Mesoporous MAl2O4 (M = Cu, Ni, Fe or Mg) spinels: Characterisation and application in the catalytic dehydrogenation of ethylbenzene in the presence of CO2

Article abstract of DOI:10.1016/j.apcata.2010.04.027

The catalytic properties of nanostructured MAl₂O₄ (M = Cu, Ni, Fe or Mg) were investigated in the dehydrogenation of ethylbenzene with CO₂. The catalysts were characterised by XRD, Raman spectroscopy, textural properties,

Full text of DOI:10.1016/j.apcata.2010.04.027

Applied Catalysis A: General 382 (2010) 148–157
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Mesoporous MAl2O4 (M = Cu, Ni, Fe or Mg) spinels: Characterisation and
application in the catalytic dehydrogenation of ethylbenzene in the presence
of CO2
Andressa H. de Morais Batistaa, Francisca S.O. Ramosa, Tiago P. Bragaa, Cleanio L. Limab,1
,
Francisco F. de Sousab,1, Eduardo B.D. Barrosb,1, Josue M. Filhob,1, Aldenor S. de Oliveirac,2
,
Jackson R. de Sousac,2, Antoninho Valentinia, Alcineia C. Oliveiraa,∗
a Universidade Federal do Ceará, Campus do Pici-Bloco 940, Langmur Lab. de Adsorc¸ ão e Catálise, 60.000.000, Fortaleza, Ceará, Brazil
b Universidade Federal do Ceará, Campus do Pici-Bloco 922, Departamento de Física, Fortaleza, Ceará, Brazil
c Universidade Federal do Ceará, Campus do Pici-Bloco 922, Departamento de Química Orgânica e Inorgânica, Fortaleza, Ceará, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 March 2010
Received in revised form 9 April 2010
Accepted 13 April 2010
Available online 21 April 2010
The catalytic properties of nanostructured MAl2O4 (M = Cu, Ni, Fe or Mg) were investigated in the dehy-
drogenation of ethylbenzene with CO2. The catalysts were characterised by XRD, Raman spectroscopy,
textural properties, acidity (TPD-NH3) and basicity (CO2-TPD) measurements, SEM and TEM analyses and
TPR techniques. XRD revealed the spinel structure of MAl2O4, except for the NiO, CuO, Fe2O3 and MgO
phases. All solids exhibited the nanostructured features of the SBA-15 template, including high values of
textural properties and morphologies characteristic of the mesoporous silica. The acidic strength follows
the electronegativity trends of the cations present on the spinel aluminates: CuAl > NiAl > FeAl and MgAl,
which is the inverse order of the basicities. The selectivities in the dehydrogenation of ethylbenzene
with CO2 showed that styrene was the predominant product for all catalysts except for those based on
Ni, which was highly selective for the production of toluene. The FeAl2O4 catalyst provided the best cat-
alytic performance among the solids studied due to the continuous oxidation of Fe3+ sites by the CO2 from
the reaction. This solid has the additional advantage of being stable over the timescale of the reaction, as
compared to the traditional Fe-based catalysts, due to its nanostructured features.
Keywords:
Ethylbenzene
Spinels
Raman
Metals
CO2
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
include a substantial decrease in the amount of energy required and
higher equilibrium yields of styrene.
The industrial synthesis of styrene is generally performed via
the dehydrogenation of ethylbenzene (EB) in the presence of super-
heated steam (reaction I), using Fe–K–Cr oxide-based catalysts at
high temperatures (e.g., 600–700 ◦C) [1]. A broad variety of materi-
als derived from the styrene that is now produced by this reaction
find key uses as intermediates in the synthesis of fibres, monomers
for the production of polystyrene and other applications [2,3].
C6H5CH2–CH3 + CO2 → C6H5CH CH2 + CO + H2O
(II)
Various catalysts have been used in the dehydrogenation of EB
with CO2, includingiron based oxides, hydrotalcite-likeoxides, zeo-
lites, spinel oxides, active carbon, pillarised clays, active carbon
containing promoters (Li, Be, Na, Mg, K, Ca) and carriers for actives
components (Fe, V, Cr, Cu, Zn, Zr, Ce, La, Ni, Co) [3–16].
Among these catalysts, the iron oxides (i.e., ␣-Fe2O3) are less
active in the reaction. Specifically, the initial yield of EB in the pres-
ence of CO2 is often lower than that in the presence of inert gases
over these catalysts [9,10,17]. Hence, the development of catalysts
with improved activities and stabilities is much needed.
C6H5CH2–CH3 → H2 + C6H5CH CH2
(I)
Currently, the dehydrogenation of ethylbenzene in the presence
of CO2 (reaction II) has been extensively investigated because of its
well-documented advantages over the steam-based process, which
In this context, previous studies have shown that spinels con-
taining transition metal ions can act as efficient catalysts for the
dehydrogenation of EB in the presence of CO2 [8,10,17–19]. It
is thought that the spinel structure provides resistance to the
reducibility and/or coking provided by the reaction media. More-
over, aluminium in an aluminate matrix was proved to be an
∗
Corresponding author. Tel.: +55 85 33 66 90 41/9982;
fax: +55 85 33 66 90 41/9982.
E-mail address: alcineia@ufc.br (A.C. Oliveira).
Tel.: +55 85 3366 90 08; fax: +55 85 3366 90 08.
Tel.: +55 85 3366 9412; fax: +55 85 3366 9412.
1
2
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.04.027

A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
149
excellent structural promoter, preventing sintering of the solid and
stabilising the active state of the Ni or Fe in a MAl2O4 catalyst
[2,18–21]. In addition, alkali earth metals such as Ca and Mg have
been introduced as promoters and the latter metal has consider-
ably improved the activity of the RGWS coupled reaction (reaction
III) because CO2 is activated in the basic sites provided by Mg while
EB is activated by the acidic sites over Al [1]
because the solution of the precursors can efficiently improve the
filling and homogeneity of the resultant replicas [29–31]. However,
these spinel oxides have been scarcely used in the dehydrogenation
of EB with CO2.
Here, we report the synthesis and characterisation of MAl2O4
(M = Mg, Cu, Ni or Fe) nanocasted from SBA-15 and the catalytic
properties of the complexes in the dehydrogenation of EB with
CO2. The spinel oxides were formed inside the mesochannels of the
SBA-15 silica, a well-known hard template, through a mild thermal
process and MAl2O4 oxides were obtained after removal of the sil-
ica template by treatment with HF. Based on these results, the roles
of the spinel structure and, deactivation mechanisms are proposed.
CO2 + H2 → CO + H2O
(III)
Although the MAl2O4 system has a favourable framework for
this particular structure and displays good activity in the dehydro-
genation of EB with CO2, the lifetime of the active sites and, more
importantly, its selectivity for styrene are the major drawbacks. For
these reasons, there are limitations for the use of spinel oxides in
this reaction.
2. Experimental
Ye et al. and Valente et al. [15,22], respectively have synthesised
Mg-M-Al (M = Fe, Cu, Ni) catalysts obtained from hydrotalcites as
precursors. After calcination of the catalyst to obtain spinel oxides,
the properties of the oxides, such as high metal dispersion, large
surface area and porosity, and good thermal stability, resulted in
better catalytic behaviours.
Therefore, for the dehydrogenation of EB in the presence of CO2
and catalysed by spinel oxides, the accessible porosity, the homoge-
neous distribution of the cations in the structures, the stabilisation
of the active components and the modulation of the acid–base char-
acter are relevant properties that can be adjusted depending on
the preparation method used. Until now, spinel oxides possessing
diverse structural/morphological characteristics have been synthe-
sised by nitrate decomposition, sol–gel technology, hydrothermal
synthesis, aerogel pyrolysis, anodic electrodecomposition [23] and
so forth. The multicomponent spinel oxides can also be obtained by
a nanocasting route that seems more attractive and provides supe-
rior properties (e.g., thermal stability and porous properties) than
those achieved by other methods.
The nanocasting pathway provides a promising route for the
preparation of porous transition metal-based materials, such as
spinel oxides, with novel framework compositions and well-
controlled mesostructures [24,25]. This method consists of the
inorganic condensation reaction that occurs in the voids of a rigid
structure mold or on its outer surface (i.e., silica, carbon, among
others) [26].
The metal oxide/mold composite material is then obtained by
thermal decomposition of the precursors, and finally a replica of the
mesoporous transition metal can be obtained after removal of the
mold template by leaching [27]. Highly ordered mesoporous oxides
of TiO2, CeO2, SnO2, Fe2O3, NiO [25,27] and mixed oxides like Fe–Cr
and Mn–Co have successfully been nanocasted using mesoporous
silica as a template [22,28–30].
Among these mixed oxides, spinel aluminates (MAl2O4; M = Mg,
Cu, Ni, Mn, Zn, Fe) prepared by this technique are interesting mate-
rials for a broad range of applications including catalysts, sensing
and electronic devices [27–31]. Specifically, in the field of catalysis,
the high catalytic efficiency of the nanocrystalline aluminate oxide
spinels is owed to their large surface-to-volume ratios, narrow
mesopore diameters and specific material morphologies that differ
from their bulk counterparts. These features are highly suitable for
processes that take place at boundaries between solid particles and
the gas phase.
However, the synthesis of the mesoporous metal aluminate
spinel oxides by a nanocasting route is not a trivial task due to
complex events related to the control of the composition, stoi-
chiometry and/or crystal structure; consequently, the metal oxides
obtained are poorly ordered because of the inherent limitations of
the method. Recent findings demonstrate that it is possible to use
the hard template nanocasting route to obtain mesoporous per-
ovskite and spinel types. These oxides contain high surface areas
2.1. Synthesis of the mesoporous SBA-15
The synthesis of mesoporous SBA-15 was performed under
hydrothermal conditions according to the procedure described
elsewhere [32]. In a typical synthesis, approximately 4.0 g of tri-
block copolymer P123 (Aldrich 98 wt%) was dissolved in 105 mL of
water and 20 mL of concentrated HCl (Vetec, 37 wt%). After 30 min
of stirring, about 8.5 g of tetraethoxysilane (TEOS, Aldrich 98 wt%)
was added to the previous mixture and the reaction was stirred at
40 ◦C for 24 h. The mixture was then submitted to hydrothermal
treatment at 100 ◦C for 72 h under static conditions. The product
obtained was filtered, dried and subsequently calcined under air-
flow at 550 ◦C for 6 h with a heating rate of 1 ◦C min−1
.
2.2. Synthesis of MAl2O4 via the nanocasting pathway
The two-solvent impregnation method was used to obtain
the MAl2O4 oxides [23]. Briefly, 10 g of a mixture of nickel and
aluminium nitrates (Ni(NO3)2·6H2O, and Al(NO3)2·9H2O,Vetec,
98 wt%) in a molar ratio of 1:2 were dissolved in distilled water
to obtain a 1.0 g mL−1 solution of the mixed nickel and aluminium
nitrate. A 1.0 g of the mesoporous SBA-15 was dispersed in 40 mL
of n-hexane (Synth, 99 wt%) with stirring for 3 h. Then, 1 mL of the
nitrate solution was added to the latter mixture dropwise with vig-
orous stirring for an additional 24 h at room temperature. After this
period passed, the silica composite powder was calcined at 600 ◦C
with a heating rate of 1 ◦C min−1 and maintained at that tempera-
ture for 6 h under nitrogen flow. The powder obtained was leached
with a 1.5 mol L−1 sodium hydroxide solution to remove the SBA-15
silica and further washed with ethanol, filtered and dried to obtain
the resultant NiAl sample. Similarly, the other MAl2O4 spinel oxides
(M = Cu, Fe, or Mg) were obtained using copper, iron and magne-
sium nitrate sources. The samples were designated as CuAl, FeAl
and MgAl, respectively.
2.3. Characterisation
The catalysts were characterised by chemical analysis, X-ray
diffraction (XRD), N2 adsorption isotherms, scanning electron
microscopy (SEM), temperature programmed reduction (TPR),
temperature programmed desorption of ammonia (NH3-TPD),
Raman spectroscopy, thermogravimetric analyses (TGA) and tem-
perature programmed desorption of CO2 (CO2-TPD) techniques.
Chemical analysis of the solids was performed by inductively
coupled plasma (ICP) optical emission spectroscopy (ICP-OES) in a
Varian equipment. A 0.1 g of the sample was dissolved in 40 wt%
hydrofluoric acid solution and heated at 200 ◦C in a sand bath.
The solution obtained was evaporated and dried, and a solution
of 3.4 mL of nitric acid and 10 mL of water were added, prior to the
measurements.

150
A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
Fig. 1. (a) Low-angle diffractograms of the samples. (b) Wide-angle XRD diffractograms of the mesoporous MeAl2O4 spinel oxides (Me = Cu, Ni, Fe or Mg) after removal of
the SBA-15.
Low-angle X-ray diffraction (XRD) patterns were obtained using
a Rigaku X-ray diffractometer (Rigaku model) apparatus with a
CuK␣ radiation at 30 mA and 40 kV. Also, wide-angle measure-
ments (ca. 2ꢀ = 3–80◦) were recorded on the same equipment.
for 2 h to remove weakly adsorbed ammonia. The temperature was
increased to 500 ◦C at a heating rate of 5 ◦C min−1 to desorb ammo-
nia.
Raman spectra of the selected spent catalysts were obtained
on a T64000 Raman spectrometer (JobinYvon triple spectrome-
ter) under ambient conditions. A 514.5 nm Ar laser was used as
the exciting source on the sample surface and a power of 20 mW.
The measurements were referenced to Si at 521 cm−1 with 16 data
acquisitions in 120 s. The lens focus was of 100 times.
The amount of carbon on spent catalysts was measured by TGA
experiments using Netzsch STA 409 PC/PG equipment coupled to
a Bruker Tensor 27 IR instrument. The measurements were per-
formed on spent catalysts to know the coking degree of the spent
samples from room temperature to 1000 ◦C at 10 ◦C min−1, under
airflow by placing approximately 15 mg of the sample in an alu-
minium pan.
The unit cell parameters were calculated using the formula a0
2d2 1 0 5 where d1 0 0 represents the d-spacing value of the (1 0 0)
reflection.
=
√
Nitrogen adsorption–desorption isotherms were measured
with an ASAP2000 adsorption analyser (Micromeritics) at −196 ◦C
nitrogen temperature. Prior to the analyses, the samples were
degassed at 150 ◦C for 6 h under vacuum. The surface area and the
total pore volume were obtained using the BET method and were
calculated from the adsorption data in the relative interval from
0.05 to 0.2 of the adsorption branch of the isotherm at P/Po = 0.99.
Pore diameters were calculated by BJH (Barret–Joyner–Halenda)
method from the adsorption branch.
Scanning electron microscopy (SEM) measurements were con-
ducted on a JEOL JEM-2010 electron microscope equipped with an
EDX Link Analytical QX-20000 system coupled to the SEM micro-
scope, using an acceleration voltage of 200 kV.
2.4. Catalytic test
The solids were evaluated in the dehydrogenation of ethyl-
benzene (EB) in the presence of carbon dioxide. The reaction was
carried out in a fixed-bed tubular stainless steel reactor where 0.1 g
of catalyst was placed. Previously, the samples were activated in
situ under a nitrogen flow from room temperature to 550 ◦C over
the course of 1 h to purge all gaseous impurities from the cata-
lyst surface. The reaction was started by introducing a gas mixture
of ethylbenzene (Fluka, 99.99 wt%), nitrogen (Aga, 99.99 wt%) and
carbon dioxide (Aga, 99.99 wt%) to the reactor. The ethylbenzene
(ca. 1.98 mmol h−1) was fed by bubbling a gas mixture of nitro-
gen (11.7 mmol h−1) and carbon dioxide (58.7 mmol h−1) into the
liquid ethylbenzene in a molar ratio of CO2/EB = 30. The reaction
was conducted at atmospheric pressure at 550 ◦C. The products of
the reaction (styrene, benzene and toluene) were condensed and
analysed by a gas chromatograph (Simple Chrom model) equipped
with a flame ionisation detector. The gaseous effluents (CO2 and
CO) were analysed in a thermal conductivity gas detector chro-
matograph (Ciola model). The ethylbenzene conversion, styrene
selectivity were expressed as mol.% on the basis of carbon atoms, as
suggested by the findings [33]. Activities were measured as moles
of ethylbenzene consumed per gram of catalyst per hour and were
Temperature programmed reduction (TPR) analysis of the cata-
lysts was performed in a home-made equipment using a quartz
tube reactor possessing a inner diameter of 6 mm and coupled
to a TCD for monitoring the hydrogen consumption. The mass of
catalyst was 20 mg and the experiment was lead in the range of
50–1000 ◦C. An 8% H2/N2 mixture, as reducing gas was used with a
rate of 100 mL min−1, after passing through a 13× molecular sieve
trap to remove water. Before the analysis, samples of ca. 0.1 g were
placed in tube reactor and heated under nitrogen at 100 ◦C for 2 h.
The temperature programmed desorption of CO2 (CO2-TPD)
measurements was carried out in the same equipment, under
nitrogen, as carrier (flow rate of 50 mL min−1). Prior to the CO2
adsorption, all catalysts were heated to 300 ◦C. Samples were then
cooled to 60 ◦C and exposed to a flow of CO2 for 1 h. Desorption
proceeded by heating (10 ◦C min−1) up to the final temperature of
950 ◦C.
Temperature programmed desorption of ammonia (NH3-TPD)
was performed in a Shimadzu TG/DTA 50 apparatus. For these
measurements, 50 mg of the samples were pretreated at 500 ◦C
for 2 h and then cooled to 120 ◦C in He flow. Pure ammonia was
injected until adsorption saturation followed by purging with He
expressed in mol2 g−1 h−1
.

A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
151
3. Results and discussion
3.1. Structure and textural properties of the solids
The low-angle XRD pattern of the template-free NiAl spinel
oxide is depicted in Fig. 1a. All diffraction patterns show the char-
acteristic broad diffraction peak at 1◦, corresponding to (1 0 0)
reflexion of the SBA-15 mesoporous 2D-hexagonal symmetry of
the pores with the space group of P6 mm. The similarity of the XRD
patterns possessing a weak intensity compared to that of the SBA-
15 suggests less structural order in the replicated material. This
result is probably due to the incomplete pore filling of the silica
template during the decomposition of the metal nitrate precursor
[25,27]. Indeed, removal of the silica template could lead to a par-
tial collapse of the structure, and thus the periodicity of the spinel
replica, as observed elsewhere [23].
The unit cell parameters calculated (Table 1) are close to
10.5 nm, which match well with that of the SBA-15 template (ca.
10.9 mm) [32] and suggest that any significant structural shrinkage
occurred during the replication process of the solids at 600 ◦C.
Fig. 1b shows the wide-angle XRD patterns of the MAl2O4 spinel.
It is clear that the diffraction peaks of the samples are significantly
broadened as in case of the NiAl and MgAl samples. They are also
weak in intensity, indicating that the framework walls of the spinel
replica are composed of crystalline nanoparticles, probably due to
their growth in a confined space.
All diffraction lines could be unequivocally assigned to a spinel-
type structure with a space group of Fd3m [32]. However, the
diffractogram of the samples shows peaks corresponding to CuO,
NiO, MgO and Fe2O3, in addition to the spinel-type lines, for the
CuAl, NiAl, MgAl, and FeAl samples, respectively. These peaks could
be rationalised by considering the variance in the cell unit param-
eter, which will depend mainly on the ionic radii of the cations in
the octahedral coordination. The ionic radii of the cations decrease
in the following order: Cu2+ > Mg2+ > Ni2+ > Fe3+ > Al3+. Thus, the
Fe3+ will probably be randomly distributed in octahedral sites
of the spinel oxide, owing to the smaller ionic radius of Fe3+
(0.49 Å) compared to those of Mg2+ (0.72 Å), Ni2+ (0.69 Å) and
Cu2+ (0.73 Å) [34,35]. Al3+ possesses an ionic radius of 0.39 Å and
will probably occupy a tetrahedral position in a spinel structure.
Accordingly, the large percentage of spinel phases detected for
FeAl and NiAl samples confirms the increased tendency for the iso-
morphous substitution of M2+ or M3+ cations in octahedral sites
of an Al2O4 matrix during the formation of spinel oxides as the
ionic radius decreases. In addition, the a0 values (Table 1) increase
upon increasing the ionic radii of the cations, which indicates a
possible isomorphous substitution of the M cation in the MAl2O4
structure.
Fig. 2. N2 adsorption isotherms of the spinel oxides.
a high relative pressure. This suggests that the metal oxides have
mesoporous character similar to that of the SBA-15 mold.
The adsorption of nitrogen increases rapidly at a relative
pressure close to unity, which is indicative of the disordered inter-
particle voids generated during the thermal transformation of
precursors to form the replica [36]. The corresponding values of
the BET surface area are shown in Table 1.
These replica materials display high surface areas, ranging from
118 to 88 m2 g−1, and large pore volumes (ca. 0.32–0.42 cm2 g−1), a
characteristic of hexagonally arranged mesoporous materials. This
indicates that the replicas were successfully retained and their tex-
tural properties are close to those of oxide replicas reported in the
literature [36]. Specifically, for the MgAl sample, the shape of the
isotherms suggests the formation of macropores in this solid. Dif-
ferences in surface area and variations in the mesopore volumes
depend on the specific density of the different metal oxides and
the number of impregnations of the metal precursor [32,36], but in
our case only one impregnation step is performed.
The chemical composition determined by chemical analysis
shows that the M2+/M3+ molar ratio values agree well with those
of the theoretical ratio (e.g., 0.60). In the case of FeAl, the Fe2+/Al3+
ratio is close to the expected experimental values (e.g, 0.50).
Fig. 2 shows the N2 adsorption–desorption isotherms for the
four samples studied. These isotherms are akin to type IV of the
IUPAC classification [23,24], with a pronounced hysteresis loop at
Additionally, all samples display a unimodal distribution of
pores with an average pore diameter of about 4.3 nm for all solids.
The structural collapse during the thermal treatment to obtain the
spinel oxides results in the textural pores (interparticle voids), and
the sizes of these interparticle voids depend on the small crystal
size of the solids [37] and this is further observed by TEM anal-
ysis. Interparticle porosity thereby contributes largely to the total
Table 1
Physicochemical properties of the catalysts studied.
a
Sample
a0 (nm)
Phases detectedb
Chemical compositionc
Sg (m2 g−1
)
Vp (cm3 g−1
)
Dp (nm)
NiAl
CuAl
FeAl
MgAl
10.8
11.3
10.3
10.5
57NiO; 43NiAl2O4
72CuO; 28CuAl2O4
23Fe2O3; 77FeAl2O4
64MgO; 36MgAl2O4
0.57
0.56
0.52
0.60
96
88
107
118
0.34
0.32
0.39
0.42
4.7
4.2
4.5
4.6
√
a
From a0 = 2d2 1 0 5.
b
c
After Rietvield refinement.
Chemical analysis of M2+/ M3+ or Fe2+/ Al3+
.

152
A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
Fig. 3. SEM images of the catalysts studied and TEM micrograph of the NiAl catalyst. The inset figure is the SAED pattern.
FeAl2O4 + H2 → Al3O3 + H2O + Fe2+xO
Fe2+xO + H2 → Fe0 + H2O
The TPR curve of NiAl shows a strong peak centred at ca. 470 ◦C.
This suggests that large particles of NiO are weakly interacting with
the NiAl2O4 phase and the reduction of the former is facilitated
[8,37,38] while the reduction of the NiAl2O4 phase is obtained at
relatively high temperatures (ca. 550 ◦C).
(VI)
pore volume. Relevant features of our materials are the high surface
areas, pore volumes and diameters found for all samples, proper-
ties that cannot be achieved by more conventional spinel oxide
synthetic methods.
(VII)
3.2. Morphological aspects
Fig. 3 shows the SEM images of the nanocasted spinel oxides
obtained. They consist of aggregated particles with irregular
shapes, which may be a result of the aforementioned structural
collapse caused by incomplete filling of the metal sources. Energy
dispersive spectra of different areas indicated removal of the silica,
as <1.8 wt% of SiO2 was found.
The TPR profile of CuAl shows two reduction peaks correspond-
ing to the reduction of Cu2+ ions. Since a large amount of the CuO
Particularly, the TEM image of NiAl exhibits the void (indicated
by an arrow) that is evidence that isolated nanoparticles are formed
because of the loss of structural fidelity in the silica template as
observed elsewhere [23]. Indeed, the SAED pattern of NiAl (Fig. 3,
inset) shows the lattice fringe distances that correspond to the
(1 1 1) and (2 2 0)-spacing of cubic spinel structures [36].
3.3. TPR studies
The TPR studies were conducted to ascertain the reducibility
aspects of the solids. Two reduction peaks in the profiles of FeAl
are observed (Fig. 4).
At the first TPR peak, a shoulder at about 400 ◦C appears, indi-
cating that the isolated ␣-Fe2O3 is reduced to Fe2+ and then to Fe0
[20,22] (reactions (IV) and (V)). The second peak has a broad feature
(≈440–670 ◦C) and corresponds to the hard reduction of FeAl2O4 to
Fe2+ and further to Fe0 (reactions (VI) and (VII))[10].
␣-Fe2O3 + H2 → Fe3O4 + H2O
Fe3O4 + H2 → Fe0 + H2O
(IV)
(V)
Fig. 4. H2-TPR profiles of the samples.

A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
153
phase was detected (from XRD analysis), it can be assumed that
the Cu2+ is mostly reduced to Cu0 at low temperatures (e.g., 250 ◦C,
first peak). The high temperature peak (ca. 450 ◦C, second peak) is
ascribed to the reduction of the CuAl2O4 phase in agreement with
previous findings [8].
The TPR peaks of the MgAl have low intensity compared to those
of the other samples. This result points to the hard reducibility of
the MgAl2O4 phase. The TPR spectra of the MgAl2O4 sample show
only one peak centred at 500 ◦C, which is indicative of Mg2+ ions
from decomposition of the MgAl2O4 phase or a lack of reducibility
in the sample until 1000 ◦C [39].
Based on these results, it can be inferred that the pure oxides
(NiO, CuO, ␣-Fe2O3) that exist concomitant to the spinel oxides, are
favourably reduced due to their weak interaction with the spinel
oxide phases. Indeed, the results point to the relative stability of
MgAl2O4, FeAl2O4 and CuAl2O4 in a hydrogen environment. In addi-
tion, the restricted reduction of MgAl2O4 may be a consequence of
the stability of the spinel phase compared to those of the other
spinel analogues.
Fig. 5. CO2-TPD curves of the spinel oxides studied.
3.4. Acid–base properties
coordinations: (i) unidentate carbonate formation requires iso-
lated O2− ions, i.e., low-coordination anions, such as those present
at the corners or edges; (ii) bidentate carbonate forms on Lewis
acid-Brønsted base pairs (Mn+–O2− sites with Mn+ or Al3+); (iii)
bicarbonate species form from surface hydroxyl groups [40]. The
order of base strength for these surface oxygen species is: low-
coordination O2− anions > oxygen in Mn+-O2− pairs > OH groups.
Our studies on MgAl spinel oxides suggest that the high basicity
of this sample is probably due to the O2− and oxygen species from
Mn+–O2− pairs, which may by predominant over the MgAl catalyst
at 300 ◦C.
As shown in Fig. 5, the TPD patterns of NiAl and CuAl samples
reveal a broad peak in the temperature range of 100–300 ◦C due
to mild and medium strength basic sites, according to FTIR studies
[41]. The CO2-TPD curve of FeAl suggests that the only high inten-
sity CO2 desorption peak (ca. 200 ◦C) exhibited by this sample is
probably assigned to the medium strength basic sites provided by
the Fen+–O2− pairs from FeAl2O4 or ␣-Fe2O3, as observed by the
findings [1,19].
The TPD of ammonia profiles have that NH3 desorbs in two over-
lapping signals at a low temperature peak and a high temperature
peak and the results are shown in Table 2.
The CuAl sample is abundant in stronger acid sites (T = 315 ◦C,
peak II) and has a moderate amount of weak acid sites (T = 123 ◦C,
peak I) in comparison to the other three catalysts. This is con-
sistent with the amount of NH3 desorbed, which is 0.015 and
0.038 mmol g−1, respectively, for the low and high temperature
peaks of CuAl. The other samples have the acidic sites mostly
centred in the region of high temperatures (T = 287–302 ◦C, peak
II), whereas the low temperature peaks (T = 97–120 ◦C, peak I)
are acidic to a lesser amount. The surface densities of the acid
sites are obtained by deconvolution and integration of NH3
from the samples and suggest the following order of acidity:
CuAl > NiAl ≈ FeAl > MgAl.
The electronegativity of the cationic metals increases in the
following order: Mg < Al < Ni ≈ Fe < Cu, and this suggests that the
coulombic attractive forces of the electrons of Cu are higher than
those of Mg. It implies a smaller polarisability of Cu ions, in other
words, the minor tendency to share or transfer electrons of another
cation. This may constitute one of the reasons for the superior acid-
ity of the CuAl, FeAl and NiAl samples compared to the MgAl solid.
The CO2-TPD experiments are performed to characterise the
surface basic sites of the samples.
When considering the amount of CO2 desorbed per gram of sam-
ple, the basic strength follows the order: MgAl > FeAl ≈ NiAl > CuAl,
which is the opposite order seen for the acidity as was previously
shown. Therefore, CuAl stands out as a solid with a relatively high
acidity while MgAl has a large number of high strength basic sites.
The MgAl sample appears to have a much higher basicity than
do the NiAl, FeAl and CuAl samples. Indeed, the CO2-TPD curve
of MgAl shows a low intensity peak at about 100–200 ◦C that is
related to the desorption of the molecular CO2, while the higher
temperature desorption peak (nearly 300 ◦C) could be ascribed
to the desorbed CO produced by the dissociative adsorption of
CO2 [40,41]. Studies on CO2 adsorption by FTIR of MgyAlOx oxides
show that these types of samples possess surfaces oxygen atoms
of a different chemical nature that bind to CO2 with different
3.5. Catalytic activity in the dehydrogenation of EB
The dehydrogenation of ethylbenzene is carried out with the
nanocasted spinel oxides in the presence of CO2. Table 3 shows the
results of conversion, selectivities, both in terms of mol% and activ-
ities (measured as moles of EB consumed per gram of catalyst per
hour), expressed in mol2 g−1 h−1. Except for NiAl catalyst, carbon
balance in the catalytic tests is above 98%.
Table 2
Catalyst characterisation by acid–base properties.
Sample
NH3-TPD temperature (◦C)
NH3desorbed (mmol g−1
)
Density of the acid sites (mmol m2)
Peak I
Peak II
Peak I Peak II
NiAl
CuAl
FeAl
MgAl
110
123
120
97
302
315
298
287
0.009
0.015
0.012
0.002
0.027
0.038
0.025
0.018
0.37
0.60
0.34
0.16

154
A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
Table 3
Catalytic properties of the solids in the dehydrogenation of EB with CO2.
Sample
% Conversiona
% Selectivities
St
Activity (mmol g−1 h−1
)
Tol
Bz
Eth
Xb
NiAl
CuAl
FeAl
MgAl
12.0
8.5
20.1
10.0
56.0
99.8
98.7
100
40.1
–
–
1.0
0.9
–
1.4
–
0.4
–
0.9
–
–
0.7
0.4
1.2
0.6
–
–
–
Reaction conditions: T = 550 ◦C; CO2/EB = 30; mcat = 0.1 g; 6 h of time on stream ; ethylbenzene flow: 1.98 mmol h−1
.
a
Values obtained from the stead state condition.
b
Polyaromatics compounds.
Me2+–O sites (M = Mg, Al) could promote the dehydrogenation of
EB with CO2 via an acid–base mechanism [5], the smaller disper-
sion of the acidic sites (density = 0.16 mmol m2, Table 2) diminishes
the catalytic performance.
The CuAl sample shows the lowest catalytic performance, with a
conversion of 8.5% and 99.7% selectivity for styrene formation. On
MAl2O4 spinel oxide aluminates, the conversion of ethylbenzene
depends mainly on the transition metal (M). When replacing Cu
for Mg or Ni, the activity steeply increases in the first minutes of
the reaction (Fig. 6a); nevertheless, the selectivity for styrene using
NiAl greatly decreases (Fig. 6b), resulting in the production of 56%
styrene, 40.1% toluene and 1.4% ethane, the latter resulting from
ethylbenzene cracking promoted by Ni species [5,38].
Although the conversion of EB is ca. 12% for NiAl, the selectivity
for styrene is 56%, which is considered too low, regarding the other
solids. By-products of cracking such as toluene, benzene and ethane
are the highest (ca. 43.4% in the total amount) among the catalysts
studied. According to Organowski et al. [12], electrophilic oxygen
atoms participate in reactions such as combustion and cracking. In
the case of the NiAl catalysts, the oxygen adsorbed from CO2 could
react with EB molecules from the gas phase or those at the cata-
lyst surface (e.g., nucleophilic oxygen), according to the Rideal or
Langmuir–Hinsewoold mechanism and combustion and cracking
by-products are formed. This fact could explain the large amount of
by-products formed. Indeed, the SEM and textural property results
of NiAl suggest that larger NiO particles are on the NiAl2O4 phase
in accordance with the findings [8,38]. In such a case, the EB and
oxygen activation might be distant spatially, which hinders direct
reaction between the adsorbed species. Thereby, the nucleophilic
oxygen sites of NiAl2O4 constitutes the active oxygen from M–O
(M = Ni2+ or Al3+) spinel frameworks, and it maintains the activity
after the reduction of NiO particles proceeds. It justifies the slightly
higher conversion of the NiAl catalyst compared to the MgAl cata-
lyst.
The FeAl catalyst is the most active solid (conversion = 20%,
Table 3) due to the combined effects of basicity and reducibility.
It is known that the oxidative dehydrogenation of EB proceeds via
an acid–base mechanism on an Fe-based catalyst [19]. According
to Mars-van-Krevelen, the mechanism of simple EB follows these
steps: first, the adsorption of EB on the Fe3+ acid site, followed by
the abstraction of the ␣-hydrogen from EB by a basic OH group
adjacent to the acid site; second, the formation of an anion vacancy
via desorption of water; third, the filling of the vacant site by O−
species formed from adsorbed CO2; and finally, the abstraction of
the ␤-hydrogen from EB by the O− species to give styrene and a
In light of the above results, it is clear that Cu2+ plays a detri-
mental role in making the CuAl catalyst less active than the other
solids due to the generation of acidic sites (NH3-TPD results) by the
Cu2+ cations. It is well known that CO2 can be activated through
either the basic sites or the redox sites on catalysts [9]. Even if the
redox mechanisms can occur over CuAl and the high acidity of the
solid, CO2 could keep the copper species at high valence and the
conversion could be a result of the simple dehydrogenation of EB.
Accordingly, the TPR experiments confirm the high reducibility of
Cu2+ species and suggest that the mesostructured features of the
CuAl catalyst are not enough to maintain the activity of the solid.
Both NiAl and MgAl catalysts are more active than the CuAl cata-
lyst, indicating that their increased basicity and lower reducibility
(from CO2-TPD and TPR measurements) account for their higher
activity in terms of EB dehydrogenation with CO2. The slight dif-
ferences in the conversion of NiAl and MgAl in the steady state
condition (Table 3) may be ascribed to the smaller acidic density
of the MgAl catalyst. Activity values in the steady state condition
account for the higher stability of FeAl as compared to the other
solids.
The surface area of MgAl is not all free (some of the fragments of
the surface are empty, as seen by SEM micrographs) and therefore
particular parts of the surface differ in their role in the catalytic
process. Thus, the catalyst surface defines the area of the pores
inaccessible to EB molecules and the uncovered surface of MgO
that is present in high amounts as a layer (i.e.,SEM-EDX analy-
sis of the Mg amount) on MgAl2O4. Although the strong O2− or
Fig. 6. Activity as a function of the time on stream for the nanostructured catalysts.

A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
155
Fig. 8. Long-term stability runs for the FeAl catalyst under different atmospheres.
Dotted lines represent the ethylbenzene conversion while straight lines are the
selectivities to styrene.
Fig. 7. Raman spectra of the spent catalysts after testing the dehydrogenation of
ethylbenzene with CO2.
Ni species to some extent. However, the high amount of NiO oxide
(XRD results) is reduced to Ni0 and the latter species produces the
coking through Boudouard and/or styrene dimerization [8,38]. This
result could explain the lower selectivity for styrene for this solid.
Even if CuAl has a high acidity, the carbonaceous deposition over
this solid is unlikely [38] and the reduction of the Cu species is the
main cause of deactivation, as discussed above.
basic OH group [5]. The FeAl catalyst is abundant in strong basic
sites compared to the CuAl and NiAl catalysts. Based on these pos-
tulations, as well as the characterisations, the highest activity of the
FeAl can be explained. Additionally, the activity values suggest the
following order of the favourable acid–base and catalytic properties
under the reactions conditions.
FeAl > NiAl > MgAl ≈ CuAl. In addition, FeAl, MgAl and CuAl are
selective for the production of styrene while NiAl has a preference
to form toluene, as observed in early studies on Ni-based catalysts
[5,8].
Generally speaking, mesoporous FeAl spinel replica oxide has
superior catalytic performance in comparison to the aforemen-
tioned aluminates. Initially, upon using Fe3+ components of FeAl,
it is found that the conversion of EB increases and the selectivity
for styrene formation is observed due to the presence of the highly
active ␣-Fe2O3 phase. Moreover, after 3 h of reaction EB conversion
(Fig. 6a) gradually decreases due to ␣-Fe2O3 reduction, whereas no
substantial decline in the selectivities is observed. The improved
catalytic performance of the FeAl catalyst after 6 h of reaction can
be ascribed to the unique structural and textural characteristics
of this spinel replica, including its high surface area, which could
expose more active sites and interconnecting pore structures. It can
favour the adsorption of EB and CO2 reactants and the desorption
of the products and thus facilitate the dehydrogenation of EB. In
addition, the moderate acid–base features of FeAl (CO2-TPD and
NH3-TPD analyses) combined with the FeAl2O4 structure provide
long-term stabilities with less than a 2% change relative to the EB
conversion after 6 h of time on stream. Further experiments with
FeAl are performed to explain its stability.
Further studies are necessary to justify the better catalytic per-
formance of FeAl.
3.5.1. Catalytic performance along the time on stream
To understand the stability of the nanocasted spinel oxides, fur-
ther experiments were done.
Fig. 6 presents the relationships between the activity and selec-
tivities to styrene. All these dependencies have distinct courses:
activities markedly decrease with the time on stream, except for the
NiAl catalyst; simultaneously, the increase of the time on stream
gives a significant increase in the selectivities for styrene produc-
tion, except for the NiAl catalyst. Initially, the CuAl catalyst shows
relatively low EB activity compared to that of FeAl, NiAl and MgAl.
This can be attributed to the easier reduction of Cu species (TPR
measurements). Indeed, at the first stage of the reaction the higher
number of medium strength basic sites of MgAl and NiAl and their
textural features provide a better catalyst as compared to CuAl.
However, the MgAl and NiAl spinel catalysts deactivate quickly and
activity decreases more than 5% after only 2 h on stream due to coke
formation, according to Raman studies (Fig. 7).
Fig. 7 clearly shows that the production of amorphous car-
bonaceous deposits is considerably high over NiAl and MgAl spent
catalysts in comparison to the FeAl catalyst. This is confirmed by the
ratio of the ID/IG band that is 0.45 for NiAl and 0.80 for MgAl, indicat-
ing a lesser degree of coking over the latter sample. This result is also
well in agreement with the NiAl selectivity results: the selectivity
for styrene drops considerably, clearly indicating the strong depen-
dence on the electronic state of Ni. If the coupled dehydrogenation
of EB and RGWS (reactions (III) and (IV)) reactions were to be oper-
ating by a redox mechanism, then CO2 will impede the reduction of
3.5.2. Catalytic recyclability studies
After burning the deposited coke at 550 ◦C in flowing air, CO2 or
N2 for 2 h, the catalytic performance can be mostly resumed (Fig. 8).
After the pretreatment with air, the conversion of EB starts at a
low value and increases over 4 h. The styrene selectivity is low while
benzene and toluene production increases and styrene selectivity
is about 82%. After 10 h of reaction, both conversion and selectivity
are unchanged; however, the low performance in airflow is due to
the fact that Fe is probably reduced to Fe0 (reactions (III)–(VI), TPR
results) and then in the presence of CO2 from the reaction media
Fe0 is converted into Fe3C carbide by coke deposition during the EB
dehydrogenation. The Fe3C is permanently deactivated and cannot
be converted to Fe3+ or Fe2+ [41] and Fe3O4 phase is less active to
produce styrene.
Fe0 + CO2 → Fe3O4 + Fe3C
(VIII)

156
A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
Accordingly, the amounts of coke deposited on the solid sur-
CuAl2O4 and MgAl2O4 are produced with high levels of textu-
ral properties, but their acid–base characteristics provide a fast
deactivation pathway in the reaction media. Specifically, the com-
bined effects of basicity and the low temperature of reducibility of
MgAl2O4 resulted in its modest performance in the reaction due to
the lower capacity for adsorption of CO2 exhibited by the strongly
basic sites. CuAl2O4 is practically inactive due to the reduction of
Cu2+ species by CO or H2 from the reaction. The FeAl2O4 phases
obtained with CO2, air, or N2 atmospheres showed different activ-
ities, but the CO2 consumption over this solid revealed the high
basicity of the material and provided the equilibrium change to
produce styrene from the coupling RGWS and the simple dehydro-
genation reaction.
face after 10 h of reaction is 17 mg g−1 (from thermogravimetric
analysis) suggesting that coking is the main cause of the catalyst
deactivation. Nevertheless, some activity is maintained due to the
concentration of oxygenated functional groups on the catalyst sur-
face (coke) which can be responsible for the catalytic performance
of the solid, as observed for the oxidative dehydrogenation of ethyl-
benzene [42].
The pretreatment of the spent catalyst with N2 results in a
decrease in the catalytic parameters. Indeed, N2 is not able to burn
the carbonaceous deposits; thus, it is obvious that the nitrogen
pretreatment does not give favourable effects to the catalytic activ-
ity of the FeAl catalyst. The selectivity for styrene is extremely
affected because of the formation of Fe3O4 (magnetite oxide) which
is not selective for styrene. The literature reports that the Fe3+ and
Fe2+ species in oxygen-deficient magnetite are not distinguished
because the electrons could hop between Fe3+ to Fe2+ in the spinel
[41]. Thus, the oxygen-deficient magnetite is a good material to
convert CO2 into carbon and this could explain the decrease in EB
conversion and styrene selectivity.
However, the CO2 pretreatment generates an increase in con-
version and selectivity that reaches a steady state value after 6 h of
reaction. In this case, the CO2 pretreatment promotes the occu-
pation of the oxygen deficiency in the neighbourhood of both
Fe3+/Fe2+ in oxygen-deficient magnetite oxide after the CO2 adsorp-
tion. Thus, the electrons of the various phases of iron possessing
both Fe3+ and Fe2+ species on the adsorption sites are donated
easily to the carbon in the CO2, which is then reduced (reactions
(XIX)–(XI)):
Acknowledgments
A.H. M.B. is grateful to the Pibic/Cnpq scholarship. We are also
acknowledged to JM Sasaki for XRD analysis.
References
[1] A. Sun, Z. Qin, S. Chen, J. Wang, J. Mol. Catal. A: Chem. 210 (2004) 189–195.
[2] A.C. Oliveira, J.L.G. Fierro, A. Valentini, P.S.S. Nobre, M.C. Rangel, Catal. Today
85 (2003) 49–57.
[3] B.S. Liu, G. Rui, R.Z. Chang, Appl. Catal. A: Gen. 335 (2008) 88–94.
[4] H. Li, Y. Yue, C. Miao, Z. Xie, W. Hua, Z. Gao, Catal. Commun. 8 (2007) 1317–1322.
[5] A.H.M. Batista, F.F. de Sousa, S.B. Honorato, A.P. Ayala, J.M. Filho, F.W. de Sousa,
A.N. Pinheiro, J.C.S. de Araujo, R.F. Nascimento, A. Valentini, A.C. Oliveira, J. Mol.
Catal. A: Chem. 315 (2010) 86–98.
[6] X.-H. Li, X.W. Ying, C. Ke, Catal. Lett. 105 (2005) 223–227.
[7] D.R. Burri, K.-M. Choi, D.-S.H. Sujandi, N. Jiang, A. Burri, S.-E. Park, Catal. Today
131 (2008) 173–178.
[8] R.M. Freire, A.L. Pinheiro, A.C. Oliveira, F.F. de Sousa, A.P. Ayala, J.M. Filho, P.T.C.
Freire, A.C. Oliveira, Appl. Catal. A: Gen. 359 (2009) 165–179.
[9] A. Sun, Z. Qin, J. Wang, Appl. Catal. A: Gen. 234 (2002) 179–189.
[10] Y. Ohishi, T. Kawabata, T. Shishido, K. Takaki, Q. Zhang, Y. Wang, K. Nomura, K.
Takehira, Appl. Catal. A: Gen. 288 (2005) 220–231.
FeAl2O4 + CO2 → FeAl2O4 + CO
Fe1−xO + CO2 → Fe3O4 + C
Fe3O4 + CO2 → Fe3O4 + C
(XIX)
(X)
(XI)
[11] F. Kus´trowski, A.R. Laschoa, D. Majda, D. Tomaszewska, R. Dziembaj, Sold. State
The oxygen in CO2 in the form of O2− is transferred into lattice
vacancy (oxygen deficiency) and CO2 is decomposed to carbon and
CO according to the reaction:
Ionics 141 (2001) 237–242.
[12] W. Oganowski, J. Hanuza, L. Kepiski, Appl. Catal. A: Gen. 171 (1998) 145–154.
[13] I. Serafin, A. Kotarba, M. Grzywa, Z. Sojka, H. Bin´ czycka, J. Catal. 239 (2006)
137–144.
[14] C.O. Areán, M.P. Mentrut, A.J.L. Lopez, J.B. Parra, Colloids Surf. A: Phys. Eng.
Aspects 180 (2001) 253–258.
[15] X. Ye, Ning Ma, W. Hua, Y. Yue, C. Miao, Z. Xie, Z. Gao, J. Mol. Catal. A: Chem.
217 (2004) 103–108.
[16] Z. Li, B.H. Shanks, Appl. Catal. A: Gen. 354 (2009) 50–56.
[17] N.J. Jebarathinam, M. Eswaramoorthy, V. Krishnasamy, Appl. Catal. A: Gen. 145
(1996) 57–74.
CO2 → CO + O2−
(XII)
Indeed, CO could decompose to coke (reaction (XII)) but the pos-
sibility is unlikely provided that the FeAl catalyst treated in the CO2
flow is active and stable towards styrene production. This indicates
that the redox coking on Fe◦ sites is suppressed by CO2 oxidation.
[18] A. Miyakoshi, A. Ueno, M. Ichikawa, Appl. Catal. A: Gen. 216 (2001) 137–146.
[19] J.C.S. de Araujo, C.B.A. Sousa, A.C. Oliveira, F.N.A. Freire, A.P. Ayala, A.C. Oliveira,
Appl. Catal. A: Gen. 377 (2010) 55–63.
CO → C + O2−
(XIII)
[20] O. Clause, M. Gazzano, F. Trifiroꢀ, A. Vaccari, L. Zatorski, Appl. Catal. (1991)
217–236.
Even though it is a mild oxidant, CO2 is able to maintain
the Fe3+ species and/or to form FeAl2O4 from magnetite in situ
during the reaction. In addition, coking is slightly removed by
the CO2 in the reaction (i.e., carbon content is about 3 wt%, TGA
analysis). Additionally, the textural properties of FeAl in CO2 are
[21] T.P. Braga, E. Longhinotti, A.N. Pinheiro, A. Valentini, Appl. Catal. A: Gen. 362
(2009) 139–146.
[22] J.S. Valente, J.H. Cortez, M.S. Cantu, G. Ferrat, E. L-Salinas, Catal. Today 150
(2010) 340–345.
[23] J. Zhu, Q. Gao, Micropor. Mesopor. Mater. 124 (2009) 144–152.
[24] A.-H. Lu, H. Tüysüz, F. Schüth, Micropor. Mesopor. Mater. 111 (2009) 117–123.
[25] T. Valdés-Solís, A.B. Fuertes, Mater. Res. Bull. 41 (2006) 2187–2197.
[26] S. Lepoutre, J.-H. Smått, C. Laberty, H. Amenitsch, D. Grosso, M. Lindén, Micro-
por. Mesopor. Mater. 123 (2009) 185–192.
slightly lower compared to those of the initial solid (Sg = 102m2 g−1
Vp = 0.39 cm3 g−1), whereas that of FeAl in oxygen is 87 m2 g−1
For FeAl in a N2 flow, the textural parameters are the lowest
(55 m2 g−1), showing the decline in the spinel properties after the
pretreatment to which the latter sample is submitted.
;
.
[27] F. Schüth, Chem. Mater. 13 (2001) 3184.
[28] Y. Sharma, N. Sharma, G.V. Subba Rao, B.V.R. Choudari, Adv. Funct. Mater. 17
(2007) 285.
Therefore, the use of CO2 in the pretreatment as well as con-
tinuously throughout the reaction brings benefits to the reaction
by producing a stable catalyst as compared to that produced when
using air or N2 in the pretreatment.
[29] G.Y. Zhang, B. Guo, J. Chem. Sensors Actuators B 114 (2006) 402.
[30] F. Liao, P.G. Bruce, Adv. Mater. 19 (2007) 657.
[31] M. Imperor-Clerc, D. Bazin, M.–D. Appay, P. Beaunier, A. Davidson, Chem. Mater.
16 (2006) 1813.
[32] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucki,
Science 279 (1998) 548.
[33] B.S. Liu, R.Z. Chang, L. Jiang, W. Liu, C.T. Au, J. Phys. Chem. C 112 (2008)
15490–15501.
4. Conclusions
[34] L.E. Smart, E.A. Moore, Solid State Chemistry—An Introduction, third ed., CRC
press, New York, 2005.
[35] R.D. Shanon, Acta Crystalogr. A 32 (1976) 751.
[36] J.-H. Smått, N. Schüwer, M. Järn, W. Lindner, M. Lindén, Micropor. Mesopor.
Mater. 112 (2008) 308–318.
This work demonstrates that nanostructured FeAl2O4 is effec-
tive at catalysing the dehydrogenation of ethylbenzene with CO2
to produce styrene. Other spinel phases such as FeAl2O4, NiAl2O4,

A.H.d.M. Batista et al. / Applied Catalysis A: General 382 (2010) 148–157
157
[37] J.S. Valente, J.H. Cortez, M.S. Cantu, R. Montiel, X. Bokhim, E. L-Salinas, J. Phys.
Chem. C 111 (2007) p.642.
[40] V.K. Díez, J.I. Di Cosimo, C.R. Apesteguía, Appl. Catal. A: Gen. 345 (2008)
143–151.
[38] A.L. Pinheiro, A.N. Pinheiro, A. Valentini, J.M. Filho, F.F. de Sousa, J.R. de Sousa,
M.G.C. Rocha, P. Bargiela, A.C. Oliveira, Catal. Commun. 11 (2009) 11–14.
[39] J. Wang, L.-F. Chen, C.L. Li, J. Mol. Catal. A: Chem. 139 (1999) 315–323.
[41] C.I. Zhang, Mater. Chem. Phys. 62 (2000) 52–61.
[42] J.J. Delgado, X. Chen, J.P. Tessonnier, M.E. Schuster, E. Del Rio, R. Schlögl, D.S. Su,
Catal. Today 150 (2010), pp. 49-45.