Magnetic linear dichroism studies of in situ grown NiO thin films

Article abstract of DOI:10.1016/j.jmmm.2006.07.020

We report thickness dependence of magnetic linear dichroism (MLD) of in situ grown NiO(0 0 1) films on Ag(0 0 1) substrate at the Ni L₂ absorption edge. Antiferromagnetic domains at the surface of NiO(0 0 1) films are found to be preferentially aligned in-plane. For films thinner than a critical thickness t_c (20-40ML), we observe a softening of the in-plane magnetic domain alignments with increasing film thickness, arising from the strain-relaxation effects. Films thicker than t_c exhibits a residual in-plane anisotropy, possibly related to the finite-thickness effects.

Full text of DOI:10.1016/j.jmmm.2006.07.020

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 310 (2007) 8–12
www.elsevier.com/locate/jmmm
Magnetic linear dichroism studies of in situ grown NiO thin films
Ã
S.R. Krishnakumara, , M. Liberatib,c, C. Graziolid, M. Veronesed, S. Turchinid, P. Luchesc,
S. Valeric, C. Carboned
aInternational Centre for Theoretical Physics (ICTP), Strada Costiera 11, 34100 Trieste, Italy
bDipartimento di Fisica, Universita di Modena e Reggio Emilia, Via Campi 213/a, 41100 Modena, Italy
cINFM–National Center on nanoStructures and bioSystems at Surfaces (S3), Via Campi 213/a, 41100 Modena, Italy
dIstituto di Struttura della Materia, Consiglio Nazionale delle Ricerche, Area Science Park, I-34012 Trieste, Italy
Received 12 May 2006; received in revised form 14 July 2006
Available online 10 August 2006
Abstract
We report thickness dependence of magnetic linear dichroism (MLD) of in situ grown NiO(0 0 1) films on Ag(0 0 1) substrate at the
Ni L2 absorption edge. Antiferromagnetic domains at the surface of NiO(0 0 1) films are found to be preferentially aligned in-plane. For
films thinner than a critical thickness tc (20–40ML), we observe a softening of the in-plane magnetic domain alignments with increasing
film thickness, arising from the strain-relaxation effects. Films thicker than tc exhibits a residual in-plane anisotropy, possibly related to
the finite-thickness effects.
r 2006 Elsevier B.V. All rights reserved.
PACS: 75.25.+z; 75.50.Ee; 75.30.Gw; 75.70.ꢀi
Keywords: Nickel oxide; Antiferromagnetism; Magnetic linear dichroism; Thin films
1. Introduction
material also for device applications. There have been
several studies on the growth and structural characteriza-
tion of the thin oxide films [3–13], but many aspects of the
magnetic properties are still unclear. It is essential to
understand the antiferromagnetic (AF) spin structure and
domain behaviors of these films if we have to manipulate
them in a controlled way to be useful for applications. In
recent years, magnetic linear dichroism (MLD) in polar-
ization dependent X-ray absorption spectroscopy (XAS)
has been used to study the antiferromagnetism of thin films
such as NiO [14], LaFeO3 [15], CoO [16] and Mn on
Fe(1 0 0) [17] etc. This technique can provide magnetic
contrast with chemical and spatial sensitivity and can be
used to study even fractions of a monolayer of materials.
Using polarization dependence at Ni L2 edge, magnetic
domains on NiO(0 0 1) single crystal surfaces have been
imaged [18,19] and the spin orientations at the interface
between NiO(0 0 1) and ferromagnetic overlayers have been
explored [20,21].
Ultra-thin oxide films have become a topic of great
interest in recent years due to their technological relevance
in the field of magnetic data storage and spintronic devices
[1,2]. There are many fundamental issues related to the
electronic structure and magnetism of 2D oxide overlayers
grown on different substrates, which still remains to be
explored. Nickel oxide (NiO) has been the model system to
study many important properties like antiferromagnetism,
electron correlations in narrow band systems etc. and is
still interestingly a model system to study the growth and
properties of low dimensional oxide layers. The appealing
electronic properties of NiO such as large bandgap
ðꢁ4 eVÞ, high-magnetic ordering temperature ðTN
¼
520 KÞ and high-thermal stability make it a favorable
Ã
Corresponding author. Present address: Surface Physics Division,
Saha Institute of Nuclear Physics, 1/AF Bidhannagar, Kolkata 700 064,
India. Tel.: +91 33 2337 5345; fax: +91 33 2337 4637.
E-mail address: krishna.menon@saha.ac.in (S.R. Krishnakumar).
It was demonstrated by Alders et al. [14] that it is possible
to study the long range AF order and nearest-neighbor
0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2006.07.020

ARTICLE IN PRESS
S.R. Krishnakumar et al. / Journal of Magnetism and Magnetic Materials 310 (2007) 8–12
9
spin–spin correlations using polarization dependent XAS,
by studying thin films of NiO(0 0 1) grown on MgO(0 0 1).
They observed that the AF domains on these samples were
preferentially aligned normal to the sample surface.
Exchange Biasing of NiO films were also studied using
MLD at Ni L2 edge by Zhu et al. [22] and observed a
repopulation of the NiO AF domains with in-plane spin
component. A recent study by Altieri et al. [23] has pointed
out that the MLD in Ni L2 XAS is reversed for NiO(0 0 1)
films grown on Ag(0 0 1), which has reversed lattice
mismatch compared to NiO(0 0 1) on MgO(0 0 1) substrate.
This points to an in-plane AF domain alignment in the case
of Ag(0 0 1) substrate and this reversal of the magnetic
domain alignment from out-of-plane to in-plane is pro-
posed [24] to be due to the effect of interfacial strain at NiO-
substrate interface. On MgO capped NiO(0 0 1) thin films
on Ag(0 0 1) substrate, they observed in-plane domains
(ꢀve MLD) for thin NiO(0 0 1) layers while out-of-plane
domains (þve MLD) were observed for thicker NiO layers,
due to the effect of the capped MgO layers. Crystal-field
effects in ultra-thin films and surfaces can also cause linear
dichroism which can interfere with the MLD signal. A
single monolayer (ML) of NiO(0 0 1) film, sandwiched
between Ag(0 0 1) substrate and MgO capping layers,
showed [25] strong linear dichroism effect in the Ni L2;3
XAS spectra, arising from the lower symmetry environment
of the Ni2þ ions in the monolayer film ðC4vÞ, in contrast to
the Oh environment in the bulk.
and annealing cycles to 723 K. Growth of NiO films were
achieved by evaporation of Ni metal in oxygen environ-
ment. The partial pressure of oxygen was optimized for the
full oxidation of the evaporated Ni metal. The Ni
deposition rates from the e-beam evaporators were
calibrated by the quartz crystal microbalance and was set
for the oxide growth rate of approximately 1ML/min.
During the growth, the sample temperature was set to
490 K, which enhances the surface mobility and improves
the structural quality of the film [7]. The films were grown
in the form of wedged steps with the help of a sharp
shutter. This enabled us to study many film thicknesses
grown under the same growth conditions and substrate.
Both the substrate and oxide films were structurally
characterized by LEED and were in good agreement with
previous studies [7].
3. Results and discussion
In Fig. 1, we show the experimental results from 14ML
NiO(0 0 1) film grown on Ag(0 0 1) substrate. The Ni L2
XAS spectra taken at room temperature is shown in the
main panel for different incident beam polar angles. The
schematic diagram in the figure shows the experimental
geometry used, with y ¼ 90ꢃ being the normal incidence
~
with the electric field vector E lying in the plane of the
crystal while y ¼ 0ꢃ making the electric field vector normal
to the surface. The spectra are normalized at the first peak
of the L2 edge at 869.7 eV. The intensity of the second peak
at 870.8 eV decreases as y decreases and this polar angle
dependence is different from those reported by Alders et al.
[14] on NiO(0 0 1) film on MgO(0 0 1) substrate. The L2
ratio, which is the ratio of the L2 low-energy peak to L2
high-energy peak intensity, increases with increasing polar
angle on MgO(0 0 1) substrate and is attributed to the NiO
AF domains having a preferential out-of-plane alignment
Most of these studies were performed on samples which
were either exposed to ambient or grown with protective
capping layers, though it is most desirable to study them
as-grown. The motivation of the present study is to
understand the magnetic properties of the NiO thin films
with strain relaxation and in absence of capping overlayers.
Here, we present our results on the thickness dependent
MLD study of NiO(0 0 1) thin films, grown in situ on
Ag(0 0 1) substrates.
2. Experimental
All the experiments were performed at the Circular
Polarized Beamline (BL 4.2) of the Elettra Synchrotron
Light Source at Trieste, Italy. X-ray absorption spectra
were collected with linear polarized light ðꢁ95%Þ from the
elliptical electromagnetic wiggler in the total electron yield
(TEY) mode. The beam size on the sample was about
200 mm ꢂ 500 mm. Reproducibility of the experimental
results was established with samples prepared at different
experimental runs and with different substrate crystals.
All the experiments were performed at room temperature
and the spectra were corrected for a linearly increasing
background.
Thin film samples used for this study were prepared in
the growth chamber and then transferred in situ to the
analysis chamber for measurements. The NiO(0 0 1) films
were grown on the Ag(0 0 1) single crystal, which was
prepared by repeated cycles of Arþ ion (700 eV) sputtering
Fig. 1. Ni L2 XAS spectra from 14ML of NiO(0 0 1) at 300 K at different
polar angles. Azimuthal angular dependence of the L2 spectra is shown in
the inset.

ARTICLE IN PRESS
S.R. Krishnakumar et al. / Journal of Magnetism and Magnetic Materials 310 (2007) 8–12
10
to the sample surface [14]. Experimental data in Fig. 1
shows decreasing L2 ratio with increasing polar angle on
Ag(0 0 1) substrate indicating a preferential alignment of
the domains in-plane of the film. Similar AF domain
structures were also observed by Altieri et al. [23] on MgO
capped 3ML of NiO(0 0 1) film on Ag(0 0 1) substrate. This
difference in the domain structures of the NiO films on
MgO(0 0 1) and Ag(0 0 1) substrates have been ascribed to
the reversed lattice mis-match between NiO–Ag compared
to NiO–MgO [24], as mentioned in the Introduction.
In the inset of Fig. 1, we show the azimuthal angle
dependence of the same 14ML of NiO film, which provides
information on the in-plane distribution of the AF
domains. Instead of rotating the sample, we rotate the
polarization of the incoming beam from horizontal to the
vertical (as shown in the schematic) to probe the two
orthogonal directions. The vertical polarized spectrum has
a smaller line-width as a consequence of the higher energy
resolution in this mode, however, with an order of
magnitude decrease in the photon flux. We see that the
L2 ratio remains unchanged at orthogonal directions,
indicating that the AF domain populations are isotropic
in-plane and anisotropic out-of-plane as seen with the
polar angle dependence. We observe similar azimuthal
angle dependence in the whole thickness range of
NiO(0 0 1) films studied here.
24ML films are shown in solid and dashed lines,
respectively. The individual spectra are area-normalized
and the difference between the spectra gives the dichroism
signal. The qualitative nature of the XAS spectra and the
dichroism signal are rather similar for films with different
thickness, as evident from the figure. For a detailed
comparison, we show the experimental results for
˚
NiO(0 0 1) films in the thickness range 6–72ML ðꢁ150 AÞ
in Fig. 3. In the main panel, we show the L2 ratio obtained
from NiO(0 0 1) films for y ¼ 30ꢃ (closed circles) and 90ꢃ
(open circles), at room temperature. The L2 ratio do not
show much variation with thickness in the thickness range
studied. We plot the percentage of ꢃMLD obtained from the
data using the relation, ð1 ꢀ ðL320 =L290ꢃ ÞÞ ꢂ 100% for the
same thickness range in the inset of Fig. 3. Now, we can see
the variation of MLD with thickness, with thick solid line
as a guide to the eye. The MLD decreases from ꢁ ꢀ 19%
with increasing film thickness to ꢁ ꢀ 15% at about 40ML.
Interestingly, further increasing the film thickness does not
have any effect on the MLD of the film. These results are in
contrast to the findings of Altieri et al. [23] where a reversal
of the sign of the MLD was observed for a thickness of
30ML NiO film, resulting from the presence of the MgO
capping layer. In absence of the capping layers, we are able
to study the effect of strain relaxation of the film on its
magnetic properties, though the magnitude of the effects
observed here are smaller compared to the MgO capped
films.
Fig. 2 shows the polar dependence of the Ni L2 XAS
spectrum taken at room temperature for NiO(0 0 1) films
with different thickness. Here, the grazing incidence ðy ¼
30ꢃÞ and normal incidence ðy ¼ 90ꢃÞ spectra for 6, 12 and
It is known that the Ne
´
el temperature dependents on the
el
thickness of the film. Alders et al. [14] puts the Ne
´
temperature of 20, 10 and 5ML NiO(0 0 1) on MgO(0 0 1)
to be 470 K, 430 K and 295 K, respectively. A change in the
Neel temperature of the film will have an effect on the
´
magnitude of the L2 ratio and will decrease the MLD.
These effects will be important for films with thickness
lower than 10ML, while for films thicker than 10ML no
Fig. 3. L2 ratio extracted from the Ni L2 edge of NiO(0 0 1) films of
different thickness at normal and grazing incidence angles. Inset shows the
percentage of MLD observed obtained from the data and the solid line is a
guide for the eyes.
Fig. 2. Ni L2 XAS spectra from 6, 12 and 24ML of NiO(0 0 1) thin films
for y ¼ 30ꢃ and 90ꢃ.

ARTICLE IN PRESS
S.R. Krishnakumar et al. / Journal of Magnetism and Magnetic Materials 310 (2007) 8–12
11
significant changes are expected. Thus, we can conclude
that the qualitative nature of the experimental results in
MLD with increase of thickness is indeed a strain induced
effect.
Fig. 3 do not depend on the Ne
´
el temperature of the films,
For 72ML film, which is thicker than the critical
thickness for the strain relaxation, it is expected to have
lattice parameters close to the bulk while the observed
MLD saturates to near ꢀ15%. However, our experiments
on freshly cleaved NiO(0 0 1) single crystal surface [26]
shows insignificant amount of MLD. This suggests that
NiO films with thickness above the critical thickness shows
a residual in-plane magnetic anisotropy and is important to
understand the origin of this anisotropy. It cannot be ruled
out that there is still small residual strain remaining in the
thick films yielding the observed in-plane anisotropy. This
residual strain should relax at even larger thickness, leading
to gradual removal of this in-plane anisotropy. Experi-
ments on films thicker than used in the present study might
help to resolve this.
According to the theoretical model by Finazzi and
Altieri [24], the magnetic dipolar anisotropy in the AF films
such as NiO depends on the combination of strain and film
thickness. In the case of NiO(0 0 1) films on Ag(0 0 1)
substrate, the compressional strain and the finite-thickness
effects are found to act cooperatively stabilizing the in-
plane domains [24]. Even in absence of strain, this model
predicts a stabilization of in-plane AF domains due to
finite-thickness effects, resembling the shape anisotropy
phenomenon in ferromagnetic thin films. This can provide
an explanation to the observed in-plane anisotropy above
tc. However, the shape-anisotropy term in the anisotropy
Hamiltonian is inversely proportional to the thickness [24]
and should have weaker contribution for thicker films.
Other factors such as the stoichiometry of the films also
cannot be ruled out as cause of the observed in-plane
anisotropy as their influence on the magnetic properties of
the films are still unexplored.
Thickness dependence of MLD in Fig. 3 suggests a
continuous evolution of the magnetic domain structure
with increasing film thickness which settles down above the
critical film thickness ðtcÞ. In case of bulk NiO, the Ni
moments can orient in 24 possible [2 1 1]-derived directions
in the crystal, and as many different AF domains are
possible [14]. On NiO(1 0 0) surface, this amounts to 16
possible in-plane and eight possible out-of-plane directions
for the Ni moments. From this simple picture of discrete
moment directions, it is difficult to perceive the observed
broad and continuous changes in MLD. Spin reorientation
between in-plane and out-of-plane moment directions
would have lead to a rather abrupt transition, as seen in
ferromagnetic systems [27]. Although the present experi-
mental results are not sufficient to fully describe the nature
of the transition, some speculations can be made. The
present results may be explained by the strain dependence
of magnetic anisotropy of the AF films as suggested by
Finazzi and Altieri [24]. As a result of the strain and
thickness dependent effects, ultrathin films have a strong
in-plane magnetic anisotropy which weakens with increase
of the film thickness and strain relaxation. Strong in-plane
at least for thickness larger than 10ML. We can also rule
out the possible influence of the crystal field effects on the
Ni L2 XAS spectra [25] as the crystal field induced splitting
occurs in low symmetry D4h point group and such an effect
should be minimal even for 6ML thin film. Moreover, we
do not observe any energy shift in the peak positions of the
L2 and L3 features, as observed by Haverkort et al. [25],
confirming the magnetic origin of the dichroism. The
results in the inset of Fig. 3 are interesting as we can
identify two thickness regimes, below ꢁ40ML where the
magnitude of MLD decreases with increase of thickness
and above ꢁ40ML range where the MLD remains
independent of the thickness. As the size of the observed
effect is small and given the size of the error bars, we assign
a range of 20–40ML for the critical thickness ðtcÞ, above
which MLD saturates. We discuss these results in the
following paragraphs.
˚
The lattice parameter for Ag, NiO and MgO are 4:086 A,
˚
˚
4:176 A and 4:212 A, respectively. This results in a lattice
mismatch at the NiO–Ag interface of about ꢀ2.2%,
yielding an in-plane tetragonal compressional strain on
the films, while the mismatch at the NiO–MgO interface is
about 0.85% with in-plane tensile strain. It is suggested
that this difference in the strain at interface is responsible
for the reversal of the domain alignment between
NiO(0 0 1) films on MgO(0 0 1) and Ag(0 0 1) substrates
[23,24]. However, as the film thickness increases, the strain
on the film relaxes and attains near-bulk lattice parameters
after a material dependent critical thickness. The decrease
of the percentage of MLD with increase of thickness below
tc can be attributed to the effect of the strain relaxation in
the film. The observed decrease of the MLD signal at
40ML is about 20% of the value at 6ML film. Though
strain at the NiO–Ag interface is small, its influence on the
magnetic domain structure is significant as NiO has weak
magnetocrystalline anisotropy and degenerate domain
directions. As the film thickness increases, the strain on
the film relaxes and alters the magnetocrystalline aniso-
tropy which disturbs the statistical population of the
magnetic domains resulting in the change of MLD signal.
It is found from LEED studies of NiO(0 0 1) thin films on
Ag(0 0 1) [7] that the surface in-plane lattice parameter is
fully relaxed at a thickness of 20ML. Another study using
Oxygen K-edge XAS [3], it was shown that the experi-
mental lattice parameters were close to that of bulk single
crystal, by about 50ML thickness of NiO(0 0 1). This
difference in the critical thickness between the two
techniques can be understood as the sample probing depth
of XAS is quite different compared to LEED. XAS sees the
average lattice parameter within the probing volume, while
LEED is most sensitive to the topmost layer. It is
interesting to note that the thickness range where MLD
saturates ðtcÞ is close to the 50ML film with near-bulk
lattice parameters [3], confirming that the decrease of

ARTICLE IN PRESS
S.R. Krishnakumar et al. / Journal of Magnetism and Magnetic Materials 310 (2007) 8–12
12
anisotropy for low film thickness points to a major
population of in-plane domains compared to that of out-
of-plane domains. As the film thickness increases, the in-
plane magnetic anisotropy energy decreases, resulting in a
decrease of in-plane domain populations relative to that of
out-of-plane. This decrease of in-plane domain populations
results in decrease of the MLD signal. The continuous
nature of the strain relaxation causes a continuous decrease
of the MLD, which attains saturation above tc. Here, the
magnetic anisotropy does not change direction to drive any
spin reorient transition, but changes in magnitude causing
small changes in the relative domain populations and
yielding small effects in MLD as seen in Fig. 3.
It is also possible that the Ni moment directions in thin
NiO films are sensitive to the presence of the strain in the
lattice. For the strained thin films, Ni moment directions
deviate from their bulk values more towards in-plane such
that the MLD is larger. As the film thickness increases, the
strain relaxes, resulting in moment directions returning to
the bulk values. Further experiments are required to
elucidate the mechanisms responsible for the observed
changes in MLD. Detailed MLD spectromicroscopy
experiments on NiO(1 0 0) films might help to resolve these
issues.
References
[1] J. Nogues, I.K. Schuller, J. Magn. Magn. Mater. 192 (1999) 203.
[2] I. Zutic, J. Fabian, S.D. Sarma, Rev. Mod. Phys. 76 (2004) 323.
[3] E. Groppo, C. Prestipino, C. Lamberti, R. Carboni, F. Boscherini,
P. Luches, S. Valeri, S. D’Addato, Phys. Rev. B 70 (2004) 165408.
[4] C. Lamberti, E. Groppo, C. Prestipino, S. Casassa, A.M. Ferrari,
C. Pisani, C. Giovanardi, P. Luches, S. Valeri, F. Boscherini, Phys.
Rev. Lett. 91 (2003) 046101.
[5] K. Marre, H. Neddermeyer, A. Chasse, P. Rennert, Surf. Sci. 358
(1996) 233.
[6] P. Luches, S. Altieri, C. Giovanardi, T.S. Moia, S. Valeri, F. Bruno,
L. Floreano, A. Morgante, A. Santaniello, A. Verdini, R. Gotter,
T. Hibma, Thin Solid Films 400 (2001) 139.
[7] C. Giovanardi, A. di Bona, S. Altieri, P. Luches, M. Liberati,
F. Rossi, S. Valeri, Thin Solid Films 428 (2003) 195.
[8] F. Muller, R. de Masi, P. Steiner, D. Reinicke, M. Stadtfeld,
S. Hufner, Surf. Sci. 459 (2000) 161.
[9] J. Wollschlager, D. Erdos, H. Goldbach, R. Hopken, K.M. Schroder,
Thin Solid Films 400 (2001) 1.
[10] M. Schulze, R. Reissner, Surf. Sci. 507 (2002) 851.
[11] M. Portalupi, L. Duo, G. Isella, R. Bertacco, M. Marcon,
F. Ciccacci, Phys. Rev. B 64 (2001) 165402.
[12] R. Reissner, U. Radke, M. Schulze, E. Umbach, Surf. Sci. 402 (1998)
71.
[13] R. Reissner, M. Schulze, Surf. Sci. 454 (2000) 183.
[14] D. Alders, L.H. Tjeng, F.C. Voogt, T. Hibma, G.A. Sawatzky, C.T.
Chen, J. Vogel, M. Sacchi, S. Iacobucci, Phys. Rev. B 57 (1998)
11623.
[15] A. Scholl, J. Stohr, J. Luning, J.W. Seo, J. Fompeyrine, H. Siegwart,
J.-P. Locquet, F. Nolting, S. Anders, E.E. Fullerton,
M.R. Scheinfein, H.A. Padmore, Science 287 (2000) 1014.
[16] S.I. Csiszar, M.W. Haverkort, Z. Hu, A. Tanaka, H.H. Hsieh,
H.-J. Lin, C.T. Chen, T. Hibma, L.H. Tjeng, Phys. Rev. Lett. 95
(2005) 187205.
4. Conclusion
We have investigated the thickness dependence of the
MLD of NiO(0 0 1) films grown on Ag(0 0 1) substrate at
Ni L2 edge. We observe in-plane magnetic domains in these
films with MLD decreasing with thickness below a critical
thickness tc, resulting from the strain relaxation of the film.
In absence of the capping layers, there was no reversal of
the MLD with thickness. For films thicker than tc, the
MLD saturates leaving a large in-plane anisotropy in the
film, possibly arising from the finite-thickness effects.
Further investigations are required for a detailed under-
standing of these observations.
[17] C. Grazioli, D. Alfe, S.R. Krishnakumar, S.S Gupta, M. Veronese,
S. Turchini, N. Bonini, A.D. Corso, D.D. Sarma, S. Baroni,
C. Carbone, Science 95 (2005) 117201.
[18] J. Stohr, A. Scholl, T.J. Regan, S. Anders, J. Luning, M.R.
Scheinfein, H.A. Padmore, R.L. White, Phys. Rev. Lett. 83 (1999)
1862.
[19] N.B. Weber, H. Ohldag, H. Gomonaj, F.U. Hillebrecht, Phys. Rev.
Lett. 91 (2001) 237205.
[20] H. Ohldag, A. Scholl, F. Nolting, S. Anders, F.U. Hillebrecht,
J. Stohr, Phys. Rev. Lett. 86 (2001) 2878.
[21] H. Ohldag, T.J. Regan, J. Stohr, A. Scholl, F. Nolting, J. Luning, C.
Stamm, S. Anders, R.L. White, Phys. Rev. Lett. 87 (2001) 247201.
[22] W. Zhu, L. Seve, R. Sears, B. Sinkovic, S.S.P. Parkin, Phys. Rev.
Lett. 86 (2001) 5389.
Acknowledgments
[23] S. Altieri, M. Finazzi, H.H. Hsieh, H.-J. Lin, C.T. Chen, T. Hibma,
S. Valeri, G.A. Sawatzky, Phys. Rev. Lett. 91 (2003) 137201.
[24] M. Finazzi, S. Altieri, Phys. Rev. B 68 (2003) 054420.
[25] M.W. Haverkort, S.I. Csiszar, Z. Hu, S. Altieri, A. Tanaka, H.H.
Hsieh, H.-J. Lin, C.T. Chen, T. Hibma, L.H. Tjeng, Phys. Rev. B 69
(2004) 020408(R).
The authors would like to acknowledge the support by
N. Zema and other personnels at Circular Polarized
Beamline, Elettra Synchrotron Center. One of the authors
(S.R.K) undertook this work with the support of the
‘‘ICTP Programme for Training and Research in Italian
Laboratories, Trieste, Italy’’.
[26] S.R. Krishnakumar, et al., to be published.
[27] P.J. Jensen, K.H. Bennemann, Surf. Sci. Rep. 61 (2006) 129.