Co-doped AlN nanowires with high aspect ratio and high crystal quality

Abstract

This study presents a systematic investigation of AlN:Co nanowires synthesis by a DC arc discharge plasma method that is catalyst- and template-free. X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM) have been used to characterize the structure, components, morphology and size of the nanowires. Raman spectroscopy revealed very symmetric and strong phonon modes, indicating very good crystal quality of the AlN:Co nanowires. Comparison of the Raman spectra of AlN and AlN:Co nanowires showed that the Co doping influenced the lattice structure, producing impurities and defects. The photoluminescence emission spectrum (PL) confirmed that the Co²⁺ ions existed in the AlN:Co nanowires and enhanced the optical light range and intensity of the material. A magnetic study suggested that ferromagnetism in the AlN:Co nanowires may be due to the Co ion clusters, accompanied by Al vacancies. In brief, 1D-AlN:Co nanowires were fabricated with high aspect ratio and high crystal quality and possessed excellent optical and magnetic properties. These features are very valuable for scaling up manufacturing opportunities of AlN-based DMSs and will be beneficial to manipulate and integrate these for nanodevice fabrication in spintronics.

---

Introduction

In the past few years, great efforts have been devoted to fabricating a material that includes charge carriers and localized spins with a higher curie temperature.^1–5 A common approach to realize the objective is to inject magnetic elements into the semiconductors to form diluted magnetic semiconductors (DMSs). Among the most promising candidates, group III–V metal nitride DMS materials have stimulated great research due to their demonstrated success in combining strong optoelectronic and unique room-temperature ferromagnetic properties that can be applied in the emerging field of spintronics.^6–10 AlN is one of the important wide band gap group III–V semiconductors, which has excellent physical and chemical properties, such as high heat conductivity, hardness, chemical stability and low thermal expansion coefficient.^11–13 Compared with AlN bulk, AlN nanowires have many unique inherent properties and low dimensional advantages (large surface-to-volume ratio, low dislocation density and low electron affinity). These offer superior performance for potential applications in high electron mobility transistors, light-emitting diodes (LEDs), laser diodes (LDs), and photodetectors.^14–16

Although the synthesis of AlN-based DMS nanowires has been investigated in recent years, some unresolved challenges still exist: (i) most of the fabrication methods rely on catalysts, leading to detrimental defects in the products and poor performance in promising applications;^17–19 (ii) a notable issue with many early efforts is the short length (<2 μm) of the AlN nanostructures produced, which makes them very difficult to manipulate and integrate for nanodevice fabrication; and (iii) the size and geometric shape of nanomagnets play a vital role in determining their magnetic properties. The anisotropy depending on the geometric shape is the most important property of the magnetic element. Therefore, one of the most attractive features of nanostructure magnetic materials is that their magnetic properties can be engineered by the choice of the geometric shape. Moreover, as mentioned in many previous studies, AlN nanostructures emit in the blue and UV regions, which is not suitable for optoelectronic applications (such as LEDs) that require white light emission consisting of three primary colors (red, green, and blue). In view of these problems, we have designed a simple, one-step, catalyst-free and template-free method for synthesizing AlN:Co nanowires to achieve simultaneous control of both optical and magnetic properties. In addition, 1D-AlN:Co nanowires with high aspect ratio and high crystal quality have not been reported to date.

Herein, we experimentally show that 1D-AlN:Co nanowires with high aspect ratio and high crystal quality have been synthesized using a DC arc discharge plasma method. The AlN:Co products simultaneously possess excellent optical and magnetic properties after the introduction of Co. The PL of Co²⁺ ions enlarge the emitting region and intensity of AlN, whereas an aggregation of the doping ions into the magnetic ion-clusters can reduce the formation energy of Al vacancies, leading to ferromagnetism. Finally, the intensity of saturation magnetization increases with an increase in Co concentration.

Experimental

The Co doped AlN nanowires were prepared by an improved DC arc discharge plasma setup. The target raw material was a mixture of high purity Al (99.99%) and Co (99.99%). The mixing column was used as both the evaporation source and deposition substrate. N₂ gas (purity 99.999%) was used as nitrogen source. Before the N₂ gas was introduced into the chamber, the chamber was evacuated to less than 5 Pa of pressure, and then the pressure increased to 30 kPa. Over the course of the reaction, the input current was maintained at 100 A and the voltage was a little higher than 25 V. The formation process of AlN:Co nanowires involved the evaporation of Al and Co, decomposition of N₂, nucleation of AlN, and substitution of Co for Al. The reaction time and air pressure were accurately controlled to produce the desired wire-shaped nanostructure. After about 10 min of reaction, the substrate was covered with a larger number of white samples. Finally, the products were passivated for about 24 h in pure Ar gas at 50 kPa.

The crystal structure of AlN:Co was characterized by powder X-ray diffraction (XRD) using a Rigaku D/Max-A diffractometer with Cu Kα radiation. The morphology and microstructure were obtained by field emission scanning electron microscopy (FESEM) using XL 30 ESEM FEG, and transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM) using TECNAI G2. The composition was tested by X-ray photoelectron spectroscopy (XPS) (ESCALAB MK II). Photoluminescence (PL) spectroscopy was performed via Perkin Elmer photoluminescence, and the Raman spectrum was obtained by Renishaw. The magnetic properties were determined using a vibrating sample magnetometer (Lakeshore Model 7410 VSM) at room temperature.

Results and discussions

The structure of AlN products was determined by analysing the XRD patterns. As shown in Fig. 1a, all of the peaks can be readily indexed to the wurtzite structure (space group: P6₃mc (186), JCPDS file no. 08-0262), suggesting that only a single phase was formed. No peaks corresponding to metallic Al or alloys of Al and Co could be found in the XRD patterns. The detailed information/composition of the AlN products was further investigated by XPS studies. Fig. 1b and c showed the Al 2p and N 1s peaks centered at 74.5 eV and 397.3 eV. The peak at 74.5 eV was assigned to the Al–N bond, indicating the existence of the AlN nanostructure.²⁰ In Fig. 1d, the XPS spectrum of Co 2p shows four peaks: a doublet and its corresponding shake-up satellites at slightly higher energies. Binding energies of 781.8 and 797.0 eV were obtained for Co 2p_3/2 and 2p_1/2, respectively. The energy difference of 15.2 eV between these two binding energy values for cobalt and the standard Co 2p binding energy agreed with the reported data on ZnO:Co.^21,22 If Co existed in the metallic form, i.e., Co (0), the binding energy of Co 2p_3/2 should be 778.1 eV.²³ Hence, we concluded that Co in our sample was in the 2+ oxidation state. Moreover, it was also concluded that Co ions replaced Al ions in the AlN:Co products.

Fig. 1 (a) X-ray diffraction (XRD) of AlN:Co nanowires, (b–d) XPS spectra of binding energy of Co 2p, Al 2p, and N 1s region.

Scanning electron microscopy (SEM) images of the as-synthesized AlN:Co nanowires (Fig. 2a and b) showed that the product consisted of 1D structures with diameters in the order of 100–200 nm and lengths greatly exceeding 5 μm, that is, high aspect ratio nanowires. These materials have very promising applications in field emission devices due to their very low electron affinity. The morphology and structure of the AlN:Co nanowires were characterized in further detail using TEM and HRTEM. Fig. 2c shows a representative TEM image of one nanowire. The nanowire was very straight with a diameter of approximately 100 nm and a smooth surface, consistent with the SEM observations. Fig. 2d presents the corresponding HRTEM lattice image taken from the nanowire surface and obtained from the circular domains in Fig. 2c. The distance between the adjacent lattice planes was about 0.239 nm, which corresponded well with the d-spacing of (101) crystal planes of hexagonal wurtzite AlN, suggesting that the direction of growth for the nanowire is [101]. The inset in Fig. 2d shows a fast Fourier transform (FFT) image, which further demonstrated that the as-grown nanowires were single crystal and grew along the [101] direction.

Fig. 2 (a) and (b) low- and high- magnification SEM images of AlN:Co nanowires; (c) TEM image of the typical nanowires; (d) HRTEM lattice image of the surface of a nanowire, with its FFT pattern inserted.

The above experimental results indicated that the morphology of nanostructure was wire-like, straight and smooth. The nucleation and growth of a crystal is a very sensitive and complex process, depending not only on the intrinsic structure of the crystal but also on its surface defect structure and external factors such as supersaturation, structure of the mother phase, temperature, pressure, and impurities. Some experimental results showed that the appropriate reaction time and N₂ pressure played an important role in the nanowire formation. Due to low N₂ pressure (5–20 kPa) and the short reaction time (10–30 min), the nitrides synthesized in previous reports were mostly nanoparticles.^24,25 However, high N₂ pressure (35 kPa) and long reaction time (100 min) employed in this report produced nitrides with novel nanostructures, consisting of trunk and branch.^26,27 So the appropriate N₂ pressure (30 kPa) and reaction time (10 min) were chosen to ensure that small aggregates of AlN crystals lead to single nucleation of the AlN crystals and the growth of AlN:Co nanowires. In this study, no metal-particle catalyst or template was used so that the formation of the nanowires should be related to the vapor-solid (VS) mechanism and the intrinsic properties of AlN:Co, such as anisotropy of the hexagonal structure. The thermal decomposition of N₂ and the evaporation of Al and Co column at 3000–5000 °C resulted in the formation of Al, Co, and N vapors. These vapors were transported by N₂ carrier gas to a zone with an appropriate temperature for the growth of AlN:Co. During this step, the heat convection and temperature gradient, produced by DC arc plasma, set up a chemical-vapor transport and condensation process. As the reaction progressed, the vapors were constantly transported in a circular flow. The flashing evaporation of Al and Co and sufficient N vapors led to an energetic growth of AlN:Co along the [101] direction.

Raman scattering is a useful tool for studying the optical and crystalline quality of the nanomaterials. The space group of the hexagonal wurtzite AlN crystal was P6₃mc, with all atoms occupying the C_3v sites. Six first-order Raman active modes, 2E₂, 1A₁ (transverse optical (TO)), 1A₁ (longitudinal optical (LO)), 1E₁ (TO), and 1E₁ (LO) may be present.²⁸ To explore the influence of the dopant, we tested AlN doped with different amounts of Co, as well as the undoped AlN product. The Co doping content was measured by EDS. Fig. 3 shows the Raman spectra of undoped AlN and AlN:Co nanowires with Co content 0.26 and 0.58 at%. Three distinct peaks centered at 604.9, 650.4, and 665.8 cm⁻¹ of the AlN:Co nanowires (0.58 at%) were correlated to the first-order vibrational modes of A₁ (TO), E₂ (high), and E₁ (TO), respectively. The low-intensity broad peak around 904.1 cm⁻¹ was assigned to the overlap of the modes A₁ (LO) and E₁ (LO). The A₁ (TO), E₂ (high), E₁ (TO) and A₁ (LO) + E₁ (LO) of the AlN:Co nanowires were slightly shifted compared with those of the undoped AlN nanowires. This may be attributed to the disorder of crystals due to the incorporation of Co. These results are in good agreement with those obtained for the Mg and Sc doped AlN nanostructures.^11,12 Moreover, Fig. 3 shows that the position of six first-order Raman active modes moved towards longer wavelengths with an increase of Co content, which proved that the doped Co can significantly affect the optical properties.

Fig. 3 The Raman spectrum of the as-grown AlN and AlN:Co nanowires.

In addition, the peaks of the Raman spectra agree well with the high quality single crystalline bulk AlN.²⁹ The E₂ (high) mode is used to analyse the stress state due to its high sensitivity to stress. Note that the sharp E₂ (high) peak indicated very good crystal quality of the AlN:Co nanowires. Moreover, the full width at half maximum (FWHM) of the A₁ (TO; 8.6 cm⁻¹) and E₂ (high; 6.6 cm⁻¹) modes of the AlN:Co (0.58 at%) nanowires were larger than that of AlN:Co (0.26 at%, 7.5 cm⁻¹ and 6.4 cm⁻¹) and AlN nanowires (6.5 cm⁻¹ and 5.6 cm⁻¹). It is well understood that the FWHM value in Raman spectra is inversely proportional to the phonon lifetime, which is strongly influenced by the impurities and defects. These FWHM values in the Raman spectra also suggested that Co doping brings in impurities and defects in AlN host structure. Moreover, the higher Co content can induce more serious lattice distortion. As a consequence, the Raman studies also indicated that the AlN:Co nanowires had very good crystal quality, and Co doping influenced the lattice structure.

Group III–V nitride compounds are promising materials for short-wavelength light emission devices. Fig. 4 presents the photoluminescence spectrum (PL) of the wire-shaped AlN:Co with Co content 0.26 and 0.58 at%, excited by 325 nm UV light from a He–Cd laser at room temperature. There appeared a strong and broad green emission, centered at about 558.0 nm and fitted by two peaks (524.6 nm and 583.9 nm), and two weak emissions around 450.2 nm and 657.5 nm. Three short wavelength emissions were attributed to the nitrogen vacancy and the radiative recombination of a photo- or electron-generated hole with an electron occupying the nitrogen vacancy.³⁰ This phenomenon was observed previously in AlN nanocone and a pine-shaped nanostructure.^31,32 The emission at 657.5 nm corresponded to E_v = 1.88 eV, on the basis of the equation E_v/eV = 1240/λ. According to Tanabe–Sugano ligand field theory,^33–35 Co²⁺ PL at ∼1.80 eV was assigned to the ⁴T₁ (P) → ²A₁ (F) transition of 3d electrons.^36,37 Thus, there should be Co²⁺ ions in the AlN products. Moreover, the emission intensity of the AlN with Co content 0.58 at% was much higher compared to that of AlN with 0.26 at% Co. This implies that the emission intensity was related to Co doping and content. The high Co content may have favored a higher level of nitrogen vacancies. The PL emission indicated that Co²⁺ ions existed in the AlN:Co nanowires and influenced the optical properties.

Fig. 4 Photoluminescence spectra (PL) of AlN:Co nanowires in absolute ethyl alcohol solution using excitation at 325 nm.

Recently, there has been intensive interest in studying AlN-based DMSs due to their potential applications in spintronics. However, the real origin of ferromagnetism is still confusing. To explore the magnetism of AlN:Co and avoid getting false-positive magnetism signals due to high-sensitivity devices,¹ we tested samples with different Co contents and undoped AlN using the vibrating sample magnetometer (VSM). The undoped AlN products showed an almost zero hysteresis loop in Fig. 5, implying that the undoped AlN was non-magnetic. However, the M–H curves of AlN:Co with different Co content at room temperature show a hysteresis loop as shown in Fig. 5. The magnified region between −2500 and 2500 Oe showed the hysteresis loop more clearly indicating ferromagnetic behavior, and achieved saturation magnetizations at high-field. The saturation magnetizations are 0.030 emu g⁻¹ and 0.078 emu g⁻¹, and the coercivities are 90 Oe and 242 Oe for the samples with Co contents of 0.26 and 0.58 at%. Moreover, the saturation magnetization increased with an increasing doped concentration. At 300 K, the saturation magnetization of AlN:Co nanowires was slightly larger than that of the nanorod arrays (0.04 emu g⁻¹),³⁸ which has larger Co content (0.9–2.1 at%). This phenomenon could be attributed to the geometric shape. Based on the previous characterization, the 1D-AlN:Co nanowires had high aspect ratio and high crystal quality with a smooth surface, and the cross section was circular. Although the AlN:Co nanorods were gathered into flower-shapes, where each branch was not a smooth surface and the cross section was a polygon.³⁸ The circular form in the AlN:Co nanowires means that they lack shape anisotropy, and they are made from an intrinsically isotropic material. It should therefore be easily to be magnetized and have larger saturation magnetization.

Fig. 5 Magnetization hysteresis loops of the undoped and Co doped AlN nanowires at room temperature.

In addition, there is a possibility that Co metal or other secondary phases could result in the ferromagnetic signal. However, the analysis of structure and composition can help to exclude the possibility that ferromagnetic behaviour comes from Co metal and other secondary phases. Thus, the observed ferromagnetism in the AlN:Co nanowires was related to the doped Co. Moreover, the doped ions concentration influenced the intensity of magnetization.^39–41 A nonhomogeneous distribution of doped Co favors the formation of areas enriched in dopant ions and some isolated dopant ions. In dopant enriched areas, a strong, short-range attraction between the magnetic dopants aggregated them into magnetic ion-clusters, which produced ferromagnetism. As the doped concentration increased, the increase of magnetic ion-clusters led to an enhanced ferromagnetism. In addition, our previous work reported that the DC arc discharge plasma method can dope Sr, Y or Mg into the AlN nanostructure, resulting in the ferromagnetic states due to the formation of Al vacancies.^11–13 The introduction of the Co into AlN significantly reduced the formation energy for Al vacancy, and caused the formation of many vacancies to enhance the ferromagnetism of the samples. The same phenomenon has been observed elsewhere.³⁸ Therefore, the observation of ferromagnetism in the AlN:Co nanowires may be due to the Co ion-clusters as well as the formation of Al vacancies.

Conclusions

The synthesis of 1D-AlN:Co with high aspect ratio and high crystal quality nanowires by DC arc discharge plasma method that was both catalyst-free and template-free was successfully demonstrated. Co²⁺ ions enhanced the optical properties and caused ferromagnetism. As a result, AlN:Co nanowires provided great opportunities for their easy transfer into any substrate for the development of new devices in spintronics.

Acknowledgements

This work was supported by the Education Department of Shaanxi Provincial Government Research Project (No. 14JK1131), the Dr Scientific Research Foundation of Shaanxi University of Technology (SLGQD13 (2)-11, SLGQD14-08), the National Science Foundation of China (No. 11074089, 51172087, NSAF. No. 10976011), and the Open Subject of State Key Laboratory of Superhard Materials, Jilin University.

Notes and references

1. S. A. Wolf, D. D. Awschalom, R. A. Buhrman, J. M. Daughton, S. von Molnar, M. L. Roukes, A. Y. Chtchelkanova and D. M. Treger, Science, 2001, 294, 1488.
2. H. Ohno, Science, 1998, 281, 951.
3. K. Ando, Science, 2006, 312, 1883.
4. Y. Ohno, D. K. Young, B. Beschoten, F. Matsukura, H. Ohno and D. D. Awschalom, Nature, 1999, 402, 790.
5. T. Dietl, H. Ohno, F. Matsukura, J. Cibert and D. Ferrand, Science, 2000, 287, 1019.
6. H. Munekat, H. Ohno, S. von Molnar, A. Segmüller, L. L. Chang and L. Esaki, Phys. Rev. Lett., 1989, 63, 1849.
7. A. Van Esch, L. Van Bockstal, J. De Boeck, G. Verbanck, A. S. van Steenbergen, P. J. Wellmann, B. Grietens, R. Bogaerts, F. Herlach and G. Borghs, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 56, 13103.
8. A. H. MacDonald, P. Schiffer and N. Samarth, Nat. Mater., 2005, 4, 195.
9. M. A. Mayerm, P. R. Stone, N. Miller, H. M. Smith, O. D. Dubon, E. E. Haller, K. M. Yu, W. Walukiewicz, X. Liu and J. K. Furdyna, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 045205.
10. Q. Liu, C. X. Liu, C. Xu, X. L. Qi and S. C. Zhang, Phys. Rev. Lett., 2009, 102, 156603.
11. W. W. Lei, D. Liu, X. Chen, P. W. Zhu, Q. L. Cui and G. T. Zou, J. Phys. Chem. C, 2010, 114, 15574.
12. Y. S. Xu, B. B. Yao, D. Liu, W. W. Lei, P. W. Zhu, Q. L. Cui and G. T. Zou, CrystEngComm, 2013, 15, 3271.
13. W. W. Lei, D. Liu, Y. M. Ma, X. Chen, F. B. Tian, P. W. Zhu, X. H. Chen, Q. L. Cui and G. T. Zou, Angew. Chem., Int. Ed., 2010, 49, 173.
14. Y. Li, J. Xiang, F. Qian, S. Gradecak, Y. Wu, H. Yan, D. A. Blom and C. M. Lieber, Nano Lett., 2006, 6, 1468.
15. X. B. Tang, G. M. Li and S. M. Zhou, Nano Lett., 2013, 13, 5046.
16. L. Rigutti, M. Tchernycheva, A. De Luna Bugallo, G. Jacopin, F. H. Julien, L. F. Zagonel, K. March, O. Stephan, M. Kociak and R. Songmuang, Nano Lett., 2010, 10, 2939.
17. J. A. Haber, O. C. Gibbons and W. E. Buhro, Chem. Mater., 1998, 10, 4062.
18. Y. J. Zhang, J. Liu, R. R. He, Q. Zhang, X. Z. Zhang and J. Zhu, Chem. Mater., 2001, 13, 3899.
19. T. Suehiro, J. Tatami, T. Meguro, K. Komeya and S. Matsuo, J. Am. Ceram. Soc., 2002, 85, 715.
20. P. Motamedi and K. Cadien, Appl. Surf. Sci., 2014, 315, 104.
21. P. Lommens, F. Loncke, P. F. Smet, F. Callens, D. Poelman, H. Vrielinck and Z. Hens, Chem. Mater., 2007, 19, 5576.
22. T. J. Chuang, C. R. Brundle and D. W. Rice, Surf. Sci., 1976, 59, 413.
23. B. J. Tan, K. J. Klabunde and P. M. A. Sherwood, J. Am. Chem. Soc., 1991, 113, 855.
24. W. W. Lei, D. Liu, L. H. Shen, J. Zhang, P. P. Zhu, Q. L. Cui and G. T. Zou, J. Cryst. Growth, 2007, 306, 413.
25. W. W. Lei, D. Liu, J. Zhang, L. H. Shen, X. X. Li, Q. L. Cui and G. T. Zou, J. Alloys Compd., 2008, 459, 298.
26. W. W. Lei, D. Liu, J. Zhang, P. W. Zhu, Q. L. Cui and G. T. Zou, Cryst. Growth Des., 2009, 9, 1489.
27. W. W. Lei, D. Liu, P. W. Zhu, X. H. Chen, J. Hao, Q. S. Wang, Q. L. Cui and G. T. Zou, CrystEngComm, 2010, 12, 511.
28. V. Y. Davydov, N. S. Averkiev, I. N. Goncharuk, D. K. Nelson, I. P. Nikitina, A. S. Polkovnikov, A. N. Smirnov, M. A. Jacobson and O. K. Semchinova, J. Appl. Phys., 1997, 82, 5097.
29. M. Kuball, J. M. Hayes, A. D. Prins, N. W. A. van Uden, D. J. Dunstan, Y. Shi and J. H. Edgar, Appl. Phys. Lett., 2001, 78, 724.
30. J. H. He, R. Yang, Y. L. Chueh, L. J. Chou, L. J. Chen and Z. L. Wang, Adv. Mater., 2006, 18, 650.
31. C. Liu, Z. Hu, Q. Wu, X. Z. Wang, Y. Chen, H. Sang, J. M. Zhu, S. Z. Deng and N. S. Xu, J. Am. Chem. Soc., 2005, 127, 1318.
32. W. W. Lei, D. Liu, P. W. Zhu, Q. S. Wang, G. Liang, J. Hao, X. H. Chen, Q. L. Cui and G. T. Zou, J. Phys. Chem. C, 2008, 112, 13353.
33. P. Lommens, P. F. Smet, C. de Mello Donega, A. Meijerink, L. Piraux, S. Michotte, S. Mátéfi-Tempfli, D. Poelman and Z. Hens, J. Lumin., 2006, 118, 245.
34. Y. Tanabe and S. Sugano, J. Phys. Soc. Jpn., 1954, 9, 753.
35. C. A. Johnson, T. C. Kaspar, S. A. Chambers, G. M. Salley and D. R. Gamelin, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 125206.
36. P. Koidl, Phys. Rev. B: Solid State, 1977, 15, 2493.
37. H. J. Schulz and M. Thiede, Phys. Rev. B: Condens. Matter Mater. Phys., 1987, 35, 18.
38. S. L. Yang, R. S. Gao, P. L. Niu and R. H. Yu, Appl. Phys. A, 2009, 96, 769.
39. T. T. Hu, M. Z. Zhang, S. D. Wang and Q. J. Shi, CrystEngComm, 2011, 13, 5646.
40. C. C. Huang, T. Y. Liu, C. H. Su, Y. W. Lo, J. H. Chen and C. S. Yeh, Chem. Mater., 2008, 20, 3840.
41. B. Martínez, F. Sandiumenge, L. Balcells, J. Arbiol, F. Sibieude and C. Monty, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 165202.

---