DSpace at VNU: Thickness dependence of the phase transformation in FePt alloy thin films

ARTICLE IN PRESS

Journal of Magnetism and Magnetic Materials 277 (2004) 187–191

Thickness dependence of the phase transformation in FePt
alloy thin films
P.T.L. Minha, N.P. Thuya,b,*, N.T.N. Chanc
a

International Training Institute for Materials Science (ITIMS), Dai hoc bach khoa, 1 Dai Co Viet, Hanoi, Viet Nam
b
Faculty of Technology, Vietnam National University, Hanoi, Viet Nam
c
Institute of Engineering Physics, Hanoi University of Technology, Hanoi, Viet Nam
Received 4 July 2003; received in revised form 17 October 2003

Abstract
FePt alloy thin films of different thickness have been prepared and studied. Both X-ray analysis and magnetization
measurements have been used to detect the FCC (A1)–FCT (L10) phase transformation due to annealing in these films.
It was found that the ordering process in the thick samples takes place at much lower temperature in comparison to the
thinner ones. The observed phenomenon can be understood taking into account the kinetics of the FCC–FCT phase
transformation and grain growth. The obtained experimental results suggest the existence of an optimal annealing
temperature for each defined sample thickness.
r 2003 Elsevier B.V. All rights reserved.
Keywords: Ordering kinetics; Grain growth; Magnetization process; Fe–Pt thin film

1. Introduction
Recently, Fe–Pt alloy thin films have attracted
much attention for their potential application as
high-density magnetic recording materials. One of
the reasons for that is the possibility to develop

high coercivity in the ordered face-centered-tetragonal (FCT) L10 FePt phase. To get this phase a
post-annealing heat treatment at relatively high
temperature (usually larger than 600 C) is required which results in large grains and thus highly
exchange coupling [1,2]. Several attempts have
been made to tailor the film microstructure and
*Corresponding author. International Training Institute for
Materials Science (ITIMS), Dai hoc bach khoa, 1 Dai Co Viet,
Hanoi, Viet Nam. Tel.: +84-4-8692518; fax: +84-4-8692963.
E-mail address: (N.P. Thuy).

magnetic properties. Non-magnetic materials such
as BN, B2O3, SiO2, Ag, C have been added to
reduce the magnetic coupling and improve the
orientation of the magnetic grains in the films
[3–6]. One of the recent trends is to reduce the
FCC–FCT phase transformation temperature of
the given alloy thus to avoid the difficulties caused
by the high temperature of post-annealing. There
are some approaches to this problem such as use of
some additives like Cu, Zr [5,7–9], preparation of
the samples with the substrates at elevated
temperature or in type of multilayers [10]. However, less attention has been paid to the question
how this annealing regime depends on the film
thickness.
In this work X-ray analysis and magnetization studies have been carried out in a series of
nearly equiatomic FePt films of different thickness

0304-8853/$ - see front matter r 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2003.10.032

ARTICLE IN PRESS
188

P.T.L. Minh et al. / Journal of Magnetism and Magnetic Materials 277 (2004) 187–191

(ranging from 20 to 100 nm). We show that the
ordering process clearly depends on sample’s
thickness. It is a phenomenon, which could
provide better understanding of their ordering
kinetics.

t = 40 nm
As-deposited

t = 20 nm
Annealed

The nearly equiatomic FePt thin films have been
prepared by RF-sputtering on the Si[1 0 0] substrate. The base pressure was 2 Â 10À6 Torr. The
films were deposited with a fixed argon pressure of
5 Â 10À3 Torr. A composite target was made
putting some Fe (99.95% purity) pieces on the Pt
target (99.99% purity). The film thickness was in
the range from 20 to 100 nm. The samples were
annealed under various conditions in Ar atmosphere with the subsequent quenching in the ice
water. The energy dispersion spectrum (EDS) and
X-ray diffractometer (XRD) have been used to
study the film composition and microstructure.
The film thickness was determined by low-angle Xray diffraction. Magnetization measurements have

been carried out using a vibrating sample magnetometer (VSM) with magnetic field up to 13 kOe.

3. Results and discussion
In Fig. 1, the X-ray patterns of the as-deposited
and annealed FePt films of different thickness are
demonstrated. The films were annealed at 500 C
for 15 min in Ar atmosphere followed by quenching in ice water. It is clearly seen that the asdeposited sample has a face-centered-cubic (FCC)
crystalline structure with the characteristic (1 1 1)
peak. After annealing for the thinnest film with
t ¼ 20 nm the only (1 1 1) peak is observed meaning that the crystalline structure in this sample still
remains FCC and no transformation has taken
place. In case of the thicker samples, with t=50
and 100 nm, which were annealed at the same
conditions the situation is quite different. The
appearance of the (1 1 0) and the split of (2 0 0)
peaks in the X-ray diagrams of these samples have
been observed which indicate the occurrence of the
phase transformation from the FCC to the FCT

Intensity (a.u)

2. Experimental procedure

t = 50 nm
Annealed

t = 100 nm
Annealed

(200)

(001)

20

(111)

(110)

30

40
2θ (degree)

(002)

50

Fig. 1. X-ray patterns of the as-deposited FePt thin films and
those annealed at 500 C for 15 min. The film thickness is
indicated in the figure. The vertical solid lines are standard
peaks for bulk polycrystalline sample of FCT structure.

crystal structure. A clear decrease of the full-width
at half-maximum (FWHM) of (1 1 1) peak with the
sample thickness is a proof of the fact that in the
thick films the grain growth is occurred at much
higher rate.
Fig. 2a,b presents the initial magnetization
curves measured in the direction parallel to the

film plane for Fe–Pt thin films of different
thickness annealed at 425 and 500 C for 15 min,
respectively. As the film thickness increases the
mechanism of the magnetization process exhibit
tendency to change from domain nucleation to
domain wall pinning type. The thin samples with
to75 nm get saturation state at very low field
whereas in the thicker samples the magnetization
slowly increases at low field and only at higher
field it augments faster. It is interesting to note a
close correspondence between the observed tendency in mechanism of the magnetization process
and the hysteresis curves as shown in Fig. 3: high


ARTICLE IN PRESS
P.T.L. Minh et al. / Journal of Magnetism and Magnetic Materials 277 (2004) 187–191
1.0

189

1200

t = 20 nm
parallel
perpendicular

0.8

1

0

M/M max

M/M max

0.6

0

t = 20 nm
t = 25 nm
t = 40 nm
t = 50 nm
t = 75 nm
t = 100 nm

0.4

0.2

-1200

1
H (kOe)

2

1
H (kOe)

2

1200

t = 25 nm

0.0

1

0

M/M max

(a)
1.0

0

-1200
1200

0.8

t = 20 nm
t = 25 nm
t = 40 nm
t = 50 nm
t = 75 nm

t = 100 nm

0.4

0.2

0

0
H (kOe)

1200

8

t = 75 nm

12

H (kOe)

Fig. 2. Initial magnetization curves for FePt thin films with
different thickness t as indicated in the figure. All the films were
annealed for 15 min at temperature of (a) 425 C, (b) 500 C.

0
1
M/M max

(b)

4

13

-1200

0.0
0

1
M/M max

M/Mmax

0.6

M (emu/cm3)

t = 40 nm

0

-1200

H (kOe)

13

H (kOe)

13

1200

t =100 nm

0

1

M/M max

coercivity value corresponds to a pinning mechanism whereas the low coercivity one—a nucleation
mechanism. As can be seen in this figure, when the
sample thickness rises a considerable increase in
the coercivity is obtained. A thin sample only has
the coercive value of hundreds oersteds while a
large value of coercivity of about 7.5 kOe has been
developed in the thick sample with t ¼ 100 nm
although they have been all annealed at the same
conditions. As mentioned above in the X-ray
analysis the phase transformation evolution presents the film thickness dependence: in the studied
ranges of thickness the thicker the sample is, the
higher the transformation rate and thus the larger

0

-1200

0

1

3

H (kOe)
Fig. 3. Hysteresis loops for FePt alloy thin films annealed at
500 C for 15 min. Two curves corresponding to measurements
parallel and perpendicular to the film plane are demonstrated.
The film thickness is indicated in the figure. The insets show
initial magnetization curves.


ARTICLE IN PRESS
190

P.T.L. Minh et al. / Journal of Magnetism and Magnetic Materials 277 (2004) 187–191

the FCT volume fraction. This suggests that the
appearance of the (1 1 0) peak and the grain
growth observed in X-ray experiments are correlated with the tendency to change to the domain
wall pinning type of magnetization process leading
to high value of coercivity in the samples as shown
in Fig. 3. These results are in good agreement with
those reported by Ristau et al. [11] on the
relationship between the high coercivity in the
annealed Fe–Pt alloy films and the volume fraction
of the FCT phase. Thus, our obtained experimental results showed that at the same annealing
temperature, there is a critical thickness above it

the fct-phase is formed leading to a high value of
coercivity.
Generally, the ordering kinetics can be described
by standard Johnson–Mehl–Avrami equation
f =1Àexp(Àktn ) where f is the transformed
volume fraction, t is time, k is a constant and n
is the Avrami exponent. Both k and n depend on
the nucleation and diffusion rates, which in their
turn depend on the activation energy and temperature through Arrhenius equation, possible
growth mechanism and spatial dimensionality of
the growing region. In thinner films transformation process appears to be limited by the sample
thickness. Thus, the Avrami exponent should be
low that leads to longer time transformation or
higher temperature required for ordering to be
completed [12]. In our case the temperature needed
for the ordering in the thick sample with
t ¼ 100 nm is about 400 C. At this temperature
the diffusion rate is rather low and in the order of
10À11 cm/h. This value cannot explain the observed phenomenon.
It is well known that in the disorder–order
transformation high activation energy is needed to
promote nucleation and growth processes. From
the obtained experimental results we deduce that
the driving force of the disorder–order transformation in thick films is larger than that in the
thinner films. In other words the barrier energy for
nucleation of new phase in the case of thicker
samples is lower than that of thin samples.
Therefore, the ordering process in the thick
samples occurs more easily at low temperature.
These results are certainly related to the thickness

dependence of the film microstructural character-

istics such as film density and film intrinsic stress.
Study on the kinetics of grain growth also
supports the above explanation. As reported by
Barmak et al. [12] the activation energy for phase
transformation is much higher than that for grain
growth in FePt thin films. At low annealing
temperatures the grain growth is dominated. In
the early stage of annealing the major driving force
for this process is surface energy reduction. For
fcc metal and alloy the (1 1 1) plane represents
the lowest surface energy plane. Then the randomly oriented grains are consumed until each
grain has mostly [1 1 1] oriented neighbors leading
to decrease of the interface energy due to
coherence between matrix and product phase.
Thus, the ordering process is triggered and
proceeded as a result of the grain coalescence in
the films [4,13].
In summary, the magnetization process and
coercivity variation in the FePt alloy thin films of
different thickness have been analyzed in close
relation to the kinetics of the FCC–FCT phase
transformation and grain growth. The phase
transformation observed at low temperature in
thick films is attributed to the relatively high
value of the driving force or low value of the
barrier to nucleation for the ordering, which
strongly depend on the film microstructure and
the kinetics of grain growth. Further study on

the phase transformation evolution with the
film microstructure and its thermodynamic aspect
is needed to fully understand the observed
phenomenon.

Acknowledgements
This work was supported by the State Program
on Fundamental Research under Grant No.
421001.

References
[1] Y.K. Takahashi, M. Ohnuma, K. Hono, Jpn. J. Appl.
Phys. 40 (2001) L1367.
[2] P.T.L. Minh, N.P. Thuy, N.D. Van, N.T.N. Chan, J.
Magn. Magn. Mater. 239 (2002) 335.


ARTICLE IN PRESS
P.T.L. Minh et al. / Journal of Magnetism and Magnetic Materials 277 (2004) 187–191
[3] P.T.L. Minh, N.D. Van, N.P. Thuy, L.T. Nguyen, J.C.
Lodder, P.D. Thang, N.T.N. Chan, Physica B 327 (2003)
360.
[4] S.C. Chen, P.C. Kuo, A.C. Sun, C.T. Lie, W.C. Hsu,
Mater. Sci. Eng. B88 (2002) 91.
[5] C.L. Platt, K.W. Wierman, E.B. Svedberg, R. van de
Veerdonk, J.K. Howard, A.G. Roy, D.E. Laughlin, J.
Appl. Phys. 92 (2002) 6104.
[6] T. Saito, O. Kitakami, Y. Shimada, J. Magn. Magn.
Mater. 239 (2002) 310.
[7] T. Maeda, T. Kai, A. Kikitsu, T. Nagase, J. Akiyama,

Appl. Phys. Lett. 80 (2002) 2147.

191

[8] Y.K. Takahashi, M. Ohnuma, K. Hono, J. Magn. Magn.
Mater. 246 (2002) 259.
[9] S.R. Lee, S. Yang, Y.K. Kim, J.G. Na, J. Appl. Phys. 91
(2002) 6857.
[10] Y. Endo, N. Kikouchi, O. Kitakami, Y. Shimada, J. Appl.
Phys. 89 (2001) 7065.
[11] R.A. Ristau, K. Barmak, L.H. Lewis, K.R. Coffey, J.
Appl. Phys. 86 (1999) 4527.
[12] K. Barmak, J. Kim, S. Shell, E.B. Svedberg, J.K. Howard,
Appl. Phys. Lett. 80 (2002) 4268.
[13] R.A. Ristau, K. Barmak, K.R. Coffey, J.K. Howard, J.
Mater. Res. 14 (1999) 3263.