Jpn. J. Appl. Phys. Vol. 38 (1999) pp. 1687–1690
Part 1, No. 3B, March 1999
c
°1999
Publication Board, Japanese Journal of Applied Physics
Initial Magnetization Curve and Magnetic Domain Pattern
of Co/Pt Multilayered Thin Film
Te-ho W U, Zu-yin Q IU, Wen-bin L AI, Jung-chun H UANG1 and Jong-ching W U2
Department of Humanities and Sciences, National Yunlin University of Science and Technology, Touliu, Taiwan 640, R.O.C.
1 Department of Physics, National Cheng-Kung University, Tainan, Taiwan, R.O.C.
2 Department of Physics, National Changhua University of Education, Changhua, Taiwan 500, R.O.C.
(Received October 14, 1998; accepted for publication December 7, 1998)
This study inquires into the relationships between the initial magnetization curve and the magnetic domain pattern in the
demagnetized states for multilayered Co/Pt thin film samples. The magnetic domain pattern for the sample demagnetized
by an in-plane magnetic field and for the sample demagnetized by a perpendicular magnetic field were found to be quite
different even though both states have zero magnetization. The former state has denser and finer domains than the latter.
The initial magnetization for the fine domains increases with an increase in magnetic field, while for the coarse domains, the
initial magnetization remains at zero for magnetic field below coercivity Hc , then rises sharply to saturated magnetization when
magnetic field is nearly equal to Hc . Moreover, the magnetic domain pattern for the sample demagnetized by an in-plane
magnetic field can be use to estimate the domain wall energy as long as the domain size is known
KEYWORDS: domain pattern, initial magnetization curve, domain size, wall energy
1. Introduction
2. Experiments
In order to understand the real character of a magnetization
reversal process, it is necessary to know every aspect of a material’s magnetization curve and its magnetic domain distributions. Magnetization measurement gives the macroscopic behavior of the film,1) while the magnetic domain observation
provides the microscopic behavior.2) Yet, often the microscopic behavior can be conjectured from bulk measurement
(which shows the macroscopic behavior) via proper experimental procedure and analysis.3)
The main purpose of this study is trying to establish
the possible relationships between the initial magnetization
curves and the magnetic domain pattern for multilayered
Co/Pt magneto-optical (MO) recording thin films. This was
done specifically through an investigation of different demagnetized states of samples demagnetized by a variety of methods. Two methods had been employed to demagnetize samples in this study. The first method to demagnetize a sample
was by applying an in-plane magnetic field (H ) perpendicular
to magnetization (M), whose strength (25 kOe) was enough
to bring M to the in-plane direction. The second way of demagnetizing a sample was by applying a perpendicular (to the
film plane) magnetic field with a direction opposite to M with
strength of coercivity Hc . The domain patterns were found
to be more condensed and finer from first method than in the
second method. In other words, in-plane field produced fine
domain distributions, while perpendicular field demagnetization yielded coarse domain patterns. In addition, both states
were studied in light of the initial magnetization curves obtained by measurements of extraordinary Hall effect (EHE).
We observe that the initial magnetization for the fine domains
increases with an increase in magnetic field (H ), while for
the coarse domains, the initial magnetization remains at zero
for H below coercivity (Hc ), then rises sharply to saturated
magnetization Ms when H is nearly equal to Hc . These observations enable us to relate the domain sizes in the demagnetized states to the initial magnetization curves of EHE measurements for MO material.
2.1 Sample preparations
Several magneto-optical recording materials have been
observed.4, 5) The sample presents here is a perpendicular anisotropy multilayered Al2 O3 /Mo(20 nm)/Pt(20 nm)/[Co
(0.3 nm)/Pt(1 nm)×30.
Multilayered of Co/Pt film was prepared using a molecular beam epitaxy (MBE) apparatus (Vacuum Product MBE930). The base pressure of the MBE system is lower
than 2 × 10−10 Torr. Buffer layer of Mo(20 nm)/Pt(20 nm)
was firstly deposited on epitaxial grade Al2 O3 (11–20) substrate (purchased from Kyocera Co.), and subsequently Co
and Pt layers were alternatively deposited at room temperature. The Mo and Pt seeding layers were grown at 900◦ C
and 500◦ C, respectively, and with deposition rates of about
0.05–0.2 Å/s. The deposited layer structures are: Al2 O3 (11–
20)/Mo(20 nm)/Pt(20 nm)/[Co(0.3 nm)/Pt(1 nm)]×30. To enable the growth of high-quality ML films, the Al2 O3 (11–20)
substrates were chemically pre-cleaned and were then introduced into the growth chamber and out-gassed at ∼1050◦ C
for 1 hr under an UHV condition before initial deposition. It
is unlikely that residual gases played a role in determining the
epitaxial orientation and surface structure. During deposition,
the growth pressure was controlled below 5 × 10−9 Torr, the
deposition rates at about 0.1 Å/s, and the substrate temperatures at room temperature. The deposition rate and sample
thickness were calibrated by a quartz crystal monitor located
very closed to the sample holder. To retain the sample uniformity the sample holder was rotated with a constant speed of
about 30 rpm.
2.2 Sample characterizations
The surface structure and epitaxial orientation of ML
films were in situ determined by RHEED. The bulk structure of the ML films was measured by X-ray diffraction (XRD) using Cu Ka1 radiation. The RHEED and
X-ray diffraction measurements confirmed the following
epitaxial relations of the sapphire substrate, buffer layer
Mo/Pt and Co/Pt multilayers ({Co/Pt}×N): Al2 O3 (11–
20)/Mo(110)/Pt(111)/{Co/Pt}×N (111).
1687
1688
Jpn. J. Appl. Phys. Vol. 38 (1999) Pt. 1, No. 3B
Magnetic property was measured by a vibrating sample
magnetometer (VSM) and the magneto-optical properties
were measured with a loop tracer (LT) using detecting signals from the extraordinary Hall effect (EHE). Magnetic domain structures were observed by employing a polar Kerr microscope and a magnetic force microscope (MFM). A Digital Instruments Nanoscope IIIa MFM, equipped with phase
extender,6) was used in this study. The magnetic tip with a
CoCr-coated Si tip magnetized along the tip axis was used to
scan the magnetic domain structures in the tapping-lift mode.
Prior to the MFM measurements, the samples were demagnetized. Two methods had been employed to demagnetize
samples in this study. The first method to demagnetize a sample was by applying an in-plane magnetic field (H ) perpendicular to magnetization (M), whose strength (25 kOe) was
enough to bring M to the in-plane direction. After that the
applied field was removed, the film would become fully demagnetized. This demagnetized state was called the Hin state.
The second way of demagnetizing a sample was by applying
a perpendicular (to the film plane) magnetic field with a direction opposite to M with strength of coercivity Hc . This
demagnetized state was called the Hc state (i.e., at H = Hc ,
M = 0). During the demagnetization processes, the films
were saturated in one direction by applying a perpendicular
magnetic field greater than coercivity Hc . Then, a magnetic
field with opposite direction near the coercivity-Hc was applied and the magnitude of the field was kept constant. During the demagnetization process, a polar Kerr microscope was
used to monitor the developing magnetic stripes and domains.
After the magnetic domains had been developed, the magnetic
field was turned off, and the sample was moved to the MFM
facility for further observation.
Fig. 1. Hysteresis loops and initial curves with (a) in-plane demagnetization and (b) perpendicular demagnetization. Notice that the difference of
initial curves between two loops.
T. W U et al.
3. Results and Discussions
3.1 Results
Figure 1 shows the initial magnetization curves obtained
by measurements of EHE signal for both the Hin state and the
Hc state for a Co/Pt sample. Although the hysteresis loops
were measured from the same thin film, we found that the
initial magnetization curve starts to change on Fig. 1(a) occurs earlier than that on Fig. 1(b). This becomes evidently
by comparing the slope of the initial curves of those loops,
i.e., Fig. 1(a) is steeper than that in Fig. 1(b). The domain
patterns are shown in Fig. 2(a) and Fig. 2(b) for the Hin
state and the Hc state respectively. The corresponding threedimensional (3-D) views of domain pattern are displayed in
Fig. 3. It can be seen that domains in Fig. 2(a) are finer than
domains in Fig. 2(b). The domain structures can be examined clearly from 3-D views. We observed that the average
domain size (about 0.68 µm) is about one-half smaller and
Fig. 2. Demagnetized state of the magnetic domain image scanned from
MFM at (a) the Hin state and (b) the Hc state. Notice that domains in (a)
are finer than in (b).
Jpn. J. Appl. Phys. Vol. 38 (1999) Pt. 1, No. 3B
Fig. 3. Demagnetized state of three-dimensional view of the magnetic domain image scanned from MFM at (a) the Hin state and (b) the Hc state.
finer (see Fig. 2(a)) in the Hin state, which is corresponding
to the in-plane demagnetization procedure, than the average
domain size (about 1.38 µm) in the Hc state (see Fig. 2(b)),
which is corresponding to the perpendicular demagnetization
procedure.
3.2 Discussions
The experimental results indicated that the initial magnetization curves for the in-plane field demagnetization procedures exhibit magnetization reversal starts from smaller applied field than the perpendicular demagnetization procedure.
The main reason for this situation appears to be the differences of distribution of domain, especially the domain size.
While, the differences of domain size is mainly due to the
different mechanisms at work in the various demagnetization
processes.
Here, we discuss the details of the various demagnetiza-
T. W U et al.
1689
tion mechanisms and address two questions: i) why does an
in-plane field produce a demagnetized state with smaller domains than a perpendicular field does; ii) why are the initial
magnetization curves different for the two different demagnetized states. To answer the first question, let us first consider
the mechanism at work when the demagnetization in the film
plane has been produced using a strong in-plane field that is
suddenly removed. Since both up and down directions are
symmetric energy minimum for the in-plane oriented magnetic dipoles, each magnetic dipole has an equal probability of being up or down. In this case, the number of seeds
that may form nucleation centers and develop into domains
reaches the maximum.7) This favors the formation of smaller
domains. The final size of the domains in the steady state
is determined by minimizing the total energy. Note that the
domain runs out into the maze-like pattern, and it is quite similar to the magnetic bubble domain pattern of uniaxial platelet
in the absence of an external magnetic field. In the case of
a perpendicular applied field, demagnetization usually begins
from a few defects that emerge as nucleation centers when
the applied field is near nucleation coercivity. As a result,
coarse domains grow up via wall motion from those defects
and construct a demagnetized state on the film. Formation of
new nucleation centers is less likely in this latter case because
in most parts of the films, the local nucleation coercivity is
higher than the wall motion coercivity.
The second question, the difference between the two
kinds of initial magnetization curves, can be answered
by the well-known minimum stable domain diameter,8–11)
d = cσw /(Ms Hc ), where c is a positive proportion factor
(whose value depends on the shape of the domain and is on
the order of 1), σw is the wall energy density, Ms is the saturation magnetization, Hc is coercivity, and d is the diameter of the domain. The domain wall feels a collapsing force,
Hcollap = −cσw /(Ms d), arising from the domain wall energy,
where the sigh means that the field is opposite to the domain
magnetization. The effect of the demagnetizing field is small
and can be neglected, because we are only concerned with a
fully or partially demagnetized state. Then, due to the domain’s collapsing field, the minimum external field Hext required to reverse a domain of diameter d is that
σw
(1)
Hext = Hc − c
Ms d
This equation shows that a small domain can be reversed by
an Hext much smaller than Hc and thus explains why the magnetization reversal starts at smaller value of Hext . If the domain is large enough so that the magnitude of Hcollap approximates to zero, the external field must reach nearly to Hc to
make any magnetic moment reversal. By definition, this field
is also strong enough to cause the magnetization across the
whole film to be reversed. This explains the flat part of the
initial curve for the demagnetized state obtained by a perpendicular field and the sharp change as magnitude of Hext near
to Hc .
4. Conclusions
Although we show experimental results only for one Co/Pt
sample in this paper, the initial magnetization curves and
magnetic domain pattern in demagnetized states of several
Co/Pt samples have been studied. In all samples, including
1690
Jpn. J. Appl. Phys. Vol. 38 (1999) Pt. 1, No. 3B
Co/Pd and TbFeCo samples as previous publish indicated,3)
it appears that the initial magnetization curves for the inplane field demagnetization procedures exhibit magnetization
reversal starts from smaller applied field than those perpendicular demagnetization procedure. The main reason for this
situation appears to be the differences of domain size. While,
the differences of domain size is mainly due to the different mechanisms at work in the various demagnetization processes. The in-plane demagnetized field produced fine domain distributions, while perpendicular field demagnetization
yielded coarse domain patterns.
In addition, by taking Hext = 0 from eq. (1), the smallest
domain size that is allowed to exist in the demagnetization
state is
cσw
(2)
dmin =
Ms Hc
Domains whose sizes are smaller than this minimum size will
collapse under zero applied field. The measured magnetic parameters of Co/Pt sample studied in this paper, Hc = 500 Oe
and Ms = 300 emu cm−3 , and if we take value of cσw = 9 erg
cm−2 as a reasonable estimate, then the minimum domain size
calculated from eq. (2) is about 0.6 µm. This is consistent
with the measured average magnetic domain size in the demagnetized state from domain image of Fig. 2(a). In other
T. W U et al.
words, the magnetic domain’s pattern for the sample demagnetized by an in-plane magnetic field can be used to estimate
the domain wall energy as long as the domain size is known.
Acknowledgement
The National Science Council of Republic of China supports this work, under contract NO. NSC 87-2112-M-224001.
1) B. E. Bernacki, Te-ho Wu and M. Mansuripur: J. Appl. Phys. 73 (1993)
6838.
2) C.-J. Lin, C. Suit and R. H. Geiss: J. Appl. Phys. 63 (1988) 3835.
3) Te-ho Wu: J. Appl. Phys. 81 (1997) 5321.
4) Te-ho Wu and M. Mansuripur: Proc. MORIS’92, J. Magn. Soc. Jpn. 17
(1993) S1, 131.
5) Te-ho Wu: Proc. MORIS’94, J. Magn. Soc. Jpn. 19 (1995) S1, 209.
6) K. Babcock, V. B. Elings, J. Shi, D. D. Awschalom and M. Dugas: Appl.
Phys. Lett. 69 (1996) 705.
7) M. W. Muller: Phys. Rev. 122 (1961) 1485 .
8) A. A. Thiele: J. Appl. Phys. 41 (1970) 1139.
9) W. D. Doyle, G. K. Goldberg and W. E. Flannery: IEEE Tran. Magn.
MAG-6 (1970) 548.
10) M. H. Kryder, W. H. Meiklejohn and R. E. Skoda: Proc. SPIE 420
(1983) 236.
11) D. Rugar, C.-J. Lin and R. Geiss: IEEE Tran. Magn. MAG-23 (1987)
2263.