Journal of

Materials Chemistry

View Online / Journal Homepage / Table of Contents for this issue

Dynamic Article Links <

Cite this: J. Mater. Chem., 2012, 22, 19276

PAPER

www.rsc.org/materials

Structural and magnetic properties of mesoporous SiO2 nanoparticles

impregnated with iron oxide or cobalt-iron oxide nanocrystals

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

Alvise Parma,a Isidora Freris,a Pietro Riello,*a Davide Cristofori,a C
esar de Juli
an Fern
andez,bc
Vincenzo Amendola,d Moreno Meneghettid and Alvise Benedettia
Received 13th April 2012, Accepted 23rd July 2012
DOI: 10.1039/c2jm32314a
Magnetic nanocomposites of FeOx@SiO2 and CoFe2O4@SiO2 were prepared via a wet-impregnation
route using mesoporous silica nanoparticles as a support matrix. The small pores in the matrix were
exploited as nanocavities for controlled growth of the embedded oxide phase, initially examined by
introducing different wt% loadings of FeOx in four different samples and sequentially treating them
under oxidising and reducing conditions. Comparative examination of the morphological and
structural properties of the FeOx@SiO2 compositions shows that a 17 wt% (nominal) loading of the
oxide phase, a mixture of Fe3O4 (magnetite) and g-Fe2O3 (maghemite), is fully embedded within the
pores. The 6070 nm dimensions of the SiO2 nanoparticles are visible in TEM micrographs which
reveal a spheroidal shape. TEM also shows a ca. 3 nm size for the crystalline oxide particles embedded
within, which agrees with the pore sizes estimated through porosimetric analysis. The measurements for
field-cooled (FC), zero-field-cooled (ZFC) magnetizations, and hysteresis loops in the temperature
range of 3 K to 300 K reveal that an enhancement in the density of magnetization is obtained for the
17 wt% FeOx@SiO2 sample following reductive thermal treatment. A CoFe2O4@SiO2 nanocomposite
prepared with a nominal 14 wt% oxide shows comparable structure and morphology to the 17 wt%
FeOx@SiO2 sample, yet superior magnetic properties. The higher density of magnetization in
CoFe2O4@SiO2 is attributed to its 40% content of magnetic material in the crystalline phase, versus
68% in FeOx@SiO2. Efficient surface functionalisation with APTES, monitored by DRIFT-IR,
implies that the magnetic nanocomposites could be used in bio-labelling applications. Data derived
from Raman spectroscopy, N2 adsorption/desorption measurements, and TGA are also used to
characterise the nanocomposite materials.

Introduction
In recent years, nanocomposite magnetic materials have gained
increasing interest due to their unique magnetic, catalytic and
chemical properties.1 Magnetic nanoparticles have already found
use in magnetic recording and ferrofluid technologies, yet today
they are also being explored for their use in biotechnological and
medical applications,2 such as DNA and RNA purification,3 cell
separation,4 drug delivery,5 magnetic recording and magnetic
resonance imaging,610 and hyperthermia cancer treatments.1114
For these purposes, magnetic nanocrystals, including iron oxides

Dipartimento di Scienze Molecolari e Nanosistemi, Universit

a Ca Foscari
Venezia e INSTM, via Torino 155/b, 30172, Mestre, VE, Italia. E-mail:
[email protected]; Fax: +39-041-2346747; Tel: +39-041-2346718
b
CNR-Istituto di Scienze e Tecnologie Molecolari, via C. Golgi 19, 20133
Milano, MI, Italy
c
INSTM-Udr Florence, Department of Chemistry, University of Florence,
via della Lastruccia 3, 50019, Sesto Fiorentino, FI, Italia
d
Dipartimento di Scienze Chimiche, Universit
a di Padova, Via Marzolo 1,
I-35131 Padova, Italia

19276 | J. Mater. Chem., 2012, 22, 1927619288

such as magnetite (Fe3O4) and maghemite (g-Fe2O3), show high

magnetic susceptibility and good dispersibility.
The most important properties of magnetic materials for
medical applications are non-toxicity and biocompatibility,
combined with injectability, stability in physiological environments and high-level accumulation in the target tissue. Hence,
for in vivo applications involving bio-imaging or hyperthermal
therapy, it is highly desirable to coat the magnetic nanocrystals
with a non-toxic material which is also a suitable host for the
desired amount of magnetic phase.15 A silica shell not only
protects and stabilises the magnetic phase, but can also prevent
its direct contact with additional agents linked to the silica
surface thus avoiding unwanted interactions.15 In biological
applications, the use of silica further renders itself interesting
since it is relatively inexpensive and demonstrates high
thermal stability, low cytotoxicity and good chemical
inertness.16
Predominantly, methods developed for the coating of
magnetic nanoparticles with silica have been based on the
St
ober method and solgel processes.1722 Also, porous silica has
This journal is The Royal Society of Chemistry 2012

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

shown promise as an inorganic matrix in magnetic composite

systems. For example, Shi et al. prepared uniform nanospheres
comprising a magnetic core and a mesoporous-silica shell, and
demonstrated their use as magnetic drug carriers for in vitro
release.23 Tartaj et al. obtained silica-coated g-Fe2O3 hollow
spherical particles, with tunable magnetic properties, by the
aerosol pyrolysis of methanol solutions containing iron
ammonium citrate and TEOS.24 Magnetic nanocomposites
formed by iron oxide particles hosted in silica aerogel pores
have been synthesised by Casas et al.,25 through solgel
processes and supercritical evacuation of the solvent, using
Fe(NO3)3$9H2O and FeNa(EDTA)$2H2O as iron precursors.
Zayat et al.26 prepared transparent and magnetic g-Fe2O3/
Vycor-glass composites by impregnating slices of porous Vycorglass (VG) rods with an iron nitrate solution, followed by a
thermal treatment and a reduction process. These composites
exhibited good mechanical, optical and magnetic properties.
Nakamura et al. obtained highly monodispersed g-Fe2O3/SiO2
and 3-Fe2O3/SiO2 nanocomposites by impregnating the mesopores of the host silica spheres with ferrocenecarbaldehyde
dissolved in furfuryl alcohol, followed by polymerisation and
calcination.27 Hence, a generally important consideration
should be taken into account with respect to the mesoporous
nature of the silica matrix that it is suitably structured to
provide ideal sites for nucleation of small magnetic nanocrystals, by limiting their aggregation and growth. The small
pore dimensions could be exploited as nanocavities for the
synthesis of the magnetic nanoparticles within the matrix. This
approach is investigated in the present Article.
Herein, we report the preparation and characterisation of two
nanocomposites comprising iron oxide and cobalt-iron oxide
nanocrystals embedded within the pores of an amorphous SiO2
matrix. The mesoporous SiO2 nanoparticles were loaded as
required with aqueous solutions of Fe(NO3)3$9H2O and
Co(NO3)2$6H2O via a wet impregnation route. The iron
oxide@SiO2 nanocomposite was obtained following two
consecutive thermal treatments, in air and 5% H2 in N2,
respectively. Four different materials with varying wt%
composition of SiO2 and iron oxide were prepared. On the basis
of results obtained for the iron oxide system, an additional
nanocomposite containing magnetic (CoFe2O4) nanocrystals
embedded in SiO2 nanoparticles was synthesised, by means of a
single thermal treatment in air, and characterised as for the
previous materials. The CoFe2O4 oxide exhibits larger anisotropy than ferrites28,29 which is dependent on the Co content.30,31
Hence, the magnetic properties of nanoparticles based on this
oxide can be tailored by controlling the size and composition.3133 The magnetic properties of the nanocomposite materials have been discussed in the present investigation. The
degree and type of magnetization with respect to nanocrystalline/nanocomposite system, particle sizes and their distributions,
blocking temperature, and related results were obtained from
the systematic analysis of the magnetic data.
The post-synthetic grafting of the composites with a silane
coupling agent, (3-aminopropyl)triethoxysilane (APTES), was
investigated by monitoring the availability of surface silanols
after the thermal treatments. Surface functionalisation is necessary in applications such as bio-labeling, in which further
immobilisation of biological molecules is required.
This journal is The Royal Society of Chemistry 2012

Experimental section
Materials
Cetyltrimethylammonium bromide (CTAB, Fluka), tetraethyl
orthosilicate (TEOS, Aldrich, 98%), absolute ethanol (J. T.
Baker, 99.9%), ammonium hydroxide solution (Fluka, 28 wt% in
water), iron(III) nitrate nonahydrate (Aldrich, 98%), cobalt(II)
nitrate hexahydrate (Acros Organics, 99%), (3-aminopropyl)
triethoxysilane (APTES, Aldrich, 99%), n-propylamine (Aldrich,
98%), cyclohexane (Fluka, 99%) and n-hexane (C. Erba, 95%)
were all used as received.
Samples preparation
Synthesis of mesoporous SiO2 nanoparticles. The synthesis of
mesoporous silica nanoparticles was adapted from the procedure
by Qiao et al.34
Distilled H2O (145.2 mL, 8.07 mol), EtOH (22.8 mL, 0.39 mol)
and CTAB powder (5.73 g, 15.72 mmol) were stirred at 60
C.
Following the complete dissolution of CTAB, as indicated by the
transparency of the solution, 1.25 mL (18.08 mmol) of a 28 wt%
NH3 solution were added into the mixture under continued
stirring and heating. After 30 minutes, TEOS (14.6 mL, 64.56
mmol) was added dropwise. Stirring at 60
C was continued for 2
hours, during which the clear solution gradually turned into a
white suspension. The suspension was cooled to room temperature and the solid product was recovered and repeatedly washed
(initially in water and finally in EtOH) with five cycles of
centrifugation (30 min at 9 krpm). The dried product was finally
calcinated in air for 2 hours at 500
C in order to remove the
organic pore template.
Synthesis of FeOx@SiO2 series: samples AD. Four different
samples were prepared with varying nominal fractions of
impregnated iron oxide inside the silica pore network (see Table
2). These fractions (respectively, 6, 17, 24 and 29% for samples A,
B, C, D) correspond to the calculated nominal weight% (on the
total weight of composite material) of Fe3O4 which should form
after the two consecutive thermal treatments. These values are
therefore indicative and considered herein as reference values.
Each sample was obtained by impregnating mesoporous SiO2
nanoparticles (0.3 g) with an aqueous solution (4 mL) of
Fe(NO3)3$9H2O salt, calculated according to the desired
nominal fraction. After stirring the mesoporous SiO2 nanoparticles in iron nitrate solutions overnight, the samples were
dried by removing the solvent under reduced pressure. The
impregnated silica powders were then annealed at 700
C for 12
hours in air (first oxidating thermal treatment) and at the same
temperature for 12 hours in N2 + 5% H2 gas atmosphere (second
reducing thermal treatment).
The respective quantities of Fe(NO3)3$9H2O salt used for the
impregnation of SiO2 to produce samples AD are reported in
Table 1.
Synthesis of CoFe2O4@SiO2 sample. Mesoporous SiO2
nanoparticles (0.3 g) were impregnated with an aqueous solution
(4 mL) of Co(NO3)2$6H2O (0.061 g, 0.21 mmol) and of
Fe(NO3)3$9H2O (0.168 g, 0.42 mmol) salts in accord with the
stoichiometric ratio of Co and Fe required to give the desired
J. Mater. Chem., 2012, 22, 1927619288 | 19277

View Online

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

Table 1 Total quantity of Fe(NO3)3$9H2O used during impregnation

for each sample of FeOx@SiO2 series
Sample

Nominal wt%
of Fe3O4

Total amount of
Fe(NO3)3$9H2O

A
B
C
D

6%
17%
24%
29%

0.101 g, 0.25 mmol

0.328 g, 0.81 mmol
0.491 g, 1.21 mmol
0.655 g, 1.62 mmol

nominal fraction of cobalt-iron oxide (CoFe2O4). This weight

fraction of CoFe2O4 was fixed to 14% of the total composite
weight (CoFe2O4@SiO2). After stirring the suspension overnight, the sample was dried by removing the solvent under
reduced pressure. The impregnated material was annealed at
700
C for 12 hours in air.
Surface functionalisation of nanocomposites with APTES. The
general procedure for surface functionalisation with APTES35 is
described. Surface grafting of APTES was achieved by addition
of a cyclohexane (5 mL) solution containing 2% v/v APTES and
2% v/v n-propylamine to the powdered sample (100 mg) previously treated at 700
C. The suspension was stirred at ambient
temperature for 2 h, after which it was recovered and purified by
repeated (3 times) centrifugation (30 min at 6 krpm), hexane
washing and sonication (30 min). The obtained sample was dried
under reduced pressure (101 mbar for 1 h).

Characterisation methods
Structural characterisation. Morphology and microstructure
of the obtained nanocomposites were studied by X-ray powder
diffraction (XRPD) and TEM. A Philips XPert vertical
goniometer working in BraggBrentano geometry, connected to
a highly stabilised generator, was used for the XRPD measurements. A focusing graphite monochromator and a proportional
counter with a pulse-height discriminator were used. Nickel and a step-by-step
filtered Cu Ka radiation (l 1.54056 A)
technique were employed (steps of 0.05
2q), with collection
times of 10 s per step. The size of the crystallites was evaluated by
Line Broadening Analysis (LBA) using a previously published
method.36
TEM images were taken with a JEOL 3010, operating at
300 kV, equipped with a GATAN (Warrendale, PA, USA) multiscan CCD camera and an Oxford EDS microanalysis detector.
TEM specimens were prepared by ultrasonically dispersing the
powdered samples in ethanol (approximately 10 mg mL1) and
depositing several drops of the suspension on a holey carbon film
grid.
Qualitative evaluation of the chemical composition of samples
was obtained using diffuse reflectance infrared Fourier transform
(DRIFT) spectroscopy (ThermoNicolet Magna-IR). The
powders were dispersed in KBr and the spectra were recorded in
the range 4000500 cm1, at a resolution of 2 cm1 using a KBr
background.
Raman spectra were recorded with a Renishaw inVia microRaman spectrometer using a 50 magnification and the 633 nm
line of a HeNe laser as exciting line. A low intensity, not
19278 | J. Mater. Chem., 2012, 22, 1927619288

exceeding 0.6 mW, was used to avoid the transformation of the

synthesised phases.
Thermogravimetric analysis (TGA) was performed using
alumina as the inert standard (Netzsch STA 409). TGA analyses
were carried out with a temperature profile of 25
C to 1000
C, a
heating rate of 10
C min1, and an air (or N2 + 5% H2) to N2
flow ratio of 1 : 2 (40 : 80 mL min1).
The structural changes as a function of temperature and
atmosphere were investigated by in situ XRD measurements
collected at the Materials Characterisation by X-ray Diffraction
(MCX) beamline37 at the Elettra Synchrotron source (Trieste,
Italy). The instrument was coupled to a home-made furnace that
provides an atmosphere and temperature (RT to 1000
C)
controlled environment for sample powders contained in
capillaries.
Porosimetric analysis was performed through N2 adsorption/
desorption measurements at the gas condensation temperature
(77 K), using a Micromeritics ASAP 2010 instrument. Before the
measurement samples were degassed (103 Torr) at 130
C for 12
hours. Surface area was calculated from the adsorption isotherm
according to the B.E.T. method. Specific pore volume was
calculated at p/p0 0.98.
Magnetic characterisation. Magnetic measurements were performed using a SQUID magnetometer (Quantum Design
MPMS) operating in the 3300 K temperature range with an
applied field up to 5.0 T. The measurements were performed
directly on a pellet made with the dried powder of the sample in
such a way to prevent preferential orientation of the nano-crystallites under the magnetic field. Zero-field cooling (ZFC)
magnetization curves were obtained by cooling the sample under
zero applied field to 5 K, then applying a small magnetic field
(2.5 mT) and measuring while the temperature was increased to
300 K. After cooling the sample to 5 K, field cooling (FC)
magnetization was carried out by increasing the temperature in
the presence of the magnetic field.

Results and discussion

Mesoporous SiO2
Thermogravimetric analysis was performed in an air atmosphere
for the as-obtained dry SiO2 nanoparticles in order to establish

Fig. 1 TG analysis in air atmosphere of dried SiO2 nanoparticles containing CTAB organic surfactant as a pore template.

This journal is The Royal Society of Chemistry 2012

View Online

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

Fig. 2 TEM micrographs of calcined mesoporous SiO2 nanoparticles.

the ideal calcination temperature for removal of the organic pore

directing template. From the TG curve (Fig. 1), 500
C was
chosen as optimal since most of the weight loss occurred prior,
suggesting significant oxidative removal of the organic template.
Hence, treatment at 500
C was expected to give a high degree of
mesoporosity to the SiO2 matrix. The thermograph (Fig. 1)
indicates that increasing the calcination temperature above
500
C (i.e. 800 or 1000
C for example) would promote complete
oxidative removal of the template, however, detrimentally it
would also initiate a sinterisation process that would strongly
decrease the mesoporosity of the material by pore shrinkage.
Therefore, calcination at 500
C was considered to be a good
compromise for significant decomposition of the organic
template whilst avoiding potential sinterisation.
TEM micrographs (Fig. 2) of the calcined SiO2 showed
nanoparticles with a spheroidal shape. The sample depicted in
Fig. 2 indicates the presence of both isolated and aggregated
particles with relative sizes of around 6070 nm diameter. The
mesoporosity of the SiO2 nanoparticles calcined at 500
C,
clearly visible in TEM micrographs, was further characterised
and corroborated by N2 adsorption/desorption measurements.
The N2 adsorption/desorption isothermal curve of SiO2
(Fig. 3) featured the typical shape of a mesoporous material (type
IV isothermal according to IUPAC classification38). From the
same measurement, the surface area of the material could be
evaluated through the B.E.T. method.39 The measured surface
area was determined to be 1040  30 m2 g1, which corresponds
to a relevant level of mesoporosity in the material. The pore size

Fig. 4 Pore size distribution curve of calcined mesoporous SiO2

nanoparticles.

distribution for the calcined SiO2 was determined by using the

B.J.H. method.40 The corresponding curve (Fig. 4) showed a
relatively narrow distribution of pore diameter, with the peak
maximum centred at 2.8 nm, herein considered as the average
value for the pore diameter. The total pore volume of the
material was found to be 1.3 cm3 g1. This latter value further
confirmed the high mesoporosity of the material and the extensive availability of accessible internal surface area for potential
embedding of various materials via impregnation methods.
Samples FeOx@SiO2 AD
The FeOx@SiO2 sample series was prepared with varying
nominal fractions of FeOx impregnated in mesoporous SiO2. On
the basis of pore volume availability, four samples were obtained
by varying the amount of iron salt in the impregnating solutions,
thereby filling the pores to varying degrees. Accordingly, the
respective nominal weight% fractions of Fe3O4 in the samples A
D (composites FeOx@SiO2) were calculated (Table 2).
Two sequential thermal treatments, in air and 5% H2 in N2,
were conducted at 700
C. This temperature was chosen by
careful consideration, taking into account the following physical
properties: the integrity of the magnetic crystals obtained at
700
C and the subsequent availability of SiO2 surface silanols
for post-functionalisation.
XRD analysis was performed on the samples AD after the
initial oxidative thermal treatment at 700
C (Fig. 5). Hereby, the
efficiency of the impregnation method was evaluated through
the structural investigation of the resulting composite material,
comprising amorphous SiO2 and a crystalline iron oxide phase.
A crystalline phase evidenced in the diffraction patterns of all
composite materials, albeit with relatively weak signals in
Table 2 Different nominal fractions of the impregnated material in
FeOx@SiO2 series of samples

Fig. 3 N2 adsorption/desorption isothermal curves of calcined mesoporous SiO2 nanoparticles.

This journal is The Royal Society of Chemistry 2012

Sample

v/v% iron salt/

pores

Nominal wt%
of Fe3O4

A
B
C
D

15%
50%
75%
100%

6%
17%
24%
29%

J. Mater. Chem., 2012, 22, 1927619288 | 19279

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Fig. 5 XRD patterns of: FeOx@SiO2 series of samples after the first
thermal treatment (air atmosphere), calcined mesoporous SiO2, a-Fe2O3
(hematite) reference from the literature (PDF #33-664).

samples A and B, was determined to be a-Fe2O3 hematite (PDF

33-664). The sharp and intense peaks typical of pure a-Fe2O3
hematite are clearly visible in the diffraction patterns of C and D
(Fig. 5). Hence, as would be expected from the oxidative treatment, a single crystalline phase of a-Fe2O3 hematite was found in
samples C and D. Alternatively, in the samples of lowest FeOx
nominal fractions (A and B), the visible peaks are weak and
broad, whereby their identification in the presence of the amorphous SiO2 pattern is rendered difficult. Since all samples were
thermally treated under identical conditions, it is expected that
samples A and B would also contain crystalline a-Fe2O3 hematite. However, the presence of a single crystalline phase in
samples A and B cannot be deduced unequivocally from their
XRD analysis. The width of the diffraction peaks were qualitatively used as an indication of the crystalline phase location with
respect to the mesoporous SiO2 network. Since the FWHM (Full
Width at Half Maximum) of the peaks is inversely proportional
to the crystallites size, the broad peaks observed in samples A
and B suggest that the crystallites are relatively small. The small
dimensions of the crystallites would also be consistent with the
presence of crystalline iron oxide inside the silica pores. The

average pore size determined by B.E.T. was 2.8 nm. Therefore,

assuming adequate impregnation of the salt precursor solutions,
the small pores can potentially limit the growth of the crystals
within to give a regularly structured nanocomposite material. In
contrast, in samples C and D the hematite peaks are sharper,
supporting the formation of larger crystallites which are likely to
exceed the dimensions of the pores. This suggests that at least a
fraction of iron oxide has formed outside the pores, yielding
composite materials with reduced regularity. It was deduced
from XRD analysis that effective impregnation of the mesoporous SiO2 used in this study could be achieved with a loading
amount equal to or below a nominal 50% v/v fraction of iron salt
for the total pore volume (corresponding to a nominal 17%
weight fraction of iron oxide with respect to the total weight). In
our study, exceeding this value proved inefficient on the basis
that a fraction of the loaded material was not contained in the
pore network.
The aforementioned results derived from XRD analysis of the
FeOx@SiO2 series (AD) were further corroborated by TEM
and EDS analyses. Representative TEM micrographs of sample
B demonstrate the presence of dark nanocrystals (ca. 3 nm size)
embedded within the porous SiO2 nanoparticles (Fig. 6). Small
domains of a crystalline phase are clearly visible in Fig. 6c, and
agree with the presence of crystal planes for iron oxide. An EDS
spectrum of the circled area in Fig. 6a is shown in Fig. 6d. The
elemental analysis identified the presence of Si, O and Fe, which
is attributed to the amorphous SiO2 matrix and iron oxide
nanocrystals.
In comparison to sample B, TEM micrographs of sample D
showed a higher density of nanocrystals embedded in the SiO2
matrix and the presence of larger nanocrystals outside the SiO2
pore network (Fig. 7a), in agreement with XRD findings. EDS
elemental analysis (Fig. 7c) confirmed that the larger nanocrystals were also iron oxide.
N2 adsorption/desorption measurements were performed for
samples AD (Fig. 8). They all display the typical shape of a
mesoporous material (type IV isothermal), notwithstanding their
impregnation and subsequent thermal treatments. The curves

Fig. 6 TEM micrographs at different magnifications (ac) and EDS analysis of circled area (d) of sample B.

19280 | J. Mater. Chem., 2012, 22, 1927619288

This journal is The Royal Society of Chemistry 2012

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Fig. 9 Pore size distribution curves of FeOx@SiO2 series of samples.

Fig. 7 TEM micrographs at different magnifications (a and b) and EDS

analysis of circled area (c) of sample D.

Fig. 8 N2 adsorption/desorption isothermal curves of FeOx@SiO2 series

of samples.

demonstrate that the composite retained mesoporosity even with

the presence of embedded iron oxide nanocrystals.
The values for specific surface area and pore volume both
decreased with the increasing iron oxide loading (Table 3).
For example, in sample D they are almost 40% lower than the
original mesoporous SiO2. In contrast, the pore size distributions
did not reveal any relevant changes throughout the series (Fig. 9).
Overall, the porosimetric results for the FeOx@SiO2 series (AD)
proved valuable, since the retained mesoporosity in the loaded
samples could potentially allow for further modification of the
composites. For example, additional loading of luminescent
materials, drugs or biomolecules would allow for the generation
of multifunctional composite materials.

On the basis of XRD, TEM, EDS and porosimetric analyses of

the FeOx@SiO2 series, sample B was selected as optimum in
terms of the regularity and quality of its composite nanostructure. Therefore, further experimental investigation and
characterisation was performed for sample B.
Sample B was further calcined in a reducing atmosphere
(700
C), in order to promote the embedded hematite phase to
magnetite. This transformation was expected following an
experimental confirmation of the hematite to magnetite transition that we obtained by means of synchrotron radiation XRD.
In this latter case, an innovative apparatus was used which allows
for the incorporation of gas flow, temperature variation and
control during the measurement. Hence, the structural behaviour
of a pure hematite bulk sample was examined while heating from
RT to 800
C in a reducing gas flow (N2 + 4% H2). The

diffraction patterns were acquired at 12 keV (l 1.034 A).
Background arises from quartz capillary.
The transformation of pure hematite into magnetite is clearly
shown in Fig. 10, in which the acquired patterns at 400
C, 500
C
and 600
C demonstrate where the phase transition occurred.
On the basis of the aforementioned result for pristine hematite,
we similarly anticipated a complete transformation of hematite
into magnetite for the systems comprising iron oxide nanocrystals embedded within the SiO2 matrix. Hence, XRD analysis
(measured with our Philips XPert vertical goniometer hereforth)
was performed for sample B after the first (oxidising) and the
second (reducing) thermal treatment and compared with

Table 3 Surface area and pore volume of FeOx@SiO2 samples AD

Sample

Surface area
(m2 g1)

Pore volume
(cm3 g1)

A
B
C
D

900  20
720  10
690  10
650  10

1.2
1.0
0.9
0.8

This journal is The Royal Society of Chemistry 2012

Fig. 10 Synchrotron radiation XRD patterns of a pure crystalline

hematite sample. Transformation to magnetite between 400 and 600
C
by annealing in N2 + 4% H2 atmosphere.

J. Mater. Chem., 2012, 22, 1927619288 | 19281

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Fig. 11 XRD patterns of: sample B after the first (oxidising) and the
second (reducing) thermal treatment, Fe3O4 (magnetite), g-Fe2O3
(maghemite) and a-Fe2O3 (hematite) references from the literature (PDF
#19629, #39-1346, #33-664 respectively).

the patterns of three pure iron oxide phases from the literature
(Fig. 11). It is evident from the diffraction patterns of sample B
that the anticipated phase transition for SiO2 embedded hematite
into magnetite did not occur (Fig. 11). Furthermore, the peaks
corresponding to the iron oxide phase are broad, thus increasing
the difficulty in their identification. A comparison of the patterns
for sample B following its respective oxidative and reductive
treatments shows that both patterns are consistent with the
presence of either magnetite or maghemite g-Fe2O3 (Fig. 11).
Distinction between magnetite and maghemite g-Fe2O3 is
often difficult since their diffraction patterns are extremely
similar. Furthermore, when the diffraction peaks are broad and
the material is embedded in SiO2, as in this material, it is difficult
to distinguish them from the diffraction pattern of hematite.
Therefore, we cannot exclude nor affirm the presence of hematite
in both instances. However, it is possible to suggest that
magnetite and/or maghemite were present prior to the reductive
treatment, contrary to the initial deduction made for samples A
and B, in which the presence of hematite was assumed on the
basis of its predominance in samples C and D (Fig. 5). The
consolidated XRD findings indicate that formation of magnetite
and/or maghemite is favoured inside the SiO2 pores during the
oxidising thermal treatment, while the opposite occurs for the
iron oxide phase outside the pores in samples C and D, which
maintain the prevalent hematite even after the subsequent
reducing thermal treatment. The preferential formation of
magnetite/maghemite inside the host matrix during the first
annealing in air could be due to the limited size of the crystals or
the presence of some organic impurities contained in the SiO2
(not totally eliminated at 700
C) which produced a reduction
oxidation reaction. The reduction of the initial Fe3+ ions to
magnetite, due to the carbonaceous species arising from the
organic fraction, is followed by the formation of maghemite via
oxidation. This behaviour has already been noted in the
literature.41
In sample B we can hypothesise the presence of both magnetite
and maghemite being embedded in SiO2 following sequential
oxidising and reducing thermal treatments, and that the reducing
treatment was less effective on the crystalline phase embedded in
the composite, compared with the transformation seen for pure
19282 | J. Mater. Chem., 2012, 22, 1927619288

Fig. 12 Raman spectra of sample B after the first (oxidising) and the
second (reducing) thermal treatment.

hematite samples (as already described with synchrotron radiation XRD). However, qualitative Raman spectroscopy performed for sample B following the respective oxidising and
reducing treatments did not unequivocally confirm the hypothesis (Fig. 12).
The Raman spectrum of sample B (Fig. 12), recorded after the
first oxidising treatment, shows a prominent band with a large
width at about 720 cm1 and two weak bands of small intensity
at 227 and 295 cm1. The band at 720 cm1 is also visible in the
spectrum recorded after the second reducing thermal treatment,
while the two weaker bands disappeared. A very broad band
visible between 300 and 500 cm1 in both spectra can be assigned
to the SiO2. The weak bands at 227 and 295 cm1 can be
attributed to the hematite phase, while their disappearance in the
second spectrum may be promoted by the reducing thermal
treatment. The origin of the band around 720 cm1 is not
unequivocally assigned. It is known that magnetite shows a
prominent broad band at 660 cm1, whereas maghemite shows a
broad band at 700 cm1 together with two other broad bands at
350 and 500 cm1.42
On the basis of literature data, the attribution of the band at
720 cm1 to maghemite is unlikely. However, in this study it must
be taken into account that the iron oxide nanoparticles are very
small and embedded in a confined environment. Such confinement can create strains in the nanocrystals and consequently the
vibrational modes can show differences with respect to a bulk
sample. Since an increase in their vibrational frequencies is
reasonable, it is possible to assign the band to maghemite, while a
minor contribution of magnetite could also be present under the
large band between 600 and 700 cm1. In terms of crystallinity
and structure, the two respectively oxidised and reduced samples
of B appear similar. The most notable difference is the minor
presence of hematite after the oxidative annealing, which
disappears following the reductive treatment. Raman analysis
confirmed that the reducing thermal treatment readily transformed the minute hematite fraction into maghemite and/or
magnetite.
The magnetic properties of sample B following the sequential
oxidative and reductive treatments were investigated. A preliminary check by visual means was performed by suspending the
powders in ethanol and placing them in the proximity of a strong
magnet for several minutes (Fig. 13). The samples showed
This journal is The Royal Society of Chemistry 2012

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Fig. 13 Left side photos: images of sample B after the first thermal treatment (in air atmosphere), with powders dispersed in EtOH (a) and in the
presence of a magnet (b); right side photos: images of sample B after the second thermal treatment (in N2/H2 atmosphere), with powders dispersed in
EtOH (c) and in the presence of a magnet (d).

comparable magnetic behaviour, as in both cases the total visible

amount of powder migrated towards the magnet within one
hour.
The magnetic properties of the sample B nanocomposites,
after oxidising and reducing thermal treatments, were assessed by
measuring their hysteresis loops. The loops measured at 3 K and
300 K are respectively shown in Fig. 14(a) and (b). The magnetic
properties of the two samples are quite similar, demonstrating
open hysteresis loops that do not close up to magnetic fields as
high as 3 T, and that the density of the magnetization does not
reach saturation even at magnetic fields of 5 T. The magnetic
moments at 5 T for the two sample B composites are several emu
g1, increasing from 2.0 emu g1 to 2.4 emu g1 from the first to
the second thermal treatment. Considering that the composites
contain a nominal 17 wt% of Fe3O4, it implies that the densities
of magnetization at 5 T for the iron oxide embedded in sample B
after the first and the second treatment are 11.7 and 14.0 emu
g1, respectively. Hence, the increase in magnetization after the
reducing thermal treatment is equivalent to 20%. Comparing
these values with those of bulk magnetite (92 emu g1) or
maghemite (82 emu g1),43 we deduce that the amount of ferrimagnetic iron oxides is between 13% and 17%, in accordance
with the oxide (sample B) or the chosen reference values used for
the estimation. The inferiority of these magnetization values with
respect to the bulk may be attributed to the presence of paramagnetic free Fe ions or very small clusters.
The temperature dependence of the magnetization of paramagnetic ions follows Curies law and hence the magnetic
contribution of these ions should be negligible at room temperature. In fact, at this temperature, the densities of the magnetization are 50% smaller than those at 3 K. Despite the low
magnetization values, we observe that in the hysteresis loop
measured at 3 K the magnetizations do not reach saturation at
high fields and that high magnetic fields are required to close the
loops.
These observations suggest that the nanoparticles exhibit a
spin glass magnetic behaviour instead of an appropriate single
domain ferrimagnetic behaviour.
Such spin glass magnetic behaviour is due to spin frustration
of the antiferromagnetic structure of these oxides arising from
This journal is The Royal Society of Chemistry 2012

structural defects and from the surface of small nanoparticles.4446 This behaviour would also give rise to a further
decrease in the magnetization.
In the loops measured at 3 K and 300 K, the densities of
magnetization were shown to be about 2030% larger in sample
B after the second thermal treatment (Fig. 14). This increase in
magnetization was investigated by analysing the superparamagnetic behaviour of the nanoparticles.

Fig. 14 Hysteresis loops of the B sample after the 1st and 2nd thermal
treatment measured at 3 K (a) and 300 K (b).

J. Mater. Chem., 2012, 22, 1927619288 | 19283

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Superparamagnetism is the magnetic behaviour that single

domain magnetic nanoparticles can exhibit.47,48 It is characterised by a lack of irreversible processes, in particular open
loops, whereby the magnetic behaviour resembles that of a
paramagnetic material. This occurs because the thermal activation processes produce a faster relaxation of the magnetization.
Consequently, it is observed only above a certain temperature
known as the blocking temperature (TB). Below this temperature
the loops exhibit hysteresis, which disappears above TB. This
behaviour was evident in the two samples (Fig. 14). The hysteresis loops at 3 K are clearly open and with coercive fields of
0.33 T, while at room temperature (Fig. 14(b)) no hysteresis is
observed. The TB depends on the particle size and on the
magnetic properties of the nanoparticles according to the
expression:4850
TB KefV/ln(s0/si)kB

(1)

where Kef is the effective magnetic anisotropy of the nanoparticles, V corresponds to the particle volume, s0 is the attempt
time of the nanoparticles, si is the measurement time and kB is the
Boltzmann constant.
The TB can be determined from the measurement of the
temperature dependence of the ZFC and FC magnetizations that
is represented in Fig. 15. The maximum in the ZFC curve is
characteristic of the average TB of the nanoparticles, while the
temperature at which the ZFC and FC curves join, Tmax,
corresponds to the temperature above which all the nanoparticles
are superparamagnetic.
In sample B, after the first thermal treatment TB 25 K and
Tmax 40 K, while the respective temperatures are larger after
the second thermal treatment TB 30 K and Tmax 80 K.
Hence, from the TB values we deduced that the nanoparticles
size increased (presumably by 20%) following the second treatment, or that their magnetic properties changed (whereby the
anisotropy increased). Alternatively, when nanoparticles are
superparamagnetic the dependence of the magnetization on the
magnetic field follows the Langevin function:
M(H,T) Msat (cotanh(CH/T)  1/(CH/T)  1)

Fig. 15 Temperature dependence of the ZFC and FC magnetization of

the B samples after the 1st and 2nd thermal treatment.

19284 | J. Mater. Chem., 2012, 22, 1927619288

where Msat is the magnetization or moment at the saturation, H

is the magnetic field, T is the temperature and C m/kB, m being
the magnetic moment of the nanoparticles. Fitting the loops
measured at 300 K to this function gave an excellent agreement
(R 0.998) to C (m) values equal to 222 (330 mB) and 250 (370 mB)
for sample B after the first and the second thermal treatment,
respectively. Since the average magnetic moment of the nanoparticles is proportional to the volume and the magnetization
(m MSV), the variation of C (m) can be related to the increase in
the magnetization and/or the volume. It is difficult to unequivocally ascertain the origin of this increase in magnetization. For
example, if we consider the simple hypothesis that the magnetization of the nanoparticles does not change after the second
treatment, that the magnetization is that of the bulk magnetite
(4 mB per unit cell), and that the nanoparticles are spherical, the
calculated particle diameters of 4.5 nm and 4.7 nm (for sample B
after the first and the second thermal treatment, respectively) are
very similar. The minute difference between the two average sizes
is difficult to ascertain by TEM or XRD analysis.
The coercive fields measured at 3 K for sample B after the first
and the second thermal treatment were equivalent, giving 0.33 T.
Considering the classical reversal process of single domain
nanoparticles,48,50,51 the coercive field is proportional to the
effective anisotropy, Kef, and inversely proportional to the
magnetization. As discussed previously, we observed an increase
in the average magnetization of the particles after the reducing
treatment; therefore, we conclude that also Kef should be
increased.
Overall, we note that the magnetic moment of the nanoparticles (and presumably also their volume) increases by 12%
from the first to the second thermal treatment. This increase in
magnetic moment is smaller than the 20% increase observed for
TB (that is dependent on the particle volume), and also smaller
than the 20% increase seen for the density of magnetization. On
the other hand, also Kef could increase. From these observations
we concluded that the second thermal treatment produced a
small improvement in the nanoparticles magnetic anisotropy
and an increase in the amount of magnetic iron oxide nanoparticles present by hematite transformation into magnetite/
maghemite. These conclusions support the observed increase for
TB distribution.
The high coercive field of 0.33 T that was measured for sample
B after the two treatments is significantly high in comparison to
the typical literature values given for iron oxide nanoparticles,
typically below 50 mT. As previously discussed, the coercive field
is proportional to the Kef. This anisotropy can be independently
calculated from the values of TB and particle volume according
to the expression (1). Considering a diameter of 4.5 nm as a
representative value for the nanoparticle volume and ln(s0/si)
25,48,50 we determine that the Kef values are 1.8  105 and 2.1 
105 J m3 after the first and the second thermal treatment,
respectively. These values are one order of magnitude larger than
those of the bulk,43 confirming that the high coercive fields are
due to the large effective anisotropy of these nanoparticles. The
enhancement of the magnetic anisotropy is often linked to
surface effects;45,48 however, in our experiments we observed that
the Kef increased after the second reducing treatment, in which a
weak increase in the particle size occurred. If the anisotropy were
due to surface effects we should observe a decrease of the former.
This journal is The Royal Society of Chemistry 2012

View Online

In fact, as previously discussed, the magnetic behaviour (the

magnetization) of these nanoparticles is mainly due to the frustration in the antiferromagnetic order typical of these oxides.
Since the reducing treatment also modifies the crystal structure,
we believe that the increase of the anisotropy after this treatment
is due to the modification of the magnetic order in the core of the
nanoparticles. We conclude that the magnetic anisotropy, well
observed in the TB, of these nanoparticles, is also linked to their
magnetic structure rather than to surface effects.

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

Sample CoFe2O4@SiO2
An additional nanostructured composite containing embedded
CoFe2O4 nanocrystals in SiO2 was prepared. The nominal fraction was based on that of sample B and adapted accordingly for
the mixed oxide. Specifically, a nominal 30% v/v fraction of salts
with respect to the total pore volume was used to obtain a
nominal 14% weight fraction of cobalt-iron oxide on the total
composite weight.
The experimental XRD pattern of the CoFe2O4@SiO2 sample
clearly shows the amorphous silica profile and the typical peaks
of the CoFe2O4 crystalline phase (Fig. 16). Unlike the previous
sample series, well defined peaks were visible thus allowing a
thorough analysis of the crystal properties. The calculated XRD
pattern was obtained by Rietveld refinement52 using the structure
reported by ICSD 160059. Quantification of the crystalline phase
content was made from the calculated pattern to give 15  1%
weight fraction of cobalt-iron oxide, which is consistent with the
nominal value (14 wt%). The dimension of the crystallites was
determined using the Line Broadening Analysis (LBA). The
average volumetric diameter of crystallites (hDivol) was found to
be 2.6 nm, in agreement with the SiO2 pore size. On this basis,
similar to the findings for the FeOx embedded phases in sample
B, it is reasonable to suggest that the nanocrystals are located
inside the silica matrix.
The successful embedding of CoFe2O4 in the SiO2 matrix was
supported by TEM and EDS analyses (Fig. 17). The TEM
micrographs (Fig. 17a and b) evidence the presence of dark
nanocrystals with an approximate 3 nm size embedded in the
porous SiO2 matrix, and the absence of crystalline material
outside the SiO2. The crystalline planes of the CoFe2O4 domains
are clearly visible in Fig. 17b. In Fig. 17c, an EDS spectrum of the

Fig. 16 XRD patterns of: CoFe2O4@SiO2 sample (experimental and

refined with the Rietveld method), mesoporous empty SiO2, CoFe2O4
reference from the literature (PDF #221086).

This journal is The Royal Society of Chemistry 2012

Fig. 17 TEM micrographs at different magnifications (a and b) and

EDS analysis of the circled area (c) of CoFe2O4@SiO2 sample.

circled area in the first micrograph is reported. The elemental

analysis confirmed the presence of Si, O, Fe and Co attributed to
the SiO2 matrix and the CoFe2O4 nanocrystals.
The porosimetric analysis of CoFe2O4@SiO2 gave comparable
results with those of the previous series. The surface area of
sample CoFe2O4@SiO2 was 770  20 m2 g1 and the pore
volume 1.1 cm3 g1. These values are positioned between the
corresponding values for samples A and B of the FeOx@SiO2
series (see Table 3).
By relating the surface area to the wt% of oxide material in the
composite, we note that a nominal 14% of CoFe2O4 in this
composite is in reasonable agreement with samples A and B of
the FeOx@SiO2 series, which have nominal FeOx wt% fractions
of 6% and 17%, respectively.
The identification of CoFe2O4@SiO2 by XRD analysis was
further corroborated by Raman spectroscopy, which evidenced
the typical features of the CoFe2O4 tetragonal phase (Fig. 18).53
As for sample B of the FeOx@SiO2 series, the magnetic
properties of CoFe2O4@SiO2 were examined. A preliminary
check by visual means was performed by suspending the powder
in ethanol and placing a strong magnet in close proximity for
several minutes (Fig. 19).

Fig. 18 Raman spectrum of the CoFe2O4@SiO2 sample.

J. Mater. Chem., 2012, 22, 1927619288 | 19285

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

View Online

Fig. 21 Temperature dependence of the ZFC and FC magnetization of

the CoFe2O4@SiO2 composite material.
Fig. 19 Images of the CoFe2O4@SiO2 sample with powder dispersed in
EtOH (a) and in the presence of a magnet (b).

The magnetic properties of CoFe2O4@SiO2 appeared superior

with respect to sample B, since the total content of visible powder
migrated towards the magnet within several minutes.
The hysteresis loops of the CoFe2O4@SiO2 composite material
were measured at 3 K and 300 K (Fig. 20).
The first measurement (3 K) shows that the density of
magnetization at 5 T is 8 emu g1. In comparison with the
respective curves of sample B (Fig. 14), the density of magnetization for CoFe2O4@SiO2 is larger and the shape of the loop is
completely different. The loop of the latter exhibits a large
coercive field of 1.16 T, and rather than being fully closed at 5 T,
it demonstrates a minor loop suggesting that a percentage of
nanoparticles exhibit anisotropy fields larger than 5 T (Fig. 20).
Alternatively, the temperature dependence of the ZFC and FC
magnetizations (Fig. 21) indicates that the blocking temperatures
of these nanoparticles are much larger than those of the FeOx
nanoparticles. The maximum in the ZFC, TB, is at 177 K, while
the ZFC and FC curves separate at room temperature indicating
that a percentage of nanoparticles are blocked. In fact, the
hysteresis loop exhibits a coercive field of 5 mT. Both the ZFC
FC curves and the hysteresis loop show that the 3 nm nanoparticles have a large effective anisotropy, at least 7 times larger
than that of the FeOx nanoparticles. This high anisotropy is a

fingerprint for the presence of high anisotropic Co2+ ions in the

octahedral sites of the ferrite spinel structure,28,30,54 being the
ferrite with the largest anisotropy.29,30,32 The anisotropy values
can depend on many correlated factors2932,55,56 such as the cobalt
content, the inversion degree of the cobalt, the chemical state of
iron, the order and disorder of the magnetic structure and on the
particle size. Notwithstanding the varying influences, it is always
very large. In these nanoparticles, despite the small 3 nm size, the
intrinsic anisotropy is dominated by surface contributions.
The density of magnetization at 3 K was accordingly corrected
by considering the mass contribution of SiO2. Hence, the density
of the magnetization for the CoFe2O4 nanoparticles at 5 T is
57 emu g1, which is smaller than that of the bulk (92 emu
g1).31 At room temperature the density of magnetization for
the composite is 4.9 emu g1, and the density of magnetization
for the magnetic fraction is equivalent to 35 emu g1. Hence, the
density of magnetization decreases by 38% when increasing the
temperature from 3 K to 300 K. Such a lowering of the magnetization density with increasing temperature was also observed in
the FeOx nanoparticles, and was attributed to free ions, clusters
and the spin-glass behaviour in the nanoparticles. However, at
room temperature the magnetic content for the FeOx nanoparticles was as low as 68%, while that of the CoFe2O4 nanoparticles was close to 40%. This further confirms the smaller
content of non-ferrimagnetic species observed in the CoFe2O4
composite, while the larger density of magnetization can explain
the greater efficiency observed in the magnetic separation of this
material (Fig. 19).

Surface functionalisation with APTES

Embedding the iron oxide nanocrystals within an amorphous
SiO2 matrix not only creates a protective coating for the
magnetic particles, it also provides a layer of superficial hydroxyl
groups which may be further functionalised via a suitable postgrafting method. The availability and type of such superficial
silanol groups vary according to the thermal treatment to which
the samples are subjected. The surface of amorphous silica is
known to contain three types of silanols: isolated, geminal and
Fig. 20 Hysteresis loops of the CoFe2O4@SiO2 composite measured at
3 K and 300 K.

19286 | J. Mater. Chem., 2012, 22, 1927619288

The bulk density of magnetization depends on the Co/Fe ratio and site
occupancy, but typically the values are larger.

This journal is The Royal Society of Chemistry 2012

View Online

hydrogen-bonded.57 Activation of amorphous silica by thermal

treatment in the range 100180
C removes the hydrogen-bonded
water molecules from the surface,58 thus rendering the surface
more reactive towards functionalisation by exposing the isolated
and geminal silanols which can form covalent linkages with
silane coupling agents. Further treatment of silica at extremely
high temperatures (>727
C) promotes complete dehydroxylation of the surface silanols and produces siloxane groups
according to reaction:57

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

SiOH + SiOH / SiOSi + H2O

In the case of our samples, complete dehydroxylation would be

undesirable for the following reasons: (1) the silica surface would
lose its ability to undergo further modification in the absence of
superficial silanols; and (2) the mesoporosity would be compromised if thermal treatment were to exceed 800
C, as discussed
before. Therefore, we established that annealing at 700
C for
12 h was suitable to maintain the integrity of the structural and
magnetic properties without causing complete surface
dehydroxylation.
In order to functionalise the silica surface for potential
immobilisation of biological molecules the widely used silane
coupling agent which contains a terminal amine group (3aminopropyl)triethoxysilane (APTES) was selected. APTES has
already demonstrated broad utility for the covalent coupling of
proteins onto the surface of silica materials.5963 Surface functionalisation with APTES was achieved using the n-propylamine
catalysed method reported in our previous study.35
The presence of silanol groups on the surface of silica in
sample B was monitored by performing DRIFT IR spectroscopy
before and after the functionalisation reaction with APTES
(Fig. 22). The presence of free silanols, which include isolated
and geminal groups, is evident from the sharp peak observed at
3750 cm1 (supported by the literature)35,57,58,64 in the DRIFT
spectra of the non-functionalised composite material (curve in
the middle), which is similar to that of non-impregnated mesoporous silica (curve at the top). The additional signals
observed for non-impregnated mesoporous silica and for

Fig. 22 DRIFT spectra of calcined mesoporous SiO2 and of sample B

before and after functionalisation with APTES. Circled areas point out
the sharp signal of active sites for functionalisation (right) and the weak
signals of aliphatic groups introduced with APTES (left).

This journal is The Royal Society of Chemistry 2012

non-functionalised sample B treated at 700
C are in accordance

with the expected spectral features of a calcinated silica. These
signals include the absorption bands in the region of 1600
400 cm1 which are attributed to the vibrational modes of the
siloxane links.64,65 The asymmetric stretching (AS) vibrational
mode of the siloxane bridges is visible in the region 1200
1000 cm1, while the symmetric stretching (SS) of the O atom
along a line which bisects an axis formed by two Si atoms is
notable at 810 cm1 (as opposed to 802 cm1 which would
characterise the SS vibrational mode in un-activated silica
sample).35,64,65
The attachment of APTES on the silica surface in sample B
was qualitatively confirmed by IR, whereby the sharp peak at
3750 cm1 was no longer present providing evidence for functionalisation (curve at the bottom). Furthermore, the introduction of aliphatic groups attributed to APTES was evidenced by
the appearance of weak signals just below 3000 cm1. The
samples were thoroughly washed via repeated centrifugation and
redispersion in organic solvent in order to remove any residual or
unbound APTES, and subsequently dried in vacuo prior to IR
measurements being recorded.
As for sample B of the FeOx@SiO2 series, the surface functionalisation of sample CoFe2O4@SiO2 with APTES was investigated through DRIFT IR spectroscopy. Similarly, the
attachment of APTES on the SiO2 surface was qualitatively
confirmed by IR (not shown). It should be noted that the presence of iron oxide or cobalt-iron oxide in the composite nanostructured systems cannot be detected through IR spectroscopy.
Therefore, the DRIFT IR spectra of the CoFe2O4@SiO2 sample
before and after its functionalisation are comparable to the
corresponding spectra of sample B reported in Fig. 22.

Conclusions
The structural and magnetic properties of mesoporous silica
nanoparticles embedded with iron oxide and cobalt-iron oxide
nanocrystals were examined. A series of FeOx@SiO2 composite
materials with increasing nominal fractions of iron oxide were
prepared via a wet impregnation procedure. The composites were
thermally treated sequentially, under oxidising and reducing
conditions. Elucidation of their phase composition, structural
and morphological homogeneity was achieved by XRD, TEM,
IR and Raman spectroscopy. The composite with a nominal 17
wt% of iron oxide was considered optimal, since the obtained
mixture of crystalline and magnetic Fe3O4 (magnetite) and
g-Fe2O3 (maghemite) was completely embedded in the silica
matrix. The magnetic properties of the composite were
examined, before and after the reducing treatment. An
enhancement in the density of magnetization was observed
following the reducing treatment, and attributed to an increase in
magnetic material content. Similarly, a CoFe2O4@SiO2
composite containing a nominal 14 wt% mixed oxide was
prepared. The composite showed comparable structural and
morphological homogeneity with the optimal sample from the
FeOx@SiO2 series. Rietveld analysis of the XRD pattern supported that crystallite dimensions were in close agreement with
the pore size, while quantification of the CoFe2O4 phase agreed
with the nominal value. The magnetic properties of the
CoFe2O4@SiO2 composite were found to be superior to those of
J. Mater. Chem., 2012, 22, 1927619288 | 19287

View Online

the FeOx@SiO2 material. The higher density of magnetization in

the former was attributed to a higher content of magnetic
material in the crystalline phase. The facile surface functionalisation of the FeOx@SiO2 and CoFe2O4@SiO2 composites was
demonstrated using APTES.

Acknowledgements

Downloaded by Universita di Padova on 24 August 2012

Published on 25 July 2012 on http://pubs.rsc.org | doi:10.1039/C2JM32314A

Mr Tiziano Finotto and Mr Loris Bertoldo are gratefully

acknowledged for conducting the XRD measurements. The
authors also thank Mrs Martina Marchiori for the N2 adsorption/desorption measurements.

Notes and references

1 E. Katz and I. Willner, Angew. Chem., Int. Ed., 2004, 43, 60426108.
2 Q. A. Pankhurst, N. K. T. Thanh, S. K. Jones and J. Dobson, J. Phys.
D: Appl. Phys., 2009, 42, 224001.
3 E. Katz, Y. Weizmann and I. Willner, J. Am. Chem. Soc., 2005, 127,
91919200.
4 D. Wang, J. He, N. Rosenzweig and Z. Rosenzweig, Nano Lett., 2004,
4, 409413.
5 M. Arruebo, R. Fern
andez-Pacheco, M. Ibarra and J. Santamar
a,
Nano Today, 2007, 2, 2232.
6 P. Trivedi and L. Axe, Environ. Sci. Technol., 2000, 34, 22152223.
7 S. Bucak, D. A. Jones, P. E. Laibinis and T. A. Hatton, Biotechnol.
Prog., 2003, 19, 477484.
8 A. Moser, K. Takano, D. T. Margulies, M. Albrecht, Y. Sonobe,
Y. Ikeda, S. Sun and E. E. Fullerton, J. Phys. D: Appl. Phys., 2002,
35, R157R167.
9 M. C. Wiltshire, J. B. Pendry, I. R. Young, D. J. Larkman,
D. J. Gilderdale and J. V. Hajnal, Science, 2001, 291, 849851.
10 O. Bomat
-Miguel, M. P. Morales, P. Tartaj, J. Ruiz-Cabello,
P. Bonville, M. Santos, X. Zhao and S. Veintemillas-Verdaguer,
Biomaterials, 2005, 26, 56955703.
11 A. Ito, K. Tanaka, K. Kondo, M. Shinkai, H. Honda, K. Matsumoto,
T. Saida and T. Kobayashi, Cancer Sci., 2003, 94, 308313.
12 I. Rabias, D. Tsitrouli, E. Karakosta, T. Kehagias,
G. Diamantopoulos, M. Fardis, D. Stamopoulos, T. G. Maris,
P. Falaras and N. Zouridakis, et al., Biomicrofluidics, 2010, 4,
024111, DOI: 10.1063/1.3449089.
13 P. Tartaj, M. D. P. Morales, S. Veintemillas-Verdaguer, T. Gonz
alezCarre~
no and C. J. Serna, J. Phys. D: Appl. Phys., 2003, 36, R182R197.
14 A. Jordan, R. Scholz, P. Wust, H. Fahling and R. Felix, J. Magn.
Magn. Mater., 1999, 201, 413419.
15 A.-H. Lu, E. L. Salabas and F. Sch
uth, Angew. Chem., Int. Ed., 2007,
46, 12221244.
16 F. Iskandar, I. W. Lenggoro, T. O. Kim, N. Nakao, M. Shimada and
K. Okuyama, J. Chem. Eng. Jpn., 2001, 34, 12851292.
17 W. St
ober, A. Fink and E. Bohn, J. Colloid Interface Sci., 1968, 26,
6269.
18 T. Tago, T. Hatsuta, K. Miyajima, M. Kishida, S. Tashiro and
K. Wakabayashi, J. Am. Ceram. Soc., 2002, 85, 21882194.
19 Y. Lu, Y. Yin, B. T. Mayers and Y. Xia, Nano Lett., 2002, 2, 183186.
20 C. Graf, D. L. J. Vossen, A. Imhof and A. van Blaaderen, Langmuir,
2003, 19, 66936700.
21 A. P. Philipse, M. P. B. van Bruggen and C. Pathmamanoharan,
Langmuir, 1994, 10, 9299.
22 C. Chan
eac, E. Tronc and J. P. Jolivet, J. Mater. Chem., 1996, 6,
19051911.
23 W. Zhao, J. Gu, L. Zhang, H. Chen and J. Shi, J. Am. Chem. Soc.,
2005, 127, 89168917.
24 P. Tartaj, T. Gonz

alez-Carre~
no and C. J. Serna, Adv. Mater., 2001,
13, 16201624.
25 L. Casas, A. Roig, E. Rodr
guez, E. Molins, J. Tejada and J. Sort, J.
Non-Cryst. Solids, 2001, 285, 3743.
26 M. Zayat, F. del Monte, M. del Puerto Morales, G. Rosa,
H. Guerrero, C. J. Serna and D. Levy, Adv. Mater., 2003, 15, 1809
1812.
27 T. Nakamura, Y. Yamada and K. Yano, J. Mater. Chem., 2006, 16,
24172419.

19288 | J. Mater. Chem., 2012, 22, 1927619288

28 J. C. Slonczewski, Phys. Rev., 1958, 110, 13411348.

29 S. Verma and D. Pravarthana, Langmuir, 2011, 27, 1318913197.
30 R. M. Bozorth, E. F. Tilten and A. J. Williams, Phys. Rev., 1955, 99,
17881798.
31 E. Fantechi, G. Campo, D. Carta, A. Corrias, C. de Juli
an Fern
andez,
D. Gatteschi, C. Innocenti, F. Pineider, F. Rugi and C. Sangregorio,
J. Phys. Chem. C, 2012, 116, 82618270.
32 Q. Song and Z. J. Zhang, J. Phys. Chem. B, 2006, 110, 1120511209.
33 G. Baldi, D. Bonacchi, C. Innocenti, G. Lorenzi and C. Sangregorio,
J. Magn. Magn. Mater., 2007, 311, 1016.
34 Z.-A. Qiao, L. Zhang, M. Guo, Y. Liu and Q. Huo, Chem. Mater.,
2009, 21, 38233829.
35 A. Parma, I. Freris, P. Riello, F. Enrichi, D. Cristofori and
A. Benedetti, J. Lumin., 2010, 130, 24292436.
36 S. Enzo, S. Polizzi and A. Benedetti, Z. Kristallogr., 1985, 170, 275287.
37 A. Lausi, E. Busetto, M. Leoni and P. Scardi, Synchrotron Radiation
in Natural Science, 2006, vol. 5, pp. 100104.
38 F. Rouquerol, J. Rouquerol and K. Sing, Adsorption by Powders &
Porous Solids. Principles, Methodology and Applications, Academic
Press, London, 1999.
39 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc., 1938, 60,
309319.
40 E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem. Soc.,
1951, 73, 373380.
41 F. del Monte, M. P. Morales, D. Levy, A. Fernandez, M. Oca~
na,
A. Roig, E. Molins, K. OGrady and C. J. Serna, Langmuir, 1997,
13, 36273634.
42 D. L. A. de Faria, S. Ven^ancio Silva and M. T. de Oliveira, J. Raman
Spectrosc., 1997, 28, 873878.
43 J. M. D. Coey, Magnetism and Magnetic Materials, Cambridge
University Press, Cambridge, U.K., 2010.
44 R. H. Kodama and A. E. Berkowitz, Phys. Rev. B: Condens. Matter
Mater. Phys., 1999, 59, 63216336.
45 Surface Effects in Magnetic Nanoparticles, ed. D. Fiorani, Springer,
New York, 2005.
46 L. Machala, R. Zboril and A. Gedanken, J. Phys. Chem. B, 2007, 111,
40034018.
47 C. P. Bean and J. D. Livingston, J. Appl. Phys., 1959, 30, 120S129S.
48 J. L. Dormann, D. Fiorani and E. Tronc, Adv. Chem. Phys., 1997, 98,
283494.
49 L. N
eel, Ann. Geophys., 1949, 5, 99136.
50 M. Knobel, W. C. Nunes, L. M. Socolovsky, E. De Biasi,
J. M. Vargas and J. C. Denardin, J. Nanosci. Nanotechnol., 2008, 8,
28362857.
51 E. C. Stoner and E. P. Wohlfarth, Philos. Trans. R. Soc. London, 1948,
240, 599642.
52 H. M. Rietveld, J. Appl. Crystallogr., 1969, 2, 6571.
53 Z. Wang, R. T. Downs, V. Pischedda, R. Shetty, S. K. Saxena,
C. S. Zha, Y. S. Zhao, D. Schiferl and A. Waskowska, Phys. Rev.
B: Condens. Matter Mater. Phys., 2003, 68, 094101.
54 L. D. Tung, V. Kolesnichenko, D. Caruntu, N. H. Chou,
C. J. OConnor and L. Spinu, J. Appl. Phys., 2003, 93, 74867488.
55 M. Sorescu, A. Grabias, D. Tarabasanu-Mihaila and
L. Diamandescu, J. Appl. Phys., 2002, 91, 81358137.
56 D. Peddis, C. Cannas, G. Piccaluga, E. Agostinelli and D. Fiorani,
Nanotechnology, 2010, 21, 125705.
57 E. A. Wovchko, J. C. Camp, J. A. Glass Jr and J. T. Yates Jr,
Langmuir, 1995, 11, 25922599.
58 L. L. Hench and J. K. West, Chem. Rev., 1990, 90, 3372.
59 Immobilized Biomolecules in Analysis A Practical Approach, ed. T.
Cass and F. S. Ligler, Oxford University Press, New York, 1998.
60 Immobilized Cells & Enzymes: A Practical Approach, ed. J.
Woodward, IRL Press, Oxford, U.K., 1985.
61 A. Subramanian, S. J. Kennel, P. I. Oden, K. B. Jacobson,
J. Woodward and M. J. Doktycz, Enzyme Microb. Technol., 1999,
24, 2634.
62 Protein Immobilization: Fundamentals and Applications, ed. R. F.
Taylor, Marcel Dekker, New York, 1991.
63 W. Tisher and F. Wedekind, Topics in Current Chemistry
Immobilized Enzymes: Method and Applications, Springer-Verlag,
Berlin, 1999.
64 A. G. Pelmenschikov, G. Morosi and A. Gamba, J. Phys. Chem.,
1991, 95, 1003710041.
65 N. A. Dhas and A. Gedanken, J. Phys. Chem. B, 1997, 101, 9495
9503.

This journal is The Royal Society of Chemistry 2012