Journal of Magnetism and Magnetic Materials 409 (2016) 3438

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials

journal homepage: www.elsevier.com/locate/jmmm

Spin wave and percolation studies in epitaxial La2/3Sr1/3MnO3 thin

lms grown by pulsed laser deposition
A. Ettay a, R. Moubah a,n, E.K. Hlil b, S. Colis c, M. Lenertz c, A. Dinia c, H. Lassri a
a
LPMMAT, Facult des Sciences Ain chock, Universit Hassan II de Casablanca, B.P. 5366 Casablanca, Morocco
b
Institut Nel, CNRS, Universit Joseph Fourier, BP 166, 38042 Grenoble Cedex 9, France
c
Institut de Physique et Chimie des Matriaux de Strasbourg (IPCMS), UMR 7504 UDS-CNRS (UDS-ECPM), 23 rue du Loess, BP 43, F-67034 Strasbourg Cedex
2, France

art ic l e i nf o a b s t r a c t

Article history: We investigate the magnetic and transport properties of high quality La2/3Sr1/3MnO3 thin lms grown by
Received 20 December 2015 pulsed laser deposition. X-ray diffraction shows that the deposited lms are epitaxial with the expected
Received in revised form pseudo-cubic structure. Using the spin wave theory, the temperature dependence of magnetization was
20 January 2016
satisfactory modeled at low temperature, in which several fundamental magnetic parameters were
Accepted 15 February 2016
Available online 23 February 2016
obtained (spin wave stiffness, exchange constants, Fermi wave-vector, MnMn interatomic distance). The
transport properties were studied via the temperature dependence of electrical resistivity [(T)], which
Keywords: shows a peak at Curie temperature due to metal to insulator transition. The percolation theory was used
Spin wave theory to simulate (T) in both the ferromagnetic and paramagnetic phases. Reasonable agreement with the
Percolation theory
experimental data is reported.
epitaxial LSMO lms
& 2016 Elsevier B.V. All rights reserved.
Magnetism
Electrical properties

1. Introduction by a ferromagnetic to paramagnetic transition evidencing the

La2/3Sr1/3MnO3 (LSMO) has been extensively studied in recent
years due to its particular physical properties, with high potential
applications in spintronic devices [1], such as magnetic memory-
storage cells or magnetic-eld sensors. This compound exhibits
some interesting properties, such as a half-metallic character, co-
lossal magnetoresistance, and a high Curie temperature (TC,bulk
370 K) [25]. In such systems, the magnetic interaction is de-
scribed by the double-exchange mechanism between the Mn3
and Mn4 ions [6], where the coupling is induced by the electron
hoping between two partially lled d shells. The existence of
Mn3 and Mn4 mixed valence is essential to interpret both the
ferromagnetism and metallic conductivity states. However, it was
demonstrated that the double-exchange interaction alone cannot
fully explain the magnetic interaction in LSMO. It was suggested
that other mechanisms may inuence the exchange interaction,
such as the charge ordering, the size and the disorder at the cation
site [7,8], the oxygen vacancies [9] or the polarons effect induced
by strong electronphonon interaction originating from the Jahn
Teller distortion [10]. On the other hand, a relationship between
electrical and magnetic properties has been demonstrated. In fact,
LSMO exhibits a metal-insulator transition, which is accompanied
n
Corresponding author.
E-mail address: reda.moubah@hotmail.fr (R. Moubah).
strong correlation between the magnetic and transport properties.
However, the simulation of the change of resistivity as a function
of temperature is limited by the metal-to-insulator transition. To
be able to model (T) in the entire temperature window, Li et al.
[11] have developed a percolation model based on phase segre-
gation, which we use in this present study. In order to better un-
derstand the properties and exploit LSMO for applications a de-
tailed fundamental understanding of the magnetic and electrical
properties is needed [12,13]. Here, we report on the magnetic and
transport properties of high quality LSMO thin lms using spin
wave and percolation theories, respectively [1416].
2. Experimental methods
LSMO thin lms were deposited using pulsed laser deposition
technique on SrTiO3(001) (STO) substrates [17]. The background
pressure was about 10  8 mbar. Before deposition, the substrate
was heated at 900 C for 30 min to clean the surface. During
growth, the substrate temperature was kept at 750 C and the
oxygen pressure at 0.5 mbar. A KrF excimer laser ( 248 nm),
with a frequency of 10 Hz and an energy density of 1.5 J/cm2 was
used for ablation [18,19]. The lm thicknesses were about 38 nm.
The LSMO target was prepared via conventional solid state reac-
tion technique using stoechiomteric mixture of La2O3, SrCO3, and
MnCO3 powders used as starting materials. The mixture was

http://dx.doi.org/10.1016/j.jmmm.2016.02.046
0304-8853/& 2016 Elsevier B.V. All rights reserved.
 A. Ettay et al. / Journal of Magnetism and Magnetic Materials 409 (2016) 3438 35

annealed at 1200 C during 16 h. Then, the obtained powder was 10

8

SrTiO3 (002)
pressed into a disk by applying a pressure of 600 bar, and sintered (a)

SrTiO3 (001)

SrTiO3 (003)
at 1450 C during 16 hours in air. The crystalline structure was 10
7

investigated by X-ray diffraction (XRD) using a Rigaku Smartlab

LSMO (024)
(9 kW) diffractometer equipped with a monochromatic Cu K1 6

LSMO (012)
10
source ( 0.154056 nm) and a Ge(220) 2-bounce front mono-

Intensity (cps)
chromator. Magnetic measurements were carried out on a super- 5

LSMO (036)
10
conducting quantum interference device (SQUID) magnetometer.
Electrical resistivity as a function of temperature was carried out 10
4

using a four-probe method.

3
10

2
3. Results and discussion 10

Fig. 1a shows the XRD pattern of a LSMO thin lm grown on 10

1

SrTiO3(001) substrate. Only the (0 k 2l) reections of the rhom- 20 40 60 80

bohedral LSMO phasei are observed suggesting that the lm has a 2 angle ()
well dened texture. No spurious phases could be identied in the
detection limit of the XRD technique. The average grain size along

SrTiO3 (-20-2)
SrTiO3 (-202)
(b)

SrTiO3 (202)

SrTiO3 (022)
the growth axis was calculated using DebyeScherrer formula: 6
G= cos where, G is the grain size, is the wavelength of the x-ray
0.9 10
source, is the full width at half maximum (FWHM) of the dif-
5
fraction peaks. In order to determine the FWHM, the LSMO (036) 10
peak was tted using a Gaussian function. The calculated grain size 90

Intensity (cps)
is found to be 257 2 nm. In order to check if there is an in-plane 10
4

structural relation between the substrate and the layer, scans

LSMO (-123) b
LSMO (-123) a
LSMO (113) b
LSMO (113) a
were performed for both the substrate and the layer. For this 3
purpose the {2,0,2} and {1,1,3} plane families of STO and LSMO 10
were analyzed, respectively. The choice of the {1,1,3} LSMO plane
2
family is motivated by the fact that no STO peak can be found 10 45
nearby. As expected the scan showed four peaks corresponding to
the cubic (fourfold symmetry) structure of STO (Fig. 1b). Although 1
10
the symmetry of the LSMO structure is lower, four peaks are ob-
served as well, suggesting the existence of at least two variants. It 0 60 120 180 240 300 360
appears that there is an in-plane relation between the substrate angle ()
and the layer, and therefore, LSMO is perfectly epitaxied on the
Fig. 1. (a) X-ray diffraction pattern of a 25-nm-thick LSMO lm grown on
STO substrate. This epitaxial relation can be written (for two LSMO SrTiO3(001) substrate. (b) Scans of the STO {2,0,2} and LSMO {1,1,3} planes
variants) as [1 710]STO(001)[2  10]LSMO(012). This relation is showing 90 periodicities for the monovariant STO and multivariant (a and b)
compatible with the very small lattice mismatch of about 0.5% LSMO.
Note: Although at rst glance the (113) and (  123) LSMO planes do not seem to be
existing at the STO(001)/LSMO(012) interface and with the almost
part of the same plane family, this is indeed the case. In the four indices notation
square lattice inside the LSMO(012) plane (angles of 90.4 and (hexagonal structure) these planes are (1123) and (  1213) which are indeed
89.6 instead of 90). part of the same plane family.
The change of magnetization as a function of temperature of a
LSMO thin lm is shown in Fig. 2. The measurements were made
upon warming the lm after a zero eld cooling (ZFC) starting
from 5 K in a eld of 500 Oe parallel to the [100]SrTiO3(001) di-
rection. The TC is found to be around 340 K. This value is lower 1.0
than that reported for LSMO bulk samples (370 K). Song, et al.
[20] have shown that TC depends sensitively on the deposition 0.8
parameters such as the oxygen pressure during growth or the laser
beam defocus on the LSMO target. Thus, the lower TC can be at-
M(T)/M(5K)

tributed to these parameters. 0.6

For ferromagnetic materials, the temperature dependence of
magnetization follows Block's law (T3/2) which is associated with
0.4
the thermal excitation of spin waves. Using spin wave theory, the
change of magnetization as a function of temperature for ferro-
magnets can be written as [21]: 0.2

M (5 K) M (T )
= BT 3/2
M (5 K) (1) 0.0
0 100 200 300 400
i
T (K)
Although the structure of LSMO is rhombohedral (R-3c), the crystalline
structure of LSMO is often described as pseudocubic. In this case the (024)LSMO Fig. 2. Temperature dependence of normalized magnetization for LSMO thin lms
peak in the rhombohedral notation corresponds to the (002)LSMO peak in the obtained at an applied magnetic eld of 500 Oe. The magnetic eld was parallel to
pseudocubic notation. the [100]SrTiO3(001) direction.
36 A. Ettay et al. / Journal of Magnetism and Magnetic Materials 409 (2016) 3438

0.008
1.0
Experimental data
Linear fit Experimental data
0.8 0.006
Ferromagnetic fit

Resistivity ( .cm)
M (T)/M (5 K)

Paramagnetic fit
0.6 Percolation theory fit
0.004

0.4 -5 -3/2
B = 5.97 10 K
0.002
0.2

0.0 0.000
0 100 200 300 400
0 1000 2000 3000 4000 5000 6000 7000
3/2 3/2 T (K)
T (K )
Fig. 4. Change of resistivity as a function of temperature of LSMO thin lms
Fig. 3. Variation of magnetization as a function of T3/2recorded by applying a eld without applying a magnetic eld. Solid red, black and green lines are the ts
of 500 Oe (sphere symbols). Red solid lines display the corresponding linear t obtained using Eqs. (7), (8), and (9), respectively.
obtained using spin wave theory.

where B is the spinwave constant and M(5 K) is the magnetiza-
tion at 5 K. Eq. (1) is a good estimate of low temperature magne-
tization for ferromagnetic materials. The B parameter expressed in
Eq. (1) is linked to the spin wave stiffness constant (D) by [21]:
g B B 3/2
B = 2.612
M (5 K) 4D
where g is the spectroscopic g-factor (gMn 2), kB is the Boltzmann's
constant and mB is the Bohr magneton. The change of magnetization
as a function of T3/2 is displayed in Fig. 3. The linear t of the ex-
perimental data allows obtaining the B parameter using Eq. (1),
which is dened as the slope of the M versus T3/2 curve. As can be
observed, Eq. (1) ts well the low temperature data. For tempera-
tures higher than 180 K, (T3/2  2415 K) a deviation from linearity is
seen. The concept of spin wave theory assumes that the local mag-
netic order uctuates with increasing temperature. Thermally in-
duced spin waves of long wavelengths are expected to dominate at
low temperature as a result of the low energy excitations [22,23].
With the increase of temperature, the uctuations increase giving
rise to a deviation from linearity of the M versus T3/2 curve.
On the basis of the obtained B values and using Eq. (2), the
constant exchange stiffness D was deduced (Table 1). The re-
lationship between D and TC has been studied by Katsuki and
Wolhfarth [24] based on the itinerant electron model. Using the
effective mass approximation, these authors have found the fol-
lowing correlation [24]:
kB TC
D=
6 2 kF2
where kF is the Fermi wave-vector.
According to Heisenberg model, D can also be expressed by the
following formula [25]:
where rMnMn is the distance between nearest magnetic ions
(Mn3 , Mn4 ) and SMn is the spin moment of Mn ions. Using Eq.
(4), it is found that rMnMn is about 3.99 70.10 . This value is in a
line with that deduced from the structural analysis using XRD
technique and which is found to be rMnMn Ea E3.87 . The good
agreement with the experimental results can be considered as a
(2) validation of the approach used in this study.
From the coefcient D mentioned above, it is possible to cal-
culate the exchange constant A (Table 1). The A parameter is
linked to D by the following formula [26]:
M (5 K) D
A (5 K) = .
2gB (5)
With the use of the mean eld theory, the exchange interaction
between a Mn ion and its nearest neighbors can be expressed as [27]:
3kB TC
Jij =
2zS (S + 1) (6)
We found that Jij 0.1mRy
In the following we describe the transport properties of LSMO
thin lms. Fig. 4 shows the experimental temperature dependence
of LSMO resistivity in the temperature range from 10 to 400 K. As
can be observed, the behavior of (T) is typical for manganites,
showing a metal to insulator transition with a peak at TMI 348 K.
We note that the magnitude of (T) is smaller than that reported
for bulk LSMO samples [28] which can be attributed to the growth
conditions or the nite size effects. At low temperature, ferro-
magnetic phase, (T) can be approximated using a formula which
(3) includes some scattering mechanisms [29]:
FM (T ) = o + AT 2 + BT.5 (7)
where o is the residual resistance, the term AT is usually attributed
2

2
to the single-magnons scattering contribution, and BT5 is a small
kB rMn Mn TC contribution due to the electron-magnon scattering processes.
D=
2 (SMn + 1) (4) For temperatures higher than that of the metal-insulator
transition (paramagnetic phase), (T) can be expressed by the
Table 1 thermo-activated law using the following formula [30]:
Some magnetic parameters of LSMO lm extracted from spin wave theory.
Ep
M (emu/cm3) at B (10  5 K  3/2) D (meV 2) A (10  8 erg/cm) kF (  1)
PM = CT exp
kB T (8)
5K at 5 K at 5 K

where Ep is the activation energy, and is a constant. As shown in

600 6 83.6 21.6 0.36
Fig. 4, (T) is well tted for both the low and high temperature
 A. Ettay et al. / Journal of Magnetism and Magnetic Materials 409 (2016) 3438 37

Table 2
Deduced parameters from the t of (T) using Eq. (7) for the ferromagnetic phase.
1.0
Equation FM(T) o AT2 BT4.5

Parameters Values
0.8
4
FM 0 [ cm] 3.11  10
FM A [ cm/K2] 7.95  10  9
FM B [ cm/K4.5] 1.14  10  15
0.6

f
Table 3 0.4
Deduced parameters from the t of (T) using Eq. (8) for the paramagnetic region.

Equation Ep
PM = CT exp 0.2
kB T

0 100 200 300 400

Parameters Values
T (K)
PM C [ cm/K] 3  10  6
PM Ep/kB [K] 606.12 Fig. 5. Thermal variation of the simulated change of the ferromagnetic phase vo-
lume fraction for a LSMO thin lm.

Ep
Table 4 ( )
(T ) = o + AT 2 + BT 5 f + C T (1 f )
kB T (9)
Deduced parameters from the t of (T) using Eq. (9) for the percolation theory.

Equation
( + AT2 + BT 5)
0
exp
(T)
U T
1 + exp 0 1
kB T TMI
Parameters
Percolation theory 0 [ cm]
A [ cm/K2]
B [ cm/K4.5]
C [ cm/K]
Ep/kB [K]
U0/kB [K]
regimes. From the ts, the obtained values of the parameters used
in Eqs. (7) and (8) are listed in Tables 2 and 3.
As mentioned above, the behavior of (T) is well described
by using Eqs. (7) and (8). However, in the vicinity of the metal
to insulator transition, there is a divergence between the two
models. This is due to the fact that the system should be con-
sidered as a phase foliated in a region of phase transition: the
ferromagnetic conducting phase with a small fraction of the
paramagnetic dielectric phase (below TMI), and the paramagnetic
insulating phase with a tiny portion of the ferromagnetic con-
ducting phase, which is prevailing above TMI. Therefore, in
such a phase-foliated system, the metal-insulator transition is a
percolation phase transition and (T) can be described using per-
colation theory [11]. By assuming a percolation character of
the metal to insulator transition, and considering that the
competition between the ferromagnetic and paramagnetic regions
plays a key role. The electrical resistivity can be expressed as:
(T) FM*f PM*(1 f). Where f and (1  f) are the volume
fractions of the ferromagnetic and paramagnetic phases, respec-
tively. The volume fraction function of the two phases satises the
1
Boltzman distribution: f
. Here, U(T) is the difference
U
1 + exp
kB T
in energy between the ferromagnetic and paramagnetic states,
which depends on temperature. We note that U(T) can be de-
veloped at the rst order to U(T)  U0 1 ( T
as the energy difference at low temperature (T{ TMI). Therefore,
(T) can be expressed in the whole temperature window using the
following formula [11]:
U0
From the t, we can observe that Eq. (9) describes satisfactory
T Ep
1
kB T TMI
*CT exp
kB T the (T) behavior in the whole temperature range, including the
U T region of phase transition. The obtained parameter values used in
1 + exp 0 1
kB T TMI
formula (9) are all presented in Table 4. These parameters are
Values comparable to those found by Khli et al. [31] in bulk
La0.8Ca0.2MnO3 using the same percolation model.
2  10  4 Fig. 5 displays the deduced temperature dependence of the
3.3  10  8 ferromagnetic phase volume fraction (f) for LSMO thin lms. One
 1.9  10  15
can notice that f is equal to 1 where the ferromagnetism is pre-
1.76  10  7
1856 vailing (low temperature), indicating the domination of the fer-
 3900 romagnetic fraction in the ferromagnetic region. With increasing
temperature, the ferromagnetic phase volume fraction decreases,
indicating that the paramagnetic contribution is increasing as the
temperature is increased. These results highlight the applicability
of the percolation model, which takes into account a transition
from ferromagnetic to paramagnetic states when the temperature
is elevated.
4. Conclusion
In summary, the magnetic and transport properties of epitaxial
LSMO thin lms were studied experimentally and theoretically.
The temperature dependence of magnetization was tted using
spin wave theory. Several magnetic constants were extracted using
simulations. The extracted results are in agreement with those
determined experimentally. The electrical resistivity as a function
of temperature was modeled in two separate regions: metallic and
insulating phases. However, these models diverge at metal-in-
sulator transition. To overcome this divergence, the electrical re-
sistivity was simulated in the entire temperature window, using a
phenomenological percolation model. Finally, this study will be
useful to understand the magnetic and transport properties of
LSMO compounds.
References
TMI ) .11 U0 is considered
[1] C. Moreno, C. Munuera, S. Valencia, F. Kronast, X. Obradors, C. Ocal, Nano Lett.
10 (2010) 3828.
[2] M. Bowen, M. Bibes, A. Barthlmy, J.P. Contour, A. Anane, Y. Lematre, A. Fert,
Appl. Phys. Lett. 82 (2003) 233.
38 A. Ettay et al. / Journal of Magnetism and Magnetic Materials 409 (2016) 3438
[3] J. Fontcuberta, B. Martnez, A. Seffar, S. Piol, J.L. Garca-Muoz, X. Obradors,
Phys. Rev. Lett. 76 (1996) 1122.
[4] A. Ozbay, E.R. Nowak, Z.G. Yu, W. Chu, Y. Shi, S. Krishnamurthy, Z.Z. Tang,
N. Newman, Appl. Phys. Lett. 95 (2009) 232507.
[5] A. Ettay, R. Moubah, A. Boutahar, E.K. Hlil, H. Lassri, J. Supercond. Nov. Magn.
29 (2016) 133.
[6] C. Zener, Phys. Rev. 82 (1951) 403.
[7] F. Damay, C. Martin, A. Martin, B. Raveau, J. Appl. Phys. 81 (1997) 1372.
[8] L.M. Rodriguez-Martinez, J.P. Atteld, Phys. Rev. B 54 (1996) 15622.
[9] A.K.M. AktherHossain, L.F. Cohen, T. Kodenkandeth, J. Mac-Manus-Driscoll, N.
M. Alford, J. Magn. Magn. Mater. 195 (1999) 31.
[10] A.J. Millis, P.B. Littlewood, B.I. Shraiman, Phys. Rev. Lett. 74 (1995) 5144.
[11] G. Li, H.-D. Zhou, S.J. Feng, X.-J. Fan, X.-G. Li, Z.D. Wang, J. Appl. Phys. 92 (2002)
1406.
[12] I. uti, J. Fabian, S. Das Sarma, Rev. Mod. Phys. 76 (2004) 323.
[13] W. Essam, Rep. Prog. Phys. 43 (1980) 833.
[14] E.A. Moujaes, A. Khater, M.A. Ghantous, J. Magn. Magn. Mater. 391 (2015) 49.
[15] A.M. Ahmed, G. Papavassiliou, H.F. Mohamed, E.M.M. Ibrahim, J. Magn. Magn.
Mater. 392 (2015) 27.
[16] M. Baazaoui, S. Hcini, M. Boudard, S. Zemni, M. Oumezzine, J. Magn. Magn.
Mater. 401 (2016) 323.
[17] R. Moubah, S. Colis, C. Ulhaq-Bouillet, M. Drillon, A. Dinia, J. Alaria, J. Phys.
Chem. C 114 (2010) 1684.
[18] R. Moubah, S. Colis, C. Ulhaq-Bouillet, M. Drillon, A. Dinia, J. Phys.: Cond.
Matter 23 (2011) 276002.
[19] R. Moubah, S. Colis, M. Gallart, G. Schmerber, P. Gilliot, A. Dinia, J. Lumin. 132
(2012) 457.
[20] J.H. Song, T. Susaki, H.Y. Hwang, Adv. Mater. 20 (2008) 2528.
[21] F. Keffer, Handbuch der Physik, vol. XVIII/2, in: F. Flgge, H.P.J. Wijin (Eds.),
Springer-Verlag, Berlin, 1966, pp. 50.
[22] N.W. Ashcroft, N.D. Mermin, Solid State Physics, Brooks- Cole, Belmont, MA,
1976.
[23] D.M. Edwards, Proc. R. Soc. Lond. Ser. A, Math. Phys. Sci. 269 (1962) 338.
[24] A. Katsuki, E.P. Wolhfarth, Proc. R. Soc. A 295 (1966) 182.
[25] S.N. Kaul, Phys. Rev. B 27 (1983) 5761.
[26] W. Dring, Handbuch der Physik vol XVIII/2, in: S. Flgge (Ed.), Springer,
Berlin, 1966, 341.
[27] C. Kittel, Introduction to Solid State Physics, John Wiley & Sons, New York,
1986.
[28] X.M. Liu, G.D. Tang, X. Zhao, L.H. Liu, D.L. Hou, C.F. Pan, X.F. Nie, C.Q. Jin, J.
Magn. Magn. Mater. 277 (2004) 118.
[29] M. Jaime, P. Lin, M.B. Salamon, P.D. Han, Phys. Rev. B 58 (1998) R5901.
[30] A. de Andres, M. Garcia-Hernandez, J.L. Martinez, C. Prieto, Appl. Phys. Lett. 74
(1999) 3884.
[31] M. Khli, M. Bejar, E. Dhahri, P. Lachkar, E.K. Hlil, J. Appl. Phys. 111 (2012)
103909.