Accepted Manuscript

Manifestations of unidirectional exchange anisotropy in ferrite-garnet films with

a 'weak' sublattice

V. Poimanov, V. Shkar, Yu. Nepochatykh, V. Sampath, V. Shavrov, V. Koledov

PII: S0304-8853(17)30970-8
DOI: http://dx.doi.org/10.1016/j.jmmm.2017.07.037
Reference: MAGMA 62968

To appear in: Journal of Magnetism and Magnetic Materials

Received Date: 22 March 2017

Revised Date: 26 June 2017
Accepted Date: 11 July 2017

Please cite this article as: V. Poimanov, V. Shkar, Yu. Nepochatykh, V. Sampath, V. Shavrov, V. Koledov,
Manifestations of unidirectional exchange anisotropy in ferrite-garnet films with a 'weak' sublattice, Journal of
Magnetism and Magnetic Materials (2017), doi: http://dx.doi.org/10.1016/j.jmmm.2017.07.037

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Manifestations of unidirectional exchange anisotropy in

ferrite-garnet films with a 'weak' sublattice

V. Poimanov1*, V. Shkar2, Yu. Nepochatykh2, V. Sampath3, V. Shavrov4, V. Koledov4
1
Donetsk National University, Universitetskaya 24, Donetsk, Ukraine
2
Donetsk Institute for Physics and Engineering named after A.A. Galkin, R. Luxemburg str. 72,
Donetsk, Ukraine
3
Department of Metallurgical and Materials Engineering, Indian Institute of Technology
Madras, Chennai-600 036, India
4
Kotelnikov IRE RAS, Moscow, Russia, 125009, Mokhovaya 11/7
*Corresponding author: [email protected]

Abstract. Ferrite-garnet films of submicron thickness were investigated by ferromagnetic

resonance method. It was found that the magnetization of the sample in perpendicular to the
surface of the film magnetic field showed an abnormal change in the magnetization and also
rotation of the magnetic axis. This behavior of a ferromagnetic material in the presence of a
magnetic field can be explained by the expressions for the linear magnetostrictive effect, and the
piezomagnetic effect arising from the para-processes operating within a “weak” rare-earth
magnetic sublattice which is located inside a unidirectional exchange field of “strong” iron
sublattice.
Keywords: piezomagnetism, linear magnetostriction, exchange anisotropy, paraprocess.

1. Introduction

Gadolinium ferrite-garnet shows the unique characteristic of unidirectional anisotropy of

the exchange interaction, which influences the “weak” gadolinium sublattice. The field strength
of this anisotropy is in excess of 200 kOe due to a “strong” (“combined”) iron sublattice [1] and
it acts in a direction opposite to that of the magnetization of the rare-earth sublattice. This leads
to the magnetic compensation appearing at a certain temperature (
 ). The reverse effect of the
exchange field from the rare-earth sublattice is very weak. It is therefore called unidirectional
exchange anisotropy. The exchange interaction within the gadolinium sublattice is also weak and
does not lead to the ordering of the magnetic moments of these cations [1]. These magnetic
moments are in paramagnetic state in a negative exchange field iron sublattice which lead to
their streamlining, known as paraprocess. With temperature decreasing (below the magnetic

1
compensation point) the exchange interaction within the rare-earth sublattice increases and at a
certain temperature,
 , it becomes ferromagnetic.
It should also be noted that while passing through the magnetic compensation point,
 , in
a magnetic field the reversal of the magnetic moment of the gadolinium sublattice
occurs.Therefore, on the one hand magnetic compensation for a ferromagnetic paraprocess
occurs (i.e., magnetization of the gadolinium sublattice is directed along the external magnetic
field), while,on the other hand, antiferromagnetic paraprocess occurs, i.e. magnetization of this
sublattice is aligned opposite to the external magnetic field.
For ferrimagnets, which do not have a 'weak' sublattice, in particular for yttrium iron
garnet, which is far from the point of magnetic ordering, magnetostriction is a quadratic function
of the magnetization. It is shown [1] that the presence of the 'weak' sublattice in rare-earth
ferrite-garnets leads to manifestation of abnormal phenomena. In particular, in gadolinium
ferrite-garnet a breach of magnetostriction “evennies” and also magnetoresistance and
magnetocaloric effect are observed. Magnetostriction shows linear dependence of magnetization
due to the influence of unidirectional exchange anisotropy. To the linear magnetostriction is
associated a thermodynamically opposite effect, namely piezomagntetism (the occurrence of the
magnetization under the action of elastic stresses in the absence of a magnetic field). The
occurrence of these phenomena is solely attributed to the paraprocess [1]. Based on the analysis
of these effects, the probability of occurrence of piezomagnetic effect is predicted in the work.
Experimental determination of this effect was not carried out in ferrites [1].
The aim of the present study was to detect piezomagnetic and other related effects given
rise to by the presence of a 'weak' sublattice in ferrite. One of the most convenient methods for
studying the magnetic phenomena and anomalies in ferrimagnets is ferromagnetic resonance
method. The experiment was supposed to becarried out in a magnetic field of strength that is
commensurate with the demagnetization field of the sample, since, as shown in studies of
piezomagnetism on other samples [2], the magnetization was reduced to a very small value, i.e. ~
0.001 Gs. If the sample is taken in the form of a thin film, the demagnetizing field from this
effect will also be very small (~0.012 Oe). In order therefore to measure this effect by FMR, it is
imperative that the width of the resonance absorption line is as small as possible. The narrowest
FMR line (~ 0.5 Oe) is observed in the well-studied yttrium iron garnet [3, 4]. It was supposed to
obtain the such samples with a “weak” sublattice, by substitution of yttrium ions by rare earth
magnetic ions. As it turned out, these thin films of submicron thickness of YIG were studied
earlier [5, 6].

2. Samples
2
 The samples were synthesized by the two-crucible liquid-phase epitaxy technique from
wetting melt-solution [7] on (111)-oriented substrate of gadolinium-gallium garnet. In the first
crucible containing a superheated melt-solution of yttrium iron garnet (YIG), overgrazing
(dissolution) of the upper layer of the substrate forms so that the various surface defects can be
gotten rid of. Then the substrate, which spent almost the entire time in the hot zone, was
transferred to second pot. During the transfer time of the substrate a submicron YIG type thin
film grows from the remaining on it solution-melt of YIG and partially etched and dissolved
substrate. In second pot was grown a uniaxial film with the required thickness for bubbles
magnetic domain devices. Both these and earlier investigations [5,6] were conducted on this film
of submicron thickness, which was grown on the substrate during the transfer time of it from the
first crucible to another. The width of ferrimagnetic resonance (FMR) line for these films was
~30-40 Oe, while that for pure YIG was ≤ 0.5 Oe. The saturation magnetization for these films
was 1250 Oe, while that for YIG was 1760 Gs. The film thickness was in the range 0.1- 0.2 µm,
and the cubic anisotropy field was ~36 Oe.
It is known [4] that if impurity ions are present in the rare-earth sublattice the width of the
FMR line of pure YIG increases. In this case, such impurity ions are magnetic gadolinium
cations, which have diffused from the dissolved substrate (non-magnetic gallium ions dilute iron
sublattice) [8]. Increased width of the line by nearly two orders of magnitude and decreased
magnetization of the sample indicate a significant concentration of these ions. Thus, the total
magnetization of the films arises due not only to the “iron” sublattice, but also to “weak”
(yttrium diluted by gadolinium) sublattice. Therefore, unusual phenomena may occur in them
[1].
To determine the form of the dependence of the saturation magnetization on temperature
and the search for the compensation point, measurements of the values of the normal field of the
transferring film to the saturation state were made for different temperatures. The measurements
were performed from the records of the change in the magnetic susceptibility of the sample at a
frequency of 5 MHz using the autodyne oscillator method [6]. The recording of the signal of
absorption of high-frequency energy in a film with an increase in the normal field by an acute
maximum fixes its transition to the saturation state. Such records were made for a sequence of
stabilized temperatures. The obtained dependence on the temperature of the measured values of
the transition field to the saturation magnetization has an acute deep minimum at a temperature
of 57 °C. The detected minimum is determined by the corresponding compensation point. By
recording the high-frequency energy absorption signal in such a film, as a function of
temperature, the Neel point is determined to be 248 °C.
3
 As an attempt to discover the piezomagnetic effect in ferrites with a “weak” sublattice is
made for the first time, the quantitative composition of the sample to produce the maximal effect
is unknown. With a view to achieving these objectives experiments were carried out on the
existing and previously studied [5,6] samples with a gadolinium sublattice.

3. Experiment

The details pertaining to the FMR apparatus that was used to carry out studies on the thin
films are given in [4,5]. Experiments were performed at the ambient temperature. Figure 1 shows
the frequency-field dependence of the fundamental mode of the FMR for the negative (H-) and
positive (H+) external magnetic fields that are normal to the surface of the film (Higher-order
modes, detected for this sample in [6], were not generated in this experiment). The frequency-
field dependence with a negative field is inverted to the positive axis H and is shown as curve H-
(invert) in Fig. 1. Figure 1 shows that both for the negative and positive directions of the field the
plots of frequency vs. field start from the zero value of the field, and with increasing absolute
value of the field, the curves gradually diverge from each other to the field of orientation of the
magnetization along the external field, which is ~1250 Oe. (In zero field the magnetization lies
almost on the plane of the film due to the demagnetizing energy. More precisely, it is set apart
from the plane by 0.8 degree due to the cubic anisotropy energy). Such behavior of the curves
can be explained by the change in the demagnetizing field during the magnetization of the
sample.

Fig.1. Experimental frequency-field dependence of FMR lines. The field is perpendicular to the
surface of the film. (Dependence with a negative field is inverted to the positive axis).

4
 This behavior of the curves can be explained by change in the demagnetizing field in the
magnetization process of the sample. With increasing field along one direction, the
demagnetizing energy grows greater, and in another increases less. Upon the sample reaching to
the maximum in magnetization, the change of the demagnetizing energy is terminated.
Demagnetizing energy is determined by the saturation magnetization of the sample. Thus, for
one direction of the field, normal to the plane of the film, the saturation magnetization increased,
while it decerased for the opposite direction.

Fig. 2.Variation of the derivative of magnetic susceptibility with quasinormal external (θ  1
degree) field for a range of azimuthal angles (φ ) separated by an interval of 2 degrees. (F = 300
MHz and T= Ambient).

Let us estimate the magnitude of change of magnetization in order to evaluate the

differences in frequency-field dependence shown in Fig. 1. The difference is found to be 36 Oe
near the saturation point. This means that for a change of magnetization of 1.43 Gs the
demagnetizing field changes to 18 Oe along each direction.
In the next experiment, the external magnetic field was allowed to subtend an angle ( ) of
1 degree with the normal to the surface of the film. Consider the behavior of two FMR lines at a
field strength that is closer to saturation at a frequency of 300 MHz and is dependent on the
azimuthal angle, φ, lying on the surface of the film. This line is L2H-in the negative field up to
the saturation point; and after the saturation point it changes to L3H-, whose position in the
frequency-field space (Fig. 1) is indicated by arrows. In the positive field, lines L2H+ and L3H+
correspond to this behavior.

5
 Since the field is tilted from the normal, which is represented by a diagonal of a cubic
crystal in a single-crystal film, i.e. the direction of type {111}, then at the azimuthal sample
rotation, the cubic magnetic anisotropy is due to the presence of the other three diagonals (one of
the methods for determining the cubic anisotropy [3]). For a given geometry and experimental
conditions, the explanation for cubic magnetic anisotropy is as follows: during rotation of the
sample and passing of the magnetic moment near one of the three edges of the cube, the distance
between the FMR lines reaches its the maximum. At a distance from the edges, they approach
and disappear. This fact is illustrated in Fig. 2, which shows the dependence of the derivatives of
the absorption lines on the magnetic field for a number of azimuthal angles close to the direction
{112}, which is representative of the projection of both the diagonal and the edge of the cube on
the plane of the film [3]. By rotating the azimuthal angle to 360 degrees, there are three such
ranges of FMR line appearing at a distance of 120 degrees from one another. One of them is
shown in Fig. 2. From this figure it follows that if we define the direction of the crystallographic
axes in accordance with the FMR spectra, they are dependent on the direction of the field. This
difference is about 4 degrees. Consequently, there is an effect, which depends on the magnetic
field, which changes the magnetization of the sample apart from rotating the magnetic axis.

4. Theory

The coordinate system in which the experimental studies and theoretical modeling were
conducted is shown in Fig. 3.

Fig.3.Figure shows the coordinate system in which the experimental studies and theoretical
modeling were carried out. The direction of the magnetization (M) and the external magnetic
field (H) are determined by the azimuthal (φ , φ ) and polar (θ , θ) angles.

6
 Let us examine how the theoretical frequency-field dependence will appear, if it is
assumed that the magnetization changes and the magnetic axes on magnetization of the sample
rotate, and compare them with those from the experiments. We make similar comparisons of the
azimuthal dependence of the resonance fields of the FMR lines for a frequency of 300 MHz. The
experimental dependence can be seen in Fig. 2. The theoretical model will be constructed on the
basis of the theoretical frequency-field dependence.
Initially, it was assumed that the film existed in the single-domain state. The modified
(the anomalies observed in the experiment were taken into account) energy density of the sample
is given by:

    √2

 4  −    + +     !3"# − #$%&'
  
2 4 3 3

− !   ( +    ( "# − #( %& −   ∙ " %

, ,
#$  #* ∙ + +  #* ∙ |"  %|,   . ∙ + +  . ∙ |"  %|
 
Here . denotes the demagnetizing field due to the change of magnetization (for a given
sample by 1.43 Gs). Here, the first term refers to the energy of demagnetization, the second
energy of the cubic magnetic anisotropy, the third Zeeman energy, while the fourth the energy of
unidirectional anisotropy, which describes the change in magnetization on magnetization of the
sample. In the absence of a field, i.e. zero field, its value is zero and does not shift the hysteresis
loop. In a direction that is parallel to the magnetic field, it is similar to uniaxial easy-plane
anisotropy, while in a direction that is opposite to that of the magnetic field, the uniaxial easy-
axis anisotropy exists. Therefore, along different directions of the field the hysteresis loop will
be of different size due to the different values of sample saturation magnetization. (We could not
introduce the term “unidirectional” anisotropy, and used . + ∆ ∙ " % instead of ,
where ., is the fixed part of the saturation magnetization of the sample, and ∆, the variable
part due to the paraprocess; in this case it is 1.43 Gs. In both cases, the theoretical curves differ a
little). Here #* and . are some adjustable parameters that characterize the anomalous effects
in the film due to the “weak” sublattice.
It is known [7] that the equilibrium orientation of the magnetization of a given field is
determined by solving a series of equations
1 1
 0, 0
1 1#
while the FMR frequency is determined by the expression

7
 4 1 1 1

3 5 ∙ − 6 7
2 1  1# 1#1

where the second derivatives in the expression for frequency are taken in the equilibrium
position (γ is gyromagnetic ratio).
The theoretical relationships between frequency and field are shown in Fig. 4. It can be
seen that they exhibit the same features as the experimental curves shown in Figure 1, namely,
both plots (for the negative and positive fields) start at zero field since on attaining a field
strength that is closer to the sample magnetization saturation field, the distance between them
increases. Beyond the saturation field the distance remains unchanged. Consequently, the change
in the absolute magnetization value in the magnetic field (an increase along one direction of the
field and a decrease along the opposite direction) at the sample saturation explains the
discrepancy between the experimental curves in Fig. 1.
The theoretical dependence of the resonance fields of the azimuth angle for L2 and L3
lines at 300 MHz, which are indicated on the theoretical curves by arrows in Fig. 4 are shown in
Fig. 5. They were constructed as follows: for the external magnetic field the direction ( ()

degree was set, and the frequency-field dependences were constructed similar to those shown in
Figure 4, for the selected sequence of azimuthal directions (#( ) of the {112} type. The values of
the resonance fields of these lines were at a frequency of 300 MHz and the dependences were
built according to the selected #( angles. As in the case of experimental studies, three intervals
of existence of these dependences were observed. One of them is shown in Fig. 5. From the
given plots it can be seen that the FMR lines behave similar to those obtained from the
experiment, namely, their azimuthal dependence for the opposite directions of quasinormal field
(deviation from normal is 1 degree) is shifted in relation to each other, and have different ranges
of existence.

8
Fig. 4. Theoretical frequency-field dependences of the ferrimagnetic resonance lineswiththe field
orientation, that is normal to the film plane. The plots are given for the absolute values of the
fields, φ 50 degrees, φ 8 4 degrees, H:. 36 Oe.

Fig. 5. Theoretical variation of the resonant field of the second and third FMR line, indicated in
Fig. 4, with the azimuth angle (φ ) for the frequency (F) of 300 MHz with θ= 1 degree in the
positive and negative fields.

Therefore, the selected form of energy density adequately describes the phenomena
observed in the ferrimagnetic film, i.e. increase of magnetization with increasing magnetic field
in a direction that is perpendicular to the surface of the film and the rotation of the magnetic axis.
As the magnetization increases along one direction of the field the axis rotates along one
direction; and along a field in the opposite direction it decreases, and the axis rotates opposite to
the direction of the magnetization.

5. Discussion

Based on the concepts developed in [1], the features observed during a study on the film
using the FMR method can be explained as follows. The technique of fabrication of the
ferromagnetic film involves a rare-earth sublattice incorporating ions of gadolinium (magnetic)
and yttrium (nonmagnetic) with the iron sublattice being diluted with gallium ions. The
exchange interaction between gadolinium ions at room temperature is so weak that it does not
lead to their ordering. (“weak” sublattice). Their ordering occurs due to influence of strong
(about 200 kOe) negative exchange field from the iron sublattice (paraprocess). The interaction
of the rare-earth sublattice with iron is weak. This unidirectional exchange interaction leads to a
linear rather than a quadratic dependence of magnetostriction on magnetization. Consequently,

9
when the sample is magnetized in a direction that is parallel to the field, it gets elongated, while
it gets compressed when it is aligned in a direction that is perpendicular to the field. As the
symmetry analysis shows, the accounting of the linear magnetostrictive effect and piezomagnetic
effect under cubic symmetry [1,8], in a coordinate system with the axis [111] along the normal
leads to the appearance of non-diagonal components in the strain tensor. In turn, this may lead
not only to the anisotropic stretching of the sample (without changing its volume), but also to a
purely mechanical torsion of crystallographic axes. Moreover, with respect to the applied
external field, this effect also shows unidirectionality. Since thermodynamically it is the opposite
effect, piezomagnetism corresponds to the linear magnetostriction. Such deformation of the
sample in a magnetic field causes an increase in magnetization due to tension and a decrease in it
due to compression and also to the rotation of the magnetic axes by twisting. Thus, as
determined from the experiment the change in magnetization and the rotation of the magnetic
axes in the magnetic field are interrelated and are defined by the simultaneous manifestation of
two effects, namely linear magnetostriction and piezomagnetic effect.

6. Conclusions

Ferromagnetic resonance method was used to study submicron-thick ferrite-garnet film

grown by the liquid-phase epitaxy technique on gallium-gadolinium garnet substrate oriented
along (111) direction. It has been found that during magnetization in a normal magnetic field an
abnormal change in the magnetization of the sample as well as rotation of the magnetic axes
occur: i.e., in the straight (one) field direction the magnetization increases and the axis of the
magnetic field rotates to align with it; on the other hand, in the opposite external field direction,
magnetization decreases and the axis rotates to align with it in the opposite direction. This
ferromagnetic behavior in a magnetic field can be explained by the simultaneous manifestation
of the linear magnetostriction and the piezomagneticeffect that operate due to the paraprocess in
a “weak” magnetic sublattice, wich is located in a unidirectional exchange field of the “strong”
sublattice [1]. Piezomagnetic effect is observed for the first time in ferrite.

This work was supported by Russian Science Foundation, grant No. 14–22-00279

References

[1] K.P. Belov, Effects of unidirectional exchange anisotropy in ferrites, Physics-Uspekhi, 42

(1999) 711–717
10
[2] А.S.Borovik-Romanov, Piezomagnetism in antiferromagnetic fluorides of cobalt and
manganese, Zhurnal Eksperimentalnoyi i Teoreticheskoi Fiziki, 38(1960)1088 [Soviet Physics
JETP 11(1960) 1346]
[3] A.M. Balbashov, F.V. Lisovsky, V.K. Raevet al. Magnetic bubble elements and devices:
Directory, Radio and Communications, Мoscow, 1987
[4] A.G. Gurevich, Ferromagnetic resonance in ferrite and antiferromagnetics, Moscow, Nauka,
1973
[5] V.F.Shkar’, V.N.Varyukhin, Iron-yttrium garnet films with magnetic vortices, JETP Letters
88 (2008) 271-274
[6] V.F.Shkar’, V.N.Varyukhin, Magnetic resonance and self-sustained oscillations in an iron-
yttrium garnet film in a vortex state, JETP Letters 92 (2010) 338-342
[7] Y. I. Nikolaev, I. А. Krasin, Obtaining ferrite garnet films with wetting solution-melt film,
Crystallography Reports 33 (1988) 478-482
[8] S. I. Yushuk, The layered structure of epitaxial yttrium iron garnet films, Zhurnal
Technicheskoi Fiziki 69(12) (1999) 62-64

11
• The abnormal change in the magnetization and the rotation of the magnetic
axes is discovered at magnetization in the normal magnetic field of the yttrium
iron garnet type film of submicron thickness.

• The rare-earth sublattice magnetic ordering occurs due to influence of strong

exchange field from the iron sublattice (paraprocess).

• The presence of the paraprocess in a “weak” rare-earth magnetic sublattice

leads to a linear magnetostrictive and piezomagnetic effect.

12