Highly infrared-transparent and p-type conductive CuSc_1−xSn_xO₂ thin films and a p-CuScO₂:Sn/n-ZnO heterojunction fabricated by the polymer-assisted deposition method

Abstract

We fabricated a series of infrared (IR)-transparent and conductive Sn-doped CuScO₂ thin films using a polymer-assisted deposition (PAD) method. Highly c-axis oriented films were grown on a-plane sapphire substrates, and the orientation relationship with respect to the substrates was confirmed to be CuScO₂:Sn[3R](0001)//a-Al₂O₃(112̄0). The films exhibited good p-type conductive characteristics, and a high conductivity of 17.86 S cm⁻¹ was achieved for CuSc_0.94Sn_0.06O₂ at room temperature, which is 19 times larger than the un-doped film at room temperature. The prepared films retain a transparency of >70% in the near-IR region and >90% in the mid-IR region, which are the best performing p-type IR transparent conductive thin films grown by chemical solution methods to date. As x changes from 0 to 0.06, the hopping activation energy varies within the range of 129–98 meV. As an application of such films, p-CuScO₂:Sn/n-ZnO heterojunction diodes were fabricated using the PAD technique. The measured ∼2 V threshold voltage of the p-CuSc_0.94Sn_0.06O₂/n-ZnO diode is in reasonable agreement with the bandgap energy of CuSc_0.94Sn_0.06O₂. The resulting films can be used in optoelectronic devices over a wide wavelength range from the visible to IR band, due to their high electrical conductivity and optical transparency.

---

1 Introduction

Transparent conductive oxides (TCOs) combine the seemingly mutually exclusive properties of electrical conductivity and optical transparency in a single material, which are needed for wide applications including solar cells, displays, light emitting diodes, and transparent electronics.^1,2 Binary oxides with large optical bandgap (>3.0 eV), such as ZnO, In₂O₃ and SnO₂, usually have excellent n-type conductivity.^1,3 Since the discovery of the p-type transparent conductive CuAlO₂ thin film in 1997,⁴ wide bandgap p-type cuprous delafossite oxides have gained significant attention.⁵ However, the development of efficient p-type TCOs remains a global material challenge. Because of the localized nature of the O2p-derived valence band, it is difficult in introducing shallow acceptors and large hole effective masses, and thus extremely difficult in converting oxides from n-type to p-type via acceptor doping.⁶ To date, several Cu⁺-based oxides, such as delafossite CuMO₂ (M = Cr,⁷ Fe,⁸ In,⁹ Y¹⁰ and Ga¹¹), have been reported, but their conductivity and transparence are quite limited. As one kind of delafossites, CuScO₂ has a valence band that mainly consists of Cu^I d¹⁰ orbitals and a conduction band consists of Sc^III d⁰ and O p orbitals, so it has the possibility of acceptor doping and high hole mobility.¹² CuScO₂ exhibits p-type conductivity through a small polaron hopping mechanism localized at Cu site. Relatively high conductivities can be achieved through the susceptibility of CuScO₂ to both extrinsic acceptor doping and oxygen interstitials.¹³ The direct-allowed bandgap of CuScO₂ has been determined to be within 3.3–3.7 eV,^14–16 which results in a low free-carrier plasma oscillation cut-off frequency and a high transparence in infrared (IR) region.

Though CuScO₂ film can be fabricated by physical vapour deposition (PVD),^13,17,18 high temperature, controlled atmosphere, and strict vacuum conditions are required by this method. Without using traditional techniques, in this letter, for the first time, we grew single-phase epitaxial CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films on a-plane sapphire substrates by polymer-assisted deposition (PAD) technique. The two used soluble polymer materials, polyethyleneimine (PEI) and ethylenediaminetetra aceticacid (EDTA), play a significant role in preparing high quality films. The formation of covalent complexes between the lone pairs on the nitrogen atoms of the polymer and the metal cations makes it possible to achieve a pure precursor solution by the commonly available chemical solution deposition techniques. In addition, the polymers not only control the desired viscosity for the process but also bind the metal ions to prevent premature precipitation and formation of metal oxide oligomers. We also adopted this method along with radio frequency (RF) magnetron sputtering technique to fabricate p-CuScO₂:Sn/n-ZnO heterojunctions. This method is more cost-effective in large-scale and high-throughput device fabrication than other techniques, e.g. the costly pulsed laser deposition (PLD) technique for CuAlO₂/ZnO transparent diodes.¹⁹

2 Experimental methods

2.1 Preparation of precursor solution and CuSc_1−xSn_xO₂ thin films

The precursor solutions for the growth of CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films were respectively prepared by mixing two separate aqueous solutions of Cu, Sc and Sn bound to polymers. In detail, copper(II) nitrate (99.9%, 0.1 mol) was dissolved in 40 mL deionized water, followed by adding 3 g PEI polymer. Scandium(III) nitrate (99.9%) and stannic(II) chloride (99.9%) with the desired molar ratio were dissolved in 40 mL deionized water, followed by adding 4 g EDTA, and then 3 g PEI was added. The two polymers are soluble and compatible with the metal precursors, and also could decompose into small gaseous organics upon heating. The first row transition metal Cu binds well to the simple PEI polymer, schematically in a manner as shown in Fig. 1(a). EDTA can form stable complexes with almost all metal ions, such as Sc and Sn, and then the EDTA complexes bind to PEI via a combination of hydrogen bonding and electrostatic attraction, as show in Fig. 1(b) and (c). The two solutions were then separately filtrated and concentrated in an Amicon filtration unit using Amicon stirred cells and 3000 molecular weight cut-off, flat cellulose filter disks (YM3) under 65 psi (1 psi = 6.89 kPa) nitrogen pressure to give a final concentration of 250 mM for Cu and 250 mM for Sc and Sn, respectively. The precursors with desired stoichiometric ratio formed by mixing the above solutions were spun coated onto a-plane sapphire substrates at 3000 rpm for 30 s and baked at 300 °C in air for 30 min, in order to remove the polymers. The spin coating and baking process were repeated for six times to get a reasonable film thickness. The films were annealed at 900 °C for 1 h in flowing N₂, and then natural cooling to room temperature.

Fig. 1 Schematic illustration of chemical coordination of (a) Cu ion with PEI, (b) Sc ion and (c) Sn ion with PEI and EDTA by hydrogen bonding and electrostatic binding. (d) XRD patterns of the CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films. (e) Enlarged XRD pattern of the CuSc_1−xSn_xO₂ films at (0009) peak. The inset in Fig. 3(e) shows the lattice constants (a and c) dependent on Sc concentration.

2.2 Characterization and measurement

The crystal structure and crystallographic orientation of the CuSc_1−xSn_xO₂ thin films were checked using conventional X-ray diffraction (XRD) technique with a Bruker D8 Advance X, Pert diffractometer (Cu-Ka: λ = 1.540 Å). The scanning speed was 4° min⁻¹ from 10° to 70°. The surface morphology of the films was examined by an atomic force microscope (AFM) instrument (Veeco DI-3100) was used to further observe the microstructure. The surface and atomic structure of the films were also investigated using a high-resolution transmission electron microscope (HRTEM, TEM 2010F). The optical properties, such as the transmission and the absorption of films, were measured using an UV-VIS-NIR spectrophotometer (Shimadzu UV-3600PC) in the wavelength range of 250–3000 nm and Fourier transform infrared spectrometer (FTIR) in the range of 400 to 4000 cm⁻¹. Photoluminescence emission and excitation were recorded on a Flouromax-4 spectrometer (HORIBA JOBIN YVON) using He–Cd laser with an excitation wavelength of 270 nm. Electrical properties were determined using a Hall-effect measurement system (ACCENT HL55OOPC) within 150–300 K.

2.3 Sample preparation for HRTEM measurement

A polymethyl methacrylate (PMMA)-assisted method was used to divert the CuSc_1−xSn_xO₂ film samples from substrate for HRTEM measurement. First, a PMMA layer was spin coated on the surface of the sample, and then the sample was baked on a heating plate for 5 minutes. A piece of scotch tape with its size larger than the sample was used as a support: open a window in the middle of a tape that was smaller than the sample surface; stick the tape on the surface of the sample; make sure that the sample surface can be seen through the window; cut the edge part of the tape to make the tape to be of the same size with the sample. Soak the sample in 5% HF solution for 1 hour-etching. After that, the PMMA layer supported by the tape was gently stripped from the substrate with a needle. The PMMA layer was rinse in deionised water, and then was posted on a Cu mesh used in the HRTEM measurement. The Cu mesh was baked under an infrared lamp for 5 minutes to bind the PMMA film and the Cu mesh together. Remove part of the PMMA film that is larger than the Cu mesh. Finally, the PMMA layer was removed with hot acetone.

3 Results and discussion

3.1 Structure characteristics

Fig. 1(d) shows the XRD patterns of the prepared CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films. In addition to the substrate peak, well-defined reflections of (0003), (0006), (0009) and (0001̲2̲) are the only observed peaks of the CuSc_1−xSn_xO₂ films, indicating delafossite crystal structure of the films. The appearance of only (0001) diffraction peaks indicates that the films are preferentially oriented along the c-axis and are perpendicular to the substrates surface. The full width at half-maximum (FWHM) of the (0009) rocking curves (Fig. 1(e)) are approximately 0.171–0.32°. With the increase of Sn content, the intensities of (0003) and (0009) peaks become weaker and shift towards smaller angles. This effect occurs due to the partial substitution of Sc3d (0.745 Å) sites with large ionic radii Sn3d (1.12 Å), leading to local lattice distortion. Using the jade 5.0 software package and CuScO₂ cif file (JCPDF#79-0599 CSO [3R]), the lattice constants (a and c) were calculated and plotted in the inset of Fig. 1(e). The Sn ions are thus deduced to have occupied the Sc site in the trivalent state because of the observed linear relation.

3.2 Morphologies and microstructure of the CSO film

The surface morphology atomic force microscope (AFM) images of the films are shown in Fig. 2(a1)–(c1). With the increase of the Sn content, the average grain size grows, and the root-mean square (RMS) values of the surface roughness are 9.11, 14.72, and 16.90 nm, respectively. The average crystalline size of a film can be calculated using the Scherrer equation

τ = kλ/β cos θ (1)

where k is a dimensionless shape factor, with a value close to unity (∼0.9), β is the full width at half maximum (FWHM), λ is the X-ray wavelength, and θ is the diffraction angle. We used the (0009) diffraction peak to calculate the film's crystallinity. The average grain size of the CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) films was calculated as ∼83, ∼117 and ∼158 nm, respectively. There are two reasons for the grain size increase with Sn doping. First, the Fermi level moves to the top of the valence band because of the acceptor impurities caused by Sn-doping. The increase of hole concentration leads to the rise of self-diffusion coefficient. Second, the impurity segregation between grain boundaries will change the interfacial energy and thus increase driving force.

Fig. 2 AFM images (a1, b1, c1), HRTEM images (a2, b2, c2) and SAD patterns (a3, b3, c3) of CuSc_1−xSn_xO₂ thin films for a doping ratio of x = 0 (a1, a2, a3), x = 0.03 (b1, b2, b3), and x = 0.06 (c1, c2, c3), respectively.

High resolution transmission electron microscopy (HRTEM) images are shown in Fig. 2(a2)–(c2). The lattice spacing between adjacent planes are found to be 4.2 Å, 4.22 Å and 4.25 Å, respectively, which correspond to the d-spacings of (12̄10) planes in the three films. The selected area diffraction (SAD) patterns in Fig. 2(a3)–(c3) shows that the CuSc_1−xSn_xO₂ thin films have single crystal structure with the c-axis orientation perpendicular to the surface of the a-plane sapphire substrates, and the diffraction spots were assigned as (101̄0) and (12̄10).

3.3 Optical measurements of the film

Fig. 3(a1, a2), (b1, b2) and (c1, c2) show the optical transmission characteristics of the CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films over the UV-Vis-IR region, where the contribution of substrate has been subtracted. All the films have an optical transmission of 40–65% in the whole visible range, and also show a high transmission of 65–90% in the near-IR region.

Fig. 3 (a1–c1) UV-Vis-near-IR transmission spectrum and (a2–c2) mid-IR transmission spectrum of CuSc_1−xSn_xO₂ thin films, where (a1 and a2) x = 0, (b1 and b2) x = 0.03 and (c1 and c2) x = 0.06. Insets in (a1–c1) show the (αhν)² ∼ hν plot and insets in (a2–c2) show absorption spectrum for CuSc_1−xSn_xO₂ thin films.

As shown in Fig. 3(a2), the CuScO₂ film has a total transmittance of >90% in the wavelength range of 2.5 (4000 cm⁻¹) to 6 μm (1667 cm⁻¹) without any sharp absorption features. The high transmittance (or low absorption), particularly in the mid-IR region, is mainly due to the free-carrier plasma edge located in the far-IR frequency.¹⁹ The sudden decrease in the transmittance at 6 μm is caused by the collective oscillations of the conduction band electron, which is known as plasma oscillations. The free-carrier plasma oscillation frequency (denoted by ω_p) determines the low cut-off frequency of transmission band. The CuSc_1−xSn_xO₂ thin films with x = 0.03 and 0.06 show a decrease of transmission wavelength range in the mid-IR range (<5 μm), because of the decrease of carrier mobility with Sn doping resulting in the increase of ω_p. So the high transmission range in the mid-IR band of the Sn-doped film becomes narrower, as shown in Fig. 3(b2) and (c2). The insets in Fig. 3(a2)–(c2) show the absorption spectrum of CuSc_1−xSn_xO₂ films. Sharp absorption edges are observed around the wavelength of 340 nm. The relationship between optical absorption coefficient (α) and photon energy electromagnetic (hν) can be expressed as

ahν = A(hν − E_g)^m (2)

where A is a constant, E_g is the optical bandgap, and m depends on the type of transition (m = 1/2 for direct band transition, m = 2 for indirect band transition).²⁰ The insets in Fig. 3(a1)–(c1) show the (αhν)² ∼ hν plots. The linear relationship of (αhν)² ∼ hν indicates that CuSc_1−xSn_xO₂ films have direct energy bandgaps, which are estimated to be 3.6 eV, 3.5 eV and 3.4 eV for x = 0, 0.03, 0.06, respectively, by extending the straight portion of the curve. These values are in the reported range of 3.3–3.75 eV.^14,18 Generally, the variation of band gap can be fitted by the equation ΔE = ΔE(0) − kp^1/3, where p is carrier concentration. There is a linear relation between the cube root of carrier concentration and the band gap.²¹ Therefore, the optical band gap shows a decrease with the increase of Sn composition resulting from the increase of the hole concentration.

Besides the influence of free carrier absorption, the film surface roughness will also affect the transmittance by causing scattering loss. Based on Fig. 2(a1)–(c1), as the increase of the doping ratio, the surface roughness of the films increases. However, no obvious difference was observed among the transmittance of the three films within the visible band and near-infrared band, e.g. 0.5–3.0 μm. So it can be declared that surface roughness causes little effect on transmittance.

The room-temperature photoluminescence (PL) spectra of the CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films are shown in Fig. 4(a). As Sn content increases, the PL peak shifts towards higher wavelength because of the variation of optical bandgap. The PL spectra exhibits violet emission at 344 nm (3.6 eV), 350 nm (3.54 eV) and 358 nm (3.46 eV) for x = 0, 0.03 and 0.06, respectively, due to the near-band edge (NBE) radiative recombination. The observed corresponding energy values from PL spectra also closely match those estimated from the (αhν)² ∼ hν plot.

Fig. 4 (a) Photoluminescence (PL) spectra of CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films with the excitation wavelength of 280 nm at room temperature. The conductivity was measured as a function of temperature for CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) films over the temperature range of 300–150 K, (b) plots of ln σ ∼ 1000/T, (c) plots of ln σ ∼ 1/T^1/4.

3.4 Hall effect measurement at low temperature

In this study, the used doping material for the growth of CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films is SnCl₂·nH₂O, where the Sn ion presents +2 state. Then the substitution of Sc³⁺ by Sn²⁺ will form acceptor doping and further improve the p-type electrical conductivity of the CuScO₂ film. The measurement results of Hall effect for CuSc_1−xSn_xO₂ thin films are shown in Table 1. The conductivity of CuSc_0.94Sn_0.06O₂ film at room temperature is 17.86 S cm⁻¹, which is 19 times larger than that of a polycrystalline counterpart. The hole concentration ranges from 1.13 × 10¹⁸ to 6.2 × 10²⁰ cm⁻³ with the corresponding mobility ranging from 5.31 to 0.18 cm² V⁻¹ s⁻¹ for x changing from 0 to 0.06. In CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films, a dopant Sn²⁺ ion will act as an acceptor when it occupies a Sc³⁺ site, which will induce an increase of hole carrier concentration. Although the carrier mobility decreases due to the ionized impurity scattering, the increase of carrier concentration is much faster than the decreasing of mobility, the conductivity can easily become higher.

Table 1 Measurement results of Hall effect for the CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films at room temperature

x	Conductive type	E a (meV)	Conductivity (S cm^−1)	Hall mobility (cm^2 V^−1 s^−1)	Carrier density (cm^−3)
0	p	129	0.96	5.31	1.13 × 10^18
0.03	p	107	5.27	0.97	3.4 × 10^19
0.06	p	98	17.86	0.18	6.2 × 10^20

---

In the Sn-doped CuScO₂ films, the substitutional defect of “Sn_Sc” will be generated, which has a negative charge. At the same time, an oxygen atom vacancy defect of “Vo” will also be produced. Therefore, the overall charges are balanced in Sn-doped CuScO₂. In addition, a dopant Sn²⁺ ion acts as a donor rather than an acceptor if it occupies a Cu⁺ site rather than a Sc³⁺ site, which will induce the compensation of the intended acceptor doping. Therefore, the conductivity increase becomes smaller for higher Sn-composition.

As shown in Fig. 4(b), the CuSc_1−xSn_xO₂ films show a thermal activation behavior within the temperature range of 150–300 K, because the plots of ln σ ∼ 1000/T show a linear relation.^5,11 In this temperature range, the hopping of holes between the nearest neighbor Cu sites determines the electrical transport properties.²² The conductivity can be expressed by σ = A exp[−E_a/(k_BT)], where E_a is the thermal activation energy, k_B is the Boltzmann constant, and A is a constant. Increased electrical conductivity with temperature confirms the semiconductor nature of all CuSc_1−xSn_xO₂ films in our study. The estimated E_a, as shown in Table 1, are 129, 107, and 98 meV for CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) films, respectively. These values are less than 10% of the optical bandgap (E_g ≈ 3.3–3.7 eV) of CuSc_1−xSn_xO₂, which indicates that the hole transport in the valence band is thermally activated from an acceptor. An obvious decrease in E_a is observed as the increase of Sn content, which should be ascribed to different carrier densities, associated with the position of the Fermi level. The doping of Sn increases the hole density and makes the Fermi level approach to the top of the valence band, then the carriers can easily jump to the Fermi level from valence band, resulting in a decrease of E_a.²³

Plots of ln σ ∼ 1/T^1/4 for CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) films can be seen in Fig. 4(c), whose linear characteristic indicates the behavior of three-dimensional variable range hopping at lower temperature (below T_cross).²⁴ In low temperature region, the thermal energy weakening may depress the hole hopping between the nearest neighbor Cu sites, and the contribution of Sc/Sn site hopping becomes more dominant. Hence, the hopping to the Cu site in other Cu layers becomes relatively dominant. Such conductivity shows the three dimensional variable range-hopping behavior.²¹ This conduction mechanism in CuSc_1−xSn_xO₂ is closely similar to that in CuAlO₂ (ref. 5) and CuGaO₂ (ref. 11) as reported previously.

3.5 Fabrication of and measurement of p-CuScO₂:Sn/n-ZnO heterojunction

As an application of such films, p-CuScO₂:Sn/n-ZnO heterojunction diodes with good rectification characteristics were fabricated using PAD technique and RF magnetron sputtering technique. Fig. 5(a) shows the schematic structure of p-CuScO₂:Sn/n-ZnO heterojunction diode. The thickness of CuScO₂:Sn film deposited by PAD technique was 200 nm, and that of the ZnO layer by RF magnetron sputtering technique was 300 nm. Fig. 5(b) shows the I–V curves of the diodes. No clear diode characteristics are observed for the diode fabricated using CuSc_1−xSn_xO₂ (x = 0, 0.03) films, probably owning to poor interface and current leakage paths because of lattice mismatch. Diodes with good rectification characteristics could be fabricated using CuSc_0.94Sn_0.06O₂ thin film, and the threshold voltage is ∼2 V, which is in reasonable agreement with the bandgap energy of CuSc_0.94Sn_0.06O₂. The illustration demonstrates the hetero-epitaxial growth at the CuSc_1−xSn_xO₂ (lower part) and ZnO (upper part) interface.

Fig. 5 (a) Schematic structure of p-CuScO₂:Sn/n-ZnO heterojunction diode. (b) Current (I)–voltage (V) characteristics of p-CuScO₂:Sn/n-ZnO heterojunction diode. The inset in (b) shows the cross-sectional HREM image of p-CuSc_0.94Sn_0.06O₂/n-ZnO interface.

4 Conclusions

In summary, using PAD technique, we have grown single-phase and p-type IR transparent conductive CuSc_1−xSn_xO₂ (x = 0, 0.03, 0.06) thin films and p-CuScO₂:Sn/n-ZnO heterojunction diodes on a-plane sapphire substrate. The obtained CuSc_1−xSn_xO₂ thin films exhibited high optical transparency in visible and IR region simultaneously with excellent p-type electrical conductivity. The p–n heterojunction diodes based on CuSc_0.94Sn_0.06O₂ film have an abrupt interface and exhibit rectifying I–V characteristics with a threshold voltage of ∼2 V, which verifies that the heterojunction formed by the wide bandgap CuSc_0.94Sn_0.06O₂ and ZnO works as a good carrier blocking contact. These unique photoelectric properties make the CuSc_1−xSn_xO₂ films a promising candidate for use in window electrodes and other novel applications over a wide wavelength range.

Acknowledgements

This work was supported in part by the National Natural Science Foundation of China (no. 11404129, 61307124), National Key Technology R&D Program (no. 2013BAK06B04, 2014BAD08B03), Science and Technology Department of Jilin Province (no. 20140307014SF), Changchun Municipal Science and Technology Bureau (no. 11GH01, 14KG022) and the State Key Laboratory of Integrated Optoelectronics, Jilin University (no. IOSKL2012ZZ12).

References

1. K. Nomura, H. Ohta, K. Ueda, T. Kamiya, M. Hirano and H. Hosono, Science, 2003, 300, 1269.
2. C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2007, 91, 1529.
3. K. H. L. Zhang, R. G. Egdell, F. Offi, S. Iacobucci, L. Petaccia, S. Gorovikov and P. D. C. King, Phys. Rev. Lett., 2013, 110, 056803.
4. H. Kawazoe, M. Yasukawa, H. Hyodo, M. Kurita, H. Yanagi and H. Hosono, Nature, 1997, 389, 939.
5. D. S. Hecht, L. B. Hu and G. Irvin, Adv. Mater., 2011, 23, 1482–1513.
6. H. Kawazoe, H. Yanagi, K. Ueda and H. Hosono, MRS Bull., 2000, 25, 28.
7. R. Nagarajan, A. D. Draeseke, A. W. Sleight and J. Tate, J. Appl. Phys., 2001, 89, 8022.
8. Y. H. Chuai, B. Hu, Y. D. Li, H. Z. Shen, C. T. Zheng and Y. D. Wang, J. Alloys Compd., 2015, 627, 299–306.
9. H. Yanagi, T. Hase, S. Ibuki, K. Ueda and H. Hosono, Appl. Phys. Lett., 2001, 78, 1583.
10. M. K. Jayaraj, A. D. Draeseke, J. Tate and A. W. Sleight, Thin Solid Films, 2001, 397, 244.
11. K. Ueda, T. Hase, H. Yanagi, H. Kawazoe, H. Hosono, H. Ohta, M. Orita and M. Hirano, J. Appl. Phys., 2001, 89, 1790.
12. R. Kykyneshi, B. C. Nielsen, J. Tate and A. W. Sleight, J. Appl. Phys., 2004, 96, 6188.
13. Y. Kakehi, K. Satoh, T. Yoshimura, A. Ashida and N. Fujimura, Thin Solid Films, 2010, 518, 3097–3100.
14. N. Duan, A. W. Sleight, M. K. Jayaraj and J. Tate, Appl. Phys. Lett., 2000, 77, 1325.
15. L. J. Shi, Z. J. Fang and J. B. Li, J. Appl. Phys., 2008, 104, 073527.
16. Y. Kakehi, K. Satoh, T. Yotsuya, K. Masuko, T. Yoshimura, A. Ashida and N. Fujimura, J. Cryst. Growth, 2009, 311, 1117–1122.
17. Q. X. Jia, T. M. McCleskey, A. K. Burrell, Y. Lin, G. E. Collis, H. Wang, A. D. Q. Li and S. R. Foltyn, Nat. Mater., 2004, 3, 529.
18. Y. Kakehi, K. Satoh, T. Yotsuya, S. Nakao, T. Yoshimura, A. Ashida and N. Fujimura, J. Appl. Phys., 2005, 97, 083535.
19. A. D. LaForge, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 125120.
20. Z. Deng, X. Fang, S. Wu, Y. Zhao, W. Dong, J. Shao and S. Wang, J. Alloys Compd., 2013, 577, 658–662.
21. T. Kawaharamura, H. Nishinaka, Y. Kamada and T. Ohshima, J. Appl. Phys., 2007, 101, 083705.
22. T. Okuda, N. Jufuku, S. Hidaka and N. Terada, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 144403.
23. D. Li, X. D. Fang, Z. X. Deng, W. W. Dong, R. H. Tao, S. Zhou, J. M. Wang, T. Wang, Y. P. Zhao and X. B. Zhu, J. Alloys Compd., 2009, 486, 462.
24. M. J. Han, Z. H. Duan, J. Z. Zhang, S. Zhang, Y. W. Li, Z. G. Hu and J. H. Chu, J. Appl. Phys., 2013, 114, 163526.

---