Magnetic properties of cobalt nanowires: Study by polarized SANS

Article abstract of DOI:10.1016/j.physb.2010.10.081

Structural and magnetic properties of two-dimensional spatially ordered system of ferromagnetic cobalt nanowires embedded into Al₂O ₃ matrix have been studied using polarized small-angle neutron scattering. A comprehensive analysis of contributions to the scattering intensity was carried out, including nonmagnetic (nuclear) contribution, magnetic contribution depending on the magnetic field, and nuclearmagnetic interference indicating the correlation between the magnetic and nuclear structures. Experiments have revealed an anomalously low value of the magnetic contribution as compared to the nuclear one. This behavior is interpreted in terms of low coherence of the magnetic structure caused by the anisotropy of Co crystallites as compared with the large coherency of nuclear structure of nanowires.

Full text of DOI:10.1016/j.physb.2010.10.081

Physica B 406 (2011) 2405–2408
Contents lists available at ScienceDirect
Physica B
journal homepage: www.elsevier.com/locate/physb
Magnetic properties of cobalt nanowires: Study by polarized SANS
a,
n
a
b
c
c
c
A.P. Chumakov , S.V. Grigoriev , N.A. Grigoryeva , K.S. Napolskii , A.A. Eliseev , I.V. Roslyakov ,
a
d
A.I. Okorokov , H. Eckerlebe
a
Petersburg Nuclear Physics Institute, Gatchina, 188300 St. Petersburg, Russia
Faculty of Physics, St. Petersburg State University, 198504 St. Petersburg, Russia
High School of Materials Science, Moscow State University, 119899 Moscow, Russia
GKSS Forschungszentrum, 21502 Geesthacht, Germany
b
c
d
a r t i c l e i n f o
a b s t r a c t
Available online 3 November 2010
Structural and magnetic properties of two-dimensional spatially ordered system of ferromagnetic cobalt
2 3
nanowires embedded into Al O matrix have been studied using polarized small-angle neutron
Keywords:
scattering. A comprehensive analysis of contributions to the scattering intensity was carried out,
including nonmagnetic (nuclear) contribution, magnetic contribution depending on the magnetic field,
and nuclear–magnetic interference indicating the correlation between the magnetic and nuclear
structures. Experiments have revealed an anomalously low value of the magnetic contribution as
compared to the nuclear one. This behavior is interpreted in terms of low coherence of the magnetic
structure caused by the anisotropy of Co crystallites as compared with the large coherency of nuclear
structure of nanowires.
Anodic aluminium oxide
Magnetic nanowires
Polarized small-angle neutron diffraction
&
2010 Elsevier B.V. All rights reserved.
The investigations of the magnetic properties of ferromagnetic
nanoparticles are one of the urgent problems in modern science and
nanoindustry. Nanoparticles can have different shapes, such as
spherical, ellipsoidal, stick-like, and others. Having a relative simpli-
city of their production and technological instrumental integration,
cylindrical nanowires are the objects of the current researchers’
attention. The main problem of these studies is the magnetic behavior
of an individual magnetic particle and their array in different spatial
environment and circumstances (T, P, H). For example, the evolution
of the magnetization reversal mechanism of nanowires arrays with
different ratios of wire’s length, L, to its diameter, D, was thoroughly
studied [1]. The authors carried out measurements hysteresis loop of
magnetization at different orientations of the nanowires axis with
respect to the magnetic field. Quantitative analysis of the M–H loops
allowed them to plot the dependence of the coercitivity on the angle
of mutual orientation of field and nanowire’s axis. This angular
dependence allows one to distinguish between two different remag-
netization mechanisms, which are realised in single defect-free
ellipsoids: coherent reversal mode [2] or curling reversal mode [3].
It was found that the remagnetization of nanowires with relatively
large length (L/DZ1) is characterized by the curling mechanism,
while upon transformation of nanowire into nanodot (L/Do1) their
remagnetization is characterized by the coherent mechanism. How-
ever, it must be noticed that the above-mentioned reversal mechan-
isms are not applicable in case when the magnetic field is applied
perpendicularly to the principal wire’s axis. For the long nanowires,
a combination of both mechanisms (the coherent rotation and curling
mode) can be suggested for the explanation [1]; however, this
problem remains unsolved. The case of nanodots arrays in the
perpendicular field has been recently studied in detail [4]. It was
shown that the magnetic vertices appear in the range of small
magnetic field with a certain chirality and polarity. Therefore, it is
the curling mode that is realized in this case. In this article the
remagnetization mechanism of the magnetic nanowires at perpen-
dicular orientation of the field with respect to the wire axis is
investigated by polarized SANS, and based on the obtained data we
have concluded that the small magnetic coherent volume and
polycrystalline nature of the nanowires play the dominating role in
their reversal mechanism.
The subject of the study is the cobalt nanowires arrays
embedded into anodic alumina. Chosen technique of polarized
small-angle neutron scattering is sensitive to the structure of the
2 3
porous Al O matrix and to the magnetic properties of the
nanowire system. As was shown in Refs. [5,6], the small angle
diffraction on these objects should be described by the scattering
theory beyond the Born approximation. Among the advantages of
the method is the fact that information about structure is obtained
over the whole volume of the sample [7]. It is worthwhile to note
that the picture of the scattering from the ordered system of the
magnetic nanowires is not trivial because the complicated system
of the demagnetizing fields appeared in the sample upon magne-
tization process [8,9].
n
Preparation of spatially ordered arrays of cobalt nanowires was
carried out by metal electrocrystallization inside the channels of
Corresponding author.
E-mail address: chumakov@lns.pnpi.spb.ru (A.P. Chumakov).
0921-4526/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.physb.2010.10.081

2
406
A.P. Chumakov et al. / Physica B 406 (2011) 2405–2408
anodic alumina films with a thickness of 110
m
m. In order to obtain
A magnetization curve with applied magnetic field in the plane
better ordering of the channels, the anodic aluminium oxide
membranes were prepared by the two-step anodization technique
of the film (perpendicular to the nanowires axis) were recorded
using SQUID magnetometer Cryogenics S-700 at temperature
290 K. Fig. 2 presents the magnetization curve, saturated at
[
10,11]. The anode oxidation of aluminium is performed in a two-
electrode electrochemical cell in a 0.3 M (COOH) solution at a DC
2
m H¼700 mT. A small hysteresis loop is observed in the range
0
voltage of 40 V for 50 and 42.5 h in the first and second anodization
stages, respectively. The procedure of fabrication of electrodes on
AAO films is described in details elsewhere [8,12].
from ꢀ300 to 300 mT. Such form of remagnetization curve with the
hysteresis loop is typical for long nanowires at the magnetic field,
applied perpendicularly to their axis [1]. We believe that in our case
the c-axis of Co crystallites is randomly oriented. The loop in Fig. 2 is
not rectangular, similar to Ni nanowires, where the shape aniso-
tropy dominates and one cannot distinguish whether shape or
magnetocrystalline anisotropy is predominant. Furthermore we
can conclude that according to the SANS data in our cobalt array, a
magnetocrystalline anisotropy plays the major role.
Electrodeposition of cobalt was carried out in three-electrode
cell from 0.3 M CoSO
4
+0.05 M CoCl
2
+0.3M H
3
BO
3
solution at
deposition potential E
trode using Autolab PGSTAT 302 potentiostat. The Pt wire serves as
a counter electrode. The duration of the Co electrodeposition was
d
¼ ꢀ0.8 V versus Ag/AgCl reference elec-
3
h, which corresponds, according to scanning electron micro-
scopy, to the formation of nanowires with a length of 100 m and a
m
A typical neutron diffraction pattern measured from Co nanowire
array grown inside AAO membrane is shown in Fig. 3a. The first and
the second order reflections have been detected, suggesting the
presence of strong correlations in the positions of the pores (or
nanowires) over significant distances. The hexagonally arranged set of
reflections demonstrates the 6-fold symmetry of the structure.
The widening of reflections to 10–201 in an azimuthal direction,
shown in Fig. 3a, indicates that samples have a block structure, with
slight disorientation of domains in the plane of the film. Never-
theless, the diffraction patterns, typical for a quasi-single crystal,
were recorded using a neutron beam of 5 mm diameter, suggesting
correlated orientation of the domains on the irradiated area
diameter of about 40 nm. Fig. 1 presents the SEM pictures of a side
view of the AAO film with arrayed Co nanowires on a bottom part.
Polarized SANS measurements were carried out at the SANS-2
beamline of Geesthacht Neutron Facility (GeNF). Polarized neutron
beam with mean wavelength of
l
¼0.5 nm, wavelength spread of
¼0.95 was used. A film with
Dl ¼10%, and initial polarization P
/
l
0
2
an area of 0.3 cm was oriented perpendicularly to the neutron
beam and was uniformly irradiated over the entire area. Such an
orientation of the sample corresponds to parallel orientation of the
pores and the long axis of the magnetic nanowires to the incident
neutron beam. The chosen geometry of the experiment allows the
observation of diffraction reflections from the ordered system of
pores or magnetic wires in a small-angle scattering range. The
sample-to-detector distance of 14 m was used with appropriate
2
(ꢁ0.3 cm ). Similar results had been observed previously for pure
anodic aluminium oxide membranes by SANS techniques [5,6].
Fig. 3b illustrates the momentum-transfer dependence of the total
(nuclear and magnetic) neutron scattering I(Q) for the nonmagnetized
ꢀ
1
collimations to cover scattering vectors Q from 0.03 to 0.18 nm
.
An external magnetic field (up to 800 mT) was applied in the
horizontal plane perpendicular to the incident beam (and perpen-
dicular to the long axis of the nanowires). The SANS data were taken
at 0 mT before the remagnetization experiment. It had been
subtracted from SANS data taken under applied magnetic field.
The data were given at temperature 290 K. We determine the total
(
m
0
H¼1 mT) and magnetized (
tion maxima are clearly observed, indicating the highly ordered
structure of the Co/Al nanocomposite. The momentum transfer
dependence I(Q) is satisfactorily reproduced by the sum of the Gaussian
m
0
H¼800 mT) sample. Several diffrac-
2 3
O
ꢀ
1
components with the centers located at Q10¼0.06970.004 nm
,
ꢀ
1
ꢀ1
Q
11¼0.11970.004 nm , Q20¼0.13870.004 nm , and fixed half-
+
–
ꢀ1
(
nuclear and magnetic) scattering (I(Q)¼I (Q)+I (Q)) and the
width of w¼0.009070.0005 nm ; coupled with a diffuse small-
+
polarization dependent part of the scattering (
D
I(Q)¼I (Q)–
angle scattering, represented by an exponential decay. The observed
diffraction maxima are classified in the hexagonal lattice with the
parameter a¼10672 nm. It should be noted that applying the
external magnetic field induces noticeable increase in the diffraction
peak intensity in the order of 6% (see Fig. 3b).
–
+
–
I (Q)), where I (Q) and I (Q) are the intensities for neutrons
polarized parallel (+) and anti-parallel (–) to the magnetic field.
The field-dependent scattering intensity is extracted as
I
H
(Q)¼I(Q,H)–I(Q,0). Data were taken at temperature 290 K. The
alpha-dependence is not statistically resolved for our measure-
ments since the obtained magnetic signal is rather weak. We are
forced to average the intensity over an azimuth angle in order to
reach a reasonable signal-to-noise ratio.
1
0
0
.0
.5
.0
0
0
0
.10
.05
.00
-
-
0.5
1.0
-
-
0.05
0.10
-40 -20
0
20
40
-
1000
-500
0
500
1000
μ0H (mT)
Fig. 1. The SEM data: a side view of the anodic alumina film with arrayed Co
Fig. 2. Magnetization curve for Co nanowires embedded into alumina membrane in
nanowires.
the field applied perpendicular to the long axis of the wires.

A.P. Chumakov et al. / Physica B 406 (2011) 2405–2408
2407
-
1
μ0H = 1 mT
μ0H = 800 mT
Q (nm )
y
60
40
(
01)
0
.069 (1-1)
2
0
0
(
10)
0
H,P0
-
-
-
20
40
60
-
0.069
-
1
Q (nm )
x
-
0.069
0
0.069
0
.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
-
1
Q (nm )
1
1
1
1
1
80
60
40
20
00
μ0H = 1 mT
μ0H = 800 mT
(
10)
μ0H = 1 mT
μ0H = 800 mT
80
6
0
8
6
4
2
0
0
0
0
40
(11)
(
20)
2
0
0
0
.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
-
20
-
1
Q (nm )
Fig. 3. Typical 2D neutron diffraction pattern from Co/Al
spatially ordered structure (a). Momentum-transfer dependence of the total
nuclear and magnetic) neutron scattering intensity at
H¼1 and 800 mT (b).
2
O
3
nanocomposite with
0.04
0.06
0.08
0.10
Q (nm )
0.12
0.14
0.16
0.18
-1
(
m
0
Fig. 4. Q-dependence of the nuclear–magnetic interference
D
I(Q) at
m
0
H¼1 and
8
00 mT (a) and magnetic contribution to the scattering I (Q) (b).
H
Fig. 4 shows the nuclear–magnetic interference
DI(Q) and the
field-dependent part of neutron scattering I (Q) as a function of
H
scattering vector. The peak positions for the magnetic contribu-
tions match the positions of nuclear peaks (at zero magnetic field in
Fig. 3b). This confirms the formation of the spatially ordered
magnetic structure, e.g. the ordered array of nanowires with
periodicity of 106 nm.
that an increase in intensity is gradual and has a tendency to the
saturation with increase in the magnetic field above 600 mT. This
behavior coincides with the data of magnetization curve obtained
from SQUID, Fig. 2.
H
The magnetic-field-dependent scattering component I (Q)
As seen from Fig. 4a, the interference contribution
zero in the entire momentum-transfer range for the unmagnetized
sample (
H¼1 mT). In high magnetic fields, the polarization-
D
I(Q) is equal to
shown in Fig. 4b represents the difference between the magnetic
cross-sections of the sample in two significantly different states, a
m
0
state close to saturation magnetization (in a field of
and a completely demagnetized state. The main component of the
magnetic cross-section at
H¼800 mT is the system of the
m H¼800 mT)
0
dependent scattering consists of two contributions: interference
maxima at Q values coinciding with the positions of diffraction
reflections in scattering on the ordered structure of nanowires and
interference in diffuse small-angle scattering. These two contribu-
tions have opposite signs in a good agreement with theory [8]. The
interference scattering can be approximated by the sum of the
Gaussians with the centers at zero (diffuse scattering) and at Q¼Qhk
m
0
magnetic reflections described by the sum of Gaussians whose
positions correspond to the maxima of the nuclear scattering cross-
section. On the contrary, the main component of the cross-section
in zero magnetic field is scattering on domains, i.e., diffuse
scattering, which is described by a Gaussian at Q¼0 with no
contribution to diffraction reflections. The magnetic field depen-
dence of the integral intensity of magnetic reflection at Q¼9Q109 is
shown in Fig. 5b. The intensity does not change within error bars in
the range from 0 up to 200 mT. Then the negligible increase occurs
up to a point of the saturation field. Thereafter, upon decreasing
field the magnetic intensity component rises up rapidly with
subsequent smooth decrease to the field value of 200 mT. A similar
(coherent scattering). The magnetic field dependence of the inter-
ference contribution integrated over the azimuthal angle at 9Q109 and
diffusion scattering are shown in Fig. 5a. Note that the interference
contribution is proportional to the average magnetization, which
exhibits no hysteresis within error bars in initial magnetization or in
subsequent remagnetization of the system of magnetic nanowires in
the field perpendicular to the long axis of wires. It is interesting to note

2
408
A.P. Chumakov et al. / Physica B 406 (2011) 2405–2408
to the fact that the nuclear scattering amplitude of cobalt
ꢀ
12
0
0
0
0
.06
.04
.02
.00
(0.28–10
cm) is twice as smaller as the magnetic scattering
ꢀ
12
amplitude (0.47–10
cm) [14]. One would expect that the
magnetic scattering from Co nanowires should be 4 times larger
than the nuclear one. This disagreement between the expected and
the measured values for magnetic contribution can be explained if
one takes into account that the coherent volume for the nuclear
scattering is much larger than the coherent volume for the
magnetic scattering [5]. Consequently, nuclear part has become
predominant over magnetic one.
To explain the small values for the coherent volume of the
magnetic scattering, we suppose that the cobalt nanowires consist
of small single crystal parts, forming a polycrystalline wire. The
magnetic moments in two neighboring crystallites pointing along
the c-axis of the HCP crystallographic structure are most probably
Arefl
ADiff
-
-
0.02
0.04
0
directed differently. The applied magnetic field (m H41 T) cannot
turn the magnetization toward the field axis thus producing out of
single wire a very inhomogeneous magnetic object with low
coherent volume. Thus one can conclude that it is crystallographic
anisotropy of Co crystallites of the wire but not the wire’s shape
anisotropy that plays the crucial role in the magnetization process
of the massive of Co nanowires.
In conclusion, we have investigated the process of the remag-
netization of the Co nanowires arrays in the magnetic field applied
perpendicularly to the long wire’s axis. It has been illustrated that
polarized neutron scattering provides unique information inacces-
sible by the standard magnetometry techniques, namely, an
anomalously low value of the magnetic contribution as compared
with nuclear one. This behavior is interpreted in terms of low
coherent volume magnetic structure as compared with the rela-
tively large coherent volume of the nuclear one.
-
800 -600 -400 -200
0
200 400 600 800
μ0H (mT)
0
.06
.05
.04
.03
.02
.01
.00
0
0
0
0
0
0
The Russian authors are grateful to the GKSS Research Centre
Germany) for hospitality and also thank RFBR for the support
Grant nos. 10-02-00634).
(
(
References
-
800 -600 -400 -200
0
200 400 600 800
[
[
1] S. Goolaup, N. Singh, A.O. Adeyeye, V. Ng, M.B.A. Jalil, Eur. Phys. J. B. 44 (2005)
μ0H (mT)
259.
2] E.C. Stoner, E.P. Wohlfarth, Philos. Trans. R. Soc. 240 (1948) 599.
Fig. 5. Magnetic field dependence (a) of the interference contributions (diffusive
and coherent) and (b) the magnetic coherent contribution obtained from the
approximation of the experimental data.
[3] W.F. Brown, Phys. Rev. 105 (1957) 1479.
[4] M. Jaafar, R. Yanes, D. Perez de Lara, O. Chubykalo-Fesenko, A. Asenjo,
E.M. Gonzalez, J.V. Anguita, Phys. Rev. B. 81 (2010) 054439.
[
5] S.V. Grigoriev, A.V. Syromyatnikov, A.P. Chumakov, N.A. Grigoryeva,
K.S. Napolskii, I.V. Roslyakov, A.A. Eliseev, A.V. Petukhov, H. Eckerlebe, Phys.
Rev. B. 81 (2010) 125405.
behavior occurred at the reverse magnetization process. It is worth
noting that such a ‘‘butterfly’’ curve of magnetic contribution in
dependence of the magnetic field are observed in the other
magnetic arrays structures [8,13]. The ‘‘butterfly’’ curve is usually
observed from different types of magnetic nanowires arrays in
polarized SANS experiment, when the field is perpendicular to the
wire axis. In our case the ‘‘butterfly’’ curve is hardly statistically
resolved; however, it can be seen that subsequent increase and
decrease in the magnetic field produce two curves: the first one lays
lower than the second one. Some instability of the magnetic sample
[
[
6] S.V. Grigoriev, N.A. Grigorieva, A.V. Syromyatnikov, K.S. Napolskii, A.A. Eliseev,
A.V. Lukashin, Yu.D. Tretyakov, H. Eckerlebe, JETP Lett. 85 (2007) 449.
7] K.S. Napolskii, I.V. Roslyakov, A.A. Eliseev, A.V. Petukhov, D.V. Byelov,
N.A. Grigoryeva, W.G. Bouwman, A.V. Lukashin, K.O. Kvashnina,
A.P. Chumakov, S.V. Grigoriev, J. Appl. Cryst. 43 (2010) 531.
[8] S.V. Grigoriev, N.A. Grigorieva, A.V. Syromyatnikov, K.S. Napolskii, A.A. Eliseev,
A.V. Lukashin, Yu.D. Tretyakov, H. Eckerlebe, JETP Lett. 85 (2007) 605.
[9] S.V. Grigoriev, A.P. Chumakov, A.V. Syromyatnikov, N.A. Grigorieva,
A.I. Okorokov, K.S. Napol’skii, I.V. Roslyakov, A.A. Eliseev, A.V. Lukashin,
H. Eckerlebe, Solid State Phys. 52 (2010) 1080.
10] H. Masuda, K. Fukuda, Science 268 (1995) 1466.
11] S. Shingubara, J. Nanoparticle Res. 5 (2003) 17.
[
[
under investigation is observed at
this instability remains unknown.
It should be noticed that value of magnetic contribution
reaches 6% from the total intensity only. This contradicts obviously
m
0
H¼ ꢀ100 mT. The nature of
[12] G. Sauer, G. Brehm, S. Schneider, K. Nielsch, R.B. Wehrspohn, J. Choi,
H. Hofmeister, U. Gosele, J. Appl. Phys. 91 (2002) 3243.
[13] N.A. Grigoryeva, S.V. Grigoriev, H. Eckerlebe, A.A. Eliseev, A.V. Lukashin,
K.S. Napolskii, J. Appl. Cryst. 40 (2007) s532.
[14] G.E. Bacon, K. Lonsdale, Rep. Prog. Phys. 16 (1953) 1.