Magnetic anisotropy in antiferromagnetic
layers affecting exchange bias of Ni-Fe/Mn-Ir
bilayers
著者
journal or
publication title
volume
number
page range
year
URL
角田 匡清
Journal of applied physics
87
9
4930-4932
2000
http://hdl.handle.net/10097/35368
doi: 10.1063/1.373206
JOURNAL OF APPLIED PHYSICS
VOLUME 87, NUMBER 9
1 MAY 2000
Magnetic anisotropy in antiferromagnetic layers affecting exchange bias
of Ni–FeÕMn–Ir bilayers
Kojiro Yagami,a) Masakiyo Tsunoda, and Migaku Takahashib)
Department of Electronic Engineering, Tohoku University, Sendai 980-8579, Japan
The magnetic anisotropy of antiferromagnetic layers (K AF) was estimated for Ni–Fe 50 Å/Mn–Ir
cr
⫽J Ks /K AF), where J Ks is the saturation value of the
d AF bilayers using Mauri’s method (d AF
unidirectional anisotropy constant (J K ). The critical thickness of the antiferromagnetic layers
cr
), at which J K took half the value of J Ks , was determined from the dependence of J K on d AF .
(d AF
cr
was found to be almost constant (35⫾2 Å) independent of J Ks . Thus, the relation of J Ks
The d AF
⬀K AF was derived, suggesting that the variation in J Ks is due to a change in the value of K AF . J Ks ,
however, was found to vary considerably for various Mn–Ir films possessing an almost identical Ir
content, and thus probably the same value of K AF . In addition, studies by x-ray diffraction,
transmission electron microscopy, and electron diffraction revealed that the change in J Ks was
independent of the microstructure and phase of the antiferromagnetic 共AF兲 Mn–Ir films, both of
which control K AF . Thus, J Ks was found to be independent of K AF contradicting the relation, J Ks
⬀K AF . This contradiction results from the assumption by Mauri that the coupling energy 共J兲 is equal
to J Ks even in the polycrystalline exchange-coupled bilayers. A model that took account of the
distribution of K AF axes of AF grains in the plane of the AF film successfully explained the behavior
of J K . J Ks was found to change independent of both J and K AF , and furthermore, it has been shown
that the dependence of J Ks on the sputtering conditions for Mn–Ir films is probably due to the
effective temperature of the films during deposition. © 2000 American Institute of Physics.
关S0021-8979共00兲35908-4兴
I. INTRODUCTION
in a range of values of J K . K AF was estimated from the
dependence of J K on d AF using Mauri’s method10 for polycrystalline exchange-coupled bilayers. The suitability of this
method was examined in connection with the microstructure
of the AF layers.
It is quite important to improve the reliability of the
dynamic response of spin valve heads to external magnetic
fields. One method is to enhance the exchange field 共⬀ the
unidirectional anisotropy constant, J K 兲 between the ferromagnetic 共F兲 layer and the adjacent antiferromagnetic 共AF兲
layer. Mn–Ir is a promising candidate for the AF layer of
spin valve heads, because it induces the strong exchange
field, even when the AF layer is very thin,1–5 and the corrosion resistance of Mn–Ir is acceptable.6 The present authors
have studied the relation between J K and sputtering conditions for Mn–Ir films using Ni–Fe/Mn–Ir bilayers.1 The
films were fabricated in an ultraclean atmosphere, because
impurities present in the atmosphere may change the microstructure or the magnetic properties of the films.7–9 As a
result, it was found that J K increased monotonously with
increasing Ir content and decreasing deposition rate.1 However, x-ray diffraction revealed no remarkable changes in the
microstructure of Mn–Ir films corresponding to the variation
in J K . 1 This work failed to address the origin of the change
in J K .
It is generally accepted that J K changes due to a change
in the magnetic anisotropy of the AF layer (K AF), the coupling energy between F/AF layers 共J兲, etc. Thus, the experimental evaluation of K AF should reveal further information
about the origin of the change in J K .
In the present study, Ni–Fe 50 Å/Mn–Ir d AF bilayers
were fabricated under typical sputtering conditions resulting
II. EXPERIMENTAL PROCEDURE
Multilayers with the substrate/Ta 50 Å/Ni–Fe 50
Å/Mn–Ir d AF /Ta 50 Å were fabricated on thermally oxidized Si wafers using the extremely clean sputtering
process.7,8 A magnetic field of 30 Oe was applied parallel to
the film plane during deposition. The substrates were displaced 200 mm from the 4 in. ␾ sputtering targets. Seventeen
at. % Ir–Mn alloy targets were used. Twelve Ir chips 共size
5⫻5 mm, purity 99.9%兲 were arranged on the target surface.
The sputtering conditions for the Ni–Fe films were fixed at
0.75 mTorr and 300 W for the gas pressure and the applied
power, respectively. Those for Mn–Ir films were changed as
shown in Table I. Five typical conditions 共a兲–共e兲 were chosen to obtain various saturation value of J K (J Ks ) for each of
the Ni–Fe/Mn–Ir bilayers.1 The resultant deposition rate and
Ir content of the Mn–Ir films is shown in Table I as well.
Conditions 共a兲 and 共b兲 were selected, as they had very different Ir content, but a similar deposition rate. On the other
hand, conditions 共c兲 and 共d兲 were selected, as they had a very
different deposition rate, but similar Ir content. J Ks reached
the maximum when the Mn–Ir film was fabricated under
condition 共e兲 in the examined range of the deposition rate
共0.04–2.6 Å/s兲 and Ir content 共8–32 at. %兲.1
a兲
Permanent address: MSCo., CNC, SONY Corp., Tagajyo 985-0842, Japan.
Author to whom correspondence should be addressed; electronic mail:
[email protected]
b兲
0021-8979/2000/87(9)/4930/3/$17.00
4930
© 2000 American Institute of Physics
Downloaded 28 Oct 2008 to 130.34.135.83. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
J. Appl. Phys., Vol. 87, No. 9, 1 May 2000
Yagami, Tsunoda, and Takahashi
4931
TABLE I. Sputtering conditions of Mn–Ir films.
共a兲
共b兲
共c兲
共d兲
共e兲
Gas pressure
共mTorr兲
Power
共W兲
Depo rate
共Å/s兲
Ir content
共at. %兲
1
10
1
10
20
50
100
200
30
50
0.49
0.47
2.59
0.04
0.06
9.0
26.9
15.6
15.4
25.1
The microstructural analysis and the identification of the
phase of the Mn–Ir films were made by electron diffraction,
transmission electron microscopy 共TEM兲, and x-ray diffraction 共XRD兲 with Co K ␣ source. M-H loops were measured
with a vibrating sample magnetometer 共VSM兲. The measurements were carried out on the as-deposited samples at room
temperature. J K was calculated using M s •d F •H ex , where
M s is the saturation magnetization of Ni–Fe films, d F is the
thickness of Ni–Fe films, and H ex is the exchange-coupling
field determined from the shift of the center of the M-H loop
along the field axis.
III. RESULTS AND DISCUSSION
cr
FIG. 2. The critical thickness of the AF layer, d AF
, as a function of the
saturation value of J K (J Ks ).
change in K AF induces a change in J Ks . The difference in
K AF for the Mn–Ir films fabricated under various sputtering
conditions is now considered. J Ks for conditions 共c兲 and 共d兲
differ by a factor of 2. According to the above result, it is
expected that K AF will also differ by this amount. The Ir
content of these two samples is, however, almost the same
共⬃15.5 at. %兲 as shown in Table I. Thus, other physical factors besides Ir content, such as a microstructure, were investigated for the variation of K AF .
A. Dependence of J K on thickness of Mn–Ir films
Figure 1 shows the dependence of J K on d AF . When the
Mn–Ir films were fabricated under condition 共e兲, J K rises
sharply at d AF⫽30 Å, and becomes almost constant 共0.14
erg/cm2兲 at d AF⭓50 Å. Using Mauri’s method,10 the critical
cr
) was determined to be 37 Å,
thickness of the AF layer (d AF
cr
where d AF is defined as the value of d AF at which J K took
cr
was determined
half of its saturation value. The value of d AF
in the same manner for the other films. Figure 2 shows the
cr
as a function of the saturation value of
change of d AF
s
J K (J K ). Here, J Ks is defined as the value of J K at d AF
⫽100 Å. The sputtering conditions for points 共a兲–共e兲 are
cr
has approximately the
listed in Table I. It is clear that d AF
same value, 35⫾2 Å, for conditions 共a兲–共e兲, even though the
deposition rate and Ir content of the Mn–Ir films fabricated
under these conditions is quite different. Taking account of
cr
⫽J Ks /K AF , derived by Mauri et al.,10,11 the
the formula, d AF
data in Fig. 2 imply that J Ks is proportional to K AF , namely a
FIG. 1. Dependence of J K on d AF for as-deposited Ni–Fe 50 Å/Mn–Ir d AF
bilayers fabricated under the various sputtering conditions for Mn–Ir films
listed in Table I.
B. Structural analysis
The microstructure and the phase of the AF layers were
investigated. The Mn–Ir 共111兲 planes were found to be
highly oriented parallel to the film plane.1 No remarkable
differences were detected between the microstructure of the
Mn–Ir films fabricated under conditions 共c兲 and 共d兲, by both
XRD analysis1 and observation of the film cross section by
TEM. K AF is expected to change when the ordered phase
Mn3Ir12 appears in disordered Mn–Ir. Electron diffraction
analysis was carried out for 200 Å Mn–Ir films fabricated on
Ta 50 Å/Ni–Fe 50 Å underlayers under conditions 共c兲 and
共d兲. Figure 3 shows a TEM plane view and an electron diffraction pattern for the Mn–Ir film fabricated under condition 共d兲. This figure reveals Mn–Ir grains 100–200 Å in
diameter. They have no preferred orientation in the film
plane 共the sheet texture is random in two dimensions兲. The
detected diffraction rings of Mn–Ir were identified as corresponding to the 共220兲, 共422兲, and 共440兲 planes, all of which
are fundamental. Superlattice lines 关e.g., 共110兲兴 were not detected. This indicates the absence of the ordered phase. The
same results were obtained for the Mn–Ir films fabricated
under condition 共c兲. These results imply that there is little
difference in the microstructure and phase of the Mn–Ir
films, upon which changes in K AF are expected to be dependent.
Consequently, it is assured that J Ks changed independently of K AF , contradicting the relation, J Ks ⬀K AF . This
contradiction is a result of the assumption by Mauri that
regarded J as being equal to J Ks . J⫽J Ks is true only when
both the magnetization of the F layer and that of the AF layer
are treated as being homogeneous. In the actual polycrystalline bilayers, however, it is impossible to regard the magnetization of the AF layer as such.
Downloaded 28 Oct 2008 to 130.34.135.83. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
4932
J. Appl. Phys., Vol. 87, No. 9, 1 May 2000
FIG. 3. 共a兲 TEM plane view and 共b兲 electron diffraction pattern of a 200 Å
thick Mn–Ir film fabricated on a Ta 50 Å/Ni–Fe 50 Å underlayer. Ni–Fe
and Mn–Ir 共111兲 planes are oriented highly parallel to the film plane. The
Mn–Ir film was fabricated under condition 共d兲. The incident electron beam
is perpendicular to the 共111兲 plane of the Mn–Ir film. All of the diffraction
rings for Mn–Ir are assigned to fundamental lines 共f 兲.
C. Effect of the distribution of K AF axes on J K
As shown in Fig. 3, the crystallographic axes of the AF
grains are randomly oriented in the film plane. It is followed
naturally that the magnetic anisotropy of the AF layer originates in the magnetocrystalline anisotropy of AF grains. The
AF layer in the polycrystalline F/AF bilayer, therefore,
should be treated as an aggregate of AF grains whose magnetic anisotropy axes are distributed in the film plane 共single
spin ensemble model13兲. Based up on this model, it is possible to calculate the dependence of J K on d AF . Figure 4
shows the change of J K /J as a function of K AFd AF /J for
various values of JS/kT, where k is Boltzmann’s constant, T
is the temperature at which the magnetic states of AF grains,
concerning their spin alignments, attained thermal equilibrium, S is the contact area of each AF grain with the F layer,
regarded as a constant for all AF grains. When JS/kT⫽⬁,
J K /J exhibits a sharp rise at K AFd AF /J⫽1, and is almost
saturated at K AFd AF /J⭓5. The saturation value, J Ks , decreases with decreasing JS/kT. This figure reveals that the
value of J Ks for the polycrystalline F/AF bilayers changes
independently of the microstructure of the AF layers or the
intrinsic physical quantities, such as K AF and J. Prior to comparing the results of this model to the experimental data
shown in Fig. 1, the meaning of the parameter JS/kT is
FIG. 4. J K calculated as a function of K AF d AF for various JS/kT, based
upon the single spin ensemble model 共see Ref. 13兲. All values are normalized to J.
Yagami, Tsunoda, and Takahashi
FIG. 5. Cross sectional TEM image. The 100-Å thick Mn–Ir film was
fabricated under condition 共a兲, growing epitaxially on the 50-Å thick Ni–Fe
film.
considered, provided that K AF and J are constant. S is also
constant for all the samples fabricated under the various
sputtering conditions. This is because the Mn–Ir grains grew
by epitaxy on the Ta/Ni–Fe underlayer which was fabricated
under the same conditions. The epitaxial growth of the
Mn–Ir films on Ni–Fe films was confirmed by TEM cross
sectional analysis 共Fig. 5兲. Consequently, the parameter
JS/kT only reflects a variation in T.
Regarding T as the effective temperature of the films
during the deposition process, it is possible to explain qualitatively the dependence of J Ks up on the sputtering conditions
for the Mn–Ir films: J Ks increased with increasing sputtering
gas pressure and decreasing applied power.1 Both of these
sputtering conditions decrease the kinetic energy of sputtering atoms incident on the substrate. As a result, the temperature of the films during the deposition process decreased,
thus increasing JS/kT. J Ks thereby increased.
In conclusion, the dependence of J K on d AF for polycrystalline Ni–Fe/Mn–Ir bilayers fabricated under various
sputtering conditions is explained qualitatively by a model
that takes into account the distribution of the K AF axes of the
AF grains within the film plane.
1
K. Yagami, M. Tsunoda, S. Sugano, and M. Takahashi, IEEE Trans.
Magn. 35, 3919 共1999兲. There were some mistakes about Ir content in this
paper. An error term is to be issued. In the present work, correct Ir content
is used.
2
K. Hoshino, R. Nakatani, H. Hoshiya, Y. Sugita, and S. Tsunashima, Jpn.
J. Appl. Phys. Part 1 35, 607 共1996兲.
3
H. N. Fuke, K. Saito, Y. Kamiguchi, H. Iwasaki, and M. Sahashi, J. Appl.
Phys. 81, 4004 共1997兲.
4
A. J. Devasahayam, P. J. Sides, and M. H. Kryder, J. Appl. Phys. 83, 7216
共1998兲.
5
A. J. Devasahayam and M. H. Kryder, IEEE Trans. Magn. 35, 649 共1999兲.
6
M. Lederman, IEEE Trans. Magn. 35, 794 共1999兲.
7
M. Takahashi, A. Kikuchi, and S. Kawakita, IEEE Trans. Magn. 33, 2938
共1997兲.
8
K. Okuyama, T. Shimatsu, S. Kuji, and M. Takahashi, IEEE Trans. Magn.
31, 3838 共1995兲.
9
M. Tsunoda, K. Uneyama, T. Suzuki, K. Yagami, and M. Takahashi, J.
Appl. Phys. 85, 4919 共1999兲.
10
D. Mauri, E. Kay, D. Scholl, and J. K. Howard, J. Appl. Phys. 62, 2929
共1987兲.
11
W. H. Meiklejhon, J. Appl. Phys. 33, 1328 共1962兲.
12
T. Yamaoka, J. Phys. Soc. Jpn. 36, 445 共1974兲.
13
M. Tsunoda and M. Takahashi, J. Appl. Phys. 共these proceedings兲.
Downloaded 28 Oct 2008 to 130.34.135.83. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp