Synthesis of Co3[Fe(CN)6]2 molecular-based nanomagnets in MSU mesoporous silica by integrative chemistry

Article abstract of DOI:10.1039/b9nj00271e

Nanometric particles of Co₃[Fe(CN)₆]₂ magnetic cyano-bridged polymers have been synthesised by the combination of the direct synthesis of mesoporous MSU-type silica and the parallel reaction of cobalt nitrate with potassiu

Full text of DOI:10.1039/b9nj00271e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Synthesis of Co3[Fe(CN)6]2 molecular-based nanomagnets in MSU
mesoporous silica by integrative chemistryw
Rola Mouawia,a Joulia Larionova,*b Yannick Guari,b Seungwon Oh,a Phil Cooka
and Eric Prouzet*a
Received (in Montpellier, France) 19th June 2009, Accepted 5th August 2009
First published as an Advance Article on the web 22nd September 2009
DOI: 10.1039/b9nj00271e
Nanometric particles of Co3[Fe(CN)6]2 magnetic cyano-bridged polymers have been synthesised
by the combination of the direct synthesis of mesoporous MSU-type silica and the parallel
reaction of cobalt nitrate with potassium ferricyanide. This synthesis, relevant to integrative
chemistry methods, allowed us to confine the particle growth during the building of the
mesostructured framework of silica. The obtained composite materials were studied by XRD,
SAXS, TEM and FT-IR. Magnetic measurements of the nanoparticles extracted after silica
dissolution by HF treatment, confirmed the expected magnetic behaviour of these Prussian Blue
analogues. Compared with bulk materials, a spin-glass like regime is observed due to the presence
of strong interparticle magnetostatic interactions between the nanoparticles which appeared as a
result of the high density of nanoparticles in the nanocomposite materials synthesised by this new
method.
method was defined as nanocasting.7 One of its drawbacks
Introduction
lies in the possibility to achieve the reaction inside the porous
Molecular-based magnets, and among them Prussian blue
framework and avoid it outside. A recent study by Folch et al.
analogues (PBA), are materials that exhibit adjustable properties
reported a smart solution that allowed them to prepare
(magnetic, optic, photo-switchable) based on their molecular
cyano-bridged polymer nanoparticles with the general formula
Mn+/[M0(CN)m]xꢀ inside SBA-15 or MCM-41 mesoporous
structure and the possibility to modify them by the correct
design of molecular entities where magnetic interactions
silica. To avoid any reaction outside the internal porous
between two different metal centres are controlled by the
structure, the silica was functionalized beforehand by
interatomic distance via a molecular connector such as a
–(CH2)2C5H4N pyridine groups that allowed a selective
anchorage of Mn+ cations and avoided any segregation in
cyano-bridge.1 Beyond the huge development of this class of
compounds, a recent trend in research in this domain has been
excess of these cations. They could react in a second step with
to generate small particles that could combine magnetic
the metallo-cyano entities, the initial grafting density being the
behaviour and nanometric-dependent properties. Since the
limiting factor for the particle growth along with the confine-
first synthesis of Prussian blue nanocrystals in a reverse
ment provided by the porous framework. Despite its great
microemulsion,2 there have been numerous methods investi-
success, this method still has some drawbacks: it requires a
gated, based on reverse micelles,3 polymers,4 ionic liquids5 or
multistep process that starts with the initial grafting of
porous silica.6
pyridine, its metal complexation via
a succession of
Among the different methods of synthesis of nanoparticles,
the use of a porous rigid matrix that will limit the growth of
particles nucleated inside has been widely explored. This
impregnations and washing, and the final reaction of the metal
cyano-complexes by similar impregnation and washing. In
addition, a limited quantity of the nanoparticles (around 3–5%)
may be inserted into the silica pores due to the limited quantity
of the pyridine groups that can be grafted.
a University of Waterloo, Chemistry Dept & Institute of
Nanotechnology, 200 University Av., W (ON) N2L 3G1, Canada.
E-mail: eprouzet@scimail.uwaterloo.ca; Fax: +1 519 746-0435;
Tel: +1 519 888-4567 x38172
It has been pointed out recently that new chemical methods
could take advantage of an integration of different shaping
processes that combine sol–gel and/or soft chemistry
techniques with the self-assembly mechanism of soft matter,
potentially helped by chemical engineering processes. This
emerging field was baptised ‘‘Integrative Chemistry’’.8 We
illustrate in the present report such a process by the integration
of different mechanisms in a one-step process that allowed us
to prepare directly nanoparticles of Co3[Fe2(CN)6]2 molecular
magnets inside a matrix of MSU-type mesoporous silica.
Synthesis of MSU mesoporous silica is achieved via a two-
step process where stable hybrid micelles made of nonionic
b Institut Charles Gerhardt Montpellier, UMR 5253
´
CNRS-UM2-ENSCM-UM1, Chimie Moleculaire et Organisation du
Solide, Universite Montpellier II, Place E. Bataillon,
´
F-34095 Montpellier Cedex 5, France.
E-mail: joulia.larionova@univ-montp2.fr; Fax: +33 467-143-852
w Electronic supplementary information (ESI) available: SI 1: FT-IR
spectrum of (a) the bulk sample of Co3[Fe(CN)6]2, and (b) the initial
reagent K3Fe(CN)6. SI 2: EDAX analysis of samples B. SI 3: EDAX
analysis of sample C. SI 4: Field-cooled and zero field-cooled magnetisation
(FC/ZFC) vs. temperature curves performed with an applied field of
10 Oe for sample B; inset: field dependence of the magnetisation for
sample B performed at 2 K. See DOI: 10.1039/b9nj00271e
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009 New J. Chem., 2009, 33, 2449–2456 | 2449

View Article Online
surfactants surrounded by a diffuse layer of silicate oligomers
interacting with the palisade of polyethylene oxide chains, are
created first.9 These ‘‘hybrid micelles’’ can be either further
destabilised by the addition of a fluoride salt that initiates
the condensation of silicate oligomers, or they can be concen-
trated by simple evaporation. In the first process, micrometric
spherical particles of mesoporous silica are obtained,10
whereas evaporation leads to a spontaneous reaction giving
a gel instead of particles. Additionally, the concentration
operation is a way to trap hetero-species such as metal ions
or particles with a high initial homogeneity, since the process
occurs in the liquid state, which leads to a good distribution of
species in the final material. We denote this process as hybrid
micelle concentration (HMC). This HMC method was initially
illustrated with the synthesis of Al-doped silica with a content
of aluminium that could be easily adjusted using this method.11
We illustrate in the following how the HMC process can be
used to combine the synthesis of PBA particles along with the
synthesis of mesoporous silica. The current synthesis reported
here combines two reactions: (i) the synthesis of PBA particles
and (ii) the parallel synthesis of the mesostructured silica that
will help to confine the PBA particle growth within the
nanometer range. This method is fully relevant to Integrative
Chemistry since the final material is obtained via a synergistic
mechanism.8 We report the single-step synthesis of PBA
nanoparticles, generated in parallel with the reaction of a
mesostructured silica framework, that ensures a confinement
mechanism controlling the final size dispersion of molecular
magnets.
was achieved by means of incipient wetness with a reaction
based on methods already reported.12 The metal salts were
dissolved in the minimum volume of deionized water and then
poured onto the silica support before being degassed under
vacuum, at room temperature. First, a solution of cobalt
nitrate (0.8 g of (Co(NO3)2ꢂ6H2O) in 1.2 mL of water at
pH = 2) was poured onto silica and left for 1 h before being
dried at 65 1C. After drying, a solution of potassium ferrihexa-
cyanate (0.29 g of (K3Fe(CN)6) in 0.8 g H2O at pH 2) was
impregnated and let to react for one day. The initial metal
ratio Co/Fe was equal to 1.5 in order to ensure that all
ferricyanate groups could react. The silica was washed with
deionized water and dried at 65 1C.
Nanocomposite prepared by a one-pot method (sample C). In
a typical preparation of PB-MSU (Si/surfactant = 20,
Si/Co = 15, Co/Fe = 1.6), a 0.1 M solution of Brij 98 was
prepared by dissolution of 5.7 g of surfactant in 50 ml of
deionized water at pH 2 followed by the addition of 20.5 g of
TEOS (0.098 mol) and 1.91 g (0.0065 mol) of Co(NO3)2ꢂ6H2O.
These were slowly dispersed in this solution under magnetic
stirring until the solution became perfectly clear. This solution
was poured in a rotovapor and 50 wt% of the total mass was
removed by evaporation at moderate temperature (30–40 1C)
under dynamic vacuum. 1.3 g of K3Fe(CN)6 (0.0039 mol)
(dissolved in the minimum amount of water and Co/Fe = 1.5)
was further added to this concentrated solution. The evaporation
was continued to completion and the solid was recovered and
dried overnight at 65 1C. The material was left as-synthesized
since the residual organics cannot disturb the magnetic properties.
Besides, they can help to protect the molecular magnets
from being affected by the external environment. After
reaction, EDAX analysis led to an average composition of
{Co3[Fe2(CN)6]2}0.07ꢂMSU0.93 (embedded surfactants not
included). The average weight percentage of surfactants
trapped in the silica is 40 wt%.10 Sample C was treated by
HF etching to dissolve silica, which allowed recovery of the
nanoparticles alone. This later sample was named C0.
Experimental
Chemicals
The nonionic PEO-based surfactant Brij 98, potassium hexacyano-
ferrate(III) (K3[Fe(CN)6]), cobalt(II) nitrate (Co(NO3)2ꢂ6H2O),
and tetraorthosilicate (TEOS: Si(OCH2CH3)4) were purchased
from Sigma Aldrich and used as received.
Synthesis
MSU silica (sample A). A typical synthesis of MSU silica
prepared with a Si/surfactant molar ratio of 20. A 0.1 mol Lꢀ1
solution of Brij 98 (C18H35(OCH2CH2)B20OH) was prepared
by dissolution of 11.5 g into 100 mL of deionized water of
pH = 2 and 41.4 g of TEOS was slowly added under magnetic
stirring. A cloudy emulsion was obtained, which became clear
within 15 min, after the total hydrolysis of TEOS. The mild
acidity of the aqueous solution prevented the silica from
precipitating. A rotary evaporator under dynamic primary
vacuum and a heating bath set at 30 1C was used to remove
60–70% of the total mass, which included most of the ethanol
generated by the TEOS hydrolysis. The solution was removed
and left to gel for 48 h in a thermostated shaking bath. After
completion of the condensation, the solid was collected,
ground and surfactants were washed out by ethanol.
Nanocomposite prepared by simple impregnation of a Co-doped
silica (sample D). First, Co-MSU was synthesised using the
one-pot method previously described until Co-doped silica
was obtained that was filtered off, dried and calcined at
450 1C. Second, 0.22 g of K3Fe(CN)6 (Co/Fe = 1.5) was
dissolved in an appropriate volume of deionized water at
pH = 2) and was impregnated into to the Co-silica at room
temperature under vacuum.
All preparations (B–D) exhibit an intense dark purple
colour.
Instrumentation
X-Ray diffraction was performed in transmission mode using
1 mm Lindemann capillaries on a laboratory SAXS apparatus
equipped with a 4 kW copper rotating anode X-ray source
(l = 1.54 A) and a multilayer focusing Osmic monochromator
giving a high flux (108 photons sꢀ1) and punctual collimation,
with a 2D Image Plate as detector. Diffraction patterns are
reported as a function of the wave vector Q = 2p/d = 4psin y/l,
Nanocomposite prepared by double impregnation (sample B).
For comparison, we prepared a sample by simple impregnation
of the mesoporous silica prepared according to the previous
method and further impregnated by precursors of PBA. This
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009
2450 | New J. Chem., 2009, 33, 2449–2456

View Article Online
where d is the correlation length (also called ‘‘d-spacing’’) for
disordered systems.
SEM images were obtained on a Leo 1530 microscope
(U. Waterloo, WATLAB) at a EHT of 5 kV with gold-coated
samples.
Samples for transmission electron microscopy (TEM)
measurements were prepared using extractive replicas or ultra-
microtomy techniques and then deposited on copper grids.
TEM measurements were carried out at 100 kV with a JEOL
1200 EXII microscope. Magnetic susceptibility data were
collected with
a Quantum Design MPMS-XL SQUID
magnetometer working in the temperature range of 1.8–300 K
and the magnetic field range of 0–50 kOe. The data were
corrected for the sample holder. FT-IR spectra were recorded
Fig. 2 TEM image of sample C showing the silica ordered meso-
by transmission on a Bruker Tensor 27 spectrometer.
¨
structure.
Results and discussion
The mesoscopic order of the silica matrix was checked by
X-ray diffraction (Fig. 1). The initial diffraction pattern dis-
plays only one peak characteristic of the 3D wormhole struc-
ture of the MSU-type silica.10,11 The correlation length
between pores is equal to 5.5 nm, which corresponds to the
sum of pore diameter plus silica wall thickness.
Both samples B and D obtained by impregnation (in one or
two steps) lost their mesoscopic order during the preparation.
Sample C obtained by the one-pot method exhibits a well-
defined peak at a distance slightly higher than that of pure
silica (5.7 instead of 5.5 nm). Unlike sample C, surfactants
were washed out from sample A (pure silica), which induces a
slight shrinkage of its structure. A stronger scattered intensity
is also observed for the nanocomposite C, compared with pure
silica. Indeed, pure silica exhibits a more ill-ordered structure
resulting from the ethanol washing. The ordered meso-
structure of C is confirmed by the TEM observation (Fig. 2).
Infrared (IR) spectral measurements of samples B and D in
the region between 2000 and 2200 cmꢀ1 exhibit an intense
peak at about 2121 cmꢀ1 (Fig. 3). This peak is ascribable to
CN stretching in the Fe(III)–CN–Co(II) structure. The mode of
Fig. 3 FT-IR spectra of samples B, C and D (signals have been
shifted).
free CNꢀ ions, initially at n(CN) = 2080 cmꢀ1, is shifted to
higher frequencies when coordinated to metal ions.6,13 We
observe that samples B and D, which exhibit larger crystals
(see Fig. 4), exhibit a single intense peak at 2121 cmꢀ1. The IR
spectrum of sample C obtained by direct synthesis (as well as
for the bulk compound (see Fig. SI 1a, ESIw), reveals a
splitting of the CN peaks into two components at 2107 and
2161 cmꢀ1. This splitting cannot be assigned to a mixture
between the new compound and unreacted K3Fe(CN)6
because the latter exhibits a strong narrow peak at 2118 cmꢀ1
(Fig. SI 1b, ESIw).
This splitting is not totally understood yet and different
explanations have been proposed so far: (i) a possible isomeri-
zation with the formation of Co(II)–CN–Fe(III), not expected
with our synthesis, or (ii) a partial reduction of Fe(III) in Fe(II)
with the formation of Fe(II)–CN–Co(II) bonds.13,14 Another
explanation could be the existence of bridging and dangling
Fig. 1 Powder diffraction patterns obtained for the pure MSU silica
(sample A) and the three nanocomposites (B–D); Q = 2p/d where d is
the correlation length.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009 New J. Chem., 2009, 33, 2449–2456 | 2451

View Article Online
Fig. 5 XRD pattern for a compound synthesised alone (bottom) and
for sample C (top).
leads to a partial leakage of cobalt ions that react onto the
surface of the silica particles, especially at the aperture of
macroporous pores. As a result, rings of crystals are observed
around the macropore openings (Fig. 4(g) and (h)).
The X-ray diffraction pattern (Fig. 5) of sample C confirms
the existence of a ferricyanide-based crystalline structure for
both the bulk synthesis and the synthesis embedded into the
silica matrix. Diffraction peaks are similar to those observed
previously in Co3[(Co(CN)6]2.14 Their average size is in perfect
agreement with the pore diameter expected from the XRD
pattern of silica (Fig. 1).
PBA nanoparticles with a mean size value of 3–4 nm can be
clearly seen after removal of silica from the nanocomposite
material by HF etching (Fig. 6). They are well dispersed and
no aggregates are observed.
Fig. 4 SEM images of the different mesoporous silicas: A (a, b);
B (c, d); C (e, f); D (g, h). The magnification in (d) corresponds to the
square in (c).
The magnetic properties of these samples have been studied
by dc and ac modes by using a SQUID magnetometer working
in the temperature range between 1.8 and 300 K. We report
the magnetic measurements of both samples C and C0
obtained after silica dissolution by HF treatment of C. The
zero field-cooled (ZFC)/field-cooled (FC) magnetisation
curves measured for sample C are displayed in Fig. 7. The
ZFC curve was obtained by recording the magnetisation when
the sample was heated under a field of 10 Oe after being cooled
in zero magnetic field. The FC data were obtained by cooling
the sample under the same magnetic field after the ZFC
experiment and recording the change in sample magnetisation
with temperature. The ZFC and FC curves reveal different
CN groups due to the formation of small particles since the
bulk powder is made of E200 nm crystals. The narrow peak at
1388 cmꢀ1 is assigned to the stretching mode of nitrate CH2
groups present in the surfactant cobalt salt, the broad one at
3500 cmꢀ1 to OH stretching and that at 1616 cmꢀ1 to OH
bending. FT-IR analysis of sample C after freeze-drying did
not reveal any modification in these components, which proves
that water is a structural component of the material.
SEM observation of the four samples (A, B, C, D) is
displayed in Fig. 4. The structure of sample A (pure silica) is
similar to that reported previously for pure silica obtained by
this method, with a monolithic aspect and specific conchoidal
fracture.11 Sample B exhibits a similar structure (Fig. 4(c)) but
additional nanocrystals are observed on the particle faces
(Fig. 4(d)). EDAX analysis confirmed that the silica matrix
does not contain PBA particles and that crystals form mostly
in the fractures on the silica monolith (Fig. SI 2, ESIw). Sample
C obtained by the co-reaction still exhibits a monolithic
structure (Fig. 4(e)) with small particles included in the matrix
(Fig. 4(f)). These particles do not exhibit the facetted shapes of
pure PBA and EDAX analysis confirms that both the matrix
and the particles exhibit the same chemical composition with a
rather constant iron and cobalt content (Fig. SI 3, ESIw) with
Co/Fe E 1.5. Finally, sample D reveals that the impregnation
of silica, which had been impregnated beforehand by cobalt
ions and dried, by an aqueous solution of [Fe(CN)6]3ꢀ ions
Fig. 6 TEM image of the cyano-bridged coordination polymer
nanoparticles (C0) obtained after HF etching of sample C.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009
2452 | New J. Chem., 2009, 33, 2449–2456

View Article Online
In order to determine the magnetic regime of the nano-
particles, alternating current (ac) susceptibility measurements
have been performed. The temperature dependence of the
in-phase, w0 (absorptive), and out-of-phase, w00 (dispersive),
components of the ac susceptibility measured in zero static
field for frequencies varying from 1 Hz to 1488 Hz for the
sample C are shown in Fig. 8. At 1 Hz, both w0 and w00
responses present a peak at 11.9 and 11.3 K, respectively.
These peaks shift toward higher temperatures as the frequency
increases. The sample C0 (nanoparticles alone) shows the same
frequency dependent behaviour as C, with w0 and w00 peaks at
11.5 and 11.0 K at 1 Hz. Such frequency-dependent behaviour
is characteristic of the presence of nanoparticles of cyano-
bridged coordination polymer presenting short range
magnetic ordering.15,16 The magnetic transition observed at
Fig. 7 Field-cooled and zero field-cooled magnetisation (FC/ZFC)
vs. temperature curves performed with an applied field of 10 Oe for the
sample C. Inset: Field dependence of the magnetisation for the sample
C performed at 2 K.
behaviours: the ZFC curve exhibits a maximum at Tmax
=
10.51 K while the FC curve increases as the temperature
decreases and tends to saturation at low temperature. Both
curves coincide at high temperatures and begin to separate at
T
sep = 11.30 K. The FC/ZFC curves performed for the sample
C0 after removing silica from C by HF etching, shows very
similar irreversible behaviour with a maximum on the ZFC
curve observed at 10.80 K (Table 1).
The field dependence of the magnetisation performed at 2 K
for samples C and C0 shows that saturation magnetisation at
50 kOe is equal to 5.34 and 35.13 emu gꢀ1 for C and C0,
respectively.
The last value corresponds to a value of 3.7 mB, relative to
the following molecular formula: Co3[Fe(CN)6]2 (Co/Fe = 1.5
compared with the initial value of 1.6). This value is in
agreement with the value of 3.5 mB calculated for the {Co3Fe2}
unit with antiferromagnetic Co2+–Fe3+ interactions through
the cyano bridges. The value of the saturation magnetisation
of 5.34 emu gꢀ1 observed for the sample C corresponds to a
total amount of 15.2 wt% of nanoparticles into the silica. The
EDAX analysis gave an atomic ratio (Fe + Co)/Si of 22 wt%.
Taking into account the existence of 40 wt% of surfactant in
the as-synthesized material, we obtain a total amount of
13 wt% of nanoparticles, in good agreement with the magnetic
measurements. Both samples (C and C0) exhibit a hysteresis
effect with a coercive field of 2.0 and 1.98 kOe, respectively
(inset of Fig. 6 and Table 1).
Fig. 8 Temperature dependence of (top) in-phase, w0, and (bottom)
out-of-phase, w00, components of the ac susceptibility performed in
zero field for the sample C. Frequencies: 1 Hz (J), 9.99 Hz (K),
125 Hz (&), 998 Hz (’) and 1498 Hz (n).
Table 1 Magnetic data for samples C, C0 and B
Arrhenius lawa
Vogel–Fulcher law
Power law
(Ea/kB)/K t0/s
Tg/K t0/s (Ea/kB)/K Tg/K t0/s
zv
6.9
8.2 11.02
10.1
Sample Tmax/K
Tg/K AT line Hc/kOe
C
C0
B
10.51
10.80
10.92
2589
2115
1262
1.92 ꢃ 10ꢀ88 10.33 1.74 ꢃ 10ꢀ12 23.2
1.21 ꢃ 10ꢀ52
11.75 1.86 ꢃ 10ꢀ13
9.65
2.00
1.98
1.04
3.42 ꢃ 10ꢀ82 10.0
1.44 ꢃ 10ꢀ12 35.2
11.4
11.8
2.17 ꢃ 10ꢀ13
8.9
2.6 ꢃ 10ꢀ12
45.6
2.65 ꢃ 10ꢀ10 10
a
Obtained at the maximum of the ZFC curve.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009 New J. Chem., 2009, 33, 2449–2456 | 2453

View Article Online
low temperature may be attributed to three limiting cases:
(i) the blocking process of isolated or weakly interacting
superparamagnetic nanoparticles usually well described with
an Arrhenius law;17 (ii) nano-cluster-glass-like transition
caused by strong magnetostatic interparticle interactions and
by randomness,18 or (iii) intraparticle spin glass-like regime
due to the particle surface spin disorder.6
First of all, the temperature dependence of the relaxation
time extracted from the maximum of the w00 component of the
ac susceptibility was fitted with an Arrhenius law:
t = t0 exp(Ea/kBT)
(1)
where Ea is the average energy barrier for the reversal of the
magnetisation, t0 is the attempt time and kB is the Boltzmann
constant. The values of the energy barrier, Ea/kB, are equal to
2589 and 2115 K for C and C0, respectively, and the values
of pre-exponential factor, t0, are equal to 1.92 ꢃ 10ꢀ88 and
3.42 ꢃ 10ꢀ82 s (Fig. 8). Usual values of t0 observed for
superparamagnetic systems are in the range of 10ꢀ8–10ꢀ12 s.
Hence, these low values have no physical meaning and they
may be interpreted as the signature of a correlation among
magnetic moments introduced by strong magnetostatic
interparticle interactions. The probable presence of these
interactions can be probed by the temperature dependence
of the relaxation time. This evolution was fitted with the
Vogel–Fulcher law commonly used for magnetically interacting
clusters:19
t = t0 exp(Ea/kB(T ꢀ T0))
(2)
In comparison with the Arrhenius law adopted for isolated
nanoparticles, the interparticle interactions deduced from this
analysis, are introduced by means of an additional parameter,
T0. The best fits for sample C (C0) give satisfactory parameters
of Ea/kB, t0 and T0 equal to 23.2 (35.2) K, 1.74 ꢃ 10ꢀ12
(1.44 ꢃ 10ꢀ12) s and 10.33 (10.0) K (Table 1), suggesting the
presence of relatively strong interparticle interactions in both
cases (Fig. 9(a)).
Fig. 9 (a) Thermal variations of the relaxation time according to the
Arrhenius law, t = t0 exp(Ea/kBT), (&) and the Vogel–Fulcher law,
t = t0 exp(Ea/kB(T ꢀ T0)), (J) for sample C and according to the
Arrhenius law, t = t0 exp(Ea/kBT), (’) and the Vogel–Fulcher law,
t = t0 exp(Ea/kB(T ꢀ T0)), (K) for sample C0. (b) Thermal variation
of the relaxation time according to a power law, t = t0[Tg/(Tmax ꢀ Tg)]zv,
for C (B) and C0 (E).
The trace of the temperature maxima of the w0 and w00
components as a function of H2/3 gives linear dependence that
corresponds to the so-called de Almeida–Thouless (AT) line.21
It is given by the equation:
In order to verify if the dynamics in these systems exhibit
critical slowing down as usually observed in canonical
spin-glass systems presenting strong interparticle interactions,
we used the critical scaling law of the spin dynamics:
t = t0[Tg/(Tmax ꢀ Tg)]zv
(3)
H p (1 ꢀ (Tmax/Tf))3/2
(4)
where Tg (a0) is the glass temperature, f is the frequency and
zv is a critical exponent.20 For the sample C (C0), the best fits
gives the following parameters, Tg = 11.75 (11.40) K, t0 =
1.86 ꢃ 10ꢀ13 ( 2.17 ꢃ 10ꢀ13) s and zv = 6.9 (8.2) (Fig. 9(b),
Table 1). The obtained zv as well as the t0 values are perfectly
in the range 4–12 and B10ꢀ13 s, respectively, expected for
classical spin glass systems.
The spin-glass transition temperature obtained by extrapolation
of the AT line back to H = 0 was found equal to 9.65 K and
11.02 K for C and C0, respectively (Fig. 10). These values are
relatively close to the Tmax values obtained by the maximum
on the ZFC curve, equal to 10.51 K. Usually, the compliance
of the data with the AT line is considered to be an evidence for
the existence of a spin-glass behaviour.22
In order to verify if the samples C and C0 present a
spin-glass regime, the temperature dependences of the ac
susceptibility at 125 Hz were measured with different applied
direct current (dc) fields (Fig. 9). For both samples, the peak
intensity of the two components w0 and w00 decreases as the
field increases. This had been observed previously for
cyano-bridged coordination polymer nanoparticles stabilised
by capping ligands in organic solvents.20
To summarise, all parameters extracted from this magnetic
study indicate the presence of strong magnetostatic intra-
particle interactions that induce a spin-glass regime. It is worth
noting that previously published cyano-bridged coordination
polymer nanoparticles obtained in mesostructured silica by
step-by-step coordination of precursors were isolated from
each other and did not exhibit magnetostatic interactions.3 On
the other hand, the presence of a surface spin frustration
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009
2454 | New J. Chem., 2009, 33, 2449–2456

View Article Online
shows the presence of strong interparticle interactions
(Table 1). This conclusion was confirmed by the power law
t = t0[Tg/(Tmax ꢀ Tg)]zv fitting of the temperature dependence
of the relaxation time giving satisfactory parameters of Tg =
11.8 K, t0 = 2.65 ꢃ 10ꢀ10 s and zn = 10 (Table 1) usually
observed for the strongly interacted nanoparticles presenting
spin-glass behaviour such as the sample C. As for sample D,
the FC/ZFC curves as well as the temperature dependence
of the ac susceptibility show no peak in the temperature
region 1.8–300 K showing that visibly no cyano-bridged
Co2+/[Fe(CN)6]3ꢀ is present.
Conclusions
We successfully synthesised nanoparticles of Prussian Blue
analogues by a direct method based on integrative chemistry
where the synthesis of the particles, Co3[Fe(CN)6]2, and of the
confinement matrix (MSU-type mesostructured silica) are run
in parallel. This method allowed us to obtain a composite
structure with a homogeneous dispersion of a high amount of
PBA nanoparticles (around 15%) with an average size ranging
between 2 and 4 nm. Magnetic measurements confirmed the
occurrence of this reaction and revealed a strong magnetostatic
component due to interparticle interactions. This synthesis,
which does not require any pre-grafting step, is currently
applied to a larger family of Prussian Blue analogues.
References
1 O. Kahn, Philos. Trans. R. Soc. London, Ser. A, 1999, 357, 3005;
M. Verdaguer, A. Bleuzen, V. Marvaud, J. Vaissermann,
M. Seuleiman, C. Desplanches, A. Scuiller, C. Train, R. Garde,
G. Gelly, C. Lomenech, I. Rosenmann, P. Veillet, C. Cartier dit
Moulin and F. Villain, Coord. Chem. Rev., 1999, 190–192, 1023;
O. Kahn, Acc. Chem. Res., 2000, 33, 647; M. Verdaguer,
Polyhedron, 2001, 20, 1115; O. Hatlevik, W. E. Buschmann,
J. Zhang, J. L. Manson and J. S. Miller, Adv. Mater., 1999, 11,
914; S. M. Holmes and G. S. Girolami, J. Am. Chem. Soc., 1999,
121, 5593; B. G. Morin, C. Hahm, A. J. Epstein and J. S. Miller,
J. Appl. Phys., 1994, 75, 5782; J. S. Miller, Adv. Mater., 1994, 6,
322; S. Ferlay, T. Mallah, R. Ouahes, P. Veillet and M. Verdaguer,
Inorg. Chem., 1999, 38, 229; W. E. Buschmann, S. C. Paulson,
C. M. Wynn, M. Girtu, A. J. Epstein, H. S. White and J. S. Miller,
Adv. Mater., 1997, 9, 645; W. R. Entley and G. S. Girolami,
Science, 1995, 268, 397; S. Ferlay, T. Mallah, R. Ouahes, P. Veillet
and M. Verdaguer, Nature, 1995, 378, 701; V. Gadet, T. Mallah,
I. Castro and M. Verdaguer, J. Am. Chem. Soc., 1992, 114, 9213;
T. Mallah, S. Thiebaut, M. Verdaguer and P. Veillet, Science,
1993, 262, 1554; O. Sato, T. Iyoda, A. Fujishima and
K. Hashimoto, Science, 1996, 271, 49; S. Ohkoshi and
K. Hashimoto, J. Am. Chem. Soc., 1999, 121, 10591.
Fig. 10 Temperature dependence of (top) in-phase, w0 and (bottom)
out-of-phase, w0 0, components of the ac susceptibility. Applied field:
100 Oe (J); 250 Oe (K), 500 Oe (&) and 750 Oe (’). Inset: Tmax (w00)
is plotted against H2/3 for the variation of the (top) w0 and (bottom) w00
parts of the ac susceptibility performed with different applied fields for
sample C in silica matrix (K) and after the silica was removed by HF
treatment (J). The solid lines show a linear fit to the data.
induced by the interactions of nanoparticles with the silica
matrix has been assumed. In the present case, it is evident that
the magnetostatic interactions are predominant; however, we
cannot completely exclude also an influence of the particles
surface spin disorder, which can also contribute to modify the
superparamagnetic regime of the nanoparticles.
2 S. Vaucher, M. Li and S. Mann, Angew. Chem., Int. Ed., 2000, 39,
1793.
The magnetic properties of the sample B are similar to what
is observed for sample C. The FC/ZFC curves also present
irreversibility and the maximum temperature on the ZFC
curve is observed at 10.9 K (Table 1 and Fig. SI 4, ESIw).
The temperature dependence of the ac susceptibility, its
in-phase, w0, and its out-of-phase, w00, components measured
in zero static field show a frequency dependence suggesting the
presence of a short range magnetic ordering. The Arrhenius
fit of the thermal variation of the relaxation time shows a
very small value of the parameter t0 = 1.21 ꢃ 10ꢀ52 s,
which has no physical meaning, while the Vogel–Fulcher fit
3 L. Catala, T. Gacoin, J.-P. Boilot, E. Riviere, C. Paulsen, E. Lhotel
and T. Mallah, Adv. Mater., 2003, 15, 826.
4 T. Uemura and S. Kitagawa, J. Am. Chem. Soc., 2003, 125, 7814;
L. Catala, A. Gloter, O. Stephan, G. Rogez and T. Mallah, Chem.
Commun., 2006, 1018; D. Brinzei, L. Catala, C. Mathoniere,
W. Wensdorfer, A. Gloter, O. Stephan and T. Mallah, J. Am.
Chem. Soc., 2007, 129, 3778; J. M. Domı
Inorg. Chem., 2003, 42, 6983.
5 G. Clavel, J. Larionova, Y. Guari and Ch. Gue
2006, 12, 3798; J. Larionova, Y. Guari, H. Sayegh and Ch. Gue
Inorg. Chim. Acta, 2007, 360, 3829; Y. Guari, J. Larionova,
C. Blanc, P. Dieudonne, A. Tokarev and C. Guerin, Langmuir,
2009, 25, 1138.
´
ngez-Vera and E. Colacio,
´
rin, Chem.–Eur. J.,
rin,
´
´
´
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009 New J. Chem., 2009, 33, 2449–2456 | 2455

View Article Online
6 B. Folch, Y. Guari, J. Larionova, C. Luna, C. Sangregorio,
C. Innocenti, A. Caneschi and C. Guerin, New J. Chem., 2008,
32, 273.
13 E. Reguera, J. F. Bertran, C. Diaz, J. Bianco and S. Rondon,
Hyperfine Interact., 1990, 53, 391.
14 D. H. M. Buchold and C. Feldmann, Chem. Mater., 2007, 19, 3376.
´
7 H. Yang and D. Zhao, J. Mater. Chem., 2005, 15, 1217;
J. J. Baumberg, Nat. Mater., 2006, 5, 2.
8 R. Backov, Soft Matter, 2006, 2, 452; E. Prouzet, S. Ravaine,
C. Sanchez and R. Backov, New J. Chem., 2008, 32, 1284.
9 C. Boissiere, A. Larbot, C. Bourgaux, E. Prouzet and
C. A. Bunton, Chem. Mater., 2001, 13, 3580.
15 J. Larionova, Y. Guari, C. Sangregorio and C. Gue
Chem., 2009, 33, 1177.
16 L. Neel, Ann. Geophys., 1949, 60, 661; J. L. Dormann, L. Bessais
´
rin, New J.
´
and D. Fiorani, J. Phys. Chem. C, 1988, 21, 2015.
17 J. A. Mydosh, Spin Glasses, Taylor and Francis, Washington DC,
1993.
10 S. A. Bagshaw, E. Prouzet and T. J. Pinnavia, Science, 1995, 269,
1242; C. Boissiere, A. Larbot, A. van der Lee, P. J. Kooyman and
E. Prouzet, Chem. Mater., 2000, 12, 2902; E. Prouzet and
C. Boissiere, C. R. Chim., 2005, 8, 579.
18 C. Djurberg, P. Svedlindh, P. Nordblad, M. F. Hansen, F. Bodker
and S. Morup, Phys. Rev. Lett., 1997, 79, 5154; G. Balaji, G. Wilde,
J. Weissmuller, N. S. Gabhiye and V. K. Sankaranarayanan, Phys.
Status Solidi B, 2004, 241, 1589.
11 D. O. de Za
´
rate, F. Bouyer, H. Zschiedrich, P. J. Kooyman,
19 J. A. Mydosh, Phys. Rev. B: Condens. Matter Mater. Phys., 1989,
40, 11243.
20 E. Chelebaeva, Y. Guari, J. Larionova, A. Trifonov and
Ph. Trens, J. Lapichella, R. Durand, C. Guillem and E. Prouzet,
Chem. Mater., 2008, 20, 1410.
12 A. Bleuzen, C. Lomenech, V. Escax, F. Villain, F. Varret,
C. Cartier dit Moulin and M. Verdaguer, J. Am. Chem. Soc.,
2000, 122, 6648; O. Sato, Y. Einaga, Y. Iyoda, A. Fujishima and
K. Hashimoto, J. Phys. Chem. B, 1997, 101, 3903; D. F. Shriver
and D. B. Brown, Inorg. Chem., 1969, 8, 42.
C. Guerin, Chem. Mater., 2008, 20, 1367.
´
21 J. R. L. Almeida and D. J. Thouless, J. Phys. A: Math. Gen., 1978,
11, 983.
22 T. Jonsson, J. Mattsson, C. Djurberg, F. A. Khan, P. Nordblad
and P. Svedlindh, Phys. Rev. Lett., 1995, 75, 4138.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2009
2456 | New J. Chem., 2009, 33, 2449–2456