Structural properties and electrical characteristics of Ho₂O₃ and HoTi_xO_y gate dielectrics for a-InGaZnO thin-film transistors

Abstract

In this study, we developed high-κ Ho₂O₃ and HoTi_xO_y gate dielectrics for amorphous indium–gallium–zinc oxide (a-IGZO) thin-film transistor (TFT) applications. X-ray diffraction, atomic force microscopy, and X-ray photoelectron spectroscopy were used to study the structural, morphological and chemical features of Ho₂O₃ and HoTi_xO_y dielectric films. Compared with the Ho₂O₃ dielectric, the a-IGZO TFT incorporating the high-κ HoTi_xO_y gate dielectric exhibited very good electrical characteristics, such as a high I_on/off ratio of 1.1 × 10⁸, a high field effect mobility of 20.6 cm² V⁻¹ s⁻¹, a low threshold voltage of 0.23 V, and a low subthreshold swing of 183 mV decade⁻¹. These results are probably due to the incorporation of Ti into the Ho₂O₃ film, resulting in the formation of a smooth surface and a low density of interface states at the oxide/channel interface. In addition, the stability of high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFTs was investigated under positive gate-bias stress (PGBS) and negative gate-bias stress (NGBS). The electron charge trapping at the dielectric–channel interface resulted from the PGBS, whereas the oxygen vacancies occurred in the a-IGZO under the NGBS.

---

Introduction

Amorphous oxide semiconductors (AOSs) are actively researched for applications to transparent electronics, backplanes of large area active-matrix organic light-emitting diode displays, and liquid-crystal displays.^1–3 Among various AOS materials, amorphous indium–gallium–zinc oxide (a-IGZO) is considered one of the most promising channel materials for thin-film transistors (TFTs) because of its high field effect mobility (μ_FE), low leakage current, good uniformity and stability, high transparency to visible light, and low process temperature.^2–7 In order to obtain excellent device performance, a-IGZO materials with high carrier mobility are critical. Moreover, a high-capacitance gate dielectric is also desired to achieve a low operating voltage to reduce the device power consumption. Insulating silicon dioxide (SiO₂) has been traditionally used as the gate dielectric material in IGZO TFTs. However, SiO₂ with a moderate dielectric constant (3.9) is not an ideal material for high-capacitance gate dielectric. In addition, SiO₂ thin-film easily induces higher leakage current due to direct tunnelling through the film,⁸ which makes it difficult to obtain high-performance IGZO TFTs. Therefore, search for stable and high-capacitance gate dielectric to replace SiO₂ is currently one of the most important tasks and challenges in IGZO TFTs. High dielectric constant (high-κ) materials as a gate dielectric have been investigated to maintain the high capacitance density and inhibit the gate direct tunnelling current.^9–12 In addition to the electrical requirements, such as a high dielectric constant, low gate current, low interface state density, and good thermal stability, are also very important for high-κ materials. IGZO TFTs featuring high-κ materials such as Al₂O₃, Ta₂O₅, or HfO₂ were already reported in the literature.^13–15 Given the physical deposition processes, interface quality is expected to be considerably worse than the one achieved with the conventional SiO₂.¹² Most of high-κ dielectrics present a polycrystalline structure and have a rough surface, leading to reduced reliability and degraded interface properties.^8,16

To solve the problems related with high-κ dielectrics, thin rare-earth (RE) oxides have attracted interest for use as gate dielectrics for TFT applications because they have high dielectric constants and good thermal stability.^17,18 Holmium oxide (Ho₂O₃) film has been studied for gate dielectric applications due to its high κ value, wide bandgap energy, high breakdown field, and low leakage current.^19–22 Exposure to air, however, causes hygroscopic lanthanide oxides to react with moisture to form hydroxides,²³ causing lower values of κ. The incorporation of Ti or TiO₂ films into RE oxides increases their stability toward moisture.²⁴ van Dover²⁵ also demonstrated that the Ti adding RE oxide exhibited excellent electrical performance. Moreover, we reported that HoTiO₃ thin-film as a gate dielectric showed excellent electrical characteristics such as a high capacitance value, a low density of interface state, a small leakage current, and almost no hysteresis voltage.^26,27 Up to date, the physical and electrical properties of Ho₂O₃ and HoTi_xO_y films as a gate dielectric for a-IGZO TFT devices are still not known. In addition, the reliability and stability are also key concerns for practical TFT applications, and the gate bias stress test is a general method that is used to determine the electrical stability of such an a-IGZO TFT device.²⁸ In this paper, we compared the structural properties and electrical characteristics of high-κ Ho₂O₃ and HoTi_xO_y gate dielectrics for a-IGZO TFT device applications. We employed X-ray diffraction (XRD), atomic force microscopy (AFM) and X-ray spectroscopy (XPS) to analyze the structural, morphological and compositional changes, respectively, of the Ho₂O₃ and HoTi_xO_y films. Finally, the electrical and reliability characteristics of high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFTs were investigated.

Experimental

In this study, we fabricated a-IGZO TFT devices with inverted-staggered structures featuring high-κ Ho₂O₃ and HoTi_xO_y gate dielectrics as shown in Fig. 1. The a-IGZO TFT device was fabricated on thermally grown SiO₂ on Si substrate. A 40 nm TaN bottom gate was deposited on the SiO₂/Si substrate by reactive sputtering. A ∼50 nm Ho₂O₃ film was deposited on the bottom gate through reactive sputtering from a Ho target in an Ar/O₂ gas ratio of 20/5 at room temperature, while a ∼50 nm HoTi_xO_y film was deposited through co-sputtering from both Ho and Ti targets in the same condition. During the sputtering of dielectrics, the dc power of Ho and Ti was 100 and 60 W, respectively, and the chamber pressure was maintained at 10 mTorr. The deposition rate of dielectric films was ∼0.2 Å s⁻¹. Then, samples were subjected to furnace annealing in O₂ ambient for 10 min at 400 °C to form Ho₂O₃ and HoTi_xO_y structures. A 20 nm IGZO active layer was deposited from a-IGZO target (In₂O₃ : Ga₂O₃ : ZnO = 1 : 1 : 1 mol%) by reactive sputtering at room temperature. After the deposition of active channel layer, TFT devices were annealed at 200 °C in N₂ ambient for 10 min. Finally, a 50 nm Al film as the source/drain electrodes was deposited by a thermal evaporation system. The active region was defined at the channel width and length (W/L = 50 μm/5 μm) by lift-off process.

Fig. 1 Three-dimensional view of the high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices.

The crystalline structure and the chemical composition of the Ho₂O₃ and HoTi_xO_y films were investigated using XRD and XPS analyses, respectively. XRD analysis was performed using grazing-incidence Cu K_α (λ = 1.542 Å) radiation. The chemical bonding of the dielectric was determined using a monochromatic Al K_α (1486.7 eV) source. The surface morphology and roughness of the films were analyzed using an NT-MDT Solver P47 (AFM). The AFM was operated in the tapping mode for imaging. The root-mean-square (R_rms) roughness was measured from the AFM height images. The dielectric constant and gate leakage of the dielectric films were evaluated from the Al/Ho₂O₃/TaN and Al/HoTi_xO_y/TaN capacitors. The capacitance-voltage (C–V) curves of Ho₂O₃ and HoTi_xO_y capacitor devices were measured in the frequency of 1 MHz using a Hewlett-Packard (HP) 4285A LCR meter. The current-voltage (I–V) characteristics of the Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices were measured using a semiconductor parameter HP 4156C.

Results and discussion

We used XRD to examine the crystalline structures of the Ho₂O₃ and HoTi_xO_y dielectric films, as shown in Fig. 2. The crystalline structure of Ho₂O₃ is body-centered cubic, whereas HoTiO₃ film is orthorhombic. A strong HoTiO₃ (122) peak and a weak HoTiO₃ (004) peak are observed for the HoTi_xO_y sample. In contrast, for the Ho₂O₃ sample, one strong Ho₂O₃ (622) peak and two weak Ho₂O₃ (400) and (440) peaks were found in the 2θ diagram.

Fig. 2 XRD patterns of Ho₂O₃ and HoTi_xO_y dielectric films.

Fig. 3a and b illustrate the AFM images of the Ho₂O₃ and HoTi_xO_y dielectric films, respectively. The surface roughness of Ho₂O₃ film is about three times larger than that of HoTi_xO_y film. It has been reported that oxygen vacancies can exist in the RE oxide films with different charged states,²⁹ including neutral state, positively charged state, or negatively charged state. The negatively charged oxygen vacancy comprises one electron (V_O⁻) or two electrons (V_O²⁻). These electrons can react with the water by the following reaction equations:

Fig. 3 AFM surface images of Ho₂O₃ and HoTi_xO_y dielectric films. Typical parameters: scan size = 3 μm, scan rate = 1.001 Hz, data type = height, number of samples = 512, and data scale = 30 nm.

The large surface roughness may be due to the moisture absorption of Ho₂O₃ to form a Ho(OH)_x layer, causing the nonuniform volume expansion of the film.³⁰ The HoTi_xO_y film has a smooth surface (0.49 nm). The strong resistance to the moisture of HoTi_xO_y film is higher compared with Ho₂O₃ film. This behavior is attributed to the presence of less water reactive TiO_x in HoTi_xO_y.³¹

We used XPS to analyze the compositional changes in the Ho₂O₃ and HoTi_xO_y dielectric films. The atomic ratio of Ho : O is 41.1 : 58.9 in the Ho₂O₃ film, whereas Ho : Ti : O is 24.6 : 8.7 : 66.7 in the HoTi_xO_y film. The atomic ratio was determined from the peak intensity areas of XPS spectra. Fig. 4a shows the Ho 4d_5/2 XPS spectra of the Ho₂O₃ and HoTi_xO_y films. The Ho 4d_5/2 peaks of the reference Ho₂O₃ located at 161.3 eV.³² It is found that the Ho 4d_5/2 peak of the HoTi_xO_y sample is shifted higher binding energy by about 1 eV as compared to the Ho₂O₃ reference position. We attribute this behavior to the reactions of Ho with the O and Ti atoms to form a HoTi_xO_y structure. Fig. 4b displays the O 1s spectra of the Ho₂O₃ and HoTi_xO_y films with appropriate curve-fitting of peaks. Each fitting peak followed the general shape of the Lorentzian–Gaussian function. In the three sets of spectra, the O 1s peaks at 531.9, 530.1, and 529.3 eV represent the Ho–OH,³³ Ho–O–Ti, and Ho–O³⁴ bonds, respectively. Since hydrogen is more electronegative than Ho metal, the oxygen atoms in Ho(OH)_x species are less negatively charged than those in the oxide, resulting in a shift toward higher binding energies. The intensity of the O 1s peak corresponding to Ho(OH)_x for HoTi_xO_y sample was lower compared to Ho₂O₃ one. This result can be attributed to the fact that the added TiO_x during co-sputtering with Ho₂O₃ increases the oxygen potential to lessen the formation of oxygen vacancies, thus reducing the reaction of water with Ho₂O₃ film. Moreover, the O 1s peak corresponding to HoTi_xO_y had a larger intensity than that of Ho₂O₃. This result may be attributed to the addition of TiO_x into the Ho₂O₃ film reducing in the oxygen vacancies, and thus forming a HoTi_xO_y structure. Therefore, HoTi_xO_y film is a stronger resistance to moisture than Ho₂O₃ one.

Fig. 4 XPS spectra of (a) Ho 4d_5/2 and (b) O 1s for Ho₂O₃ and HoTi_xO_y dielectric films.

Fig. 5a shows the C–V curves of the Al/Ho₂O₃/TaN and Al/HoTi_xO_y/TaN capacitor devices. The Al/HoTi_xO_y/TaN device exhibited a higher capacitance density value of 228 nF cm⁻² than Al/Ho₂O₃/TaN one. Fig. 5b displays the effective dielectric constant (κ_eff) of the Ho₂O₃ and HoTi_xO_y capacitors as a function of frequency. The results show that the κ_eff value of Ho₂O₃ and HoTi_xO_y dielectric films keeps around 10 and 13, respectively, at frequency range from 100 kHz to 10 MHz. The HoTi_xO_y film has a higher κ_eff value compared with Ho₂O₃ film. This result is due to the large dielectric constant of TiO_x incorporating the Ho₂O₃ film.¹⁶ Fig. 5c illustrates the current density–voltage characteristics of the Al/Ho₂O₃/TaN and Al/HoTi_xO_y/TaN devices. It is generally believed that the leakage current of the dielectric film is associated with the surface roughness, crystal defect, and grain boundary.¹⁴ The Al/Ho₂O₃/TaN device exhibited a higher leakage current of 8.8 × 10⁻⁵ A cm⁻² at 5 V in comparison with the Al/HoTi_xO_y/TaN one (6.4 × 10⁻⁷ A cm⁻²). This result is attributed to the enhancement of leakage conduction path related to the surface roughness, defect states at the interface between channel and insulator layers, and oxygen vacancies in the film. In contrast, a lower leakage current in the HoTi_xO_y film is due to the less defective microstructure and the low surface roughness, as evidenced by AFM image.

Fig. 5 (a) Capacitance–voltage and (b) κ_eff–frequency characteristics of Al/Ho₂O₃/TaN and Al/HoTi_xO_y/TaN capacitor devices. (c) Current–voltage curves of Al/Ho₂O₃/TaN and Al/HoTi_xO_y/TaN capacitors.

Fig. 6a shows the transfer characteristics of the high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices. The threshold voltage (V_th) was defined at a fixed normalized drain current (10⁻⁹ × W/L). The V_th of the a-IGZO TFT devices featuring Ho₂O₃ and HoTi_xO_y gate dielectrics is 0.47 and 0.23 V, whereas the I_on/off ratio is 9.3 × 10⁵ and 1.1 × 10⁸, respectively. This high I_on/off ratio is due to the high capacitance density value and low gate leakage current. In contrast, the low I_on/off ratio may be attributed to the rough surface and the presence of oxygen vacancies in the Ho₂O₃ film. Furthermore, the μ_FE was determined by the maximum transconductance at constant drain voltage. The μ_FE of a-IGZO TFT devices incorporating Ho₂O₃ and HoTi_xO_y gate dielectrics is 15.5 and 20.6 cm² V⁻¹ s⁻¹, respectively. The high mobility characteristic may be attributed to the smooth surface between the dielectric film and IGZO channel, resulting in the low density of interface states at the dielectric–IGZO interface and small amount of bulk traps in the dielectric. The quality of interface at the dielectric–IGZO of the IGZO TFT can be indirectly evaluated using the subthreshold slope (SS) as is described by the following equation:³⁵

where k is the Boltzmann's constant, T is the absolute temperature, q is the elementary charge, D_it is the trap density at the dielectric/channel interface, C_B is the depletion capacitance per unit area, and C_ox is the capacitance per unit area of the gate oxide. If we want to turn on the TFT by applying low voltage, we must obtain a small SS value; this demonstrates that a little bias can greatly increase the output current. The small value of SS may be attributed to both the high gate capacitance density and the good interface charge density. The small SS of 183 mV decade⁻¹ obtained from the IGZO TFTs with a HoTi_xO_y dielectric can be ascribed to the reduced D_it at the interface of the IGZO–dielectric because of the smooth surface. On the other hand, the high SS of 520 mV decade⁻¹ is noticed in the IGZO TFT with a Ho₂O₃ dielectric. The moisture absorption of Ho₂O₃ film may be attributed to the oxygen vacancies in the film and thus increasing its surface roughness. The deposition of IGZO layer on the rough surface produces the high interface trap densities, such as vacancies and dangling bonds, at the IGZO–dielectric interface, which results in SS degradation. The output characteristics of the high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices are shown in Fig. 6b. The a-IGZO TFT device using a HoTi_xO_y gate dielectric has a larger driving current compared to Ho₂O₃ dielectric, suggesting the higher mobility and smaller threshold voltage.

Fig. 6 (a) Transfer (I_DS–V_GS) and (b) output (I_DS–V_DS) characteristics of high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices.

Only a few papers have discussed positive gate-bias stress (PGBS) and negative gate-bias stress (NGBS) conditions concurrently,³⁶ and the relationship of instability of the high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFTs under PGBS and NGBS conditions is not clear. Fig. 7 depicts the threshold voltage shift (ΔV_th) as a function of stress time for the high-κ Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices under PGBS and NGBS. The gate voltage stress was performed at V_GS = ±10 V for 1000 s. The positive shift in threshold voltage is due to the fact that accumulated electrons near the dielectric-channel interface were trapped by shallow acceptor-like trap states.³⁷ Under PGBS condition, the large V_th shift (1.31 V) of the a-IGZO TFT device using a Ho₂O₃ dielectric film indicates that more electrons are trapped near/at the dielectric and IGZO interface, while the low V_th shift (0.57 V) of TFT device using a HoTi_xO_y dielectric film shows suppressed the trapped charge in the film due to low interface states at the dielectric and channel interface. Moreover, the negative V_th shift may be attributed to the fact that the state creation is negligible and more interface states are filled during NGBS. Therefore, the negative V_th shift is related to extra free electrons from oxygen vacancies in the IGZO film.³⁸ The NGBS is more degradation in IGZO TFT than PGBS. To demonstrate the validity of stretched-exponential model to describe the PGBS and NGBS results, we followed a similar methodology that has been developed for PGBS and NGBS experiments.³⁶ The stretched-exponential model describes the ΔV_th by the following formula:³⁸

where ΔV_th0 is the threshold voltage shift for infinite time, τ is the constant characteristic trapping time, and β is the stretched-exponential exponent. The ΔV_th0 value of Ho₂O₃ TFT under PGBS and NGBS is 5.6 and 2.6 V, respectively. The τ value of a-IGZO TFT device under PGBS and NGBS is determined to be 1.1 × 10⁴ and 1 × 10⁴ s, whereas the β value is 0.48 and 0.49, respectively. In contrast, the ΔV_th0 value of HoTi_xO_y TFT under PGBS and NGBS is 4.9 and 2.2 V, respectively. The τ value of a-IGZO TFT under PGBS and NGBS is evaluated to be 1.2 × 10⁴ and 1.3 × 10⁴ s, while β value is 0.49 and 0.48, respectively. The results were in good agreement with previously reported data.³⁸

Fig. 7 (a) Threshold voltage shift as a function of stress time for Ho₂O₃ and HoTi_xO_y a-IGZO TFT devices.

Conclusions

In this paper, we have successfully demonstrated a high performance a-IGZO TFT device incorporating a high-κ HoTi_xO_y gate dielectric. The a-IGZO TFT device using a HoTi_xO_y gate dielectric exhibited better electrical characteristics, such as a low threshold voltage of 0.23 V, a high I_on/off ratio of 1.1 × 10⁸, large field effect mobility of 20.6 cm² V⁻¹ s⁻¹, and a small SS of 183 mV decade⁻¹, in comparison with that of Ho₂O₃ dielectric. We attribute this behaviour to the HoTi_xO_y film forming a smooth surface and reducing the density of interface states at the oxide/channel interface. Moreover, the V_th stability on Ho₂O₃ and HoTi_xO_y a-IGZO TFTs was studied under PGBS and NGBS. We found that NGBS is more degradation in a-IGZO TFT than PGBS. The electron charge trapping in the gate dielectric arises from the PGBS, while the oxygen vacancies generated under the NGBS can make a balance of electron trapping for PGBS.

Acknowledgements

This work was supported by the National Science Council, Taiwan, Republic of China, under contract no. NSC-102-2221-E-182-072-MY3.

References

1. H. Yabuta, M. Sano, K. Abe, T. Aiba, T. Den, H. Kumomi, K. Nomura and H. Hosono, Appl. Phys. Lett., 2006, 89, 112123.
2. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater, 2012, 24, 2945.
3. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano and H. Hosono, Nature, 2004, 432, 488.
4. J. Jeong, G. J. Lee, J. Kim and B. Choi, J. Phys. D: Appl. Phys., 2012, 45, 135103.
5. T. Kamiya, K. Nomura and H. Hosono, Sci. Technol. Adv. Mater., 2010, 11, 044305.
6. G. H. Kim, B. D. Ahn, H. S. Shin, W. H. Jeong, H. J. Kim and H. J. Kim, Appl. Phys. Lett., 2009, 94, 233501.
7. H. Hosono, J. Non-Cryst. Solids, 2006, 352, 851.
8. G. D. Wilk, R. M. Wallace and J. M. Anthony, J. Appl. Phys., 2001, 89, 5243.
9. J. S. Lee, S. Chang, S. M. Koo and S. Y. Lee, IEEE Electron Device Lett., 2010, 31, 225.
10. L. Yuan, X. Zou, G. Fang, J. Wan, H. Zhou and X. Zhao, IEEE Electron Device Lett., 2011, 32, 42.
11. N. C. Su, S. J. Wang and A. Chin, IEEE Electron Device Lett., 2009, 30, 1317.
12. P. Barquinha, L. Pereira, G. Goncalves, R. Martins, E. Fortunato, D. Kuscer, M. Kosec, A. Vila, A. Olziersky and J. R. Morante, J. Soc. Inf. Disp., 2010, 18, 762.
13. J. I. Song, J. S. Park, H. Kim, Y. W. Heo, J. H. Lee, J. J. Kim, G. M. Kim and B. D. Choi, Appl. Phys. Lett., 2007, 90, 022106.
14. C. J. Chiu, S. P. Chang and S. J. Chang, IEEE Electron Device Lett., 2010, 31, 1245.
15. W. Lim, S. H. Kim, Y. L. Wang, J. W. Lee, D. P. Norton, S. J. Pearton, F. Ren and I. I. Kravchenko, J. Electrochem. Soc., 2008, 155, H383.
16. High-k Gate Dielectrics, ed. M. Houssa, Institute of Physics Publishing, Bristol and Philadelphia, 2004.
17. C. W. Chang, C. K. Deng, J. J. Huang, H. R. Chang and T. F. Lei, IEEE Electron Device Lett., 2008, 29, 96.
18. F. H. Chen, J. L. Her, S. Mondal, M. N. Hung and T. M. Pan, Appl. Phys. Lett., 2013, 102, 193515.
19. O. Engstrom, B. Raeissi, S. Hall, O. Buiu, M. Lemme, H. D. B. Gottlob, P. K. Hurley and K. Cherkaoui, Solid-State Electron., 2007, 51, 622.
20. J. Paivasaari, M. Putkonen and L. Niinisto, Thin Solid Films, 2005, 472, 275.
21. T. M. Pan, W. T. Chang and F. C. Chiu, Thin Solid Films, 2010, 519, 923.
22. Rare Earth Oxide Thin Film: Growth, Characterization, and Applications, ed. M. Fanciulli and G. Scarel, Springer, Berlin, 2007.
23. S. Jeon and H. Hwang, J. Appl. Phys., 2003, 93, 6393.
24. T. Schroeder, G. Lupina, J. Dabrowski, A. Mane, C. Wenger, G. Lippert and H. J. Mussig, Appl. Phys. Lett., 2005, 87, 022902.
25. R. B. van Dover, Appl. Phys. Lett., 1999, 74, 3041.
26. T. M. Pan, L. C. Yen and S. H. Su, Appl. Surf. Sci., 2009, 256, 1534.
27. T. M. Pan, C. H. Chen, J. H. Liu, J. L. Her and K. Koyama, IEEE Electron Device Lett., 2014, 35, 66.
28. R. N. P. Vemuri, W. P. Mathews, M. Marrs and T. L. Alford, J. Phys. D: Appl. Phys., 2013, 46, 045101.
29. Y. Zhao, K. Kita, K. Kyuno and A. Toriumi, Jpn. J. Appl. Phys., 2007, 46, 4189.
30. Y. Zhao, M. Toyama, K. Kita, K. Kyuno and A. Toriumi, Appl. Phys. Lett., 2006, 88, 072904.
31. S. Jeon and H. Hwang, Appl. Phys. Lett., 2002, 81, 4856.
32. J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Physical Electronics, Chanhassen, MN, 1995.
33. B. S. Jeong, Y. G. Ha, J. Moon, A. Facchetti and T. J. Marks, Adv. Mater., 2010, 22, 1346.
34. Y. A. Teterin and A. Y. Teterin, Russ. Chem. Rev., 2002, 71, 347.
35. S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, John Wiley & Sons, 3rd edn, 2007.
36. A. Suresh and J. F. Muth, Appl. Phys. Lett., 2008, 92, 033502.
37. W. T. Chen, S. Y. Lo, S. C. Kao, H. W. Zan, C. C. Tsai, J. H. Lin, C. H. Fang and C. C. Lee, IEEE Electron Device Lett., 2011, 32, 1552.
38. J. M. Lee, I. T. Cho, J. H. Lee and H. I. Kwon, Appl. Phys. Lett., 2008, 93, 093504.

---