Highly transparent and conducting C:ZnO thin film for field emission displays

Abstract

Incorporation of carbon into a zinc oxide (ZnO) thin film has led to a potential new application as a transparent and conductive oxide thin film. Here, we fabricated carbon doped zinc oxide (C:ZnO) films by RF sputtering at a low temperature, aiming to produce a transparent and conductive screen for transparent field emission displays. The incorporation of highly inert carbon atoms into ZnO thin films can tune the intrinsic defect sites. A co-sputtering technique was used for carbon doping into ZnO, where the dopant amount was controlled by varying the number, size and position of the graphite plates used as the carbon source. A sheet resistance of 37 Ω □⁻¹ and a transmittance of 84% at a wavelength of 550 nm were achieved for the C:ZnO thin film with a nominal carbon concentration of 2.7%. The bonding of carbon with ZnO was analyzed using de-convolution of XPS peak C1s. We also demonstrated the light emission properties of the C:ZnO thin films by fabricating a field emission device. A high current density of 1 mA cm⁻² was obtained with a lower applied electric field of 5 V μm⁻¹. Our findings showed that the C:ZnO thin films can be a promising transparent and conducting phosphor screen for a field emission display.

---

1. Introduction

Transparent conducting oxide (TCO) films are gaining significant importance for a wide range of electronic device applications, such as solar cells,^1,2 flat-panel displays,^3,4 light emitting diodes^5–7 and thin film transistors.^8–10 In the prospect of fabricating a transparent field emission (FE) display, a transparent and conducting phosphor film is the most essential component. In the present state of the art, the display is normally covered with a phosphor material, such as, ZnGa₂O₄, ZnS, Y₂O₃:Eu, CdSe or several others, however these materials significantly reduce the transparency of the display device.^11–13 Thus, a potential material for phosphor films that possesses a high transmittance property and can emit light by itself would be the most suitable. ZnO is one of the promising materials for transparent thin film applications, and is widely known to have green light emission due to the presence of native defects.^14–16 However, undoped ZnO thin films have a limitation with regards their conductivity, even though high transmittance can be achieved. In many reports, trivalent metal ions (Al³⁺, Ga³⁺ and In³⁺) have been employed into ZnO wurtzite crystals to increase the free carrier concentration and consequently improve the conductivity of ZnO thin films.^17–20 However, even though conductivity and transparency could be achieved for the metal doped ZnO thin films, the light emission property is still an issue. Thus, the aim of this work is to improve the conductivity of highly transparent ZnO thin films by doping them with a non-metal element, which also contributes to modifying the intrinsic defects. Graphite plates were used as the carbon source, which was co-sputtered with ZnO by RF sputtering at a low temperature. The sputtering method could be a reliable method to fabricate a thin film without forming any agglomeration of the doping material. The fine distribution of particles on the thin film would help to avoid coulombic degradation, which has been one of the problems in phosphor screens.²¹

We investigated the effect on the transparency and conductivity of C:ZnO thin films with variation of the number, size and position of the graphite plates used as the carbon source along with the ZnO target. The ferromagnetism^22,23 and the p-type properties^24,25 of C:ZnO have been widely reported, and yet there is no report on the optical properties of the thin film. Carbon is an amphoteric impurity in II–IV compound semiconductors that could improve the electrical properties of ZnO thin films by forming defects such as oxygen vacancies or Zn interstitials. If a higher transparency and conductivity can be achieved, the thin film can be one of the promising TCO films for a myriad of applications in electronic industries. Besides its advantages of being a cheap and abundant material that does not undergo oxidation, the ability to emit light from the carbon doped thin film can be one of the merits of C:ZnO films when applied as a transparent phosphor screen for field emission displays.

2. Experimental details

2.1. Thin film preparation

A highly pure ZnO ceramic target (99.99%) and graphite plates (1 × 1 × 0.1 and 1.5 × 1.5 × 0.1 cm) were used as the sources for ZnO and carbon, respectively. The size of the ZnO target is 101.6 cm in diameter and 5 cm in thickness. The substrate was 1 × 1 cm corning glass, which was first cleaned with ethanol for 10 min followed by rinsing in distilled water for another 10 min in an ultrasonic bath. Then, the glass substrate was blow dried with air using a squeezer hand pump. The C:ZnO thin films were deposited by RF sputtering (ANELVA SPF-210HS) for 20 min using an RF power supply (13.56 MHz) with Ar as the sputtering gas under the sputtering conditions as shown in Table 1. The glass substrate was placed on a substrate holder at a fixed distance of 5 cm facing the target. Prior to the film deposition, the surface of the target was cleaned for 10 min by a pre-sputtering process using the same conditions as given in Table 1, while keeping the shutter closed.

Table 1 Sputtering conditions of the C:ZnO thin films

Condition	Value
Ar flow rate	1 sccm
Background pressure	3 × 10^−5 Pa
Working pressure	0.5 Pa
Substrate temperature	200 °C
RF power	200 W

---

A schematic of the sputtering process is shown in Fig. 1. The carbon concentration of the C:ZnO thin films was calculated as the nominal dopant concentration based on the percentage of the total area of the graphite plate on the ZnO target, as illustrated in Fig. 2. Naturally, carbon cannot be avoided during sample fabrication even for the undoped thin film. It can easily be detected from analyses such as X-ray photoelectron and energy dispersed X-ray. However, unlike other materials, the difficulty in quantify an atomic percentage of pure carbon concentration in the thin films by a co-sputtering process has led to the calculation of nominal carbon concentration. The percentage of nominal carbon concentration will be used throughout this paper as the carbon composition in the C:ZnO thin films. It seems difficult to reproduce the doped thin films without introducing an accurate concentration that corresponds to the atomic percentage of the doping material. However, with the same experimental setup, background pressure, and experimental parameters, doped thin films with almost the same film properties are reproducible. There were three samples of C:ZnO thin film prepared in this experiment, in which the calculated percentages of the graphite area were: 2.41%, with a distance of about 4 cm between two 1 × 1 cm graphite plates that were placed at a high erosion area of the ZnO target (denoted as CZnO1); 2.41%, with two 1 × 1 cm graphite plates with no distance between them placed at the center of the ZnO target (denoted as CZnO2); and 2.7%, with a single 1.5 × 1.5 cm graphite plate at the center of the ZnO target (denoted as CZnO3), respectively. As a reference to study the concentration of carbon in the ZnO thin films, an undoped ZnO thin film was prepared under the same conditions as for the C : ZnO thin films.

Fig. 1 Schematic diagram of RF sputtering.

Fig. 2 Top view of the graphite plate position on the ZnO targets.

2.2. Characterization

The surface morphology and roughness of the undoped and carbon doped ZnO thin films were characterized using a scanning probe microscope (JEOL JSPM-5200TM). The sheet resistance was measured using 4 point probes (Napson RT-70V/RG-7C) without deposited contact on the thin films. The measurement was done at 10 random points on each of the thin films and the average sheet resistance was taken as the result. Raman spectra were measured by Raman spectroscopy (Jasco NRS-3300) with an excitation wavelength of 534.08 nm. The average thickness of the thin films was measured using a Dektak profilometer at two different edges for each thin film. The transmittance property was analyzed using a UV-vis spectrophotometer (Jasco V-670K) and the film transparency was defined at a visible wavelength of 550 nm. X-ray photoelectron spectroscopy (XPS) of the thin films was analyzed using a focused monochromatized Mg-Kα X-ray source (1253.6 eV) in an XPS instrument (Shimadzu ESCA-3300KM). The luminescence property of the thin films was measured using SEM-CL spectroscopy (JEOL JSM 7800F) at room temperature with an accelerating voltage of 5 kV. The field emission property was measured by using a parallel plate configuration in which the C:ZnO thin film was used as an anode material, while the cathode was a single wall carbon nanotube (SWCNT) thin film. The size of the emission area was 0.1 cm² and the spacer was a layer of Teflon of 100 μm thickness. The measurement was done in a vacuum when the background pressure reached 3 × 10⁻⁴ Pa and was carried out in a dark room to observe for light emission.

3. Results and discussion

3.1. Morphology, electrical and optical properties

The surface homogeneity of the ZnO thin films was improved with C doping. The AFM image in Fig. 3(a) shows ZnO particles of the undoped ZnO thin film. The surface roughness of that 5 × 5 μm² scanning area was 41 nm. The non-homogeneous surface of the undoped ZnO film was due to a large distance between the top and valley areas. With the incorporation of carbon, the distance measured between the top and valley areas was smaller compared to that of the undoped film. The surface roughness of CZnO1, CZnO2 and CZnO3 as shown in Fig. 3(b–d) was 35.1 nm, 29.4 nm and 25.6 nm, respectively. The higher the concentration of carbon introduced into the thin film, the more homogenous the morphology of the fabricated film. This result indicates that the carbon atoms in the ZnO thin films contribute to improving the surface homogeneity of the deposited films by forming a homogenous coalescence structure that can be clearly seen from the AFM images. As a result, the C:ZnO thin film surfaces could provide a smoother electron path, since the grain boundary is decreased, resulting in a lower sheet resistance of the C : ZnO films.

Fig. 3 AFM images of the undoped and carbon doped ZnO thin films.

The sheet resistance of the ZnO thin films decreases when carbon is introduced into the thin film, as shown in Table 2. The undoped ZnO thin film showed very high sheet resistance, which indicates the property of an insulator. As mention in the introduction section, the conductivity of undoped ZnO thin films can be improved by doping with trivalent metal materials such as Al and Ga. To achieve this for a phosphor and conductive screen film, carbon is the best candidate for a non-metal doping material in this experiment due to its properties to tune the film conductivity with a very low doping concentration. According to the literature, incorporation of carbon atoms into a ZnO thin film could modify the intrinsic native deep levels in ZnO, i.e. oxygen vacancies (Vo), Zn interstitials (Vzn_i), oxygen interstitials (Oi), Zn vacancies (Vzn), oxygen anti-sites (Ozn) and zinc anti-sites (Zno). These native point defects often control the doping and luminescence efficiency of ZnO directly or indirectly. A ZnO thin film with higher Vo is said to possess a higher conductivity.^26,27 The Raman spectra shown in Fig. 4 reveal the higher Vo in the ZnO thin films determined by A₁ (LO) peak intensity. There are two dominant peaks that can be observed from the Raman spectra. The E₂ (high) mode peak at 435–437 cm⁻¹ was attributed to a wurtzite lattice of ZnO with high crystallinity, and the A₁ (LO) mode at 570 cm⁻¹ can be associated with the oxygen deficiency in the matrix of ZnO. The E₂ (high) peak is often related to oxygen atoms. For the undoped ZnO film, the E₂ (high) peak and A₁ (LO) peak show almost the same intensity, indicating the stoichiometry of the undoped ZnO film. As carbon was introduced into the ZnO film, samples CZnO1, CZnO2 and CZnO3 show higher intensities of the A₁ (LO) mode compared to the E₂ (high) mode. The higher intensity peak of the A₁ (LO) mode for the carbon doped ZnO films indicates a higher oxygen deficiency in these films.²⁸

Table 2 Electrical and transmittance properties of the undoped and C:ZnO thin films

Sample	Electrical and optical properties
	Sheet resistance (Ω □^−1)	%T	Thickness (μm)
Undoped ZnO	Very high	77	1.3
CZnO1	1.4k	78	1.52
CZnO2	646	78	1.5
CZnO3	37	84	1.5

---

Fig. 4 Raman spectra of the undoped and C doped ZnO thin films.

The sheet resistance can also be affected by the thickness of the thin film. In this experiment, we observed that all the carbon doped samples had the same average thickness of 1.5 μm, while the undoped ZnO film had an average thickness of 1.3 μm. The thickness result shows that the low sheet resistance of the carbon doped thin films is not directly related to the film thickness. It can be assumed that the contribution of carbon atoms in the ZnO film could be responsible for the modification in the conductivity of the ZnO thin films.

The position of the graphite plate can also be a factor in the difference in sheet resistance, due to the number of sputtered atoms condensed on the substrate. The CZnO1 and CZnO2 thin films were deposited with the same nominal carbon concentration of 2.41%, but with a different position of the graphite plate on the ZnO target. The difference in sheet resistance of these two samples can be attributed to the difference in the scattering effect inside the sputter chamber during the deposition process. The CZnO1 thin film showed a sheet resistance of 1.4 kΩ □⁻¹, while the CZnO2 thin film showed a lower value of 646 Ω □⁻¹. A higher scattering effect can occur at the high erosion area, i.e. between the two permanent magnets at the sputter target. Due to this high scattering effect, the number of sputtered carbon atoms that are condensed on the substrate could be less, because the atoms would lose their energy due to the impact of collision with other particles inside the chamber. The graphite plate at the center of the ZnO target would have less collision, thus resulting in the good quality of the carbon doped thin film. The CZnO3 thin film with the highest nominal carbon concentration of 2.7% showed the lowest value of sheet resistance at 37 Ω □⁻¹. The lower sheet resistance of the CZnO3 thin film could be due to the high probability of more carbon atoms being able to condense on the substrate. Since the larger size of graphite plate was located at the center of the ZnO target, less collision is expected, and this supports the idea of more carbon atoms in the CZnO3 thin film.

The optical properties were measured in the wavelength range of 300–800 nm and the percentage of transmittance was taken at the visible wavelength of 550 nm. The transparency property can be affected by the surface morphology of the thin film. A smoother surface morphology will give a higher transmittance property due to less reflection of light scattering from the thin film surface.²⁹ Fig. 5 shows the transmittance of the undoped and C:ZnO doped thin films. The wavy curve of the transmittance spectra is because of the interference in light propagation during the measurement between the thick ZnO based film and the glass substrate. If the thickness of the deposited thin film is less than 500 nm, the wavy curve can not be observed. The undoped ZnO film possesses a transmittance of 77%, which is lower than that of the C:ZnO thin films. The CZnO3 thin film shows the highest transmittance property of 84%, yet the appearance color of the thin film becomes slightly brownish. The CZnO1 and CZnO2 thin films have the same value of transmittance at 78%. This result shows that having the same area of graphite plates located at different positions on the ZnO target did not have a significant effect on the transparency of the C:ZnO thin film.

Fig. 5 Transmittance spectra of the undoped and carbon doped ZnO thin films.

3.2. X-ray photoelectron spectrometry (XPS)

The chemical state and surface composition of the C:ZnO thin films were determined from the XPS spectra. Fig. 6 shows the wide survey scans of the X-ray photoelectron spectra of the undoped ZnO, CZnO1, CZnO2 and CZnO3 thin films, respectively. The C1s and O1s spectra were deconvoluted to study the binding energy of their components, as shown in Fig. 7 and 8, respectively. The relative amount of any component peak can be calculated from the area under the curve by integrating the related peak. In the C1s spectra, the deconvoluted peak at 286 eV corresponds to O–C–Zn in the carbon doped thin films.³⁰ The calculated relative amounts of the O–C–Zn peak for CZnO1, CZnO2 and CZnO3 are 53.74%, 34.22% and 85.15%, respectively. The peaks at higher binding energies of 289.9 eV (O=C–O) and 287.4 eV (C–O–C) were attributed to carbon–oxygen bonds.

Fig. 6 Wide survey scans of the XPS spectra.

Fig. 7 Deconvolution of the C1s peak of the undoped and carbon doped ZnO thin films.

Fig. 8 Deconvolution of the O1s peak of the undoped and carbon doped ZnO thin films.

The peak at ∼531 eV from the O1s spectra of CZnO1, CZnO2 and CZnO3 was attributed to O²⁻ ions in the wurtzite structure of ZnO. The peak at a binding energy of 531.4 eV observed in the CZnO3 thin film was reported as the presence of oxygen ions (O²⁻) in the oxygen deficiency regions within the matrix of ZnO, which indicates the existence of oxygen vacancies.^31,32 That peak at 531.4 eV was also reported to correspond to the bonding of C–O–Zn.³³ The strong bonding of Zn–O–C and the existence of oxygen vacancies were found to contribute to the higher conductivity of this sample compared to the others. The higher oxygen ion concentration may trap electrons and reduce the free carriers of the ZnO thin film. The peak at ∼533 eV was identified as the presence of loosely bound oxygen on the surface.

3.3. Luminescence emission properties

Fig. 9 shows the cathodoluminescence (CL) spectra of the undoped and carbon doped ZnO thin films (CZnO1) analyzed in the range of 250–900 nm. The luminescence emission properties are strongly related to the defect states in ZnO. The observed CL spectra of both the undoped and CZnO1 samples exhibit band-edge positions at 378 nm (3.28 eV) and 384 nm (3.25 eV), respectively. These peaks are attributed to the UV emission originating from recombination of free excitons from the conduction band (CB) to the valence band (VB). The peak observed at 752 nm for the undoped ZnO sample is attributed to the infra-red (IR) wavelength. A significant difference in the CL spectrum was observed for the carbon doped ZnO thin film. The broad peak centered at 646 nm (1.92 eV) indicated complex defects of substitutional carbon (Co), oxygen vacancy and interstitial Zn (2Co-Vo-Zn_i), which is responsible for red-orange emission.³⁴

Fig. 9 CL spectra of the undoped and carbon doped ZnO thin films.

3.4. Field emission application

In the field emission measurements, we employed CZnO1 as the phosphor film to observe light emission. The measurement was done by using two parallel electrodes in which a SWCNT film was used as the field emitter. Preparation of the SWCNT film as the field emitter was explained elsewhere.³⁵ Fig. 10 shows the FE property of the carbon doped ZnO film. The SWCNT emitter possesses an excellent efficiency owing to the high field enhancement factor of 4492 calculated from the linear behavior of the FN plot shown in the inset of Fig. 10(a). High emission current was also achieved at a lower threshold voltage. An electric field at a current density of 1 mA cm⁻² was achieved at 5 V μm⁻¹. The experimental result of light emission is shown in the inset of Fig. 10(b). It shows a UV-violet emission. The mechanism can be explained as typical exciton emission when photo-generated electrons recombine with holes in the valence band. The light emission became brighter as the electric field increased with a higher current density of more than 1 mA cm⁻². The presence of carbon in the ZnO thin film may modify the intrinsic defects and contribute to the light emission.³⁶

Fig. 10 The FE property of C doped ZnO film as the phosphor layer and SWCNT as the field emitter. The inset figures are (a) the FN plot and (b) the light emission from the C doped ZnO film.

4. Conclusions

We have successfully fabricated a highly transparent and conductive C:ZnO thin film and employed the carbon doped ZnO film as a phosphor screen. The position, size and number of graphite plates that act as the carbon source played a crucial role in a co-sputtering technique to fabricate the conductive and transparent C:ZnO thin film. A lower sheet resistance of the carbon doped ZnO thin film was achieved for sample CZnO3, which had the higher nominal carbon concentration of 2.7%. Carbon atoms could be condensed more on the glass substrate from a graphite plate positioned at the center of the ZnO target. Less collision of particles in the plasma at the center of the ZnO target compared to the high erosion area resulted in a lower sheet resistance of the CZnO2 thin film, even though the same nominal concentration was introduced in the chamber. All the carbon doped ZnO thin films did not show a significant difference in the transmittance properties. A CL spectrum shows a broad peak centered at 646 nm indicating red-orange luminescence emission from the carbon doped ZnO film (CZnO1). The experimental result shows UV-violet light emission, which is brighter when the current density increased above 1 mA cm⁻². Thus, depending on the doping and deposition conditions, highly conductive and transparent C:ZnO thin films can be obtained, which also act as a promising material for phosphor screens of transparent field emission displays.

Acknowledgements

The authors would like to thank Chubu University for the XPS equipment and their help with the measurements. The authors would also like to thank Mr Yanagihara Toshinari for his kind help in measuring CL from the FE-SEM instrument.

References

1. M. L. Grilli, A. Sytchkova, S. Boycheva and A. Piegari, Phys. Status Solidi A, 2013, 210, 748.
2. G. E. Eperon, V. M. Burlakov, A. Goriely and H. J. Snaith, ACS Nano, 2014, 8, 591.
3. C. W. Hsu, B. Zhen, W. Qiu, O. Shapira, B. G. DeLacy, J. D. Joannopoulos and M. Soljacic, Nat. Commun., 2013, 5, 3152.
4. K. H. Kim, H. M. An, H. D. Kim and T. Geun Kim, Nanoscale Res. Lett., 2013, 8, 507.
5. S. Ju, J. Li, J. Liu, P. C. Chen, Y. G. Ha, F. Ishikawa, H. Chang, C. Zhou, A. Facchetti, D. B. Janes and T. J. Marks, Nano Lett., 2008, 8, 997.
6. G. W. Kim, R. Lampande, J. Boizot, G. H. Kim, D. C. Choe and J. H. Kwon, Nanoscale, 2014, 6, 3810.
7. S. H. K. Park, C. S. Hwang, M. Ryu, S. Yang, C. Byun, J. Shin, J. L. Lee, K. Lee, M. S. Oh and S. Im, Adv. Mater., 2009, 21, 678.
8. P. Gorrn, M. Sander, J. Meyer, M. Kroger, E. Becker, H. H. Johannes, W. Kowalsky and T. Riedl, Adv. Mater., 2006, 18, 738.
9. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater., 2012, 24, 2945.
10. A. Klein, J. Am. Ceram. Soc., 2013, 96, 331.
11. P. F. Smet, I. Moreels, Z. Hens and D. Poelman, Materials, 2010, 3, 2834.
12. P. Thiyagarajan, M. Kottaisamy, N. Rama and M. S. R. Rao, J. Phys. D: Appl. Phys., 2009, 42, 155301.
13. M. A. Schreuder, J. D. Gosnell, N. J. Smith, M. R. Warnement, S. M. Weiss and S. J. Rosenthal, J. Mater. Chem., 2008, 18, 970.
14. Y. Gong, T. Andelman, G. F. Neumark, S. O’Brien and I. L. Kuskovsky, Nanoscale Res. Lett., 2007, 2, 297.
15. H. S. Kang, J. S. Kang, J. W. Kim and S. Y. Lee, J. Appl. Phys., 2004, 95, 1246.
16. H. Faber, J. Hirschmann, M. Klaumunzer, B. Braunschweig, W. Peukert and M. Haliks, ACS Appl. Mater. Interfaces, 2012, 4, 1693.
17. M. C. Jun, S. U. Park and J. H. Koh, Nanoscale Res. Lett., 2012, 7, 639.
18. J. Sun, Y. Huang and H. Gong, J. Appl. Phys., 2011, 110, 023709.
19. J. Jia, A. Takasaki, N. Oka and Y. Shigesato, J. Appl. Phys., 2012, 112, 013718.
20. A. K. Srivastava and J. Kumar, Sci. Technol. Adv. Mater., 2013, 14, 065002.
21. S. A. Bukesov, D. Y. Jeon and B. I. Lee, Appl. Phys. Lett., 2005, 87, 253503.
22. B. J. Nagare, S. Chacko and D. G. Kanhere, J. Phys. Chem. A, 2010, 114, 2689.
23. M. Subramanian, Y. Akaike, Y. Hayashi, M. Tanemura and D. L. S. Ping, Phys. Status Solidi B, 2012, 249, 1254.
24. S. T. Tan, X. W. Sun, Z. G. Yu, P. Wu, G. Q. Lo and L. Kwong, Appl. Phys. Lett., 2007, 91, 072101.
25. T. S. Herng, S. P. Lau, L. Wang, B. C. Zhao, S. F. Yu, M. Tanemura, A. Akaike and K. S. Teng, Appl. Phys. Lett., 2009, 95, 012505.
26. C. F. Yu, C. W. Sung, S. H. Chen and S. J. Sun, Appl. Surf. Sci., 2009, 256, 792.
27. M. S. Arnold, P. Avouris, Z. W. Pan and Z. L. Wang, J. Phys. Chem. B, 2003, 107, 659.
28. D. Das and P. Mondal, RSC Adv., 2014, 4, 35735.
29. P. Ghosh, S. Satou, T. Tsuchiya, Y. Hayashi and M. Tanemura, Phys. Status Solidi RRL, 2012, 6, 184.
30. D. K. Mishra, J. Mohapatra, M. K. Sharma, R. Chattarjee, S. K. Singh, S. Varma, S. N. Behera, S. K. Nayak and P. Entel, J. Magn. Magn. Mater., 2013, 329, 146.
31. P. T. Hsieh, Y. C. Chen, K. S. Kao and C. M. Wang, Appl. Phys. A, 2008, 90, 317.
32. B. Panigrahy, M. Aslam and D. Bahadur, Nanotechnology, 2012, 23, 115601.
33. J. Ding, M. Wang, X. Zhang and C. Ran, RSC Adv., 2013, 3, 14073.
34. L. T. Tseng, J. B. Yi, X. Y. Zhang, G. Z. Xing, H. M. Fan, T. S. Herng, X. Luo, M. Ionescu, J. Ding and S. Li, AIP Adv., 2014, 4, 067117.
35. D. Ghosh, P. Ghosh, G. Kalita, T. Noda, C. Takahashi and M. Tanemura, Phys. Status Solidi RRL, 2013, 7, 489.
36. Z. G. Wang and X. T. Zu, J. Mater. Sci., 2006, 41, 3729.

---