See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224489905

Microstructure and magnetic properties of strained La0.7Sr0.3MnO3 thin films

Article  in  Journal of Applied Physics · November 2000

DOI: 10.1063/1.1309040 · Source: IEEE Xplore

CITATIONS READS
137 94

9 authors, including:

Anne-Marie Haghiri-Gosnet Jerome Wolfman

French National Centre for Scientific Research University of Tours
161 PUBLICATIONS   3,389 CITATIONS    74 PUBLICATIONS   922 CITATIONS   

SEE PROFILE SEE PROFILE

B. Mercey
National Graduate School of Engineering and Research Center (Caen)
186 PUBLICATIONS   3,941 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ANR NanoPiC View project

DIMIPOLE View project

All content following this page was uploaded by Anne-Marie Haghiri-Gosnet on 09 November 2020.

The user has requested enhancement of the downloaded file.

Microstructure and magnetic properties of strained La0.7Sr0.3MnO3 thin
films
A. M. Haghiri-Gosnet, J. Wolfman, B. Mercey, Ch. Simon, P. Lecoeur et al.

Citation: J. Appl. Phys. 88, 4257 (2000); doi: 10.1063/1.1309040

View online: http://dx.doi.org/10.1063/1.1309040
View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v88/i7
Published by the American Institute of Physics.

Related Articles
Spin precession modulation in a magnetic bilayer
Appl. Phys. Lett. 101, 262406 (2012)
Alternating domains with uniaxial and biaxial magnetic anisotropy in epitaxial Fe films on BaTiO3
Appl. Phys. Lett. 101, 262405 (2012)
Effect of a change in thickness on the structural and perpendicular magnetic properties of L10 ordered FePd
ultra-thin films with (001) texture
J. Appl. Phys. 112, 113919 (2012)
Spacer-less, decoupled granular L10 FePt magnetic media using Ar–He sputtering gas
J. Appl. Phys. 112, 113916 (2012)
Development of FeNiMoB thin film materials for microfabricated magnetoelastic sensors
J. Appl. Phys. 112, 113912 (2012)

Additional information on J. Appl. Phys.

Journal Homepage: http://jap.aip.org/
Journal Information: http://jap.aip.org/about/about_the_journal
Top downloads: http://jap.aip.org/features/most_downloaded
Information for Authors: http://jap.aip.org/authors

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
JOURNAL OF APPLIED PHYSICS VOLUME 88, NUMBER 7 1 OCTOBER 2000

Microstructure and magnetic properties of strained La0.7Sr0.3MnO3

thin films
A. M. Haghiri-Gosnet,a) J. Wolfman,b) B. Mercey, Ch. Simon, P. Lecoeur, M. Korzenski,
and M. Hervieu
Laboratoire de Cristallographie et de Sciences des Matériaux, CRISMAT-ISMRA, UMR6508,
6 bd du Maréchal Juin, 14050 Caen Cedex, France
R. Desfeux
Laboratoire de Physico-Chimie des Interfaces et Applications, Université d’Artois, rue Jean Souvraz, SP18,
62307 Lens Cedex, France
G. Baldinozzi
Laboratoire Structures, Propriétés et Modélisation des Solides, SPMS-UMR8580, Ecole Centrale de Paris,
92295 Chatenay Malabry, France
共Received 19 April 2000; accepted for publication 13 July 2000兲
The lattice deformation of dense strained La0.7Sr0.3MnO3 共LSMO兲 films is shown to control the easy
direction of the magnetization. Optimized pulsed laser deposited conditions allow the fabrication of
dense LSMO thin films which present an exceptional flatness with a peak–valley roughness (R p – v )
of 1 Å, associated to epitaxial grains as large as 1 ␮m. Electron microscopy coupled with x-ray
diffraction have been used to study the unit cell distortion of both tensile and compressive dense
LSMO films as a function of the thickness. No relaxation of the lattice distortion imposed by
substrate has been observed in the thickness range 10–60 nm. The Curie temperature is not
significantly affected by the nature of the substrate: a T C of 350 K is observed for both SrTiO3
共STO兲 and LaAlO3 共LAO兲 substrates, i.e., close to the bulk material 共369 K兲. In contrast, the easy
direction of magnetization depends on the substrate. For tensile films deposited on the STO
substrate, the unit cell is elongated along the film’s plane (a in-plane⫽3.905 Å) with a reduced
perpendicular parameter (c perp⫽3.85 Å): an easy direction of magnetization M in the plane of the
film is observed. For compressive films deposited on LAO substrate, the situation is reversed with
a unit cell elongated along the direction of growth 共c perp⫽4.00 Å and a in-plane⫽3.79 Å兲 and an easy
axis for M along this perpendicular out-plane direction. It is thus demonstrated that the larger cell
parameter, a in-plane for films deposited on STO and c perp for films deposited on LAO, is fully
correlated to the direction of the easy magnetization. © 2000 American Institute of Physics.
关S0021-8979共00兲08020-8兴

I. INTRODUCTION From zero-field magnetic force microscopy images, Kwon

et al. got a signature of the easy directions of magnetization
The doped perovskite manganites Ln1⫺x AEx MnO3, for tensile films deposited on SrTiO3 共STO兲 and compressive
where the lanthanide trivalent ion 共Ln兲 is substituted by a films deposited on LaAlO3 共LAO兲.7 Suzuki et al. proposed
divalent ion (AE⫽alkaline earth), can exhibit very large torque magnetometry measurements on tensile 关100兴 and
magnetoresistance.1,2 This effect, which has been called ‘‘co- 关110兴 LSMO films deposited respectively on STO and
lossal’’ magnetoresistance 共CMR兲,3 has renewed the interest LaGaO3 substrates.8 The influence of thickness on the resis-
of the thin-film community to these CMR materials. With a tive properties were also reported by Lu et al.9 for compres-
high Curie temperature T C of 369 K,4 La0.7Sr0.3MnO3 sive 关100兴 LSMO films on LAO. Here we propose a more
共LSMO兲 appears an attractive material for magnetic field complete study in which the easy direction of magnetization
sensing and magnetic storage applications at room is fully correlated to the film lattice distortion.
temperature.5,6 First, the morphology of the pulsed laser deposited
The properties of CMR films are known to be very sen- 共PLD兲 LSMO film is shown to follow the ‘‘morphology zone
sitive to the nature of the substrate, due to strains that are model’’ proposed by Thornton10,11 for the sputtering deposi-
imposed by the lattice mismatch between the film and the tion technique. Also for the PLD technique, at a micrometer
substrate. Different studies on the magnetoresistive proper- scale, the film morphology can be predicted as a function of
ties of strained LSMO films have been previously reported. the most prominent deposition parameters. A brief guideline
will be proposed concerning the optimal choice of the PLD
a兲
parameters in order to avoid columnar growth and to thus get
Author to whom correspondence should be addressed; electronic mail:
thin films with a dense morphology as well as large grains.
anne-marie.haghiri@ismra.fr
b兲
Present address: IBM Almaden Research Center, 650 Harry Road, San In a second step, at an atomic scale, the microstructure
José, CA 95120. of each large dense grain also has to be studied to determine

0021-8979/2000/88(7)/4257/8/$17.00 4257 © 2000 American Institute of Physics

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
4258 J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al.

the lattice distortion. Both STO 共a p STO⫽3.905 Å cubic兲 and III. MORPHOLOGY AND DEPOSITION PARAMETERS
LAO 共a p LAO⫽3.789 Å pseudocubic兲 substrates were used to
produce tensile and compressive strained states respectively, For the sputtering deposition technique the evolution of
in the LSMO (a p bulk LSMO⫽3.889 Å) films. The microstruc- the film’s morphology has been well studied and described
ture of these two series of strained materials was investigated by Thornton10 as a function of both the inert sputtering gas
using x-ray diffraction 共XRD兲 and high resolution electron pressure and the homologous temperature T s /T m , where T s
microscopy 共HREM兲 in order to precisely analyze the unit is the substrate temperature and T m is the melting point for
cell distortion. the depositing material. At low T s /T m , two different mor-
Finally, for our two series of tensile and compressive phologies can be obtained as a function of the pressure.
films, angular dependant magnetization measurements com- 共a兲 At high pressure, scattering occurs in the plasma so
bined with the film’s thickness dependence will be used to that the thermalized particules impinge on the surface with a
evidence the magnetic anisotropy governed by the lattice dis- random angle of incidence inducing a self-shadowing effect.
tortion. Due to this shadowing effect, the deposited film presents a
fine vertical fibrous morphology with ‘‘wide boundaries’’
that are voids. This is the so-called zone I columnar porous
morphology.
II. EXPERIMENTAL METHODS
共b兲 If the pressure is sufficiently low to avoid scattering
The dense target of LSMO was synthetized using stan- in the plasma, the particules will keep their energy from the
dard ceramic methods. La2O3, MnO2, and SrCO3 powders target to the growth surface. Thus, the film becomes dense
were mixed in appropriate ratios and intimately ground using with a bimodal distribution of grain size, corresponding to
a semiplanetary ball mill. The mixed powder was first an- large scattered grains surrounded by small ones 共zone T mor-
nealed at 900 °C for 24 h and then at 1200 °C for 24 h with phology兲. At a sufficient reduced pressure, the columnar
intermediate grindings. A pellet having a final diameter of morphology disappears and the film presents an equiaxed
2.5 cm was isostatically cold pressed at 950 kg/cm2 and sin- dense structure of each grain in the total thickness of the film
tered at 1450 °C for at least 72 h. In order to avoid microc- 共zone II兲.
rack formation during heating and cooling slow temperature One should also note that another relevant parameter for
variation rates were used. scattering in the plasma is the target-to-substrate distance D.
The LSMO thin films were deposited using a standard For a fixed pressure, increasing D towards larger values will
PLD apparatus12 working with a KrF excimer laser at ␭ favor the columnar morphology shifting the zone I/zone T
⫽248 nm. Several important parameters are: laser power transition towards low pressures.13 In the PLD technique, the
density ⬇1–2 J/cm2, pulse rate 2 Hz and temperature of plasma created between target and substrate, referred to as
580 °C. Before deposition the optical quality single-crystal the ‘‘plume,’’ is transported through high reactive gas pres-
substrates of 共100兲 SrTiO3 and LaAlO3 were cleaned in ul- sure due to a quasifree jet expansion governed by hydrody-
trasonic baths of acetone and ethyl alcohol and were then namics. Even if the plume is a dynamic plasma, similarities
attached to the heater using silver paste. The heater can be between standard high pressure PLD and static plasma depo-
translated for proper target-to-substrate positioning. A sition techniques concerning the energy of the species 共a few
K-type thermocouple was mechanically attached to the tens of electron volts兲, the pressure range and the target-to-
heater block close to the substrate and the temperature mea- substrate distance D values are observed.14 Thus, the study
sured at this location is referred to as the deposition tempera- of the applicability of Thornton’s morphology predictions to
ture. During deposition the substrate was held at 640 °C in a PLD appears of great interest. Both target-to-substrate dis-
dynamic vacuum having an O2 pressure ranging from 0.4 to tance and oxygen pressure were successively varied in the
0.55 mbar. After complete deposition, the films were cooled LSMO ablation experiments for this morphology study.
to room temperature at 20 °C/min under a static oxygen pres- Two series of depositions were realized at oxygen pres-
sure of 500 mbar. sures p(O2) of about 0.53 and 0.46 mbar, with a target-to-
X-ray diffraction patterns were recorded on a high reso- substrate D distance ranging from 40 to 65 mm. All of the
lution two axes goniometer equipped with a Cu rotating an- other deposition parameters, i.e., laser fluence, repetition fre-
ode of 18 kW at Ecole Centrale Paris. The angles 共2␪ and ␼兲 quence, and cooling under oxygen, were maintained at our
were accurately measured by means of incremental optical traditional values given in Sec. II. Note that standard high
encoders. The samples were mounted in a cryostat, cooled by pressure 关 p(O2)⬇0.5 mbar兴 ablation systems for oxides or-
gaseous convection with a thermal stability of about 0.1 K. dinarily use similar conditions. Thus, all of our results that
The measuring precision was better than 0.1 K. are presented here could be easily applied to numerous stud-
The magnetization of the LSMO films was measured on ies on ablation of CMR materials. After deposition each film
a MPMS 5S superconducting quantum interference device surface was systematically observed using scanning electron
magnetometer from Quantum Design. We used a rotating microscopy 共SEM兲 and atomic force microscopy 共AFM兲 in
sample holder, the susceptibility of which being quite inde- order to determine the morphology.
pendent of the temperature and the magnetic field amplitude. The morphology study of films deposited at the high
It depends slightly on the angle between the field and the pressure value p(O2)⫽0.53 mbar is presented in Fig. 1. The
sample, so at each angle its magnetization was subtracted film deposited at the highest D value (D⫽65 mm) exhibits a
from the measured signal. typical zone I columnar morphology as shown on the three-

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al. 4259

FIG. 1. Morphologies of films deposited at an oxygen pressure of 0.53 FIG. 2. Morphologies of films deposited at a reduced oxygen pressure of
mbar. For a large D value of 65 mm, a zone I columnar morphology is 0.46 mbar. at D⫽50 mm, the AFM scanning 共a兲 is typical of zone T. at
observed on both 共a兲 3D AFM scanning and 共b兲 the corresponding SEM D⫽40 mm, the AFM scanning 共b兲 and the SEM image 共c兲 show a excep-
image. For a reduced to 50 mm D value, the surface morphology is typical tional flatness typical of zone II morphology.
of zone T as shown on 共c兲 the 3D AFM scanning.

and R p – v roughnesses present comparable values to the co-

dimensional 共3D兲 AFM scanning of Fig. 1共a兲. The diameter lumnar zone I film.
of the small juxtaposed columns is approximately 600 Å. AFM images of films deposited at the smaller pressure
This can be also observed on the corresponding SEM top p(O2)⫽0.46 mbar are shown in Fig. 2. The films deposited
view 关Fig. 1共b兲兴. This value is in agreement with our grain at D⫽50 mm and D⫽40 mm exhibit a zone T morphology
size measurements made using cross-sectional electron mi- 关Fig. 2共a兲兴 and a zone II dense morphology 关Fig. 2共b兲兴, re-
croscopy 共EM兲 at low magnification: each small column with spectively. In the latter case, the grain size is larger than 1
a diameter of 600 Å is constituted of one single grain. If the ␮m with a very smooth surface. A typical SEM image of this
mean roughness R ms (10.5 Å) presents a reasonable value, surface can be seen in Fig. 2共c兲. Both R ms and R p – v values
the peak-to-valley roughness R p – v is as high as 73.3 Å. This are extremely low: R ms ⫽0.9 Å and R p – v ⫽10 Å. This is the
is the signature of the column top roughness. The AFM im- signature of the disappearance of the columns as well as the
age of the film deposited at a lower D value 关D⫽50 mm, presence of a very smooth, dense, and equiaxed grain struc-
Fig. 1共c兲兴 shows a typical zone T morphology with a bimodal ture.
grain size distribution: the small columns 共600 Å in diam- These observations show that a dense morphology, asso-
eter兲 are surrounded by dense and smooth grains as large as ciated to an exceptional flatness of the film surface, is fa-
1 ␮m. Roughness is reduced in the surface regions of large vored by a reduced oxygen pressure as well as a small target-
grains compared to regions of small columnar grains. If this to-substrate distance D. Small D values correspond to
film appears smoother in the regions of the large grains, R ms substrates placed inside the plume. On the opposite, at D

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
4260 J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al.

TABLE I. Variation of the perpendicular c perp LSMO parameter as a function

of the LSMO film’s thickness. The parameter c perp LSMO is deduced from the
position of the 关002兴 XRD peak that is indexed by reference to the perov-
skite structure.

Thickness c parameter values 共Å兲

from Kiessig
Number franges On STO On LAO
of pulses 共Å兲 (a⫽3.905 Å) (a⫽3.789 Å)

5000 510 3.849 4.004

2500 255 3.852 4.005
500 120 3.856 4.012
1000 100 3.87⫾0.01 4.011⫾0.01
250 60 Unmeasurable Unmeasurable

FIG. 3. One experimental XRD spectrum around the 关002兴 peak of a LSMO
film deposited on LAO. The simulated spectrum for thickness determination
distances far from the end of the plume, zone I columnar is shown for comparison.
porous morphology is obtained. This morphology evolution
is in excellent agreement with previous results of Thornton
A. XRD measurements
for sputtered materials: an increasing gas pressure stabilizes
zone I for a fixed T s /T m . A more systematic and exhaustive From the structural phase diagram of bulk
study should be necessary to predict the influence of all the La1⫺x Srx MnO3, a rhombohedral pseudocubic symetry
other ablation parameters on the morphology, such as the (R3̄c) with hexagonal lattice parameters a h ⫽a p 冑 2
laser fluence, the repetition rate and the reduced temperature ⬇5.52 Å and c h ⫽2 a p 冑 3⬇13.36 Å is expected for the com-
T s /T m . The magnetic properties of the ablated LSMO films position x⫽0.3. For simplicity, and also for comparison be-
should be strongly influenced by the morphology. In a tween the film and the substrate, the indexing will be referred
voided columnar film, the small juxtaposed tubes have a high to the cubic perovskite subcell. On the XRD patterns, besides
vertical shape anisotropy which should favor an out-plane the 关001兴 reflections of the substrate, film peaks are observed
easy direction of the magnetization of the LSMO. This is the allowing us to calculate the thin film interreticular distance
reason why all the thin films studied in the following are along the growth direction. The variation of the c perp param-
deposited at sufficient low target-to-substrate distances D eter perpendicular to the interface, i.e., along the direction of
and reduced oxygen pressure to assure a dense morphology growth, was determined from the 关 002兴 P peak angle posi-
共zone II兲 with very large and smooth grains. tions in the XRD spectra and is reported in the Table I.
Thicknesses deduced from deposition rate are confirmed us-
ing XRD simulations. The diffused intensity formula used
IV. MICROSTRUCTURE INDUCED BY LATTICE for simulations is
MISMATCH I⫽sin2 关 2 ␲ e LSMO sin共 ␪ /␭ 兲兴 /sin2 关 2 ␲ c perp sin共 ␪ /␭ 兲兴 ,
The internal microstructure of each large grain was stud- where ␭ is the wavelength, ␪ is the XRD angle, and e LSMO
ied as a function of the lattice stressed state for dense 共zone and c perp correspond to the thickness and the ‘‘perpendicu-
II兲 films. PLD conditions were chosen in order to assure a lar’’ parameter of the film, respectively. One experimental
dense and smooth zone II growth: T s ⫽640 °C, D⫽40 mm, spectrum, that shows the 关 002兴 P peak at 45.24° and besides
and p(O2)⫽0.46 mbar. It is well established that the stress two finite size oscillations, is reported in Fig. 3 with the
pattern in films depends on the film’s thickness. It is thefore corresponding simulation. The thickness deduced from simu-
of interest to produce and study films with different patterns lation is about 255 Å. A good agreement between experi-
by varying the thickness. LSMO films were deposited with mental and theoretical spectra is observed for both amplitude
different thickness on two substrates, namely STO and LAO. and angle position. The precision on thickness determination
For ease of comparison, each series of two samples were using these simulations is about 5% in this range of thick-
deposited at the same time in the experiment chamber. The ness.
epitaxial LSMO films 共a p LSMO bulk⬇3.889 Å where a p is the In a first global approach, let us consider that the in-
perovskite subcell parameter兲 deposited on LAO (a p LAO plane parameter, a in-plane , is constant and equal to the bulk
⬇3.789 Å) and STO (a p STO⬇3.905 Å) are, respectively, lattice parameter of the substrate, a STO⫽3.905 Å and a LAO
under a compressive stress and a tensile stress. The lattice ⫽3.789 Å. The c perp /a in-plane ratio has been calculated using
mismatch ␦ along the interface is given by ␦ ⫽(a p substrate the XRD values of c perp given in Table I. The evolution of
⫺a p LSMO)/a p substate . This mismatch on LAO is negative, the shape of the LSMO unit cell can be analyzed from the
⫺2.6%, whereas it is positive ⫹0.41% on STO. This sub- c perp /a in-plane . The variation of c perp /a in-plane as a function of
strate mismatch strain induces a lattice deformation. The unit the thickness is reported, for both substrates, in Fig. 4. For a
cell distortion has been investigated here as a function of the better understanding the deformation of the unit cell is also
thickness of the film, using first XRD measurements and, in indicated. Note that c perp /a in-plane⫽1 is expected for a per-
a second step, electron diffraction 共ED兲 and HREM. fect cubic unit cell. Two regions are observed 共see Fig. 4兲:

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al. 4261

FIG. 4. Variation of the c perp /a in-plane ratio 共determined from XRD measure-
ments兲 as a function of the thickness for films deposited on STO and LAO. FIG. 5. Variation of both STO substrate (c STO) and film (c perp) perpendicu-
lar parameters as a function of temperature. These parameters are deduced
from the 关004兴 XRD peaks.
共1兲 on STO, for the tensile state, the c perp /a in-plane ratio is
lower than 1; the pseudocubic 共rhombohedral兲 unit cell of
LSMO is elongated in the direction of the interface along the 6共a兲. On this image only one grain is shown. It confirms that,
a in-plane parameter. in our deposition regime, large dense grains are obtained.
共2兲 on LAO, for compressive films, a reverse situation is The film is epitaxial from the sharp flat interface to the sur-
observed with a c perp /a in-plane ratio higher than 1; the unit cell face. The structure is monocrystalline without any columns,
here is elongated in the growth direction along the c perp pa- dislocations, or secondary phase inclusions. The HREM im-
rameter. age 关Fig. 6共b兲兴 of the interface shows that the LSMO film is
perfectly coherent across the interface which is completely
For both strained states, only a very small decrease of
free of any defects. The variation of the c perp /a in-plane ratio
the c perp /a in-plane ratio is observed as the thickness of the film
determined previously by XRD is too small to be precisely
increases. A decrease of this c perp /a in-plane ratio has also been
measured on selected area electron diffraction 共SAED兲 pat-
previously observed by Lebedev and co-workers.15,16 How-
ever, in our experiments, the variation of the c perp /a in-plane
ratio appears very small and ⌬(c/a) does not exceed 0.04 in
the thickness range 10–60 nm. This means that the lattice
distortion due to the mismatch between substrate and film is
not relaxed even for the thicker 60 nm films. Local determi-
nation of both in-plane parameters of substrate a substrate and
film a in-plane by ED and HREM is necessary to precisely ana-
lyze these cell distortions at an atomic scale.
The SrTiO3 substrate exhibits a structural phase transi-
tion at low temperature. The transition takes place at 103 K.
The high temperature phase is cubic whereas the low tem-
perature phase is tetragonal.17 To study the consequence of
this transition on the structural behavior of the film structure,
XRD measurements have been performed on STO films
from room to low temperatures. The variation of both 共sub-
strate and film兲 perpendicular parameters 共c STO and c perp兲
were deduced from high angle 关004兴 XRD peaks. A linear
dependence of both these parameters is shown in Fig. 5 from
the 430 to the 105 K transition temperature. At the transition
and for lower temperatures, a slope variation is observed
with a roughly constant value of both parameters. The most
important result is the similarity in these two parameters evo-
lution: the LSMO film ‘‘follows’’ exactly the structural
changes of the STO substrate. This result is of importance
for the magnetic properties as it will be demonstrated in
Sec. V.

B. HREM on tensile LSMO films deposited on STO

substrates
FIG. 6. Electron microscopy on a 138 nm thick film deposited on STO: 共a兲
A low magnification of a cross-section bright field image at low magnification and 共b兲 a zoom of the interface between the STO and
of a 138 nm thick LSMO deposited on STO is given in Fig. the film at an atomic scale using HREM.

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
4262 J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al.

atomic rows being shifted by a p /2 crossing the defects 关see

the inset of the Fig. 7共a兲兴. The contrast variations observed in
HREM images and the translations of the atoms columns
suggest the local formation of two adjacent 关共La, Sr兲O兴 lay-
ers, according to a well known mechanism of rocksalt-type
layers. Such defects have often been reported in perovskite
derivative structures, in bulk materials18 as well as in thin
films.16,19,20 They have been observed in LSMO16 and
Pr0.5Ca0.5MnO3 共PCMO兲20 thin films deposited on LAO. Dif-
ferent hypotheses can be put forward about the origin of
these defects. They could be generated by composition inho-
mogeneities during the deposition process17 or stress relax-
ation due to the mismatch between the film and the
substrate,21 and also local rearrangements of the film to ac-
commodate the structural distortion due to the ‘‘cubic/
rhombohedral’’transition of the LAO substrate at ⬇544 °C.
Similar defects have been observed in weak-stressed PCMO
thin films20 deposited on LAO in the same PLD system with
comparable conditions 共zone II without columnar growth兲.
The mismatch between PCMO and LAO is very small, about
␦ ⬇⫺0.4%, compared to the LSMO/LAO mismatch
( ␦ ⬇⫺2.6). Thus, the relaxation of the lattice cell distortion
is not sufficient to explain the presence of these defects. The
authors want to point out the importance of the cubic/
rhombohedral transition which occurs at ⬇544 °C. During
deposition at T⬎544 °C, the substrate is cubic with a facet-
ted surface due to the disappearance of the twins: the origin
of the defects observed at room temperature can be the pres-
ence of facets on the growth surface16 as well as the rhom-
bohedral transformation of the LAO during cooling after
growth.22 Aside from these planar defects, the structure is
epitaxial from the interface to the surface without any sec-
FIG. 7. 共a兲 HREM of a compressive film deposited on LAO. Both interface
between LAO and surface of the 51 nm thick film are shown. One magnified ondary phase inclusions.
defect is magnified in the inset: the vertical dark line shows the a p /2 shift of One should remember that the lattice mismatch is six
the atomic rows perpendicular to the defect. 共b兲 Localized ED pattern show- times larger on LAO than on STO ( ␦ ⬇⫹0.4%). This larger
ing the 关400兴 splitting which allows the a LSMO in-plane determination.
mismatch induces a stronger distortion of the film cell for
films deposited on LAO as compared with these deposited on
STO. Thus, the variation of the c perp /a in-plane ratio, deter-
terns: c perp /a in-plane is decreasing only from 1.5% when the mined previously by XRD, can be confirmed by localized
thickness of the film is increasing from 10 to 60 nm. No spot
ED: the ED pattern of the 51 nm thick LSMO film 关Fig. 7共a兲兴
splitting is observed in the SAED pattern for both directions,
is given in Fig. 7共b兲. A small splitting of the 关400兴 spots 关Fig.
along the interface (a in-plane) and parallel to the growth
7共b兲—bottom right兴 allows for the determination of both film
direction (c perp). Parallel to the interface, both substrate
and substrate parameters along the interface: the a LSMO in-plane
and film have the same lattice parameter a substrate
and a substrate are, respectively, 3.84 and 3.789 Å. Along the
⫽a in-plane LSMO . Moreover, in film as thick as 130 nm, the
关002兴 direction, that is the growth direction, the c LSMO perp is
perpendicular lattice parameter is found to remain constant
4.010 Å value is in total agreement with our XRD measure-
in the whole film thickness.
ments 共Table I兲.
Note that these results provide two important facts. First,
C. HREM on compressive films deposited on LAO calculations of the subcell volumes show that both compres-
substrates sive and tensile films keep a quite similar value, ⬇58.8 Å3,
A cross-sectional image of the thicker LSMO deposited which is the same as that the bulk one. As previously
on LAO 共e⫽51 nm with c perp⫽4.004 Å兲 is given in Fig. reported,21 it shows that the cell volume remains roughly
7共a兲. As mentioned before, large dense grains are obtained in constant though the strain effects of the substrate. Second,
the Zone II deposition regime. However, extended defects since the in-plane film parameters accommodate the sub-
are systematically observed. A few of them are indicated in strate parameters along the equivalents 兵 100其 P directions, the
Fig. 7共a兲 by white arrows. They run a few tens of nanometers new materials deposited in the thin film form do not exhibit
long, perpendicular to the growth direction. Dislocation-like anymore a rhombohedral cell, but a tetragonal one with a
mechanisms are associated with these planar defects, the I-type lattice.

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al. 4263

FIG. 8. In-plane magnetization under 100 G vs temperature for three films

of LSMO deposited on STO for three different thickness. The critical tem-
perature decreases slightly for the thinner film.

V. MAGNETIC PROPERTIES OF STRAINED LSMO FIG. 10. Comparison between the angular dependence of the magnetization
FILMS at 300 K for the films grown on SrTiO3 and on LaAlO3 showing that the
easy magnetization is in plane (⌽⫽0) for SrTiO3 and out of plane (⌽
In the bulk samples of the same composition, the system ⫽90°) for LaAlO3 substrates. The film deposited on STO is 27 nm thick,
is ferromagnetic whith a critical temperature of 369 K. The and the film deposited on LAO is 100 nm thick.
hysteresis cycles present a small hysteresis of about 50 G and
a magnetization of 3.8 ␮ B per Mn atom. No measurement of
the anisotropy is known since no untwinned single crystal of
this phase exists. netic field of 30 G was chosen to be between the in-plane
and out-of-plane coercive fields兲. The easy magnetization is
A. LSMO films deposited on STO substrate clearly parallel to the plane of the film, a direction where the
The films deposited on the STO substrate present a criti- cell parameter a in-plane is larger than the perpendicular param-
cal temperature of 350 K, very similar to that of the bulk eter 共see Fig. 4兲, and is independant of the thickness of the
samples of the same composition 共Fig. 8兲. A small decrease film.
of this temperature is observed for the film of the thinner Below 100 K, a structural phase transition occurs for the
thickness 共6 nm兲. In Fig. 9, we have reported the hysteresis STO substrate 共Fig. 5兲. It can be seen in Fig. 11 that the
cycles at 150 K, with the magnetic field applied in the sub- magnetization with a field of 100 G applied in the plane is no
strate plane, showing that the coercitive field is about 10 G, longer saturated below 200 K. This increase of the coercive
smaller than in the bulk samples. In the thinner film, it is field can be attributed to a rotation of the easy axis of the
even smaller, approaching the non-measurable limits within magnetization out of the substrate plane. This effect becomes
our resolution. The anisotropy of the coercive field was also dramatically important below 100 K due to the substrate
studied by rotating the sample in the magnetic field. In Fig. phase transition of the STO substrate.
10, we have reported a typical curve of the angular depen-
dence of the magnetization under 30 G at 300 K 共the mag-

FIG. 9. Hysteresis cycles at 150 K for the in-plane magnetization for four FIG. 11. The in-plane magnetization of the 12 nm thick film on SrTiO3 at
different thicknesses. 100 G 共field cooled and zero field cooled兲 and 3000 G.

Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
4264 J. Appl. Phys., Vol. 88, No. 7, 1 October 2000 Haghiri-Gosnet et al.

shown to control the direction of the easy magnetization of

the LSMO films, whatever the thickness of the films may be.
The direction of the easy magnetization is always that of the
larger cell parameter, a in-plane in the case of STO films, c perp
in the other case. This effect, very different in nature from
the interface effect which exists in thin metallic giant mag-
neto resistive layers can be used here in thicker layers 共here
up to 100 nm兲. For this reason, it makes this system very
promising for perpendicular recording applications.

1
R. Von Helmott, B. Holzapfel, L. Schutz, and K. Samwer, Phys. Rev.
Lett. 71, 2331 共1991兲.
2
M. Mac Cormack, S. Jin, T. Tiefel, R. M. Fleming, J. Philipps, and R.
Ramesh, Appl. Phys. Lett. 64, 3045 共1994兲.
FIG. 12. Hysteresis cycle at 150 K for the in-plane magnetization of the 58 3
B. Raveau, A. Maignan, C. Martin, and M. Hervieu, in Colossal Magne-
nm thick film deposited on LaAlO3. toresistance Charge Ordering and Related Properties of Manganese Ox-
ides, edited by C.N.R. Rao and B. Raveau 共World Scientific, Singapore,
1998兲, pp. 43–82.
4
B. LSMO films deposited on LaAlO3 substrate A. Urushibara, Y. Morimoto, T. Arima, A. Asamitsu, G. Kido, and Y.
Tokura, Phys. Rev. B 51, 14 103 共1995兲.
5
The films deposited on the LaAlO3 substrate also present T. Venkatesan, M. Rajeswari, Z-W. Dong, S. B. Ogale, and R. Ramesh,
a critical temperature of 350 K, again very similar to that of Philos. Trans. R. Soc. London, Ser. A 356, 1661 共1998兲.
6
J. Z. Sun and A. Gupta, Annu. Rev. Mater. Sci. 28, 45 共1998兲.
the bulk samples of the same composition. In Fig. 12, we 7
C. Kwon et al., J. Magn. Magn. Mater. 172, 229 共1997兲.
have reported one hysteresis cycle at 150 K, showing that the 8
Y. Suzuki, H. Y. Hwang, S-W. Cheong, and R. B. Van Dover, Appl. Phys.
in-plane coercivity is about 500 G larger than in the STO Lett. 71, 140 共1997兲.
9
H. L. Ju, M. Krishnan, and D. Lederman, J. Appl. Phys. 83, 7073 共1998兲.
samples. The anisotropy of the coercive field was also stud- 10
J. A. Thornton, J. Vac. Sci. Technol. 11, 666 共1974兲.
ied by rotating the sample in the magnetic field. In Fig. 10, 11
J. A. Thornton, Proc. SPIE 821, 95 共1987兲.
12
we have reported a typical curve of the angular dependence P. A. Salvador, T. D. Doan, B. Mercey, and B. Raveau, Chem. Mater. 10,
of the magnetization under 500 G at 300 K 共the magnetic 2592 共1998兲.
13
A. M. Haghiri-Gosnet, F. R. Ladan, C. Mayeux, and H. Launois, Appl.
field of 500 G was also chosen to be intermediate between Surf. Sci. 38, 295 共1989兲.
the in-plane and out-of-plane coercive fields兲. The easy mag- 14
G. K. Hubler, in Pulsed Laser Deposition of Thin Films, edited by D. B.
netization is clearly along the direction perpendicular to the Chrisey and G. K. Hubler, 共Wiley-Interscience, New York, 1984兲, Chap.
film, a direction where the parameter c perp is larger than the 13, p. 346.
15
O. I. Lebedev, G. Van Tendelo, S. Amelinckx, H. L. Ju, and K. M. Krish-
other in-plane parameters 共see Fig. 4兲. nan, Philos. Mag. A 80, 673 共2000兲.
16
G. Van Tendeloo, O. I. Lebedev, and S. Amelinckx, J. Magn. Magn.
VI. CONCLUSION Mater. 211, 73 共2000兲.
17
K. A. Muller, W. Berlinger, and F. Waldner, Phys. Rev. Lett. 21, 814
Dense LSMO films presenting an exceptional flatness of 共1968兲.
18
R. J. D. Tilley, J. Solid State Chem. 21, 293 共1977兲.
1 Å and grains as large as 1 ␮m can be synthetized with 19
B. Mercey, A. Gupta, M. Hervieu, and B. Raveau, J. Solid State Chem.
optimized PLD parameters. These should be deduced from 116, 37 共1984兲.
Thornton’s model which can be applied to the PLD tech- 20
A. M. Haghiri-Gosnet, M. Hervieu, Ch. Simon, B. Mercey, and B.
nique. Even for the thicker films 共60 nm兲, the lattice distor- Raveau, J. Appl. Phys. 共to be published兲.
21
B. Mercey, J. Wolfman, W. Prellier, M. Hervieu, Ch. Simon, and B.
tion which is imposed by both tensile and compressive Raveau, Chem. Mater. 共to be published兲.
stresses from STO and LAO substrate mismatches, respec- 22
M. Hervieu, A. M. Haghiri-Gosnet, W. Prellier, B. Mercey, Ch. Simon,
tively, is not relaxed. The strained lattice deformation is and B. Raveau 共unpublished兲.

View publication stats Downloaded 30 Dec 2012 to 128.148.252.35. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions