The magnetic properties of Fe-based amorphous ribbons coated with various oxides using the sol-gel process

Article abstract of DOI:10.1023/A:1018615104519

The metal oxides MgO, BaO, Al₂O₃, SiO₂, MgO-Al₂O₃, CaO-Al₂O₃, SrO-Al₂O₃, BaO-Al₂O₃ and MgO-SiO₂ have been coated onto ribbo

Full text of DOI:10.1023/A:1018615104519

JOURNAL OF MATERIALS SCIENCE 32 (1997) 3219—3225
The magnetic properties of Fe-based amorphous
ribbons coated with various oxides
using the sol–gel process
S . H. LIM , T. H. N O H
Magnetic Materials Laboratory, Korea Institute of Science and Technology,
P.O. Box 131, Cheongryang, Seoul, 130 —650, Korea.
Y. J . BAE, H. K. CHAE
Departm ent of Chem istry, Hankuk University of Foreign Studies, Seoul, 130—791, Korea
Y. S . CHO I
Departm ent of Metallurgical Engineering, Korea University, Seoul, 136—701, Korea
The m etal oxides MgO, BaO, Al O , SiO , MgO—Al O , CaO—Al O , SrO—Al O , BaO—Al O and
2 3
2
2 3
2 3
2 3
2 3
MgO—SiO have been coated onto ribbons of the Fe-based am orphous alloy Metglas
2
2605S3A using a sol—gel process. The effects of the surface coating on the m agnetic
properties of the alloy are investigated. The d.c. hysteresis loop for ribbons coated with MgO
and MgO—Al O is m ore square shaped than that for the uncoated ribbon. For ribbons coated
2 3
with BaO, SrO—Al O and BaO—Al O it is m ore inclined than for the uncoated ribbon.
2 3
2 3
Significant differences in the frequency dependence of the effective perm eability are
observed that depend on the nature of the coated oxides. The core loss is also affected by the
coating. These results m ay be explained in term s of a stress induced by the coating and the
m odification of the dom ain structure via elastic and/or m agnetoelastic interactions. It is
thought from the m agnetoelastic interactions that the MgO and MgO—Al O coatings induce
2 3
tensile stresses whilst those of BaO, SrO—Al O and BaO—Al O induce com pressive stresses.
2 3
2 3
1. Introduction
1—1000 kHz and also the d.c. hysteresis loops were
inclined by the surface coating. These results were
thought to be a result of the tensile stress induced by
the coating, which in turn modifies the domain struc-
ture. Since the ribbon used had a nearly zero mag-
netostriction, the stress effect was considered to be
mainly elastic. It is expected that, if a coating could be
produced on a ribbon with a large magnetostriction
then the effects of the coating on the magnetic proper-
ties would be even more pronounced, since the prop-
erties would now be influenced by both magnetoelas-
tic and elastic interactions. It is therefore of interest to
investigate the effects of surface coating on the mag-
netic properties of an amorphous ribbon with a large
magnetostriction.
Coatings are frequently applied to the surfaces of
sheets or ribbons of magnetic materials in order to
improve their properties. One important use of coat-
ings is the reduction of eddy currents when the coating
provides insulation between layers in stacked or tape
wound cores. The coating can be produced by simply
dipping the ribbons into a mixture of a fine oxide
powder in water or organic solvents [1]. Another
important use for coatings is the modification of
domain structure via stresses induced by the coating.
An example of this type of coating is that which is
applied to silicon steel sheets in order to refine its
domain structure [2—5], since the coated layer/matrix
interface is able to transfer induced stress at the coated
layer to the matrix. This type of coating is generally
produced by more sophisticated methods.
In this work, we employ a previously developed [6]
sol-gel process to coat various oxide layers onto rib-
bons of an Fe-based amorphous alloy with a large
magnetostriction and investigate the effects of the
coatings on the magnetic properties of the alloy.
Coatings applied to amorphous magnetic ribbons
have so far been produced in order to achieve the first
goal, namely, a reduction in eddy currents. Recently,
however, a magnesium oxide coating on a zero
magnetostrictive Co-based amorphous ribbon, cre-
ated using the sol—gel process, was observed to also
affect the a.c. and d.c. magnetic properties in addition
to the eddy current loss [6]. As well as the reduction
in core loss, it was found that the a.c. effective perme-
ability was improved over the frequency range of
2. Experimental procedure
The amorphous alloy used was Metglas 2605S3A
(Fe
Cr B Si C ) with a magnetostriction of
ꢀꢁꢂꢁ ꢃ ꢄꢁ ꢅ ꢆꢂꢅ
\ꢁ
20;10 which was supplied by AlliedSignal Corp.,
USA. The width and thickness of the ribbon were
0022—2461 ( 1997 Chapman & Hall
3219

TABLE I The type of oxides and their precursor solutions, together with the concentrations of the metal alkoxides and water in ethanol and
the thickness of the coated oxide layers
Type of
oxide layer
Precursor solution
in 2-methoxyethanol
Concentration of
metal alkoxide (M)
Concentration of
Thickness of
coated layer (lm)
H O in ethanol (M)
ꢃ
MgO
BaO
Al O
Mg(OCH CH OCH )
0.5
0.5
0.5
0.06
0.5
0.5
0.5
0.5
0.5
0.5
0.5
no water
0.24
no water
no water
no water
no water
no water
0.70
0.40
0.39
0.65
0.75
1.92
1.68
1.80
1.39
ꢃ
ꢃ
ꢇ ꢃ
Ba(OCH CH OCH )
ꢇ ꢃ
Al(OCH CH OCH )
ꢇ ꢇ
ꢃ
ꢃ
ꢃ
ꢇ
ꢃ
ꢃ
SiO
MgO—Al O
Si(OCH CH )
ꢃ
ꢃ
ꢇ ꢈ
MgAl (OCH CH OCH )
ꢇ ꢉ
CaAl (OCH CH OCH )
ꢇ ꢉ
SrAl (OCH CH OCH )
ꢇ ꢉ
BaAl (OCH CH OCH )
ꢇ ꢉ
MgSi(OCH CH OCH )
ꢇ ꢁ
ꢃ
ꢇ
ꢃ
ꢃ
ꢃ
CaO—Al O
ꢃ
SrO—Al O
ꢃ
ꢇ
ꢃ
ꢃ
ꢃ
ꢇ
ꢃ
ꢃ
ꢃ
BaO—Al O
ꢃ
ꢇ
ꢃ
ꢃ
ꢃ
MgO—SiO
ꢃ
ꢃ
ꢃ
4.48 mm and 15.5 lm, respectively. A variety of metal
oxides; MgO, BaO, Al O , SiO , MgO—Al O ,
thickness of single cation oxides is generally much
smaller than that of the mixed cation oxides.
The ribbons were wound onto toroidal cores with
an inner diameter of 19 mm. The cores were annealed
ꢃ ꢇ
ꢃ
ꢃ ꢇ
and
CaO—Al O ,
SrO—Al O ,
BaO—Al O ,
ꢃ ꢇ
ꢃ
ꢃ ꢇ
ꢃ ꢇ
MgO—SiO were coated onto the alloy ribbons using
the sol—gel process. The synthesis of the metal
°
at 435 C for 2 h under an N gas atmosphere and
ꢃ
then air-cooled. The a.c. effective permeability (l ) was
%
alkoxide precursors and complexes is described in
detail elsewhere [7]. In this publication we
present only a brief discussion of the technique.
The solution of magnesium methoxyethoxide
[Mg(OCH CH OCH ) ] for the MgO coating was
measured using an impedance analyser at an applied
\ꢄ
field of 0.795 Am and over a wide frequency range
of 0.1—1000 kHz. D.c. magnetic properties were meas-
ured with a hysteresis loop tracer. The core loss (¼ )
ꢃ
ꢃ
ꢇ ꢃ
*
prepared by the reaction of an excess (8 mol) of
2-methoxyethanol (CH OCH CH OH, Aldrich) with
was measured by using a B—H analyser (Iwatsu Co.
Model No. SY-8216) at frequencies up to 200 kHz and
flux densities of 0.1, 0.3, 0.5, 0.7 and 1.0 T.
ꢇ
ꢃ
ꢃ
one mole of magnesium ethoxide [Mg(OCH CH ) ,
ꢃ
ꢇ ꢃ
Aldrich]. Similarly, the other precursor solutions for
single cation oxides were prepared by the exchange
reaction of metal alkoxides with excess 2-methoxy-
ethanol (in the cases of BaO and Al O ) or ethanol (in
3. Results and discussion
ꢃ ꢇ
the case of SiO ). For bimetallic precursors, a solution
ꢃ
In Fig. 1(a and b) we present the results for l as
%
of one metal alkoxide was added equivalent to a solu-
a function of frequency. The results obtained on rib-
tion of the other metal alkoxide and the mixed solu-
tions were then refluxed. Hydrolysis was carried out
by adding a solution of water and alcohol to the
precursor. The concentrations of the metal alkoxide
and water in the precursor solution, which are impor-
tant parameters in determining the thickness of the
coated layer, were adjusted to give reasonable thick-
ness values. For example, the concentration of water
in ethanol was 0.5 M for MgO and BaO, and 0.24 M for
bons coated with single cation metal oxides are shown
in Fig. 1a, whilst those for ribbons coated with the
double metal oxides are shown in Fig. 1b. The results
obtained for an uncoated ribbon are also shown on
both figures. It can be seen from Fig. 1(a and b) that
significant differences are observed in the frequency
dependencies of the permeability. The ribbon coated
with MgO—Al O exhibits a substantially higher
ꢃ ꢇ
value of the permeability than the uncoated ribbon
SiO , whereas the concentration of metal alkoxide
over the whole frequency range with the difference
being larger at lower frequencies. The values for the
ꢃ
was 0.5 M for MgO and BaO, and 0.06 M for SiO . It
ꢃ
was not necessary to add water to the other coatings,
ribbons coated with MgO and Al O are also slightly
ꢃ ꢇ
because the viscosity of the precursor solutions were
already satisfactory to produce a reasonable thick-
ness. The concentrations of the metal alkoxides and
water used in this work are summarized in Table I.
After dipping the alloy ribbon into the precursor solu-
higher than for the uncoated ribbon. The rest of the
coated ribbons exhibit smaller values of the permeab-
ility than the uncoated ribbon at nearly all the
frequencies. In the case of ribbons coated with BaO,
BaO—Al O and SrO—Al O , a peculiar behaviour in
ꢃ ꢇ
ꢃ ꢇ
°
tion, the ribbon was dried at 200 C. All the coating
the frequency dependence of the permeability is
observed; the effective permeability is very low at low
frequencies, remains nearly constant up to 100 kHz
and then slowly decreases with a further increase in
the frequency. It is worth noting that, at the very high
frequency of 1 MHz, the permeability of the coated
ribbons is always higher than that of the uncoated
ribbon. Some values of the permeability between
50—100 kHz are observed to be irregular, particularly
procedures were carried out in a dry nitrogen atmo-
sphere. The thickness of a coated layer was measured
using a digital micrometer. In order to increase the
accuracy of the measurement, a stack consisting of
about 10—20 ribbons was first measured and the thick-
ness of each ribbon was then obtained by dividing the
total stack thickness by the number of ribbons in the
stack. The results for the thickness of the coated oxide
layer are also given in Table I. The measured thickness
ranges from 0.39—1.92 lm. It is worth noting that the
for the ribbons coated with MgO and MgO—Al O .
ꢃ ꢇ
This phenomenon is considered to be due to
3220

Figure 2 The value of core loss of Metglas 2605S3A at the flux
density of 0.1 T as a function of frequency for the ribbons coated
with (a) the single metal oxides; (᭿) Al O , (᭡) SiO ; (᭢) BaO and
Figure 1 The value of effective permeability of Metglas 2605S3A as
a function of frequency for the ribbons coated with (a) the single
metal oxides (E) MgO, (᭿) Al O , (᭡) SiO and (᭢) BaO and (b) the
ꢃ
ꢇ
ꢃ
ꢃ
ꢇ
ꢃ
(b) the double metal oxides; (E) MgO—Al O , (᭿) MgO—SiO , (᭡)
double metal oxides, (E) MgO—Al O , (᭿) MgO—SiO , (᭡)
ꢃ
ꢇ
ꢃ
ꢃ
ꢇ
ꢃ
CaO—Al O (᭢) BaO—Al O and (⅙) SrO—Al O . The results for the
CaO—Al O (᭢) BaO—Al O and (⅙) SrO—Al O . The results for the
ꢃ
ꢇ
ꢃ
ꢇ
ꢃ ꢇ
ꢃ
ꢇ
ꢃ
ꢇ
ꢃ ꢇ
uncoated ribbon are shown in both figures using the symbol (᭜) .
uncoated ribbon are shown in both figures using the symbol (᭜)
Note: no core loss date for MgO coating at 0.1 T.
magnetomechanical resonance, which often occurs for
highly magnetostrictive alloys [8].
BaO. The same behaviour is observed at the other
measured flux densities. In the case of ribbons coated
with double metal oxides (Figs 2b and 3b), the
frequency dependence of the core loss is somewhat
different from that observed in Figs 2a and 3a in the
case of ribbons coated with a single metal oxide. The
The results for the core loss are shown in Figs 2(a
and b) and 3(a and b) as a function of frequency at the
fixed flux densities of 0.1 and 0.7 T, respectively. The
results for the ribbons coated with a single oxide are
shown in Figs 2a and 3a and those for the ribbons
coated with the double metal oxides are shown in
difference between the ¼ values is small and
*
depends on the individual coating oxide. In addition
Figs 2b and 3b. The values of ¼ were measured at
the relative magnitude of ¼ of the coated ribbons
*
*
various flux densities ranging between 0.1 to 1.0 T and
compared to the uncoated one changes with the flux
the way in which ¼ varies with frequency at the flux
density. At a flux density of 0.1 T, the uncoated ribbon
*
densities, not shown in this paper due to space limita-
has the lowest value of ¼ at nearly all of the investi-
*
tions, is qualitatively similar to that shown in Figs 2(a
and b) and 3(a and b) for the flux densities of 0.1 and
0.7 T. In the case of the ribbons coated with a single
gated frequency range. However, the relative
magnitude of ¼ of the uncoated ribbon progress-
*
ively increases with the flux density and, at flux densit-
metal oxide, the difference in ¼ is rather large
ies of 0.7 T and higher, the uncoated ribbon exhibits
*
depending on the oxides used in the coating. The core
the largest value of ¼ at nearly all the measured
*
loss is higher than the uncoated ribbon for the ribbons
frequencies. This indicates that the importance of hav-
coated with MgO and SiO but is lower than the
ing a surface coating increases with the flux density.
The same behaviour, namely a steady increase in the
ꢃ
uncoated ribbon for those coated with Al O and
ꢃ ꢇ
3221

coated ribbon exhibits the lowest value of ¼ at
*
nearly all the measured frequencies and flux densities.
From the present results for l and ¼ , it is clear
%
*
that there is no obvious correlation between the two
properties, indicating that the major factors affecting
the two individual properties differ from each other.
The present results can be compared with those ob-
served for a Co-based amorphous alloy, where both
l and ¼ were improved by an MgO coating [6].
%
*
Although l is often referred to as being the single
%
most important property of a soft magnetic material,
it is a highly complex function of both material
parameters and microstructure. This quantity, l , is
%
closely related to the magnetization behaviour. The
magnetization is created by domain wall movement
and spin rotation, the relative contribution of each
being a function of the frequency and the angle
between the applied field and spin directions. The
contribution to the magnetization by the spin rotation
is known to be large when the frequency is high and
the applied field direction is perpendicular to the spin
direction [9, 10]. The magnetization mode of domain
wall movement is significantly affected by the domain
wall energy. The magnetization will be mainly created
by the motion of a domain wall when the wall energy
is large [11—13]. In this case the potential well
produced by obstacles impeding the wall motion is an
important factor. The obstacles include defects and
internal (residual) stress, the latter being important for
a material with a large magnetostriction such as the
Metglas 2605S3A used in this work. The magnetiz-
ation will be largely generated by wall bending if the
wall energy is small and in this case the domain
structure (refinement) is an important factor [11—13].
Magnetization by wall bending may be operating in
the present amorphous alloy, since the wall energy is
considered to be small due to a small anisotropy. The
Figure 3 The value of core loss of Metglas 2605S3A at the flux
density of 0.7 T as function of frequency for the ribbons coated with
(a) the single metal oxides; (E) MgO, (᭿) Al O , (᭡) SiO and (᭢)
present results for l may therefore be explained by
%
considering the variation of domain structure with
ꢃ
ꢇ
ꢃ
BaO and (b) the double metal oxides, (E) MgO—Al O , (᭿)
ꢃ
ꢇ
surface coating when the magnetization is mainly
created by domain wall movement, namely the
frequency is low and the proportion of the domains
that are aligned in the length direction is high. The loss
of cores subjected to an a.c. magnetic field may consist
of a hysteresis loss and an eddy current loss. The core
loss due to eddy currents may be further divided into
classical eddy current loss and an anomalous loss
[14, 15]. The classical eddy current loss is due to
macroscopic eddy currents induced from homogene-
ous magnetization and it is independent of domain
structure. On the other hand, the anomalous loss is
due to microscopic eddy currents, which are induced
from the dynamic movement of domain walls and are
proportional to the wall velocity. Since amorphous
alloy ribbons have a high resistivity and the thickness
is very small, the classical eddy current loss is much
smaller than the anomalous loss [15—18]. The total
loss of amorphous alloy ribbons may therefore be
understood by considering hysteresis loss and the
anomalous loss. The hysteresis loss and the anom-
alous loss are proportional to the frequency and
the frequency squared, respectively, so that the former
will dominantly contribute to the total loss at low
MgO—SiO , (᭡) CaO—Al O (᭢) BaO—Al O and (⅙) SrO—Al O .
ꢃ
ꢃ
ꢇ
ꢃ
ꢇ
ꢃ ꢇ
The results for the uncoated ribbon are shown in both figures using
the symbol (᭜).
relative magnitude of ¼ with the flux density, is also
*
expected to occur for the ribbons coated with a single
metal oxide. Actually, a closer examination of the core
loss dependence on the frequency results for the
ribbons coated with a single metal oxide does show
the same behaviour. For example, the difference in
¼
values between the SiO coated and uncoated
*
ꢃ
ribbons tends to decrease with the flux density, but,
the gap between the BaO coated and uncoated
ribbons increases with the flux density. This indicates
that as the flux density increases the core loss charac-
teristics of the coated ribbons improve to a greater
extent than that of the uncoated ribbon. However, the
change in the relative values between the coated and
uncoated ribbons with the flux density does not occur
in the case of ribbons coated with a single metal oxide.
One reason for this may be that the difference in
¼ values is large in this case. Amongst the ribbons
*
coated with double metal oxides, the BaO—Al O
ꢃ ꢇ
3222

TABLE II D.c. magnetic properties of the uncoated ribbon and
the ribbons coated with the oxides
Type of coated
oxide
l
l
B /B
H (7.958
;10\ꢃAm\ꢄ)
G
K
ꢀ
ꢄꢆ
#
Uncoated
MgO
BaO
20800
9400
44 800
66 400
14 800
38 800
50 500
77 700
29 700
12 400
11 300
49 000
0.40
0.74
0.20
0.34
0.57
0.71
0.34
0.16
0.17
0.50
67
80
83
51
89
90
85
69
80
69
12200
27700
18900
10300
23300
10000
8600
Al O
ꢃ
ꢇ
SiO
MgO—Al O
ꢃ
ꢃ
ꢇ
CaO—Al O
ꢃ
SrO—Al O
ꢃ
ꢇ
ꢇ
BaO—Al O
ꢃ
ꢇ
MgO—SiO
15400
ꢃ
frequencies but the latter will be dominant at high
frequencies.
This brief discussion indicates that, in order to
understand the results for l and ¼ , detailed
%
*
information on the domain structure is required. In an
effort to obtain this information d.c. magnetic proper-
ties were measured and the results are summarized in
Table II. The hysteresis results are plotted in Fig.
4(a—e) for some typical cases, namely for the
uncoated ribbon and the ribbons coated with MgO,
MgO—Al O , BaO, SrO—Al O , respectively. The
ꢃ ꢇ
ꢃ ꢇ
values of coercivity H listed in Table II were obtained
#
\ꢄ
by applying a maximum field of 79.58 Am . A very
large difference in the d.c. magnetic properties is ob-
served, which is vividly reflected in the shapes of the
hysteresis graphs. The hysteresis loop for the ribbons
coated with MgO and MgO—Al O is square-shaped
ꢃ ꢇ
(more squared than the uncoated ribbon) with a large
value of B /B (B is the remanent flux density and
ꢀ
ꢄꢆ
ꢀ
B
is the flux density at the magnetic field of
ꢄꢆ
\ꢄ
795.8 Am ) and the maximum permeability, l , but
ꢁ
a small value of the initial permeability, l . The
*
squared-shape (and hence large B /B ) may be re-
ꢀ
ꢄꢆ
lated to the fact that more domains are aligned in the
length direction [19]. The low value of l may be
*
related to a low domain density, since magnetization
by domain wall movement and domain wall bending
is expected to be small when the domain density is
small. From this, the ribbons coated with these single
and double metal oxides are considered to have
a small number of domains and a large proportion of
these domains are aligned in the length direction. The
shape of the hysteresis graphs for the ribbons coated
with BaO, SrO—Al O and BaO—Al O are signifi-
Figure 4 D.c. hysteresis loops for (a) the uncoated ribbon and the
ribbons coated with (b) MgO, (c) MgO—Al O , (d) BaO, (e)
ꢃ
ꢇ
SrO—Al O .
ꢃ ꢇ
ꢃ ꢇ
ꢃ
ꢇ
cantly inclined with a small value of B /B . The
ꢄꢆ
values of l and l are small as is expected. It is
ꢀ
ꢁ
*
important to note that the difference in the values
teresis loop for the ribbon coated with SiO is squared
ꢃ
and with Al O it is slightly inclined, whilst no sub-
ꢃ ꢇ
between l and l is small for these ribbons. These d.c.
ꢁ
*
results may indicate that more domains are aligned in
the transverse direction. Detailed information on the
domain density cannot be drawn from the present d.c.
results. However, from geometrical considerations the
domains of these ribbons are considered to be more
refined than for the ribbons with squared hysteresis
loops, since a large domain width results in a very
large magnetostatic energy when the domains are alig-
ned in the transverse direction. The shape of the hys-
stantial change in the d.c. properties was observed for
the MgO coating. In most cases, the value of H was
#
appreciably increased by the coating with the notable
exception of the Al O coating.
ꢃ ꢇ
The results for the d.c. properties and information
drawn from these d.c. results may offer some explana-
tion for the l and ¼ results shown in Figs 1 and 2.
%
*
The l results can be explained by using the correla-
%
tion between l and B /B investigated by Lim et al.
%
ꢀ
ꢄꢆ
3223

for amorphous alloys [10, 18]. The results indicate
induced anisotropy energy. Work in this direction is in
progress and will be published in due course.
that a material with a high B /B has a high value of
ꢀ
ꢄꢆ
l at low frequencies and that l steeply decreases
%
%
with the frequency resulting in a low value of l at
%
high frequencies. A material with a low B /B , on the
other hand, has a low value of l at low frequencies
and this small value of l is maintained up to very high
frequencies. In fact, in most cases, the current l
4 Conclusions
ꢀ
ꢄꢆ
The metal oxides MgO, BaO, Al O , SiO ,
%
ꢃ ꢇ
ꢃ
ꢃ ꢇ
MgO—Al O , CaO—Al O , SrO—Al O , BaO—Al O ,
%
ꢃ ꢇ
ꢃ ꢇ
ꢃ ꢇ
and MgO—SiO have been coated onto ribbons of the
Fe-based amorphous alloy Metglas 2605S3A using
%
ꢃ
results are satisfactorily explained by the correlation
behaviour although some minor disagreements are
a sol—gel process. The effects on the magnetic proper-
ties of the alloy of the surface coating have been
investigated. A more square-shaped hysteresis loop
was observed for ribbons coated with MgO and
seen in the cases of ribbons coated with Al O and
ꢃ ꢇ
SiO , where some other factor that affects l must be
ꢃ
%
introduced. For example, the slightly improved per-
meability of the Al O coated ribbon compared to the
uncoated ribbon may be related to the low value of
MgO—Al O but a more inclined loop was observed
ꢃ ꢇ
ꢃ ꢇ
for the ribbons coated with BaO, SrO—Al O and
ꢃ ꢇ
BaO—Al O than for the uncoated ribbon. The fre-
ꢃ ꢇ
H . The high values of ¼ observed for the ribbons
#
*
coated with MgO, SiO and MgO—Al O are thought
quency dependence of the effective permeability is also
ꢃ
ꢃ ꢇ
to result from the low domain density which results in
a large anomalous loss. Similarly, the low loss values
observed for the ribbons coated with BaO and
significantly affected by the oxide coating. A ribbon
coated with MgO—Al O , for example, exhibits sub-
ꢃ ꢇ
stantially higher values of the permeability than the
BaO—Al O may result from a refined domain struc-
ture. The very low losses for the ribbon coated with
uncoated ribbon over the whole measured frequency
range, whilst, in the case of ribbons coated with BaO,
BaO—Al O and SrO—Al O , the permeability is low
ꢃ ꢇ
Al O does not appear to be explained in terms of the
ꢃ ꢇ
ꢃ ꢇ
ꢃ ꢇ
domain structure; again, this may be related to the
at low frequencies but it remains relatively constant
up to very high frequencies. The core loss is also
affected by the coating, although the difference is not
so significant dependant on the type of the coated
oxide. These results may be explained by a stress
induced by the coating and the modification of the
domain structure via elastic and/or magnetoelastic
interactions. It is considered from the magnetoelastic
very low value of H which reduces the hysteresis loss
#
component. As previously mentioned the hysteresis
loss is one of the two important components that
contribute to the total loss.
It would be of interest to consider the reasons
behind the large change in the magnetic properties
caused by the surface coating. The main reason is
thought to result from the magnetoelastic effect which
is caused by coupling the magnetostriction to the
stress induced by the coating. Since the present
Fe-based amorphous alloy has a positive magneto-
interactions that the MgO and MgO—Al O coatings
ꢃ ꢇ
induce tensile stresses whilst those of BaO,
SrO—Al O and BaO—Al O induce compressive
ꢃ ꢇ
ꢃ ꢇ
stresses.
\ꢁ
striction (whose value is relatively high at 20;10 ),
a tensile stress in the length direction causes domains
to align in the length direction, thus resulting in
a square-shaped hysteresis loop with a high value of
B /B [20]. Similarly, a compressive stress leads to an
References
1. D. M. NATHASINGH, C. H. SMITH and A. DATTA, IEEE
¹rans. Magn. MAG-20 (1984) 1332.
ꢀ
ꢄꢆ
inclined hysteresis graph with a low value of B /B
2. S. TAGUCHI, T. YAMAMOTO and A. SAKAKURA, Ibid.
MAG-10 (1974) 123.
ꢀ
ꢄꢆ
[20]. It is therefore thought that the MgO and
MgO—Al O coatings induce a tensile stress while
3. D. R. THORNBURG and W. M. SWIFT, Ibid. MAG-15
(1979) 1592.
4. Y. INOKUTI, K. SUZUKI and Y. KOBAYASHI, J. Jpn.
Inst. Met. 59 (1995) 213.
ꢃ ꢇ
those of BaO, SrO—Al O and BaO—Al O induce
ꢃ ꢇ
ꢃ ꢇ
a compressive stress. The other oxides do not seem to
induce any appreciable amounts of stress.
5. Y. INOKUTI, Ibid. 59 (1995) 347.
The various induced stress types that are obtained
depending on the type of the coating oxide used may
be of practical importance. The coatings play a similar
role to field and stress annealing. Considering that the
formation of induced anisotropy in the transverse
direction by field annealing is known to be difficult
and the induced anisotropy energy is small in the case
of Fe-based amorphous alloys [21], the inclination of
the hysteresis loop by the coating is particularly
important. In applications such as choke cores,
constant permeability over a wide applied field is an
important property, which may be achieved by induc-
ing anisotropies in the transverse direction. A rough
estimate shows that the energy of induced anisotropy
achieved by the present coating is of the order of
6. S. H. LIM, Y. S. CHOI, Y. J. BAE, H. K. CHAE, T. H. NOH
and I. K. KANG, IEEE ¹rans. Magn. MAG-31 (1995)
3898.
7. D. C. BRADLEY, R. C. MCHROTRA and D. P. GAUR
in ‘‘Metal alkoxides’’, (Academic, London, 1978) Ch. 2 and
Ch. 5.
8. M. MITERA, H. FUJIMORI and T. MASUMOTO, in ‘‘Rap-
idly quenched metals’’, Vol. II, edited by T. Masumoto and
K. Suzuki (The Japan Institute of Metals, Sendai, Japan, 1982)
p. 1035.
9. H. FUJIMORI, H. MORITA, M. YAMAMOTO and J.
ZHANG, IEEE ¹rans. Magn. MAG-22 (1986) 1101.
10. S. H. LIM, Y. S. CHOI, T. H. NOH and I. K. KANG, J. Appl.
Phys. 75 (1994) 6937.
11. M. KERSTEN, Physik. Z. 39 (1938) 860.
12. S. CHIKAZUMI in ‘‘Physics of magnetism’’, (J. Wiley, New
York, 1964) Ch. 13.
13. D. JILES, in ‘‘Introduction to magnetism and magnetic mater-
ials’’, (Chapman and Hall, London, 1991) Ch. 7.
14. R. H. PRY and C. P. BEAN, J. Appl. Phys. 29 (1958) 532.
\ꢇ
100 J m . A field annealing of a coated ribbon in the
transverse direction is expected to further increase the
3224

15. H. PFUTZNER, P. SCHONHUBER, B. ERBIL, G.
HARASKO and T. KLINGER, IEEE ¹rans. Magn. MAG-27
(1991) 3426.
16. R. F. KRAUSE and F. E. WERNER, Ibid. MAG-17 (1981)
2686.
20. B. D. CULLITY, in ‘‘Introduction to magnetic materials’’,
(Addison-Wesley, Reading MA, 1972) Ch. 12.
21. A. HERNANDO and M. VAZQUEZ, in ‘‘Rapidly solidified
alloys’’, edited by H. H. Lieberman (Marcel Dekker, New York
1993) Ch. 17.
17. V. R. V. RAMANAN, J. Mater. Engng. 13 (1991) 119.
18. S. H. LIM, Y. S. CHOI, T. H. NOH and I. K. KANG,
(unpublished results).
19. S. CHIKAZUMI, in ‘‘Physics of magnetism’’, (J. Wiley, New
York, 1964) Ch. 12.
Received 7 March
and accepted 31 July 1996
.
3225