Microstructure and magnetic properties of colloidal cobalt nano-clusters

Article abstract of DOI:10.1016/j.jmmm.2010.07.008

The magnetic response of nanometer sized Co nanoparticles (NP) prepared using reverse micelle solutions are presented. The use of complementary structural and morphological probes (like transmission electron microscopy, high resolution electron microscopy, X-ray absorption spectroscopy) allowed to relate the magnetic properties to the size, morphology, composition and atomic structure of the nanoparticles. All data agree on the presence of a coreshell structure of NPs made of a metallic Co core surrounded by a thin Co-oxide layer. The coreshell microstructure of NPs affects its magnetic response mainly raising the anisotropy constant.

Full text of DOI:10.1016/j.jmmm.2010.07.008

Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials
journal homepage: www.elsevier.com/locate/jmmm
Microstructure and magnetic properties of colloidal cobalt nano-clusters
Ã
R. Torchio a,b,c, C. Meneghini a,b, , S. Mobilio a,b,d, G. Capellini a, A. Garcia Prieto e, J. Alonso f,
M.L. Fdez-Gubieda f, V. Turco Liveri g, A. Longo h, A.M. Ruggirello g, T. Neisius i
a
ꢀ
Dipartimento di Fisica ‘‘E. Amaldi’’, Universita di Roma Tre, via della Vasca Navale 84, I-00146 Roma, Italy
b CNR-TASC c/o GILDA-ESRF Grenoble, France
c European Synchrotron Radiation Facility, 6 Rue Jules Horowitz, BP220, 38043 Grenoble Cedex, France
d Laboratori Nazionali di Frascati INFN, via E. Fermi 40, I-00044 Frascati, Roma, Italy
e Departamento de F´ısica Aplicada I, Universidad del Pa´ıs Vasco, Spain
f Departamento de Electricidad y Electro´nica, Universidad del Pa´ıs Vasco, Spain
g
ꢀ
Dipartimento di Chimica Fisica ‘‘F. Accascina’’, Universita di Palermo, Viale delle Scienze, Parco d’Orleans II, Edificio 17, 90128 Palermo, Italy
h ISMN, Istituto per lo Studio dei Materiali Nanostrutturati, CNR, Via U. La Malfa 153, 90146 Palermo, Italy
i F´ede´ration des Sciences Chimiques de Marseille, Universite´ Paul Ce´zanne, Faculte´ des Sciences et Techniques Campus de Saint Je´roˆme av. Escadrille Normandie Niemen 13397
Marseille Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The magnetic response of nanometer sized Co nanoparticles (NP) prepared using reverse micelle
solutions are presented. The use of complementary structural and morphological probes (like
transmission electron microscopy, high resolution electron microscopy, X-ray absorption spectroscopy)
allowed to relate the magnetic properties to the size, morphology, composition and atomic structure of
the nanoparticles. All data agree on the presence of a core–shell structure of NPs made of a metallic Co
core surrounded by a thin Co-oxide layer. The core–shell microstructure of NPs affects its magnetic
response mainly raising the anisotropy constant.
Received 28 January 2010
Received in revised form
2 July 2010
Available online 13 July 2010
Keywords:
Cobalt nanoparticles
TEM
& 2010 Elsevier B.V. All rights reserved.
XAFS
Magnetism
1. Introduction
complex interplay among several parameters intrinsic of NPs
such as size and morphology, atomic structure and chemical
composition, or related to the synthesis parameters, environ-
mental variables, interactions with substrates and matrices. In
order to reliably comprehend the correlation among the many
parameters involved all these aspects must be considered and this
require complementary probes [2] and careful investigations.
Moreover the development of massive methods for the synthesis
of nano-structured materials with well defined and controlled
physical properties is a topic of extreme interest for practical
applications. Chemical synthesis routes generally allow massive
production of NPs with rather well defined NP sizes and chemical
compositions [3–6]. In particular methods based on the chemical
reduction in micro-heterogeneous systems [7–10], using reverse
micelles solutions as reaction and stabilizing media, represent a
valuable way to get a precise and reproducible control on size,
shape and distribution of the NPs simply acting on the nature of
the surfactant, on the composition of the solution and on the
sample ageing.
Nano-sized particles (NP) made of a few to some thousand
atoms attract widespread interest for their peculiar physical
properties. These special properties are mainly due to the finite
size and huge surface to volume ratio of NPs, which imply a high
number of surface atoms having a coordination definitively
different from that of bulk ions. Moreover the small size of NPs,
comparable with (or even smaller than) the coherence lengths of
the electrons, modifies their electronic properties and their
response to external fields. In particular metallic NPs may show
an ample variety of phenomena such as super-paramagnetism,
magneto-resistance, magnetic anisotropy [1] which are relevant
for innovative applications in very different fields ranging from
medicine to microelectronics, magnetic storage devices and
sensors. Understanding the origin of these properties is
a
fundamental issue in order to finely tune and control the physical
response of NP based materials. This is a challenging task, not
straightforward, since their physical response originates from a
This work is aimed to characterize the magnetic response of Co
NPs prepared by reverse micelles synthesis and relate the
magnetic properties to sizes and chemical composition of the
NPs, as a function of the preparation route. Co NPs were
synthesized from colloidal solutions using different surfactants
Ã
ꢀ
Corresponding author at: Dipartimento di Fisica ‘‘E. Amaldi’’, Universita di
Roma Tre, via della Vasca Navale 84, I-00146 Roma, Italy.
E-mail address: meneghini@fis.uniroma3.it (C. Meneghini).
0304-8853/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2010.07.008

3566
R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
(AOT and DEHP) and drying methods. Complementary experi-
mental probes were used to characterize the morphology,
structure and magnetic response of the NPs: transmission
electron microscopy (TEM) was employed to determine the
particle morphology and size distribution. High resolution TEM
(HR-TEM) was used to selectively identify the atomic structure of
single particles. Co K edge X-ray absorption spectroscopy (XAS) in
the near edge region (XANES) was used to determine and quantify
the chemical nature of the different Co containing phases present
in the samples. Finally analysis of the hysteresis loops (HL) and
temperature dependent magnetization in the field cooling and
zero field cooling (FC–ZFC) modes allowed characterizing the
magnetic properties of NPs.
Combining those complementary information permitted to
understand the magnetic response of the NPs as a function of their
structure and composition: the structural characterization (TEM
and XAS) suggests NPs have a core shell structure made of a
metallic Co-fcc core surrounded by a thin Co-oxide layer, such a
core shell structure raises the magnetic anisotropy constant with
respect to pure Co NPs. The synthesis route parameters (surfac-
tant, drying procedures, ageing) affect the size and composition of
NPs modifying the magnetic response of the samples.
which demonstrated the presence of NP with sizes (diameter) in
the 1–3 nm range. The absence of Bragg peaks in the wide angle
X-ray diffraction demonstrate that big particles (larger than few
nm) are definitively rare in these samples.
In our samples the size distribution and structure of the NP
were accurately determined by TEM and HR-TEM. TEM measure-
ments were performed at the LIME (Laboratorio Interdipartimen-
tale di Microscopia Elettronica) of the University Roma TRE
(Rome, Italy) using a Philips CM 120 microscope. HR-TEM images
were collected at the IM2NP-UMR CNRS laboratory of the
´
Universite Paul Cezanne (Marseille, France) using
a TITAN
˚
300 keV instrument with nominal resolution of 0.7 A. The
chemical nature of the different Co containing phases present in
the samples were investigated by X-ray absorption spectroscopy
of the Co K edge. Spectra were collected at the BM08-GILDA
beamline [12] at ESRF (European Synchrotron Radiation Facility)
in Grenoble (France).
TEM images (Fig. 1) reveal the presence of small NPs roughly
spherical shaped. The particle size distribution was fitted with a
log-normal distribution obtaining average sizes of D ¼ 3–4.5 nm
(Table 1). The NPs size obtained by TEM is larger compared with
SAXS [11] data (1–3 nm). This weak discrepancy is not surprising
as smallest particles (sub-nanometer size) are hindered in TEM
analysis, as well documented in the literature [2,13]. The DEHP
capped particles resulted to be systematically smaller than AOT
capped ones. Within the experimental uncertainties, the particle
2. Sample synthesis, morphological, structural and chemical
characterization
size distribution broadening
samples.
s, is around 0.8–1.3 nm for all the
Samples were prepared following the procedure well detailed
in [11] which consists of Co2+ reduction in the confined space
of cobalt bis(2-ethylhexyl) sulfosuccinate Co(AOT)2 and cobalt
bis(2-ethylhexyl) phosphate Co(DEHP)2, reverse micelle solutions
dispersed in n-heptane. Reduction was carried out by adding a
solution of sodium borohydride (NaBH4) in ethanol to the micellar
solutions:
The HR-TEM permits to selectively investigate single particles
in the sample, with atomic resolution. The HR-TEM images (Fig. 2)
shows that the nanoparticles are roughly spherical shaped. The
average size (diameter) of the NPs measured by HR-TEM is around
2.5–3 nm, smaller than in TEM images; this is likely due to the
higher resolution allowing to better individuate also the smaller
particles. Nevertheless it must be noted that fewer particle are
counted by HR-TEM this giving a poorer statistics and a larger
uncertainty. The average inter-particle distance (center to center)
is 5–7 nm.
CoðXÞ2 þ2NaBH4-Coþ2NaðXÞþB2H6 þH2
with X¼AOT or DEHP. The initial solution is rose but soon after
addition of NaBH4 the mixture releases H2 gas and turns black due
The NPs have generally poor crystallinity and the background
coming from the polymeric matrix worsens the interpretation of
HR-TEM images. However the Fourier transforms calculated from
HR-TEM images of selected particles clearly show the square and
hexagonal symmetries of the fcc lattice oriented along 100 and
111 directions (insets in Fig. 2). The lattice spacings are closely in
to the formation of
a fine metallic Co powder. From this
suspension, two different types of samples were prepared: type
A samples were obtained by evaporation under vacuum of the
liquid components; one sample (0–Au) was also aged two months
under N2 inert atmosphere. Samples B were obtained leaving for
18 h the suspension at room temperature, then removing and
drying the black precipitate formed from the suspension. The
investigated samples are listed in Table 1. Analogous samples
were characterized [11] by small angle X-ray scattering (SAXS)
Table 1
Left columns: list of the Co-NP samples investigated as a function of the synthesis
methods.
Sample Surfactant Type Co phases (%)
Co in NPs (%)
D (nm)
Cofcc Co3O4 Co2+ Cofcc
Co3O4
(0)
(1)
(2)
(3)
(4)
AOT
A0
A
B
26
51
61
31
45
38
21
19
26
40
36
29
21
44
15
41
71
76
54
53
59
29
24
46
47
4.5
4
AOT
AOT
4.5
3
DEHP
DEHP
A
B
3.5
Second columns (Co phases (%)): the relative concentration of Co in the different
phases as obtained by Co K edge XANES analysis (the error bar is around 5%). Third
columns (Co in NPs (%)): the relative amount of Co containing magnetic phases
(Cofcc and Co3O4) is reported. Fourth column ðDÞ: average particle size (diameter)
from TEM measurements; the standard deviation of particle size distribution is in
between 0.8 and 1.3 nm.
Fig. 1. TEM results: the NP size distribution histogram of sample (3) and the TEM
image in the inset.

R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
3567
agreement with that of metallic Cofcc (being d110 ꢀ 2:422:5 A).
˚
Looking at the borders of NPs we found lighter absorbers and
changes in the lattice spacing suggesting Co oxide phases
surrounding the metallic core (Fig. 2b). In all the (1)–(4) samples
isolated NPs with structure consistent with Co-oxide phase are
very rarely observed. Only in the sample (0) some large particle
(some tens nm) is recognized as Co3O4 (Fig. 2c), this is the sample
with the larger fraction of Co-oxide (see the XAS analysis below).
A close look into the HR-TEM image points out some defective NPs
as, for instance, in the right panel of Fig. 2b where a stacking fault
resulting in 100 and 111 orientations is clearly visible. In this
figure, at the surface of the defective NP, a light element shell with
thickness of 0.4–0.6 nm appears, showing lattice planes distances
˚
consistent with a Co-oxide phase (around 2 A).
The analysis of HR-TEM images suggests that NPs are mainly
formed by metallic Co phase with fcc structure, surrounded by a
thin shell of Co-oxide phase while isolated Co-oxide NPs are rarely
observed. In order to have a deeper insight into the nature of Co
containing phases, the samples were investigated by X-ray
absorption spectroscopy at the Co K edge. Absorption spectra
were collected in transmission geometry at the Co K edge in the
7690–7800 eV energy range keeping the samples at low tempera-
ture (77 K). The following samples were also measured as
reference: Co-hcp (metal foil), CoO and Co3O4 as references for
the Co oxide phases and Co(AOT)2 unreduced micellar solution as
reference for the Co+2 bound to the surfactant. The Co fcc
spectrum of Ref. [14] was used as reference for metallic Cofcc
phases.
The raw absorption spectra were corrected subtracting a pre-
edge linear background and were normalized to the jump edge
discontinuity. The NP normalized spectra
linear combination of normalized spectra of the reference
m
expwere fitted to a
P
compounds mrjef
:
mth
¼
jajmref where the sum runs over the
reference compound used. Thej aj parameters (with the constraint
P
jaj ¼ 1) directly quantify the fraction of the different phases in
which Co is present in the samples. The refinement has been
performed [15] minimizing the square residual sum:
P
2
R2 ¼ iðmexpðEiÞꢁmthðEiÞÞ in which the sum runs over the experi-
mental points. Three contributions were necessary to fit the
experimental data: the Cofcc metallic phase, the Co3O4 oxide phase
and the contribution of Co2+ ions bound to the surfactant. The
refined aj are reported in Table 1, an example of XANES
refinement is presented in Fig. 3. We underline that attempts to
refine the data using the spectrum of Co hcp foil instead of Cofcc
,
definitively worsen the quality of the fit for all the investigated
samples. This is not surprising because it is well known that fcc
phase is favored with respect to the hcp in nanometer sized Co
particles [16] and is consistent with our HR-TEM observations.
The XANES analysis also demonstrates that Co-oxide is mainly
Co3O4-like while CoO fraction is negligibly in these samples.
The fraction of Cofcc increases from wet (A) to dried
(B) samples. On the contrary the fraction of Co+2 bound to the
surfactant is higher in the wet samples (A) than in the dried ones
(B) meaning that the drying procedure also removes part of the
surfactant. The magnetic data (see below) suggest that the Cofcc
and Co3O4 phases are the main responsible for the magnetic
response of our samples, then in order to highlight the effect of
synthesis and drying procedures on these magnetic phases, the
relative amounts of Co in Cofcc and Co3O4 phases are separately
reported in Table 1 (Co in NPs (%)): note that the AOT capped NPs
present a higher amount of Cofcc showing that AOT surfactant is a
more efficient capping agent against oxidation. In DEHP capped
Fig. 2. High resolution TEM images (colour online): sample (4) (panel a), sample
(2) (panel b) and sample (0) (panel c). In the insets the Fourier Transform (FT) of
selected particles show square (fcc) and hexagonal symmetries of the fcc lattice.
The panel b shows the image of a single NP with a stacking fault resulting in 100
and 111 orientations, the FT (in the inset) shows the square and hexagonal
symmetries superimposed. The high resolution images of sample (0) show large
particles in which the Co3O4 phase is recognized.
NPs, the drying procedure (samples B) preserves the overall Cofcc
/
Co3O4 ratio while in AOT capped samples it slightly increases after
drying. In sample (0) the Cofcc to Co3O4 ratio decreases with
respect to sample (1), as a consequence of the long time ageing, in

3568
R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
3. Magnetic properties
(1)
The magnetic response of the samples as a function of
fit
temperature and applied magnetic field has been characterized
by ZFC–FC magnetization curves and hysteresis loop analysis
(Figs. 5 and 6). The ZFC–FC data were performed using a SQUID
magnetometer: the samples were zero-field cooled (ZFC) from
300 to 5 K; then a magnetic field of H0 ¼ 100 Oe was applied and
the magnetization measured while raising the temperature up to
300 K. Afterwards, keeping H0 ¼ 100 Oe, the samples were field
cooled (FC) down to 5 K and the magnetization was measured
again while raising the temperature up to 300 K.
Co fcc
Co2+
Co3O4
The ZFC curve behavior is typical of superparamagnetic (SPM)
NPs: the broad maximum signals the blocking temperature TB. For
isolated SPM particles TB is related to the particle size and
anisotropy constant: for a given anisotropy constant, smaller
particles would have lower TB, and for a given size, higher
anisotropy constant results in higher TB. However, it must be
noticed that the values of TB reported in literature for Co NPs of
similar size and shape depict widely spread values in the range
50–200 K [10,18,9,17]; these differences have been ascribed to
many effects: inter-particle interactions [9], enhancement of
anisotropy constant due to surface effects, core–shell structure
of the NPs [10] and so on. The TB values found in our samples
ranges in between 120 and 220 K and are slightly higher for B
samples than for A ones. ZFC–FC curves closely resemble those
reported in literature [17] for Co NPs embedded into polymeric
membranes. A noticeable increase of the magnetization in ZFC
and FC curves at temperatures below ꢀ 10 K is observed in
sample (0) and also, but much weaker, in sample (4). Similar
features have been reported in literature [17] and ascribed [13] to
ultrafine particles (sub-nanometer size and/or few atoms Con-O
molecules) which retain their paramagnetic behavior in the whole
temperature range or have a very low blocking temperature
(below 5 K in this case). For example Co3O4 is antiferromagnetic
Res.
Fig. 3. (colour online) Refinement of Co K edge XANES of the sample 1 as a linear
NP
combination of reference compound XANES. Sample
1
XANES data
m
are
presented as dots, best fit (
m
mod) is reported as full line. The contributions of the
different models, each weighted by their ai parameter, are reported (gray).
The residual (res ¼ mNP mod) is also shown.
ꢁm
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
ꢀ
with Neel temperature TN ¼ 33 K in the bulk form but, in the form
of nanosized particles, it becomes superparamagnetic
[20,21,23,22,24] with very low blocking temperatures which
depends on the particles size: TB around 10 K have been reported
for 2 nm sized particles [21].
The hysteresis loops were measured at T¼5 and 300 K using a
SQUID (samples 0 and 1) and a vibrating sample magnetometer
(VSM) (samples 2, 3 and 4) up to a maximum applied field of
0H ¼ 5 ꢂ 104 Oe (up to 105 Oe for the VSM measurements)
0
10
20
30
40
D[Å]
50
60
70
m
(Fig. 5, right panels). The 300 K hysteresis loops show very
low coercivity, weak remanence and unsaturated behavior
Fig. 4. Fraction of Co ions on the surface of spherical fcc Co clusters, the curve has
even at
a
magnetic field of 5 ꢂ 104 Oe, typical of weakly
been simply calculated cutting spherical clusters from
and counting bulk and surface atoms.
a fcc bulk model
interacting superparamagnetic systems [19]. The magnetization
curves of all the samples at 5 K present a nonlinear unsaturated
behavior and larger hysteresis loops. This finding, accordingly to
Ref. [13], is consistent with the presence of ultrafine magnetic
particles behaving superparamagnetic even at these low tem-
peratures, while the larger particles become saturated at high
fields. The anomalous behavior of hysteresis loop measured at 5 K
for sample (0) must be related to the presence of two
ferromagnetic phases with different anisotropies, one from the
Co nanoparticles and the other to blocked Co3O4 large NPs as
those observed by HR-TEM.
The values of coercivities measured at 5 K result larger for wet
samples (1) and (3) (respectively, Hc ¼ 1550 and 1850 Oe) than for
the corresponding dry samples 2 and 4 (respectively, Hc ¼ 1150
and 1550 Oe). This trend originates from the inter-particles
magnetic coupling which favors the magnetization reversal thus
decreasing the coercive fields: the inter-particles magnetic
coupling is expected more intense in the dried samples as they
have higher NP concentration.
agreement with HR-TEM that showed the presence of large Co3O4
particle in this sample. This finding suggests that the ageing of
samples rises the probability of NPs oxidation even in inert
atmosphere, this is likely due to the oxygen ions still dispersed in
the sample after evaporation of the solvents.
XAS and HR-TEM analyses coherently demonstrate the forma-
tions of core shell NPs having Cofcc metallic core surrounded by
thin Co-oxide shells.
In Fig. 4 we show the fraction of Co surface ions for spherical
particles as a function of the diameter: NPs having size around
3 nm, like those observed in our DEHP samples, would have about
25–30% of Co atoms at the surface forming the oxide shell. But the
fraction of Co in the Co3O4 phase is about twice. This suggests that
the oxide shell is about two layers thick, consistent with the
0.4–0.5 nm thickness estimated by HR-TEM image inspection.

R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
3569
(0) Co(AOT)2 A’
(0) Co(AOT)2 A’
2.0
1.0
0.0
0.4
0.2
1
0
−1.0
−2.0
−1
3
0
3
−2.0 10
0.0 10
2.0 10
0.0
0.4
(1) Co(AOT) A
2
(1) Co(AOT)2 A
2.0
1.0
1
0
0.0
0.2
−1.0
−2.0
−1
3
0
3
−2.0 10
0.0 10
2.0 10
(2) Co(AOT)2 B
0.0
4.0
(2) Co(AOT)2 B
20
10
10
0
2.0
0
−10
−20
−10
3
0
3
−2.0 10
0.0 10
2.0 10
0.0
0.4
3.0
(3) Co(DEHP)2 A
(3) Co(DEHP)2 A
1.5
1
0
0.0
0.2
−1.5
−3.0
−1
−2.0 103
0.0 100
2.0 103
0.0
4.0
(4) Co(DEHP)2 B
20
(4) Co(DEHP)2 B
10
10
0
0
2.0
−10
−20
−10
−2.0 103
0.100
H [Oe]
0.0 100
2.104
2.0 103
4.104
−4.104
−2.104
0
100
200
300
T [K]
Fig. 5. Left panels: ZFC–FC measurements (ZFC are blue online, FC are red online) for all the investigated samples. Right panels: hysteresis loops at 5 K (red online) and
300 K (blue online) for all the investigated samples. In the inserts is presented the low H region. Magnetization is normalized to sample mass. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
The hysteresis loops measured at 300 K were quantitatively
analyzed assuming the presence of two magnetic terms: a SPM
one coming from the Co NPs and a linear one arising from the
Co3O4 oxide phase, which behaves PM at room temperature [22],
the Co2+ bound to the surfactant and the weak diamagnetic
contribution from the surfactant (wDIA ¼ ꢁ8 ꢂ 10ꢁ8 emu=g Oe).
In order to take into account the residual weak hysteretic
behavior of the samples, we have calculated the average between
the two branches Me+xp (raising H) and Mꢁ (decreasing H) as
suggested in Ref. [25] in the case of interacting systems:
The Mexp curves were fitted to the following expression [14]:
Z
1
Mth ¼ MsSPM
fðDÞLðaÞ dDþwlinear ꢃ H
ð2Þ
0
where MSsPM is the saturation magnetization of SPM metallic Co
and L(a) is the Langevin function LðaÞ ¼ cothðaÞꢁaꢁ1, describing
the SPM response of the metallic Cofcc NPs. Here a ¼
mH=kBT,
fcc
where
m
¼ MsCo ꢃ V is the magnetic moment of the Co nanopar-
ticles, MsCo ¼ 1449 emu=cm3 being the saturation magnetization
of fcc Co and V the volume of the nanoparticle. The f(D) is the log-
normal distribution which takes into account the NP size
fcc
1
Mexp
¼
ðMeþxp þMeꢁxp
Þ
ð1Þ
2

3570
R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
0.15
0.10
0.05
0.00
2.0
(0) Co(AOT)2 A’
1.0
(0) Co(AOT)2 A’
1
0
0.0
−1.0
−2.0
−1
−2.0 103
0.0 100
0.0 100
0.0 100
0.0 100
2.0 103
2.0 103
2.0 103
2.0 103
2.0
(1) Co(AOT)2 A
1.0
0.10
0.05
0.00
(1) Co(AOT)2 A
(2) Co(AOT)2 B
(3) Co(DEHP)2 A
1
0
0.0
−1.0
−2.0
−1
−2.0 103
−2.0 103
−2.0 103
20
(2) Co(AOT)2 B
10
0.10
0.05
0.00
10
0
0
−10
−20
−10
3.0
(3) Co(DEHP)2 A
1.5
0.10
0.05
1
0
0.0
−1.5
−3.0
−1
0.00
0.15
20
(4) Co(DEHP)2 B
10
10
0
0.10
0.05
(4) Co(DEHP)2 B
0
−10
−20
−10
3
0
3
−2.0 10
0.0 10
2.0 10
−4.104
−2.104
0.100
H [Oe]
2.104
4.104
0.0
5.0
10.0
15.0
DM [nm]
Fig. 6. Left panels: NP size distribution functions f(D) obtained refining 300 K hysteresis loops. The tics point out the mean particle sizes D. Right panels: experimental
magnetization Mexp (dots, red online) measured at T¼300 K and best fit Mth (full lines, blue online) curves for all the investigated samples. In the insets the low field region
is presented to highlight the best fit quality. Magnetization is normalized to the sample mass. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)
distribution:
1
variance of the distributions sM2
:
!
!
DM2
ln2ðD=
aÞ
2=2
pffiffiffiffiffiffi
DM
¼
a
eb
and s2M ¼ D2M
ꢁ1
:
ð4Þ
fðDÞ ¼
exp
ꢁ
ð3Þ
2
2
b2
a
2pD
b
The term wlinear ꢃ H in Eq. (2) describes the linear contribution
coming from the paramagnetic Co2+ bound to the surfactant and
from the Co-oxide phase Co3O4. As the two contributions are in
principle different, the linear susceptibility parameter wlinear has
are reported in Table 2: DM are larger for AOT than for DEHP
samples, in agreement with TEM analysis. We notice that the size
of NPs obtained from magnetic analysis are systematically larger
than those obtained by TEM. Magnetic sizes larger than those
found by TEM have been reported in literature and interpreted as
an effect of inter-particle interactions [9] or core shell structure
of NPs [10]. We expect both the effects in our samples as the NPs
been left free. The other free parameters are: MsSPM
, a and b. The
best fits are shown in Fig. 6 together with the f(D) resulting from
the refinement. The mean magnetic particle diameters DM and the

R. Torchio et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 3565–3571
3571
Table 2
NP size distribution parameters as obtained from 300 K hysteresis loops analysis.
Acknowledgments
We are grateful to Dr. In~aki Orue from the Servicios Generales
Sample
sM (nm)
Keff (106 erg/cm3)
DM (nm)
´
de Investigacion of the University of Basque Country, for the
accurate technical support during magnetic measurements.
(0)
(1)
(2)
(3)
(4)
7.23(6)
8.13(7)
10.5(3)
6.23(5)
7.1(2)
5.2(1)
6.2(2)
9.8(7)
4.3(1)
4.5(2)
4.4
6.3
5.7
8.2
8.6
References
[1] J. Bansmann, S.H. Baker, C. Binns, J.A. Blackman, J.P. Bucher, J. Dorantes-Da´vila,
V. Dupuis, L. Favre, D. Kechrakos, A. Kleibert, K.H. Meiwes-Broer, G.M. Pastor,
A. Perez, O. Toulemonde, K.N. Trohidou, J. Tuaillon, Y. Xie, Surf. Sci. Rep. 56
(2005) 189–275.
[2] P. Canton, P.F. Fazzini, C. Meneghini, A. Benedetti, G. Pozzi, Nanoscale
characterization of metal nanoclusters by means of XRay diffraction (XRD)
and transmission electron microscopy (TEM) techniques, in: B. Corain,
G. Schmid, N. Toshima (Eds.), Metal Nanoclusters in Catalysis and Materials
Science, vol. 1, Elsevier, Amsterdam, 2007.
appear quite close together, 5–7 nm being the interparticle
distances determined from TEM and HR-TEM analysis. The
broadening of magnetic size distribution is smaller for DEHP
samples than for AOT samples, consistently with TEM data,
suggesting that DEHP is more efficient in selecting the particle size.
The anisotropy constants of NPs were calculated from TB
accordingly to [26] as Keff ¼ 25kBTB=V (kB being the Boltzmann
constant, V the particle volume obtained from TEM measure-
ments) and reported in Table 2. Due to the surface effects, the
values we found are definitively higher than Keff of bulk fcc Co
(2:7 ꢂ 106 erg=cm3). However, these values are also higher than
published values for Co nanoparticles having similar sizes
[18,27,28], which strengthens the conclusion that the fcc Co
nanoparticles are surrounded by a Co-oxide shell.
[3] S.H. Baker, S.C. Thornton, K.W. Edmonds, M.J. Maher, C. Norris, C. Binns, Rev.
Sci. Instrum. 71 (2000) 3178.
[4] R.P. Methling, V. Senz, E.D. Klinkenberg, Th. Diederich, J. Tiggesa¨umker,
G. Holzhu¨ter, J. Bansmann, K.H. Meiwes-Broer, Eur. Phys. J. D 16 (2001) 173.
[5] G. Gantefo¨r, K.H. Meiwes-Broer, H.O. Lutz, Phys. Rev. A 37 (1988) 2716.
[6] K.H. Meiwes-Broer, Metal Clusters at Surfaces, Springer-Verlag, Berlin, 2000.
[7] C. Chiang, J. Colloid Interface Sci. 239 (2001) 334.
[8] P. Calandra, V. Turco Liveri, A. Longo, Colloid Polym. Sci. 279 (2001) 1112.
[9] X.M. Lin, C.M. Sorensen, K.J. Klabunde, G.C. Hadjipanayis, Langmuir 14 (1998)
7140.
[10] J.P. Chen, C.M. Sorensen, K.J. Klabunde, G.C. Hadjipanayis, Phys. Rev. B 51
(1995) 11527.
[11] A. Longo, F. Giordano, F. Giannici, A. Martorana, G. Portale, A. Ruggirello,
V. Turco Liveri, J. Appl. Phys. 105 (2009) 114308.
4. Conclusions
[12] S. Pascarelli, F. Boscherini, F. D’Acapito, J. Hrdy, C. Meneghini, S. Mobilio,
J. Synchrotron Radiat. 3 (1996) 147.
[13] X. Chen, S. Bedanta, O. Petraic, W. Kleeman, S. Sahoo, S. Cardoso, P.P. Freitas,
Phys. Rev. B 72 (2005) 214436.
[14] A. Garcia Prieto, M.L. Fdez-Gubieda, C. Meneghini, A. Garcia-Arribas,
S. Mobilio, Phys. Rev. B 67 (2003) 224415.
Nanostructured colloidal systems made of Co NPs embedded
into surfactant matrices were accurately characterized using
advanced complementary techniques: (HR-)TEM, Co K edge XAS
and magnetic probes.
TEM images reveal nanocluster sizes having the size in the
3–4.5 nm range as a function of surfactant and drying procedures.
Co K edge XAS data demonstrate the presence of Co in three main
phases: Co-fcc, Co3O4 plus a fraction of Co2+ which remains
bound to the surfactant. The effect of surfactant on the particle
morphology and magnetic response is slightly different: DEHP
determine smaller particles and sharper size distribution, but AOT
surfactant appear more efficient against oxidation.
The analysis of HR-TEM images demonstrates that NPs mainly
have Co-fcc structure. Therefore, combining (HR)TEM and XAS
results it is possible to conclude that NPs have a core shell
structure formed by a Co-fcc core surrounded by a thin Co-oxide
shell about 0.5 nm thick. The core–shell model explains the high
magnetic anisotropy constant of these NPs and the raising of TB
with respect to pure Co NPs. Our results suggest that differences
in the core–shell structure of the particles play a significant role in
determining the ample differences in the TB reported in literature
for Co NPs having similar size and shape.
[15] F. Bardelli, C. Meneghini, S. Mobilio, Sugata Ray, D.D. Sarma, J. Phys.: Condens.
Matter 21 (2009) 195502.
[16] O. Kitakami, H. Sato, Y. Shimadaet, F. Sato, M. Tanaka, Phys. Rev. B 56 (1997)
21.
[17] I.W. Park, M. Yoon, Y.M. Kim, Y. Kim, H. Yoon, H.J. Song, V. Volkov, A. Avilov,
Y.J. Park, Solid State Commun. 126 (2003) 385.
[18] E. Cattaruzza, F. Gonella, G. Mattei, P. Mazzoldi, D. Gatteschi, C. Sangregorio,
M. Falconieri, G. Salvetti, G. Battaglin, Appl. Phys. Lett. 73 (1998) 1176.
[19] M. Knobel, L.M. Socolovsky, J.M. Vargas, Rev. Mex. Fis. 50 (2004) 8–28.
[20] Y. Ichiyanagi, Y. Kimishima, S. Yamada, J. Magn. Magn. Mater. 272–276
(2004) e1245–e1246.
[21] Y. Ichiyanagi, S. Yamada, Polyhedron 24 (2005) 2813–2816.
[22] P. Dutta, M.S. Seehra, S. Thota, J. Kumar, J. Phys.: Condens. Matter 20 (2008)
015218.
[23] S.A. Makhlouf, J. Magn. Magn. Mater. 246 (2002) 184–190.
[24] Y. Hayakawa, S. Kohiki, M. Sato, Y. Sonda, T. Babasaki, H. Deguchi, A. Hidaka,
H. Shimooka, S. Takahashi, Physica E 9 (2001) 250–252.
[25] P. Allia, M. Coisson, M. Knobel, P. Tiberto, F. Vinai, Phys. Rev. B 60 (1999)
12207.
[26] B.D. Cullity, Introduction to Magnetic Materials, Addison-Wesley, 1972.
[27] A. Garcı´a Prieto, M.L. Fdez-Gubieda, Physica B 354 (2004) 92.
~
[28] B.R. Pujada, E.H.C.P. Sinnecker, A.M. Rossi, C.A. Ramos, A.P. Guimaraes, Phys.
Rev. B 67 (2003) 024402.