Magnetic anisotropic properties of Pd/Co/Pd trilayer films studied by X-ray absorption spectroscopy and magnetic circular dichroism

Abstract

We study the magnetic anisotropic properties of as-grown and annealed Pd/Co/Pd trilayer films based on their atomic and electronic structures using extended X-ray absorption fine structure (EXAFS), X-ray absorption near edge structure (XANES) spectroscopy and magnetic circular dichroism (XMCD) measurements. The annealed film exhibits interesting perpendicular magnetic anisotropy (PMA) whereas the as-grown film exhibits in-plane anisotropy. Cross-sectional transmission electron microscopic analysis together with the Co K-edge EXAFS results confirm the formation of an ordered-alloy CoPd phase in the annealed film, whereas the as-grown film has an hcp Co-like phase. Co L_3,2-edge XMCD measurements reveal an enhanced ratio of the Co 3d orbital to spin moments of the annealed film providing evidence of the observed PMA upon annealing.

---

Introduction

Magnetic thin films with perpendicular magnetic anisotropy (PMA) have great research interest due to their applications in featured high-density non-volatile memory and magnetic memory devices with the possibility of current-induced magnetization switching.^1,2 In these films, they are able to facilitate faster and smaller data-storage magnetic bits, compared with the normal in-plane ones.³ In a magnetic material, the spin–orbit interaction of the electrons plays a major role in determining the anisotropy behavior.⁴ To attain perpendicular anisotropy, a number of material systems have been explored by various research groups and these materials include rare-earth/transition-metal alloys, L1₀-ordered (Co, Fe)–Pt alloys and Co/(Pd, Pt) multilayers.^2–7 In such alloys, the electronic hybridization of the valence levels of a 3d transition metal, with large magnetic moments but relatively small spin–orbit coupling (Fe or Co), with the valence levels of 4d, 5d transition metals (Pd or Pt) with small magnetic moments but a large spin–orbit coupling, alters the spin–orbit interaction which could enhance the PMA.^8,9 Among the alloys, the Co/Pd composites exhibit attractive magnetic properties and have been studied extensively by many research groups.^10–13

Yakushiji et al. have observed an enhanced PMA with a large anisotropy energy in [Co/Pt]_n and [Co/Pd]_n superlattice films consisting of 0.14–0.20 nm thick Co and Pt(Pd) layers.¹⁴ These [Co/Pt(or Pd)]_n superlattice films had a post-annealing stability of up to 370 °C due to the formation of the ordered-alloy-like structure, in which the inter-diffusion of Co and Pt(Pd) atoms was suppressed. Carrey et al. have studied the perpendicular anisotropy in Co/Pd multilayers with the structure: [(∼0.3 nm Co co-deposited with y nm Pd)/(2.0 − y) nm Pd]₁₅, with y = 0, 0.1, 0.3, 0.5, 1, 1.5 and 2.0.¹³ The perpendicular anisotropy is almost unaffected by alloying up to y = 0.5 nm. With further increase of y, the magnetic coercivity, saturation field and squareness are decreased exponentially. Most of Co/Pd multilayers exhibit PMA when the Co thickness is very less. Specifically, Co film of thickness ∼0.8 nm is highly suggested for enhanced perpendicular magnetization.^15,16 This is mainly due to tensile strain at the interface due to the lattice mismatch with the Pd layer. On the other hand, The Pd layer thickness also plays a significant role in tuning the spin orientation in the CoPd composites.^16–19 Kim et al. have studied the magnetization property of Pd(50 Å)/CoPd(20 Å)/Pd(t_Pd) film as a function of Pd layer thickness (t_Pd).¹⁸ These authors observed that the magnetization direction smoothly changes from in-plane orientation (at t_Pd = 0 Å) through canted out of plane (0 < t_Pd < 30 Å) to perpendicular (30 < t_Pd < 60 Å). From these results, it is clear that a well controlled thickness of Co, Pd films and multilayered structures are essential to achieve the desired PMA in CoPd composites. Yabuhara et al. have also observed that in-plane anisotropy in CoPd films prepared by rf magnetron sputtering.²⁰ These authors concluded that the in-plane magnetic anisotropy of CoPd films is due to disordered CoPd alloy crystal. Hence, a detailed analysis is necessary to clarify the correlation between the structure and magnetic anisotropy in this alloy system. Many research groups have observed that high-temperature annealing enhanced PMA,^9,21 however, the effect of temperature on alloy formation, particularly the relationships among the atomic and electronic structures, magnetic moments and PMA of magnetic ion such as Co are not fully understood. To solve these problems, a simple method for synthesizing CoPd alloy with better understanding of factors that enhancing the PMA is necessary.

The present work explores a simple method to synthesis CoPd alloy nanostructure, which shows strong PMA behavior, by in situ annealing of Pd(4 nm)/Co(5 nm)/Pd(4 nm) trilayer film. The film thicknesses are considerable large in comparison with the earlier reports.^13–17,22 It is observed that the high temperature annealing rigorously alters the arrangement of atoms in the layer structure. As the spin–orbit interaction plays a major role in PMA, the quantification of spin, orbital moments and electronic and atomic structural properties are essential for better understanding of anisotropy behavior of a material. We have studied the atomic and electronic structural properties of the as-grown and annealed films using synchrotron radiation based X-ray absorption measurements, which are highly proven techniques to study the local atomic, electronic structure and magnetic properties of a material.^8,23 The observed PMA behavior of annealed film has been discussed in detail with the help of these sensitive techniques.

Experimental details

Investigation of the Pd/Co/Pd films was performed at room temperature at National Synchrotron Radiation Research Center (NSRRC) in Hsinchu, Taiwan. The electronic structure of the Pd/Co/Pd films was measured by Pd L₃-edge X-ray absorption near-edge structure (XANES) at beamline BL-16A in total fluorescence yield (TFY) mode with an angle of 45° between the incident X-rays and the surface normal. The local atomic structure of the Co atoms in the films were measured by extended X-ray absorption fine structure (EXAFS) at beamline BL-17C. For probing the magnetic properties of samples, Co L_3,2-edge XANES and X-ray magnetic circular dichroism (XMCD) were performed at beamline BL-11A at NSRRC as well. The spectra were recorded in total electron yield (TEY) mode. For the XMCD measurement, the incident angle of X-rays was fixed at 30° from the sample normal, and a magnetic field of 1 Tesla was applied parallel and antiparallel to the sample normal.

The Pd/Co/Pd thin trilayers were deposited on Al₂O₃ (0001) substrate by similar molecular beam epitaxy (MBE) method under ultra high vacuum condition with a base pressure better than 3 × 10⁻⁹ torr. The deposition rate was maintained ∼1 nm min⁻¹ for uniform deposition of films. With identical condition, two Pd(4 nm)/Co(5 nm)/Pd(4 nm) samples were prepared at room temperature. After deposition, one of them was in situ annealed at 700 K for 250 seconds. In order to understand the mechanism of the PMA behavior, Pd-rich (approximately Pd₇₅Co₂₅) and Co-rich (approximately Pd₂₅Co₇₅) alloys were also prepared by similar MBE using co-evaporation from individual Co and Pd targets with controlled rate of deposition to achieve the desired composition. The cross-sectional view of the deposited films were analyzed using transmission electron microscopy (TEM) equipped with energy dispersive spectroscopy (EDS) facility. Please note that the spatial resolution of TEM-EDS was determined by both size of the incident e-beam and its relative spreading in the sample. In our TEM measurement, the e-beam size is ∼1 nm and its spreading in the sample is in few nm. Therefore, the spatial resolution of TEM-EDS analysis is within few nanometers scale. The magnetic properties of all the films were ex situ investigated using magneto-optical Kerr effect (MOKE). The MOKE measurement system was composed of four electromagnet poles installed in a vacuum chamber, so that the MOKE hysteresis loops were measured in both perpendicular and in-plane geometries by simply switching the polarity of electromagnet without changing the optical setup.¹¹

Results and discussion

Fig. 1(a) and (b) show the in-plane and out-of-plane MOKE magnetic hysteresis loops of as-grown and annealed Pd/Co/Pd films, respectively. The insets show the corresponding cross-sectional TEM images and EDS scan across the film interface. It is clear that in as-grown film the Co and Pd films have individual layer structures, whereas in annealed film they are mixed with each other. The elemental line profile based on the EDS scan across the film also confirms a mixing of the Co and Pd layers upon annealing. This is in contrast with earlier reports of ultrathin Co/Pd multilayers where the inter-diffusion was suppressed.¹⁴ The as-grown Pd/Co/Pd film exhibits in-plane magnetization [Fig. 1(a)], whereas annealed film exhibits a strong PMA with a weak in-plane magnetization [Fig. 1(b)]. Generally in a magnetic thin film, the magnetic dipolar interaction causes shape anisotropy where the magnetization lies within the plane of the film, while the spin–orbit interaction in certain crystalline structures could cause magnetocrystalline anisotropy which is perpendicular to the plane.^23,24 The effective anisotropy is the resultant of these two competitive interactions in a magnetic thin film. In the present work, the in-plane anisotropy observed in as-grown film could be simply attributed to the dominated shape anisotropy owing to a thick layer of Co film (∼5 nm).⁴ In case of annealed film, the Pd layers have strong effect on the magnetization orientation. This result may be correlated with a formation of CoPd alloy upon annealing. To further elucidate the correlation between the magnetic anisotropy and alloy structure, MOKE measurements have also been performed on Pd-rich and Co-rich alloy films and are shown in Fig. 1(c) and (d), respectively. The inset shows the corresponding cross-sectional TEM images and the EDS analysis. The thickness of Pd-rich alloy (∼16.1 nm) is close to the annealed film (∼14.4 nm), whereas the thickness of Co-rich alloy is large (∼38.3 nm). The MOKE measurements reveal that the Co-rich alloy has in-plane magnetization while the Pd-rich alloy has in-plane as well as perpendicular magnetizations. Furthermore, the PMA of annealed film was also found to differ from the disordered CoPd alloy which exhibits in-plane anisotropy.²⁰ These findings suggest that the magnetic anisotropy of CoPd film is highly sensitive to the alloy composition and structural orderings.

Fig. 1 In-plane and out-of-plane MOKE magnetic hysteresis loops of (a) as-grown and (b) annealed Pd/Co/Pd films. Insets show the corresponding cross-sectional TEM images and EDS scans across the interface of film. (c) and (d) are the in-plane and out-of-plane MOKE magnetic hysteresis loops of Pd-rich alloy and Co-rich alloys, respectively. Insets present the corresponding cross-sectional TEM images and the EDS analysis.

In order to observe the local atomic structure of the synthesized films, EXAFS measurements have been carried out at Co K-edge. Fig. 2(a) represents the Fourier transformed (FT) spectra of the Co K-edge EXAFS of as-grown and annealed Pd/Co/Pd films along with the references. The inset shows the corresponding EXAFS k²χ data. For the as-grown film, the spectral features and the main feature ∼2.1 Å (without phase correction), which is correspond to the nearest-neighbor (NN) Co–Co bond distance, are same as reference Co foil, indicating no random mixing between Co and Pd.²⁵ However, the peak intensity is slightly reduced because of slight disorder in the film due to thickness effect. In contrast to the as-grown film, the annealed film exhibits two main FT features in the region ∼1.6–3.0 Å, which corresponds to NN Co–Co and Co–Pd bond distances and are characteristic of the Co–Pd alloy phase.^12,17,18,22 In Fig. 2(a), it is noticed that the spectral feature of the annealed film is relatively close to the Pd-rich alloy than the Co-rich alloy films. Such a finding can be explained by the uniform mixing of Co atoms in the surrounding Pd atoms, as evidenced by TEM-EDS analysis. By comparing the intensity of the EXAFS features, it is inferred that in annealed film, the NN Co–Co bond is more in comparison with Pd-rich alloy while the Co–Pd bond is slightly lower than in Pd-rich alloy. The TEM-EDS and EXAFS analyses reveal that the high temperature annealing results an alloying CoPd phase in Pd/Co/Pd trilayer film.

Fig. 2 (a) Fourier-transformed Co K-edge EXAFS spectra of as-grown and annealed Pd/Co/Pd films and reference samples. Inset shows corresponding EXAFS k²χ data, where k is from 3.5 to 11 Å⁻¹. (b) Normalized Pd L₃-edge XANES spectra. Inset magnifies region of feature A.

To find the origin of the magnetism in the synthesized CoPd composites, electronic structural studies have been performed using XANES at Pd L_3,2-edge and the normalized spectra are shown in Fig. 2(b). All XANES spectra exhibit a broad white line (WL) feature A at ∼3174 eV, which is attributed primarily to electron transitions from Pd 2p orbitals to unoccupied 4d conduction band states. High intensity and a slight energy shift of the WL feature of the studied samples relative that to the reference Pd foil are observed. The inset in Fig. 2(b) shows the magnified region of feature A, and reveals that the as-grown film and Pd-rich alloy have similar WL features, whereas the annealed film and Co-rich alloy have blue shifted WL features. According to literature,²⁶ this blue-shift of the WL feature is associated with charge transfer from Co to Pd atoms, which is caused by strong hybridization of Pd 4d and Co 3d orbitals and the different electro-negativities of the atoms.²⁷ This process causes higher energy of X-ray to be required to excite the electrons from Pd 2p orbitals to unoccupied 4d states. The transfer of charge from Co to Pd atoms promotes the generation of a magnetic moment in the CoPd alloy and acts as the origin of the magnetism in such alloy system.²⁸ It should be noted that in Co-rich alloy, the shift is large, this is because of the Pd atom is surrounded by more Co atoms which can provide more charge to the Pd site.

To understand the anisotropy behavior of the synthesized films, XMCD measurements at Co L_3,2-edge have been performed. The extracted orbital (m_orb) and spin (m_spin) magnetic moments are the essential quantities to understand the anisotropy properties of the present systems. Fig. 3 displays the Co L_3,2-edge XANES spectra of the as-grown and annealed films along with the reference samples, with the photo-helicity of incident X-rays parallel (μ₊) and anti-parallel (μ₋) to the direction of magnetization. All the films exhibit two broad features in the ranges 777–785 eV and 792–800 eV, which are attributed to Co 2p_3/2 → 3d_5/2 and 2p_1/2 → 3d_3/2 dipole transitions, respectively. Fig. 3 also presents Co L_3,2-edge XMCD spectra (bottom panels), (μ₊ − μ₋)/(μ₊ + μ₋), in which the general line shapes are consistent with those reported elsewhere.²⁹ All the XMCD curves, integrated over the respective L_3,2-edges, are related to the m_orb and m_spin via the sum rules,^29–31

andwhere n_3d is the 3d electron occupation number of the specific transition metal atom. p and q are the integrated values of the XMCD curves correspond to the L₃ and L₂ transitions, respectively. r is the integrated value of the sum (μ₊ + μ₋) from XAS spectra, as shown in Fig. 3. 〈T_z〉 is the expectation value of the magnetic dipole operator, and 〈S_z〉 equals half of m_spin in Hartree atomic units. Generally, the magnetic dipolar term 〈T_z〉 is less than 10% of the value of 〈S_z〉. Thus, the ratio of m_orb and m_spin values can be derived from the XMCD data by neglecting these terms in eqn (2), and using the value n_3d = 7.51 for Co.²⁹ Similarly, the magnetic moments of the Co foil, Co-rich and Pd-rich alloys were also calculated.

Fig. 3 The Co L_3,2-edge XAS spectra of the as-grown and annealed films and Co foil, Co-rich alloy and Pd-rich alloy with the sample magnetization parallel (μ₊) and antiparallel (μ₋) to direction of incident X-ray. Dashed lines represent integrated XAS spectra. Bottom panels show XMCD and integrated XMCD spectra.

Table 1 summarizes the m_orb and m_spin moments and their ratio of all samples measured along out-of-plane direction. The estimated m_orb and m_spin values of the standard Co foil are in agreement with the reported value of hcp Co.²⁹ The estimated values of m_orb and m_spin of as-grown film are larger than that of reference Co foil. This may due to the strain generated by the lattice mismatch at Pd–Co interfaces.^4,27 Interestingly, the m_spin value of Co atoms in all films (except the as grown film) is relatively lower than that of the Co foil. In particular, the value is drastically reduced in Pd-rich alloy. This effect could be attributable to the formation of CoPd alloy. Generally in a magnetic material, the value of magnetic moment is associated with the degree of long-range magnetic order, which is determined by a competition between exchange coupling among electrons and thermal fluctuation. Consequently, the m_spin would exhibit a higher value at low temperature owing to a less thermal fluctuation, but decrease rapidly as it close to the Curie temperature (T_C). For a bulk Co, the T_C is ∼1288 K,⁴ and therefore sizeable magnetic moments can be simply obtained at room temperature. As the Co concentration decreases, the exchange interaction among Co atoms in the alloy becomes weaker, and the T_C is reduced significantly from 1288 K to 1200 K for 75% Co alloy to 295 K for 8.9% Co alloy. For the present Pd-rich alloy film, the T_C is estimated to be ∼560 K.³² Such a low T_C could lead to a weak long-range magnetic ordering of the Co atoms at room temperature. Consequently, the m_spin of Co atoms in Pd-rich alloy is drastically reduced compared to the reference Co foil. Following the above discussion, the reduction of m_spin and m_orb values of annealed film, in comparison with as-grown film, could be attributed to a formation of CoPd alloy. On the other hand, although the atomic structure of annealed film is similar to the Pd-rich alloy, the values of the m_orb and m_spin are much higher than those of Pd-rich alloy. This can be due to the formation of CoPd alloy with high order structure of annealed film, which generally posses a higher moment than that of the non-crystallographic morphologies.³³

Table 1 Orbital (m_orb), spin (m_spin) moments and ratio of m_orb/m_spin derived from the XMCD curves and sum-rules (in μ_B per atom unit)

Sample	morb	mspin	morb/mspin
hcp-Co^29	0.15	1.50	0.1
Co foil	0.20 ± 0.05	1.50 ± 0.05	0.13 ± 0.05
As-grown film	0.37 ± 0.05	1.95 ± 0.05	0.19 ± 0.05
Annealed film	0.33 ± 0.05	1.48 ± 0.05	0.23 ± 0.05
Co-rich alloy	0.16 ± 0.05	1.37 ± 0.05	0.12 ± 0.05
Pd-rich alloy	0.10 ± 0.05	0.37 ± 0.05	0.27 ± 0.05

---

According to previous reports, the m_orb/m_spin ratio measured along out-of-plane direction could roughly be considered as an indication for the presence of the perpendicular K_cry in a magnetic film.^34–37 Nevertheless, the presence of PMA is not only determined by the K_cry but also the total magnetic moment (i.e. m_orb + m_spin) of the film. This is because, the total magnetic moments would contribute the shape anisotropy (K_shape = −2πM², where M is saturation magnetization) which prefers in-plane direction. Thus, only if the magnitude of positive K_cry can overcome the negative in-plane shape anisotropy energy, the magnetization of thin film would tend to align out-of-plane direction.²⁴ In the present work, as shown in Table 1, it is clear that the m_orb/m_spin ratio of the annealed Pd/Co/Pd film is higher than the as-grown film. This gives clear evidence to the enhanced perpendicular crystalline anisotropy of Pd/Co/Pd film upon annealing.

Conclusion

We have demonstrated a simple method to synthesis CoPd alloy, by in situ annealing of Pd(4 nm)/Co(5 nm)/Pd(4 nm) trilayer film, which shows strong perpendicular magnetic anisotropy. In this method, any traditional conditions, like ultra thin Co layer or multilayer structured films, have not been adopted to achieve the desired PMA behavior. TEM-EDS together with the EXAFS analyses confirm the formation of high structural ordered CoPd alloy in the annealed film. Pd L₃-edge XANES analysis shows that there is an enhanced charge transfer from Co to Pd site in the annealed film in comparison with the as-grown film, and the charge transfer induces the magnetic moment in the film. The enhanced Co 3d orbital/spin moment ratio plays major role in the perpendicular anisotropy observed in the annealed film.

Acknowledgements

The authors (BYW and WFP) are grateful to the Ministry of Science and Technology of Taiwan for financial support in this research under Grant no. NSC 102-2112-M-018-006-MY3 and 102-2112-M-032-007-MY3.

References

1. S. Mangin, D. Ravelosona, J. A. Katine, M. J. Carey, B. D. Terris and E. E. Fullerton, Nat. Mater., 2006, 5, 210.
2. M. Jamali, K. Narayanapillai, X. Qiu, L. M. Loong, A. Manchon and H. Yang, Phys. Rev. Lett., 2013, 111, 246602.
3. A. D. Kent, B. Ozyilmaz and E. del Barco, Appl. Phys. Lett., 2004, 84, 3897.
4. J. Stöhr and H. C. Siegmann, Magnetism: From fundamentals to Nanoscale Dynamics, Springer, illustrated edition, 12 Sep 2006.
5. K. Mizunuma, S. Ikeda, J. H. Park, H. Yamamoto, H. Gan, K. Miura, H. Hasegawa, J. Hayakawa, F. Matsukura and H. Ohno, Appl. Phys. Lett., 2009, 95, 232516.
6. B. Carvello, C. Ducruet, B. Rodmacq, S. Auffret, E. Gautier, G. Gaudin and B. Dieny, Appl. Phys. Lett., 2008, 92, 102508.
7. G. Kim, Y. Sakuraba, M. Oogane, Y. Ando and T. Miyazaki, Appl. Phys. Lett., 2008, 92, 172502.
8. C. Boeglin, E. Beaurepaire, V. Halte, V. L. Flores, C. Stamm, N. Pontius, H. A. Dürr and J. Y. Bigot, Nature, 2010, 465, 458.
9. T. Das, P. D. Kulkarni, S. C. Purandare, H. C. Barshilia, S. Bhattacharyya and P. Chowdhury, Sci. Rep., 2014, 4, 5328.
10. S. Tsunashima, K. Nagase, K. Nakamura and S. Uchiyama, IEEE Trans. Magn., 1989, 25, 3761.
11. W. C. Lin, C. J. Tsai, B. Y. Wang, C. H. Kao and W. F. Pong, Appl. Phys. Lett., 2013, 102, 252404.
12. S. K. Kim, V. A. Chernov, J. B. Kortright and Y. M. Koo, Appl. Phys. Lett., 1997, 71, 7.
13. J. Carrey, A. E. Berkowitz Jr, W. F. Egelhoff and D. J. Smith, Appl. Phys. Lett., 2003, 83, 5259.
14. K. Yakushiji, T. Saruya, H. Kubota, A. Fukushima, T. Nagahama, S. Yuasa and K. Ando, Appl. Phys. Lett., 2010, 97, 232508.
15. P. F. Carcia, A. D. Meinhaldt and A. Suna, Appl. Phys. Lett., 1985, 47, 178.
16. F. J. A. den Broeder, H. C. Donkersloot, H. J. G. Draaisma and W. J. M. de Jonge, J. Appl. Phys., 1987, 61, 4317.
17. S. K. Kim and S. C. Shin, J. Appl. Phys., 2001, 89, 3055.
18. S. K. Kim, J. W. Lee, J. R. Jeong, J. Kim and S. C. Shin, Appl. Phys. Lett., 2001, 79, 1652.
19. D. G. Stinson and S. C. Shin, J. Appl. Phys., 1990, 67, 4459.
20. O. Yabuhara, M. Ohtake, K. Tobari, T. Nishiyama, F. Kirino and M. Futamoto, Thin Solid Films, 2011, 519, 8359.
21. M. Gottwald, K. Lee, J. J. Kan, B. Ocker, J. Wrona, S. Tibus, J. Langer, S. H. Kang and E. E. Fullerton, Appl. Phys. Lett., 2013, 102, 052405.
22. S. K. Kim, Y. M. Koo, V. A. Chernov, J. B. Kortright and S. C. Shin, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 62, 3025.
23. D. Weller, J. Stöhr, R. Nakajima, A. Carl, M. G. Samant, C. Chappert, R. Megy, P. Beauvillain, P. Veillet and G. A. Held, Phys. Rev. Lett., 1995, 75, 3752.
24. L. Néel, J. Phys. Radium, 1954, 15, 225.
25. J. Miyawaki, D. Matsumura, A. Nojima, T. Yokoyama and T. Ohta, Surf. Sci., 2007, 601, 95.
26. T. K. Sham, Phys. Rev. B: Condens. Matter Mater. Phys., 1985, 31, 1903.
27. N. Nakajima, T. Koide, T. Shidara, H. Miyauchi, H. Fukutani, A. Fujimori, K. Iio, T. Katayama, M. Nývlt and Y. Suzuki, Phys. Rev. Lett., 1998, 81, 5229.
28. S. Barman, D. G. Kanhere and G. P. Das, J. Phys.: Condens. Matter, 2009, 21, 396001.
29. C. T. Chen, Y. U. Idzerda, H.-J. Lin, N. V. Smith, G. Meigs, E. Chaban, G. H. Ho, E. Pellegrin and F. Sette, Phys. Rev. Lett., 1995, 75, 152.
30. B. T. Thole, P. Carra, F. Sette and G. van der Laan, Phys. Rev. Lett., 1992, 68, 1943.
31. P. Carra, B. T. Thole, M. Altarelli and X. D. Wang, Phys. Rev. Lett., 1993, 70, 694.
32. D. M. S. Bagguley, W. A. Crossley and J. Liesegang, Proc. Phys. Soc., London, 1967, 90, 1047.
33. M. E. Gruner and A. Dannenberg, J. Magn. Magn. Mater., 2009, 321, 861.
34. D. Weller, Y. Wu, J. Stöhr, M. G. Samant, B. D. Hermsmeier and C. Chappert, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49, 12888.
35. E. C. Stoner and E. P. Wohlfarth, Philos. Trans. R. Soc., A, 1948, 240, 599.
36. P. Bruno, Phys. Rev. B: Condens. Matter Mater. Phys., 1989, 39, 865.
37. B. Y. Wang, J. Y. Hong, K. H. Ou Yang, Y. L. Chan, D. H. Wei, H. J. Lin and M. T. Lin, Phys. Rev. Lett., 2013, 110, 117203; Y. Matsuo, J. Phys. Soc. Jpn., 1972, 32, 972.

---