Direct observation of a propagating spin wave
induced by spin-transfer torque
M. Madami1†*, S. Bonetti2†*, G. Consolo3,4, S. Tacchi1, G. Carlotti1,5, G. Gubbiotti1,6, F. B. Mancoff7,
M. A. Yar8 and J. Åkerman2,9
Spin torque oscillators with nanoscale electrical contacts1–4 are
able to produce coherent spin waves in extended magnetic
films, and offer an attractive combination of electrical and
magnetic field control, broadband operation5,6, fast spin-wave
frequency modulation7–9, and the possibility of synchronizing
multiple spin-wave injection sites10,11. However, many potential
applications rely on propagating (as opposed to localized) spin
waves, and direct evidence for propagation has been lacking.
Here, we directly observe a propagating spin wave launched
from a spin torque oscillator with a nanoscale electrical
contact into an extended Permalloy (nickel iron) film through
the spin transfer torque effect. The data, obtained by wave-
vector-resolved micro-focused Brillouin light scattering, show
that spin waves with tunable frequencies can propagate for
several micrometres. Micromagnetic simulations provide the
theoretical support to quantitatively reproduce the results.
Much effort has recently been devoted to a better understanding
of the details of the spin waves excited in magnetic films by nano-
contact-based spin torque oscillators (STOs)12,13. In particular, it
has been predicted that the spatial characteristics of spin-wave exci-
tations have a critical dependence on the direction of the magnetiza-
tion angle in the out-of-the plane film direction (and therefore on the
externally applied field angle)14–16. Only very recently has it been
demonstrated experimentally (by means of electrical microwave
detection) that above a certain critical angle a propagating spin-
wave mode can be excited, and both localized and propagating spin
waves can be excited alternately below this critical angle17.
Although it was possible to elucidate a number of important charac-
teristics of both propagating and localized spin-wave modes using
only electrical detection (for example, the current and field depen-
dencies of the frequency, the linewidth and the output power),
direct evidence of their propagating nature is still lacking.
Micro-focused Brillouin light scattering (m-BLS)18 is a powerful
technique for resolving the spatial profile of spin waves in magnetic
nanostructures, and has recently been used in pioneering studies19,20
to experimentally observe spin waves caused by spin transfer torque
(STT) in in-plane magnetized nanocontact STOs. However, the pro-
pagating character of the radiated spin waves has not been
proven experimentally.
Here, we usem-BLS to study spinwaves emitted in an out-of-plane
magnetized nanocontact STO and provide experimental proof that
propagating spin waves are radially emitted from the nanocontact
region into the continuous ferromagnetic thin film up to several
micrometres away from the nanocontact.
The sample under investigation comprises a pseudo spin valve
stack with the layer structure Co81Fe19(20 nm)/Cu(6 nm)/Ni80
Fe20(4.5 nm), patterned into an 8× 26 mm2 mesa. The thicker CoFe
layer is considered the ‘fixed’ magnetic layer, and the thinner NiFe
plays the role of the low-dissipation ‘free’ magnetic layer in which a
steady STT-driven spin wave can be sustained. The thickness of the
copper spacer (6 nm) ensures that there is negligible interlayer
exchange coupling between the two magnetic layers. A circular
contact of diameter d¼ 200 nm is patterned in the middle of the
spin valve mesa, and a thick (400 nm) aluminium ground–signal–
ground waveguide is deposited on top of the mesa, allowing for
the injection of a high, spin-polarized, current density21 and the
subsequent extraction of the generated microwave voltage.
Optical m-BLS access to the active region of the free layer was
achieved through a combination of focused ion-beam (FIB) and
Hext
Cu spacer
Ni80Fe20 free layer
Co81Fe19 fixed layer
Optical window
Pd-Cu
top electrode
Pd-Cu
bottom electrode
Al coplanar
waveguide
SiO2
insulator
d.c. source
Probing
laser light
−+
+ /
Nanocontact
Figure 1 | Schematic sample layout. Cross-section of the sample, revealing
the layers of the spin valve mesa and the active area of the STO device. An
aluminium coplanar waveguide is deposited onto the spin valve mesa, and
an optical window is etched into the central conductor of the waveguide
close to the nanocontact.
1CNISM, Unita` di Perugia and Dipartimento di Fisica, Universita` di Perugia, Via A. Pascoli, I-06123 Perugia, Italy, 2Materials Physics, School of Information
Communication Technology, KTH – Royal Institute of Technology, Electrum 229, 164 40, Kista, Sweden, 3Dipartimento di Scienze per l’Ingegneria
e l’Architettura, Universita` di Messina C.da di Dio, I-98166 Messina, Italy, 4CNISM, Unita` di Ferrara, Via G. Saragat 1, I-44100 Ferrara, Italy, 5Centro S3,
CNR-Istituto di Nanoscienze, Via Campi 213A, I-41125 Modena, Italy, 6Istituto Officina dei Materiali del CNR (CNR-IOM), Unita` di Perugia,
c/o Dipartimento di Fisica, Via A. Pascoli, I-06123 Perugia, Italy, 7Everspin Technologies, Inc., 1347 N. Alma School Road, Suite 220, Chandler, Arizona
85224, USA, 8Functional Materials Division, School of Information Communication Technology, KTH – Royal Institute of Technology, Electrum 229, 164 40,
Kista, Sweden, 9Department of Physics, University of Gothenburg, 412 96 Gothenburg, Sweden; †These authors contributed equally to this work.
*e-mail: marco.madami@fisica.unipg.it; bonetti@kth.se
LETTERS
PUBLISHED ONLINE: 28 AUGUST 2011 | DOI: 10.1038/NNANO.2011.140
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 1
© 2011 Macmillan Publishers Limited. All rights reserved.

selective wet-etching processes to open an optical window (dimen-
sions, 4× 6 mm2) in the aluminium waveguide at a distance of
500 nm from the nanocontact (Fig. 1). An example of such a
modified sample is shown in Fig. 2a. To validate the effectiveness
of this process we performed two separate checks, including
energy dispersive spectroscopy (EDS) and m-BLS measurements
of thermal spin waves in two regions of the sample. The entire
procedure is described in the Methods, and the results prove that
we are able to access a portion of the free magnetic layer close to
the nanocontact region.
In our first set of measurements, we focused the laser spot in the
middle of the optical window. We detected a strong spin-wave signal
only for one current polarity, which corresponds to electrons drift-
ing from the free to the fixed magnetic layer. No signal was detected
for the opposite current polarity (shown in the inset of Fig. 3).
Remarkably, not only was the intensity of the emitted spin wave
much larger than that of the thermal spin waves, but the frequency
f of the excited mode was also considerably higher than that of the
ferromagnetic resonance (FMR) mode. This is consistent with the
expected blueshift of propagating spin waves excited by STT in a
nanocontact geometry when the free layer is magnetized out of
the film plane14,15. Figure 3 shows f as a function of the injected
direct current I for two different values of the perpendicular field
m0Hext. In both cases, f exhibits an almost linear increase with I,
resulting in current and field tunabilities consistent with all-electri-
cal results obtained in previous works on the same samples6,21.
These behaviours have been reproduced by micromagnetic simu-
lations, the results of which are also shown in Fig. 3. Simulations
also reveal that, as I is increased, the wavelength of the excited
spin waves decreases from 300 nm at 40 mA to 200 nm at
80 mA (Fig. 5, inset).
As a second step of the dynamic characterization of our sample,
we demonstrated that the emitted spin waves have indeed a propa-
gating character. To this aim, we performed wave-vector-resolved
m-BLS measurements by using the procedure described in ref. 22
and illustrated in Fig. 4. Because the detected spin wave is excited
by STT within the nanocontact region and propagates away from
it, its wave vector (KSW) has a well-defined direction in the plane
of the free layer. As a consequence, photons that undergo a Stokes
(anti-Stokes) process (in other words that create (destroy) a spin-
wave quantum, or magnon) will be scattered at opposite angles
with respect to the sample normal. This is analogous to the positive
or negative Doppler shift, which affects light beams diffracted in
opposite directions, after interaction with propagating acoustic
waves in Raman-Nath acousto-optic modulators23. To understand
whether such a process occurs in our system, we use a beam
shutter by setting, alternately, its full or half aperture towards the
interferometer. When the full beam is sent to the interferometer
(Fig. 4, lower spectrum), two spin-wave peaks are present, one on
the Stokes side of the spectrum and the other on the anti-Stokes
side. Alternatively, by selecting half of the beam, one of the two
peaks disappears, depending on which half of the beam is selected
(Fig. 4, middle and upper spectra). This result clearly demonstrates
that the emitted mode propagates with a uniquely defined wave-
vector direction. In fact, in the presence of a stationary or localized
(not propagating) wave, counts are expected on both the Stokes and
anti-Stokes regions of the spectrum, whatever the selected half of the
10 μm
r
Nanocontact
Optical
window
O Si Cu Co Pd Ni Fe Al
Element
Re
la
tiv
e
at
om
ic
c
on
te
nt
(
%
)
In
te
ns
ity
(
a.
u.
)
Outside
Inside
Outside
Inside
Frequency (GHz)
−12 −10 −8 −6 −4
35
30
25
20
15
10
5
0
a
b
Figure 2 | Characterization of the optical window. a, Scanning electron
microscope image of a processed device, showing the optical window in the
central conductor of the aluminium waveguide, the nanocontact approximate
position and the line (dotted arrow) across which the m-BLS laser spot was
scanned. b, EDS data acquired in regions outside and inside the etched
window (indicated by dashed and solid squares in a, respectively). Inset:
experimental m-BLS spectra (Stokes side), measured outside and inside the
etched optical window in the absence of any injected current and within an
applied perpendicular field of 2.0 kOe, revealing the presence of thermal
spin waves.
20
10
0
FMR
−20 −15 −10
I = −70 mA
I = +70 mA
−5
Frequency (GHz)
In
te
ns
ity
(
a.
u.
)
μ0Hext = 0.6 T
20
FMR
10
0
40
μ0Hext = 0.7 T
50
Modulus of the applied d.c. |I| (mA)
Sp
in
-w
av
e
fr
eq
ue
nc
y,
f

(G
H
z)
Sp
in
-w
av
e
fr
eq
ue
nc
y,
f

(G
H
z)
60 70 80 90 100
40 50
Modulus of the applied d.c. |I| (mA)
60 70 80 90 100
Figure 3 | Spin-wave frequencies as a function of the injected d.c.
intensity. Measured (filled symbols) and simulated (open symbols) spin-
wave frequency dependence on d.c. intensity for two different values of the
magnetic field. Dashed lines represent the calculated FMR frequency.
Inset: m-BLS spectra (Stokes side) recorded at m0Hext¼0.6 Tand for
different signs of the current |I|¼ 70 mA.
LETTERS NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.140
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology2
© 2011 Macmillan Publishers Limited. All rights reserved.

collected light beam. Note, however, that in the present case of
m-BLS, one also has to carefully consider the effect of uncertainty
in the wave-vector conservation due to the sub-micrometric
dimension of the illuminated area. This is quantitatively discussed
in the Methods.
At the final stage of our study, evolution of the emitted spin-wave
intensity as a function of distance from the nanocontact (r) was
measured by performing a scan of the laser spot across the open
optical window (dotted arrow in Fig. 2a). At a constant current of
I¼ 70 mA and a bias magnetic field of m0Hext¼ 0.6 T, we measured
a detectable spin-wave signal in the entire optical window, that is, up
to a maximum distance from the contact of 4 mm. The results are
shown in Fig. 5, where a marked reduction of the spin-wave inten-
sity (J(r)) with distance from the nanocontact (r) is clearly observed.
This decreasing behaviour can be accounted for if one considers
both the cylindrical symmetry of the emitted spin wave and the
finite propagation length in the free layer due to the intrinsic
damping of Permalloy. The former can be easily described by a
1/r dependence, and the latter can be accounted for by an exponen-
tial term exp(–r/lr) (ref. 19). Here, lr¼ vg/(2av) represents
the decay length related to the spin-wave group velocity (vg), a is
the Gilbert damping parameter and v¼ 2pf is the spin-wave
angular frequency ( f¼ 15.3 GHz). By representing the spin-wave
intensity as J(r)¼ ((J0/r)e2r/lr), in the range of values correspond-
ing to the open optical window, and by extracting the characteristic
parameters (vg¼ 3.3 mm ns21, a¼ 0.008) from the micromagnetic
simulations already exploited to fit the frequencies in Fig. 3, we
obtained a very good agreement with the experimental data. The
decay length value lr¼ 2.1 mm thus obtained is very similar to
literature values for Permalloy24.
We have provided experimental evidence that spin waves emitted
from an out-of-plane magnetized nanocontact STO into a continu-
ous NiFe film propagate unidirectionally several micrometres away
from the nanocontact. Our findings show that STOs represent a
very attractive nanoscale current- and field-controlled broadband
spin-wave generator for future use in magnonic devices. This
study is also relevant for the ongoing attempts to synchronize
large arrays of spin torque oscillators by means of the emission of
propagating spin waves10,11,25, with the aim of achieving the
minimum emitted power required for moving these devices out of
the laboratory and into actual microwave applications.
Methods
Experimental. Giant magnetoresistance (GMR) films were sputter-deposited in
ultrahigh vacuum on silicon wafers coated with SiO2. The GMR film stack consisted
of a seed layer of 5 nm palladium, a 25 nm copper bottom electrode, a 20 nm
Co81Fe19 fixed layer, a 6 nm copper spacer layer, a 4.5 nm Ni80Fe20 free layer, a 2 nm
copper top electrode and a 3.5 nm palladium cap. The GMR films were patterned on
an 8× 26 mm2 mesa using optical lithography and then coated with a SiO2 interlayer
dielectric deposited by chemical vapour deposition. The point contact area was
defined using electron-beam lithography and reactive ion etching through the SiO2.
Finally, a 400 nm aluminium top electrode was patterned by optical lithography,
sputter deposition and lift-off.
For FIB processing of the optical window, we used a dual-beam FEI Quanta-3D
field-emission gun (FEG) FIB-scanning electron microscope system, where both
electron (5 kV) and Gaþ (30 kV, 50 pA) beams were operated simultaneously to
monitor the etching process in real time. EDS was used to compare the chemical
composition of two regions of the sample: inside the optical window (black square in
Fig. 2a) and between the signal and ground lines (dashed square in Fig. 2a). The results
of this analysis, shown in Fig. 2b, demonstrate that aluminium has been completely
removed, but the SiO2 insulating layer (transparent to visible light) is still present. We
also performed m-BLS (ref. 26) measurements of thermal spin waves in the same two
regions of the sample. Results are shown in the inset in Fig. 2b. We detected thermal
spin waves with the same f and virtually the same amplitude both inside the optical
windowand between the signal and ground aluminium lines. Once all the preliminary
analyseswere performed, a projected field electromagnetwas placed below the sample,
in close proximity to it, with the aim of providing a tunable source of a perpendicular-
to-plane bias magnetic field (m0Hext) up to 0.7 T. In addition, by means of a low-noise
electric d.c. source, an electric direct current (I ) was allowed to flow through the
nanocontact. To stabilize the sample properties against irreversible heating effects
during the set of measurements, these were performed, for each value of the applied
field, by decreasing the applied current from an initial higher value (80 mA).
Another important issue is the uncertainty in the in-plane component of the
wave vector of scattered photons resulting from the limited spatial extent of the laser
spot on the sample. In a previous work27 we estimated the full-width at half-
maximum of our laser spot to be 235 nm. Using this value as DX in the wave
uncertainty relation DX.DK≈ 2p, one finds the uncertainty in the values of the
in-plane component of the photon wave vector involved in the scattering process to
be DK≈ 27× 104 cm21. This uncertainty is almost equal to the modulus of the
spin-wave wave vector KSW¼ 2p/lSW because, from the simulated wavelength
(Fig. 5, inset) it varies in the range 22× 104 to 31× 104 cm21. It follows that there is
a spread in the scattering angle: u¼ arctan[(KSW+DK )/Klight], where Klight≈ 12×
104 cm21 (the scattering angle, measured against the normal to the sample, is
indicated in Fig. 4). For instance, at I¼ 70 mA, u varies from 08 to 758. In the
Stokes
process
Anti-Stokes
process
θ
−20
Frequency shift (GHz)
−10 0 10 20
KL
KSW
a
b
Figure 4 | Proof of spin-wave propagation. a, Schematic of the experimental
procedure used to prove the propagating character of the detected spin
wave. KL and KSW represent the wave vectors of the incoming light and of
the emitted spin wave, respectively. b, Measured m-BLS spectra (I¼ 70 mA
and m0H¼0.6 T) corresponding to the case of fully opened (bottom
spectrum) and partially closed (upper spectra) collected beam.
2.0
5
10
15
20
25
2.5
In
te
gr
at
ed
in
te
ns
ity
, J
(r
) (
a.
u.
)
3.0 3.5
Distance from the contact r (μm)
40 60
300
200
I (mA)
λ
SW
(
nm
)
80
μ-BLS measurements
Calculated intensity
4.0
Figure 5 | Spin-wave attenuation as a function of distance from the STO.
Integrated intensity (symbols) of the spin-wave excitations detected using
m-BLS as a function of distance from the centre of the point contact (r).
Analytical calculation (line) of the decay obtained using the function
described in the text. Inset: simulated spin-wave wavelength as a function
of applied d.c. intensity.
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.140 LETTERS
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 3
© 2011 Macmillan Publishers Limited. All rights reserved.

presence of a collecting angle of our objective of about+508 (NA¼ 0.75), this
demonstrates the feasibility of our wave-vector selection method (illustrated in
Fig. 4), because the uncertainty does not cause the Stokes and anti-Stokes cones
of emission to overlap spatially.
Using the same argument, one finds that it is important to use a relatively large
contact size because our m-BLS setup can detect spin waves, with sizable intensity,
down to wavelengths of 200 nm.
Numerical. Micromagnetic simulations were performed using a three-dimensional
code that integrates the Landau–Lifshitz–Gilbert–Slonczewski equation of motion by
means of a finite-difference time-domain (FDTD) approach15,28,29. The effective field
includes all the main standard micromagnetic contributions arising from
demagnetizing, exchange, Zeeman and Oersted fields. We neglect thermal
fluctuations (which only affect the linewidths of the output signal) and
magnetocrystalline anisotropy (as usual for Permalloy-based materials). A uniformly
distributed current density within the current-carrying region, which exhibits
an abrupt cutoff outside that area, has been considered. We also assume that the
spin-transfer torque perturbation acts only on the thinner free layer. Because of
obvious constraints on computational time and memory allocation, a reduced
computational region with sides of 1 mm and with the nanocontact in the middle
was simulated. To reduce the spurious effect of spin-wave reflection from the
computational boundaries, absorbing boundary conditions have been
implemented28. The computational domain was discretized in prismatic cells of
4× 4× 5 nm3. The material parameters used in our setup are: a saturation
magnetization m0MS¼ 0.5 T, Gilbert damping constant a¼ 0.008, spin-torque
efficiency¼ 0.25, exchange constant A¼ 1.0× 10211 J m21 and nanocontact radius
(RC¼ 120 nm). We believe that the slightly reduced values of the exchange constant
and saturation magnetization with respect to their nominal values are a consequence
of the limited NiFe thickness and possibly also due to the effect of local heating in the
area of the nanocontact, leading to local oxidation and copper interdiffusion. Similar
values were recently used in micromagnetic simulations of spin torque excitations in
thin NiFe nanowires30. The increased value of the nanocontact radius takes into
account the local current spreading inside the extended free layer20. The magnetic
field implemented in the simulations was allowed to be 10% higher than the nominal
experimental value and applied at 858 with respect to the sample plane, reflecting the
uncertainty in the intensity and direction of the field projected by the magnet.
Received 20 June 2011; accepted 22 July 2011;
published online 28 August 2011
References
1. Slonczewski, J. C. Current-driven excitation of magnetic multilayers. J. Magn.
Magn. Mater. 159, L1–L7 (1996).
2. Berger, L. Emission of spin waves by a magnetic multilayer traversed by a
current. Phys. Rev. B 54, 9353–9358 (1996).
3. Tsoi, M. et al. Generation and detection of phase-coherent current-driven
magnons in magnetic multilayers. Nature 406, 46–48 (2000).
4. Silva, T. J. & Rippard, W. H. Developments in nano-oscillators based upon
spin-transfer point-contact devices. J. Magn. Magn. Mater. 320,
1260–1271 (2008).
5. Rippard, W. H., Pufall, M. R., Kaka, S., Silva, T. J. & Russek, S. E. Current-driven
microwave dynamics in magnetic point contacts as a function of applied field
angle. Phys. Rev. B 70, 100406(R) (2004).
6. Bonetti, S., Muduli, P. K., Mancoff, F. B. & Åkerman, J. Spin torque oscillator
frequency versus magnetic field angle: the prospect of operation beyond 65 GHz.
Appl. Phys. Lett. 94, 102507 (2009).
7. Pufall, M. R., Rippard, W. H., Kaka, S., Silva, T. J. & Russek, S. E.
Frequency modulation of spin-transfer oscillators. Appl. Phys. Lett. 86,
082506 (2005).
8. Muduli, P. K. et al. Nonlinear frequency and amplitude modulation of a
nanocontact-based spin-torque oscillator. Phys. Rev. B 81, 140408(R) (2010).
9. Pogoryelov, Ye. et al. Frequency modulation of spin torque oscillator pairs.
Appl. Phys. Lett. 98, 192501 (2011).
10. Kaka, S. et al. Mutual phase-locking of microwave spin torque nano-oscillators.
Nature 437, 389–392 (2005).
11. Mancoff, F. B., Rizzo, N. D., Engel, B. N. & Tehrani, S. Phase-locking in double-
point-contact spin-transfer devices. Nature 437, 393–395 (2005).
12. Slonczewski, J. C. Excitation of spin waves by an electric current. J. Magn. Magn.
Mater. 195, 261–268 (1999).
13. Slavin, A. N. & Tiberkevich, V. S. Spin-wave mode excited by spin-polarized
current in a magnetic nanocontact is a standing self-localized wave bullet.
Phys. Rev. Lett. 95, 237201 (2005).
14. Gerhart, G., Bankowski, E., Melkov, G. A., Tiberkevich, V. S. & Slavin, A. N.
Angular dependence of the microwave-generation threshold in a nanoscale
spin-torque oscillator. Phys. Rev. B 76, 024437 (2007).
15. Consolo, G. et al. Micromagnetic study of the above-threshold generation regime
in a spin-torque oscillator based on a magnetic nanocontact magnetized at an
arbitrary angle. Phys. Rev. B 78, 014420 (2008).
16. Hoefer, M. A., Silva, T. J. & Stiles, M. D. Model for a collimated spin-wave
beam generated bya single-layer spin torque nanocontact. Phys. Rev. B 77,
144401 (2008).
17. Bonetti, S. et al. Experimental evidence of self-localized and propagating
spinwave modes in obliquely magnetized current-driven nanocontacts.
Phys. Rev. Lett. 105, 217204 (2010).
18. Demidov, V. E., Demokritov, S. O., Hillebrands, B., Laufenberg, M. & Freitas, P.
Radiation of spin waves by a single micrometer-sized magnetic element.
Appl. Phys. Lett. 85, 2866–2868 (2004).
19. Demidov, V. E., Urazhdin, S. & Demokritov, S. O. Direct observation and
mapping of spin waves emitted by spin-torque nano-oscillators. Nature Mater.
9, 984–988 (2010).
20. Demidov, V. E., Urazhdin, S., Tiberkevich, V. S., Slavin, A. N. &
Demokritov, S. O. Control of spin-wave emission from spin-torque nano-
oscillators by microwave pumping. Phys. Rev. B 83, 060406(R) (2011).
21. Mancoff, F. B., Rizzo, N. D., Engel, B. N. & Tehrani, S. Area dependence of
high-frequency spin-transfer resonance in giant magnetoresistance contacts up
to 300 nm diameter. Appl. Phys. Lett. 88, 112507 (2006).
22. Neumann, T., Schneider, T., Serga, A. A. & Hillebrands, B. An electro-optic
modulator-assisted wavevector-resolving Brillouin light scattering setup.
Rev. Sci. Instrum. 80, 053905 (2009).
23. Raman, C. V. & Nath, N. S. N. The diffraction of light by high frequency sound
waves: Part III. Proc. Indian Acad. Sci. 3, 75–84 (1936).
24. Demidov, V. E. et al. Self-focusing of spin waves in Permalloy microstripes.
Appl. Phys. Lett. 91, 252504 (2007).
25. Slavin, A. N. & Tiberkevich, V. S. Theory of mutual phase locking of spin-torque
nanosized oscillators. Phys. Rev. B 74, 104401 (2006).
26. Madami, M. et al. Experimental evidence of field-induced localization of spin
excitations in NiFe elliptical rings by micro-focused Brillouin light scattering.
IEEE Trans. Magn. 46, 1531–1536 (2010).
27. Gubbiotti, G. et al. Setup of a new Brillouin light scattering apparatus with
submicrometric lateral resolution and its application to the study of spin modes
in nanomagnets. J. Appl. Phys. 105, 07D521 (2009).
28. Consolo, G. et al. Excitation of self-localized spin-wave bullets by spin-polarized
current in in-plane magnetized magnetic nanocontacts: a micromagnetic study.
Phys. Rev. B 76, 144410 (2007).
29. Romeo, A. et al. A numerical solution of the magnetization reversal modeling in
a Permalloy thin film using fifth order Runge–Kutta method with adaptive step
size control. Physica B 403, 464–468 (2008).
30. Berkov, D. V., Boone, C. T. & Krivorotov, I. N. Micromagnetic simulations of
magnetization dynamics in a nanowire induced by a spin-polarized current
injected via a point contact. Phys. Rev. B 83, 054420 (2011).
Acknowledgements
This work was supported by CNISM under the m-BLS INNESCO project. Authors
acknowledge the European Community’s Seventh Framework Programme (FP7/2007-
2013, grant agreement no. 228673, MAGNONICS). Support from the Swedish Foundation
for Strategic Research (SSF), the Swedish Research Council (VR) and the Knut and Alice
Wallenberg Foundation is gratefully acknowledged. J.Å. is a Royal Swedish Academy of
Sciences Research Fellow supported by a grant from the Knut and Alice Wallenberg
Foundation. The authors gratefully acknowledge S. Redjai Sani at the Royal Institute of
Technology for help with the wet etching process, F. Magnusson and W. Michelsen at
NanOsc AB for their help in designing the printed circuit boards, and S. Gunnarsson,
S. Sandelin and K. Penkkila¨ at Sivers IMA AB for performing the wire bonding.
S.B. gratefully acknowledges support from the C.M. Lerici foundation.
Author contributions
M.M., G.G., S.T. and G.Ca. performed m-BLS measurements. S.B., M.A.Y. and J.Å. realized
the procedure to open the optical access to the sample and performed EDS measurements.
F.B.M. fabricated the original samples. G.Co. performed numerical simulations. All authors
co-wrote the manuscript.
Additional information
The authors declare no competing financial interests. Reprints and permission information is
available online at http://www.nature.com/reprints. Correspondence and requests for materials
should be addressed to M.M. and S.B.
LETTERS NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.140
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology4
© 2011 Macmillan Publishers Limited. All rights reserved.