Characterisation of high surface area nanocomposites for glycerol transformation: Effect of the presence of silica on the structure and catalytic activity

Article abstract of DOI:10.1016/j.cattod.2012.09.013

The catalytic properties of the nanocomposites were evaluated at various temperatures during the gas-phase dehydration of glycerol, an environmental friendly reaction. A pressure of 1 atm and a glycerol/water molar ratio of 0.25 were applied. Nanocomposites possessing high surface area were synthesised, and silica-containing CeO₂-MnO₂, NiO-Al₂O ₃ and CeO₂-Al₂O₃ nanoparticles were obtained. The materials were characterised by XRD, Raman spectroscopy, TPR, TEM and SEM-EDX analysis, and the textural properties of the materials were determined. The effect of the structure, surface properties and textural parameters was evident in the unique properties of the nanoparticles, and the accessibility of the active sites dictated the catalytic activity. The basicity and reducibility of the nanoparticles on silica plays a critical role in the catalytic properties of the oxides.

Full text of DOI:10.1016/j.cattod.2012.09.013

Catalysis Today 212 (2013) 127–136
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Characterisation of high surface area nanocomposites for glycerol
transformation: Effect of the presence of silica on the structure and
catalytic activity
Francisco de Assis A. Barrosa,1, Helvio S.A. de Sousaa,1, Alcemira C. Oliveirab,2, Manoel C. Juniorb,2
,
Josué M. Filhob,2, Bartolomeu C. Vianac,3, Alcineia C. Oliveiraa,∗
a Universidade Federal do Ceará, Campus do Pici-Bloco 940, Departamento de Química Analítica e Físico-Química, Langmuir Lab de Adsorc¸ ão e Catalise, Fortaleza, Ceará, Brazil
b Universidade Federal do Ceará, Campus do Pici-Bloco 922, Departamento de Física, Fortaleza, Ceará, Brazil
c Universidade Federal do Piaui, Departamento de Física, Teresina, Piauí, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
The catalytic properties of the nanocomposites were evaluated at various temperatures during the
gas-phase dehydration of glycerol, an environmental friendly reaction. A pressure of 1 atm and a
glycerol/water molar ratio of 0.25 were applied. Nanocomposites possessing high surface area were syn-
thesised, and silica-containing CeO2–MnO2, NiO–Al2O3 and CeO2–Al2O3 nanoparticles were obtained.
The materials were characterised by XRD, Raman spectroscopy, TPR, TEM and SEM-EDX analysis, and the
textural properties of the materials were determined. The effect of the structure, surface properties and
textural parameters was evident in the unique properties of the nanoparticles, and the accessibility of
the active sites dictated the catalytic activity. The basicity and reducibility of the nanoparticles on silica
plays a critical role in the catalytic properties of the oxides.
Received 16 February 2012
Received in revised form 7 August 2012
Accepted 2 September 2012
Available online 3 November 2012
Keywords:
Characterisations
Nanocomposites
Structural properties
Catalytic activity
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
According to the abovementioned reaction, glycerol is firstly
protonated at the secondary hydroxyl group by a catalyst, and
The catalytic conversion of renewable resources for the produc-
tion of fuels and chemicals has attracted intense attention in recent
years to limit global warming and climate change [1–3]. Glycerol
dehydration is a reliable process for the conversion of renewable
biomass into value-added chemicals or intermediates such as 1-
hydroxyacetone and acrolein and [3–5], according to the reactions
[6]:
the protonated glycerol is dehydrated to form 3-hydroxypropanal
(3-HPA). The latter intermediate is a reactive hydroxyl carbonyl
compound, which can be easily dehydrated to acrolein [1–8]. The
glycerol protonation occurs at a terminal hydroxyl group and then
1-hydroxyacetone is obtained by dehydration and deprotonation of
its protonated intermediate [6,9]. Typical uses of acrolein were for
production of acrylic acid ester, super absorber polymers, methio-
nine synthesis and detergents whereas 1-hydroxyacetone is largely
used as a starting material for production of polyols [6].
The conversion of glycerol can be conducted in the aqueous or
gas phase; however, the latter is suitable for the production of
acrolein due to the tunable catalytic activity and selectivity. Var-
ious homogeneous and heterogeneous acid catalysts have been
investigated in the dehydration of glycerol, since the homogeneous
catalysts are more disadvantageous and they cause corrosiveness
and safety problems. Thereby, unsupported solid catalysts, such as
oxides, zeolites and molecular sieves as well as supported solids,
e.g., heteropolyacids dispersed on oxides, metal sulphides or phos-
phates, and metal supported on active carbon, have been studied
for the conversion of glycerol in the gas phase [4–16].
∗
Corresponding author. Tel.: +55 85 3366 99 82; fax: +55 85 3366 99 82.
E-mail address: alcineia@ufc.br (A.C. Oliveira).
Tel.: +55 85 3366 99 82; fax: +55 85 3366 99 82.
Tel.: +55 85 3366 90 08; fax: +55 85 3366 90 08.
Tel.: +55 86 3215 58 33; fax: +55 86 3215 58 33.
Although the aforesaid catalysts had sufficiently high dehydra-
tion activity, the rapid catalyst deactivation due to coke deposition,
phase transformations as well as the formation of large amounts
1
2
3
0920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2012.09.013

128
F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
of undesired by-products were the major drawbacks. Indeed, the
literature reports important production of other by-products such
as propanaldehyde, acetaldehyde, acetone, glycerol polyconden-
sation products, cyclic glycerol ethers, adducts of acrolein with
glycerol, among others that contribute to coking [17]. Therefore,
there have been several attempts to avoid these problems, namely
using basic solids, co-injection of oxygen into the feed, mild reac-
tion conditions or introducing a moving-bed reactor [6,7,12,13,16].
Among the aforementioned materials, binary metal oxides have
been used successfully, and high conversion and acrolein selectiv-
ity have been obtained [14–20]. Nevertheless, the low surface area
as well as the low stability of these solids limited their catalytic per-
formance in the conversion of glycerol, as glycerol is considered to
be a strong reductant, and mostly due to carbonaceous deposition
[17,18].
mesoporous silica. The amount of the remaining mesoporous silica
was about 20 wt%, as determined by chemical analysis. To finish,
the solid was dried and calcined under air atmosphere to 600 ◦C
at a constant rate of 2 ◦C min−1 and the solid was kept at that
temperature for 5 h to obtain the CeMn sample. CeAl and NiAl
nanocomposites were obtained by the abovementioned procedure,
by using Ni(NO3)2·6H2O, Al(NO3)3·9H2O and (NH4)2Ce(NO3)6 pre-
cursors.
For catalytic performance comparisons purposes, nanoparti-
cles of CeMn, CeAl and NiAl were also produced by using a
mesoporous carbon, according to a method previously published
[14].
2.2. Characterisations
Therefore, binary oxides composites could be suitable solids to
produce catalysts that are more active and less susceptible to phase
degradation and carbon deposition than conventional binary oxide
catalysts. Indeed, binary oxides composites can provide a large
surface area for the formation and stabilisation of small oxide crys-
tallites and can improve the accessibility of binary oxides catalysts
[21,22]. Additionally, CeO2 is of particular interest due to its abil-
ity in preventing the loss of surface area of bare alumina, silica and
other supports and thus, avoiding sintering during various catalytic
reactions such as glycerol steam reforming [16,23–27]. Moreover,
the presence of Mn or Al in Mn4+/Mn3+–Ce–O or Mn–Al–O mixed
oxide catalysts can enhance the catalytic activity by redox and
acid–base properties compared to their monoxides counterparts
[28,29]. Thus, the use of Ce-containing Mn or Al as catalysts for
glycerol transformation could provide an alternative to balance
the acidity in the solids and also enhance their activities by CeO2
redox properties. Ni–Al compounds have been successively used to
improve the catalytic performance of the solids by stabilisation of
Ni in NiAl2O4 spinel phase effects [30,31].
Powder X-ray diffraction was carried out on a X-ray powder
diffractometer Xpert MPD (PANalytical) using Cu K␣ radiation
(40 kV, 30 mA). The measurements were taken at low and wide-
angles. Average diameters of crystallites were not estimated from
XRD patterns applying the Scherrer’s equation due to the broaden-
ing of the XRD lines. The phases were identified according to Joint
Committee on Powder Diffraction Standards (JCPDS) and Interna-
tional Centre of Diffraction Data (ICDD) files.
The specific surface areas and pore structure parameters of
the solids were determined by the adsorption isotherms of nitro-
gen at −196 ◦C utilizing a ASAP 2000 Micromeritics surface area
analyser. Samples were degassed at 150 ◦C for 24 h prior to mea-
surement, and these conditions of degassing were enough to desorb
all adsorbed molecules. The BET equation was used to calculate the
total specific surface areas. The pore volumes were calculated from
the adsorption branch of the isotherm at P/Po = 0.99.
Inductively Coupled Plasma Optic Emission Spectroscopy (ICP-
OES) was performed in a Varian instrument for determination of
the chemical analysis of the calcined catalysts. Previously, the solids
were digested with a mixture of hydrofluoric, nitric and chloridric
acids at 90 ◦C in a sand bath. Me1 and Me2 represent the first and
the second element in the sequential order of the samples nomen-
clature: CeAl, NiAl and CeMn.
High-resolution transmission electron microscopy (TEM) was
performed on a FEI TECNAI G2S-Twin transmission electron micro-
scope equipment operating at 200 kV.
Therefore, in the present study, we focused on the synthesis,
characterisation and catalytic properties of high surface area silica-
containing CeO2–MnO2, NiO–Al2O3 and CeO2–Al2O3 nanoparticles
for the dehydration of glycerol. The effect of the presence of sil-
ica on the stability of the nanoparticles was also determined and
compared with that of conventional oxides.
Scanning electron microscopy (SEM) measurements were con-
ducted on a TESCAN VEGA XMU electron microscope equipped with
an EDX Bruker QUANTAX system coupled to the SEM microscope,
using an acceleration voltage of 20 kV.
2. Experimental
2.1. Preparation of the solids
Temperature programmed reduction (TPR) analysis of the cat-
alysts was performed in a home-made equipment using a quartz
tube reactor coupled to a TCD for monitoring the hydrogen con-
sumption. The mass of catalyst was 50 mg and the experiment
was conducted in the range of 50–1000 ◦C. A 8% H2/N2 mixture,
as reducing gas was used with a rate of 100 mL min−1, after pass-
ing 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 were carried out in a homemade equipment. Prior
to the CO2 adsorption, all catalysts were preheated in He flow at
300 ◦C for 2 h and then exposed to 5% of CO2 in N2 flow. Finally,
CO2 desorption was performed from 50 to 700 ◦C.
Raman spectra of the catalysts were obtained on a T64000
Raman spectrometer (JobinYvon triple spectrometer) under ambi-
ent conditions. A 514.5 nm Ar laser was used as the exciting source
on the sample surface using a power of about 1 mW. The measure-
ments were referenced to Si at 521 cm−1 with 16 data acquisitions
in 120 s. The lens focus was of 100 times.
The nanocomposites were prepared by a nanocasting route
using the mesoporous SBA-15 as a sacrificial template [31]. Briefly,
an equimolar mixture of cerium ((NH4)2Ce(NO3)6) and manganese
(Mn(NO3)2·6H2O) methanolic solutions were added drop-wise
through a peristaltic pump to a beaker containing acetic acid that
was stirred for 2 h. Then, about 1 g of SBA-15 silica was added to
the previous solution and mixed for 5 min, according to the method
described in the literature, with modifications [23]. Subsequently,
propylene oxide (C3H6O) was added to the suspension (molar ratio
C3H6O/Mn(NO3)2·6H2O = 11:1) and stirred for 20 min. The solid
was separated by filtration, and the SBA-15 containing hydrated
cerium and manganese salts, acetic acid, and propylene oxide, was
left for 16 h under steam of methanol. After this period was passed,
the solid was washed with methanol and separated by filtration.
The solid was washed, dried and calcined in air flow at 500 ◦C for
2 h to obtain the SBA-15-supported CeMn oxide.
Silica-containing cerium–manganese oxide composite was pre-
pared from SBA-15-supported CeMn oxide material. The latter solid
was washed several times with water and exposed to 1 mol L−1
NaOH solution under stirring for 1 h to remove partially the

F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
129
2.3. Catalytic tests
that they are not pure Mn2O3 and MnO2 anymore and are possi-
bly doped by cerium, as in MnCeOx phase. The ionic radius of Mn3+
and Mn4+ (0.066 and 0.0053 nm, respectively) is smaller than that of
Ce4+ (0.098 nm), and the incorporation of Mnn+ into the fluorite lat-
The gas-phase dehydration of glycerol was used to evaluate the
catalytic properties of the solids at 250 ◦C and P = 1 atm. A 100 mg of
catalyst was placed into a stainless steel tubular reactor possessing
quartz wool to fix the bed. The system was then preheated at 300 ◦C
in an oven for 1 h in a nitrogen flow of 20 mL min−1. After this period
was passed, the reactant mixture consisting of a 20% by weight of
an aqueous solution of glycerol was fed through a peristaltic pump
to the reactor. Then, the reaction temperature was decreased to
250 ◦C. Temperature of the reaction was varied from 150 to 350 ◦C.
It is important to note that at temperatures as low as 290 ◦C, the
catalytic line was heated to assure the gas-phase dehydration of
glycerol and avoid glycerol condensation, since the boiling point of
glycerol is close to the abovementioned temperature. The products
were collected in a cold ice trap and were subsequently dissolved
in an ethanol/methanol internal standard. The products of the reac-
tion were analysed using a gas chromatograph (GC) connected to a
flame ionisation detector (FID) detector and GC 300 integrator. In
addition, the products of reaction were identified by GC–MS. Con-
versions of glycerol as well as selectivities were obtained according
to the previous calculations [11,14,16,18].
˚
tice reduced the lattice parameters from 5.38 to 5.04 A. This means
that Mn ions can be incorporated in fluorite structure; however,
the MnCeOx phase was not observed by XRD. Moreover, system-
atic peak width broadening was indicative of a nanocrystalline size
and suggested a mixed Ce–Mn lattice.
The XRD patterns of CeAl suggested that crystalline Al2O3 oxide
was formed, however, due to the amorphous characteristics of the
gamma-alumina phase, which is obtained at temperatures greater
than 550 ◦C, only a CeO2 phase (ICDD 00-034-0394) was detected.
Indeed, the typical broad diffusion peak at 2ꢀ ranging from 21 to
28◦ due to the presence of amorphous silica was superimposed by
several narrow peaks that were attributed to the fluorite unit cell
of CeO2. The stabilisation of Al2O3 and CeO2 may have occurred
due to the dispersion of the aforesaid oxides nanoparticles in silica.
This is consistent with the broad features of the gamma-alumina
diffractogram and was further confirmed by SEM–EDX analysis.
According to the findings, temperatures above 600 ◦C favour ther-
modynamically the nucleation of CeAlO3 by the easy diffusion of
Al3+ ions into the partially CeO2 lattice [37]. Moreover, the forma-
tion of CeAlOx phase cannot be excluded, as observed previously
[25,37]. However, the CeAlOx phase detection was not possible by
XRD technique.
3. Results and discussion
3.1. Physicochemical properties of the solids
The XRD patterns of NiAl displayed peaks of NiO cubic struc-
ture (ICDD 4-0835) and ␥-Al2O3 (ICDD 50-741). Cubic Fm3m NiO
has diffraction angles at 2ꢀ equal to 43.4, 62.7 and 75.5◦ [27,30]
whereas the main peaks of crystallised ␥-Al2O3 phase are observed
at about 36.9, 45.2 and 60.5◦ [27]. Besides, the aforesaid phases,
XRD pattern of NiAl revealed the appearance of the reflections typ-
ical of cubic NiAl2O4 phase, which are attributable to the inverse
spinel (2ꢀ = 31.4, 37.4, 44.8, 59.6 and 65.5◦). The latter was observed
because the ionic radius of Al3+ (0.068 nm) was similar to that of
Ni2+ (0.067 nm); thus, aluminium occupied a tetrahedral position
in an almost inverse spinel, and Ni2+ ions were preferentially dis-
tributed over the octahedral cation sites (A-site) in an AB2O4 binary
spinel [30]. Fewer diffraction peaks were observed in the XRD pat-
tern of NiAl; however, NiO and NiAl2O4 peaks were detected at
10.1◦ and 29.0◦, and were attributed to (1 1 1) and (2 0 0) reflections.
A slight broadening of the peaks indicated that small particles exist
and the wide-angle region of the XRD pattern revealed that NiAl
was crystalline. The particle size of CeAl, CeMn and NiAl could be
not estimated by XRD due to the broad features of the diffraction
lines.
The CeMn, CeAl and NiAl molar ratios were 0.48, 0.42 and 0.45,
respectively, as determined by chemical analyses.
X-ray diffractograms of the solids are shown in Fig. 1.
The XRD patterns of CeAl, CeMn, and NiAl displayed broad peaks
that were characteristic of nanostructured oxides [18,32]. The
characteristic diffraction lines of CeO2 fluorite [33,34] (2ꢀ = 28.5,
33.4, 36.4, 47.6 and 56.5) were observed in CeAl and CeMn. The
XRD pattern of pure manganese oxide (Mn2O3) displays the main
diffraction peaks at 2ꢀ equal to 23.0, 32.7, 38.2, 45.7, 49.3 and 55.3◦
[35] while MnO2 exhibits a variety of phases, since it is difficult to
identify the main tetragonal peaks [36]. Due to the broadness of the
diffraction peaks, it was difficult to determine whether the CeMn
peaks were due to Mn2O3 cubic bixbyte with Pcab space group
(ICDD 6-540) or tetragonal MnO2 (ICDD 44-0141) phases [28,33].
Also, it was found by closely comparing the diffraction pattern and
the ICCD files that the diffraction peaks from the Mn2O3 and MnO2
phases were slightly shifted from the ICCD files, which suggests
The aim of using the synthesis strategy was to produce
nanocomposites of oxides with interaction with the silica for glyc-
erol conversion, since the nanoparticles themselves are not stable
for this purpose. Thereby, in accordance with the SEM–EDX mea-
surements (not shown) it was revealed that a significant amount
of silicon form SBA-15 mold, e.g., 20% remained in NiAl and CeMn
while a lesser amount of silica (e.g., 13%) was observed over CeAl.
Moreover, SEM images showed that NiAl, CeAl and CeMn possessed
the same morphology of the silica template. The nanoparticles of
the elements were agglomerated in some regions of the materials,
and this is in agreement with the broad features of the XRD patterns
that were indicative of the presence of nanoparticles. Moreover,
the EDX pattern of CeMn, NiAl and CeAl indicated that the concen-
tration of these elements in all of the regions was not uniform;
thus, these metals were not evenly distributed throughout the
SiO2 phase. It is important to note that independently of the sil-
icon species, there is no evidence whether the SiO2 species are
crystalline.
The TEM image of CeMn suggested that the nanoparticles were
distributed over the surface of the silica (Fig. 2).
Fig. 1. XRD patterns of the solids.

130
F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
Fig. 2. TEM images of the solids: CeMn, CeAl, and NiAl.
Due to electronic density contrasts, the nanoparticles appeared
obtained for NiAl (2–17 nm). Even after partially dissolving the sil-
ica template and calcinating in air, nanoparticles of CeO2, Mn2O3
and MnO2 were obtained, in accordance with XRD and SEM–EDX
analysis results. In case of CeAl, nanoparticles with an average par-
ticle size of 30 nm were observed by TEM; the micrograph of CeAl
also revealed that the particles of attached metals were seen after
partially removal of silica, as indicated further by the textural prop-
erties.
in black while silica walls were in grey. Crystal nanoparticles were
agglomerated, however, in some regions, in accordance with the
SEM images (not shown). Planes with d spacing of 0.28 nm related
to CeO2 were observed whereas that with 0.31 and 0.23 nm were
associated with MnO2 and Mn2O3, respectively. This result is in
accordance with the literature data [38]. The crystallites of CeMn
presented a particle size of 1–15 nm, which is close to the values

F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
131
Fig. 2 shows also high-resolution images of the edge of CeAl and
NiAl nanoparticles. The result indicated that CeAl nanocomposite
consisted of short and disordered planes. Some nanoparticles of
CeO2 and Al2O3 phases with d spacings of 0.28 (JCPDS 43-1002) and
0.25 nm (JCPDS 10-0425), respectively were observed. The disor-
dered layers on the surface can create many defects, which resulted
in the formation of active sites and nanoparticle aggregation [2].
Indeed, NiAl exhibited disordered interconnections between adja-
cent planes due to the patterning of NiO, NiAl2O4 and Al2O3 oxides
in the disordered intra-wall defects of the silica, as observed in
other oxides [22,39]. Three sets of lattice stripes can be well dis-
tinguished and planes of 0.21, 0.64 and 0.34 nm were attributed
to low concentrations of NiO (JCPDS 71-1179), NiAl2O4 (JCPDS
71-0965) and Al2O3, respectively; after conversion by Fourier trans-
form, the diffraction spots confirmed the unit cell parameters of
these phases. The sizes of the nanoparticles varied from 13.6 nm.
Some regions had carbon species that were not well-defined as
shaped finger but contrasted with some carbon nanotubes. They
were formed as a result of the carbon source used and the syn-
thesis conditions, probably. The EDX analyses revealed that the
bulk grain of the solid consisted of silicon frameworks contain-
ing not well-dispersed Al and Ni nanoparticles in the silica matrix,
which suggested that the preparation method favoured the for-
mation of nanoparticles over crystalline NiAl. In addition, empty
mesopores as well as nanoparticles (i.e., the black dots) trapped
within the pores of the silica were also visible in NiAl micrograph.
The latter observation was not surprising as the amount of Ni and
Al introduced in this sample did only allow partial filling of the
mesoporosity.
Although the solids were produced according to the same pro-
cedure, their physicochemical and structural properties were quite
different. In particular, the chemical and crystalline properties
of the materials were diverse, and their growth and migration
abilities in either mesopores of the silica or silica surface were
unique because the metal oxides interacted differently with the
silica walls, which strongly influenced the migrating capacity. For
instance, the crystallisation temperature of CeMn led to relatively
weak interactions between the oxides and the silica walls; there-
fore, the oxide tended to aggregate and crystallises to form a
CeO2 and MnOx framework, as for nanostructured oxides [40].
Although CeAl had a weak interaction with the silica walls, the
high crystallisation temperature to form oxides with high crys-
tallinity is required; consequently, CeAlOx crystalline framework
is difficult to obtain by using our synthesis condition and separated
CeO2 and Al2O3 phases were obtained, as evidenced by the XRD
results.
The size difference among the nanoparticles of the CeMn, CeAl
and NiAl nanocomposites was attributed to the nature of the pre-
cursor. In silica-based nanocomposites, the metallic salt precursor
is homogeneously distributed in the pores of the silica template,
and subsequent calcination leads to the migration to nucleation
seeds [33]. Because the porosity of the silica template is formed
by non-connected pores, the formation of nanoparticles is highly
likely.
Fig. 3. Laser Raman spectra of the nanocomposites.
3.2. Structural determination by Raman spectroscopy
Laser Raman spectroscopy is complementary to XRD and allows
additional information about framework vibrations of the struc-
ture and phases of the oxides in more details. Raman spectra of the
samples are shown in Fig. 3.
and form oxygen vacancies in the CeO2 lattice [41]. This was con-
firmed by the Raman mode at 646 cm−1,which is related to the
oxygen vacancies in CeO2 lattice attributed to the presence of a
defective structure [42]. This mode appeared in the same region
observed for Mn2O3 peak at 642 cm−1 and may be attributed to the
out-of-plane bending modes of MnOx [43]. In addition, the pure
MnO2 exhibits peaks at 648 cm−1 and the mode was attributed to
the stretching mode of the MnO6 octahedrafrom MnO2, as observed
in manganese bulk oxides [42–44]. Moreover, the intensity of F2g
Raman spectrum of CeMn displayed a broad peak at 460 cm−1
,
which was attributed to the single Raman active mode of F2g
symmetry of the perfect CeO2-based cubic fluorite lattice [25].
However, the shift of F2g mode by 5 cm−1 indicates an interaction
between Mn and Ce that could deform the fluorite-like structure

132
F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
Table 1
band is strong probably as the result of the structure of a MnCeOx
solid solution.
Physicochemical properties of the solids.
The small shoulder at 380 cm−1 was also assigned to the partial
breaking of the Raman-active oscillation mode of the CeO2 fluo-
rite structure [44]. In general, the Raman spectroscopy results of
CeMn supported the XRD observations. According to the findings,
the substitution of a part of Ce by Mn ions would cause either
the formation of oxygen vacancies straining the CeO2 structure or
binary species [39,44]. The Raman peaks at about 251 cm−1 and
583 cm−1 could confirm these assignments, however, they could
be not observed due to the broadening of the peaks. In addition,
the peak at 583 cm−1 may overlap with the peak due to the Mn
species. These modes are attributed to be from the defect space
including an O2− vacancy [28,34,43].
a
b
Sample
Sg (m2 g−1
)
Vp (cm3 g−1
)
Dp (nm)
CeAl
CeMn
NiAl
391
183
400
1.17
0.55
1.20
1.6
2.1
2.4
a
BET method.
Adsorption branch at P/Po = 0.99.
b
the vacancies or easier oxygen mobility from bulk to the solid
surface.
3.3. Textural properties and surface features of the solids
In the Raman spectrum of NiAl, peaks at about 89 and 305 cm−1
were attributed to NiO whereas broad bands centred at 250 and
518 cm−1 were attributed to NiAl2O4 vibrations [45]. Al3+ ions can
be incorporated in an amorphous nickel aluminate phase or a sep-
arate amorphous alumina phase, according to previous findings
[30,45]. These results are in line with those of XRD that confirmed
the existence of the aforesaid phases.
The textural properties of the solids are listed in Fig. 4 and
Table 1.
N2 adsorption–desorption isotherms (Fig. 4a) show typical type-
IV curves. Specifically, the NiAl sample possessed an H1-type
hysteresis loop, suggesting regularly cylindrical mesopores while
both CeAl and CeMn showed an H2-type hysteresis loop. The
shape of the adsorption isotherm for NiAl suggests the devel-
opment of microporosity, and a pronounced hysteresis in the
adsorption–desorption isotherms at relative pressure of 0.8 indi-
cated that the sample also had mesoporosity. For CeMn and CeAl,
the hysteresis close to P/Po = 0.40 was due to the tensile strength
effect [21,22]. A unimodal porosity was obtained for all the sam-
ples (Fig. 4b), with a narrow pore size distribution ranging from 1
to 2.5 nm.
Mesoporous nanocomposites with variable surface areas
(183–400 m2 g−1), pore volumes (0.55–1.20 cm3 g−1) and pore
sizes (1.6–2.4 nm) were obtained, according to Table 1. The text-
ural properties of the CeAl and NiAl were slight lower than those of
pure silica (BET surface area and total pore volume of 650 m2 g−1
and 1.25 cm3 g−1, respectively). However, compared to the par-
ent mesoporous silica, the surface area (183 m2 g−1) and pore
volume (0.55 cm3 g−1) of CeMn were relatively low due to the
agglomeration of the particles on solid surface. Specifically, because
mesoporous silica contains micropores, the infiltration of metal
precursors into this type of pore is limited, and continuous connec-
tions on the pore walls in the resultant oxide are difficult to form
[21,40]. The latter is inevitably deposited on the external surface of
silica and contributes to a reduction in the surface area and pore
volume. Another factor that affects the pore volume and surface
area is the partial removal of the silica from the surface of CeMn, as
In the spectra of CeAl, the CeO2 peak was observed at 462 cm−1
and this Raman mode was attributed to the triply degenerate
F2g mode of fluorite CeO2, as discussed previously. Indeed, the
sharp and intense peak at 460 cm−1 suggested that the geomet-
rical arrangement between the three structures of CeO2, CeAlO3,
and transition Al2O3 could favour the formation of CeAlOx phase
[25,29]. Moreover, a weak and broad peak at 615 cm−1 was also
due to the location of the zone-centre longitudinal optical phonon
mode due to oxygen vacancies in the cerium–aluminium mixed
phase [25]. Furthermore, the inhomogeneous distribution of the
crystal size and strain could account for the broad and asymmet-
ric character of the weak peaks, and the increase in strain due to
the redox Ce4+/Ce3+ ions was attributed to an increase in the molar
fraction of oxygen vacancies [25]. However, the peak ascribed to the
oxygen vacancies (i.e., 580 cm−1) was not well-defined because it
may overlap with the peaks of alumina.
In addition, the crystallite size of CeAl increased (from TEM
results) in comparison to that of CeMn and NiAl. According to the
findings, this fact could be associated to effects of peaks shifting
to higher energies in Raman spectrum of Ce based compounds
[25]. These facts were strongly correlated to the vacancy con-
centration and were attributed to strain, defects and variations
in phonon relaxation with particle size [25,43]. The aforemen-
tioned effects could explain either the oxygen adsorption on
Fig. 4. (a) Nitrogen adsorption isotherms and (b) the corresponding BJH pore size distribution of CeAl, NiAl and CeMn.

F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
133
NiAl
CeMn
CeAl
Fig. 5. CO2-TPD curves of the solids.
evidenced by chemical analysis. Among the studied oxides, CeMn
possessed the lowest surface area and pore volume.
200
400
600
800
1000
The textural properties and chemical analysis results of CeAl and
NiAl composites clearly indicated that the metal precursors were
well dispersed on silica surface. Even if the silica mold remained,
CeAl and NiAl displayed a significantly larger surface area and pore
volume than CeMn. Moreover, NiAl displayed the highest surface
area (e.g., 400 m2 g−1) and largest pore volume (1.20 cm3 g−1), and
these textural parameters are comparable to those of SBA-15 type
material.
Fig. 6. TPR profiles of the oxides studied.
Moreover, NiAl displayed a broad band with two superposed
desorption shoulders centred at 200 and 385 ◦C. Thus, the basicity of
NiAl due to NiO phases was lower than that of CeAl. The appearance
of these features corresponded to the formation of new, more basic
sites due to NiAl2O4 in NiAl, as observed for bulk nickel aluminates
[30].
The pore size of CeMn, NiAl and CeAl varied between 1.6 and
2.4 nm, which is comparable to the wall thickness of the sil-
ica template (2.2 nm) [31]. However, the pore diameters of the
nanocomposites are lower than that of SBA-15, which is charac-
teristic of the preparation method. These results revealed that the
size of the pores in the template plays an important role in the
replication of the nanocomposites. Due to the presence of micro-
pores (mainly that of 0.1 cm−3 g−1) on CeMn, NiAl and CeAl, ordered
mesostructures could not be obtained. Additionally, residual silica
may exert a stabilisation effect on the nanoparticles.
It is well-known that silica–alumina composite oxides are
acidic [6,17]. The presented catalysts are considered to be
SiO2–CeO2–MnOx and SiO2–CeO2–Al2O3 nanocomposite oxides,
which seem to be solid acids. However, basicity measurements
show that these solids have weak acidic sites due to CeO2 or
MnOx addition. This result confirmed that the basicity of CeMn was
greater than that of CeAl and NiAl. In fact, Ce promotes more basic
solids than Ni or Al.
The TPR technique was employed to establish the type of
reducible species formed in the catalysts after calcination. Fig. 6
depicts the TPR profiles of the samples.
3.4. Basicity measurements by CO2-TPD and TPR analyses
The typical TPR profile of Ce-containing silica materials was
observed for CeMn. Namely, a maximum appeared at 550 ◦C, and
two broad peaks were detected. The peak observed at about 400 ◦C
was attributed to the reduction of surface oxygen from Ce4+ [34]
whereas that observed at 700 ◦C was attributed to the reduction of
bulk oxygen, and may be ascribed to the CeO structure, in agree-
ment with the findings [49]. Remarkably, the broad peak in the
profile of CeMn was assigned to the reduction of MnOx at 650 ◦C
[34,49]. These results suggest that the small crystallites of CeO2
and MnOx or even MnCeOx (from Raman analysis) nanoparticles
on the silica surface were not mobile enough to agglomerate and
undergo reduction. Indeed, MnCeOx favours the mobility of oxygen
species and thus facilitates the reduction of manganese oxide [49].
Because Al2O3 does not reduce until a temperature of 1000 ◦C
is attained, broad peaks centred at 275, 386, 574 and 775 ◦C were
observed in the profile of CeAl. The peaks below 386 ◦C are probably
associated with adsorbed oxygen on alumina [25]. The shoulder at
574 ◦C could be due to the reduction of surface oxygen of Ce species
in strong interactions with aluminium and silica. The peak at 775 ◦C
was attributed to the further reduction of Ce3+ to Ce0. Moreover, the
presence of a reduction peak at 775 ◦C suggested that the reduction
of these species leads to the formation of or CeAlO3 on surface due
to difficulties in the reduction of CeO2, according to an early work
[25].
The basicity of the nanocomposites was evaluated by perform-
ing temperature-programmed desorption (CO2-TPD). According to
the findings, adsorption of CO2 is interesting due to the fact that
CO2 acts as an inhibitor of CO oxidation [46]. Although CO2 appears
to be the proper sonde molecule because of its acidic nature, the
CO2-TPD profile in the literatures varies depending on the adsorp-
tion condition of CO2 [47,48]. As a matter of fact, only a broad
desorption peak appeared when too much of CO2 was adsorbed.
However, over CeO2-based compounds, the adsorbate adsorbed on
weaker sites is desorbed at lower temperatures and that adsorbed
on stronger sites is desorbed at higher temperature [46,47]. For
instance, CeMn displayed two desorption peaks (Fig. 5). One peak
was centred at 350 ◦C and was attributed to medium-strength basic
sites. In addition, another peak was observed at 425 ◦C and was
ascribed to strongly basic sites such as that of O2− vacancies, as
suggested by Raman analysis.
The aforementioned peaks were attributed to CO2 adsorbed on
Ce sites or/and CeO2 vacancies, which were supplied by oxygen
from CO2. In the curve of CeAl, the peaks were shifted to lower
values, and an additional peak was observed at 340 ◦C. Thus, the
number of medium and strongly basic sites in CeAl decreased,
and weakly basic sites were created due to the presence of Al3+
.

134
F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
50
40
30
20
60
results) can dehydrate glycerol; however, their number and acces-
CeAl
NiAl
sibility plays a key role in achieving higher degrees of conversion
[7]. Although the nanoparticles become agglomerated in NiAl, the
activity of the catalyst may be due to the presence of medium to
strong basic sites.
Accessible medium to strong basic sites in CeMn were respon-
sible for the initiation of the catalyst. Therefore, Mn species can be
used as an alternative to Al, which is frequently used to add Lewis
acidic sites to CeO2 in dehydration reactions. Thus, the superior
catalytic performance of CeMn was due to the basic properties of
CeO2 and the stability provided by Mn2O3 and MnO2 phases (from
Raman and HRTEM results). Indeed, in CeMn, the reducibility of
these species in the reaction media was unlikely, as predicted by
TPR analysis.
50
40
80
CeMn
60
40
20
100
150
200
250
300
Temperature (ºC)
Selectivity to 1-hydroxyacetone and other products are shown
in Table 2.
Fig. 7. Glycerol dehydration as a function of the reaction temperature. Reaction
conditions: P = 1 atm; 0.1 g of catalysts; glycerol to water molar ratio of 0.25; time
on stream 1 h.
As expected, the selectivity for 1-hydroxyacetone was elevated
at temperatures as low as 200 ◦C. Moreover, light C2 and C3 by-
products such as acetone, propanediols, acetaldehyde, acrolein and
COx were detected. At temperatures greater than 200 ◦C, the selec-
tivity for 1-hydroxyacetone decreased drastically, reaching values
less than 5% at 300 ◦C due to the formation and occurrence of
secondary reactions of heavy by-products from 1-hydroxyacetone
side reactions. Moreover, at higher temperatures, greater amounts
of acetone, allyl alcohol, COx, formaldehyde, aromatics, heavy
compounds and unknown products (e.g., other products) were
detected, leading to a severe reduction in the selectivity to 1-
hydroxyacetone.
TPR profile of NiAl consisted of three reduction peaks at 513,
642 and 900 ◦C. The first peak was attributed to the reduction of
free NiO nanoparticles on the surface of the solid, and the second
peak was ascribed to hard to reduce Ni2+ species that interact with
SiO2. Lastly, the third peak was assigned to the reduction of NiAl2O4
[30].
3.5. Catalytic properties in the dehydration of glycerol
Usually, the dehydration of glycerol can follow two distinct
pathways. The first main pathway of the reaction is the dehy-
dration of glycerol to acrolein, which is much easier than the
dehydration of glycerol to 1-hydroxyacetone [6,17], the sec-
ond pathway. As described in the literature, although most of
the dehydration sites could be acidic sites, this hypothesis is
valid for the formation of acrolein whereas the selectivity to 1-
hydroxyacetone is due to the activity of the basic sites. Also,
in the absence of strong Brønsted acidic sites, 1-hydroxyacetone
or acrolein reactions lead to nonselective products and facili-
tate rapid deactivation of the catalyst [6,17]. If basic sites are
introduced, the deactivation of the catalysts decreases dramat-
ically, and the selectivity of the products over prolonged times
increases at relatively low temperatures [6,7]. Thus, basic sites
may not be transformed at temperatures below 250 ◦C. However,
as the temperature of the reaction increases above 250 ◦C, the
selectivity to 1-hydroxyacetone decreased due to thermal cracking
of the products and formation of heavy aromatics, namely other
products.
Therefore, these observations could be used as a basis for
explaining the superior performance of the catalysts. CeMn dis-
played improved conversion and product selectivity compared to
the other catalysts. The superior activity of CeMn may be attributed
to a better acid–base and/or redox properties displayed by the
nanocomposite. Additionally, CeMn exhibited high selectivity for 1-
hydroxyacetone due to the strength of its basic sites. Namely, when
the number of basic sites is low, the production of other products
over CeAl is enhanced.
The dependence of the conversion of glycerol on the tempera-
ture of the reaction under otherwise identical conditions is shown
in Fig. 7.
Pure and SiO2-supported CeO2, MnO2, NiO and Al2O3 displayed
relatively low activity, achieving conversions of approximately
5% at temperatures greater than 250 ◦C, as observed in an ear-
lier work with ternary composites [16]. The conversion of silica
template-based catalysts increased with an increase in the reaction
temperature and was higher on CeMn than CeAl. NiAl presented
intermediate activity due to the stability of highly exposed NiAl2O4
nanoparticles. At temperatures as low as 200 ◦C, all of the catalysts
displayed modest catalytic performance. However, at higher tem-
peratures, the conversion increased suddenly as the temperature
increased from 250 ◦C to 300 ◦C, and constant rates of conversion
were achieved at 250 ◦C due to the stability of the nanoparticles on
silica.
The catalytic activity was strongly dependent on the compo-
sition of the solids. Although CeMn and NiAl catalysts were very
active at temperatures of 250–300 ◦C, conversion did not exceed
55% at temperatures less than 250 ◦C. Alternatively, CeAl catalysts
achieved conversions close to 40% at 250 ◦C. The slightly low activ-
ity of CeAl can be attributed to several factors: (i) the instability of
Ce species and the inactivity of Al2O3 reduced the activity of the
catalysts. (ii) Although these species are accessible (HRTEM and
textural properties results), their sites are less accessible than dis-
persed CeMn nanoparticles. (iii) After accounting for the basicity of
the solids, the strength of basic sites on CeAl was lower than that of
CeMn. Thus, weak and medium basic sites of CeAl (from CO2-TPD
Table 2
Selectivity to 1-hydroxyacetone and other products at different temperatures. Reaction conditions: P = 1 atm; 0.1 g of catalysts; glycerol to water = 0.25; reaction time on
stream 1 h.
Sample
(%) Selectivity to 1-hydroxyacetone
(%) Selectivity to other products
150 ◦
C
200 ◦
C
250 ◦
C
300 ◦
C
350 ◦
C
150 ◦
C
200 ◦
C
250 ◦
C
300 ◦
C
350 ◦
C
CeMn
CeAl
NiAl
60
17
42
58
12
36
56
9
19
27
3
13
2
0
5
40
83
58
42
88
64
44
91
81
73
97
87
98
100
95

F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
135
Fig. 9. Raman spectrum of spent CeMn catalyst.
[9], followed by no appreciable changes in 1-hydroxyacetone selec-
tivity. NiO and Al2O3 are inactive in glycerol dehydration; thus, the
NiAl2O4 active phase present in NiAl nanoparticles (from XRD and
TPR results) could enhance the catalytic activity and stability of the
solid. Spinel phases are hard to reduce and display good stability;
however, coke formation deteriorated the catalytic performance of
the material.
Thereby, the highest activity and better stability of CeMn was
attributed to the strength of the basic sites on the solid. However,
NiAl catalyst with similar basic strength (Fig. 5) was also stable, and
glycerol conversion was close to 60% versus 71% for CeMn. There-
fore, the existence of a CeO2 and MnOx redox couple should be
considered to explain the catalytic results. The redox ability of Ce
and the high mobility of oxygen in the crystal structure are impor-
tant properties of CeO2 [24]. As a result, these oxides are capable of
reversibly adsorbing oxygen from glycerol or water followed by the
deoxygenation of these molecules. Studies on the dehydration of
alcohols over CeO2 have shown that reduced CeO2 surface is readily
oxidized by water at about 300 ◦C [13]. This result was confirmed
by the Raman spectrum of the spent CeMn catalyst that suggested
that a large amount of water was adsorbed on CeMn (the included
Fig. 9).
Fig. 8. Plots of the conversion against the time on stream for CeMn, NiAl, and CeAl
samples. Reaction conditions: P = 1 atm; 0.1 g of catalysts; glycerol to water molar
ratio of 0.25; time on stream 6 h. T = 250 ◦C.
The performance of the catalysts in the dehydration of glyc-
erol at 250 ◦C is presented in Fig. 8 as a function of the time on
stream.
For CeAl, the conversion of glycerol decreased with an increase
in the time on stream due to a deactivation phenomenon that
negates the expected conversion stabilisation effects. The conver-
sion of glycerol under steady state conditions was 71% for CeMn
while that of NiAl was 60% (Fig. 8). Thus, the presence of basic
sites of medium and strong strength on CeMn and NiAl (Fig. 5)
was directly related to the activity of these catalysts, as for rare
earth and alkaline-doped metal oxide catalysts [7,14]. However,
this relationship was not observed when the basicity decreased. If
the alkaline property was critical to reaction, CeAl should be more
active than NiAl. However, other factors such as the stability of CeO2
and MnOx, and NiO, Al2O3, NiAl2O4 nanoparticles dispersed on sil-
ica contribute to the catalytic performance, and this could explain
the stable performance in the later stages of catalytic run.
All the Raman modes of CeMn after the reaction were shifted to
higher wavenumber and the peak at 580 cm−1 (Fig. 9), which was
ascribed to oxygen vacancies that had greatly decreased; thus, it can
be suggested that the adsorption of the water by the oxides defects
may be occurring. Moreover, due to the insufficiency of oxygen
vacancies in the core of the nanoparticles, Mn species are dis-
tributed as uncoupled spins or are mediated through oxygen ions,
as observed on supported ZrO2 [49]. Therefore, synergistic effects
between Ce and Mn were relatively strong, particularly when SiO2
was present. As a result, the activity of CeMn catalysts was superior.
In order to make clear the correlation between the stabilisation
effect of the silica and the activity of the solids, the catalytic per-
formance of the CeMn, CeAl and NiAl without silica in the glycerol
dehydration is shown in Table 3.
CeAl presented modest performance over time on stream due
to the instability of Ce species in the reaction media, which can
be attributed to the degradation of the former to Al2O3 or CeO2.
Although ceria modifiers prevent the loss of alumina and limit
conversion to Al2O3, the latter species are ineffective in the dehy-
dration of glycerol, as previously mentioned [6,17]. Silica is an
inert oxide in dehydration reactions; thus, CeO2 particles inter-
act with Al2O3 and form preferentially stable nanoparticles, as
suggested by the TPR results. On the other hand, the low activ-
ity of the CeAl catalyst could be attributed to the presence of
Al2O3 and CeO2 on the surface of the material, which results in the
sintering of the nanoparticles. Moreover, among the studied mate-
rials, CeAl displayed remarkable textural properties, however, the
relatively low number basic sites could account for the modest per-
formance of CeAl in the reaction. Another fact to be considered is
that over the less stable CeAl catalyst, 1-hydroxyacetone transfor-
mation by decomposition and consecutive reactions to form cyclic
compounds could be likely, as observed by the findings [19].
Conversely, in CeMn, the interactions among Ce, Si and Mn
species may occur (from TPR results), which increased the stabil-
ity of CeMn nanoparticles in the reaction media, and lead to high
activity and stability. Additionally, based on the results obtained
from catalysts with both acidic and basic sites, the conversion over
CeMn increased gradually reached a maximum (ca. 4–6 h) due to
the presence of an induction period and the activation of active sites
Obviously, there is a close relationship between the conversion
of glycerol and the presence of silica in the catalysts. The glycerol
Table 3
Conversion of glycerol and products distribution over CeMn, NiAl and CeAl nanopar-
ticles synthesised without silica. Reaction conditions: P = 1 atm; 0.1 g of catalysts;
glycerol to water molar ratio of 0.25; time on stream 1 h. T = 250 ◦C.
Catalyst
CeMn
NiAl
CeAl
% Conversion of glycerol
1-Hydroxyacetone
Others
12
52
48
7
24
78
0
–
–

136
F.d.A.A. Barros et al. / Catalysis Today 212 (2013) 127–136
conversion over CeMn nanoparticles exhibited higher values than
that over NiAl and CeAl, and this behaviour was quite similar
from that of silica-containing CeMn, CeAl and NiAl composites. The
synergetic effect previously observed when silica is present was
restrained herein and activity of the solids was low due to sintering
of the nanoparticles, as glycerol is considered a strong reductant at
relatively high temperature reactions [6]. With regard to the prod-
ucts distribution, 1-hydroxyacetone was formed in major amounts.
Others contain allyl alcohol, acetic acid, propanediols, cyclic-ether
products, gases and unknown products, as previously observed
upon using these reaction conditions [17,50].
Thus, stability of the solids depends on the silica presence and
good catalytic activity was attributed to a high concentration of
surface defects and large amounts of Ce and Mn ions on the sur-
face of the material that were stabilised by silica, as for silica-based
compounds [51]. Hence, Ce coupled Mn ions limit nanoparticles
sintering, which results in high activity over time on stream.
[11] C.L. Lima, H.S.A. de Sousa, S.J.S. Vasconcelos, J.M. Filho, F.F. de Sousa, A.C.
Oliveira, Reaction Kinetics Mechanisms and Catalysis 102 (2011) 487–500.
[12] Y.T. Kim, K.-D. Jung, E.D. Park, Applied Catalysis A: General 393 (2011)
275–287.
[13] S. Sato, R. Takahashi, T. Sodesawa, N. Honda, Journal of Molecular Catalysis A:
Chemical 221 (2004) 177–183.
[14] C.L. Lima, S.J.S. Vasconcelos, J.M. Filho, B.C. Neto, M.G.C. Rocha, P. Bargiela, A.C.
Oliveira, Applied Catalysis A: General 399 (2011) 50–62.
[15] P. Lauriol-Garbay, J.M.M. Millet, S. Loridant, V. Bellière-Baca, P. Rey, Journal of
Catalysis 280 (2011) 68–76.
[16] H.S.A. Sousa, F.A.A. Barros, S.J.S. Vasconcelos, J.M. Filho, C.L. Lima, A.C. Oliveira,
A.P. Ayala, M.C. Junior, A.C. Oliveira, Applied Catalysis A: General 406 (2011)
63–72.
[17] M. Pagliaro, M. Rossi, The Future of Glycerol: New Uses of a Versatile Raw
Material, RSC Green Chemistry Book Series, 2008, pp. 54–64.
[18] S.J.S. Vasconcelos, C.L. Lima, J.M. Filho, A.C. Oliveira, E.B. Barros, F.F. de Sousa,
M.G.C. Rocha, P. Bargiela, A.C. Oliveira, Chemical Engineering Journal 168 (2011)
656–664.
[19] J. Deleplanque, J.-L. Dubois, J.-F. Devaux, W. Ueda, Catalysis Today 157 (2010)
351–358.
[20] A. Ulgen, W.F. Hoelderich, Applied Catalysis A: General 400 (2011) 34–38.
[21] Y. Wang, Y. Wang, J. Ren, Y. Mi, F. Zhang, C. Li, X. Liu, Y. Guo, G. Lu, Journal of
Solid State Chemistry 183 (2010) 277–284.
[22] G. Satishkumar, L. Titelman, M.V. Landau, Journal of Solid State Chemistry 182
(2009) 2822–2828.
4. Conclusions
[23] M. Abecassis-Wolfovich, M.V. Landau, A. Brenner, M. Herskowitz, Journal of
Catalysis 247 (2007) 201–213.
[24] B. Puertolas, B. Solsona, S. Agouram, R. Murillo, A.M. Mastral, A. Aranda, S.H.
Taylor, T. Garcia, Applied Catalysis B: Environmental 93 (2010) 395–405.
[25] J.Z. Shyu, W.H. Weber, H.S. Gandhi, Journal of Physical Chemistry 92 (1988)
4964–4970.
[26] I.O. Mazali, B.C. Viana, O.L. Alves, J. Mendes Filho, A.G. Souza Filho, Journal of
Physics and Chemistry of Solids 68 (2007) 622–627.
[27] A. Iriondo, V.L. Barrio, J.F. Cambra, P.L. Arias, M.B. Guemez, M.C. Sanchez-
Sanchez, R.M. Navarro, J.L.G. Fierro, International Journal of Hydrogen Energy
35 (2010) 11622–11633.
[28] D. Jiang, M. Zhang, G. Li, H. Jian, Catalysis Communications 17 (2012) 59–63.
[29] H. Li, G. Lu, Q. Dai, Y. Wang, Y. Guo, Y. Guo, Applied Catalysis B: Environmental
102 (2011) 475–483.
[30] R.M. Freire, F.F. de Sousa, A.L. Pinheiro, E. Longhinotti, J.M. Filho, A.C. Oliveira,
P.T.C. Freire, A.P. Ayala, A.C. Oliveira, Applied Catalysis A: General 359 (2009)
165–179.
[31] A.H.M. Batista, F.S.O. Ramos, T.P. Braga, C.L. Lima, F.F. de Sousa, A.S. Oliveira,
E.B. Barros, J.R. de Souza, A. Valentini, J.M. Filho, A.C. Oliveira, Applied Catalysis
A: General 382 (2010) 148–157.
[32] C.-J. Jia, Y. Liu, W. Schmidt, A.-H. Lu, F. Schüth, Journal of Catalysis 269 (2010)
71–79.
High surface area silica-containing CeMn, NiAl and CeAl
nanocomposites were produced in the present study. Residual sil-
ica stabilised the nanoparticles of the oxides, and materials with
distinct phases, good textural properties and basic sites with dis-
tinct strengths and accessibilities were obtained. NiAl and CeAl
catalysts were less effective in the dehydration of glycerol due to
the relatively low number and accessibility of the basic sites. Syn-
ergistic effects were observed between cerium and manganese,
particularly when SiO2 was used. As a result, the activity of the
catalysts was greater than that of cerium and aluminium or nickel
and aluminium catalysts. CeO2, Mn2O3, and MnO2 phases gen-
erated surface defects, and the relatively high number of cerium
and manganese species on the surface of the material enhanced
glycerol dehydration due to the formation of acrolein and 1-
hydroxyacetone at 250 ◦C. Moreover, the sintering of cerium and
manganese nanoparticles was limited due to the silica presence
and the activity and stability of the catalyst over time on stream
was high.
[33] J. Roggenbuck, H. Schafer, T. Tsoncheva, C. Minchev, J. Hanss, M. Tiemann,
Microporous and Mesoporous Materials 101 (2007) 335–341.
[34] Yu Dai, Xingyi Wang, Qiguang Dai, Dao Li, Applied Catalysis B: Environmental
111-112 (2012) 141–149.
[35] J.I. Gutierrez-Ortiz, B. de Rivas, R. Lopez-Fonseca, S. Martın, J.R. Gonzalez-
Velasco, Chemosphere 68 (2007) 1004–1012.
Acknowledgments
[36] J. Cao, Q. Mao, L. Shi, Y. Qian, Journal of Materials Chemistry 21 (2011)
16210–16215.
[37] B.M. Reddy, K.N. Rao, G.K. Reddy, A. Khan, S.-E. Park, Journal of Physical Chem-
istry C 111 (2007) 18751–18758.
[38] F. Zhou, H. Zheng, X. Zhao, Q. Guo, X. Ni, T. Shen, C. Tang, Nanotechnology 16
(2005) 2072–2076.
The authors are grateful for financial support from CNPq for this
research project (MCT/CNPq 482364/2010-6) as well as FUNCAP
with Contract 0011-00206.01.00/09. We acknowledge CETENE for
providing the HRTEM images.
[39] J. Zhu, Q. Gao, Microporous and Mesoporous Materials 124 (2009) 144–152.
[40] T. Valdés-Solís, A.B. Fuertes, Materials Research Bulletin 41 (2006)
2187–2197.
References
[41] J.R. McBride, K.C. Hass, B.D. Poindexter, W.H. Weber, Journal of Applied Physics
76 (1994) 2435–2441.
[42] Z.Y. Pu, J.Q. Lu, M.F. Luo, Y.L. Xie, Journal of Physical Chemistry C 111 (2007)
18695–18702.
[43] F. Buciuman, F. Patcas, R. Craciun, D.R.T. Zahn, Physical Chemistry Chemical
Physics 1 (1999) 185–190.
[44] W.-J. Hong, S. Iwamoto, S. Hosokawa, K. Wada, H. Kanai, M. Inoue, Journal of
Catalysis 277 (2011) 208–216.
[45] M.A. Laguna-Becero, M.L. Sanjuán, R.I. Merino, Journal of Physics: Condensed
Matter 19 (2007) 186217–186227.
[46] G. Avgouropoulos, T. Ioannides, Journal of Molecular Catalysis A: Chemical 296
(2008) 47–53.
[47] Istadi, N.A.S. Amin, Journal of Molecular Catalysis A: Chemical 259 (2006)
61–66.
[48] X. Wang, X. Ma, V. Schwartz, J.C. Clark, S.H. Overbury, S. Zhao, X. Xu, C. Song,
Physical Chemistry Chemical Physics 14 (2012) 1485–1492.
[49] X. Wu, S. Liu, D. Weng, F. Lin, R. Ran, Journal of Hazardous Materials 18 (2011)
7283–7290.
[1] L.-Z. Tao, S.-H. Chai, Y. Zuo, W.-T. Zheng, Y. Liang, B.-Q. Xu, Catalysis Today 158
(2010) 310–316.
[2] N. Hamzah, N.M. Nordin, A.H.A. Nadzri, Y.A. Nik, M.B. Kassim, M. Ambar Yarmo,
Applied Catalysis A: General 419–420 (2012) 133–137.
[3] S. Panyad, S. Jongpatiwut, T. Sreethawong, T. Rirksomboon, S. Osuwan, Catalysis
Today 174 (2011) 59–64.
[4] D. Roy, B. Subramaniam, R.V. Chaudhari, Catalysis Today 156 (2010)
31–37.
[5] A. Witsuthammakul, T. Sooknoi, Applied Catalysis A: General 413–414 (2012)
109–116.
[6] C.-H. Zhou, J.N. Beltramini, Y.-X. Fan, C.Q. Lu, Chemical Society Reviews 37
(2008) 527–549.
[7] A.K. Kinage, P.P. Upare, P. Kasinathan, Y.K. Hwang, J.-S. Chang, Catalysis Com-
munications 11 (2010) 620–623.
[8] A. Iriondo, J.F. Cambra, M.B. Güemez, V.L. Barrio, J. Requies, M.C. Sánchez-
Sánchez, R.M. Navarro, International Journal of Hydrogen Energy 37 (2012)
7084–7093.
[9] A.S. de Oliveira, S.J. Vasconcelos, J.R. de Sousa, F.F. de Sousa, J.M. Filho, A.C.
Oliveira, Chemical Engineering Journal 168 (2010) 765–774.
[10] Z. Huang, F. Cui, H. Kang, J. Chen, C. Xia, Applied Catalysis A: General 366 (2009)
288–298.
[50] Y. Feng, H. Yin, A. Wang, L. Shen, L. Yu, T. Jiang, Chemical Engineering Journal
168 (2011) 403–412.
[51] P. Lauriol-Garbey, S. Loridant, V. Bellière-Baca, P. Rey, J.-M.M. Millet, Catalysis
Communications 16 (2011) 170–174.