Intrinsic magnetic properties of In2O3 and transition metal-doped-In2O3

Article abstract of DOI:10.1016/j.jmmm.2007.10.002

This paper reports on the influence of the sintering temperature and atmosphere and transition-metal doping on the magnetic properties of nanocrystalline and bulk In₂O₃. Undoped nanocrystalline In₂O₃ is diamagnetic whatever the sintering temperature and atmosphere. All single-phase transition-metal-doped In₂O₃ samples are paramagnetic, with a paramagnetic effective moment originating from weakly interacting transition metal ions. No trace of ferromagnetism has been detected even with samples sintered under argon, except extrinsic ferromagnetism for samples with magnetic dopant concentrations exceeding the solubility limit.

Full text of DOI:10.1016/j.jmmm.2007.10.002

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
www.elsevier.com/locate/jmmm
Intrinsic magnetic properties of In O and transition metal-doped-In O
2 3
2
3
Ã
David Berardan , Emmanuel Guilmeau, Denis Pelloquin
´
CRISMAT Laboratory, UMR 6508 CNRS-ENSICAEN, 6 Boulevard du Mar e´ chal Juin, 14050 CAEN Cedex, France
Received 6 August 2007; received in revised form 18 September 2007
Available online 10 October 2007
Abstract
This paper reports on the influence of the sintering temperature and atmosphere and transition-metal doping on the magnetic
properties of nanocrystalline and bulk In . Undoped nanocrystalline In is diamagnetic whatever the sintering temperature and
atmosphere. All single-phase transition-metal-doped In O samples are paramagnetic, with a paramagnetic effective moment originating
2
O
3
2 3
O
2
3
from weakly interacting transition metal ions. No trace of ferromagnetism has been detected even with samples sintered under argon,
except extrinsic ferromagnetism for samples with magnetic dopant concentrations exceeding the solubility limit.
r 2007 Elsevier B.V. All rights reserved.
PACS: 75.20.ꢀg; 75.50.Pp
Keywords: Indium oxide; Magnetic material; DMS
1
. Introduction
percolation threshold x . Typically, x is below 10%,
p
whereas xp exceeds 20–25%. Therefore, the origin of
ferromagnetism is still controversial. Moreover, the results
strongly depend on the samples’ preparation conditions
and their dimensionality (nanocrystals, thin films or bulk),
and the reproducibility of the magnetic response from run
to run may be poor.
Diluted magnetic semiconductors (DMS) are semicon-
ducting compounds where some atoms are randomly
substituted by magnetic atoms. They have been extensively
studied due to their potential spintronic applications [1]
since theoretical prediction by Dietl et al. [2] highlighted
that ferromagnetism could be obtained above room
temperature in manganese-doped semiconductors. More
specifically, significant attention has been focused on the
electronic and magnetic properties of DMS oxides since the
discovery of room-temperature ferromagnetism coupled
with good electrical conductivity in cobalt-doped anatase
Among the wide variety of potential DMS that have
been reported, several belong to the transparent conduct-
ing oxides (TCOs) family, such as transition-metal (TM)-
doped ZnO [5], TiO [3], SnO [6] and In O . For the
2
2
2
3
development of a transparent conductive ferromagnetic
material, In O is one of the most promising candidates
2
3
2
+
Co :TiO2 thin films [3]. In the past few years, many
candidates have been discovered that could fulfil the
requisite of high Curie temperature and independently
controllable carrier density and magnetic doping (see for
example Ref. [4]). However, established theoretical models
of magnetism in oxides failed to explain this magnetic
ordering, as the fraction x of 3d dopant that carries a
magnetic moment in these systems usually falls below the
as host material as it is a widely industrially used
TCO. Indeed, In O is a wide bandgap semiconductor
2
3
(3.5–4.3 eV) [7] with a cubic bixbyite crystal structure [8].
Its electrical conductivity can be easily tuned by introdu-
cing oxygen vacancies and/or Sn doping. In the recent
years, many authors have reported that room-temperature
ferromagnetism can be achieved in TM-doped indium
oxide or indium tin oxide, TMs including Cr [9,10],
Mn [11], Fe [12,13], Co [14] and Ni [15]. The proposed
origins of the ferromagnetic response include, notably,
isolated clusters with charge transfer superexchange [16],
mobile-electrons-mediated coupling [17] or nanoscale
Ã
Corresponding author. Tel.: +33 1 69 15 70 20; fax: +33 1 69 15 48 19.
´
E-mail addresses: david.berardan@u-psud.fr (D. Berardan),
emmanuel.guilmeau@ensicaen.fr (E. Guilmeau).
0304-8853/$ - see front matter r 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2007.10.002

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
984
ferromagnetic clusters [13]. However, the magnetic proper-
ties strongly depend on the substrate nature and/or the
synthesis process. It seems that the ferromagnetism is
somehow linked to oxygen vacancies since a surprising
magnetic ordering has been observed in nanocrystals [18]
or thin films [19] of undoped In O , similar to the so-called
X-ray powder diffraction (XRD) was employed for
structural characterization using a Philips X’Pert Pro
diffractometer using Cu-Ka radiation in the 2y range
10–901. The XRD patterns were analysed using the
Rietveld method with the help of the FullProf software
[23]. Grain sizes were estimated using the Scherrer formula.
Thermogravimetric analysis (TGA) was performed on a
Setaram TGA 92, and differential thermal analysis (DTA)
was performed on a Setaram Labsys.
2
3
0
d
ferromagnetism observed in HfO2 [20]. Even more
surprisingly, a degradation of the magnetic ordering in
In O thin films with Mn or Cu doping has been recently
2
3
reported [21]. Nevertheless, the origin of the ferromagnet-
ism in these materials and the physics underlying it remain
controversial. Moreover, despite the dramatic influence
of the synthesis conditions on the magnetic properties,
most of these reports deal with thin films or nanocrystalline
samples and very few studies have been devoted to
bulk materials that could be less sensitive to the synthesis
process. We have recently shown that bulk iron-
doped indium oxide samples belonging to the series
In2 Fe O are paramagnetic, even when sintered under
The morphology and size of crystallites have been also
scanned by means of transmission electron microscopy
(TEM) techniques. This work has been performed with a
JEOL 2010Cx microscope operating at 200 kV (point
˚
resolution 2.7 A, CS ¼ 1.0 mm) and equipped with an
INCA analyser (EDS measurements).
Thermal variations of the magnetic susceptibility (w(T))
and magnetization versus field curves (M(H)) have been
measured using a Quantum Design MPMS SQUID
magnetometer between 2 and 350 K. The electrical
resistivity measurements have been performed using a
Quantum Design PPMS between 5 and 400 K Highly
resistive samples were characterized at room temperature
using a SEFELEC DM500A megohmmeter. This measure-
ment requires electrodes to be painted on two opposite
faces with In–Ga eutectic paste.
ꢀx
x
3
argon [22], which was not consistent with previous results
reported by several groups who observed ferromagnetism
in thin films annealed under argon or with Cu co-doping
[
netic properties of TM-doped bulk In O with various
12,13]. Therefore, we decided to investigate the mag-
2
3
dopants, to study the influence of the sintering atmos-
phere on the magnetic properties, and to carefully
investigate the magnetic properties of nanocrystalline
In O in order to determine the intrinsic properties of
3. Results and discussion
2
3
these systems.
3
.1. Nanocrystalline undoped In O
2
3
2
. Experiments
The addition of ammonium hydroxide to the indium
nitrate solution leads to the formation of a white
precipitate constituted of hydrated indium hydroxide.
The transformation of this precipitate to indium oxide
requires a thermal treatment. Fig. 1 shows the thermal
analysis of the precipitate. The maximum weight loss is
close to 20%, higher than the theoretical value of 16.3%
All bulk samples belonging to the series In2ꢀx
M O and
x 3
In2ꢀ2x
M Sn O (M=transition metal) were prepared using
x x 3
2
a standard solid reaction route. Starting powders, In O3
(
Neyco, 99.99%), Fe O (Rectapur, 99%), Mn O (Cerac,
2 3 2 3
99.9%), Cr O (Prolabo, 99.7%), NiO (Aldrich, 99.8%),
2 3
CuO (Diopma, 99.99%) and SnO (Cerac, 99.9%) were
1
3
2
corresponding to the reaction In(OH) - In O + H O. It
3 2 3 2
2 2
weighed in stoichiometric amounts and ground together by
ball milling using agate balls and vial. The resulting
powders were pressed uniaxially under 300 MPa, using a
polyvinyl alcohol binder to form parallelepipedic samples.
Then they were sintered at 1300 1C (Cr), 1350 1C (Fe,
Ni+Sn and Mn) or 1400 1C (Cu+Sn) for 24 h under air or
flowing argon on platinum foils to avoid any contamina-
tion from the alumina crucible.
shows that the precipitate is still hydrated despite the 24 h
The nanocrystalline In O samples were prepared via a
2
3
liquid route. Indium nitrate pentahydrate (Alfa Aesar
9.99%) was diluted in distilled water. Ammonium
hydroxide was then slowly added until pH=9.5. A white
9
precipitate of In(OH) was obtained after centrifugation
3
and carefully washed with distilled water and absolute
ethanol. It was dried at 80 1C for 24 h, and then calcined at
3
nanocrystals. Nanocrystalline In O was then heated at
50 1C under flowing argon for 1 h to form In O3
2
2
3
400 1C or 600 1C in alumina crucibles under flowing
oxygen, argon or argon/hydrogen.
Fig. 1. Thermal analysis of the hydrated indium oxide precipitate. The
solid line corresponds to DTA (left) and the dashed line to TGA (right).

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
985
drying at 80 1C. The DTA curve exhibits a sharp
endothermic peak at 240 1C. The temperature of this peak
corresponds to the maximum weight loss on the TGA
the particles obtained by various thermal treatments are
summarized in Table 1.
As it seems that the surprising ferromagnetism observed
in In O nanocrystals is somehow linked to oxygen
curve. Therefore, In(OH) undergoes a complete transfor-
3
2
3
mation into In O at 240 1C. Our result slightly differs from
vacancies [18], we have performed several thermal treat-
ments (from 400 to 600 1C) under various atmospheres
(oxygen, argon or argon/hydrogen) to change the oxygen
vacancies concentration. All samples annealed under argon
or oxygen were constituted of In O nanocrystals, as well
2
3
that of Pramanik et al. [24], who observed a broad peak in
the DTA curve between 120 and 210 1C in addition to the
sharp peak, and a significantly higher weight loss. This
different behaviour may originate from different prepara-
tion processes, as the centrifugation step may lead to drier
samples. Nevertheless, this thermal analysis shows that
annealing at a temperature higher than about 250 1C leads
to In O particles.
2
3
as the sample annealed at 400 1C for 2 h under argon/
hydrogen. However, the samples annealed under argon/
hydrogen at 600 or 400 1C for 10 h also contained nano-
crystals of metallic indium, evidencing a partial reduction
of indium oxide. Despite these different treatments, all
samples were diamagnetic, as it is exemplified in Fig. 3 for a
sample annealed at 600 1C for 1 h under flowing argon.
Even the samples that contained indium metal were
diamagnetic although their oxygen vacancies concentration
was certainly high. Therefore, neither the nanocrystalline
character nor the high oxygen vacancies concentration
seem to be sufficient to lead to magnetic ordering in In O .
2
3
This result was confirmed by XRD analysis. Fig. 2a
shows the XRD pattern of the dried white precipitate. A
few peaks can be seen that all correspond to In(OH) Bragg
3
positions. The intensity is very small and the peaks are
broad, denoting a poor crystallinity of the sample. After a
1
becomes yellow and all diffraction peaks fit to In O Bragg
h annealing at 350 1C under flowing argon, the precipitate
2
3
positions (Fig. 2b). The Rietveld refinement shows the
absence of secondary phases within the XRD detection
limit and more especially the absence of In(OH) . The size
2
3
It is noteworthy that the ferromagnetic indium oxide
sample presented in Ref. [18] was prepared under flowing
oxygen (as one of our samples, see Table 1) and may have a
3
of the crystallites is estimated close to 8 nm using the
Scherrer formula. This nanocrystalline size is corroborated
by the TEM image in Fig. 2c showing two agglomerated
particles significantly smaller than 10 nm.The grain sizes of
Fig. 2. (a) XRD pattern of the indium hydroxide precipitate. (b) XRD
nanocrystals. (c) High-resolution
Fig. 3. Magnetization versus temperature (H ¼ 100 Oe) for In
2 3
O
pattern and Rietveld refinement of In
TEM image of In
2
O
3
nanocrystals sintered at 600 1C for 1 h under flowing argon. Inset:
Magnetization versus field at 10 K.
O
2 3
nanocrystals.
Table 1
Size of the crystallites and magnetic behaviour for various thermal treatments
Atmosphere
Temperature ( 1C)
Time
Composition
Grain size (nm)
Magnetic behaviour
As precipitated
–
–
In(OH)
3
ꢁ (H
O)
2 x
–
Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic
Ar
Ar
350
600
400
400
400
600
1 h
1 h
10 h
2 h
10 h
2 h
In
In
In
In
In
In
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
8
12
8
O
2
Ar/H
Ar/H
Ar/H
32
34
61
+In
+In
The size of the crystallites was deduced from the Rietveld refinement using the Scherrer formula and the magnetic behaviour was deduced from
magnetization versus field curves at 10 K.

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
986
lower oxygen vacancies concentration than samples sin-
tered under flowing argon or argon/hydrogen. Therefore,
the discrepancy between the results of Sundaresan et al.
refinement of the XRD pattern for a sample with the
nominal composition In1.9Mn0.1O sintered under air. All
3
peaks can be indexed in the bixbyite structure type (space
group Ia3, no. 206) and no trace of secondary phases can
be detected within the XRD detection limit. For all studied
TMs, the lattice parameter of TM-doped In O is smaller
[
18] and this work cannot be linked to different oxygen
vacancies concentration, nor to particles sizes differences,
as their reported grain sizes are similar to our values. To
rule out the possibility that the magnetic ordering could be
linked to unreacted indium hydroxide, we also measured
the magnetic behaviour of the dried precipitate, which was
diamagnetic. From these results, we firmly believe that the
intrinsic magnetic behaviour of nanocrystalline In O is
2
3
than that of undoped In O , which is consistent with the
2
3
3
smaller ionic radius of Cr , Mn , Fe , Ni and Cu
+
3+
3+
2+
2+
3
+
as compared to In [25]. For example, the inset of Fig. 4
shows the evolution of the lattice parameter with the
nominal manganese fraction in In Mn O , for samples
2
3
2ꢀx
x
3
diamagnetism, whatever the oxygen vacancies concentra-
tion. However, we fail to explain the origin of the
ferromagnetism reported by Sundaresan et al [18].
sintered under air. The Vegards law is well followed for
x ¼ 0.05 and x ¼ 0.1, and then the lattice parameter
saturates at a constant value for xX0.15. From this
3
+
picture, it seems that the solubility limit for Mn in In O
2
3
3
.2. Bulk transition metals doped In O
2
with sintering under air is slightly lower than x ¼ 0.15
7.5 at%). The solubility limits determined for the
studied dopants are summarized in Table 2. It is
3
(
As it seems that the In O matrix is not responsible for
2
3
2
+
the magnetic ordering observed in TM-doped indium
oxide, we have studied the influence of the TM on the
magnetic properties. It is noteworthy that our samples were
prepared by a conventional solid-state route, which is an
equilibrium sintering process. Careful attention was given
to the handling and storage of the samples to avoid any
contamination by magnetic impurities.
The synthesis process presented above leads to well-
crystallized samples, most of them being single phase as
exemplified in Fig. 4. This figure shows the Rietveld
noteworthy that we were not able to substitute Cu
2
or
in significant amounts. Nevertheless, as it
+
3+
Ni
for In
3
+
has been shown by Bizo et al [26] that In
can be
and Sn simultaneously with the
Sn O , we have performed
x 3
2
+
4+
substituted by M
general formula In2ꢀ2x
M
x
coupled Ni+Sn and Cu+Sn substitutions.
Above these solubility limits, secondary phases can be
detected in the XRD patterns. For example, Fig. 5 shows a
magnified view of the XRD patterns in the series
Fig. 5. Magnified views of XRD patterns for samples with different x in
the series In2ꢀxCr
(from bottom to top: x ¼ 0.05, 0.1, 0.2, 0.3, 0.4, 0.5). The circle symbol
corresponds to Cr Bragg reflexion, the square symbol corresponds to
Mn Bragg reflexion.
x
O
3
(left) and In2ꢀxMn
x 3
O (right) sintered under air
Fig. 4. XRD pattern and Rietveld refinement for In1.9Mn0.1
3
O sintered
under air. The inset shows the lattice parameter versus manganese fraction
2
O
3
in the series In2ꢀxMn
x
O
3
(the dotted lines are guide for the eyes).
3 4
O
Table 2
Solubility limits of the TM dopants in In2ꢀx
M O
x 3
and In2ꢀ2xNi
x
x
Sn O
3
sintered under oxygen, air or argon
3+
Cr
3+
Mn
3+
Fe
2+ 4+
Ni +Sn
2+ 4+
Cu +Sn
Dopant
Solubility limit (air or O
Solubility limit (Ar)
2
)
x ¼ 0.15
x ¼ 0.15
x ¼ 0.15
x ¼ 0.05
x ¼ 0.55
x ¼ 0.30
x ¼ 0.55
x ¼ 0.35

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
987
In2 Cr O (left) and In Mn O (right) sintered under
x
ꢀx
3
2ꢀx
x
3
air with x increasing from 0.05 to 0.5 from bottom to top.
In these two cases, the samples with x40.1 unambiguously
contain little amounts of Cr O (peaks indicated by a circle
2
3
symbol on the XRD patterns) or Mn O (peak indicated by
3
4
a square symbol on the XRD pattern). In the same manner,
samples with x exceeding the solubility limit contain little
3
+
2+
amounts of Fe O (M ¼ Fe ) or NiO (M ¼ Ni ). In the
2
3
+
2
case of M ¼ Cu , the secondary phase was not identified.
As most of these secondary phases are magnetic materials,
we will see that small residual amounts in the samples will
dramatically influence their magnetic properties.
To check that no contamination by magnetic impurities
occurred during the synthesis process, the magnetic
behaviour of undoped In O has been recorded. Both
samples, sintered under air or argon, are diamagnetic, in
the same manner as nanocrystalline samples. All samples
2
3
Fig. 7. Magnetization versus temperature (H ¼ 100 Oe) for a sample with
nominal composition In Mn
1
.85
0.15 3
O sintered under oxygen atmosphere
(square symbols: wFC, triangle symbols : wZFC). Inset : Magnetization
versus field curves at 10 K for two samples with same final composition
belonging to the series In2 M O and In
M Sn O
x x 3
ꢀx
x
3
2ꢀ2x
with x below the solubility limit are paramagnetic down to
3 4
but different Mn O fractions.
5
typical example of the magnetic behaviour observed in
these compounds, with In1.9Mn0.1O sintered under oxygen
3
atmosphere. The inverse magnetic susceptibility (1/w)
linearly increases with temperature in the whole tempera-
ture range, field-cooled and zero-field-cooled curves are
identical, and the paramagnetic behaviour can be easily
described using a classical Curie-Weiss law. Moreover, the
curve does not display any characteristic feature of a
transition to a magnetic order. The paramagnetic effective
K, whatever the sintering atmosphere. Fig. 6 shows a
inset shows a magnetization versus field curve, which is
linear too. The result is very similar for M ¼ Cr, Fe, Ni or
Cu, whatever x below the solubility limit.
The magnetic properties are very different for x
exceeding the solubility limit. For example, in Fig. 7 is
displayed the temperature dependence of zero field-cooled
susceptibility wZFC and field-cooled susceptibility wFC for a
sample exceeding the solubility limit, In1.85Mn0.15O3
sintered under oxygen atmosphere. The sample is para-
magnetic down to about 40 K, and then undergoes a
magnetic transition. The transition temperature corre-
sponds well to the ferrimagnetic Curie point of Mn O at
moment is 1.5 m per formula unit, which is in agreement
B
3
+
with Mn
concentration. The results are similar to the
other studied dopants, with the paramagnetic effective
moment that only originates from the doping ion. It is
3
4
41.9 K [27]. As a small amount of Mn O has been detected
3
4
3
+
noteworthy that the magnetic interactions between Mn
ions are very weak, with the paramagnetic temperature
in the XRD pattern of this compound, the magnetic
ordering in this compound is unambiguously connected
with the presence of this secondary magnetic phase.
Moreover, the inset of Fig. 7 shows the magnetization
versus field curve recorded at 10 K for the same sample
3
+
y close to 0 K. y is also close to 0 K with TM ¼ Cr
,
or Cu . It is negative with Fe , where it stays
p
p
2
+
2+
3+
Ni
small, with a weak antiferromagnetic interaction [22]. The
(closed symbols) and for another sample with the same
final composition but with a slightly higher Mn O fraction
3
4
as deduced from XRD refinement (open symbols). The
observed values for the saturation moment can be reason-
ably explained assuming the presence of a small percentage
of Mn O in the samples, which is consistent with XRD
3
4
patterns. Moreover, a slightly higher Mn O fraction for
3
4
similar In2 Mn O composition corresponds to a slightly
ꢀx
higher saturation moment.
x
3
In the studies that report a ferromagnetic behaviour in
TM-doped In O thin films, the presence of magnetic
2
3
ordering is most frequently coupled with high electrical
conductivity (see for example Ref. [13]). Fig. 8 shows the
influence of the TM fraction on the room-temperature
electrical resistivity. All samples belonging to the series
In2 M O have a semi-conducting behaviour and are
ꢀx
x
3
Fig. 6. Magnetization versus temperature (H ¼ 100 Oe) for In1.9Mn0.1
O
3
highly resistive, with resistivity values exceeding that of
undoped In O by several orders of magnitude. The
sintered under oxygen atmosphere (the solid line is a fit assuming a
Curie–Weiss law). Inset: Magnetization versus field at 10 K (the dashed
line is a guide for the eyes).
2
3
samples belonging to the series In2 M Sn O have an
x
ꢀ2x
x
3

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
988
concentration of magnetic dopants that locally exceeds
the percolation threshold, increasing the interaction
between magnetic ions and leading to ferromagnetic
ordering. Therefore, it would be of greatest interest to
study the time and temperature stability of the magnetic
ordering in ferromagnetic TM-doped In O thin films.
2
3
4. Conclusions
We have shown that nanocrystalline indium oxide is
1
0
diamagnetic, as expected for a d element, whatever the
sintering atmosphere. This result strongly disagrees with
previous reports that showed ferromagnetic ordering in the
same material. We have also shown that bulk TM-doped
(M ¼ Cr, Mn, Fe, Ni, Cu) indium oxide is intrinsically
paramagnetic, with a paramagnetic effective moment
originating from the dopant, whatever the sintering atmo-
sphere and the electron concentration. The magnetic
transitions that appear for high dopant concentrations
are linked to the presence of small amounts of magnetic
secondary phases in the samples. Therefore, it seems that
the magnetic ordering observed in thin films of TM-doped
In O is directly connected to the thin film form and might
Fig. 8. Influence of the transition metals fraction on the room-
temperature electrical resistivity.
opposite behaviour, with a metal-like temperature depen-
dence and low resistivity values, which is in agreement with
the results reported by Bizo et al. [26]. All these samples are
paramagnetic (below the solubility limit of the dopant)
whatever their electrical conductivity. Therefore, it seems
that, in our case, high electrical conductivity, and therefore
heavy electron concentration, does not necessarily lead to
magnetic ordering.
2
3
originate from an inhomogeneous distribution of the
dopant in the matrix due to the non-equilibrium synthesis
process.
From our results, we can conclude that bulk TM-doped
indium oxide is paramagnetic, even with argon sintering or
in metallic Sn co-doped samples. The observed magnetic
transitions for high dopant concentrations clearly originate
from the presence of randomly dispersed inclusions of
magnetic phases. These results strongly disagree with those
of several groups who reported room-temperature ferro-
magnetism in TM-doped indium oxide thin films. There-
fore, it seems that the magnetic ordering is directly
connected with the synthesis process and more especially
with the thin films form. As nanocrystalline undoped In O
Acknowledgements
We acknowledge S. D’Astorg for the resistivity measure-
ments using the megohmmeter. E.G. gratefully acknowl-
edges the French Ministe
Technologie and the Delegation Re
et a la Technologie—region Basse Normandie—for finan-
`
re de la Recherche et de la
´
´
´
gionale a la Recherche
`
`
´
cial support.
References
2
3
is diamagnetic whatever the sintering atmosphere, this
discrepancy does not originate from different oxygen
vacancies concentrations between the samples. It is note-
worthy that most samples with ferromagnetic behaviour
have been prepared by pulsed laser deposition, which is a
non-equilibrium technique, contrary to our samples that
are prepared by a conventional solid reaction route.
Therefore, we believe that the ferromagnetism observed
in thin films, which is known to be very sensitive to the
synthesis conditions and poorly reproducible, may origi-
nate from inhomogeneous distribution of the dopants in
the matrix at a nanometre scale, in agreement with the
results reported by Kim et al. [28], who observed
ferromagnetism in inhomogeneous samples and paramag-
netism in homogeneous samples of Co-doped ZnO. This
inhomogeneous repartition could lead to the presence of
nanoscale ferromagnetic clusters as it has been reported in
Co-doped TiO2 films [29]. It could also lead to a
[
1] S.A. Wolf, D.D. Awschalom, R.A. Buhrman, J.M. Daughton, S. von
Molnar, M.L. Roukes, A.Y. Ctchelkanova, D.M. Treger, Science 294
´
(2001) 1488.
[
[
2] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287
(
2000) 1019.
3] Y. Matsumoto, M. Murakami, T. Shono, T. Hasegawa,
T. Fukumura, M. Kawasaki, P. Ahmet, T. Chikyow, S. Koshihara,
H. Koinuma, Science 291 (2001) 854.
[4] J.M.D. Coey, Curr. Opin. Solid. State Mater. 10 (2006) 83.
[5] M. Venkatesan, C.B. Fitzgerald, J.G. Lunney, J.M.D. Coey, Phys.
Rev. Lett. 93 (2004) 177206.
[
[
[
6] S.B. Ogale, et al., Phys. Rev. Lett. 91 (2003) 077205.
7] R.L. Weiher, R.P. Ley, J. Appl. Phys. 37 (1996) 299.
8] M. Marezio, Acta Crystallogr 20 (1966) 723.
[9] J. Philip, A. Punnoose, B.I. Kim, K.M. Reddy, S. Layne,
J.O. Holmes, B. Satpati, P.R. Leclair, T.S. Santos, J.S. Moodera,
Nat. Mate. 5 (2006) 298.
[
[
10] H.S. Kim, S.H. Ji, H. Kim, S.K. Hong, D. Kim, Y.E. Ihm,
W.K. Choo, Solid State Commun 137 (2006) 41.
11] J. Philip, N. Theodoropoulou, G. Berera, J.S. Moodera, B. Satpati,
Appl. Phys. Lett. 85 (2004) 777.

ARTICLE IN PRESS
D. B e´ rardan et al. / Journal of Magnetism and Magnetic Materials 320 (2008) 983–989
989
[
[
[
[
12] J. He, S. Xu, Y.K. Yoo, Q. Xue, H.C. Lee, S. Cheng, X.D. Xiang,
G.F. Dionne, I. Takeuchi, Appl. Phys. Lett. 86 (2005) 052503.
13] T. Ohno, T. Kawahara, H. Tanaka, T. Kawai, M. Oku, K. Okada,
S. Kohiki, Jpn J. Appl. Phys. 45 (2006) L957.
[20] M. Venkatesan, C.B. Fitzgerald, J.M.D. Coey, Nature 430 (2004)
630.
´
[21] V. Brize, J. Sakai, N.H. Hong, Physica B 392 (2007) 379.
[22] D. Berardan, E. Guilmeau, J. Phys.: Condens. Matter 19 (2007)
´
14] N.H. Hong, J. Sakai, N.T. Huong, A. Ruyter, V. Brize
Condens. Matter 18 (2006) 6897.
15] G. Peleckis, X. Wang, S.X. Dou, Appl. Phys. Lett. 89 (2006)
22501.
´
, J. Phys.:
236224.
[23] J. Rodriguez-Carjaval, Physica B 192 (1993) 55.
[24] N.C. Pramanik, S. Das, P.K. Biswas, Mate. Lett. 56 (2002) 571.
[25] R.D. Shannon, Acta Crystallogr. A 32 (1976) 751.
[26] L. Bizo, J. Choisnet, R. Retoux, B. Raveau, Solid State Commun 136
(2005) 163.
0
[
[
16] Y.K. Yoo, et al., Appl. Phys. Lett. 86 (2005) 042506.
17] Z.G. Yu, J. He, S. Xu, Q. Xue, O.M.J. van’t Erve, B.T. Jonker,
M.A. Marcus, Y.K. Yoo, S. Cheng, X.D. Xiang, Phys. Rev. B 74
[27] K. Dwight, N. Menyuk, Phys. Rev. 119 (1960) 1470.
[28] J.H. Kim, H. Kim, D. Kim, Y.E. Ihm, W.K. Choo, J. Appl. Phys. 92
(2002) 6066.
(2006) 165321.
[
[
18] A. Sundaresan, B. Bhargavi, N. Rangarajan, U. Siddesh, C.N.R.
Rao, Phys. Rev. B 74 (2006) 161306.
´
19] N.H. Hong, J. Sakai, N. Poirot, N. Brize, Phys. Rev. B 73 (2006)
[29] S.R. Shinde, S.B. Ogale, J.S. Higgins, H. Zheng, A.J. Millis,
V.N. Kulkarni, R. Ramesh, R.L. Greene, T. Venkatesan, Phys.
Rev. Lett. 92 (2004) 166601.
132404.