated by RF magnetron sputtering with different nitrogen partial pressures (NPPs) of 5%~40%. The elemental composition and band structures of the ITO TFSGs were examined by the energy dispersive X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS), respectively. Results show that the Fermi energy level shifts closer to the valence band maximum energy (Ev) gradually with the growth of NPPs, causing a reduction in the number of electrons ionized to the conduction band. The smallest content change rates of nitrogen (3.8%) and oxygen (1.6%) after subjecting to the thermal strain test were observed in the 20%N2 ITO TFSG. In consequence, the 20%N2 ITO TFSG exhibits the lowest resistance drift rate (DR) at high temperatures due to its stable elemental composition. Moreover, we found that the band structures and elemental composition of the ITO TFSGs are the main factors affecting their piezoresistive response at different temperatures. The band structures play a major role in the gauge factors (GFs) of the ITO TFSGs at room temperature and 600 ℃. The element variation takes responsibility for the different GFs of the ITO TFSGs at 800℃, 900℃ and 1000℃. In addition, the piezoresistive stability is also dependent on the elemental composition affected by the dynamic equilibrium between the diffusion amount of oxygen and the escape number of the nitrogen in the ITO thin films at high temperatures.
Research Article
Investigation on Sensing and Stability of Nitrogen-doped ITO Thin-film Strain Gauges at High Temperatures
https://doi.org/10.21203/rs.3.rs-907891/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Nitrogen-doped ITO
Thin-film strain gauges
Piezoresistive response
High temperature
Ultra-high temperature thin-film sensors (UTTSs) have been widely used to measure the structural behavior of the hot sections in harsh environments including elevated temperature, high pressure and rigorous mechanical loading[1–3]. The UTTSs can be directly fabricated on the surface of the hot components since they are non-intrusive for the multiple physical parameters measurements such as strain[4–6], temperature[7–10] and heat flux[11–13]. As one of the UTTSs, thin-film strain gauges (TFSGs) are robust enough to in-situ measure the strain of hot components in harsh environment since they have negligible mass and considerably less thickness than that of the gas-phase boundary layer[14, 15]. Therefore, the surface airflow fields and vibration patterns will be minimally disturbed by these sensors. Recently, indium tin oxide (ITO) has been applied in the TFSGs due to its high temperature endurance, excellent piezoresistivity and antioxidation at elevated temperatures. However, ITO TFSGs exhibit resistance drift with the time at high temperatures, which will limit the improvement of the measurement accuracy and stability [1].
Literature shows that the piezoresistivity of the ITO TFSGs was conditional upon their microstructures and electrical properties, whereby these characteristics were significantly affected by their elemental composition. For example, Fang and his coworkers prepared the ITO TFSGs with different oxygen concentrations using pulsed laser deposition[16]. Their results suggest that the piezoresistivity of ITO TFSGs increased with the growth of oxygen content. Tao You et al introduced aluminum into the ITO films by co-sputtering with Al2O3 and ITO targets to investigate the stability of the aluminum-doped ITO TFSGs at high temperatures[17]. Results indicate that aluminum-doped ITO TFSGs have good thermal stability at elevated temperatures up to 1450℃. However, it is hard to achieve homogeneous doping in the ITO TFSGs, which will affect their property uniformity.
Some researchers had reported that nitrogen-doped ITO TFSGs could have more stable piezoresistive performance than undoped ITO TFSGs. Gregory et al prepared ITO strain sensors using reactive sputtering in nitrogen-rich plasma ambient[18]. Compared with the ITO TFSGs prepared in an argon/oxygen environment, the nitrogen-doped sensors exhibited better performance in electrical stability and piezoresistive response at high temperatures due to the formation of the controlled nano-porosity. Yang and his co-authors fabricated the ITO TFSGs using magnetron sputtering under argon/oxygen ambient and introduced nitrogen into the ITO TFSGs via heat-treatment in nitrogen [19, 20]. Their results suggest that the nitrogen-doped ITO TFSGs had a smaller but more stable gauge factor (GF) than undoped sensors. However, these methods cannot control the nitrogen doping content and are difficult to accurately modify the performance of the ITO TFSGs. The effect of different nitrogen doping concentration on the piezoresistive stability and the GFs of the ITO TFSGs have not been comprehensively illuminated in those works. More details between the nitrogen incorporation and the piezoresistive response of ITO TFSGs still need further clarification.
The purpose of this paper is to further investigate the effect of nitrogen doping concentration on the sensing and stability, various ITO TFSGs were prepared using RF magnetron sputtering with different N2/Ar ratios ranging from 5%~40%. The elemental composition and band structure of the ITO TFSGs were examined, and their piezoresistive response was also recorded. The influence mechanism of the nitrogen-doping content on the piezoresistive stability and the GFs of the ITO TFSGs were revealed. Results show that 20%NPP was the optimal nitrogen incorporation parameter due to its high stability and high GF at elevated temperatures
2.1 TFSGs preparation
Different ITO TFSGs were prepared on high-purity alumina substrates using MEMS fabrication technology, as shown in Fig.1. The alumina substrates were cut into the constant strain beams according to the designed shape and size (Fig.2d), which can provide adequate electrical insulation at elevated temperatures. The alumina constant strain beams were ultrasonic cleaned successively with Acetone, Ethanol, and Distilled Water and blow-dried by nitrogen (Fig.1a). They were then baked at 150℃ for 120 minutes prior to the TFSGs preparation. Subsequently, different ITO films were deposited onto the substrates using MW-ECR plasma enhanced magnetron sputtering system with a high-purity ITO target (99.99 wt.%, Ф68*3mm) comprised of 90 wt.% Indium oxide and 10 wt.% Tin oxide (Fig.1b). The sputtering parameters are listed in Table.1. The mixed gas with different NPPs of 5%~40% was leaked into the sputtering chamber and working pressure was set to 0.6 Pa. The sputtering time is one hour. After the deposition, these sputtered films were spin-coated with a positive photoresist (BP212, Kempur Microelectronics Ins, Beijing, China) at a speed of 2400r/min (Fig.1c). The photoresist was soft baked at 85℃ for 30minutes (Fig.1d), followed by patterning with a film mask via contact exposure (Fig.1e) and developing in 0.5 wt.% sodium hydroxide solution for the 30s (Fig.1f). Subsequently, these patterned samples were etched in Aqua regia to delineate the final TFSGs structures (Fig.1g). Ultimately, the patterned TFSGs were placed in acetone to remove the residual photoresist (Fig.1h). The ITO TFSGs were heat-treated at 1000°C in a programmable muffle furnace (TongKe, TCXQ-1700, China) for two hours.
Table 1 Sputtering parameters of the ITO TFSGs under different NPPs.
|
Nitrogen partial pressure (%) |
The flow ratio between N2 and Ar (sccm) |
Working power (W) |
Base pressure (Pa) |
Working pressure (Pa) |
Sputtering temperature (℃) |
|
|
5%N2 ITO TFSG |
1/19 |
300 |
3×10-3 |
0.6 |
Room temperature |
|
|
10%N2 ITO TFSG |
2/18 |
|||||
|
20%N2 ITO TFSG |
4/16 |
|||||
|
30%N2 ITO TFSG |
6/14 |
|||||
|
40%N2 ITO TFSG |
8/12 |
|||||
2.2 Characterization and static strain testing
The valence band structures of the various ITO TFSGs were examined by X-ray photoelectron spectroscopy (XPS, K-Alpha+, America) calibrated using C1s line (284.8 eV). The energy dispersive X-ray spectroscopy (EDS, NOVA Nano SEM450, FEI, America) was used to analyze the elemental composition before and after the thermal strain test. The piezoresistive response of the different ITO TFSGs was tested in a high-temperature furnace programmed from room temperature to 1000℃, as shown in Fig.2a-2c. The furnace was ramped at 30℃/min to the specified temperature and maintain for at least 30 minutes to establish thermal equilibrium. The strain was induced by an alumina rod driven by a linear motor to compress the constant strain beam fixed by the Al2O3 ceramic clamping. A laser displacement sensor was used to record the applied strain of the constant strain beams by measuring their deflection and interfaced to an acquisition card via an I/O board. The corresponding resistance variations of the ITO TFSGs were noted using a digit multimeter (DMM6500 6½, TEKTRONIX, America).
The piezoresistive response of the ITO TFSGs is described by gauge factor (GF) as shown in equation (1) , which is dependent on the initial resistance and the resistance change (△R) induced by the applied strain (ε).
In addition, the stability of the ITO TFSGs is evaluated by drift rate (DR), and it is defined as equation (2):
Where Rref is the initial resistance at a specified testing temperature, μ is the Poisson ratio related to the geometric dimension of TFSGs, ρ is the resistivity. Generally, μ is a constant. Therefore, the GF of ITO TFSGs mainly depends on the variation of resistivity, which is determined by the carrier concentration and mobility at elevated temperatures. △t is the time during the test of resistance change △R.
3.1 Characterizations of the ITO TFSGs
Fig.3a shows the XPS valence band spectra of the different ITO TFSGs, and the mechanism of nitrogen incorporation is revealed. A peak at the binding energy of 4.65eV is observed, which might be due to the In5p-O2p interaction and correspond to the orbital state of p-like valence band[21, 22]. The peak at 4.65 eV shifts to 3.5 eV with the increase of NPPs, which can be attributed to a weaker electronegativity of nitrogen (3.04) than that of oxygen (3.44). This can be deduced that more nitrogen atoms were incorporated into the ITO TFSGs and substituted a portion of oxygen atoms with the increase of NPPs. The valence band maximum energy (Ev) of the ITO TFGSs was obtained from the linear fitting of the leading edge of the valence band and the flat energy distribution based on the XPS data, and finding the crossover point of these two lines[23]. As shown in Fig.3a, the Ev of the ITO TFSGs decreases from 2.88 to 1.85 eV when the NPP increases from 5% to 40%. Intrinsic ITO was produced by introducing tin into indium oxide, which forms a donor impurity level Ed, as shown in Fig3b. Each tin atom doping in In2O3 has a valence of four and indium is three, which can contribute one electron to the conduction band. When nitrogen atoms replace oxygen atoms in the process of nitrogen doping, each nitrogen atom will produce an electron hole. In consequence, an acceptor impurity level Ea was formed in the nitrogen-doped ITO TFSGs, as shown in Fig3c. The donor impurity level Ed will compensate a part of the acceptor impurity level Ea, which causes the Fermi energy level (EF) of the nitrogen-doped ITO TFSGs to shift closer to the Ev with the increase of NPPs[22], as shown in Fig.3c. Moreover, according to the results in the literature, the nitrogen incorporation will not influence the band gap(Eg, as shown in Fig3b and 3c) of ITO[24]. In that case, the energy difference between the Fermi energy level (EF) and the conduction band (Ec) will become larger with the increase of NPPs. Therefore, this will cause a reduction of the carrier concentration due to the decrease of the number of electrons ionized to the conduction band in the ITO TFSGs with the growth of NPPs.
The elemental composition is an important factor affecting the piezoresistive performance. Fig.4 shows the nitrogen and oxygen content of the ITO TFSGs prepared with various NPPs before and after the thermal strain test. Herein, the element content is defined as the proportion of nitrogen and oxygen in the sum of the four elements containing nitrogen, oxygen, indium and tin. Initially, the nitrogen content decreases and the oxygen content of the ITO TFSGs grows with the increase of NPPs before the thermal strain test. After the thermal strain test from room temperature to 1000℃, the nitrogen content declines (Fig 4a) and the oxygen content increases at different degrees (Fig 4b). This is because some metastable state nitrogen will escape from the grain boundary and the originally occupied oxygen vacancies at high temperatures [18]. On the other hand, more oxygen in the air will also diffuse into the ITO TFSGs at elevated temperatures [20]. Consequently, the unbalance between these two processes will affect the content of nitrogen and oxygen in the ITO TFSGs. Notably, the smallest content change rates of nitrogen (3.8%) and oxygen (1.6%) among the ITO TFSGs are observed in the 20%N2 ITO TFSG, suggesting this TFSG has the most stable chemical composition.
3.2 Strain testing
The piezoresistive responses of the ITO TFSGs at room temperature are shown in Fig.5a-e. The resistances of all the ITO TFSGs decrease with the growth of the applied strain and increase with the decline of the strain. All the ITO TFSGs exhibit excellent stability in the resistance response during three loading and unloading cycles, and almost no resistance drift (maximum drift rate is only 0.0059/h) is found. Herein, the GFs of the ITO TFSGs were calculated according to equation (1). The GFs of the ITO TFSGs decrease from 2.78 to 1.94 with the increase of the NPPs, as shown in Fig5f. This can be attributed to the larger energy difference between the Fermi energy level (EF) and the conduction band (Ec) with the increase of NPPs, causing fewer ionized electrons transfer to the conduction band in the ITO TFSGs under the same strain.
The furnace was then ramped from room temperature to 600℃. The piezoresistive responses of the ITO TFSGs at 600℃ are shown in Fig.6a-e, and the resistances can negatively change in pace with the applied strain. Compared with the testing results at room temperature, the different resistance drift appears in the ITO TFSGs at 600℃. As shown in Fig.6f, the DRs of the ITO TFSGs will initially decrease from 0.027/h to 0.0015/h and then grow from 0.0015/h to 0.02/h with the increase of the NPPs. This is because both inadequate and excessive nitrogen doping concentration will influence a dynamic equilibrium between the nitrogen escape from the ITO TFSGs and the oxygen diffusion into the TFSGs. Therefore, this will cause a destabilization of piezoresistive response. Meanwhile, it also suggests that the piezoresistivity of the 20%N2 ITO TFSG is the most stable one at 600℃. In addition, different ITO TFSGs have various GFs, as shown in Fig.6f. The GF declines from 2.79 to 1.48 with the enhancement of NPPs due to the decrease of carrier concentration, which depends on the decline of the amounts of electrons ionized to the conduction band since the shift of EF in the ITO TFSGs. However, the GF of 10%N2 ITO TFSG is smaller than that of the 20%N2 ITO TFSG. This is because the EF of these two TFSGs is similar. In that case, the Hall mobility might decrease when the temperature was raised to 600℃ due to the enhanced scattering effect induced by the unbalance between the nitrogen escape and the oxygen diffusion occurred in the 10% N2 ITO TFSG[25]. Thus, this will play a dominant role in the smaller GF of the 10% N2 ITO TFSG than that of the 20% N2 ITO TFSG.
The furnace was continually ramped from 600℃ to 800℃ and the piezoresistive responses of the ITO TFSGs at 800℃ are shown in Fig.7a-e. The resistances of all the ITO TFSGs changed synchronously with the applied strain, and the variation trend of resistance is consistent with the results discussed in Fig5 and 6. However, the piezoresistive stability decreases with the increase of temperature and the high resistance drift also appears in the ITO TFSGs. As shown in Fig.7f, all the DRs of the ITO TFSGs at 800℃ are larger than those at 600℃. Notably, the DR of the 20%N2 ITO TFSG is the lowest at 0.015/h compared with other TFSGs at 800℃, suggesting the TFSG prepared with 20% NPP still presents the most stable piezoresistivity. This further indicates that the nitrogen doping concentration should be moderate. Slight or redundant nitrogen content in the ITO TFSGs is not conducive to the stability of piezoresistive response at high temperatures. In addition, the ITO TFSGs also have a great difference in the GFs, as shown in Fig.7f. The GFs of the ITO TFSGs no longer monotonically changes with the increase of NPPs at 800 ℃. The GF of the 20%N2 ITO TFSG is the highest at 2.66 among all the TFSGs. Comparatively, the GFs of the 5%N2 and 10%N2 ITO TFSGs are lower than those of 30%N2 and 40%N2 ITO TFSGs. This might be attributed to the disequilibrium between the amount of nitrogen escape and the number of oxygen diffusion in the ITO TFSGs. The oxygen diffusion amount of the 5%N2 and 10%N2 ITO TFSGs is greater than their amount of nitrogen escape, while the unbalanced relationship will inversely exist in the 30% N2 and 40% N2 ITO TFSGs. Thus, this will result in the lower carrier concentration and the smaller resistance change in the 5%N2 and 10%N2 ITO TFSGs than those in the 30% N2 and 40% N2 ITO TFSGs. In conclusion, the carrier concentrations affected by the element variation become dominant in the piezoresistive response of the different ITO TFSGs and the band structures is no longer take effect at 800℃.
The strain test results of the ITO TFSGs at 900℃ are shown in Fig.8a-e. It can be seen that the piezoresistive stability of the ITO TFSGs at 900℃ are similar to that of the ITO TFSGs at 800℃. The DR of the 20%N2 ITO TFSG is also the lowest one (0.013/h), which is consistent with the results at 800℃. In addition, the GFs of the ITO TFSGs also bear resemblance to the results at 800℃. The 20% ITO TFSG has the highest GF of 2.42 among all the TFSGs at 900℃, as shown in Fig.8f. The reasons for the change trends of the DRs and GFs in the ITO TFSGs are the same as those have been discussed at 800℃, which suggests that the ITO TFSGs have the similar piezoresistive performance at 800℃ and 900℃.
The testing temperature was ultimately raised to 1000℃, and the resistance changes of the ITO TFSGs with the applied strain are shown in Fig.9a-e. The waveform of the piezoresistive response has severely degraded. The DRs of the ITO TFSGs at 1000℃ increase significantly compared with those of the TFSGs below 900℃ (Fig.9f). This is because the electrical properties were seriously influenced by the elemental composition at 1000℃. More oxygen in the air was activated to diffuse into the ITO TFSGs at 1000 ℃ and occupied oxygen vacancies, leading to the compensation of the doubly charged oxygen vacancies and the decline of the carrier concentration [20]. At the same time, some metastable nitrogen incorporated in the ITO TFSGs will diffuse out and some eventually trapped at triple junctions and grain boundaries [18]. In consequence, these two processes will result in the decline of nitrogen content (Fig 4a) and the increase of oxygen content (Fig 4b), exacerbating the chaos of the carrier concentration and Hall mobility of the TFSGs at 1000℃. This will lead to the severe destabilization of the piezoresistive response. However, the 20%N2 ITO TFSG can still maintain the relatively low DR of 0.015/h due to the smallest content change rates of nitrogen (3.8%) and oxygen (1.6%), which indicates the superiority of this TFSG in piezoresistive stability at high temperatures. Moreover, the GF of the 20%N2 ITO TFSG is also the highest one compared with other TFSGs because of its stable elemental composition.
This paper aims to investigate the effect of the NPPs on the piezoresistive response of the ITO TFSGs from room temperature to 1000℃. Different ITO TFSGs were prepared by RF magnetron sputtering with various N2/Ar ratios of 5%~40%. The XPS valence band spectra of the ITO TFSGs suggest that the EF shifts to the Ev with the growth of NPPs due to the compensation between the donor impurity level (Ed) and acceptor impurity level (Ea). The elemental analysis indicates that the 20%N2 ITO TFSG has the smallest content change rates of nitrogen (3.8%) and oxygen (1.6%) after the thermal strain test. Consequently, the 20%N2 ITO TFSG exhibits the lowest DR at high temperatures due to its stable elemental composition. In addition, the mechanism of the piezoresistive response in the different ITO TFSGs at high temperatures was revealed. Results show that the band structures play the main role in the GFs of the ITO TFSGs at room temperature and 600 ℃. The element variation is responsible for the different GFs of the ITO TFSGs at 800℃, 900℃ and 1000℃. The piezoresistive stability also depends on the elemental composition influenced by the dynamic equilibrium between the diffusion amount of oxygen and the escape amount of nitrogen in the ITO TFSGs at high temperatures.
ACKNOWLEDGEMENTS
This research is financially supported by the Dalian Science and Technology Innovation Fund (No.2019J12GX042), National Natural Science Foundation of China (Nos.51675085), Aviation Science Fund (No.2019ZD063001) and the Fundamental Research Funds for the Central Universities.
[1] Gregory O J, Luo Q, Crisman E E. High temperature stability of indium tin oxide thin films. Thin solid films, 2002,406(1):286-293.
[2] Pollak R T, Lambros J. Special issue on strain measurement in extreme environments. The Journal of Strain Analysis for Engineering Design, 2014,49(4):202-203.
[3] Liu Z, Liang J, Su S, et al. Preparation of defect-free alumina insulation film using layer-by-layer electrohydrodynamic jet deposition for high temperature applications. Ceramics International, 2021.
[4] Gregory O J, Chen X, Crisman E E. Strain and temperature effects in indium–tin-oxide sensors. Thin solid films, 2010,518(19):5622-5625.
[5] Liu H, Mao X, Cui J, et al. Effect of thickness on the electrical properties of PdCr strain sensitive thin film. Journal of Materials Science: Materials in Electronics, 2019,30(11):10475-10482.
[6] Liu H, Mao X, Yang Z, et al. High temperature static and dynamic strain response of PdCr thin film strain gauge prepared on Ni-based superalloy. Sensors and actuators. A. Physical., 2019,298:111571.
[7] Tougas I, Amani M, Gregory O. Metallic and Ceramic Thin Film Thermocouples for Gas Turbine Engines. Sensors, 2013,13(11):15324-15347.
[8] Chen X, Gregory O J, Amani M. Thin‐Film Thermocouples Based on the System In2O3–SnO2. Journal of the American Ceramic Society, 2011,94(3):854-860.
[9] Lin X, Zhang C, Yang S, et al. The impact of thermal annealing on the temperature dependent resistance behavior of Pt thin films sputtered on Si and Al2O3 substrates. Thin Solid Films, 2019,685:372-378.
[10] Han J, Cheng P, Wang H, et al. MEMS-based Pt film temperature sensor on an alumina substrate. Materials Letters, 2014,125:224-226.
[11] Zhang C, Huang J, Li J, et al. Design, fabrication and characterization of high temperature thin film heat flux sensors. Microelectronic Engineering, 2019,217:111128.
[12] Li X, Sun D, Cui Z, et al. The influence of spatial arrangement of endpoints on output characteristics of ITO/In2O3 heat flux gauge. Sensors and Actuators A: Physical, 2021,322:112587.
[13] Tian W, Wang Y, Zhou H, et al. Micromachined Thermopile Based High Heat Flux Sensor. Journal of Microelectromechanical Systems, 2020,29(1):36-42.
[14] Gregory O J, Luo Q, Bienkiewicz J M, et al. An apparent n to p transition in reactively sputtered indium–tin–oxide high temperature strain gages. Thin solid films, 2002,405(1):263-269.
[15] Gregory O J, Luo Q. A self-compensated ceramic strain gage for use at elevated temperatures. Sensors and actuators. A. Physical., 2001,88(3):234-240.
[16] Fang H, Miller T, Rogers B R, et al. Effect of oxygen content on piezoresistivity of indium tin oxide thin films prepared by pulsed laser deposition. Journal of applied physics, 2005,97(8):83502.
[17] Gregory O J, You T, Crisman E E. Effect of aluminum doping on the high-temperature stability and piezoresistive response of indium tin oxide strain sensors. Thin solid films, 2005,476(2):344-351.
[18] Gregory O J, You T. Piezoresistive properties of ITO strain sensors prepared with controlled nanoporosity[J]. Journal of the Electrochemical Society, 2004,151(8):H198-H203.
[19] Yang S, Zhang C, Chang X, et al. Effect of heat treatment atmosphere on the piezoresistivity of indium tin oxide ceramic strain sensor. Ceramics International, 2019,45(14):17048-17053.
[20] Yang S, Zhang C, Yang Z, et al. Effect of nitrogen doping temperature on the resistance stability of ITO thin films[J]. Journal of Alloys and Compounds, 2019,778:90-96.
[21] Skuridina D, Dinh D V, Lacroix B, et al. Polarity determination of polar and semipolar (112¯2) InN and GaN layers by valence band photoemission spectroscopy. Journal of Applied Physics, 2013,114(17):173503.
[22] Tian L, Cheng G, Wang H, et al. Effect of nitrogen doping on the structural, optical and electrical properties of indium tin oxide films prepared by magnetron sputtering for gallium nitride light emitting diodes. Superlattices and Microstructures, 2017,101:261-270.
[23] Chaker Fares, Ren F, Eric L, et al. Valence and conduction band offsets for sputtered AZO and ITO on (010) (Al0.14Ga0.86)2O3. Semicond. Sci. Technol. 2019, 34:025006
[24] Avelar-Muñoz F, Berumen J A, Aguilar-Frutis M A, et al. Enhancement on carrier mobility in amorphous indium tin oxynitride (ITON) thin films. Journal of Alloys and Compounds, 2020,835:155353.
[25] Klaassen D.B.M. A unified mobility model for device simulation-Ⅱ. Temperature dependence of carrier mobility and lifetime. Solid-State Electronics. 1992, 35:961-967