Surface and structural features of Co-Fe oxide nanoparticles deposited on a silica substrate

Article abstract of DOI:10.1002/ejic.200600778

Mono- and dimetallic cobalt- and iron oxide nanoparticles deposited on the surface of a silica substrate have been prepared by an impregnation technique. Both the bulk and surface structures of these particles have been characterised by different physical and chemical techniques. The results provided by X-ray diffraction, Moessbauer and X-ray photoelectron spectroscopy show the formation of separate Co₃O₄ and Fe₂O ₃ nanoparticles in oxide samples, but in no case were cobalt-iron mixed oxides detected. Quantitative data also showed that the dispersion degree of cobalt- and iron oxides is rather low. It was also observed that pretreatment of the supported metal oxide nanoparticles under a hydrogen atmosphere does not promote the formation of a metal-support interaction, although a cobalt-iron interaction is observed in the dimetallic systems. The diffraction patterns and photoelectron and Moessbauer spectra of these dimetallic samples provide conclusive proof for the formation of both metallic Co⁰ and iron-cobalt (Co₇Fe₃) alloy phases in the hydrogen-reduced samples. It was also found that the crystallite size of the alloyed Co ₇Fe₃ phase increases with increasing iron content, i.e. 11 and 23 nm for samples containing 1 % and 5 % Fe added to the base Co sample, respectively, while that of Co⁰ was constant (10 nm). Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejic.200600778

FULL PAPER
DOI: 10.1002/ejic.200600778
Surface and Structural Features of Co-Fe Oxide Nanoparticles Deposited on a
Silica Substrate
[
a,b]
[a]
[c]
Víctor A. de la Peña O’Shea,*
M. Consuelo Álvarez-Galván,
[a]
[c]
José M. Campos-Martin, Nieves N. Menéndez, Jesús D. Tornero, and
José L. G. Fierro*[
a]
Keywords: Cobalt / Iron / Oxides / Metal–metal interactions / Materials science
Mono- and dimetallic cobalt- and iron oxide nanoparticles
interaction, although a cobalt–iron interaction is observed in
deposited on the surface of a silica substrate have been pre- the dimetallic systems. The diffraction patterns and photo-
pared by an impregnation technique. Both the bulk and sur-
face structures of these particles have been characterised by
different physical and chemical techniques. The results pro- Co and iron-cobalt (Co
electron and Mössbauer spectra of these dimetallic samples
provide conclusive proof for the formation of both metallic
0
7
Fe
3
) alloy phases in the hydrogen-
vided by X-ray diffraction, Mössbauer and X-ray photoelec-
tron spectroscopy show the formation of separate Co and
Fe nanoparticles in oxide samples, but in no case were
cobalt–iron mixed oxides detected. Quantitative data also
showed that the dispersion degree of cobalt- and iron oxides
is rather low. It was also observed that pretreatment of the
supported metal oxide nanoparticles under a hydrogen atmo-
sphere does not promote the formation of a metal–support
reduced samples. It was also found that the crystallite size
of the alloyed Co Fe phase increases with increasing iron
3
O
4
7
3
2
O
3
content, i.e. 11 and 23 nm for samples containing 1% and
5% Fe added to the base Co sample, respectively, while that
0
of Co was constant (10 nm).
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Introduction
ness impregnation methods from nitrate precursors have
[
6,14]
also been reported.
Dimetallic catalysts prepared by the
Nowadays, composite materials give rise to numerous
studies with a view to improving their mechanical, thermal,
optical and other properties, which hold promise for novel
technological applications. Transition metal oxides are be-
ing widely used, and Fe-Co materials are of particular inter-
est in catalytic reactions such as the Fischer–Tropsch syn-
incipient wetness method show reactivities different from
that of equivalent Fe- or Co single-metal catalysts, thus in-
dicating that the intimate interaction between Fe and Co
[24,25]
is important.
A synthesis of iron-cobalt nanoparticles
dispersed in a silica matrix has also been reported, and the
authors noted the formation of magnetic cobalt ferrite nan-
ocomposites, Co O and ferrihydrite or Co-Fe alloys, de-
[
1–6]
[7–9]
thesis (FTS),
methanol decomposition,
NH synthe-
3
3
4
sis,[ carbon nanotube synthesis.
10]
[11–13]
Furthermore, their
[19]
pending on the Fe and Co precursors.
mixtures have been prepared on several supports, including
alumina, carbon,
Fe/Co catalyst
interesting magnetic properties[
14]
suggest that they will
have many applications in magneto-optical recording me-
dia, displays and devices such as wave-guides, insulators,
modulators and switches.[
[24]
[26,27]
[28]
[14,25,29,30]
zirconia and titania.
A viable methodology that has been developed for con-
15–17]
[31–35]
trolling the property of a metal is that of alloying.
This fact has provoked several reports in which these
mixtures are prepared by different synthetic methods such
Three structures have been found for the metal in the iron-
cobalt alloy, depending on the synthesis temperature and
the subsequent thermal treatment: a bcc structure, an α-
[
18,19]
[20]
[21]
as sol–gel,
tering
co-precipitation,
and thermal decomposition.
supported systems prepared by wetness- or incipient wet-
plasma coating,
sin-
[
22]
[23]
Co- and/or Fe-
Mn-like structure and a fcc structure that appears after
[36,37]
thermal treatment.
The bcc and fcc structures are clas-
sical structures for iron- and cobalt-based alloys. These al-
loys exhibit outstanding properties, particularly the cobalt-
enriched composites. Indeed, these latter are active in the
FTS and favour the formation of olefins without producing
[
[
a] Instituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco,
8049 Madrid, Spain
2
b] Departament de Química Inorgànica, Universitat de Barcelona,
C/Martí i Franquès 1–11, 08028 Barcelona, Spain
Fax: +34-93-490-7725
[
4,38]
[6]
high amounts of CO2
and alcohol production.
A large body of work has been developed over the past
decades on the performance of supported cobalt catalysts,
E-mail: victor@ qi.ub.es
[
c] Universidad Autónoma de Madrid,
[
39–43]
Cantoblanco, 28049 Madrid, Spain
especially with regard to the effect of promoters,
sup-
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5057

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
ports[44–46] and preparation[47–52] methods. Cobalt catalysts Results and Discussion
produce high yields of long-chain alkenes in the FTS. They
are also characterised by a low ability to yield oxygenates. Chemical Composition and Textural Properties
In addition, cobalt catalysts exhibit considerable stability in
The chemical composition of the oxide samples was de-
the metallic state.
termined by ICP-AES; results are compiled in Table 1.
Iron is the other active phase that is most often used in
These results indicate that cobalt- and/or iron loadings are
the hydrogenation reactions of carbon oxides. It is know
close to the nominal values. The BET specific areas and
that several forms of iron (iron, iron carbides, iron oxide)
pore volumes of the fresh metal oxide loaded samples and
exist in iron-based catalysts when subjected to FTS, and
the bare SiO support, also summarised in Table 1, indicate
2
there are many studies dealing with the role of iron phases
that incorporation of cobalt and/or iron onto the SiO sub-
[
53–55]
2
in the FTS.
Iron catalysts used in CO hydrogenation
strate leads to a drop in both specific area and pore volume.
A decrease in the BET area of around 20–30% is found
upon incorporation of iron and cobalt, and a similar drop
is found for the other two dimetallic (CoFe10/1-c and
CoFe10/5-c) samples. The pore-size distributions of the
metal-oxide-loaded samples and SiO2 reference are dis-
played in Figure 1. The pore-size distributions of the sam-
ples do not change upon incorporation of cobalt- and/or
iron oxides during the preparation step. The simultaneous
drop in BET area and pore volume and the almost un-
changed pore distribution of the metal-oxide-loaded sam-
are usually promoted to obtain a high stability and catalytic
activity. These systems produce higher yields of oxygenated
products and can perform the water gas-shift reaction. The
incorporation of cobalt and iron phases onto a support
substrate results in substantial changes in both activity and
product distribution as, when used together, they do not
simply give the additive properties (activity, selectivity) ex-
pected from knowledge of the properties of the individual
metals. These dimetallic CoFe catalysts proved to be much
more attractive in terms of alcohol formation and give
rise to the production of ethanol and propanol, depending
on the iron content.[ Taking this into account, the
possible interaction of cobalt and iron, in the form of cobalt
ferrite or alloys, when used together as active phases for
catalytic reactions such as FTS or as a possible magnetic
composite should be studied by different characterisation
techniques.
6]
In light of the above, the present work was undertaken
with the aim of analysing the effect of the incorporation of
both cobalt and iron species supported on a silica matrix
on the structural, textural and morphological characteris-
tics relative to the properties shown by the monometallic
systems. In particular, emphasis is placed on the formation
of CoFe intermetallic compounds after catalyst activation.
Some clues to the nature of the bulk and surface structures
generated on the silica-supported cobalt-iron catalysts were
obtained by XRD, X-ray photoelectron spectroscopy
(XPS), Mössbauer spectroscopy and temperature-pro-
grammed reduction (TPR). The synergism between cobalt
and iron, two FT active metals, will be explored in a subse-
quent work that will consider whether the above metals,
when used together, give the additive properties (activity
and selectivity) expected of the individual metals or not.
Figure 1. Pore distribution for the (1) support, (2) CoFe10/5-c and
(3) Co10-c.
Table 1. Elemental analysis and textural characteristics of the CoFe/SiO
2
system.
Elemental analysis
2
N adsorption
Metal loading
Metal loading
(% Fe)
nominal
(% Co)
2
–1
3 –1
Catalyst
nominal
ICP
ICP
S
BET [m g ]
Vads [cm g ]
SiO
SiO
Co10-c
Fe10-c
CoFe10/5c
CoFe10/1c
2
-c
–
–
10
–
10
10
–
–
10.2
–
9.3
9.4
–
–
–
10
5
1
–
–
–
10.2
5.1
0.9
289
300
219
216
221
210
1.19
1.18
0.80
0.93
0.96
0.88
2
5058
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068

Co-Fe Oxide Nanoparticles Deposited on a Silica Substrate
FULL PAPER
ples point to a preferential location of big cobalt- and/or
iron oxide particles in the interparticle voids between silica shows all the lines of crystalline Fe O3
The diffraction pattern of the calcined Fe10-c sample
[
58,59]
(JCPDS card
2
particles. Thus, the decrease in the pore volume as a result 85-0599), together with the broad feature of silica at about
of pore locking does not change the pore-size distribution 22° 2θ (see above). No peaks of any other iron-containing
and is in accordance with the greater decrease in pore vol- compound are observed.
ume, which is higher than the weight fraction of deposited
The XRD patterns of calcined CoFe10/5-c and CoFe10/
metal phases. However, deposition of very small metal ox- 1-c samples are also included in Figure 2A. Both samples
ide particles within the smallest pores of silica, and more show diffraction lines that correspond to the Fe O phase;
2
3
specifically for the Co10-c sample, cannot be ruled out.
these lines, as expected, are more intense in the sample with
a higher Fe loading (CoFe10/5-c). Some of the Fe O lines
2
3
[
58,60]
are overlapped by those of the Co O spinel,
and the
3
4
rather low intensity of the lines corresponding to Fe O in-
2
3
Crystalline Structure
dicates the low crystallinity of the Fe O particles. A weak
2
3
interaction between Fe O and Co O phases is possible
and could affect the crystallinity of the iron oxide particles
2
3
3
4
Calcined Samples
The crystal structures of the supported cobalt and iron without producing any modification in the structure of the
phases of oxide samples were determined by X-ray diffrac- oxides. These results indicate that the two oxides remain as
tion (Figure 2A). The monometallic Co10-c sample displays separate phases at the scale of the XRD technique, which
sharp and narrow diffraction lines that fit very well with is above 3–4 nm. The lattice parameters found for the di-
[56]
that of Co O spinel (JCPDS card 78-1970).
The ob- metallic systems CoFe10/5-c and CoFe10/1-c are a = 8.090
tained lattice parameter for this phase is a = 8.088 Å, which and 8.089 Å, respectively, and are very similar to that of
is similar to that obtained for the cobalt oxide spinel phase. Co , which indicates that a CoFe mixed composite has
3
4
O
4
3
A broad band centred at about 22° 2θ belongs to the not been formed. The presence of a mixed cobalt–iron ox-
amorphous structure of silica. A detailed examination of ide phase would give a higher value of the lattice parameter.
the spectrum was done to ensure that no other crystalline
The mean particle size of the supported cobalt- and/or
Co-containing phases, such as Co SiO , had developed. As iron oxides was calculated from the Scherrer equation by
2
4
Co O and CoSiO4 species have similar interplanar dis- selecting the most intense diffraction line for each oxide —
3
4
tances, differentiation between both compounds is not easy the crystal planes (311) for Co O and (220) for Fe O —
3
4
2
3
unless they are very well crystallised. It should be empha- and by assuming a spherical geometry for the oxide par-
sised that formation of a crystalline Co SiO phase is only ticles. The results compiled in Table 2 reveal that the larger
2
4
possible upon calcination of the impregnate at temperatures size corresponds to Co O particles in the monometallic
3
4
substantially higher than that (773 K) employed in this catalyst (21 nm), while it is slightly reduced (17 nm) for the
[
57]
work.
Note that the formation of a cobalt-oxide–silica parent Fe10-c sample. It can also be seen that the size of
interacting layer cannot be ruled out from this XRD the Co O particles decreases upon adding iron, and this
3
4
pattern.
drop is even more marked for sample CoFe10/5-c. Finally,
2 2
Figure 2. X-ray diffraction pattern of Co/Fe(SiO ): (A) calcined, (B) after H reduction at 773 K.
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5059

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
no attempt was made to calculate the size of the iron oxide obtained for the reduced CoFe10/5-c sample and is in ac-
particles in the CoFe10/1-c sample because of the extreme cordance with the relative proportions of Co and Fe in this
broadening of the Fe O diffraction lines. This fact indi- sample. Both metallic species found in the dimetallic sys-
2
3
cates that the particle size should be below 5 nm, which is tems present very similar lattice parameters. The following
the limit for applying the Scherrer equation.
lattice parameters are found for both systems: a = 2.842 Å,
which corresponds to Co , and a = 3.541 Å, which corre-
sponds to the CoFe alloy.
The mean particle size was determined by X-ray line
broadening according to the Scherrer equation (Table 2).
For this calculation, the (111) plane of metallic Co, the
0
Table 2. Crystallite size of supported nanoparticles for the calcined
and reduced systems.
Sample
XRD phases
Co
2θ [°]
Crystal size [nm]
Co10-c
Fe10-c
CoFe10/5-c
3
O
4
44.38
44.75
44.22
45.03
44.33
42.49
44.38
44.75
44.22
45.03
44.33
42.49
21
18
14
14
17
–
17
22
10
23
10
11
Fe
Co
2 3
O
(110) plane for metallic Fe and the (110) plane for the Co-
3
O
4
Fe alloy phases were selected. A comparison of the crystal-
lite sizes for both reduced and calcined samples shows an
increase in crystallite phase during the reduction because of
sintering of the metal phase. In addition, the extent of sin-
tering is higher in Fe10-c than in Co10-c. For the dimetallic
catalysts, an increase in alloy particle size is observed in
Fe
2
O
3
CoFe10/1-c
3
Co O
4
2 3
Fe O
0
Co10-c (H
Fe10-c (H
CoFe10/5-c (H
2
)
)
Co
Fe
0
2
0
2
2
)
)
Co
Fe Co
Co
Fe Co
3
7
7
0
0
CoFe10/5-c. Finally, the crystallite size of Co in both di-
CoFe10/1-c (H
3
metallic samples is similar but lower than for Co10-c, which
suggests that alloy formation inhibits the sintering of Co0
particles.
Reduced Samples
The pattern of the Co10-c sample reduced in H at 773 K
2
Morphology and Phase Distribution
shows diffraction lines belonging to the (111), (220) and
0
(
1
311) planes of the fcc Co crystallites (JCPDS card 1-
0
Calcined Samples
255). The lattice parameter found for cubic Co is a =
3.541 Å. No hexagonal structure is observed because the
The morphology of the samples was studied by electron
reduction temperature is too high to achieve this type of microscopy and line profile analysis (LPA). The LPA tech-
crystal structure. The diffraction lines of CoO are still ob- nique allows the determination of the catalyst surface top-
served, which indicates than the cobalt phase has not been ology as well as the active-phase disposition, the metallic
reduced completely, although their intensity is lower than particle size and the surface state of the support. Fig-
that of the oxide sample. A broad feature at about 22° 2θ ures 3A and 3C show SEM micrographs of cobalt, while
arising from the amorphous silica substrate is also found.
Figures 3B and 3D display micrographs of monometallic
The reduction profile of the Fe10-c sample is shown in silica-supported samples taken at two different magnifica-
Figure 2B. Diffraction lines of metallic α-Fe (JCPDS card tions. No differences were found for these two types of sam-
6
-696) and FeO (JCPDS card 75-1150) are observed, al- ples. The surface is made up of particles of irregular shape
though the intensity of the peaks of the latter phase is weak. and size ranging from 2 to 25 µm. The particles are de-
The absence of hematite (Fe O ) or magnetite (Fe O ) limited by well-defined faces formed by laminar exfoliation
2
3
3
4
phases indicates that the reduction is in the final stage. The (A and B). Higher magnification images (C and D) indicate
lattice parameter found for this species is a = 2.863 Å.
The diffraction profiles of the dimetallic CoFe10/5-c and
CoFe10/1-c samples are also displayed in Figure 2B. Besides
the species found in the monometallic samples, new diffrac-
tion lines belonging to a FeCo phase are observed. These
new peaks — the most intense one in both diffraction pat-
terns occurrs at 2θ ≈ 45° — correspond to the (110) plane of
the Co Fe alloy (JCPDS card 48-1818) with a bcc crystal
7
3
structure. This indicates that most of the cobalt and iron
phases are involved in this alloy. Some additional peaks ob-
served in both reduced samples can be indexed to metallic
Co and cobalt oxide, although apparently no lines for a
metallic iron phase are present. This fact could be due to
an overlapping of these lines with those assigned to cobalt
and to the Co Fe alloy, since the ionic radii are very sim-
7
3
ilar. The intensity of the main diffraction line characteristic
of the alloy in the reduced CoFe10/1-c sample is lower than
that of metallic Co , which contrasts with the proportion (B and D).
Figure 3. SEM images of calcined Co10-c (A and C) and Fe10-c
0
5060
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068

Co-Fe Oxide Nanoparticles Deposited on a Silica Substrate
FULL PAPER
the presence of uniformly developed spherical particles, support. The TEM images of Co10-c show spherical par-
about 0.2 µm in size, with compact packing. The SEM ticles with non-uniform sizes within each SiO particle. The
2
micrographs of the dimetallic samples do not show any dif- same is true for other SiO particles. The size of the cobalt
2
ferences with respect to those of the monometallic samples. oxide particles falls between 0.1 and 1 µm (Figure 4A). Riva
[
61]
The LPA analysis indicates a very irregular profile for et al. have performed a TEM analysis of silica-supported
silicon, with holes and protuberances of different sizes, cobalt oxides and found that metal oxide particles form
depths and heights. However, the profile belonging to co- spherical aggregates both inside and on the surface of silica
balt is more homogeneous and is consistent with a well- particles, with non-uniform sizes and a mean diameter of
dispersed phase. Similarly, LPA analysis of the monometal- between 0.3 and 0.5 µm. On the contrary, the micrograph
lic Fe10-c sample shows a silicon profile resembling that of of Fe10-c shows that iron oxide particles are not spherical
the cobalt counterpart (Co10-c), although the iron profile as in the case of Co10-c (Figure 4B). In addition, non-uni-
reflects a wider size distribution of the iron oxide particles. form aggregates, with sizes ranging from 0.1 to 1 µm, are
In addition, the LPA profiles for the dimetallic CoFe10/5-c formed on each silica particle. Elemental analysis was also
and CoFe10/1-c samples indicate that both cobalt and iron performed on individual particles, and the results indicated
profiles are more homogeneous than in the monometallic that they are mainly made up of iron and oxygen. XRD,
ones, thereby suggesting a better distribution of the Co O4 XPS and Mössbauer spectroscopy revealed Fe O to be the
3
2
3
and Fe O phases across the surface.
only phase present. Finally, the TEM images of CoFe10/5-
2
3
The cobalt and iron mono- and dimetallic systems were c and CoFe10/1-c revealed that the particle size diminishes
also analysed by TEM with the aim of determining the with respect to the monometallic samples, which is in good
form, shape and distribution of the particles on the silica agreement with the Mössbauer data.
2
Figure 4. TEM images of (A) Co10-c and (B) Fe10-c; (C) TEM images of CoFe10/5-c reduced in H at 773 K, with a high resolution
image on the right. Inset: the electron diffraction pattern.
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5061

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
Reduced Samples
lyst surface was determined by photoelectron spectroscopy.
The iron-containing samples were characterised by Möss-
bauer spectroscopy with the aim of determining the local
environment of the iron species and their concentration.
Only the two calcined Fe10-c and CoFe10/5-c samples were
analysed by this technique. The other dimetallic sample
Co10-c, Fe10-c and dimetallic samples reduced in hydro-
gen at 773 K were analysed by TEM with the aim of com-
paring the size, shape and distribution of the metal particles
with respect to the fresh counterparts. The TEM images of
Co10-c show that the size and shape of some particles re-
main unchanged upon reduction. In addition, a certain
fraction of particles lose their spherical form and form
other non-uniform aggregates. The reduced Fe10-c sample
behaves similarly. The TEM image of reduced CoFe10/5-c
displayed in Figure 4B reveals that the particles present an
irregular shape and appear better dispersed on the silica
substrate. Electron nanodiffraction taken on a large particle
(
(
CoFe10/1-c), which contains a much lower amount of iron
1 wt.-%), was not studied since its signal is too weak to
get reliable information. Figure 5A displays the Mössbauer
spectrum of sample Fe10-c. It includes two components: a
paramagnetic doublet assigned to highly dispersed Fe3 and
a magnetic sextet related to iron(III) oxide particles with a
particle size larger than that responsible for the doublet.
The hyperfine field corresponding to the sextet (50.4 T) is
+
(Figure 4B) confirms the formation of an iron-cobalt alloy.
[
62,63]
lower than that of bulk α-Fe O (51.5 T).
This fact
This phase was also confirmed by XRD, XPS and Möss-
bauer spectroscopy.
2
3
points to the presence of small α-Fe O particles with a dia-
2
3
[
64]
meter of between 7 and 12 nm.
α-Fe O species are the
2 3
only species detected by Mössbauer analysis of the calcined
Fe10-c sample, in agreement with the XRD pattern analysis
presented above. The isomer shift (δ), quadrupolar shift
Bulk Structures of Iron-Containing Catalysts (Mössbauer
Spectroscopy)
(
QS), quadrupolar magnetic field (H ) and weight concen-
hf
tration of each species were measured and are reported in
Table 3. It can be seen that α-Fe O crystallites form the
major phase (71%), while the proportion of Fe nanopar-
2
3
Oxide Samples
3+
The bulk structure of iron-containing catalysts was re- ticles is substantially lower (29%). In addition, these results
vealed by Mössbauer spectroscopy, and both the chemical indicate that the dispersion degree of the major phase (α-
state and relative abundance of iron and cobalt at the cata- Fe O ) is rather low.
2 3
Figure 5. Mössbauer spectra of calcined (A) Fe10-c and (B) CoFe10/5-c; Mössbauer spectra of (C) Fe10-c and (D) CoFe10/5-c reduced
in H at 773 K.
2
5062
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068

Co-Fe Oxide Nanoparticles Deposited on a Silica Substrate
FULL PAPER
Table 3. Mössbauer parameters for the systems Fe10-c and in the H -reduced Fe10-c sample are collected in Table 3
2
CoFe10/5-c calcined and reduced with H
2
at 773 K.
and are similar to the ones reported by these authors. The
assignment of the parameters was done on the basis of
Sample
Components
δ
QS
[mms–1]
H
hf
%
mms–1]
[T]
50.4 71
29
50.7 69
31
33.2 39
[70]
[66]
[
studies by Cagnoli et al. and Hobson et al., in which
the doublet with δ = 0.76 mms–1 and ∆QS = 0.98 mms
was assigned to small FeO crystallites supported on the
–1
Fe10-c
α-Fe
(nano-size)
α-Fe
2
O
3
0.37
0.33
0.37
0.32
0.00
0.33
0.79
1.12
0.01
0.32
0.21
0.71
0.21
0.71
0.01
0.92
0.87
2.07
0.01
0.71
α-Fe
α-Fe
α-Fe
2
2
2
O
3
–
CoFe10/5-c
2 3
O
SiO substrate. For the second doublet, the large differences
2
O
3
(nano-size)
–
between the obtained quadrupole shift and the one re-
ported in the literature for Fe1–xO species suggests the exis-
tence of Fe2 particles interacting strongly with the support.
The quadrupole shift of this second doublet is similar to
that reported in the literature for different Fe SiO spe-
Fe10-c (H
2
)
α-Fe
O
3
(nano-size)
–
–
–
20
38
3
+
Fe1–x
O
Fe2+ (Fe
Co-Fe
(nano-size)
2
4
SiO )
CoFe10/5-c (H
2
)
33.9 76
24
2
4
α-Fe
2
O
3
–
[
71,72]
cies.
Indeed, the reduction of these iron silicate species
to iron metal [Equation (1)] is extremely difficult and hence
requires high temperatures.
The Mössbauer spectrum of the calcined CoFe10/5-c Fe SiO + H Ǟ 2Fe + SiO + H O
(1)
2
4
2
2
2
sample (Figure 5B) is quite similar to that of the monome-
tallic Fe10-c sample. The appearance of a sextet suggests
the presence of ferromagnetic particles that correspond to
Figure 5D depicts the Mössbauer spectrum of the
CoFe10/5-c sample reduced under hydrogen at 773 K for
h. The Mössbauer spectrum of this sample arises from
–
1
iron(III) oxide, while the doublet (δ = 0.32 mms , QS =
2
–
1
3+
0
.71 mms ) is characteristic of highly dispersed Fe ions,
thus revealing the existence of nanometre-sized particles (5–
0 nm). From the data collected in Table 3, it can be in-
the convolution of a magnetic sextet and a paramagnetic
doublet. The values of the hyperfine field (H) obtained for
the sextet points to the appearance of the CoFe alloy. John-
1
III
[
73]
ferred that the proportion of α-Fe O and nano-sized Fe
2
3
son et al. have collected a sequence of values for different
CoFe alloys between 34.6 and 35.6 T. The value of the hy-
oxide particles in the calcined CoFe10/5-c sample is similar
to that found in monometallic Fe10-c.
perfine field (33.9 T) found for the H -reduced CoFe10/5-c
2
sample is similar to that reported by other authors.[
74,75]
Reduced Samples
2
The paramagnetic doublet can be assigned to small Fe O3
particles highly dispersed on the silica surface. In favour of
[
76]
The local environment of the iron atoms in H -reduced this assignment is the work of Huang et al.,
who pro-
2
iron-containing samples was revealed by Mössbauer spec- posed that this paramagnetic doublet arises from iron(III)
troscopy. The Mössbauer spectrum of Fe10-c after re- particles exchanged with the silica, which are, indeed, very
duction with hydrogen at 773 K for 2 h is displayed in Fig- difficult to reduce.
ure 5C. This spectrum can be fitted by four components: (i)
a magnetic sextet that corresponds to α-Fe, (ii) a paramag-
netic doublet due to small iron oxide particles (α-Fe O )
2
3
Surface Structures
similar to the one found in the calcined sample, and (iii)
2
+
two paramagnetic doublets corresponding to Fe species.
It should be emphasised that one of these latter doublets
has a low isomer shift, which suggests that the Fe2 ions
Oxide Samples
+
The chemical state and relative abundance of cobalt-
have a low coordination number (tetrahedral coordination), and/or iron oxides at the silica surface were determined by
while the other doublet has an isomer shift characteristic photoelectron spectroscopy. Figure 6A shows the energy re-
of octahedral coordination.[
65,66]
No magnetic component gion corresponding to the Co 2p3/2 core levels in calcined
belonging to Fe O was found after hydrogen reduction. Co10-c. This peak can be resolved, after curve-fitting pro-
3
4
2
+
This observation agrees with previous studies by Yuen et cedures, into two components belonging to the Co and
al.[ and Weilers et al.
67]
[68]
Thus, for hydrogen-reduced sil- Co ions present in the two samples. From this compari-
3+
ica-supported iron samples, Yuen et al.[
54]
observed the son, it is evident that the Co O spinel phase is present in
3 4
2
+
presence of two doublets corresponding to Fe species the sample Co10-c. The first peak at 780.0 eV can be as-
with different coordination environments: an inner doublet cribed to Co3 in an octahedral environment, while that
+
2+
attributed to low-coordinate cations strongly interacting at 782.5 eV can be assigned to Co ions in a tetrahedral
with the substrate and an external doublet assigned to high- environment.[57] Similarly, the XP spectrum of the Fe
2
+
coordinate Fe cations present in highly dispersed iron ox- 2p3/2 core level of calcined Fe10-c is displayed in Figure 6B.
ide particles. A somewhat different interpretation of these This suggests that hematite (Fe O ) is the only species pres-
2
3
two doublets was proposed by Bjerne et al.[ These au- ent in the calcined Fe10-c sample.
69]
thors suggested that both components correspond to super-
Figures 6C and 6D show the energy regions correspond-
ficial iron silicates with different symmetry (tetrahedral and ing to the Co 2p3/2 and Fe 2p3/2 core-levels of the calcined
octahedral environment). The Mössbauer parameters found CoFe10/5-c sample. The similarity between the peak posi-
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5063

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
Reduced Samples
The chemical state of the elements and their relative
abundances on the catalyst surface after reduction were re-
vealed by photoelectron spectroscopy. Figure 6A depicts
the Co 2p3/2 peaks of Co10-c after reduction. The calcined
Co10-c sample shows two components: one associated with
3+
Co in an octahedral position and the other assigned to
2+
[77,78]
Co in the tetrahedral position of the Co O spinel.
3
4
Upon reduction with H , three components are observed:
2
3+
octahedral Co
ions present in an unreduced Co O
octahedral Co
Figure 6B shows the Fe 2p3/2 core-level spectrum
3
4
[
57,77]
2+
[79–81]
phase,
ions in CoO
and
0
[81–83]
Co .
of Fe10-c after reduction. The fresh sample exhibits one
component assigned to Fe3 (Fe O ), whereas the reduced
+
2
3
3+
sample displays three components, one due to Fe in a
[
84]
2+
non-reduced Fe O
phase, the second due to Fe
and the third due to metallic iron.
2
3
[
85,86]
[84,87]
(FeO),
Figure 6C shows the Co 2p3/2 and Fe 2p3/2 regions of
CoFe10/5-c. Some differences are observed for the reduced
samples. Although the profiles of the dimetallic and mono-
0
metallic samples are similar, the binding energies of the Co
0
and Fe species are different, which suggests that a Co-Fe
alloy is formed in the dimetallic samples. It is apparent from
the binding energies that the reduction of iron oxide is facil-
[
88]
itated in the presence of cobalt,
while the reduction of
cobalt oxide in close interaction with iron oxide is partly
inhibited. This observation agrees with the results reported
[
89]
in the literature
and is consistent with the TPR data.
Table 4 shows the metal/support surface ratios. An increase
in the amount of surface metallic phase is observed for the
dimetallic catalyst after reduction.
Table 4. Surface metal/Si ratios obtained by XPS.
Sample
Oxide
Co/Si
2
H reduced (773 K)
Fe/Si
Co/Si
Fe/Si
Co10-c
Fe10-c
CoFe10/5-c
0.01
–
0.01
–
0.01
0.01
0.01
–
0.02
–
0.01
0.02
Figure 6. XPS profiles of (A) Co10-c, (B) Fe10-c and (C and D)
CoFe10/5-c calcined and reduced with H
2
at 773 K.
tions and line profiles in the Co 2p3/2 and Fe 2p3/2 regions
for the calcined CoFe10/5-c sample and those of the respec-
tive Co10-c and Fe10-c monometallic samples is an indica-
tion that the cobalt- and iron oxide phases are the same as
Redox Properties
An insight into the structure of the supported metal ox-
in the monometallic systems, that is, Co O and Fe O are ide phase can be derived from programmed reduction ex-
3
4
2
3
the only phases detected. As is the case with Co10-c, a shift periments. The reduction profile of Co10-c and a bulk
in the binding energy of the Co 2p3/2 peak of CoFe10/5 is Co O reference are shown in Figure 7A. The reduction
3
4
observed relative to bulk Co O because of the presence of profile of Co10-c shows two peaks that are similar to those
3
4
metal–substrate interactions. As the binding energies of the observed in the bulk Co O oxide. These profiles point to
3
4
Co 2p3/2 and Fe 2p3/2 levels for CoFe10/5-c are practically a two-step reduction process: the first one involving low H2
the same as those found for the monometallic samples, it consumption starts at about 475 K and overlaps with the
can be inferred that the Co–Fe interaction in the calcined more intense second one whose maximum is placed at
sample, if any, is very weak. Finally, it should be empha- about 630 K. Thus, the reduction of Co O can be de-
3
4
3+
sised that both CoFe10/5-c and CoFe10/1-c were analysed scribed by the reduction of Co ions present in the spinel
but the signal-to-noise ratio in the Fe 2p3/2 region for the structure to Co2 [Equation (2)], with the subsequent struc-
latter sample was too low to be measured accurately and tural change to CoO, followed by the reduction of CoO to
+
[
90–92]
hence no further attention was paid to this level.
metallic cobalt [Equation (3)].
5064
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068

Co-Fe Oxide Nanoparticles Deposited on a Silica Substrate
FULL PAPER
Figure 7. A comparison of the TPR profiles: (A) Co
3
O
4
(dashed line) vs. Co10-c (solid line); (B) Fe
2
O
3
(dashed line) vs. Fe10-c (solid
line); (C) CoFe101/5-c (dashed line) and CoFe10/1-c (solid line).
Co
3
O
4
+ H
2
Ǟ 3CoO + H
2
O
(2) ternal cavities of the support, where the reduction process
[90,93]
should be limited by H O diffusion through the pores.
2
3
CoO + 3H
2
Ǟ 3Co + 3H
2
O
(3) This interpretation is supported by the XPS results, which
2+
point to a low surface concentration of Co . The hydrogen
consumption between 950 and 1000 K could be due to the
reduction of a cobalt silicate phase, which could not be de-
tected by other techniques. Therefore, the similarity of the
The reduction profile of Co10-c shows a very intense
peak, which includes contributions at 558 and 587 K, asso-
ciated to the two reduction steps found in bulk Co O , al-
though shifted to lower temperatures as a consequence of
lower crystallite sizes of supported phase. Deconvolution of
the reduction peak into two components reveals that the
ratio between the hydrogen consumption for the first peak
and the second peak is 1:3, which is consistent with the
two-step reduction depicted in Equations (2) and (3) and
H consumption still
continues in the range 770–1000 K, and then tends to in-
crease again at higher temperature. The broad H consump-
tion peak in the range 770–1000 K could reasonably be at-
3
4
reduction profiles for Co10-c and bulk Co O up to 770 K
3
4
can be taken as an indication that most of the cobalt oxide
species in Co10-c interact weakly with the silica substrate.
Figure 7B shows the reduction profiles of calcined Fe10-
c and bulk Fe O as reference. The reduction of bulk hema-
2
3
tite (α-Fe O ) occurs via magnetite (Fe O ) and wustite
2
3
3
4
[
99,100]
agrees with literature findings.[
93–98]
(FeO) to zero-valent metallic iron.
FeO formation is
2
not observed as wustite is a metastable phase below
[
99,100]
0
843 K
and disproportionates into Fe O and Fe . Two
3 4
2
well-defined peaks are observed in the Fe O reduction pro-
2
3
file: the first is narrow, with a maximum at 654 K, and the
tributed to the reduction of cobalt species placed in the in-
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5065

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
second is wider and asymmetric, with a maximum at 886 K. ticles under a hydrogen atmosphere does not promote the
Thus, the two-stage reduction of Fe O can be described by formation of a metal–support interaction; (iv) a cobalt–iron
2
3
Equations (4) and (5).
interaction is observed in dimetallic systems after reduction,
and as a result of this interaction, the reducibility of cobalt
oxide is partly inhibited while that of iron oxide follows an
opposite trend. The XRD patterns and photoelectron and
Mössbauer spectra are conclusive for the formation of an
iron-cobalt alloy in the hydrogen-reduced samples.
3
Fe
2
O
3
+ H
2
Ǟ 3Fe
3
O
4
(4)
(5)
Fe
3
O
4
+ 4H
2
Ǟ 3Fe + 4H O
2
FeO formation is not detected under the reduction condi-
tions employed in this work. However, the reduction profile
of Fe10-c exhibits three peaks arising from the reduction of
iron species through the consecutive steps Fe O Ǟ Fe O
4
Experimental Section
2
3
3
[99]
0
Ǟ FeO Ǟ Fe according to Equations (6), (7) and (8).
Catalyst Preparation: Silica-supported cobalt- and/or iron oxides
were prepared by the wetness impregnation method on a silica car-
3
Fe
Fe
FeO + H
2
O
3
+ H
2
Ǟ 3Fe
3
O
4
(6)
(7)
(8)
2
–1
rier (Grace Davison; specific surface area of 310 m g , pore vol-
2
–1
ume of 1.22 m g ) with aqueous solutions of cobalt nitrate
Co(NO ·6H O, Merck reagent grade] and iron nitrate [Fe-
NO ·9H O, Merck reagent grade]. Monometallic cobalt and iron
3
O
4
+ H
2
Ǟ 2FeO
[
3
)
2
2
(
3
)
3
2
2
Ǟ 2Fe + H
2
O
catalysts were prepared by the appropriate concentration of the
above salts to achieve a final metal amount of 10 wt.-%. These
samples are labelled Co10-c and Fe10-c, respectively. Two dimet-
allic FeCo-supported catalysts containing a fixed amount of cobalt
The similarity in reduction temperatures for both bulk
Fe O and Fe10-c suggests a weak metal–support interac-
2
3
tion in the first reduction stage, as for Co10-c. The only
major difference between them is the absence of the re- CoFe10/1-c and CoFe10–5c, respectively) were prepared by simul-
duction stage of magnetite to wustite.
taneous impregnation with a solution containing both cobalt and
(
10 wt.-%) and different amounts of Fe (1 and 5 wt.-%, labelled
The reduction profiles of mono- and dimetallic cobalt iron salts. All impregnates were dried at 393 K overnight and cal-
cined at 773 K under flowing air for 2 h.
samples were also recorded and are compared in Figure 7C.
Quantitative analysis of these reduction profiles is difficult
Catalyst Characterisation: The cobalt content of the catalyst was
because the iron oxide reduction steps are overshadowed by determined by the ICP technique using a Perkin–Elmer Optima
the cobalt oxide ones. Nevertheless an interaction between 3300 DV apparatus. Specific areas were calculated using the BET
both oxides can be inferred from the changes in the re- method from the nitrogen adsorption isotherms, recorded at the
temperature of liquid nitrogen using a Micromeritics apparatus
duction of both iron- and cobalt oxides. It appears that the
2
(
model ASAP-2000) with a value of 0.162 nm for the cross-sec-
cobalt oxide reduction is inhibited by iron oxide — this is
more evident for CoFe10/5-c — while the iron oxide re-
duction is favoured by the presence of cobalt oxide. As can
2
tional area of the N molecule adsorbed at 77 K. Samples were
degassed at 423 K prior to adsorption measurements.
be seen by Mössbauer spectroscopy, a CoFe alloy is formed The powder XRD patterns of the precursor and calcined samples
were recorded with a Seifert 3000 P diffractometer using nickel-
in this sample upon reduction. Thus, it is likely that this
interaction between both metal oxides is responsible for al-
loy formation.
filtered Cu-Kα1 (λ = 0.15406 nm) radiation. A scanning step of
0.02° was taken between 5 and 80° Bragg angles.
An important observation for all reduction profiles is Temperature-programmed reduction (TPR) experiments were car-
ried out with a Micromeritics TPD/TPR 2900 apparatus equipped
that H2 consumption continues above 1100 K. As men-
with a thermal conductivity detector. Reduction profiles were ob-
tioned above, the species that are reduced at high tempera-
2
tained by passing a 10% H /Ar flow at a rate of 50 mL (STP) per
ture could be metallic silicates, i.e. Co SiO and Fe SiO , or
ionic species placed inside the support cavities.
2
4
2
4
minute through the sample (weight about 30 mg). The temperature
–
1
was increased from 300 to 1273 K at a rate of 10 Kmin , and the
amount of hydrogen consumed was determined as a function of
temperature. Under these conditions, the line profile and peak posi-
tion can be measured accurately. The effluent gas was passed
through a cold trap placed before the TCD in order to remove
water from the exit stream of the reactor.
Conclusions
A detailed analysis of the data derived from mono- and
dimetallic silica-supported cobalt-iron systems has allowed
us to draw the following conclusions: (i) the textural proper- SEM images were recorded with an ISI DS-130 microscope cou-
pled to a solid-state Si/Li Kevex detector and a SUN SparcStation
ties of the supported oxide phases are not influenced by the
5
for acquiring and processing energy-dispersive X-ray (EDX) spec-
calcination of the oxide precursors; (ii) the results obtained
by XRD, XPS and Mössbauer spectroscopy do not show
the development of a cobalt–iron interaction in the calcined
tra. Powder samples were converted into flat pellets and coated
with a thin graphite layer to prevent the accumulation of static
charge derived from the electron beam. Transmission electron
micrographs were recorded with a Fei Tecnai G30 microscope. The
acceleration voltage was set at 200 kV. The powdered sample was
first suspended in acetone, after which a drop of the suspension
oxides — the dimetallic systems only show separate Co O4
3
and Fe O phases and in no case are cobalt–iron mixed
2
3
oxides detected. The quantitative data also show that the
dispersion degree of the cobalt- and iron oxides is very low; was deposited on a copper grid covered with a fine carbon film
iii) pre-treatment of the supported cobalt oxide nanopar- evaporated under vacuum.
(
5066
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068

Co-Fe Oxide Nanoparticles Deposited on a Silica Substrate
FULL PAPER
Surface analysis was carried out with a VG Escalab 200R electron
spectrometer equipped with an Mg-K X-ray source and a hemi-
α
spherical electron analyser. The powder samples were pressed in 8-
mm-diameter copper troughs and then mounted on a sample rod
placed in a pre-treatment chamber and heated under vacuum at
[19] A. Corrias, M. Casula, G. Ennas, S. Marras, G. Navarra, G.
Mountjoy, J. Phys. Chem. B 2003, 107, 3030.
[
20] Chemierohstoffe aus Kohle (Eds.: C. D. Fröhning, H. Kölbel,
M. Ralek, W. Rottig, F. Schnur, H. Schulz J. Falbe), G. Thieme
Verlag, Stuttgart, 1977, p. 247.
[
[
[
21] A. K. Dalai, N. N. Bakhshi, M. N. Esmail, Ind. Eng. Chem.
Res. 1992, 31, 1449.
22] M. E. Dry, in Catalysis Science and Technology (Eds.: J. R. An-
derson, M. Boudart), Springer, Berlin, 1981, p. 159.
23] S. Zwart, R. Snel, J. Mol. Catal. 1985, 30, 305.
373 K for 1 h prior to being moved into the analysis chamber. The
base pressure in the analysis chamber was maintained below
4
ϫ10–9 mbar during data acquisition. The area under analysis was
2
about 2.4 mm , and the pass energy of the analyser was set at 50 eV,
for which the resolution as measured by the full width at half maxi-
mum (FWHM) of the Au 4f7/2 core level was 1.7 eV. The binding
energies were referenced to the C 1s peak at 284.9 eV due to adven-
titious carbon. Data processing was performed with the XPS peak
program, the spectra were decomposed with the least-squares fit-
ting routine provided with the software with Gaussian/Lorentzian
[24] T. V. Reshetenko, L. B. Avdeeva, A. A. Khassin, G. N. Kus-
tova, V. A. Ushakov, E. M. Moroz, A. N. Shmakov, V. V. Kriv-
entsov, D. I. Kochubey, Y. T. Pavlyukhin, A. L. Chuvilin, Z. R.
Ismagilov, Appl. Catal. A 2004, 268, 127.
[
[
25] D. J. Duvenhage, N. J. Coville, J. Mol. Catal. A 2005, 235, 230.
26] S. Hagen, R. Barfod, R. Fehrmann, C. J. H. Jacobsen, H. T.
Teunissen, I. Chorkendorff, J. Catal. 2003, 214, 327.
27] C. Pham-Huu, N. Keller, C. Estournes, G. Ehret, M. J. Ledoux,
Chem. Commun. 2002, 1882.
(90/10) product function and after subtracting a Shirley back-
[
ground. Atomic fractions were calculated using peak areas normal-
ised on the basis of sensitivity factors.[
101]
[
[
28] Y. Bi, A. K. Dalai, Can. J. Chem. Eng. 2003, 81, 230.
29] D. Banerjee, D. K. Chakrabarthy, Indian J. Technol. 1992, 30,
81.
Mössbauer spectra were recorded in a sinusoidal mode using a
57
transmission spectrometer with a Co/Rh source. Spectral analyses
[
[
30] H. Arai, K. Mitsuishi, T. Seiyama, Chem. Lett. 1984, 1291.
31] G. Ennas, M. F. Casula, A. Falqui, D. Gatteschi, G. Marongiu,
G. Piccaluga, C. Sangregorio, G. Pinna, J. Non-Cryst. Solids
were performed by non-linear fit using the NORMOS[
102]
program,
and energy calibrations were accomplished with an α-Fe (6 µm)
foil.
2
001, 239, 1.
32] M. F. Casula, A. Corrias, G. Paschina, J. Mater. Chem. 2002,
2, 1505.
[
[
1
Acknowledgments
33] A. Corrias, M. F. Casula, A. Falqui, G. Paschina, J. Sol-Gel
Sci. Technol. 2004, 31, 83.
A fellowship (VPO) granted by the Repsol-YPF Foundation is ac-
knowledged. This work was partly supported by the MCYT (Spain;
projects MAT2001-2215-C03-01 and ENE2004-07345-C03-01/
ALT).
[34] A. Hutlova, D. Niznansky, J. Plocek, J. Bursik, J.-L. Reh-
springer, J. Sol-Gel Sci. Technol. 2003, 26, 473.
[
[
35] T. Caillot, G. Pourroy, D. Stuerga, J. Solid State Chem. 2004,
77, 3843.
36] A. Malats i Riera, G. Pourroy, P. Poix, J. Alloys Compd. 1993,
02, 113.
1
2
[
1] S. L. González-Cortés, S. M. A. Rodulfo-Baechler, A. Oliveros,
J. Orozco, B. Fontal, A. J. Mora, G. Delgado, React. Kinet.
Catal. Lett. 2002, 75, 3.
[37] G. Pourroy, S. Läkamp, S. Vilminot, J. Alloys Compd. 1996,
244, 90.
[38] F. Tihay, G. Pourroy, M. Richard-Plouet, A. C. Roger, A. Kien-
nemann, Appl. Catal. A 2001, 206, 29.
[39] G. Jacobs, T. K. Das, Y. Zhang, J. Li, G. Racoillet, B. H. Davis,
Appl. Catal. A 2002, 233, 263.
[
[
[
[
[
[
[
[
[
[
[
[
2] A. J. Dalai, N. N. Bakhshi, M. N. Esmail, Fuel Process. Tech-
nol. 1997, 51, 219.
3] F. Tihay, A. C. Roger, G. Pourroy, A. Kiennemann, Energy
Fuels 2002, 16, 1271.
4] F. Tihay, A. C. Roger, A. Kiennemann, G. Pourroy, Catal. To-
day 2000, 58, 263.
5] C. Cabet, A. C. Roger, A. Kiennemann, S. Lakamp, G. Pour-
roy, J. Catal. 1998, 173, 64.
[40] J. Panpranot, J. G. Goodwin Jr, A. Sayari, J. Catal. 2002, 211,
530.
[41] S. Sun, K. Fujimoto, Y. Yoneyama, N. Tsubaki, Fuel 2002, 81,
1583.
[42] G. W. Huber, C. H. Bartholomew, Stud. Surf. Sci. Catal. 2001,
6] V. A. de la Peña O’Shea, N. N. Menéndez, J. D. Tornero,
J. L. G. Fierro, Catal. Lett. 2003, 88, 123.
136, 283.
[
[
[
[
[
[
43] S. Ali, B. Chen, J. G. Goodwin, J. Catal. 1995, 157, 35.
44] R. L. Chin, D. M. Hercules, J. Phys. Chem. 1982, 86, 360.
45] H. M. Bruce, G. Baker, Appl. Catal. A 1995, 123, 23.
46] E. Coulter, A. G. Sault, J. Catal. 1995, 154, 56.
47] S. Sun, L. Fan, K. Fujimoto, Chem. Lett. 1999, 28, 343.
48] E. Iglesia, S. C. Reyes, R. J. Madon, S. T. Soled, Adv. Catal.
7] L. Pettersson, K. Sjöström, Combust. Sci. Technol. 1991, 80,
265.
8] J. Agrell, B. Lindstroem, L. J. Pettersson, S. Jaras, Catalysis
002, 16, 67.
2
9] E. Manova, T. Tsoncheva, D. Paneva, I. Mitov, K. Tenchev, L.
Petrov, Appl. Catal. A 2004, 222, 119.
1
993, 39, 221.
49] E. Iglesia, S. L. Soled, R. A. Fiato, G. H. A. Via, J. Catal. 1993,
43, 345.
10] S. Hagen, R. Barfod, R. Fehrmann, C. J. H. Jacobsen, H. T.
Teunissen, I. Chorkendorff, J. Catal. 2003, 214, 327.
11] Z. Kóya, I. Vesseléyi, K. Lázár, J. Kiss, I. Kiricsi, IEEE Trans.
Nanotechnol. 2004, 3, 73.
12] N. Nagaraju, A. Fonesca, Z. Konya, J. B. Nagy, J. Mol. Catal.
A 2000, 181, 57.
[
1
[
[
[
50] R. C. Reuel, C. H. Bartholomew, J. Catal. 1985, 85, 63–77.
51] R. C. Reuel, C. H. Bartholomew, J. Catal. 1985, 85, 78–88.
52] B. A. Sexton, A. E. Hughes, T. W. Turney, J. Catal. 1986, 97,
390.
13] K. Hernadi, A. Fonseca, J. B. Nagy, B. Bernaerts, A. A. Lucas,
Carbon 1996, 34, 1249.
[
[
53] D. B. Bukur, K. Okabe, M. P. Rosynek, C. Li, D. Wang,
K. R. P. M. Rao, G. P. Huffman, J. Catal. 1995, 155, 535.
54] C. S. Kuivila, P. C. Stair, J. B. Butt, J. Catal. 1989, 118, 299–
311.
[
[
[
[
14] D. J. Duvenhage, N. J. Coville, Appl. Catal. A 2005, 289, 231.
15] M. Sugimoto, J. Am. Ceram. Soc. 1999, 82, 269.
16] G. Bate, J. Magn. Magn. Mater. 1999, 100, 413.
17] R. C. Pedroza, S. W. da Silva, M. G. Soler, P. P. C. Sartoratto,
D. R. Rezende, P. C. Morais, J. Magn. Magn. Mater. 2005, 289,
[55] B. H. Davis, Catal. Today 2003, 84, 83.
[56] B. K. Sharma, M. P. Sharma, S. Kumar, S. K. Roy, S. K. Roy,
S. Badrinarayanan, S. R. Sainkar, A. B. Mandale, S. K. Date,
Appl. Catal. A 2001, 211, 203.
139.
[
18] X. Huang, Z. Chen, Solid State Commun. 2004, 132, 845.
Eur. J. Inorg. Chem. 2006, 5057–5068
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5067

V. A. de la Peña O’Shea and J. L. G. Fierro et al.
FULL PAPER
[
57] B. Ernst, A. Bensaddik, L. Hilarie, P. Chaumette, A. Kinne-
[81] Y. Okomoto, T. Imanaka, S. Teranishi, J. Catal. 1980, 65, 44.
[82] B. K. Sharma, M. P. Sharma, S. Kuma, S. K. Roy, S. Badrina-
rayana, S. R. Sainkar, A. B. Mandale, S. Date, Appl. Catal. A
2001, 211, 203.
man, Catal. Today 1998, 39, 329.
[
[
58] F. P. Larkins, A. Z. Khan, Appl. Catal. 1989, 47, 209.
59] K. Jothimurugesan, J. G. Goodwin Jr, K. Santosh, K. Ganwal,
J. J. Spivey, Catal. Today 2000, 58, 355.
[83] J. Łojewska, W. Makowski, T. Tyszewski, R. Dziembaj, Catal.
Today 2001, 69, 409.
[
[
60] D. J. Duvenhage, N. J. Coville, Appl. Catal. A 1997, 15, 43.
61] R. Riva, H. Miessner, R. Vitali, G. del Piero, Appl. Catal. A
[84] J. P. Baltrus, J. R. Dile, M. A. McDonald, M. F. Zarochak,
Appl. Catal. 1989, 48, 199.
2
000, 196, 111.
62] H. M. Gager, M. C. Hobson, Catal. Rev. Sci. Eng. 1975, 11,
17.
[
[
[
[
85] M. Oku, K. J. Hirokawa, Appl. Phys. 1979, 50, 6303.
86] J. B. Lambert, L. Xue, J. M. Weydert, J. H. Winter, Archaeome-
try 1990, 32, 47.
1
63] X. Gao, J. Shen, Y. Shia, Y. Chen, J. Chem. Soc., Faraday
Trans. 1993, 89, 1079.
[
87] R. V. Siriwardene, J. M. Cook, J. Colloid Interface Sci. 1985,
[
[
64] A. M. van der Kraan, Phys. Status Solidi A 1973, 18, 215.
65] M. C. Hobson Jr, H. M. Gager, J. Colloid Interface Sci. 1970,
108, 414.
[
88] D. D. Hawnm, B. M. Dekoven, Surf. Interface Anal. 1987, 10,
34, 357.
63.
[
[
66] M. C. Hobson, A. D. Campbell, J. Catal. 1967, 8, 294.
67] S. Yuen, Y. Chen, J. E. Kubsh, J. A. Dumesic, N. Topsøe, H.
Topsøe, J. Phys. Chem. 1982, 86, 3022.
68] A. F. H. Wielers, A. H. J. M. Dock, C. E. C. A. Hop, J. W.
Geus, A. M. van der Kraan, J. Catal. 1989, 117, 1.
69] S. C. Bjerne, H. Topsøe, S. Mørup, Appl. Catal. 1989, 48, 327.
70] M. V. Cagnoli, S. G. Marchetti, N. G. Gallegos, A. M. Alvarez,
A. C. Mercader, A. A. Yeramian, J. Catal. 1989, 123, 21.
71] W. Kündig, J. A. Cape, R. H. Lindquist, G. J. J. Constabaris,
Appl. Phys. 1967, 38, 947.
[
[
89] D. J. Duvenhage, N. J. Coville, Appl. Catal. A 2002, 233, 63.
90] Y. Okamoto, T. Nagata, T. Adachi, T. Imanaka, K. Inamura,
T. Takyu, J. Chem. Phys. 1991, 95, 310.
[
[91] G. J. Haddad, J. G. Goodwin, J. Catal. 1995, 157, 56.
[92] E. van Steen, G. S. Sewell, R. A. Makhothe, H. Micklethwaite,
H. Manstein, M. de Lange, C. T. O’Connor, J. Catal. 1996,
162, 220.
[
[
[
[
[
[
[93] D. G. Castner, P. R. Watson, I. Y. Chan, J. Phys. Chem. 1990,
94, 819.
72] G. M. Bancroft, A. G. Maddock, R. G. Burns, Geochim. Co-
smochim. Acta 1967, 31, 2219.
[94] M. P. Rosynek, C. A. Polanky, Appl. Catal. 1991, 73, 97.
[95] L. B. Backman, A. Rautiainen, A. O. I. Krause, M. Lindblad,
Catal. Today 1998, 43, 11.
[96] S. Bessell, Appl. Catal. A 1993, 96, 253.
[97] J. G. Choi, Catal. Lett. 1995, 35, 291.
[98] H. Ming, B. G. Baker, Appl. Catal. A 1995, 123, 23.
[99] A. J. H. M. Kock, H. M. Fortuin, J. W. Geus, J. Catal. 1985,
96, 261.
[100] X. Gao, J. Shen, Y. Hsia, Y. Chen, J. Chem. Soc., Faraday
Trans. 1993, 89, 1079.
[101] C. D. Wagner, L. E. Davis, M. V. Zeller, J. A. Taylor, R. H.
Raymond, L. H. Gale, Surf. Interface Anal. 1981, 3, 211.
[102] R. A. Brand, Nucl. Instrum. Methods Phys. Res., Sect. B 1987,
28, 398.
73] C. E. Johnson, M. S. Ridout, T. E. Cranshaw, P. E. Madsen,
Phys. Rev. Lett. 1961, 6, 450.
74] J. P. Pinheiro, P. Gadelle, C. Jeandey, J. L. Oddou, J. Phys.
Chem. Solids 2001, 62, 1023.
[
[
[
75] R. M. Stanfield, W. N. Delgass, J. Catal. 1981, 72, 37.
76] Y. Y. Huang, J. R. Anderson, J. Catal. 1975, 40, 143.
77] B. Ernst, S. Libs, P. Chaumette, A. Kiennemann, Appl. Catal.
A 1999, 189, 145.
[
[
[
78] A. Bensaddik, B. Ernst, F. Garin, G. Maire, A. Kiennemann,
J. Phys. IV 1996, 6, 645.
79] B. J. Tan, K. J. Klabunde, P. M. A. Sherdwood, J. Am. Chem.
Soc. 1991, 113, 855.
80] T. J. Chuang, C. R. Brundle, D. W. Rice, Surf. Sci. 1976, 59,
Received: August 23, 2006
Published Online: October 24, 2006
413.
5068
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 5057–5068