Ferromagnetic-antiferromagnetic Fe/NiO (1 0 0) interface studied by non-linear Kerr effect

Article abstract of DOI:10.1016/j.susc.2007.06.011

In this work, the magnetic properties of the Fe/NiO (1 0 0) interface at room temperature is investigated using non-linear magneto-optical Kerr effect. Nickel oxide is a promising material for technological applications due to its Néel temperature (525 K) occurring well above the room temperature. Non-linear Kerr rotation measurements have demonstrated to be an effective way to investigate a ferromagnetic/antiferromagnetic interface. The measurements here reported demonstrate that the magnetic properties of the studied interface strongly depend on the Fe layer thickness. In fact, a clear hysteresis loop is observed only for a thickness higher than 5 ML, whereas for the 2 ML sample the hysteresis loop completely disappears. These results are in agreement with previous structural characterizations of the interface.

Full text of DOI:10.1016/j.susc.2007.06.011

Surface Science 601 (2007) 4362–4365
www.elsevier.com/locate/susc
Ferromagnetic–antiferromagnetic Fe/NiO (100) interface studied
by non-linear Kerr effect
a,
a
a,c
b
b
c
c
*
S. Colonna , F. Ronci , A. Cricenti , P. Luches , S. Valeri , J. Qi , Y. Xu ,
c
c
J.K. Miller , N. Tolk
a
CNR–Istituto di Struttura della Materia, Via del Fosso del Cavaliere 100, I-00133 Roma, Italy
S3-INFM Dipartimento di Fisica, Universit a` di Modena e Reggio Emilia, via G. Campi 213/a, I-41100 Modena, Italy
b
c
Department of Physics and Astronomy, Vanderbilt University, Nashville, TN 37235, USA
Available online 13 June 2007
Abstract
In this work, the magnetic properties of the Fe/NiO (100) interface at room temperature is investigated using non-linear magneto-
optical Kerr effect. Nickel oxide is a promising material for technological applications due to its N e´ el temperature (525 K) occurring well
above the room temperature. Non-linear Kerr rotation measurements have demonstrated to be an effective way to investigate a ferro-
magnetic/antiferromagnetic interface.
The measurements here reported demonstrate that the magnetic properties of the studied interface strongly depend on the Fe layer
thickness. In fact, a clear hysteresis loop is observed only for a thickness higher than 5 ML, whereas for the 2 ML sample the hysteresis
loop completely disappears. These results are in agreement with previous structural characterizations of the interface.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Exchange bias; Non-linear optics; Kerr effect
1
. Introduction
(100) surface on top oxygen atoms thus the surface crystal
cell is rotated 45° with respect to the underling [100] direc-
tion of the substrate [2,3]. A comprehensive understanding
of the chemical and physical properties of the interface is of
paramount importance for future applications. Several
studies have demonstrated that the Fe/NiO interface is
not entirely abrupt but an oxygen diffusion takes place
resulting in a reduced NiO layer [2–5]. A structural charac-
terization of the Fe/NiO interface previously conducted by
X-ray absorption spectroscopy in combination with den-
sity functional calculations has provided evidence that a
FeO-like layer is formed at the interface between the metal
and the oxide [6]. As a consequence the magnetic moment
of the Fe atoms at the interface are increased. Uncompen-
sated moments coming from the interfacial FeO layer,
which may couple ferromagnetically with the Fe layer,
are expected to dramatically influence the exchange inter-
action at the Fe/NiO interface, with deep consequences
also for the exchange bias mechanism.
The interface formed between a ferromagnetic (FM) and
an antiferromagnetic (AFM) material has been a subject of
study since the discovery of the exchange bias phenome-
non, which is characterized by a shift of the hysteresis loop
of the ferromagnetic layer in the direction of the applied
field [1]. This subject has recently received strong renewed
interest due to the application of the exchange bias in
recording media and reading heads in data storage devices.
Different models for the description of the exchange bias
mechanism have been proposed, however, a general quan-
titative description is still lacking.
Nickel oxide is a promising material for technological
applications due to its N e´ el temperature (525 K) which is
well above room temperature. The Fe grows on the NiO
*
Corresponding author. Tel.: +39 06 4993 4163; fax: +39 06 4993 4153.
E-mail address: colonna@ism.cnr.it (S. Colonna).
0
039-6028/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2007.06.011

S. Colonna et al. / Surface Science 601 (2007) 4362–4365
4363
Non-linear optical techniques have been employed to
study the interfaces and surfaces of different materials
of Fe was checked by low energy electron diffraction. Pre-
vious measurements have demonstrated that the capping
layer is effective against oxidation induced by air [15].
For SHG measurements the output beam from a mode-
locked Ti:sapphire laser (150 fs pulses duration, 800 nm
wavelength, 76 MHz repetition rate) was focalized to a
spot of about 100 lm at the sample surface (average laser
power 190 mW). A glan polarizer and a quarter-wave plate
were used to obtain a pure ^s polarized pump beam. The
sample was installed on a rotating stage located in the
gap of an electro-magnet. The radiation beam coming from
the sample passed through a prism to separate the funda-
mental from the second harmonic response. Two polarizers
were used to measure the polarization state of the two
beams. The reflected beam intensity was measured by a
photodiode whereas the second harmonic intensity was
measured by a photomultiplier. The two signals were de-
tected by means of a lock-in amplifier because the pump
beam was modulated by an acoustic optical modulator.
Before the SHG measurements we verified that a bare
MgO substrate produced an undetectable amount of second
harmonic signal. Taking into account that the Au capping
layer is not magnetic, it is possible to say that the magnetic
signal comes from the NiO (100)/Fe interface only.
[
7,8]. In particular second harmonic generation (SHG)
has proven to be effective in the study of the magnetic state
of metal surfaces and thin films [9].
The second harmonic polarization (P(2x)) induced by
the electric field of the incoming light beam, in the electric
dipole approximation can be written as
ð2Þ
Pð2xÞ ¼ v EðxÞEðxÞ
ð1Þ
(2)
where E(x) is the fundamental electric field and v is the
second harmonic susceptibility tensor. In centrosymmetric
(2)
media the bulk v is zero, but a second harmonic signal
can be produced through a higher order term in (1), namely
the magnetic quadrupole term. At the surface, the inver-
sion symmetry is broken and SHG takes place. The region
where the second harmonic radiation is produced extends
few atomic layers around the surface or interface [10].
The magnetization of the surface lowers the surface sym-
(
2)
metry then modifies the v tensor. The susceptibility ten-
sor can be written as a sum of two terms, even and odd
in the magnetization
ð2Þ
þ
ð2Þ
ꢀ
ð2Þ
ijk
v ¼ v þ v ; i; j; k ¼ x; y; z
ð2Þ
ijk
ijk
the nonvanishing elements of the susceptibility tensor de-
pend on the symmetry of the investigated surface [9]. As
a consequence, by SHG is possible to investigate the mag-
netization state of the FM film in a region of few ML
around a FM/AFM interface. This information is of cru-
cial importance in the understanding of the exchange bias
mechanism because the magnetic coupling between the
two materials takes place near the interface. Experimental
evidence of non-linear magneto-optic effects has been re-
ported [11–13] and the non-linear magneto-optical Kerr ef-
fect (NOMOKE) has proven an effective way to investigate
FM/AFM interface for the polycrystalline Ni Fe /
3
. Results and discussion
The reported SHG measurements were performed using
a ^s polarization configuration for the pump laser impinging
the sample surface. In Fig. 1 the polarization dependence
H
s
p
8
1
19
Fe0.5Mn0.5 system [14].
In this work we present SHG measurements performed
on the Fe/NiO (100) system. In particular we measured the
NOMOKE induced by the magnetization at the FM/AFM
interface.
2
. Experiment
The samples were prepared ex situ in an ultra high vac-
uum chamber equipped with Ni, Fe and Au evaporating
sources, following a well established procedure [4]. The
NiO (100) surface was produced by evaporating Ni in a
suitable O atmosphere using MgO (100) as substrate,
2
0
50
100
150
200
250
300
350
1
0 ML of NiO were deposited in this way. On the freshly
Angle (deg.)
prepared NiO surface Fe was deposited at different thick-
nesses, then 10 nm of Au were deposited as a capping layer
to avoid oxidation of the sample. Three different samples
were prepared with 2, 5 and 10 ML Fe deposits, respec-
tively. The nickel oxide stoichiometry was verified by X-
ray photoelectron spectroscopy and the epitaxial growth
Fig. 1. Polarization dependence of the SHG measured on the 5 ML Fe
sample in longitudinal configuration. The inset shows the magnetic field
direction with respect to the incidence plane. The two curves were acquired
for the positive (squares) and negative (circles) magnetization, respec-
tively. The continuous lines are cos(h) fitting curves of the experimental
points.
2

4364
S. Colonna et al. / Surface Science 601 (2007) 4362–4365
of SHG signal for the 5 ML Fe sample is reported. The two
curves are measured by applying two values of magnetic
field of ± 900 Oe, respectively in the longitudinal configura-
tion (see inset in Fig. 1), along the [110] direction and
clearly show the strong sensitivity of the second harmonic
signal to the magnetic state of the surface. In fact, the
two polarization patterns are shifted by about 17°, indicat-
ing a NOMOKE Kerr angle by far larger than the typical
linear MOKE values measured in bulk ferromagnetic
metals. For comparison, in Fig. 2 the measurement of the
fundamental reflected radiation intensity as a function of
the polarizer angle is reported for 10 ML Fe on NiO
10ML Fe/NiO(100)
(
100). In this case the effect of the magnetic field is hardly
detectable demonstrating that the NOMOKE technique
is particularly well suited for the investigation of the
samples studied in this work. In fact, the linear radiation
from the sample is dominated by the light reflected from
the Au capping layer. The use of the SHG, then, opens
the possibility to easily probe the capped FM/AFM
interface.
5
ML Fe/NiO(100 )
The evolution of the second harmonic signal as a func-
tion of the applied magnetic field was measured by fixing
the polarizer angle and by measuring the SHG as a func-
tion of the magnetic field. In Fig. 3 this measurement is re-
ported for the 5 ML and 10 ML samples. In the two
measurements a hysteresis loop is observable. The reported
measurements demonstrate that even at this low thickness
the iron film is magnetically active.
With appropriate combinations of input and output
beam polarizations and geometries it is possible to probe
different elements of the susceptibility tensor, thereby
-400 -300 -200 -100
0
100 200 300 400
Magnetic Field (Oe)
Fig. 3. SHG measurements as a function of the applied magnetic field,
measured for the 10 ML and 5 ML samples, respectively. These curves
were acquired for the longitudinal configuration of the magnetic field.
obtaining a comprehensive characterization of the inter-
face. A detailed analysis of the non-linear optical measure-
ments will be reported in a forthcoming paper [16]. Here we
focus on the qualitative aspects of the magnetic behavior of
the Fe/NiO interface as a function of the metal thickness.
Comparing the 5 ML and the 10 ML measurements we no-
tice a strong difference. At high Fe thickness a distinct hys-
teresis loop is observed, which is a clear signature of
ferromagnetic behavior. Decreasing the thickness of the
Fe deposit, the hysteresis loop is smaller and is superim-
posed on a steep background. This behavior can be under-
stood in the light of the structural characterization of the
Fe/NiO interface [6,15]. In fact, the Fe film in the initial
stage of the growth is strained on the NiO substrate and
about one monolayer is oxidized. The relaxation of the me-
tal film starts at 10 ML, indicating a completely formed
0
50
100
150
200
250
300
350
Angle (deg.)
Fig. 2. Polarization dependence for the reflected beam measured on the
0 ML Fe sample in longitudinal configuration. The two curves were
acquire for the positive (squares) and negative (circles) magnetization,
respectively. The continuous lines are cos(h) fitting curves of the
experimental points.
1
2

S. Colonna et al. / Surface Science 601 (2007) 4362–4365
4365
4. Conclusions
In this paper we have presented non-linear optical mea-
surements of the Fe/NiO (100) interface. These results fea-
tured the detection of non-linear Kerr rotation of the
second harmonic beam as a function of the applied mag-
netic field. To our knowledge this is the first time that the
SHG technique has been applied to the study of the very
interesting Fe/NiO interface. This technique has proven
to be extremely sensitive to the magnetic properties of the
buried interface. The reported measurements demonstrate
that the magnetic properties of the studied interface
strongly depend on the Fe thickness. In fact, a clear hyster-
esis loop is observed only for a thickness higher than 5 ML.
At 2 ML the hysteresis loop completely disappears. These
results are in agreements with previous structural charac-
terizations of the interface.
0
50
100
150
200
250
300
350
Acknowledgements
Angle (deg.)
The stay of S.C. at the Vanderbilt University was par-
tially funded by the CNR Short Term Mobility program.
We are grateful to Y. Glinka for stimulating discussions.
Fig. 4. Polarization dependence of the SHG measured on the 2 ML Fe
sample in longitudinal configuration. The two curves were acquired for the
positive (squares) and negative (circles) magnetization, respectively The
2
continuous lines are cos(h) fitting curves of the experimental points.
References
[
[
1] W.H. Meiklejohn, C.P. Bean, Phys. Rev. 102 (1956) 1413.
2] R. de Masi, D. Reinicke, F. M u¨ ller, P. Steiner, S. H u¨ fner, Surf. Sci.
metallic film. Below this thickness probably the influence of
the oxidized Fe layer at the interface is strong thus resulting
in the observed magnetization curve for the 5 ML sample.
This picture is confirmed by the measurements at 2 ML re-
ported in Fig. 4. In this case an almost undetectable influ-
ence of the magnetic field on the SHG polarization
dependence is observed. Clearly, at this low thickness the
metal film is not ferromagnetic any more. It is arguable
that the FeO-like layer formed at the interface presents
an AFM character. On the other hand, the formation of
superparamagnetic islands, which would depress the ferro-
magnetic character of the film at room temperature, ap-
pears unlikely in this case because the structural
characterization of the 2 ML film evidenced the formation
of an epitaxial growth [15].
Interestingly in the 5 ML measurements of the SHG
polarization dependence (Fig. 1), the magnetic field not
only affects the polarization angle but also the intensity
of the SHG. As pointed out in Ref. [13] in the longitudinal
configuration of the magnetic field one should have only a
Kerr rotation of the second harmonic beam. A modula-
tion of the intensity is observed in the transversal configu-
ration. A further careful analysis of this aspect is in
progress.
5
15 (2002) 523.
[
3] S. Benedetti, P. Luches, M. Liberati, S. Valeri, Surf. Sci. 572 (2004)
L348.
[4] P. Luches, M. Liberati, S. Valeri, Surf. Sci. 532–535 (2003) 409.
[
5] T.J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. L u¨ ning, J. St o¨ hr,
R.L. White, Phys. Rev. B 64 (2001) 214422.
[6] P. Luches, V. Bellini, S. Colonna, L. Di Giustino, F. Manghi, S.
Valeri, F. Boscherini, Phys. Rev. Lett. 96 (2006) 106106.
[7] G. L u¨ pke, Surf. Sci. Rep. 35 (1999) 75.
[8] J.F. McGilp, Prog. Surf. Sci. 49 (1995) 1.
[9] R. Pan, H.D. Wei, Y.R. Shen, Phys. Rev. B 39 (1989) 1229.
[
10] H.A. Wierenga, W. de Dong, M.W.J. Prins, Th. Rasing, R. Vollmer,
A. Kirilyuk, H. Schwabe, J. Kirschener, Phys. Rev. Lett. 74 (1995)
1
462.
[11] J. Reif, J.C. Zink, C.-M. Schneider, J. Kirschner, Phys. Rev. Lett. 67
(1991) 2879.
12] H.A. Wierenga, M.W.J. Prins, D.L. Abraham, Th. Rasing, Phys.
Rev. B 50 (1994) 1282.
13] Th. Rasing, M. Groot Koerkamp, B. Koopmans, Hv.d. Berg, J. Appl.
Phys. 79 (1996) 6181.
[14] L.C. Sampaio, A. Mougin, J. Ferr e´ , P. Georges, A. Brun, H. Bernas,
S. Poppe, T. Mewes, J. Fassbender, B. Hillebrands, Europhys. Lett.
[
[
6
3 (2003) 819.
[
[
15] S. Colonna, P. Luches, S. Valeri, F. Boscherini, Nucl. Instrum.
Methods Phys. Res. B 246 (2006) 131.
16] S. Colonna, J. Qi, P. Luches, Y. Xu, J.K. Miller, A. Cricenti, S.
Valeri, N. Tolk, in preparation.