PDF hosted at the Radboud Repository of the Radboud University
Nijmegen
The following full text is a publisher's version.
For additional information about this publication click this link.
http://hdl.handle.net/2066/29749
Please be advised that this information was generated on 2022-05-31 and may be subject to
change.
A combined nonlinear and linear magneto-optical microscopy
V. Kirilyuk, A. Kirilyuk, and Th. Rasinga)
Research Institute for Materials, University of Nijmegen, Toernooiveld 1, 6525 ED Nijmegen,
the Netherlands
(Received 5 December 1996; accepted for publication 25 February 1997)
New possibilities for magnetic domain studies are demonstrated using a combination of nonlinear
magneto-optical microscopy and a conventional linear polarizing microscope. The use of an optical
response that is governed by a higher rank tensor offers sensitivity to additional combinations of
magnetization directions and optical wave vector and polarization, which is demonstrated in
magnetic garnet films of different crystallographic orientations. We observed a nontrivial modulated
domain structure in a (210) film and a clear domain contrast for a (111) film, where the linear image
only indicated simple up-down domains and no domain contrast for these two situations,
respectively. © 1997 American Institute of Physics. [S0003-6951(97)01717-8]
Recently, magnetization induced optical second har
monic generation (MSHG) has been shown to be a very sen
sitive tool to probe magnetic surfaces and interfaces.1-3 This
nonlinear optical technique has great capabilities in studying
magnetic properties of antiferromagnets,4,5 stratified metal
structures and burried interfaces.6 Because the symmetry
properties of the nonlinear interactions differ essentially
from those in linear optics it would be interesting to visualize
magnetic domain patterns by nonlinear magneto-optical mi
croscopy, particularly for situations where the surface/
interface magnetization is (expected to be) different from the
bulk.
In this letter, nonlinear magneto-optical microscopy and
its application to the visualization of domain structures in
epitaxially grown magnetic garnet films is demonstrated. The
proposed method allows images to be obtained both from the
SH and linear response. This combination appears to be very
powerful because the nonlinear and linear images contain
complementary magnetic information. As all our experi
ments were done in transmission at normal incidence, only
the magnetization component perpendicular to the film sur
face could be directly visualized with the fundamental light
(via the linear magneto-optical Faraday effect), whereas in
plane components were probed by the second harmonic. In
the nonlinear microscope the MSHG response was imaged
using the very same setup, after filtering out the fundamental
light. The linear polarization of the incoming light was ro
tated between 0° and 180° with respect to the sample sym
metry plane m . In this way, subdomains that have different
in-plane components, were distinguished. The observed
MSHG contrast was correlated with the in-plane magnetiza
tion component using recent studies of MSHG rotational
anisotropy.7,8
The experimental set-up of our nonlinear magneto
optical microscope is schematically presented in Fig. 1. As a
light source we used a Ti:sapphire laser operating at a rep
etition rate of 82 MHz with a pulse width of about 100 fs and
at the wavelength of 775 nm. A half-wavelength plate was
used to rotate the linear polarization of the incoming light.
The laser beam was focused on the sample onto a spot of
about 70 ¡xm diameter. The average power of the pump
a)Electronic mail: theoras@sci.kun.nl
beam on the sample was 100 mW, resulting in a peak power
of nearly 4 GW/cm2. We magnified the exposed area by a
X 40 (N.A.=0.65) objective in combination with an achro
matic concave lens. After appropriate filtering the generated
second harmonic intensity was imaged with a cooled charge
coupled device (CCD) camera. The subtraction of the
gaussian-like background was applied afterwards to remove
the spot-profile inhomogeneity in the image intensity due to
the pump beam.
As test structures for our nonlinear magneto-optical mi
croscope, differently oriented magnetic garnet films with
thicknesses around 10 ¡xm were probed. Garnet films are
interesting subjects because of the fact that their structural,
magnetic and magneto-optical properties may be widely var
ied by changing the film composition and the composition
and orientation of the substrate. Magnetic garnet films were
grown by a liquid phase epitaxial method, on (210) and (111)
gadolinium gallium garnet (GGG) substrates. These sub
strates are centrosymmetric, nonmagnetic and transparent at
the fundamental and SHG wavelengths. Because of a growth
anisotropy and a lattice mismatch between the substrate and
the magnetic film, the inversion symmetry in the thin garnet
films is broken.9 The point group symmetry of these mag
netic films results from the substrate orientation and is 3 m
(C3v) for (111) and m (C1h) in the case of (210).10 In these
noncentrosymmetric structures both crystallographic and
magnetization-induced bulk contributions to the nonlinear
polarization are allowed.7 Therefore they can interfere in a
magnetized sample or in a sample with a spontaneous mag
netic (domain) structure: I2ffl« lxcr±Xmagnl2, where ± de
pends on the direction of M and thus changes from domain
to domain. It is this interference term that will be responsible
for the magnetic contrast.
FIG. 1. Schematic microscopy setup.
2306 Appl. Phys. Lett. 70 (17), 28 April 1997 0003-6951/97/70(17)/2306/3/$10.00 © 1997 American Institute of Physics
Downloaded-16-Jun-2008-to-131.174.20.161.-Redistribution-subject-to-AIP-license-or-copyright;-see-http://apl.aip.org/apl/copyright.jsp
FIG. 2. Faraday hysteresis loops for (a) (210) oriented and (b) (111) ori
ented films.
Faraday hysteresis measurements (Fig. 2) show that both
films have a remanence magnetization, but only in the (111)-
film is the magnetization exactly perpendicular to the film
surface. In the (210)-sample the magnetization is tilted at an
angle of about 18° with respect to the film normal. This
value was derived from the difference between the saturation
and the remanence magnetization, taking into account that at
remanence the sample is in a single-domain state, which is
supported by direct imaging.
The domain pattern of the demagnetized films was ini
tially tested using our setup as a linear Faraday microscope.
Fig. 3(a) shows a typical labyrinth type domain structure for
the (210) film where the dark/light areas indicate ‘‘up’’ and
‘‘down’’ domains.
Next, second harmonic images of this very same domain
structure were taken, for various values of the incoming lin
ear polarization with respect to the crystal symmetry plane
m [Figs. 3(b)-(f)]. The SH images were recorded without
analysing the outgoing light polarization. Remarkable
changes in the magnetic contrast and in the SH intensity for
the (210) garnet film were thus found. At 0° the SHG do
main pattern appears to be exactly the same as in the linear
light. To follow all subsequent changes appearing in the
magnetic structure, the domain walls in this image are
marked with dashed lines [Fig. 3(b)]. Rotating the incoming
light polarization by 10°, a subdivision of the original do
mains was clearly observed [Fig. 3(c)]. This subdivision is
even more sharp at larger angles and at 35° the SH intensity
in the subdomains I and III becomes equal [Fig. 3(d)]. At
90° and at 145° the magnetic structure looks very similar to
the cases 3(b) and 3(d) respectively, but the contrast appears
to be shifted by half a domain width [Figs. 3(e),(f)]. To anal
yse the observed images, one should recall that in this con
figuration of normal incidence (and without polarization
analysis), MSHG can only probe in-plane magnetization
components, as follows from the MSHG selection rules.8
This means that different SH intensities correspond to differ
ent in-plane magnetizations. Therefore, from Fig. 3 we can
conclude that four domain types appear to exist in the (210)-
film (M/# M//# M///# M /v).
To analyse all recorded images (taken in steps of 5°) we
started from the rotational anisotropy measurements (see
Ref. 7), i.e. from the dependencies of the SH intensity on the
azimuthal position of the sample with respect to the incom
ing light polarization and fixed (in-plane) magnetization. By
FIG. 3. (a) Linear and (b)-(f) second harmonic images of the magnetic
domain structure in (210) oriented film. Input polarization was: (b) 0°; (c)
10°; (d) 35°; (e) 90°; (f) 145° with respect to the symmetry plane m.
a straightforward transformation, we obtain similar depen
dencies as a function of the incoming light polarization, with
the in-plane M fixed in the sample. It was then possible to
estimate the in-plane magnetization in every subdomain. In
total, an appropriate model of the (210) domain structure was
derived (Fig. 4) with every ‘‘up’’ or ‘‘down’’ domain subdi
vided into two subdomains with different in-plane compo
nents. The in-plane magnetizations of the neighbouring do
mains are not collinear and the absolute value of the in-plane
magnetization is the same in every subdomain. We should
note that the observed image appeared to exist through the
whole film, as was checked by changing the focusing depth,
and can thus not be related to closure domains at the surface.
FIG. 4. Magnetization directions in four different domains in (210) film, as
derived from the images of Fig. 3. Symmetry plane m coincides with the
side-face plane.
Appl. Phys. Lett., Vol. 70, No. 17, 28 April 1997 Kirilyuk, Kirilyuk, and Rasing 2307
Downloaded-16-Jun-2008-to-131.174.20.161.-Redistribution-subject-to-AIP-license-or-copyright;-see-http://apl.aip.org/apl/copyright.jsp
FIG. 5. Linear (a) and second harmonic (b) images of the magnetic domain
structure in (111) oriented film obtained with crossed input and output po
larizers.
Thus, a nontrivial modulated domain structure is shown
to exist in the (210) oriented garnet film, which becomes
only distinguishable using combined linear and nonlinear
magneto-optical microscopy.
Similar measurements on the (111) film showed no mag
netic contrast which means that no in-plane magnetization
exists in the (111) sample. An analyser was then used to
study the MSHG polarization rotation. To avoid an influence
of the linear polarization rotation effect, a carefully crossed
polarizer/analyzer configuration was used. Surprisingly, the
outcoming SH signal showed a strong contrast in this case
[Fig. 5(b)], though the linear picture showed no contrast be
tween the domains at all [Fig. 5(a)]. Hence, a different Far
aday rotation value for the SH light is found in the different
domains. Indeed, the resulting SH polarization is a vector
sum of crystallographic and magnetic contributions, the lat
ter having different sign for the two opposite domains. As
soon as the polarization of the crystallographic part is not
equal to that of the fundamental light (except for certain
high-symmetry directions), the final SH polarization states
are not symmetric with respect to the incoming light polar
ization (see insets in Fig. 5). This configuration clearly dem
onstrates the difference in the mechanism responsible for the
magneto-optical contrast in the two cases, and may be used
to make a correlation between the domain structure and crys
tallographic axes. The contrast disappears when the incom
ing light polarization coincides with one of the sample sym
metry planes.
In conclusion, we have demonstrated a new type of non
linear magneto-optical microscopy that, in combination with
standard linear microscopy, yields a wealth of additional and
complementary information about magnetic domain struc
tures. In particular, the higher rank tensor and a polarization
analysis of the incoming fundamental and/or second har
monic light gives information about the magnetic structure
that cannot be obtained in a single configuration with linear
microscopy. It can also be shown that magnetization gradient
near domain walls can give rise to additional contrast8. The
sensitivity of MSHG to the breaking of crystallographic in
version symmetry also gives new possibilities for the obser
vation of interface domain structures. This, together with the
already demonstrated sensitivity for antiferromagnetic
domains5, makes the further development of the nonlinear
magneto-optical microscopy very promising.
The authors thank R.V. Pisarev and A. Petukhov for
helpful discussions and A.F. van Etteger for technical assis
tance. Part of this work was supported by HCM institutional
Fellowship Nos. ERBCHBGCT930444 and ERB-
CHRXCT940563, INTAS 94-2675, and European network
ERBFMRXCT960015.
1 J. Reif, J. C. Zink, C. M. Schneider, and J. Kirschner, Phys. Rev. Lett. 67,
2878 (1991).
2H. A. Wierenga, W. de Jong, M. W. J. Prins, Th. Rasing, R. Vollmer, A.
Kirilyuk, H. Schwabe, and J. Kirschner, Phys. Rev. Lett. 74, 1462 (1995).
3B. Koopmans, M. Groot Koerkamp, Th. Rasing, and H. van den Berg,
Phys. Rev. Lett. 74, 3692 (1995).
4M. Fiebig, D. Fröhlich, B. B. Krichevtsov, and R. V. Pisarev, Phys. Rev.
Lett. 73, 2127 (1994).
5M. Fiebig, D. Frohlich, G. Sluytermann, and R. V. Pisarev, Appl. Phys.
Lett. 66, 2906 (1995).
6H. A. Wierenga, M. W. J. Prins, and Th. Rasing, Physica B 204, 281
(1995).
7R. V. Pisarev, V. V. Pavlov, A. Kirilyuk, and Th. Rasing, J. Magn. Soc.
Jpn. 20, 23 (1996).
8 A. V. Petukhov, I. L. Lyubchanskii, and Th. Rasing, J. Appl. Phys. (in
press).
9 B. B. Krichevtsov, V. V. Pavlov, and R. V. Pisarev, Sov. Phys. Solid State
31, 1142 (1989).
10R. V. Pisarev, B. B. Krichevtsov, V. N. Gridnev, V. P. Klin, D. Fröhlich,
and Ch. Pahlke-Lerch, J. Phys. C 5, 8621 (1993).
2308 Appl. Phys. Lett., Vol. 70, No. 17, 28 April 1997 Kirilyuk, Kirilyuk, and Rasing
Downloaded-16-Jun-2GG8-to-131.174.2G.161.-Redistribution-subject-to-AIP-license-or-copyright;-see-http://apl.aip.org/apl/copyright.jsp