Room temperature ferromagnetism in Fe doped CuO nanorods

Article abstract of DOI:10.1016/j.jmmm.2010.04.020

One dimensional CuO and Fe doped CuO nanorods have been synthesized by template free solution phase hydrothermal methods. The typical diameter and the length of the Cu_1-xFe_xO nanorods ( x = 0, 0.02, 0.05, 0.10) are 20-25 and 300-400

Full text of DOI:10.1016/j.jmmm.2010.04.020

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 322 (2010) 2749–2753
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials
journal homepage: www.elsevier.com/locate/jmmm
Room temperature ferromagnetism in Fe doped CuO nanorods
Ã
S. Manna, S.K. De
Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata 700 032, India
a r t i c l e i n f o
a b s t r a c t
Article history:
One dimensional CuO and Fe doped CuO nanorods have been synthesized by template free solution
phase hydrothermal methods. The typical diameter and the length of the Cu1ꢀxFexO nanorods (x¼0,
0.02, 0.05, 0.10) are 20–25 and 300–400 nm. Pure CuO nanorods show weak ferromagnetism and the
introduction of Fe within CuO lattice improves significantly the ferromagnetic property with the Curie
temperature far above room temperature. The shape anisotropy is the key point to understand
ferromagnetism in Fe doped CuO nanorods.
Received 8 October 2009
Received in revised form
19 January 2010
Available online 13 April 2010
Keywords:
Magnetic semiconductor
Hydrothermal synthesis
Nanorod
& 2010 Elsevier B.V. All rights reserved.
Magnetic properties
Jahn-Teller distortion
Mixed valency
1. Introduction
coupling in Li doped CuO and CuO nanoparticles. Punnose et al.
[12,13] explained the particle size dependence exchange bias and
large coercivity in CuO nanoparticles on the basis of uncompen-
sated surface spin of the nanocrystal. Li et al. [14] observed
ferromagnetism in Mn doped bulk CuO and ferromagnetism
enhanced by the co-doping of Al ion with low Curie temperature
(below 100 K). Anomalous ferromagnetic behavior have observed
in CuO nanorods [15]. Nanorods also have a number of advantages
over thin films or nanoparticles with respect to studying
ferromagnetism in DMSs. Specifically, they offer thermodynami-
cally stable features and are typically single-crystalline and defect
free. They can thus safely exclude the effect of defects and
nonuniform distribution of dopants that are typically observed in
DMSs prepared by nonequilibrium processing (e.g., molecular
beam epitaxial process). All spin rotations in the nanorods could
be limited to a single axis. Possible tetragonal distortion and
shape anisotropy in the nanorods could also be helpful for the
evolution of ferromagnetism [16,17]. Nanorods themselves are
attractive building blocks for electronic devices. So nanorods
could further fuel the development of ferromagnetism in DMSs.
In this article, we report the room temperature Ferromagnetic
properties of Fe doped CuO nanorods. It was observed that the
ferromagnetic coupling interaction increases with Fe concentra-
tion. The effect of Fe doping is discussed and relevant mechanism
is proposed.
The potential of exploiting the spin of the electron (in addition
to its charge) in novel new electronic devices and the associated
opportunities for new science have led to extensive search for
viable magnetic semiconductors with room temperature ferro-
magnetism [1,2]. Some success for ferromagnetism has been
reported in dilute magnetic semiconductors (DMS) and dilute
magnetic oxides (DMO) containing a few percent of transition
metal ions such as (Ga,Mn)As [3,4], (Zn,M)O [5,6], and (Ti,M)O2
[7,8] with M¼V, Cr, Mn, Co, Ni, and Cu. Most of the earlier works
were concentrated on transition metal doped non-magnetic
semiconductors. Nevertheless much more detailed investigations
are required to understand the origin of ferromagnetism and
possible applications in spintronic devices.
Transition metal monoxides MO, M¼Mn, Co. Ni and Cu reveal
complicated magnetic as well as electronic structure. Ferromag-
netic property was observed in nanosized MO and 3d metals
doped MO by modifying the original magnetic order. Among
transition metal monoxides,Cupric oxide CuO is
a strongly
correlated electron system which exhibits Mott insulating
behavior and anti ferromagnetic (AFM) with two magnetic transi-
tions near 215 and 230 K [9], possesses complicated monoclinic
tenorite structure. The magnetic structure of CuO consists of Cu–O
parallel sheet of plane in which oxygen atom is located at the
center of a copper distorted tetrahedron in (1 1 0) plane. The
exchange in the Cu–O–Cu chains along the ð1 0 1Þ direction is
strongly and completely antiferromagnetic. Recently Zheng and
his co workers [10,11], observed dramatic suppression of AFM
2. Experimental
Pure CuO and Fe doped CuO nanorods were synthesized by
hydrothermal method. For the hydrothermal synthesis of the CuO
nanorods, a closed cylindrical teflon linked autoclave with 23 ml
Ã
Corresponding author. Tel.: +91 33 2473 3073; fax: +91 33 2473 2805.
E-mail addresses: msskd@iacs.res.in, msskd@mahendra.iacs.res.in (S.K. De).
0304-8853/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2010.04.020

ARTICLE IN PRESS
2750
S. Manna, S.K. De / Journal of Magnetism and Magnetic Materials 322 (2010) 2749–2753
capacity was used. In a typical preparation, 0.01 M cupric acetate
CuðCH3COOÞ2 ꢁ 6H2O, 1.25 mM NaOH and the required amount of
ferric nitrate ½FeðNO3Þ3 ꢁ 9H2Oꢂ were dissolved in deionised water
to form an alkali solution. The mixture was stirred for 45–50 min
and then it was transferred to an autoclave. The hydrothermal
synthesis was conducted in an electric oven at 140 3C for 12 h.
After the reaction was over, black crystalline products were
harvested by centrifugation and through washing with deionised
water and ethanol for several times. Finally the samples were
dried at 60 3C for 24 h.
can be replaced by Fe ions. Here we have adopted modified Rietveld
refinement procedure to obtain the refined values of structural and
microstructural parameters. The whole pattern fitting has been done
with the help of the software MAUD [18]. Initial simulation of the
diffraction pattern was carried out with the CuO phase having space
group C2/c. Then the Cu atom is accordingly substituted with the Fe
atom for the doped samples. Fig. 1(b) shows the variation of unit cell
volume as a function of dopant concentration. Clearly, the unit cell
volume contraction by very small amount confirm the incorporation
of Fe ion in CuO host lattice.
X-ray diffraction patterns of the samples were recorded by
high resolution X’Pert PRO PANalytical X-ray diffractometer in the
The atomic compositions of constituent elements in Cu1ꢀxFexO
were determined from ICP-AES spectra with instrument resolu-
tion of 1 ppm. The concentration of Fe in pure CuO nanorods is
about 100 ppm. The atomic ratio of Fe to Cu was quantitatively
determined as 0.016 for CuO:2% Fe 0.041 for CuO:5%Fe and 0.098
for CuO:10%Fe nanorods.
Morphology of as synthesized samples was obtained using
high-resolution transmission electron microscope studies. TEM
and HRTEM images of the 10% Fe doped CuO nanorods are
presented in Fig. 2. Inset shows a single nanorod grown along
(1 1 0) direction. The average diameters of the nanorods vary
between 20 and 25 nm. Length of nanorods are in the range
300–800 nm. The aspect ratio of the nanorods lies between 15 and
40. The high resolution TEM image is shown in Fig. 2(b) which
exhibits clear lattice fringes. Homogeneous and parallel lattice
fringes imply the single crystalline behavior of prepared samples.
The interval between two adjacent lattice fringes is distinctly
range 302903 using Cu K radiation. Microstructure and crystal
a
structures of the nanorods were obtained using transmission
electron microscope (TEM) and high-resolution TEM (HRTEM,
JEOL 2010) studies. The metal contents in the nanorods were
determined by an inductively coupled plasma atomic emission
spectrometer (ICP-AES), (Perkin Elmer). Doped samples were
characterized by X-ray photoelectron spectroscopy (5400 ESCA).
DC Conductivity was measured by high resistance electrometer
(Keithley 6517A) in two probe configuration. Magnetization
measurements were carried out by SQUID magnetometer,
quantum design, MPMS XL (evercool).
3. Results and discussion
˚
shown by arrows in Fig. 2b is 2.749 A which corresponds to the
Fig. 1(a) shows the XRD patterns of the Cu1ꢀxFexO samples. It can
(1 1 0) plane of the monoclinic structure of CuO. The oriented
(1 1 0) nanorods are formed from dehydration of Cu(OH)2 in
solution medium by oxolation process [19].
be seen that all the samples contain a single CuO monoclinic phase.
3+
Because the radius difference between Cu2+ ion 0.73 A and Fe ion
˚
2+
2+
˚
˚
0.64 A or Fe 0.74 A is not remarkable, Cu ions in lattice structure
82.70
82.65
82.60
82.55
82.50
82.45
82.40
Cu0.90Fe0.10
O
Cu0.95Fe0.05
O
Cu0.98Fe0.02
O
CuO
30
40
50
60
70
80
90
0.00
0.02
0.04
0.06
0.08
0.10
2θ
(degree)
Fe concentration (X)
Fig. 1. (a) XRD patterns of the Cu1ꢀxFexO [x¼0, 0.02, 0.05, 0.10] at room temperature. The calculated and observed diffraction profiles are shown at the top with the color
solid line and open circle, respectively. The lower trace is a plot of the difference between calculated and observed intensities. (b) Fe Concentration dependence of unit-cell
volume. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

ARTICLE IN PRESS
S. Manna, S.K. De / Journal of Magnetism and Magnetic Materials 322 (2010) 2749–2753
2751
CuO
16.0
15.5
15.0
14.5
14.0
13.5
13.0
12.5
12.0
11.5
Cu0.98Fe0.02
Cu0.95Fe0.05
Cu0.90Fe0.10
O
O
O
(110)
(110)
2.749 Å
Fig. 2. (a) Typical TEM image of Cu0.90Fe0.10O nanorods. Inset shows a single
nanorod. (b) HRTEM image of a nanorods.
Fe2P3/2
Fe2P1/2
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
Fe2p3/2
Fe+3
1000/T (1/K)
Fe+2
Fig. 4. Variation of logarithm resistivity as a function of inverse temperature for
pure and iron-substituted CuO samples.
The dc resistivity ðrdcÞ for all compositions have been mea-
sured. The values of rdc at room temperature are 4:69 ꢃ 106,
2:38 ꢃ 106, 1:43 ꢃ 106 and 7:95 ꢃ 105 Ocm for x¼0, 0.02, 0.05 and
0.10, respectively. Room temperature resistivity decreases by
about one order of magnitude with Fe doping. This clearly
indicates that the incorporated Fe ions are donating extra elec-
trons to the host CuO lattice. Resistivity measurements suggest
that some doped Fe ions exist in Fe+3 state for the enhancement
of carrier concentration. Fig. 4 represents the resistivity variation
with temperature for four different samples. Monotonous
increase of resistivity with decrease of temperature indicates
semiconducting behavior. Temperature dependence of resistivity
is generally described by hopping transport mechanism [23],
Raw
fitted
Background
716 712 708 704 700
732 728 724 720 716 712 708 704 700
Binding Energy (eV)
Fig. 3. Fe 2p core level X-ray photoelectron spectrum of 10% Fe doped CuO
nanorods. Inset shows fitted X-ray photoelectron spectrum of Fe2P1/2 peak.
Fig shows that the Fe2P1/2 peak around 710.23 eV contains the contribution from
the Fe+2 and Fe+3 ion.
ꢀꢀ ꢁ ꢁ
g
T0
rðTÞ ¼
r0exp
ð1Þ
T
The chemical valence state of Cu and Fe in doped samples were
characterized by X-ray photoelectron spectroscopy (XPS) (5400
ESCA). The characteristic peaks of Cu 2p1/2 and 2p3/2 appear at
954.3 and 934.4 eV which are consistent with Cu+2 valence
state [20]. The Fe 2p core level spectrum for the x¼0.10 sample is
depicted in Fig. 3. This shows that Fe 2p spectrum is split into
Fe 2p1/2 at 721.44 eV and Fe 2p3/2 at 710.23 eV due to spin orbit
interaction. The absence of peak at 706–707 eV rules out the
presence of metallic Fe in doped samples. The broad spectrum of
Fe 2p3/2 with binding energy (B.E.) 710.23 eV has been analysed
by fitting the experimental data using Gaussian–Lorentzian
functions as shown in inset. The main peak is deconvoluted into
two peaks centered at 709.01 and 710.84 eV. The best fitted
two peaks are in agreement with that for Fe+2 (B.E.¼709–710 eV)
and Fe+3 (B.E.¼710–711 eV) ion [21,22]. Thus, the present XPS
investigation gives an evidence for the coexistence of the Fe+2
and Fe+3 ions in Cu1ꢀxFexO. The tetrahedral Fe3+ ion is expected
to form CuO through the oxidation of the Fe+2 ion.
where r0 and T0 are constants. The exponent
dimension. The values of
two and one dimensional system. Experimental data are fitted to
Eq. (1) as shown in Fig. 4. The best fitted value of is 0.48
for undoped CuO nanorods which supports one dimensional
conduction process. Straight line behavior of ln
ðTÞ Vs. 1000/T
plots as depicted in Fig. 4 shows thermally activated conduction
for
¼ 1. Low resistive Fe doped samples favors band conduction
mechanism.
Fig. 5 illustrates the field (H) dependent magnetization (M) of
pure CuO (inset) and Fe doped CuO nanorods at room tempe-
rature. The M–H curve for pure CuO reveals hysteresis loop but it
does not saturate up to 5 kOe. The almost linear behavior of M at
higher H suggests that ferromagnetic and antiferromagnetic
components are present in pure CuO nanorods. The coercive
field is 421 Oe. Shape anisotropy plays an important role to affect
magnetic properties. The shape anisotropy constant is propor-
tional to the aspect ratio. The coercive filed (Hc) increases with the
g
depends on the
are 0.25, 0.33 and 0.50 for three,
g
g
r
g

ARTICLE IN PRESS
2752
S. Manna, S.K. De / Journal of Magnetism and Magnetic Materials 322 (2010) 2749–2753
0.66
0.66
0.53
0.40
0.26
0.13
0.00
-0.13
-0.26
-0.40
-0.53
-0.66
5K
100K
300K
Cu0.98Fe0.02
O
O
T = 300K
0.53
0.40
Cu0.95Fe0.05
O
Cu0.90Fe0.10
0.26
0.13
0.00
-0.13
-0.26
-0.40
-0.53
-0.66
20
15
10
5
CuO
0
-5
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
-10
-15
-20
H (Tesla)
Fig. 6. Hysteresis loops (M–H) obtained at different temperatures for 10% Fe
-5 -4 -3 -2 -1
0 1 2 3 4 5
H (KOe)
doped sample. Inset shows temperature dependent coercivity (Hc).
-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0 .5 1.0 1 .5 2.0 2 .5
Cu1ꢀxFexO is formed even for the highest x¼0.10. The XPS results
H (Tesla)
exclude the presence of metallic Fe. At room temperature, the
electrical resistivity of Fe3O4 is about 10ꢀ3 Ocm and the
g- Fe2O3
Fig. 5. Magnetization (M) versus magnetic field (H) of Fe doped CuO nanorods
is an insulator. The electrical resistivity of Fe3O4 increases by two
orders of magnitude at around 120 K due to Verway transition
[25]. The resistivity of CuO nanorods is much higher than that of
Fe3O4 and also it does not reveal any sharp rise up to 100 K. Hence
the temperature dependence of resistivity discards the presence
of Fe3O4. The magnetite Fe3O4 is a ferromagnet with Curie
temperature 860 K and magnetic moment 1:33 mB=Fe. The sus-
ceptibility decreases sharply around 120 K characteristic of
Verway transition. The Fe content of pure CuO nanorod is
measure at room temperature. Inset shows for pure CuO nanorods.
increase of aspect ratio. Spherical nanoparticles do not posses
any shape anisotropy, hence have smaller Hc as reported by
Punnoose et al. [12] The aspect ratio of the present nanorods is
about five times larger than the previous studied nanorods [15].
This clearly indicates that larger Hc is consistent with higher
shape anisotropy.
The M–H curve tends to saturate with increase of Fe
concentration as depicted in Fig. 5. The highest Fe doped sample
(x¼0.10) shows distinct M–H loop which almost saturates at
100 ppm. The saturation moment for g- Fe2O3 is 0:5 mB=Fe. The
contribution from Fe to saturation magnetization is about two
orders magnitude smaller than the present values. The thermal
2.5 T. The saturated magnetic moment is 0:428
mB=Fe. The
behavior of wðTÞ for all samples can not be accounted for Fe oxides.
saturation magnetization increases significantly with increase of
Fe concentration. This clearly indicates that the incorporated Fe
ions enhance ferromagnetic coupling. The M–H hysteresis curve
at lower magnetic field for 10% Fe doped CuO at different
temperatures are shown in Fig. 6 which exhibits well-defined
hysteresis at room temperature with a coercive field of 33 Oe
All the experimental evidences suggest that the samples are not
contaminated with Fe oxides. Secondary phases as sources for the
observed ferromagnetism can be ruled out.
The electronic configurations of Cu+2, Fe+3 and Fe+2 are 3d9
3d5 and 3d6. The magnetic moments of Cu+2, Fe+3 and Fe+2 are
1:73 mB, 5:9 mB and 4:89 mB, respectively, in high spin state. The
and a remnant magnetization of 0:065
ing values at 5 K are 846 Oe and 0:281
m
B=Fe. The correspond-
B=Fe. It can be clearly
magnetic moments of Fe ions are larger than Cu ions. The incorpo-
ration of Fe in CuO lattice enhances the magnetic moment.
Ferromagnetism has not been obtained in Fe+3 doped bulk CuO
although its magnetic moment is enhanced greatly [26].
CuO nanocrystals with size larger than 10 nm behave like bulk
CuO without ferromagnetic order [12]. Antiferromagnetic nano-
crystals reveal ferromagnetic property due to the uncompensated
surface spins. Nanorods with large aspect ratio have more un-
compensated surface spins which assist the formation of ferro-
magnetic behavior. In octahedron crystal field, eg levels of Cu+2
ions are more than half filled which renders Jahn–Teller distor-
tion. First principle self-interactions-free density functional
calculation suggest that the hybridizations between Cu dz2 and
O p orbitals are responsible for antiferromagnetic Cu–O–Cu chain
along z-axis and ferromagnetic Cu–O chain in (x,y) plane [27]. The
antiferromagnetic state is stabilised by the structural distortion.
Both Fe+3 and Fe+2 are not Jahn–Teller ions and do not produce
local lattice distortion. Antiferromagnetic exchange interaction is
weaken in Fe doped CuO samples. The oriented nanorods along
m
seen that the coercivity increases significantly with decreas-
ing temperature. Inset shows the variation of coercivity with
temperature.
The temperature dependence of the susceptibility
w
ðTÞ ¼ MðTÞ=H for all samples are shown in Fig. 7. The thermal
variation of
ðTÞ for undoped nanorods is similar to pure CuO [12].
The increase of
ðTÞ below 50 K (FC) and 75 K (ZFC) is due to the
w
w
presence of some paramagnetic centers arising from the defects
and surface layers [24]. The substitution of Cu ions by Fe ions
shifts the magnetic order of CuO towards low temperature.
The paramagnetic contribution is totally suppressed due to the
prominent ferromagnetic component. The Fe doping modifies the
shape as well magnitude of
The possible impurities in all samples are metallic Fe,
magnetite Fe3O4 and maghemite - Fe2O3. The most intense
peaks in XRD for Fe3O4 and - Fe2O3 phases are not observed. The
absence of noticeable peaks indicates that a solid solution of
wðTÞ.
g
g

ARTICLE IN PRESS
S. Manna, S.K. De / Journal of Magnetism and Magnetic Materials 322 (2010) 2749–2753
2753
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
H=1000 Oe
90
80
70
60
FC
50
40
30
20
10
Cu0.98Fe0.02
Cu0.95Fe0.05
Cu0.90Fe0.10
O
O
O
ZFC
CuO
300
0
50
100
150
200
250
300
350
0
50
100
150
200
250
T (K)
T (K)
Fig. 7. (Left) Field cool (FC) and zero field cool (ZFC) susceptibility of pure CuO nanorods. (Right) Field cooled susceptibility for all doped samples.
(1 1 0) direction may strengthen the ferromagnetic coupling in
(x,y) plane. The mixed valency of Fe ions gives rise to double
exchange interaction which favors the ferromagnetic spin order.
[6] 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.
[7] K. Ueda, H. Tabata, T. Kawai, Appl. Phys. Lett. 79 (2001) 988.
[8] K.R. Kittelstved, N.S. Norberg, D.R. Gamelin, Phys. Rev. Lett. 94 (2005)
147209.
[9] M.S. Seehra, Z. Feng, R. Gopalakrishnan, J. Phys. C 21 (1988) L1051.
[10] X.G. Zheng, Y. Kodama, K. Saito, E. Tanaka, Y. Tomokiyo, H. Yamada, C.N. Xu,
Phys. Rev. B 69 (2004) 094510.
4. Conclusion
[11] X.G. Zheng, T. Mori, K. Nishiyama, W. Higemoto, C.N. Xu, Solid State Commun.
132 (2004) 493.
[12] A. Punnoose, H. Magnone, J. Bonevich, M.S. Seehra, Phys. Rev. B 64 (2001)
174420.
CuO nanorods with high aspect ratio 15–20 have been
successfully prepared by template free hydrothermal method.
Analysis of XPS and resistivity data suggest that Fe ions in CuO
nanorods exist in mixed valence state. Room temperature
ferromagnetism in CuO nanorods is significantly enhanced by
the substitution of Cu by Fe ion. Shape anisotropy, mixed valency
of Fe ions and non-Jahn–Teller property of Fe ions play important
roles to produce ferromagnetism with high Curie temperature in
Fe doped CuO nanorods.
[13] A. Punnoose, M.S. Seehra, Solid State Commun. 128 (2003) 299.
[14] K.H. Gao, Z.Q. Li, T. Du, E.Y. Jiang, Y.X. Li, Phys. Rev. B 75 (2007) 174444.
[15] H.M. Xiao, L.P. Zhu, X.M. Liu, S.Y. Fu, Solid State Commun. 141 (2007) 431.
[16] X.Y. Cui, B. Delley, A.J. Freeman, C. Stampfl, Phys. Rev. Lett. 97 (2006)
016402.
[17] Z.M. Liao, Y. Li, J. Xu, J. Zhang, K. Xia, D. Yu, Nano Lett. 6 (2006) 1087.
[18] L. Lutterotti, MAUD, Version 2.046, 2006, /http://www.ing.unitn.it/maud/S.
[19] Y. Cudennec, A. Lecerf, Solid State Sci. 5 (2003) 1471.
[20] J. Ghijsen, L.H. Tjeng, J. Van Elp, H. Eskes, J. Westerink, G.A. Sawatzky,
M.T. Czyzyk, Phys. Rev. B 38 (1988) 11322.
[21] D. Castro, V. Ciampi, Surf. Sci. 331 (1995) 294.
References
[22] M. Muhler, R. Schlogl, G. Ertl, J. Catal. 138 (1992) 413.
[23] N.F. Mott, E. Devis, Electronic Processing Non Crystalline Materials, second
ed., Clarendon, Oxford, 1979.
[1] H. Ohno, Science 281 (1998) 951.
[2] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287 (2000) 1019.
[3] A.H. Macdonald, P. Schiffer, N. Samarth, Nat. Mater. 4 (2005) 195.
[4] K. Ando, Science 312 (2006) 1883.
[5] W.K. Park, R.J. Ortega-Hertogs, J.S. Moodera, A. Punnoose, M.S. Seehra, J. Appl.
Phys. 91 (2002) 8093.
[24] T.V. Chandrasekhar, V.C. Sahni, J. Phys. Condens. Matter 6 (1994) L423.
[25] F. Walz, J. Phys. Condens. Matter 14 (2002) R285.
[26] R.A. Borzi, S.J. Stewart, G. Punte, R.C. Mercader, G.A. Curutchet, R.D. Zysler,
M. Tovar, J. Appl. Phys. 87 (2000) 4870.
[27] A. Filippetti, V. Fiorentini, Phys. Rev. Lett. 95 (2005) 086405.