ARTICLE IN PRESS

Journal of Magnetism and Magnetic Materials 309 (2007) 132–138

www.elsevier.com/locate/jmmm

Structural and magnetic properties of Co2MnSn films

and Co2MnSn/Cr multilayers
W. Zhanga,b,, N. Jikoc, T. Okunoc,1, K. Mibua,2, K. Yoshimurab
a
Research Center for Low Temperature and Materials Sciences, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan
b
Department of Chemistry, Graduate School of Science, Kyoto University, Sakyoku, Kyoto 606-8502, Japan
c
Institute for Chemical Research, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan
Received 1 March 2006; received in revised form 26 May 2006
Available online 12 July 2006

Abstract

Thin films of Co2MnSn were prepared on MgO(0 0 1) substrates with a Cr-buffer layer from three elemental targets by ultrahigh-
vacuum alternate deposition. These films showed good crystallinity, and the epitaxial relations were found to be MgO(0 0 1)//Cr(0 0 1)//
Co2MnSn(0 0 1) and MgO[1 0 0]//Cr[1 1 0]//Co2MnSn[1 1 0]. [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 multilayers were also prepared with high
crystallographic quality. The magnetic properties of Co2MnSn films were studied by magnetization measurements and 119Sn Mössbauer
spectroscopy. Hyperfine fields observed by Mössbauer effect measurements showed that the degree of short range order around Sn in the
samples grown at 200
C was as good as that in bulk samples.
r 2006 Elsevier B.V. All rights reserved.

PACS: 73.61.At; 75.50.Cc; 75.70.i; 76.80.+y

Keywords: Heusler alloy; Half-metallic ferromagnets; Ferromagnetic thin film; Mössbauer effect

1. Introduction TMR value has been reported in heterostructures of

Since the discoveries of giant magnetoresistance (GMR)
and tunneling magnetoresistance (TMR) in layered struc-
tures of magnetic metals and non-magnetic materials [1–3],
a new field of spin electronics has emerged up aiming at
their applications in magnetic information storage [4]. The
increased interest in this field has intensified research on
half-metallic materials, in which a band gap opens at the
Fermi level for one spin direction, leading to 100% spin
polarized conduction electrons [5]. In recent years, high
Corresponding author. Department of Engineering Physics, Electro-
nics and Mechanics, Graduate School of Engineering, Nagoya Institute of
Technology, Gokiso-cho, Showa-ku, Nagoya, Aichi, 466-8555, Japan.
Tel.: +81 527357906; fax: +81 527357904.
E-mail addresses: wzhang@nitech.ac.jp, wzhang@kuchem.kyoto-u.ac.jp
(W. Zhang).
1
Present address: Sumitomo Electric Industries, Ltd., Itami, Hyogo 664-
0016, Japan.
2
Present address: Graduate School of Engineering, Nagoya Institute of
Technology, Showa-ku, Nagoya 466-8555, Japan.
Heusler alloys Co2 Cr1x Fex Al [6] and Co2MnAl [7]. One
of the issues in this field is to prepare half-metallic
ferromagnets with high crystallographic quality.
Heusler alloy Co2MnSn attracts much attention because
it is expected to be half-metallic ferromagnet at low
temperatures and to remain highly spin polarized at room
temperature [8]. Bulk Co2MnSn crystallizes in an L21
structure (space group Fm3̄mÞ, which consists of four
interpenetrating face-centered cubic (fcc) sublattices [9].
The thin film of Co2MnSn has been grown by sputtering at
470
C [10]. However, TMR and GMR junctions using
Co2MnSn have not yet reported. The optimization of the
thin film preparation is just in the beginning stage. The
lattice of Co2MnSn matches well with that of Cr. It is
known that Cr grows epitaxially on MgO(0 0 1) with the
structural relations MgO(0 0 1)//Cr(0 0 1) in the growth
direction and MgO[1 0 0]//Cr[1 1 0] in the film plane [11].
Therefore, it is expected that epitaxial MgO/Cr/
Co2MnSn thin films can also be prepared by e-gun
deposition.

0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2006.06.023
 ARTICLE IN PRESS
W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138 133

To address these challenges, in this paper we present an
investigation on Co2MnSn films deposited from Co, Mn,
and Sn metallic sources alternately on a MgO(0 0 1)
substrate with a Cr-buffer layer at 200
C and room
temperature using e-gun deposition in an ultrahigh-
vacuum system (109 Torr). We also tried to prepare the
multilayers [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 and tested a
possible appearance of the GMR effect in these structures.
The experimental methods are presented in Section 2. The
magnetic properties of Co2MnSn thin films and Co2MnSn/
Cr multilayers are reported in Section 3. The results are
summarized in Section 4.
2. Experimental methods
Four Co2MnSn films shown in Table 1, which are
denoted as CMS1, CMS2, CMS3, and CMS4, respectively,
were grown on MgO(0 0 1) substrates (18  18 mm2 ) using
ultrahigh-vacuum deposition technique with e-gun heating.
Chemically etched MgO single crystalline (0 0 1) substrates
were first baked in a vacuum for 1 h at 400
C in order to
remove the surface hydroxide layer. A Cr buffer layer with
5.0 nm thick was deposited at 200
C on the MgO(0 0 1)
substrate with a deposition rate of 1.2 nm/min, prior to the
deposition of the Co2MnSn film in order to relieve the
strain resulting from the MgO substrate. Afterwards the
temperature was set at 200
C or room temperature for
the film growth. The three alloy components Co, Mn, and
Sn were evaporated from three e-guns alternately. For
CMS1 and CMS3 samples, Co was first deposited for one
atomic layer (0.2 nm), secondly Mn and thirdly Sn was
deposited for half an atomic layer (0.1 nm), and the process
was repeated for the designed cycles as shown in Table 1.
For CMS2 and CMS4, Sn was secondly deposited and Mn
Table 1
Designed nominal structures of Co2MnSn films
Sample Designed structure (nm)
CMS1 MgO/Cr(5.0)/[Co(0.2)/Mn(0.1)/Sn(0.1 )]100/Co(0.2)
CMS2 MgO/Cr(5.0)/[Co(0.2)/Sn(0.1)/Mn(0.1)]100/Co(0.2)
CMS3 MgO/Cr(5.0)/[Co(0.2)/Mn(0.1)/Sn(0.1)]100/Co(0.2)
CMS4 MgO/Cr(5.0)/[Co(0.2)/Sn(0.1)/Mn(0.1)]100/Co(0.2)
Substrate temperature (T sub ) was set at 200
C when CMS1 and CMS2
were deposited, whereas, CMS3 and CMS4 were deposited at room
temperature (RT).
thirdly. In order to prepare each layer with an accuracy of
0.1 nm, the deposition rate was set around 0.6 nm/min. The
vacuum during the deposition was in the 109 -Torr range.
We also prepared multilayers including the Co2MnSn
Heusler alloy as shown in Table 2. Although the four samples
have the same nominal structure of [Co2MnSn(2.2 nm)/
Cr(0.9 nm)]10, the interfacial atoms were designed to be
different by changing the deposition sequence of one or half
an atomic layer. In CMS5 and CMS6, the interfaces of
Co2MnSn with Cr are designed to be composed of Co atoms
and in CMS7 and CMS8 the interfaces of Co2MnSn are Mn
and Sn atoms. By comparing the results for these samples, we
can see if the interface atoms have a significant influence on
the structural and physical properties of [Co2MnSn(2.2 nm)/
Cr(0.9 nm)]10. The substrate temperature during the prepara-
tion of the multilayers was set at 200
C, to grow the film
expitaxially and avoid excessive interdiffusion at the
Co2MnSn/Cr interfaces.
The quality of the surface crystallographic structure was
monitored by in situ reflection high-energy electron
diffraction (RHEED) during the growth. The crystal-
lographic structure of the films was characterized by X-ray
diffraction (XRD) in high- and low-angle regions using Cu
K a radiation, with the scattering vector perpendicular to
the film plane. The composition of the samples was
checked by scanning electron microscopy (SEM) with
energy dispersive X-ray analysis (EDX).
The magnetization was measured by using a super-
conducting quantum interference device (SQUID) magnet-
ometer. Current-in-plane giant magnetoresistance (CIP-
GMR) measurements were carried out with a conventional
four-probe method. In order to check the local magnetic
properties 119Sn Mössbauer spectra were measured at
room temperature by means of conversion electron
Mössbauer spectroscopy using a Ca119mSnO3 g-ray source
and a He þ 1%ðCH3 Þ3 CH gas-flow counter. The direction
of the incident g-rays was set perpendicular to the film
plane. The Sn in the samples was enriched with 119Sn to
obtain sufficiently strong signals. The isomer shift (d) is
T sub ð
CÞ
expressed relative to that of CaSnO3.
200
200 3. Results and discussion
RT
RT
3.1. Co2MnSn films
The RHEED patterns for Co2MnSn films are shown in
Fig. 1. For CMS1 and CMS2, the streaked RHEED

Table 2
Designed nominal structures of [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 multilayers

Sample Designed structure (nm) T sub (
C)

CMS5 MgO/Cr(5.0)/{[Co(0.2)/Mn(0.1)/Sn(0.1)]5/Co(0.2)/Cr(0.9)}10 200

CMS6 MgO/Cr(5.0)/{[Co(0.2)/Sn(0.1)/Mn(0.1)]5/Co(0.2)/Cr(0.9)}10 200
CMS7 MgO/Cr(5.0)/{[Mn(0.1)/Sn(0.1)/Co(0.2)]5/Mn(0.1)/Sn(0.1)/Cr(0.9)}10 200
CMS8 MgO/Cr(5.0){[Sn(0.1)/Mn(0.1)/Co(0.2)]5/Sn(0.1)/Mn(0.1)/Cr(0.9)}10 200

The substrate temperature (T sub ) during the preparation of the multilayers was set at 200
C.
 ARTICLE IN PRESS
134 W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138
Fig. 1. RHEED patterns of Co2MnSn films deposited at 200
C (CMS1
and CMS2) and room temperature (CMS3 and CMS4), after the
deposition of the final Co (0.2 nm) layer.
patterns were observed during the growing process from
the beginning. For CMS3 and CMS4, the streaked patterns
were observed after 10 repetitions of the Co/Mn/Sn or Co/
Sn/Mn unit, indicating that the surface is rough at the
beginning stage of the growing process and then becomes a
little smoother as the film grows.
The low-angle XRD patterns for the samples grown at
200
C and at room temperature are shown in Fig. 2. The
flatness of the surface is proven by the appearance of
wiggles in the low-angle XRD patterns. The total thickness
of Cr/Co2MnSn ðttotal Þ is estimated from the wiggle
patterns in Fig. 2. If we assume that the thickness of
buffer tCr equals to 5.0 nm as designed, the thickness
of Co2MnSn films can be estimated as tCo2 MnSn ¼
ttotal  5:0 nm, which is shown in Table 3.
The high-angle XRD measurements for y22y scans with
the scattering vector normal to the film plane are shown in
Fig. 3. Bulk Co2MnSn has an L21 type structure, where 4a
and 4b sites are occupied by Mn and Sn, respectively, and
4c, 4d sites are occupied by Co atoms. In disordered
structures, ions occupy the sites other than those specified
for the ordered structure. When the disorder occurs
exclusively between Mn and Sn, the structure is alterna-
tively described as a B2-type structure. The observation of
the peaks at 2y ¼ 29:9
and 2y ¼ 62:1
indicates that
Co2MnSn with a B2 or L21 -type structure grows epitaxi-
ally with the [0 0 1] direction perpendicular to the film
plane. In principle we can qualitatively check the degree of
order by comparing L21 002 (or B2 001) and L21 004 (or
B2 002) peaks. When the perfect disorder occurs among
Co, Mn and Sn, the alloy becomes an A2-type structure,
and the L21 002 (B2 001) peak would disappear. It is
difficult to obtain the order parameter numerically, since
the intensities in the y22y scans are also influenced by the
107
106
Intensity (arbitrary units)
105
104
CMS4
103
CMS3
102
CMS2
101
100 CMS1
1.0 1.5 2.0 2.5 3.0 3.5 4.0
2ϑ (degrees)
Fig. 2. XRD patterns of Co2MnSn films deposited at 200
C (CMS1 and
CMS2) and room temperature (CMS3 and CMS4) in low-angle region.
The estimated thicknesses of the Co2MnSn films are shown in Table 3.
quality of the crystallographic orientation. The disappear-
ance of L21 006 peak for CMS3 and CMS4 indicates that
they have lower quality in crystallographic orientation.
When the degree of crystallographic perfection decreases,
the width in the rocking curve increases and the intensity in
the y  2y scan decreases for the higher order peak. For
CMS1 and CMS2, there seems to be a very small amount
of (1 1 0) oriented portion.
The values of saturated magnetic moment for
Co2MnSn films (M s ) determined from the magnetization
curves at room temperature (the in-plane magnetization
curve of sample CMS1 is shown in Fig. 4) are listed in
Table 3. The values of saturated magnetic moment for the
Co2MnSn films grown at 200
C were found to be about
5:0 mB /formula unit (f.u.). Errors may arise from inaccu-
racy of the estimated thickness of the films. The values of
magnetization agree well with the bulk value of 5.1 mB /f.u.
determined from magnetization and neutron diffraction
measurements by Webster [9]. The values of saturation
magnetization for Co2MnSn films grown at room tem-
perature were, on the other hand, lower than the bulk
value. This decrease in magnetization may be attributed to
the larger structural or site disorder in the films grown at
room temperature than in the films grown at 200
C.
Hysteresis loops at the low-field region were also measured
at room temperature with the field parallel to the sample
plane (CMS1’s curve is shown in the inset of Fig. 4). The
samples possess in-plane magnetization due to the shape
anisotropy. The obtained loops are square with a coercive
field of 1.5 mT and high remanent magnetization (90%) at
room temperature, demonstrating that the samples are soft
ferromagnets.
A magnetoresistance (MR) curve for Co2MnSn (CMS1)
measured with magnetic fields between 9 and 9 T applied
 ARTICLE IN PRESS
W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138 135

Table 3
Measured structural and physical parameters of Co2MnSn films

Sample t (nm) a (nm) M s (mB /f.u.) d (mm/s) Bhf (T)

CMS1 58.2 0.595 5.0 1.50 7.7

CMS2 60.5 0.594 4.9 1.50 7.8
CMS3 54.8 0.601 3.9 1.65 4.7
CMS4 55.4 0.602 3.2 1.65 4.7
Bulk 0.598a 5.1a 1.39b 6.9b

The thicknesses (t) and lattice parameters (a) of the Co2MnSn films were estimated from the X-ray diffraction patterns. The saturated magnetic moments
M s were calculated from the magnetization measurements. The isomer shift (d) and average hyperfine field ðBhf Þ were estimated by fitting the
Mössbauer spectra. These values were measured at room temperature. The values for bulk Co2MnSn are also shown for comparison.
a
Ref. [9].
b
Ref. [12].

5
CMS4

4 4

M (µB/f.u.)
M (µB/f.u.)
3
0

CMS3 2 -2

-4
1 -8 -6 -4 -2 0 2 4 6 8
(x10-3
Counts

B T)
CMS2 0
0.0 0.2 0.4 0.6 0.8 1.0
B (T)

Fig. 4. Magnetization curve of a Co2MnSn film (CMS1) measured at

room temperature. The magnetic field is applied in the plane of the sample.
MgO (002)
Inset: the hysteresis loop at the low-field region.
L21 004 MgO (004)
L21 002
B2 002
B2 001
CMS1
ðBhf Þ. The spectra are symmetric and, therefore, were fitted
L21 220
L21 006 assuming that the value of d is the same for all the
B2 110 Cr (002)
B2 003 components with different Bhf and that the effect due to the
electric quadrupole interaction is negligibly small. For the
fitting, the angle between the Bhf and normal direction to
the film plane was set to be 90
. The average Bhf and d
40 60 80 100 120
(relative to CaSnO3) for the Co2MnSn thin films are listed
2ϑ (degrees)
in Table 3. The d are found to be about 1.50 mm/s for the
Fig. 3. XRD patterns of Co2MnSn films deposited at 200
C (CMS1 and films grown at 200
C and 1.65 mm/s for the films grown at
CMS2) and room temperature (CMS3 and CMS4). room temperature. The difference in isomer shifts between
CMS1, 2 and CMS3, 4 may be related to the degree of the
disorder in these films. The distributions of Bhf for CMS1
in the film plane and parallel to the current is shown in and CMS2, indicate that the magnetic circumstances
Fig. 5. We found a 0:1% isotropic negative MR up to a around the Sn sites in the films grown at 200
C are similar
magnetic field of 9 T, as a characteristic feature of Co- to that in the bulk sample [12]. Although ideal ordered
based Heusler thin films [10]. The isotropic MR may come Heusler alloys have only one Sn site, even in bulk samples
from the forced MR effect. the spectra of Co2MnSn consist of two sets of sextets [12]
The 119Sn Mössbauer spectra collected at room tem- reflecting some imperfection of the alloy from an ideal
perature for the Co2MnSn films are shown in Fig. 6. Each ordered structure. Because of the disorder in the films
spectrum was fitted with six Lorentzians with 1 mm=s in samples, we also found that there are some imperfection in
linewidth, convoluted with a distribution of hyperfine field the Sn sites. The reduction of Bhf in CMS3 and CMS4 in
 ARTICLE IN PRESS
136 W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138

58.18

58.16
ρ (µΩ • cm)

58.14

58.12

58.10
-5 0 5
B (T)

Fig. 5. Resistivity versus magnetic field measured at 5 K for a thin film of

Co2MnSn (CMS1).
Fig. 7. RHEED patterns of [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 films de-
posited at 200
C, after the deposition of the final Cr (0.9 nm) layer.

CMS4 105

-1
1
104

-2
Counts (arbitrary units)

CMS3 2
Counts

103 -4 -3 3
4
CMS8
CMS2 102
CMS7

101 CMS6
CMS1

CMS5
-10 0 10 0 4 8 12 16
100
(a) Velocity (mm/s) (b) Bhf (T) 50 55 60 65 70 75
2θ (degrees)
Fig. 6. (a) Mössbauer spectra of Co2MnSn films measured at room
temperature and (b) the distributions of hyperfine field ðBhf Þ. Fig. 8. Patterns from out-of-plane X-ray scans of the multilayers
[Co2MnSn(2.2 nm)/Cr(0.9 nm)]10. The numbers in the figure denote the
order of satellites.

comparison with the bulk sample indicates that the films

grown at room temperature have more disorders around
the Sn sites. This tendency is consistent with the X-ray
patterns in Fig. 3.
3.2. [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 multilayers
The RHEED patterns of the multilayers [Co2MnSn
(2.2 nm)/Cr(0.9 nm)]10 are shown in Fig. 7. Clear streaked
RHEED patterns were also observed during the growing pro-
cess at 200
C for all [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10
multilayers. The out-of-plane X-ray scans of the multi-
layers [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 reveal that the
multilayers possess (0 0 1) epitaxy for Co2MnSn and Cr
layers. Fig. 8 shows the scans around the 004 fundamental
Bragg reflection of Co2MnSn/Cr. Besides the fundamental
Bragg peak from the average lattice spacing, the multi-
layers exhibit satellite peaks caused by the chemical
modulation. The satellites up to the order of l ¼ 4
can be resolved, indicating that the superstructures grow
well in the growth direction (0 0 1). The modulation
periods were estimated from the positions of the satellite
peaks using the relation L ¼ 2p=dQ, where dQ is the
average spacing between the Q values of the peaks
ðQ ¼ 4p sin y=lÞ.
 ARTICLE IN PRESS
W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138 137

In order to investigate CIP-GMR in [Co2MnSn(2.2 nm)/

Cr(0.9 nm)]10, we measured the MR with the magnetic field
CMS8
between 5 and 5 T applied in the film plane and parallel
to the current. All the samples in the Table 4, independent
of the deposition sequence, reveal very small values of the
MR effect. One possible reason why GMR effect is not
observed is that the spin polarization at Fermi level
CMS7 decreases at the interfaces. There is also a possibility that
Counts

the Cr-layer thickness is not optimum, and further

systematic investigation on Cr-layer thickness dependence
should be made in the future.
CMS6
4. Summary and conclusions

High-quality single-crystalline thin films of Co2MnSn

CMS5 have been prepared on MgO(0 0 1) substrates by ultra-
high-vacuum deposition from three elemental targets
-10 0 10 0 4 8 12
alternately on a Cr-buffer layer at room temperature and
(a) Velocity (mm/s) (b) B (T)
200
C. The epitaxial growth relation was found to be
Fig. 9. (a) Mössbauer spectra of the multilayers [Co2MnSn(2.2 nm)/ Co2MnSn(0 0 1)//Cr(0 0 1)//MgO(0 0 1) and Mg[1 0 0]//Cr
Cr(0.9 nm)]10 measured at room temperature and (b) the hyperfine field [1 1 0]//Co2MnSn[1 1 0]. Co2MnSn films grown at 200
C have
ðBhf Þ distribution.
saturation magnetization of 5 mB /f.u. at room tempera-
ture, which is comparable to the bulk sample. The Bhf for
the samples grown at 200
C is larger than that for the
Table 4 samples grown at room temperature, reflecting a better
Structure and physical properties of the four [Co2MnSn(2.2 nm)/ crystallographic quality.
Cr(0.9 nm)]10 multilayers; the multilayer periods (L) estimated from X-
ray diffraction patterns, the isomer shift (d) and the average hyperfine field The multilayers combining Co2MnSn alloys with Cr
ðBhf Þ estimated by fitting the Mössbauer spectra. These values are interlayer can also be grown with high crystallographic
measured at room temperature quality. The Mössbauer measurements reveal that the size
of Bhf is comparable with that for the thin films of
Sample L (nm) d (mm/s) Bhf (T)
Co2MnSn deposited at the same temperature, almost
CMS5 3.96 1.49 6.7 independent of the deposition sequence of the component
CMS6 3.74 1.42 5.5 metals in Co2MnSn. A CIP-GMR effect has not been
CMS7 3.85 1.53 7.4 observed in these [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 multi-
CMS8 3.94 1.49 7.0
layers at this stage.

The Mössbauer spectra measured at room temperature
for multilayers [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 are shown
in Fig. 9(a). The spectra were fitted with six-line
components with a distribution of Bhf . The fitting results
are shown by the solid line in Fig. 9(a), and the
distributions of Bhf are shown in Fig. 9(b). The average
Bhf for the [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10 multilayers
are listed in Table 4. We also found that there are
distribution in hyperfine field in [Co2MnSn(2.2 nm)/
Cr(0.9 nm)]10 as shown in Fig. 9(b). The Bhf distribution
for multilayers [Co2MnSn(2.2 nm)/Cr(0.9 nm)]10, in com-
parison with that for CMS1 and CMS2, indicates that the
magnetic circumstances around the Sn sites in the multi-
layers grown at 200
C are similar to those in the
Co2MnSn thin films grown at 200
C. There are slight
difference in Bhf distribution between the samples with
different deposition sequence, but it is not clear at the
present stage whether the difference is physically significant
or not.
Acknowledgments
We have benefited from discussion with Prof. T. Ono.
We also thank Dr. T. Ohkochi for simulating the thickness
of the films during this work. This study was supported by
Grants-in-Aid for COE Project ‘‘Elements Science’’ (Grant
no. 12CE2005) and Science Research on Priority Area,
‘‘Invention of anomalous quantum materials’’ (Grant no.
16076210), from the Ministry of Education, Science, Sports
and Culture of Japan.
References
[1] M.N. Baibich, J.M. Broto, A. Fert, F. Nguyen van Dau, F. Petroff,
P. Etienne, G. Creuzet, A. Friederich, J. Chazelas, Phys. Rev. Lett. 61
(21) (1988) 2472.
[2] T. Miyazaki, N. Tezuka, J. Magn. Magn. Mater. 139 (1995) L231.
[3] J.S. Moodera, L.R. Kinder, T.M. Wong, R. Meservey, Phys. Rev.
Lett. 74 (16) (1995) 3273.
[4] G.A. Prinz, Science 282 (1998) 1660.
[5] R.A. de Groot, F.M. Mueller, P.G. van Engen, K.H.J. Buschow,
Phys. Rev. Lett. 50 (25) (1983) 2024.
 ARTICLE IN PRESS
138 W. Zhang et al. / Journal of Magnetism and Magnetic Materials 309 (2007) 132–138

[6] K. Inomata, S. Okamura, R. Goto, N. Tezuka, Jpn. J. Appl. Phys. 42
(4B) (2003) L419.
[7] H. Kubota, J. Nakata, M. Oogane, Y. Ando, A. Sakuma,
T. Miyazaki, Jpn. J. Appl. Phys. 43 (7B) (2004) L984.
[8] J. Kübler, A.R. Williams, C.B. Sommers, Phys. Rev. B 28 (4) (1983) 1745.
[9] P.J. Webster, J. Phys. Chem. 32 (1971) 1221.
[10] U. Geiersbach, A. Bergmann, K. Westerholt, J. Magn. Magn. Mater.
240 (2002) 546.
[11] P. Etienne, J. Massies, S. Lequien, R. Cabanel, F. Petroff, J. Crystal
Growth 111 (1991) 1003.
[12] W. Zhang, N. Jiko, K. Mibu, K. Yoshimura, J. Phys.: Condens.
Matter 17 (2005) 6653.