- Email: [email protected]
Improving exchange-coupling field in the same thickness of pinned magnetic layer
ARTICLE IN PRESS Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Improving exchange-coupling field in the same thickness of pinned magnetic layer X.L. Tang , H. Wu Zhang, H. Su, Y.L. Jing, Z. Yong Zhong State Key Laboratory of Electronic Thin Films and Integrated Devices, University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China
a r t i c l e in fo
abstract
Article history: Received 17 November 2008 Received in revised form 20 March 2009 Available online 3 May 2009
A conventional Ta/NiFe/Cu/NiFe/FeMn spin valve was prepared to investigate the exchange bias properties with the variations of deposition field. By enhancing the deposition magnetic fields from 50 to 650 Oe, increase of exchange bias fields at a given thickness of the pinned NiFe layer has been found in the spin valves. In this paper, we show that this increase is due to the change of magnetic moment distribution at the ferromagnetic and antiferromagnetic interface by comparison of measured results with the interfacial uncompensated model. Therefore, by enhancing deposition magnetic fields, a large exchange-coupling field can be achieved in relatively thicker magnetic films for application. & 2009 Elsevier B.V. All rights reserved.
PACS: 75.70.cn 75.75.+a 85.75.d Keywords: Exchange bias Spin valve
1. Introduction Exchange coupling at interface between a ferromagnetic (FM) layer and an antiferromagnetic (AFM) layer can cause exchange bias, which is often described as an in-plane unidirectional anisotropy. It has been known for more than 50 years [1,2] and extensively studied by both experiments and theoretical analysis [3–12]. However, its physical mechanism remains unappreciated. Considerable interest in FM/AFM exchange coupling has been revived because of its key role in giant magnetoresistance (GMR) spin valve (SV). For magnetic random access memory (MRAM) and magnetic sensor applications, a large exchange bias between the pinned FM layer and the AFM pinning layer promotes magnetic stability of the pinned layer, which is the key to production of reliable, highquality SV structure with increased sensitivity. It is well known that the pinning achieved in an AFM/FM bilayer is produced either by heating the sample to above the blocking temperature of the AFM, applying a magnetic field and then cooling to room temperature with the field on, or by applying a magnetic field during sample growth [13]. Using field-cooling process, a large exchange bias field Hex can be achieved in the sample comparing with field growth method [14]. However, field-cooling process may bring interface diffusion, oxidize and destroy the properties of the sample. Therefore, using field-cooling process to get large Corresponding author.
E-mail address: [email protected] (X.L. Tang). 0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2009.04.050
Hex is disadvantage in device preparing procession. Since the exchange bias field Hex is usually inversely proportional to the FM layer thickness [13,15], a more direct way to get large exchangecoupling field is based on decreasing the thickness of pinned FM film when applying a magnetic field during sample growth. However, thin films cannot supply enough spin polarization, which is responsible for the effect of decreasing GMR. Therefore, many researches on the enhancement of exchange bias for relatively thicker FM film have been conducted, such as changing seed layers, adding surfactant [16–18]. In addition, surveying most of the experimental results on the exchange coupling between FeMn and NiFe, the deposition magnetic fields for developing the necessary exchange bias are in a large area, from several oersted to several hundred oersted [3,19–21]. Little attention has been given to the magnitude of deposition magnetic fields. Therefore, we have studied in great details on exchange bias with a variation of magnetic field when applied during film deposition, and find a feasible way to achieve large exchange bias for application.
2. Experiment Multilayer thin films with a composition Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (t)/FeMn (15 nm) with t ¼ 4–10 nm were deposited by DC magnetron sputtering using LS500 automatic sputtering system at a base vacuum better than 7 108 Torr, and working pressure 5 104 Torr. Deposition rates of all the layers were around 0.1 nm/s and controlled by a quartz crystal
ARTICLE IN PRESS 2856
X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
deposition monitor. A magnetic field was applied during the film deposition to define the easy axis and the unidirectional axis of the film and changed from 50 to 650 Oe. The exchange bias field was obtained from the hysteresis loop measured by a BHV-525 vibrating sample magnetometer (VSM).
300
250
Hex (Oe)
3. Results and discussions The magnetization curves for a series of samples with four different deposition magnetic fields of 50, 200, 400 and 650 Oe are displayed in Fig. 1. In Fig. 1 the two parts of the hysteresis loop match well with the relative thickness and magnetizations of the two NiFe layers. Notice one interesting feature of the curves: Hex increased with increasing deposition magnetic field for a given NiFe layer thickness. Furthermore, the exchange-coupling fields and the deposition magnetic fields for the samples are also plotted as a function of pinned NiFe layer thickness in Fig. 2. It is obvious that Hex is proportional to 1/tFM (tFM: the thickness of pinned magnetic
4
6
8
10
pinned NiFe thickness (nm) Fig. 2. Exchange bias variation of the deposition fields as a function of pinned NiFe layer thickness.
tFM=6nm
1.0
0.5
M/Ms
M/Ms
150
50
0.5
0.0
0.0
-0.5
-0.5
-1.0
-1.0
-200
0
200 H(Oe)
400
-100
tFM=8nm
1.0
1.0
0
100 H(Oe)
200
tFM=10nm
0.5
M/Ms
0.5
M/Ms
200
100
tFM=4nm
1.0
50Oe 200Oe 400Oe 650Oe
0.0
0.0
-0.5
-0.5
-1.0
-1.0
-50
0
50
H(Oe)
100
150
-25
0
25
50
75
100
H(Oe)
Fig. 1. The magnetization hysteresis loops of Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (t)/FeMn (15 nm) multilayer at different deposition fields, for (a) t ¼ 4 nm, (b) t ¼ 6 nm, (c) t ¼ 8 nm and (d) t ¼ 10 nm.
ARTICLE IN PRESS X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
2857
a 1.0
Fig. 3. Schematic diagram shows the spin orientations at the FM/AFM interface.
M/Ms
0.5
0.0
-0.5
-1.0 -50
0
50 H (Oe)
-200
0
200 H (Oe)
100
150
b 1.0
0.5 M/Ms
layer) at a certain applying magnetic fields, which is consistent with other experiments and the theoretical analyses. However, it increases when a large deposition field is applied. In addition, the influence of deposition fields on Hex is more remarkable in thinner pinned magnetic films than in relatively thicker pinned magnetic films, which is consistent with the notion that the exchange coupling is an interfacial effect. It also indicates that the results may be related to deposition field effect on interface properties. The experimental results are very important for spintronic applications, because a large Hex achieved in relatively thicker pinned magnetic layer can offer enough spin polarization and ensure high GMR effect. Many theoretical models and experimental results assume that the exchange coupling is probably an interfacial effect and the change of interfacial magnetic properties is responsible for the effect on exchange coupling [22,23]. Therefore, the different exchange-coupling fields observed for different deposition fields for a given thickness of pinned NiFe layer could be due to the deposition fields change of distribution of magnetic moments at the FM/AFM interface. To properly understand the enhanced exchange bias in our samples, we introduce the interfacial uncompensated model: a few AFM atomic layers at the FM/AFM interface may have a parallel net spin ordering in the direction of the FM layer. This parallel net spins provide the net spin moments required for exchange bias, which is displayed in Fig. 3. In our samples, by enhancing the deposition field, the magnetic moments at the interface layer may be further turned parallel to the direction of deposition field, resulting in change of the distribution of the magnetic moments at the AFM/FM interface, and finally affecting the exchange-coupling field. To further prove it, we have tested the angular dependence of exchange coupling in Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (4 nm)/FeMn (15 nm) multilayer with 50 and 650 Oe deposition fields. The hysteresis loops, which are shown in Fig. 4, were measured in a VSM by physically rotating the sample about an axis perpendicular to the sample plane, and the applied field was at an angle y to the deposition field. In Fig. 4, the two series of hysteresis loops display different anisotropies, especially when the angles y are small. At y ¼ 01 and y ¼ 151, the hysteresis loops for the multilayer fabricated at 50 Oe deposition field are almost the same, but for the sample deposited at 650 Oe field, they are significantly different. It indicates that the easy axis for pinned layer fabricated under large field tends to be the direction of deposition field, but for small deposition field, the easy axis presents an angular deviation from the direction of deposition field. The different anisotropy illustrates that the magnetic moments distribution in the pinned layer is different. The different magnetic distribution of pinned layer can influence the interfacial moments. Since the exchange bias is sensitive to the distribution of interfacial moments, we consider that the film deposited under a large field may provide more net spins at the FM/AFM interface for
0.0
-0.5
-1.0 400
600
Fig. 4. Hysteresis loops of Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (4 nm)/FeMn (15 nm) multilayer at various angle y for different deposition fields of (a) 50 Oe and (b) 650 Oe.
achieving lager exchange bias. In addition, the exchange bias disappears at y ¼ 601 in the sample fabricated at 50 Oe, but it still exists for the sample fabricated at 650 Oe. Combining the measured results with the analyses above, it can be concluded that the magnetic moment distribution was changed by the different deposition fields at the interface, and is responsible for the different Hex achieved at a given NiFe layer thickness.
4. Conclusion In summary, the deposition field effects on the exchange bias fields have been systematically studied. At a given thickness of pinned magnetic layer, enhanced exchange bias fields can be achieved with larger deposition magnetic fields. Angular dependence of exchange-coupling analyses suggests that different deposition fields are playing a key role in affecting the anisotropic magnetic properties at ferromagnetic/antiferromagnetic interface, which has been shown clearly from the testing results. Therefore, by increasing deposition magnetic fields, a large exchange-coupling field can be achieved in relatively thick magnetic films.
ARTICLE IN PRESS 2858
X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
Acknowledgments This work was supported by the National Natural Science Foundation of China (nos. 60721001, 60771047 and 60801027), and National High Technology Research and Development Program of China (863 Program, no. 200901AAZ111). References
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]
[1] W.P. Meiklejohn, C.P. Bean, Phys. Rev. 102 (1956) 1413. [2] W.P. Meiklejohn, C.P. Bean, Phys. Rev. 105 (1957) 904. [3] V.K. Sankaranarayanan, S.M. Yoon, D.Y. Kim, C.O. Kim, C.G. Kim, J. Appl. Phys. 96 (2004) 7428. [4] G.H. Yu, M.H. Li, F.W. Zhu, H.W. Jiang, W.Y. Lai, C.L. Chai, Appl. Phys. Lett. 82 (2003) 94. [5] J.H. Lee, H.D. Jeong, C.S. Yoon, C.K. Kim, B.G. Park, T.D. Lee, J. Appl. Phys. 91 (2002) 1431. [6] M. Pakala, Y. Huai, G. Anderson, L. Miloslavsky, J. Appl. Phys. 87 (2000) 6653.
[18] [19] [20] [21] [22] [23]
D. Mauri, H.C. Siegmann, P.S. Bagus, E. Kay, J. Appl. Phys. 62 (1987) 3047. A.P. Malozemoff, Phys. Rev. B 35 (1987) 3679. N.C. Koon, Phys. Rev. Lett. 78 (1997) 4865. T.C. Schulthess, W.H. Butler, Phys. Rev. Lett. 81 (1998) 4516. Y. Sakurai, H. Fujiware, J. Appl. Phys. 93 (2003) 8615. J. Spray, U. Nowak, J. Phys. D: Appl. Phys. 39 (2006) 4536. J. Nogue´s, I.K. Schuller, J. Magn. Magn. Mater. 192 (1999) 203. H. Sang, Y.W. Du, C.L. Chien, J. Appl. Phys. 85 (1999) 4931. K. Liu, C.L. Chien, IEEE Trans. Magn. 34 (1998) 1021. H.W. Jiang, M.H. Li, G.H. Yu, F.W. Zhu, J.W. Cai, W.Y. Lai, J. Magn. Magn. Mater. 264 (2003) 1. L. Ritchie, X.Y. Liu, S. Ingvarsson, G. Xiao, J. Du, J.Q. Xiao, J. Magn. Magn. Mater. 247 (2002) 187. M. Fujita, K. Yamano, A. Maeda, T. Tanuma, M. Kume, J. Appl. Phys. 81 (1997) 4909. G.C. Han, P. Luo, J.J. Qiu, K.B. Li, Y.H. Wu, Appl. Phys. A 75 (2002) 655. H. Sang, Y.W. Du, C.L. Chien, J. Appl. Phys. 85 (1999) 4931. K. Noma, H. Kanai, J. Appl. Phys. 95 (2004) 6669. C. Tsang, K. Lee, J. Appl. Phys. 52 (1998) 2471. S. Bae, J.H. Judy, W.F. Egelhoff Jr., J. Chen, J. Appl. Phys. 87 (2000) 6650.
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Improving exchange-coupling field in the same thickness of pinned magnetic layer X.L. Tang , H. Wu Zhang, H. Su, Y.L. Jing, Z. Yong Zhong State Key Laboratory of Electronic Thin Films and Integrated Devices, University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China
a r t i c l e in fo
abstract
Article history: Received 17 November 2008 Received in revised form 20 March 2009 Available online 3 May 2009
A conventional Ta/NiFe/Cu/NiFe/FeMn spin valve was prepared to investigate the exchange bias properties with the variations of deposition field. By enhancing the deposition magnetic fields from 50 to 650 Oe, increase of exchange bias fields at a given thickness of the pinned NiFe layer has been found in the spin valves. In this paper, we show that this increase is due to the change of magnetic moment distribution at the ferromagnetic and antiferromagnetic interface by comparison of measured results with the interfacial uncompensated model. Therefore, by enhancing deposition magnetic fields, a large exchange-coupling field can be achieved in relatively thicker magnetic films for application. & 2009 Elsevier B.V. All rights reserved.
PACS: 75.70.cn 75.75.+a 85.75.d Keywords: Exchange bias Spin valve
1. Introduction Exchange coupling at interface between a ferromagnetic (FM) layer and an antiferromagnetic (AFM) layer can cause exchange bias, which is often described as an in-plane unidirectional anisotropy. It has been known for more than 50 years [1,2] and extensively studied by both experiments and theoretical analysis [3–12]. However, its physical mechanism remains unappreciated. Considerable interest in FM/AFM exchange coupling has been revived because of its key role in giant magnetoresistance (GMR) spin valve (SV). For magnetic random access memory (MRAM) and magnetic sensor applications, a large exchange bias between the pinned FM layer and the AFM pinning layer promotes magnetic stability of the pinned layer, which is the key to production of reliable, highquality SV structure with increased sensitivity. It is well known that the pinning achieved in an AFM/FM bilayer is produced either by heating the sample to above the blocking temperature of the AFM, applying a magnetic field and then cooling to room temperature with the field on, or by applying a magnetic field during sample growth [13]. Using field-cooling process, a large exchange bias field Hex can be achieved in the sample comparing with field growth method [14]. However, field-cooling process may bring interface diffusion, oxidize and destroy the properties of the sample. Therefore, using field-cooling process to get large Corresponding author.
E-mail address: [email protected] (X.L. Tang). 0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2009.04.050
Hex is disadvantage in device preparing procession. Since the exchange bias field Hex is usually inversely proportional to the FM layer thickness [13,15], a more direct way to get large exchangecoupling field is based on decreasing the thickness of pinned FM film when applying a magnetic field during sample growth. However, thin films cannot supply enough spin polarization, which is responsible for the effect of decreasing GMR. Therefore, many researches on the enhancement of exchange bias for relatively thicker FM film have been conducted, such as changing seed layers, adding surfactant [16–18]. In addition, surveying most of the experimental results on the exchange coupling between FeMn and NiFe, the deposition magnetic fields for developing the necessary exchange bias are in a large area, from several oersted to several hundred oersted [3,19–21]. Little attention has been given to the magnitude of deposition magnetic fields. Therefore, we have studied in great details on exchange bias with a variation of magnetic field when applied during film deposition, and find a feasible way to achieve large exchange bias for application.
2. Experiment Multilayer thin films with a composition Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (t)/FeMn (15 nm) with t ¼ 4–10 nm were deposited by DC magnetron sputtering using LS500 automatic sputtering system at a base vacuum better than 7 108 Torr, and working pressure 5 104 Torr. Deposition rates of all the layers were around 0.1 nm/s and controlled by a quartz crystal
ARTICLE IN PRESS 2856
X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
deposition monitor. A magnetic field was applied during the film deposition to define the easy axis and the unidirectional axis of the film and changed from 50 to 650 Oe. The exchange bias field was obtained from the hysteresis loop measured by a BHV-525 vibrating sample magnetometer (VSM).
300
250
Hex (Oe)
3. Results and discussions The magnetization curves for a series of samples with four different deposition magnetic fields of 50, 200, 400 and 650 Oe are displayed in Fig. 1. In Fig. 1 the two parts of the hysteresis loop match well with the relative thickness and magnetizations of the two NiFe layers. Notice one interesting feature of the curves: Hex increased with increasing deposition magnetic field for a given NiFe layer thickness. Furthermore, the exchange-coupling fields and the deposition magnetic fields for the samples are also plotted as a function of pinned NiFe layer thickness in Fig. 2. It is obvious that Hex is proportional to 1/tFM (tFM: the thickness of pinned magnetic
4
6
8
10
pinned NiFe thickness (nm) Fig. 2. Exchange bias variation of the deposition fields as a function of pinned NiFe layer thickness.
tFM=6nm
1.0
0.5
M/Ms
M/Ms
150
50
0.5
0.0
0.0
-0.5
-0.5
-1.0
-1.0
-200
0
200 H(Oe)
400
-100
tFM=8nm
1.0
1.0
0
100 H(Oe)
200
tFM=10nm
0.5
M/Ms
0.5
M/Ms
200
100
tFM=4nm
1.0
50Oe 200Oe 400Oe 650Oe
0.0
0.0
-0.5
-0.5
-1.0
-1.0
-50
0
50
H(Oe)
100
150
-25
0
25
50
75
100
H(Oe)
Fig. 1. The magnetization hysteresis loops of Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (t)/FeMn (15 nm) multilayer at different deposition fields, for (a) t ¼ 4 nm, (b) t ¼ 6 nm, (c) t ¼ 8 nm and (d) t ¼ 10 nm.
ARTICLE IN PRESS X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
2857
a 1.0
Fig. 3. Schematic diagram shows the spin orientations at the FM/AFM interface.
M/Ms
0.5
0.0
-0.5
-1.0 -50
0
50 H (Oe)
-200
0
200 H (Oe)
100
150
b 1.0
0.5 M/Ms
layer) at a certain applying magnetic fields, which is consistent with other experiments and the theoretical analyses. However, it increases when a large deposition field is applied. In addition, the influence of deposition fields on Hex is more remarkable in thinner pinned magnetic films than in relatively thicker pinned magnetic films, which is consistent with the notion that the exchange coupling is an interfacial effect. It also indicates that the results may be related to deposition field effect on interface properties. The experimental results are very important for spintronic applications, because a large Hex achieved in relatively thicker pinned magnetic layer can offer enough spin polarization and ensure high GMR effect. Many theoretical models and experimental results assume that the exchange coupling is probably an interfacial effect and the change of interfacial magnetic properties is responsible for the effect on exchange coupling [22,23]. Therefore, the different exchange-coupling fields observed for different deposition fields for a given thickness of pinned NiFe layer could be due to the deposition fields change of distribution of magnetic moments at the FM/AFM interface. To properly understand the enhanced exchange bias in our samples, we introduce the interfacial uncompensated model: a few AFM atomic layers at the FM/AFM interface may have a parallel net spin ordering in the direction of the FM layer. This parallel net spins provide the net spin moments required for exchange bias, which is displayed in Fig. 3. In our samples, by enhancing the deposition field, the magnetic moments at the interface layer may be further turned parallel to the direction of deposition field, resulting in change of the distribution of the magnetic moments at the AFM/FM interface, and finally affecting the exchange-coupling field. To further prove it, we have tested the angular dependence of exchange coupling in Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (4 nm)/FeMn (15 nm) multilayer with 50 and 650 Oe deposition fields. The hysteresis loops, which are shown in Fig. 4, were measured in a VSM by physically rotating the sample about an axis perpendicular to the sample plane, and the applied field was at an angle y to the deposition field. In Fig. 4, the two series of hysteresis loops display different anisotropies, especially when the angles y are small. At y ¼ 01 and y ¼ 151, the hysteresis loops for the multilayer fabricated at 50 Oe deposition field are almost the same, but for the sample deposited at 650 Oe field, they are significantly different. It indicates that the easy axis for pinned layer fabricated under large field tends to be the direction of deposition field, but for small deposition field, the easy axis presents an angular deviation from the direction of deposition field. The different anisotropy illustrates that the magnetic moments distribution in the pinned layer is different. The different magnetic distribution of pinned layer can influence the interfacial moments. Since the exchange bias is sensitive to the distribution of interfacial moments, we consider that the film deposited under a large field may provide more net spins at the FM/AFM interface for
0.0
-0.5
-1.0 400
600
Fig. 4. Hysteresis loops of Ta (5 nm)/NiFe (10 nm)/Cu (4 nm)/NiFe (4 nm)/FeMn (15 nm) multilayer at various angle y for different deposition fields of (a) 50 Oe and (b) 650 Oe.
achieving lager exchange bias. In addition, the exchange bias disappears at y ¼ 601 in the sample fabricated at 50 Oe, but it still exists for the sample fabricated at 650 Oe. Combining the measured results with the analyses above, it can be concluded that the magnetic moment distribution was changed by the different deposition fields at the interface, and is responsible for the different Hex achieved at a given NiFe layer thickness.
4. Conclusion In summary, the deposition field effects on the exchange bias fields have been systematically studied. At a given thickness of pinned magnetic layer, enhanced exchange bias fields can be achieved with larger deposition magnetic fields. Angular dependence of exchange-coupling analyses suggests that different deposition fields are playing a key role in affecting the anisotropic magnetic properties at ferromagnetic/antiferromagnetic interface, which has been shown clearly from the testing results. Therefore, by increasing deposition magnetic fields, a large exchange-coupling field can be achieved in relatively thick magnetic films.
ARTICLE IN PRESS 2858
X.L. Tang et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2855–2858
Acknowledgments This work was supported by the National Natural Science Foundation of China (nos. 60721001, 60771047 and 60801027), and National High Technology Research and Development Program of China (863 Program, no. 200901AAZ111). References
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]
[1] W.P. Meiklejohn, C.P. Bean, Phys. Rev. 102 (1956) 1413. [2] W.P. Meiklejohn, C.P. Bean, Phys. Rev. 105 (1957) 904. [3] V.K. Sankaranarayanan, S.M. Yoon, D.Y. Kim, C.O. Kim, C.G. Kim, J. Appl. Phys. 96 (2004) 7428. [4] G.H. Yu, M.H. Li, F.W. Zhu, H.W. Jiang, W.Y. Lai, C.L. Chai, Appl. Phys. Lett. 82 (2003) 94. [5] J.H. Lee, H.D. Jeong, C.S. Yoon, C.K. Kim, B.G. Park, T.D. Lee, J. Appl. Phys. 91 (2002) 1431. [6] M. Pakala, Y. Huai, G. Anderson, L. Miloslavsky, J. Appl. Phys. 87 (2000) 6653.
[18] [19] [20] [21] [22] [23]
D. Mauri, H.C. Siegmann, P.S. Bagus, E. Kay, J. Appl. Phys. 62 (1987) 3047. A.P. Malozemoff, Phys. Rev. B 35 (1987) 3679. N.C. Koon, Phys. Rev. Lett. 78 (1997) 4865. T.C. Schulthess, W.H. Butler, Phys. Rev. Lett. 81 (1998) 4516. Y. Sakurai, H. Fujiware, J. Appl. Phys. 93 (2003) 8615. J. Spray, U. Nowak, J. Phys. D: Appl. Phys. 39 (2006) 4536. J. Nogue´s, I.K. Schuller, J. Magn. Magn. Mater. 192 (1999) 203. H. Sang, Y.W. Du, C.L. Chien, J. Appl. Phys. 85 (1999) 4931. K. Liu, C.L. Chien, IEEE Trans. Magn. 34 (1998) 1021. H.W. Jiang, M.H. Li, G.H. Yu, F.W. Zhu, J.W. Cai, W.Y. Lai, J. Magn. Magn. Mater. 264 (2003) 1. L. Ritchie, X.Y. Liu, S. Ingvarsson, G. Xiao, J. Du, J.Q. Xiao, J. Magn. Magn. Mater. 247 (2002) 187. M. Fujita, K. Yamano, A. Maeda, T. Tanuma, M. Kume, J. Appl. Phys. 81 (1997) 4909. G.C. Han, P. Luo, J.J. Qiu, K.B. Li, Y.H. Wu, Appl. Phys. A 75 (2002) 655. H. Sang, Y.W. Du, C.L. Chien, J. Appl. Phys. 85 (1999) 4931. K. Noma, H. Kanai, J. Appl. Phys. 95 (2004) 6669. C. Tsang, K. Lee, J. Appl. Phys. 52 (1998) 2471. S. Bae, J.H. Judy, W.F. Egelhoff Jr., J. Chen, J. Appl. Phys. 87 (2000) 6650.