Effects of annealing ambient on oxygen vacancies and phase transition temperature of VO₂ thin films

Abstract

VO₂ thin films are prepared on Si substrates by direct-current (DC) magnetron sputtering at room temperature and annealed in vacuum at different argon pressures. The VO₂ thin films annealed in vacuum and in Ar are all polycrystalline with a monoclinic structure; after annealing in Ar, the particle size is reduced compared to the films annealed in vacuum. There are more boundaries for smaller particles; therefore, oxygen can easily diffuse through the boundaries, resulting in more oxygen vacancies. Annealing under Ar further prevents the samples from oxidation. As the Ar pressure increases, the V 2p_3/2 peak broadens and shifts to a lower binding energy, implying that there are more oxygen vacancies after annealing. The features in the Raman spectra acquired at room temperature shift to lower frequencies after annealing in Ar, further corroborating the existence of oxygen vacancies in the thin films. Raman scattering and resistance measurements show that the critical temperature of the phase transition from monoclinic to tetragonal is reduced from 341 K to 319 K. This can be ascribed to the weaker hybridization between V 3d and O 2p orbitals as a result of oxygen vacancies. Oxygen vacancies affect the phase transition in VO₂ thin films, and the optical and electrical properties as well.

---

1. Introduction

Vanadium oxide has various structures due to the multivalent characteristics of vanadium such as V²⁺, V³⁺, V⁴⁺, and V⁵⁺ and the metal-insulator transition (MIT) can take place at different temperatures depending on the structure. In particular, VO₂ has attracted considerable interest because of the low MIT transition temperature (T_MIT) of 68 °C.¹ Below the critical temperature, VO₂ has a monoclinic phase (M₁, P2₁/c) behaving as an insulator, but it transforms to the metallic phase with a tetragonal rutile lattice (R, P4₂/mnm) at a temperature higher than the critical value. Since the structural transition is commonly accompanied by abrupt changes in the resistivity, magnetic susceptibility, and optical transmittance,^2–5 VO₂ has potential applications in sensors,^6,7 optical switches,^8,9 and smart windows.^10–12 The phase transition can also be induced by light, electric field, magnetic field, temperature, and stress. Many methods, such as atomic layer deposition (ALD),¹³ pulsed laser deposition (PLD),¹⁴ chemical vapor deposition (CVD),¹⁵ and sol–gel,¹⁶ have been employed to prepare VO₂ thin films and magnetron sputtering is particularly attractive because of the low deposition temperature, high film adhesion, and smoothness.

In spite of these attractive properties, the use of VO₂ thin films in smart windows is still limited since T_MIT is much higher than room temperature. Much research has been conducted to reduce the phase transition temperature by introducing stress/strain,^17,18 changing the chemical stoichiometry,^19,20 and conducting metal doping.²¹ Doping is generally considered to be the most effective method; however, it tends to make the process complex and may also introduce stress. Recently, it has been reported that the oxygen content can affect the physical properties of VO₂ thin films.^22,23 Fan et al.²² attempted to tune the oxygen vacancy density by changing the oxygen flux rate in the molecular beam epitaxy (MBE) process and Jeong et al.²³ produced oxygen vacancies in VO₂ thin films by applying an electric field. The transition temperature decreases with increasing oxygen vacancies. In previous reports, annealing in a protected atmosphere was shown to produce a less crystalline structure than that annealed in vacuum.²⁴ In this study, vanadium oxide thin films are prepared by direct current (DC) magnetron sputtering at room temperature and then annealed in Ar. Compared to the films annealed in vacuum, the films annealed in Ar have a smaller grain size and more grain boundaries as Ar plays a key role in the formation of oxygen vacancies. The vacancy density increases with Ar pressure and as a result, the transition temperature is lowered substantially. The mechanism is discussed.

2. Experimental details

Vanadium oxide thin films were deposited on Si (100) substrates using direct-current (DC) reactive magnetron sputtering of a V target (99.9% in purity) in a mixed ambient consisting of high-purity oxygen and argon (Ar). Prior to deposition, the substrates were ultrasonically rinsed in acetone and alcohol for 15 min and dried with nitrogen. After the vacuum chamber was evacuated to 1.2 × 10⁻⁴ Pa, the substrates were cleaned by Ar sputtering at 0.5 Pa for 5 min with a power of 300 W. Film deposition was then conducted at room temperature for 20 min using a DC power of 220 W at 0.63 Pa. The partial pressures of O₂ and Ar were controlled by mass flow controllers. After deposition, the films were annealed at 450 °C for 2 hours in vacuum as well as in an Ar atmosphere with a pressure of 44, 72, and 100 Pa through changing the gas flow from 90 to 190 sccm. The base pressure before thermal annealing is 5 Pa. The important experimental parameters are listed in Table 1.

Table 1 Samples annealed at different Ar pressures

Sample	1	2	3	4
O2/Ar sputtering	32.5%	32.5%	32.5%	32.5%
Ar pressure (Pa) (annealing)	0	44	72	100

---

The phases of the thin films were determined by grazing-incidence X-ray diffraction (GIXRD, Shimadzu XRD-7000) with Cu K_α radiation (λ = 0.154 nm) at a fixed incident angle of 0.5° and a 2θ scanning rate of 8° min⁻¹. X-ray photoelectron spectroscopy (XPS, Physical Electronics PHI 5802) with Al K_α irradiation was performed to determine the chemical states and composition. The surface morphology was examined by field-emission scanning electron microscopy (FESEM, JSM-7000F) and atomic force microscopy (AFM, Bruker dimension Icon) employing the tapping mode. Raman scattering was performed on a micro-Raman system (LabRAM HR Evolution, France). A 50× objective lens was used to focus the HeNe laser beam with a wavelength of 514.5 nm onto the VO₂ films and the scattered light was monitored using a 1800 g mm⁻¹ grating. To avoid unintentional heating during Raman analysis, the incident laser power was small but sufficient to produce an optimal signal-to-noise ratio. The samples were heated from 303 to 343 K at a temperature step of 5 K and the Physical Property Measurement system (PPMS, Quantum Design PPMS-9) was used to determine the electrical resistivity of the films when the temperature was increased from 293 to 363 K and subsequently reduced back to 293 K with 2 K steps. The sample was fixed and connected to the electrode separately by a double-side tape and a conductive silver adhesive. The film thickness was measured by SEM and used to calculate the resistivity of the thin films.

3. Result and discussion

The vanadium oxide thin films prepared by DC-reactive magnetron sputtering are homogeneous with a relatively smooth surface. Fig. 1 shows the SEM images revealing the morphological differences between the as-deposited (Fig. 1(a)) and annealed thin films (Fig. 1(b) and (c)). It is obvious that annealing leads to an appreciable change in the surface morphology. Fig. 1(a) reveals that the as-deposited thin films are continuous and smooth, whereas Fig. 1(b) shows that the films annealed in vacuum have a porous and fuzzy morphology. As shown in the cross-sectional image in the inset of Fig. 1(b), the films annealed in vacuum consists of columnar particles with an average width of about 25 nm and length of 80 nm. The films annealed in Ar are composed of relatively homogenous and continuous spheroidal nanoparticles with an average diameter of ∼14 nm, as shown in Fig. 1(c). This is consistent with previously reported results.^24,25 In fact, thermal annealing is accompanied by the reaction of . Therefore, the nucleation of crystalline VO₂ is more difficult in the oxygen-rich atmosphere. As compared to thermal annealing in vacuum, the oxygen content in the Ar atmosphere is reduced, and consequently the nucleation of VO₂ is promoted. It becomes more significant under higher Ar pressure and leads to smaller grain sizes. No noticeable difference in the surface roughness is observed from the thin films annealed at different Ar pressures. There are more boundaries when the particles are smaller and hence, it is easier for oxygen to diffuse through the boundaries to increase the oxygen vacancies. Moreover, annealing in Ar prevents the films from oxidizing.

Fig. 1 SEM images of VO₂ thin films: (a) as-deposited and (b)–(c) annealed in vacuum and Ar, respectively. The inset in (b) is the corresponding cross-sectional image showing columnar particles with an average width of 25 nm and length of 80 nm. (c) The film has relatively homogenous and continuous spheroidal nanoparticles with an average diameter of ∼14 nm.

The as-deposited thin films are amorphous, and no XRD peak of vanadium oxide could be identified. Furthermore, only a peak at 517.2 eV for V⁵⁺ is observed from the XPS spectra. Therefore, the as-deposited thin films are indeed amorphous V₂O₅. After thermal annealing, some oxygen atoms escape from the thin film with only one peak at 515.8 eV for V⁴⁺ being observed in the XPS spectra, which is due to the thin film becoming crystallized VO₂ in the M1 phase, as illustrated in the XRD patterns in Fig. 2. Samples 1–4 were prepared at room temperature for 20 min using a DC power of 220 W and annealed at different Ar pressures. The O₂/Ar ratio for the fabrication of samples 1–4 is 32.5%, as shown in Table 1. Sample 1, which was annealed in vacuum, is designated as 1-V, and samples 2–4, annealed in Ar at pressures of 44, 72, and 100 Pa, are labeled 2-44, 3-72 and 4-100, respectively. Fig. 2(a) shows that the diffraction peaks at 27.86°, 37°, and 55.54° correspond to the (011), (21−1) and (220) planes of VO₂. The four samples are polycrystalline having a monoclinic structure and (011) preferential orientation. The full-width at half-maximum (FWHM) of the (011) peaks in samples 1-V and 2-44 are 0.574° and 0.732°, respectively. According to Scherrer's formula, the grain size of the VO₂ in the thin films annealed in Ar is smaller than that annealed in vacuum, which is consistent with the SEM images. However, the grain size slightly changes after annealing at different Ar pressures. Fig. 2(b) shows the patterns with the VO₂ (011) peak for samples 1-V and 2-44 marked by a light blue square in Fig. 2(a). The red lines are the fitted curves of the experimental data. Compared to sample 1-V, the peak positions for sample 2-44 annealed in Ar shift towards the small-angle direction. The smaller intensity and larger FWHM of the VO₂ (011) peak suggests less crystallization in sample 2-44. Stress in thin films can be categorized into two types: extrinsic stress and intrinsic stress. Thermal stress is an extrinsic one. Because the coefficient of thermal expansion of VO₂ [4.9 × 10⁻⁶ (a_‖) or 2.6 × 10⁻⁵ K⁻¹ (a_⊥)] is larger than that of Si (3.1 × 10⁻⁶ K⁻¹),²⁶ tensile thermal stress will be induced when the annealing temperature is lowered down to room temperature. However, compressive stress is closely related to the smaller grain size as well as the densification of thin films.²⁷ Since the particle size in the thin films annealed in Ar is considerably smaller than that annealed in vacuum, compressive stress is produced in the in-plane direction, resulting in a decreased in-plane lattice constant and increased out-of-plane lattice constant.²⁸ Therefore, the a-axis on the preferred (011) plane is compressed. The a-axis of the monoclinic structure (a_M) corresponds to the c-axis of the rutile structure (c_R). In such a case, the phase transition temperature should be lowered to some degree.²⁹

Fig. 2 (a) X-ray diffraction patterns of the vanadium oxide thin films 1 to 4 deposited with the ratio of O₂/Ar of 32.5% followed by annealing at different Ar pressure of 0, 44, 72, and 100 Pa, respectively. Sample 1–4 are designated as 1-V, 2-44, 3-72, and 4-100, respectively. (b) Fine pattern of (011) peak for sample 1-V and 2-44. The red lines are the fitted curves of the experiment data.

The concentration of residual O₂ in the annealing atmosphere decreases with increasing Ar pressure, and the formation of vanadium oxide in higher valence states will be suppressed. This promotes the transformation from V₂O₅ to VO₂ during thermal annealing. In particular, oxygen vacancies are produced in the thin films in order to satisfy the condition of electrical neutrality.³⁰ Actually, oxygen vacancies are produced through the overflowing of oxygen atoms from the thin films to the environment, as . According to Le Chatelier's principle, the reaction becomes favorable, resulting in the formation of oxygen vacancies, if the thermal annealing is carried out in an Ar atmosphere. Thus, as the Ar pressure is increased in the thermal annealing process, the oxygen vacancies increase.^31,32 The oxygen vacancies²³ in the VO₂ thin films annealed in Ar are confirmed by XPS. Fig. 3 shows the XPS spectra of samples 1-V, 2-44 and 4-100. The core level binding energy of V 2p_3/2 is usually used to characterize the oxidation state of vanadium and the peak position is fitted by a Gaussian function with the provided XPS software.³³ Compared to that of the VO₂ thin film annealed in vacuum, the V⁴⁺ peak shifts to lower energy. When the Ar pressure in the annealing process is further increased, a weak peak at 513.1 eV due to V³⁺ is observed in addition to the strong peak at 515.8 eV for V⁴⁺.^34–37 Both the slight shift of V⁴⁺ peak and appearance of V³⁺ confirm the existence of oxygen vacancies in the thin films annealed in Ar.^38,39 The O : V ratio of the VO₂ thin films annealed in Ar at a pressure of 72 Pa is calculated to be 1.92 and it is 2 for the VO₂ thin films annealed in vacuum. Accordingly, oxygen vacancies exist in the thin films annealed in Ar. However, no V₂O₃ phase is observed from the XRD pattern. When the oxygen partial pressure in the sputtering process is further increased (not show here) and annealed in Ar, the XPS spectrum is nearly the same as that of sample 1-V, confirming the existence of oxygen vacancies in samples 2-44 to 4-100.

Fig. 3 Comparison between the high-resolution V 2p XPS spectra of thin films 1-V, 2-44, and 4-100. The V⁴⁺ peak shifts to lower energy as the Ar pressure is raised.

The oxygen vacancies in the VO₂ thin films annealed in Ar are further assessed by Raman scattering. Fig. 4(a) shows the room-temperature Raman spectra of samples 1-V and 4-100. At 303 K, there are 7 Raman peaks at 195 (A_g), 224 (A_g), 261 (B_g), 310 (A_g), 391 (A_g), 439 (B_g), and 616 cm⁻¹ (A_g), characteristic of the M1 phase of VO₂; the peak at 520 cm⁻¹ originates from the Si substrate. The three prominent phonon modes at frequencies of 195, 224, and 616 cm⁻¹ are denoted ω_v1, ω_v2, and ω₀. ω_v1 and ω_v2 are attributed to the V–V vibration, and the high-frequency mode, ω₀, is due to the V–O vibration. The peaks of sample 4-100 marked by the black squares in Fig. 4(a) shift to lower frequencies compared to sample 1-V, as shown in Fig. 4(b). Parker et al.⁴⁰ found that the Raman peaks shifted to lower frequencies in V-rich thin films. Fig. 4(c) presents the Raman spectra of sample 4-100 in the temperature range from 303 to 338 K. As the annealing temperature increases, the intensity of all the Raman peaks decrease gradually. When the temperature is higher than 323 K, the M1 phase is no longer detected and the VO₂ thin films transform into the tetragonal phase, with no Raman peak accompanying the semiconductor-to-metal transition. Fig. 5 shows the electrical resistivity of the samples as a function of temperature and the corresponding derivatives of resistivity with respect to temperature. The black and red curves correspond to the heating and cooling cycles, respectively. The thin films are heated in vacuum from room temperature to 363 K and then cooled to room temperature in steps of 2 K. T_MIT is defined as the minimum value of dR/dT and the hysteresis width (ΔH) is defined as the difference in T_MIT in the heating and cooling processes. All the samples exhibit a phase transition from a high resistance state to a metallic state and thermal hysteresis. Fig. 6 shows the T_MIT and hysteresis width obtained from the resistivity curve as a function of temperature. The resistance of sample 1-V changes by more than three orders of magnitude in the process, and T_MIT is 341 K with a smaller hysteresis width of 4 K, which is consistent with previously reported results.^1,41 However, when the thin films are annealed in Ar, both T_MIT and the change in resistance decrease. As reported before, the grain size plays a role in broadening the hysteresis width and reducing T_MIT.⁴² The smaller grain size commonly leads to a larger hysteresis width due to the small density of nucleating defects and the large interfacial energies.⁴³ However, the samples have similar grain sizes when they are annealed at different Ar pressures. Hence, the grain size is not the dominant factor in determining T_MIT. Goodenough et al.⁴⁴ pointed out that oxygen vacancies in non-stoichiometric thin films may reduce the transition temperature due to extra free electrons.^45,46 As shown in Fig. 5, the films annealed in Ar (samples 2-44 to 4-100) are more conductive than those annealed in vacuum at room temperature, indicating electron doping. The chemical states and hybridization of V 3d and O 2p electrons may reveal information about the influence of oxygen vacancies on the phase transition in VO₂. Below the MIT transition temperature, the V and O atoms in the monoclinic phase are strongly bonded with each other via hybridization between the V 3d and O 2p orbitals and the films are insulating. After annealing in Ar, oxygen vacancies are generated, weakening the hybridization between the V 3d and O 2p, and the partial loss of V⁴⁺–V⁴⁺ bonding lowers the temperature for the metal-insulator transition. It was also found that oxygen vacancies could stabilize the R phase even at room temperature, thus reduce the activation energy and phase transition temperature. Hence, bond breaking and reforming become easier and the phase transition temperature is reduced substantially.

Fig. 4 (a) Raman spectra of sample 1-V and 4-100 obtained at room temperature. (b) Fine pattern of ω_v1 peaks of sample 1-V and 4-100. (c) Raman spectra of sample 4-100 as a function of temperature between 303 and 323 K.

Fig. 5 (a) and (c) Resistivity of the VO₂ thin films as a function of temperature for samples 1-V and 4-100. The differential curves (b) and (d) are used to determine the phase transition characteristics including the T_MIT and hysteresis width.

Fig. 6 Comparison of the phase transition properties derived from the differential of the resistivity–temperature curves: (a) T_MIT and (b) hysteresis width.

4. Conclusion

In summary, polycrystalline VO₂ thin films with a monoclinic structure are prepared on Si substrates by DC magnetron sputtering. The oxygen content can be reduced by annealing in Ar and so more oxygen vacancies are produced. Raman scattering and electrical resistance measurements in the temperature range between 293 K and 363 K demonstrate that the T_MIT of the VO₂ thin films decreases from 335 K to 319 K with increasing oxygen vacancies. Our study reveals that annealing in Ar is a simple and effective approach to produce oxygen vacancies in VO₂ thin films and lower the phase transition temperature.

Acknowledgements

This study was jointly supported by the National Natural Science Foundation of China (Grant No. 51271139, 51471130, 51302162), the Fundamental Research Funds for the Central Universities, and the City University of Hong Kong Applied Research Grant (ARG) Nos. 9667104 and 9667122.

References

1. F. J. Morin, Phys. Rev. Lett., 1959, 3, 34.
2. A. Cavalleri, T. Dekorsy, H. H. W. Chong, J. C. Kieffer and R. W. Schoenlein, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70, 161102.
3. W. Li, J. J. Zhu, J. R. Liang, Z. G. Hu, J. Liu, H. D. Chen and J. H. Chu, J. Phys. Chem. C, 2011, 15, 23558–23563.
4. W. T. Liu, J. Cao, W. Fan, Z. Hao, M. C. Martin, Y. R. Shen, J. Wu and F. Wang, Nano Lett., 2011, 11, 466–470.
5. R. Lopez, T. Haynes, L. Boatner, L. Feldman and R. Haglund, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 65(224113), 1–5.
6. J. Wu, Q. Gu, B. S. Guiton, N. P. de Leon, L. Ouyang and H. Park, Nano Lett., 2006, 6, 2313–2317.
7. Z. Li, H. Zhang, W. Zheng, W. Wang, H. Huang, C. Wang, A. G. MacDiarmid and Y. Wei, J. Am. Chem. Soc., 2008, 130, 5036–5037.
8. W. X. Huang, X. G. Yin, C. P. Huang, Q. J. Wang, T. F. Miao and Y. Y. Zhu, Appl. Phys. Lett., 2010, 96, 261908.
9. C. Chen and Z. Zhou, Appl. Phys. Lett., 2007, 91, 011107.
10. C. G. Granqvist, B. Pehlivan, Y. X. Ji, S. Y. Li and G. Niklasson, Thin Solid Films, 2014, 559, 2–8.
11. C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2007, 91, 1529–1598.
12. S. Y. Li, G. A. Niklasson and C. G. Granqvist, Thin Solid Films, 2012, 520, 3823–3828.
13. K. Martens, I. P. Radu, S. Mertens, X. Shi, L. Nyns, S. Cosemans, P. Favia, H. Bender, T. Conard, M. Schaekers, S. De Gendt, V. Afanas'ev, J. A. Kittl, M. Heyns and M. Jurczak, J. Appl. Phys., 2012, 112, 124501.
14. J. Sakai, M. Zaghrioui, V. T. Phuoc, S. Roger, C. Autret-Lambert and K. Okimura, J. Appl. Phys., 2013, 113, 123503.
15. D. Vernardou, D. Louloudakis, E. Spanakis, N. Katsarakis and E. Koudoumas, Sol. Energy Mater. Sol. Cells, 2014, 128, 36–40.
16. J. Wu, W. X. Huang, Q. W. Shi, J. H. Cai, D. Zhao, Y. B. Zhang and J. Z. Yan, Appl. Surf. Sci., 2013, 268, 556–560.
17. J. H. Park, J. Coy, T. Kasirga, C. Huang, Z. Fei, S. Hunter and D. Cobden, Nature, 2013, 500, 431–434.
18. J. Cao, E. Ertekin, V. Srinivasan, W. Fan, S. Huang, H. Zheng, J. W. L. Yim, D. R. Khanal, D. F. Ogletree, J. C. Grossman and J. Wu, Nat. Nanotechnol., 2009, 4, 732–737.
19. S. X. Zhang, I. S. Kim and L. J. Lauhon, Nano Lett., 2011, 11, 1443–1447.
20. J. Jeong, N. Aetukuri, T. Graf, T. D. Schladt, M. G. Samant and S. S. P. Parkin, Science, 2013, 339, 1402–1405.
21. N. R. Mlyuka, G. A. Niklasson and C. G. Granqvist, Appl. Phys. Lett., 2009, 95(171909), 1–3.
22. L. L. Fan, S. Chen, Y. F. Wu, F. H. Chen, W. S. Chu, X. Chen, C. W. Zou and Z. Y. Wu, Appl. Phys. Lett., 2013, 103, 131914.
23. J. Jeong, N. Aetukuri, T. Graf, T. D. Schladt, M. G. Samant and S. S. P. Parkin, Science, 2013, 339, 1402–1405.
24. D. W. Liu, Y. Y. Liu, B. B. Garcia, Q. F. Zhang, A. Pan, Y. Jeong and G. Z. Cao, J. Mater. Chem., 2009, 19, 8789–8795.
25. D. Zhang, L. B. Su, H. J. Li, X. B. Qian and J. Xu, J. Cryst. Growth, 2006, 294, 437.
26. K. Okimura, T. Watanabe and J. Sakai, J. Appl. Phys., 2012, 111, 073514.
27. N. H. Azhan, K. Su, K. Okimura and J. Sakai, J. Appl. Phys., 2015, 117, 185307.
28. D. P. Zhang, M. D. Zhu, Y. Liu, K. Yang, G. X. Liang, Z. H. H. Zheng and X. M. Cai, J. Alloys Compd., 2016, 659, 198–202.
29. L. L. Fan, S. Chen, Z. L. Luo, Q. H. Liu, Y. F. Wu, L. Song, D. X. Ji, P. Wang, W. S. Chu, C. Gao, C. W. Zou and Z. Y. Wu, Nano Lett., 2014, 14, 4036–4043.
30. X. X. Liu, S. H. W. Wang, F. L. Chen, L. M. Yu and X. S. H. Chen, J. Phys. D: Appl. Phys., 2015, 48, 265104.
31. B. Lucas-Granados, R. Sanchez-Tovar, R. M. Fernandez-Domene and J. Garcia-Anton, Sol. Energy Mater. Sol. Cells, 2016, 153, 68–77.
32. B. Sarma, A. L. Jurovitzki, R. S. Ray, Y. R. Smith, S. K. Mohanty and M. Misra, Nanotechnology, 2015, 26, 265401.
33. G. Silversmit, D. Depla, H. Poelman, G. B. Marin and R. D. Gryse, J. Electron Spectrosc. Relat. Phenom., 2004, 135, 167–175.
34. Y. Q. Lv, M. Hu, M. Wu and Z. H. Liu, Surf. Coat. Technol., 2007, 201, 4969–4972.
35. N. Alov, D. Kutsko, I. Spirovova and Z. Bastl, Surf. Sci., 2006, 600, 1628–1631.
36. D. H. Youn, H. T. Kim, B. G. Chae, Y. J. Hwang, J. W. Lee, S. L. Maeng and K. Y. Kang, J. Vac. Sci. Technol., A, 2004, 22, 719–724.
37. G. A. Sawatzky and D. Post, Phys. Rev. B: Condens. Matter Mater. Phys., 1979, 20, 1546.
38. Z. T. Zhang, Y. G. Gao, Z. Chen, J. Du, C. Cao, L. Kang and H. J. Luo, Langmuir, 2010, 26, 10738–10744.
39. R. V. Vovk, M. A. Obolenskii, A. A. Zavgorodniy, A. V. Bondarenko, I. L. Goulatis, A. V. Samoilov and A. Chroneos, J. Alloys Compd., 2008, 453, 69–74.
40. J. C. Parker, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 42, 3164.
41. Z. F. Luo, X. Zhou, D. W. Yan, D. Wang, Z. Y. Li, C. B. Yang and Y. D. Jiang, Thin Solid Films, 2014, 550, 227–232.
42. M. J. Miller and J. Wang, J. Appl. Phys., 2015, 117, 034307.
43. H. F. Zhang, Z. H. M. Wu, Q. He and Y. D. Jiang, Appl. Surf. Sci., 2013, 277, 218–222.
44. J. B. Goodenough, J. Solid State Chem., 1971, 3, 490–500.
45. D. Ruzmetov, S. D. Senanayake, V. Narayanamurti and S. Ramanathan, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 195442.
46. C. C. Y. Kwan, C. Griffith and H. K. Eastwood, Appl. Phys. Lett., 1972, 20, 93–95.

---