Dehydrogenation of methylcyclohexane to toluene over partially reduced silica-supported Pt-Mo catalysts

Article abstract of DOI:10.1016/j.molcata.2016.04.011

A series of bifunctional Pt/Mo(x)-SiO₂ catalysts were investigated in the context of the hydrogen storage using the methylcyclohexane (MCH)-toluene-hydrogen cycle. The performance of partially reduced catalysts for hydrogen production was evaluated in the MCH dehydrogenation reaction carried out in a fixed-bed flow reactor at 673 K and total hydrogen pressure of 2.2 MPa. The catalysts with different acidity and Mo loading (4.1-12.7 wt.%) were prepared by impregnation of the calcined Mo(x)-SiO₂ samples with Pt precursor. The oxide catalyst precursors were characterized by chemical analysis (ICP-AES), N₂ physisorption at 77 K, X-ray diffraction (XRD), Raman spectroscopy and temperature-programmed reduction (H₂-TPR) techniques whereas the partially reduced samples were characterized by temperature-programmed adsorption of ammonia (TPD-NH₃₎, high-resolution transmission electron microscopy (HRTEM) and X-ray photoelectron spectroscopy (XPS). The catalyst having optimized Mo loading (8.0 wt.%) was the most active and its best catalytic performance was attributed to the high dispersion of MoO₂ and Pt° phases and its lowest deactivation by coke formation. The activity drop of the catalysts having larger Mo loading (10.6 and 12.7 wt.%) was linked with formation of MoO_x-Pt core-shell nanoparticles, as confirmed by HRTEM.

Full text of DOI:10.1016/j.molcata.2016.04.011

Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Dehydrogenation of methylcyclohexane to toluene over partially
reduced silica-supported Pt-Mo catalysts
N. Boufadena,∗, R. Akkaria, B. Pawelecb, J.L.G. Fierrob, M. Said Zinaa, A. Ghorbela
a Laboratoire de Chimie des Matériaux et Catalyse, Département de Chimie, Faculté des Sciences de Tunis, Université de Tunis El Manar, Campus
Universitaire Farhat Hached, Rommana 1068, Tunis, Tunisia
b Instituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of bifunctional Pt/Mo(x)-SiO2 catalysts were investigated in the context of the hydrogen stor-
age using the methylcyclohexane (MCH)-toluene-hydrogen cycle. The performance of partially reduced
catalysts for hydrogen production was evaluated in the MCH dehydrogenation reaction carried out in
a fixed-bed flow reactor at 673 K and total hydrogen pressure of 2.2 MPa. The catalysts with differ-
ent acidity and Mo loading (4.1–12.7 wt.%) were prepared by impregnation of the calcined Mo(x)-SiO2
samples with Pt precursor. The oxide catalyst precursors were characterized by chemical analysis
(ICP-AES), N2 physisorption at 77 K, X-ray diffraction (XRD), Raman spectroscopy and temperature-
programmed reduction (H2-TPR) techniques whereas the partially reduced samples were characterized
by temperature-programmed adsorption of ammonia (TPD-NH3), high-resolution transmission electron
microscopy (HRTEM) and X-ray photoelectron spectroscopy (XPS). The catalyst having optimized Mo
loading (8.0 wt.%) was the most active and its best catalytic performance was attributed to the high dis-
persion of MoO2 and Pt◦ phases and its lowest deactivation by coke formation. The activity drop of the
catalysts having larger Mo loading (10.6 and 12.7 wt.%) was linked with formation of MoOx-Pt core-shell
nanoparticles, as confirmed by HRTEM.
Received 22 February 2016
Received in revised form 12 April 2016
Accepted 13 April 2016
Available online 20 April 2016
Keywords:
Hydrogen storage
Methylcyclohexane
Toluene
Pt-Mo catalysts
Silica
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
[16–18]. However, because of the high cost of platinum and its lim-
ited terrestrial reserve, the major problem of those catalysts is the
Recently, there has been a great interest in the concept of
hydrogen storage in liquid organic hydrocarbons (LOH), such as
cyclohexane, methylcylohexane or decalin [1–4]. In this concept,
the hydrogen production from organic hydrocarbons occurs via
a catalytic dehydrogenation reaction, being the most favorable,
the methylcyclohexane-toluene-hydrogen cycle [5]. In parallel,
non-oxidative catalytic dehydrogenation of light alkanes demon-
strated to be effective method to produce unsaturated short-chain
hydrocarbons. Among different light alkanes, the catalytic dehy-
drogenation of propane is of particular interest being mixed oxides,
such as In2O3-Ga2O3 [6] and In2O3-Ga2O3-Al2O3 [7], promising
catalysts to fill the supply-demand gap for propylene.
high Pt content needed for their preparation (5–10 wt.%). For these
reasons, it was essential to find solutions that minimize the cata-
lyst cost while trying to keep the same catalytic performances as
monometallic Pt-based ones.
Thus, with the objective to minimize the catalyst cost, some
attempts have been made to substitute Pt by non-noble met-
als, such as Ni [19,20] or Mo [21,22]. In particular, the partially
reduced MoOx species were found to be effective in the dehydro-
genation/isomerization of MCH at varied temperature and pressure
conditions [21,22]. This was explained by the presence of Brønsted
acidic groups Mo-OH which led to the formation of bifunctional
MoO2(Hx)ac phase [21,22]. However, as a consequence of the high
metal loading, it was difficult to obtain a good dispersion of the
active phases on the surface of the support [21]. Indeed, it was
reported that the enrichment of the Mo surface decreased or even
disappeared in the presence of aromatic hydrocarbons strongly
adsorbed on the surface of metal particle [23–25].
Since the pioneer works by Sinfelt [5,8], the most suitable cat-
alysts studied for dehydrogenation of methylcylohexane (MCH)
were mainly platinum-based ones, such as Pt/Al2O3 [9–14], Pt/HY
[15], Pt/V2O5, Pt/Y2O3 [4,16,17], and alumina-supported Pt-Re
Taking into account the good catalytic response of Mo-SiO2
catalysts prepared by sol-gel method in the reaction of MCH
dehydrogenation [21], the promotion of Mo-based catalysts with
∗
Corresponding author.
E-mail address: nesrinebfd@gmail.com (N. Boufaden).
http://dx.doi.org/10.1016/j.molcata.2016.04.011
1381-1169/© 2016 Elsevier B.V. All rights reserved.

N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
97
platinum seems to be a good alternative for reducing the platinum
content of the Pt monometallic ones. Since the catalyst dehydro-
genation activity increases with the increase of the dispersion of
active phases [26], the approach undertaken in this work consists
in the preparation of silica-supported bifunctional Pt-Mo catalysts
with optimized Mo loading and therefore to enhance the intrinsic
activity of partially reduced Mo-SiO2 catalysts studied previously
[21] by their promotion with Pt. In fact, platinum crystallites have
the advantage to remove rapidly the hydrogen from the reaction
system thanks to their ability to easily dissociate hydrogen atoms,
which induces the shift of the chemical equilibrium in favor of
the products [27–29]. Moreover, the combination of the ability of
Mo atoms to adsorb hydrocarbon species with the rapid dissocia-
tion of hydrogen atoms on Pt crystallites [30] may improve both
activity and selectivity with respect to monometallic Pt-based cat-
alysts [23–25]. In fact, Pt-Mo catalysts were reported to be active
in many hydrotreating reactions such as hydrogenolysis, hydro-
genation and dehydrogenation [23–25,31]. In this sense, the large
body of the existing work on the dehydrogenation of cycloalkanes
allowed to define the catalyst composition and preparation pro-
cedures, which are appropriate to minimize the undesired coke
deposition on the catalyst surface and also to enhance the stabil-
ity during on-stream operations [32], which pointes out a possible
hydrogen spill-over effects. Indeed, the importance of spilled-over
hydrogen in the enlargement of surface area was confirmed for the
catalyst prepared by physical mixing of MoO3 with Pt black [33], as
well for the Pt-Mo phases supported on alumina, TiO2 and Mg-Al
[34–36].
The use of a Pt-Mo bimetallic catalyst formulation would create
new catalytic sites [30]. In fact, it was found that the reduction of
Pt/MoO3 catalyst with H2 at 673 K proceeds via the formation of
hydrogen molybdenum bronze phase, HxMoO3 [33], which could
be responsible for the generation of new catalytic sites as it was
reported by Belatel et al. [34]. Concerning the metal function, the
enhancement of hexane hydrogenolysis over the Pt-Mo catalyst
with respect to monometallic platinum one was explained also
in terms of Pt-Mo interaction [31]. Similarly, the enhancement of
ethane hydrogenolysis over Mo-Pt/SiO2 catalysts was explained
also as a consequence of a change in the electronic properties of
platinum due to its interaction with molybdenum [35].
The silica-supported Mo catalysts studied in this work were
prepared by a sol-gel method and not by impregnation because
silica exhibits a low amount of basic OH groups that are able
to react with molybdates during impregnation [37]. Indeed, the
catalysts prepared by impregnation exhibit a poor dispersion of
the active phases on silica substrate which has been considered
as the main factor responsible for the low activity [38]. Moreover,
the molybdenum oxide structures formed on the support surface
after calcination generate more basal planes and lower quantity of
defects. This can be disadvantageous as a low fraction of defects
inhibit to some extent the reduction of MoO3 into MoO2 and after
catalyst activation by partial reduction they are the active sites for
dehydrogenation of MCH [21]. This is explained by the fact that
the kink and step active sites of MoO2 phase are important for the
catalyst activity [39].
In view of the observed discrepancies between the factors influ-
encing the dehydrogenation activities of Pt-Mo catalysts, this work
was undertaken with the objective to cover the surface of Mo-SiO2
catalysts as much as possible by incorporating a large amount of
MoO3 prior to platinum addition. Therefore, the partially-reduced
molybdenum oxide phase should be regarded as a support for
the platinum active phase, enabling the promoter to be optimally
dispersed. Thus, the objective was to elucidate the mutual influ-
ence of the modified SiO2 carrier by MoO3 and platinum, which
could be helpful in increasing our knowledge of this type of cata-
lysts. For the simultaneous control of the textural properties and
homogeneity of Mo species, the Mo-based catalysts were prepared
by the sol–gel process. The catalysts were characterized by a vari-
ety of physic-chemical techniques in an attempt to explain the
relationship between the catalyst performance and structure.
2. Experimental part
2.1. Catalysts preparation
Four bimetallic Pt/Mo(x)-SiO2 catalysts, coded hereafter as
Pt/Mo(x), where x is the Mo/Si molar percentage, were prepared
by wet impregnation of the calcined Mo(x)-SiO2 samples with an
aqueous solution of Pt[(NH2)6NO3]2 which volume was selected to
achieve Pt content of 5 wt.%. In the first step, Mo(x)-SiO2 samples
were prepared through a sol-gel method according to the pro-
cedure described previously [21]. Briefly, appropriate amounts of
HNO3 (used as a catalyst for hydrolysis of the silica precursor) and
C2H5OH (solvent; Sigma-Aldrich, 99.8% purity) were mixed in order
to adjust the HNO3/Si molar ratio at 0.12 with a concentration of
silica precursor in EtOH at the value of 0.2 mol L−1. Then, the tetra-
ethoxy-silane silica precursor (TEOS, Si(OC2H5)4, Acros, 98% purity)
was added dropwise, and then appropriate amount of molybdenum
precursor (molybdenum acetylacetonate, MoO2(C5H7O2)2, Merck-
Schuchardt) was incorporated in order to obtain Mo*100/Si molar
ratio of 5, 10, 12 and 15. The mixture was then stirred until gelation
while maintaining the temperature at 333 K. The obtained gel was
dried in an autoclave under ethanol supercritical conditions (521 K;
6.14 MPa), and then calcined at 773 K under a flux of 30 mL min−1 of
oxygen for 3 h. Finally, the Pt/SiO2 reference sample (coded here-
after as Pt/SiO2) was prepared by wet impregnation of the SiO2
substrate with an aqueous solution of Pt [(NH2)6NO3]2, Alfa Aesar)
whose volume was selected to obtain a final Pt content of 5 wt.%. For
this catalyst, the bare silica substrate was also prepared by sol-gel
method.
2.2. Catalyst characterization techniques
The synthesized Pt-Mo silica catalysts prepared by consec-
utive wet impregnation of calcined Mo-SiO2 precursors were
characterized by chemical analysis (ICP), textural analysis (N2-
physisorption), X-ray powder diffraction (XRD), H2-Temperature-
Programmed Reduction (H2-TPR), Temperature-Programmed Des-
orption of ammonia (TPD-NH3), High Resolution Transmission
Electron Microscopy (HRTEM), Raman and X-ray Photoelectron
Spectroscopic (XPS) techniques.
The Pt and Mo loadings of the calcined catalysts were
determined by Inductively Coupled Plasma Atomic Emission Spec-
troscopy (ICP-AES) with a Perkin Elmer Optima 3300DV equipment.
The solid samples were first digested for 2 h in a mixture of HF,
HCl and HNO3 carried out in a microwave oven. Then, aliquots of
solution were diluted to 50 mL using deionized water.
N2 adsorption-desorption isotherms were measured at 77 K on
a Micromeritics ASAP 2020 physical adsorption analyzer after sam-
ple degassing at 473 K for 3 h. The specific surface areas were
calculated according to the Brauner-Emmett-Teller (BET) model
[40] while the porous distributions were evaluated using the
Barrett-Joyner-Halenda (BJH) method [41] applied to the corre-
sponding nitrogen desorption isotherms.
The XRD measurements were performed in the reflection mode
within the wide angular range of 4–70◦ (2␪), using a Siemens D-
5000 powder diffractometer operated at 40 kV and 30 mA (Cu-K␣
radiation ␭ = 1.5406 Å) with a scan rate of 0.04◦ s−1. An estimation
of the crystalline size was obtained using Scherrer equation.
In order to determine the vibrational and rotational imprint of
the samples, Raman spectra were recorded using a Renishaw in

98
N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
Velocity (WHSV) of 92.4 h−1 and H2/MCH ratio of 250 L(N) L−1. The
liquid samples were analyzed using a GC Agilent 6890A equipped
with a FID and HP-Innowax column (Crosslinked Polyethylene
Glycol) of 30 m length and 0.25 mm of inner diameter and thickness
of 0.25 mm (split 300:1). The percent conversion was determined
based upon the GC results in terms of mass percent of the products
relative to the total mass. The major products resulting from the
MCH dehydrogenation were toluene and cyclohexane. Other minor
reaction products were isomerization and hydrocracking products:
trans- and cis-1,4-dimethylcyclohexane; trans- and cis-1,2-
dimethylcyclohexane; 2,4-dimethyl-hexane; ethyl-cyclohexane;
1,1-dimethyl-cyclopentane; 1,3-dimethyl-cyclopentane; trans-
1,2-dimethyl-cyclopentane; cis-1,2-dimethyl-cyclopentane, and
a low molecular weight paraffins. The selectivity of the catalysts
towards these products with respect to the time was determined
by:
Via Raman microscope spectrometer equipped with a laser beam
emitting at 532 nm and 100 mW output power. The photons scat-
tered by the sample were dispersed by a 1800 (o 1200) lines/mm
grating monochromator and simultaneously collected on a CCD
camera; the collection optics was set at 50× objective. The spectral
resolution was 1 cm−1
.
The reducibility of the samples was studied by Temperature-
Programmed Reduction with H2 (TPR-H2) using a semi-automatic
Micromeritics TPD/TPR 2900 apparatus interfaced to a computer.
The samples (0.04 g) were placed in a tubular quartz reactor and
swept by a mixture of 5 vol.% H2 in Ar at a total flow rate of
30 mL min−1 and then heated at a rate of 15 K min−1 up to a final
temperature of 1073 K.
The acidity of the partially-reduced samples was determined
by temperature-programmed desorption (TPD) of ammonia mea-
surements carried out with the same apparatus described for
TPR. Before TPD-NH3 experiment, 0.04 g of each sample was
pre-reduced in a quartz reactor at 673 K for 1 h in a flow of 5
vol.% H2 in Ar mixture. Following this, the sample was cooled
to 373 K and ammonia-saturated in a 5% NH3/He (Air Liquide)
flow (50 mL min−1) for 30 min. Then, the sample was subjected
to a flow of He (100 mL min−1) at 373 K for 15 min to remove any
physisorbed NH3 species from the surface. After catalyst equilibra-
tion in a helium flow at 373 K, ammonia was desorbed using a linear
heating rate of 10 K min−1 from 373 K to 873 K. In order to obtain the
total acidity of the catalysts, the areas under the curves in the NH3
desorption profiles were integrated while the Gaussian deconvo-
lution of the peaks allowed a semi-quantitative comparison of the
acid strength distribution.
Si(%) = ci × ꢀci
ꢀ
where ci is the molar concentration of product (i) and
sum of the molar concentrations of all the products.
A comparison of repeated experiments showed the experiments
to be highly reproducible. In addition, carbon balance in all exper-
iments was >99%.
ci is the
3. Results and discussion
3.1. Chemical analysis
The chemical compositions of the catalysts determined by ICP-
AES technique are summarized in Table 1. It is seen that the real
Pt and Mo contents are relatively close to the nominal ones. This
is because the solids were prepared via a combination of an acid
hydrolytic sol–gel route with an impregnation method. The former
method led to the immobilization of molybdenum species through
a kinetically controlled hydrolytic polycondensation around the
silicon atom leading to the formation of an amorphous silica frame-
work [42]. During the TEOS hydrolysis step, the protonation of the
alkoxide group occurs followed by a nucleophilic attack of the water
molecule leading to the formation of a pentacoordinate intermedi-
ate species. As a consequence of the electron withdrawn from the
silicon atom, the silicon atom became more electrophilic and sus-
ceptible to be attacked by water, and the leaving group behaved as
alkoxide [43]. Then, the polyanion of molybdate might destabilize
the positive intermediatespecies forming molybdenum-containing
The pre-reduced catalysts were studied by High Resolution
Transmission Electronic Microscopy (HRTEM) using a JEM 2100F
microscope operating with a 200 kV accelerating voltage and fitted
with an INCA X-sight (Oxford Instruments). Before analysis, all the
catalysts were crushed into powder, mixed with ethanol and then
dispersed in an ultrasonic bath for 10 min. A drop of the suspension
was then placed on a carbon-coated Cu grid. To evaluate the parti-
cle size distribution, several micrographs of the same sample were
analyzed.
X-ray photoelectron spectroscopy was performed in order to
reveal the chemical state and the surface composition of calcined,
fresh reduced and spent samples. A VG Scientific Ltd., system
equipped with a hemispherical electron analyzer and an Mg K␣
(h␯ = 1253.6 eV) X-ray source. The residual pressure in the ion-
pumped analysis chamber was kept below 9.3 × 10−13 MPa during
data acquisition. Peak intensities were estimated by calculating the
integral of each peak after subtraction of an S-shaped background
and fitting the experimental curve to a combination of Gaussian
and Lorentzian lines of varying proportion (G/L = 7–30%). The C 1s
peak at 284.8 eV (C C/C H) was used as the internal standard for
determining peak positions.
siloxane species (e.g., Si
O Mo O Si species) [42]. Additionally,
some inert Mo polyanions might be formed by their trapping
inside the silica during the TEOS hydrolysis and condensation steps.
After calcination, the MoO3-containing precursors were used for
the preparation of binary Pt-Mo catalysts. The ICP-AES analyses
(Table 1) also show that the impregnation of the Mo-SiO2 samples
with the platinum solution led to relatively small loss of molybde-
num phase.
2.3. Methylcyclohexane dehydrogenation
The catalytic performance of the Pt/Mo(x)-SiO2 catalysts was
evaluated in the methylcyclohexane (Spectrophotometric Grade
99%, Sigma-Aldrich) dehydrogenation reaction performed in
a bench scale high-pressure laboratory set-up equipped with
down-flow fixed bed reactor. In the first step, 0.10 g of the pow-
dered catalyst was dried at 433 K for 30 min with a N2 flow of
100 mL min−1. After that, the catalyst was reduced in-situ at 673 K
for 3 h with H2/N2 gaseous mixture (molar ratio H2:N2 = 1:5). The
optimum temperature for the catalyst reduction was adopted from
H2-TPR analysis. For activity tests, the MCH was introduced into
reactor via a high pressure pump (Knauer HPLC), through the pre-
heated line for mixing with H2 stream. The reaction was carried out
at 673 K, 2.2 MPa of total hydrogen pressure, Weight Hourly Space
3.2. Textural properties
The textural properties of the catalysts were examined by N2
adsorption-desorption isotherms at 77 K. Fig. 1(A) shows the N2
adsorption-desorption isotherms of all samples. The bimetallic cat-
alysts are mesoporous according to the IUPAC nomenclature as they
all show type IV isotherms [44]. However, Pt/Mo5, Pt/Mo10 and
Pt/Mo12 samples show a unique type H2 hysteresis loop which is
usually attributed to different size of pore mouth and pore body
i.e. bottle shaped pores [45]. Only Pt/Mo15 sample shows two
hysteresis loops, one in relative pressure range of 0.4–0.8 corre-
sponding to type H2 and another one in the relative pressure (P/P0)

N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
99
Table 1
Elemental analysis and textural properties of calcined Pt/Mo(x)-SiO2 samples.
c
c
Sample
Moa (wt.%)
Mob (wt.%)
Ptb (wt.%)
SBET (m2 g−1
)
Vtotal (cm3 g−1
)
Pore diameterc (nm)
Pt/Mo5
7
4.1
8.0
10.6
12.7
4.0
3.8
3.5
3.7
458
546
635
489
0.31
0.32
0.53
0.33
3.4
3.6
3.8
3.6
Pt/Mo10
Pt/Mo12
Pt/Mo15
13
15
18
a
Theoretical (wt.% Pt theoretical is fixed at 5).
As determined by ICP technique.
As determined by N2 adsorption-desorption isotherms at 77 K.
b
c
voluminous Pt and/or Mo aggregates blocking the pore mouths of
the support.
The specific surface area (SBET), total pore volume and aver-
age pore diameter of the prepared solids are listed in Table 1.
All samples exhibited a very high specific surface (SBET) area (in
the range 458–635 m2 g−1). From this table, it is evident that
the Pt/Mo(x)-SiO2 samples show the increase of surface area
as the Mo loading is increased: Pt/Mo12 (635 m2 g−1) > Pt/Mo10
(546 m2 g−1) > Pt/Mo15 (489 m2 g−1) > Pt/Mo5 (458 m2 g−1). A sim-
ilar increase in the specific surface of Pt-MoO3 catalysts supported
on amorphous alumina-silica, Al2O3 and mesoporous silica-
alumina was explained in terms of the redispersion of the MoO3
phase during Pt incorporation by impregnation with platinum salt
precursor [37,46,47]. The lower than expected SBET of Pt/Mo15 sam-
ple could be explained as due to the pore mouths blocking by a large
amount of Mo incorporated on the support surface, as suggested by
the large drop of total pore volume when going from Pt/Mo12 to
Pt/Mo15 sample.
Fig. 1(B) shows the pore size distribution (dV/dlog(D)) for all
catalysts, where V is the pore volume (cm3 g−1) and D is the pore
diameter (nm). All the catalysts exhibited only one type of pores
whose maximum ranges from 3.4 nm to 3.8 nm. The pore volume
follows the same trend as SBET: the total pore volume increases to
a maximum for Pt/Mo12 sample and then decreases for Pt/Mo15
one. The small change of pore diameter from Pt/Mo5 to Pt/Mo12
(Table 1) may indicate of the presence of PtO and MoO3 crystallites
larger than the pore mouths of the silica support which induces
the external blocking of the pores without filling them. Meanwhile,
the wide curve recorded for the Pt/Mo12 catalyst, would indicate
a large pore volume and a remarkable heterogeneity for its porous
distribution which fits well with the results reported earlier.
3.3. Wide-angle X-ray diffraction (XRD)
The XRD analysis of the samples gives useful information about
the crystalline phases present on the support surface. The X-ray
diffraction patterns of the four Pt-Mo(x) catalysts are shown in
Fig. 2. Pt/Mo5, Pt/Mo10 and Pt/Mo15 catalysts exhibit clear diffrac-
tion lines at 2␪ = 12.7◦; 23.3◦; 25.7◦; 27.3◦; 38.9◦ and 55.7◦ which are
attributed to the crystalline orthorhombic ␣-MoO3 phase (JCPDS
card 00-005-0508). The MoO3 crystal size estimated by the Debye-
Scherrer equation for the XRD line at 27.3◦ 2ꢁ follows the trend:
Pt/Mo15 (108.7 nm) > Pt/Mo5 (102.3 nm) > Pt/Mo10 (78.9 nm). On
the contrary, Pt/Mo12 did not show clear diffraction peaks belong-
ing to MoO3 phase indicating that the MoO3 particles have a crystal
size below the detection limit of XRD technique (4 nm) which fits
well the information deduced from N2-physisorption confirming
the highest dispersion of MoO3 species in this sample. The redis-
persion of MoO3 phases during wet impregnation of the calcined
Mo5-SiO2 sample with Pt precursor was also inferred from the N2
physisorption data (vide supra). In addition, all catalysts Exhibit 2
diffraction lines at 2ꢁ of 40.0◦ and 46.5◦ corresponding to (111) and
(200) reflections, respectively, of the fcc Pt metal phase (JCPDS card
Fig. 1. N2 adsorption-desorption isotherms at 77 K (A), and pore size distribution
(B) of calcined Pt/Mo(x)-SiO2 samples.
of 0.8–1.0 range which belongs to H4 type related to solids con-
sisting of agglomerates or aggregates forming slit-shaped pores
[45]. Thus, for samples prepared with low amounts of molybde-
num (Mo × 100/Si = 5–12 molar), the shape of N2 isotherms points
out to the formation of the mesopores. On the contrary, the hystere-
sis loop of the Pt/Mo15 sample suggests the formation of relatively

100
N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
Fig. 2. X-ray diffraction patterns of calcined Pt/Mo(x) samples.
Fig. 4. H2-TPR patterns of calcined Pt-SiO2 and Pt/Mo(x)-SiO2 samples.
be related to the triply coordinated oxygen (Mo3 O) stretching
mode. Otherwise, the 820 cm−1 peak is assigned to the doubly
coordinated oxygen (Mo2 O) stretching mode, which results from
corner-sharing oxygen atoms common to two octahedral. Although
the band at 820 cm−1 is related to the symmetric stretching of
the Si
O Si bonds, the simultaneous presence of a sharp band
at 994 cm−1, a strong and sharp band 820 cm−1 and a weak band
at 666 cm−1 agree with bibliography predictions for crystalline
MoO3 nanoparticles [50,51]. Moreover, the position of Raman band
at 994 cm−1 agrees with that reported for symmetric O Mo
O
stretching vibration of di-oxo Mo species [52]. According to such
assignments, the band at 994 cm−1 can be attributed to the stretch-
ing vibration mode of such species. The suggested description of the
structure of MoO3 is a layer structure based on the octahedral coor-
dination of Mo atoms (a crystal structure with the symmetry D2h
[52].
)
Fig. 3. Raman spectra of calcined Pt/Mo(x)-SiO2 samples and bulk MoO3.
All these bands are clearly present on the Pt/Mo5 even though
almost no peaks were detected in the Raman spectrum of the
monometallic Mo5-SiO2 sample prepared by sol-gel method [21].
The formation of ␣-MoO3 crystals for the Pt/Mo5 sample strongly
suggests the important changes in the dispersion of Mo species
brought about by the consecutive impregnation of the calcined
Mo5-SiO2 precursor with platinum salt precursor. The absence of
a significant difference in the terminal Mo-O stretching frequency
between bulk MoO3 and Pt/Mo5 sample suggests the presence of
the same types of surface molybdenum oxide species.
Contrary to the Pt/Mo5 catalyst, the two Pt/Mo10 and Pt/Mo15
ones exhibited only two Raman bands (994 cm−1 and 820 cm−1).
Taking into account the presence of the most intensive peak of
␣-MoO3 species (located at 820 cm−1), the formation of MoO3
crystallites is evident [50,51,53]. Thus, the absence of the band
at 666 cm−1 corresponding to the triply coordinated oxygen
could be explained by the re-dispersion of molybdenum species
00-001-1194) [48]. In all the patterns, a broad band in the 20◦–30◦
2ꢁ range characteristic of amorphous silica appeared.
3.4. Raman spectroscopy
The nature of platinum and molybdenum species formed after
the calcination process was also monitored by Raman spectroscopy.
Fig. 3 shows the Raman spectra of Pt/Mo(x) samples together with
bulk MoO3. With the exception of Pt/Mo12 sample, the Raman spec-
tra of the different samples show bands similar to that of bulk MoO3
at 155, 290, 340, 660, 820 and 994 cm−1. As reported by Zhen Ou
et al. [49], the peak at 155 cm−1 corresponds to the lattice mode,
while the one at 290 cm−1 represents the bending mode for the
double bond (Mo O) vibration. In addition, the peak at 341 cm−1
is assigned to Mo3 O bending mode while that at 666 cm−1 can

N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
101
Fig. 5. HRTEM images of partially reduced Pt/Mo(x)-SiO2 catalysts.
during incorporation of high amounts of platinum which, combined
to molybdenum, led to destruction of the large tridimensional
MoO3 species into smaller ones. However, none of these bands are
observed on Pt/Mo12 catalyst. Consistently with the N2 sorption
and XRD data (vide supra) this observation indicates that, during the
incorporation of Pt precursor onto Mo12-SiO2, Mo species became
redispersed leading the total rearrangement of the large poly-
molybdates structures and the formation of isolated Mo species
[54]. Finally, it is important to note that all Pt/Mo(x) catalysts did
not show bands corresponding to the platinum species suggesting
that Pt crystallite size is under the detection limits of this technique.
487 K would be attributed to the complete reduction of Pt species
having a weak interaction with the support [58]. These three peaks
appear on the reduction patterns of all bimetallic catalysts. Remark-
ably, the peak intensity at 799 K decreased sharply for Pt/Mo12 and
Pt/Mo15 catalysts and almost disappeared for Pt/Mo10 one which
fits well the findings of Raman spectroscopy. This suggests that the
incorporation of platinum would cause the rearrangement of the
molybdenum species into isolated ones having smaller sizes than
those present in bulk MoO3 phase which is reported to reduce at
higher temperatures [59]. It should be also noted the appearance of
a new peak at 886 K on the pattern of Pt/Mo12 sample. Cordero et al.
[60] attributed the appearance of more than one peak in the region
of 763 K–873 K to the presence of MoO3 clusters of different size,
whereas Jiang et al. [61] explained this in terms of the coexistence
of MoO3 polymolybdates which are easily reduced than crystalline
MoO3. In the case of Pt/Mo12 catalyst, and in agreement with XRD
and Raman results, it can be inferred that impregnation with by
Pt led to the rearrangement of large MoO3 polymolybdates into
smaller ones as judging by the appearance of the peak at 886 K.
3.5. Temperature-programmed reduction (H2-TPR)
In order to obtain information about the eventual relation
between the reducibility of the species present on the catalysts and
the activity of the partially reduced catalysts in the MCH dehydro-
genation, temperature-programmed reduction experiments were
carried for selected Pt-Mo(x) samples together with that of the
monometallic Pt/SiO2 reference. The reduction patterns of analyzed
samples are displayed in Fig. 4. It is seen that Pt/SiO2 sample shows
one small reduction peak at 347 K which is usually attributed to
the reduction of surface platinum species and another broad one at
663 K related to the reduction of platinum species strongly inter-
acting with the support [55,56].
It is generally known that the reduction pattern of molybde-
num based catalysts are rationalized in the light of the stepwise
process Mo(VI) → Mo(IV) → Mo(0) [57]. For Pt/Mo(x) catalysts and
despite the similar Pt loading, they showed different TPR profiles
(Fig. 4). In fact, for Pt/Mo5 catalyst, three peaks in addition to those
recorded for Pt/SiO2 appeared. Thus, two clear peaks are present at
799 K and 1120 K which would correspond, respectively, to the two
reduction steps of the molybdenum oxides. Meanwhile, the peak at
3.6. Characterization of partially reduced samples
3.6.1. HRTEM
HRTEM measurements were conducted with the objective to
show differences in the active phase dispersion of the catalysts
after reduction at 673 K. Fig. 5 shows TEM micrographs of partially-
reduced bimetallic catalysts. They reveal ordered crystal structures
clearly distinguishable. In fact, the interlayer lattice distance cor-
respond to the [111] diffraction planes typical of the cubic Pt
crystalline system (d = 0.226 nm; JCPDS card 00-004-0802) and to
MoO2 crystalline system (d = 0.171 nm; JCPDS card 00-001-0615).
Thus, the coexistence of platinum crystallites and MoO2 clusters
was observed in all the samples. This result could then confirm the

102
N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
molybdenum loadings results in a decrease of coordinatively unsat-
urated Mo6+ but increasing the Si
Mo bridges and Mo
O
O
terminals, which may also lead to a modification of the type of
the acid sites. For the Pt/Mo10 catalyst, the formation of weak and
medium strength acid sites would be beneficial for the reactivity in
the target reaction.
3.6.3. X-ray photoelectron spectroscopy (XPS)
Before the MCH dehydrogenation, the catalyst activation by par-
tial reduction is necessary. To obtain Mo4+ species and to avoid
total reduction of MoO3 phase to metallic Mo, the conditions of
pre-reduction were soft (673 K for 3 h). XPS experiments allowed
evaluating the part of Mo and Pt encountered on the catalysts
surface before and after the catalyst activation by reduction. The
identification of the Mo and Pt oxidation states by XPS technique
is based on the binding energies of the Mo (3d5/2, 3d3/2) spin-orbit
components and the Pt 4f7/2 core level, respectively. Table 3 lists
the binding energies (BE) corresponding to the Mo 3d5/2 and Pt 4f7/2
core levels for calcined and partially reduced catalysts. As expected,
all calcined and partially reduced samples show binding energies of
Mo3d5/2 level around 232.1 eV (Table 3) corresponding to Mo oxo-
species in the highest oxidation state (Mo6+) [65]. It is observed
that both partially reduced Pt/Mo10 and Pt/Mo12 catalysts show
binding energies of Mo 3d5/2 component shifted to higher val-
ues compared to bulk MoO3 (232.3 vs 231.7 eV) [66] suggesting a
stronger interaction between the oxo-molybdenum species and the
hydroxyl groups of the silica support and/or an electronic transfer
between Pt and Mo particles [30].
The partial reduction of all catalysts under a flow of a H2/N2
gas mixture at 673 K for 3 h led to the formation of Mo4+ species,
as shown by the appearance of a peak at a BE of 229.4 eV (Table 3),
since the same BE value was measured by Peterson et al. [67] for the
MoO2. For Pt/Mo10 catalyst, the reduction conditions were very soft
as the major portion of molybdenum species (80%) was still present
as Mo6+ species. The Mo 3d doublet of this sample (not show here)
exhibit broadening which could be explained by the coexistence
of several molybdenum species such as Mo6+, Mo4+ and CUS Mo␦+
surface site (0 < ␦ < 4) [68]. The partially reduced Pt/Mo15 catalyst
was unique showing a very well resolved Mo 3d doublet corre-
sponding to Mo6+ species. Considering the XRD results (Table 1),
this is probably due to formation of large MoO3 crystals (108.7 nm)
on the surface of this sample, which do not reduce under the soft
reduction conditions employed.
Concerning the Pt 4f7/2 core level spectra, it can be noted that
oxide precursors of Pt/Mo5, Pt/Mo10 and Pt/Mo15 catalysts exhibit
the Pt 4f7/2 line located at a binding energy of 71.2 eV indicating
the formation of metallic Pt species (Table 3). Additionally, the
two former samples exhibited a component at binding energy of
73.4 0.2 eV indicating the presence of PtOx oxide species [69].
However, after catalyst activation by reduction, all catalysts exhib-
ited only metallic platinum species.
The comparison of the surface Pt/Si and Mo/Si atomic ratios of
calcined and pre-reduced samples indicates that the concentration
of both Pt and Mo on the support surface after catalyst activation by
partial reduction decreases with respect to calcined counterparts
(see Table 3). Such a decreasecan be caused by the migrationof met-
als into bulk of material during catalyst reduction. The absence of
Raman bands of the molybdenum-containing siloxane species, such
Fig. 6. NH3-TPD profiles of partially reduced Pt/Mo(x)-SiO2 catalysts.
presence of MoO3 crystallites in the oxide precursor prior to partial
reduction of the catalysts. The medium particle size was calculated
for all bimetallic catalysts for the two types of particles. The sta-
tistical calculation of several particles indicated that the medium
particle size varies as follows:
1.) For
Pt
particles:
Pt/Mo12
(2.21 nm) < Pt/Mo10
(3.8 nm) < Pt/Mo15 (4.29 nm) < Pt/Mo5 (5.3 nm).
2.) For MoO2 particles: Pt/Mo15 (5 nm) < Pt/Mo12 (6.4 nm)
Pt/Mo10 (6.8 nm) < Pt/Mo5 (7.1 nm).
These trends indicate that platinum crystallite size is always
smaller than that of MoO2 clusters. In addition, particle sizes con-
firm the fact that the incorporation of platinum on solids with
higher molybdenum loading induced the shrinkage of molybde-
num oxide crystallites due to their re-distribution on the surface.
It is also important to note that for Pt/Mo5, Pt/Mo12 and Pt/Mo15
catalysts, Pt crystallites are not homogeneously dispersed on MoO2
while for Pt/Mo10 the number of platinum particles is higher than
that of molybdenum oxide, which allows the formation of well
dispersed active sites.
3.6.2. TPD-NH3
The acidity of partially reduced catalysts was evaluated from the
temperature-programmed desorption profiles of ammonia (Fig. 6).
As seen in Fig. 6, a remarkable difference in the acidic properties
is observed among all the samples. Remarkably, for the Pt/Mo10
catalyst, a shift of the peak maxima from 433 K to 448 K and its
large and broadened profile suggests the presence of the largest
amount of medium strength acid sites. In order to compare the total
acidity of the samples, Gaussian functions were used to deconvo-
lute the NH3 desorption profiles. The total acidity of the samples
is listed in Table 2. The total acidity of the partially reduced sam-
ples follows the trend: Pt/Mo10 > Pt/Mo15 > Pt/Mo12 ≈ Pt/Mo5. It
should be noticed that the superficial silanol groups are known to
be weak acid sites. Thus, the deposition of Mo species is expected
to give rise to a remarkable increase in the acidity. NH3-TPD results
show that the although the platinum loading is fixed for all the
catalysts, the Mo loading could be responsible not only of the dif-
ference of the amount of acid sites but also of the acid sites strength
and distribution. In fact, it has been reported that the develop-
ment of acid sites on molybdenum containing catalysts is usually
related to the presence of coordinatively unsaturated Mo6+ species
on the surface (Lewis acid sites) [62] and to the presence of hydroxyl
as a Si
O Mo O Si, indicates that Mo atoms are not embedded in
the SiO2 framework. These results suggest then that both Pt and
Mo species are mainly located on the support surface. Noticeably,
the most active Pt/Mo10 sample exhibited the most homogeneous
Pt and Mo species dispersion among the catalysts studied. For both
Pt/Mo12 and Pt/Mo15, the large Pt and Mo species surface exposure
determined by XPS is in line with HRTEM information (vide supra)
groups formed by protonating the bridging Si
O Mo or terminal
Mo O bonds on the surface (Brønsted acid sites) of the samples
[63,64]. Thus, regarding the results provided by the other charac-
terization techniques, it appears that the incorporation of higher

N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
103
Table 2
Cristal size of particlesa and amount of acid sitesb of partially reduced Pt/Mo(x)-SiO2 samples, and amount of carbon on the surface of spent catalysts.c
a
Sample
MoO2 (nm)
Pt◦a (nm)
Acid sitesb (␮moles NH3 g−1
)
C/Sic atomic ratio
Pt/Mo5
7.1
6.8
6.4
5.0
5.3
3.8
2.2
4.3
1.67
3.72
1.64
2.81
0.321
0.167
0.195
0.242
Pt/Mo10
Pt/Mo12
Pt/Mo15
a
As determined by HRTEM for partially reduced samples.
As determined by TPD-NH3 for partially reduced samples.
As determined by XPS for spent catalysts.
b
c
suggesting the formation of Pt(core)-Mo(shell) structures during
catalyst activation.
species [71]. To gain further insight into this decrease in activity at
higher Mo content, in Fig. 8(B) are compared the initial turnover
frequencies (TOF), expressed as moles of MCH converted per moles
of Pt + Mo (from elemental analysis) and per hour. As seen in this
figure, the initial TOF of the catalysts follows the same trend as the
specific reaction rates (Pt/Mo10 ꢀ Pt/Mo5 ꢀ Pt/Mo12 > Pt/Mo15)
indicating that the same factors influence on the activity of the
catalysts studied.
Finally, the XPS spectra of spent catalysts (data not reported
here) confirmed the presence of Pt◦, MoO2 (BE at 228.8 eV) and
MoO3 (BE in range 232.1–232.3 eV) phases only. Thus, under the
reaction conditions employed (T = 675 K PH2 = 2.2 MPa), no Mo car-
bide phase is formed in the course of MCH dehydrogenation over
Pt/Mo(x)-SiO2 catalysts. Contrary to our work, Mo carbide forma-
tion was reported during the natural gas conversion over MoO3
supported on zeolites [51].
For all catalysts, the GC–MS analysis of reaction products
revealed the presence of toluene as the main product with the
appearance of traces of substituted cyclohexanes and cyclopen-
tanes. The yields of toluene and isomerization products at TOS = 0
are given in Fig. 8(C). From this figure, it is evident that the
Pt/Mo10 catalyst shows the largest yields of toluene and iso-
merization products among the catalysts studied being yields of
toluene 8-times larger than those of the isomerization products.
The reaction scheme is shown in Fig. 9. Considering the products
detected, the transformation of MCH on silica-supported Pt/Mo(x)
catalysts occurs through a bifunctional mechanism. It is known
that the isomerization products are formed on acidic sites while
toluene is formed on metallic sites of the catalysts after direct
dehydrogenation of MCH molecules [4,30]. In fact, the results
obtained show clear evidence that the MCH dehydrogenation reac-
tion over Pt/Mo(x)-SiO2 catalysts occurs mainly on MoO2(Hx) and
Pt◦ sites [16,21]. A linear correlation between the initial dehy-
drogenation activity and the surface Pt◦/Mo4+ atomic ratio was
confirmed (Fig. 10), suggesting that the activity is proportional to
the number of surface Pt and Mo4+ species. In this sense, for binary
Pt-Mo catalysts, it has been suggested that dehydrogenation of
MCH has a rate-determining step occurring preferably on Pt sites
[53]; unlikely to hydrogenolysis, which was demonstrated to occur
much faster on Pt-Mo mixed sites known to be the preferential
adsorption sites. The experimental data of the MCH dehydrogena-
tion over Pt catalysts was found to be best fitted using single-site
Langmuir-Hinshelwood-Hougen-Watson (LHHW) kinetic mecha-
nism [10,12], which consider the loss of the first hydrogen molecule
as the rate determining step and the surface coverage by the
methylcyclohexadiene intermediate as negligible.
3.7. Catalytic performance in MCH dehydrogenation
Pt/Mo(x)-SiO2 catalysts were evaluated in the reaction of
MCH dehydrogenation at 673 K and under 2.2 MPa of total H2
pressure. Before reaction, the catalysts were activated by reduc-
tion with gas mixture H2:N2 (1:5) at 673 for 3 h. Fig. 8(A)
shows the rates of MCH dehydrogenation with time on stream
(TOS) at selected reaction conditions. Regardless of TOS, the
dehydrogenation activity of the catalysts follows the trend:
Pt/Mo10 > Pt/Mo5 > Pt/Mo12 ≈ Pt/Mo15. Thus, despite the higher
Mo loading in the Pt/Mo12 and Pt/Mo15 catalysts, the Pt/Mo10
showed the best catalytic response at the reaction conditions stud-
ied. As expected, all binary Pt/Mo(x) catalysts were more active
than their monometallic Mo(x)-SiO2 counterparts [21]. Moreover,
considering the effect of Mo loading, the activity of binary samples
follows the same trend as those of monometallic Mo(x)-SiO2 ones
[21]. All the catalysts exhibited a good stability during time course
of reaction. This is probably because MCH dehydrogenation was
carried out at relatively high H2 pressure of 2.2 MPa, which might
diminish the catalyst deactivation by the carbonaceous deposits,
known to mainly occur on the acid sites of the catalysts [70]. Notice-
ably, despite of the largest acidity (Table 2), the spent Pt/Mo10
catalyst exhibited the lowest carbon deposition among the cata-
lysts studied, as deduced from its lowest C/Si atomic ratio (Table 2).
Thus, from the information offered in Fig. 7, the lowest coke forma-
tion on the Pt/Mo10 catalyst could be tentatively explained by the
enhancement of hydrogen dissociation on the well dispersed active
sites. Additionally, the dilution effect of MoO2 might prevent exces-
sive deactivation of the Pt sites of this catalyst. This is because a
polymerization type-reaction occurs mainly on continuous Pt sites
(n > 3) which are known to be extensively dehydrogenated surface
For all catalysts, a linear correlation between the total amount
of acid sites of the samples and the catalyst initial activity was
found (Fig. 10) confirming the participation of acid sites in the
reaction mechanism. However, the much higher selectivity toward
Table 3
Binding energies (eV) of core levels and surface metal atomic ratios of calcined, partially reduced and spent Pt/Mo(x)-SiO2 samples.
a
a
Catalyst
Pt/Mo5
Pretreatment
Mo 3d5/2
232.3
Pt 4f7/2
Mo/Si at
Pt/Si at
Pt◦/Mo4+at
Calcined
Partially reduced
Calcined
Partially reduced
Calcined
Partially reduced
Calcined
71.2 (55) 73.6 (45)
0.306
0.075
0.202
0.116
0.293
0.183
0.747
0.227
0.276
0.086
0.244
0.088
0.276
0.172
0.648
0.412
–
229.3 (62) 231.1 (38)
232.1
229.4 (20) 232.3 (80)
232.3
229.4 (73) 232.3 (27)
232.1
229.4 (77) 231.2 (23)
71.1
1.849
–
3.793
–
1.287
–
2.357
Pt/Mo10
Pt/Mo12
Pt/Mo15
71.2 (46) 73.4 (54)
71.2
73.1
71.1
71.2
71.1
Partially reduced
a
Between parenthesis peak percentages.

104
N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
Fig. 7. Correlation between XPS Mo/Si and Pt/Si atomic ratios of partially reduced samples and bulk Mo/Si and Pt/Si atomic ratios of calcined samples (from ICP-OES).
dehydrogenation product (toluene) than toward cracking and iso-
merization products indicates that the metal function dominates
over the acidic one. Since a correlation between the Pt/Mo4+ atomic
surface ratio and the initial activity was also found (Fig. 10), it
is clear that all catalysts exhibit bifunctional (metallic + acidic)
character. In other words, it could be inferred that the observed
differences in the initial activity of Pt/Mo(x) catalysts may be asso-
ciated to the different number of initial Pt and MoO2 surface species
among the samples as well as with their acidity. Considering the
metallic function, it is also suggested that platinum addition to
Mo(x)-SiO2 catalysts might enhance the metallic character of the
MoO2−x(OH)y bifunctional system [72], which would facilitate the
loss of first H-atom from MCH. However, the structure sensitivity
of MCH dehydrogenation at high H2 pressure should be different
because the difference in the concentrations of the adsorbed car-
bonaceous deposits generated under different H2 pressures [73].
On the other hand, both XPS and HRTEM data point out to a
strong dependence of the catalyst initial activity on the dispersion
of Pt and Mo on the support surface. Thus, the best catalytic activity
of Pt/Mo10 catalyst can be explained considering the most homo-
geneous exposure of their Pt and Mo species on the support surface.
In fact, the higher Pt and Mo dispersion can increase the population
of the kink and step active sites [39]. This finding is in agreement
with Alhumaidan et al. [11], who reported that the increasing in
Pt loading in Pt/Al2O3 catalysts negatively influenced Pt disper-
sion and morphology leading to a decrease in the dehydrogenation

N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
105
Fig. 10. Influence of the acidity (from TPD-NH3) and the surface exposure of active
phases (expressed as Pt◦/Mo4+ atomic ratio) of the partially reduced samples on the
initial catalyst activity.
to lowering the dehydrogenation activity (when x > 10). Hence, the
best initial activity of the Pt/Mo10 catalyst could be associated with
the specific morphology of the MoO2 and Pt clusters on the silica
surface. In this sense, density functional theory calculations (DFT)
reported by Gao et al. [74,75] suggested that two adjacent Pt atoms
are required for adsorbing an acetylene (H
C C H) molecule and a
third neighboring Pt one for stabilizing the reacting H-atom during
the transformation. However, concerning the geometric require-
ments needed for MCH dehydrogenation over Pt/Mo-SiO2 catalysts,
our experimental results do not allow us to accurately determine
the type of active sites working upon reaction conditions employed.
Further reaction kinetic studies and/or DFT calculations should be
needed to predict possible geometric requirements for MCH trans-
formation on the clean Pt and MoO2 surfaces.
4. Conclusions
In this work, the silica-supported partially reduced bimetallic
Pt/Mo catalysts were tested in the dehydrogenation reac-
tion of MCH, in the context of the hydrogen storage using
methylcyclohexane-toluene-hydrogen cycle. It was concluded that
the reaction occurs mainly on MoO2(Hx) and Pt◦ sites. The best
activity and selectivity toward toluene was obtained on the
Pt/Mo10 catalyst showing optimized Mo × 100/Si molar ratio of 10.
This was linked to its bifunctional character, its homogeneous dis-
tribution of Pt◦ and Mo(VI) on the support surface, and the lowest
carbon deposition during time course of the reaction preventing
the deactivation on the catalytic reaction. However, larger Mo load-
ings (x = 12 and 15) induced the formation of MoOx-Pt core-shell
nanoparticles which caused a drop in the activity of these catalysts.
Fig. 8. Variation in the activities of partially reduced Pt/Mo(x)-SiO2 catalysts with
time on stream (A), initial TOF (B), and initial yields of products (C) for MCH dehy-
drogenation reaction (T = 673 K, PH 2 = 2.2 MPa). TOF calculated as moles of MCH
converted per mole of (Pt + Mo) per second.
Acknowledgements
Thanks to the Tunisian Ministry of Higher Education, Scientific
Research and Information and Communication Technologies for
financial support. Financial support by the Community of Madrid
(Project S213/MAE-2882) is kindly acknowledged.
Fig. 9. Reaction scheme of the MCH dehydrogenation over partially reduced
Pt/Mo(x)-SiO2 catalysts.
activity. On the contrary to Pt/Mo10, the XPS and HRTEM results
suggest that the very low activities of both Pt/Mo12 and Pt/Mo15
catalysts can be caused by the formation of core-shell particles hav-
ing Pt covered by Mo species (Fig. 5). This fits well the results of
Leclercq et al. [23], who reported that the presence of Pt particles
on the catalyst surface may be hindered by molybdenum species
located on the catalyst surface. Thus, an increase in Mo loading neg-
atively influenced Pt and Mo dispersion and morphology leading
References
[1] A. Pradhan, A. Shukla, J. Pande, S. Karmarkar, R. Biniwale, Int. J. Hydrogen
Energy 36 (1) (2011) 680–688.
[2] M. DeLuchi, Int. J. Hydrogen Energy 14 (1989) 81–130.
[3] F. Alhumaidan, D. Cresswell, A. Garforth, Energy Fuel 25 (2011) 4217–4234.
[4] F. Alhumaidan, D. Tsakiris, D. Cresswell, A. Garforth, Int. J. Hydrogen Energy
38 (2013) 14010–14026.
[5] J.H. Sinfelt, J. Mol. Catal. A: Chem. 163 (2000) 123–128.

106
N. Boufaden et al. / Journal of Molecular Catalysis A: Chemical 420 (2016) 96–106
[6] S. Tan, L. Briones Gil, N. Subramanian, D.S. Sholl, S. Nair, Ch.W. Jones, J.S.
[39] S.M. Davis, G.A. Samorjai, J. Catal. 65 (1980) 78–83.
[40] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[41] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.
[42] A. Bail, V.C. dos Santos, M. Roque de Freitas, L. Pereira Ramos, W. Herwig
Schreiner, G. Pimenta Ricci, K.J. Ciuffi, S. Nakagak, Appl. Catal. B: Environ.
130–131 (2013) 314–324.
Moore, Y. Liu, R.S. Dixit, J.G. Pendergast, Appl. Catal. A: Gen. 498 (2015)
167–175.
[7] S. Tan, S.-J. Kim, J.S. Moore, Y. Liu, R.s. Dixit, J.G. Pendergast, D.S. Sholl, S. Nair,
Ch.W. Jones, ChemCatChem 8 (2016) 214–221.
[8] J.H. Sinfelt, Adv. Chem. Eng. 5 (1964) 37–54.
[9] M. El-Sawi, F. Infortuna, P. Lignola, A. Parmaliana, F. Frusteri, N. Giordano,
Chem. Eng. J. 42 (3) (1989) 137–144.
[10] M. Usman, D. Cresswell, A. Garforth, Ind. Eng. Chem. Res. 51 (2012) 158–170.
[11] F. Alhumaidan, D. Cresswell, A. Garforth, Ind. Eng. Chem. Res. 49 (2010)
9764–9770.
[43] A.S. Alfaya, L.T. Kubota, Quim. Nova 25 (2002) 835–841.
[44] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
[45] G. Leofanti, M. Padovan, G. Tozzola, B. Venurelli, Catal. Today 41 (1998)
207–219.
[12] F. Alhumaidan, D. Cresswell, A. Garforth, Ind. Eng. Chem. Res. 50 (2011)
2509–2522.
[13] A. Rochefort, F. Le Peltier, J. Boitiaux, J. Catal. 138 (1992) 482–490.
[14] W. Doolittle, N. Skoularikis, R. Coughlin, J. Catal. 107 (2) (1987) 490–502.
[15] D.K. Cromwell, P.T. Vasudevan, B. Pawelec, J.L.G. Fierro, Catal. Today 259
(2015) 119–129.
[46] Z. Paál, T. Koltai, K. Matusek, J.M. Manoli, C. Potvin, M. Muhler, U. Wild, P.
Tetenyi, Phys. Chem. Chem. Phys. 3 (2001) 1535–1543.
[47] D. Gulkova, Y. Yoshimura, Z. Vit, Appl. Catal. B: Environ. 87 (2009) 171–180.
[48] B. Gurau, R. Viswanathan, R.X. Liu, T.J. Lafrenz, K.L. Ley, E.S. Smotkin, E.
Reddington, A. Sapienza, B.C. Chan, T.E. Mallouk, S.J. Sarangapani, Phys. Chem.
B 102 (1998) 9997–10003.
[16] A. Shukla, J.V. Pande, R.B. Biniwale, Int. J. Hydrogen Energy 37 (2012)
3350–3357.
[49] L. Zhen Ou, J.L. Cambell, D. Yao, W. Wlodarski, K. Kalantar-zadeh, J. Phys.
Chem. C 115 (2011) 10757–10763.
[17] A.A. Shukla, P.V. Gosavi, J.V. Pande, V.P. Kumar, K.V.R. Chary, R.B. Biniwale, Int.
J. Hydrogen Energy 35 (2010) 4020–4026.
[18] R. Coughlin, K. Kawakami, A. Hasan, J. Catal. 88 (1984) 150–162.
[19] R.B. Biniwale, N. Kariva, M. Ichikawa, Catal. Lett. 105 (2005) 83–87.
[20] S.P. Patil, J.V. Pande, R.B. Biniwale, Int. J. Hydrogen Energy 38 (2013)
15233–15241.
[50] E.L. Lee, I.E. Wachs, J. Phys. Chem. C 111 (2007) 1410–14425.
[51] J. Gao, Y. Zheng, J.-M. Jehng, I.E. Wachs, S.G. Podkolzin, Science 348 (6235)
(2015) 686–690.
[52] S. Chempath, Y. Zhang, A.T. Bell, J. Phys. Chem. C 111 (2007) 1291–1298.
[53] Y. Miao, G. Lu, X. Liu, Y. Guo, Y. Wang, Y. Guo, J. Mol. Catal. A: Chem. 306
(2009) 17–22.
[21] N. Boufaden, R. Akkari, B. Pawelec, J.L.G. Fierro, M. Zina, A. Ghorbel, Appl.
Catal. A: Gen. 502 (2015) 329–339.
[54] R.D. Roark, S.D. Kohler, J.G. Ekerdt, D.S. Kim, I.E. Wachs, Catal. Lett. 16 (1992)
77–83.
[22] H. Belatel, H. Al-Kandari, F. Al-Kharafi, F. Garin, A. Katrib, Appl. Catal. A: Gen.
318 (2007) 227–233.
[23] G. Leclercq, S. Pietrzyk, T. Romero, A. El Gharbi, L. Gengembre, J. Grimblot, F.
Aissi, M. Guelton, A. Latef, L. Leclercq, Ind. Eng. Chem. Res. 36 (1997)
4015–4027.
[55] T. Ren-Yuan, W. Rong-An, L. Li-Wu, Appl. Catal. 10 (1984) 163–172.
[56] H. Lieske, G. Lietz, H. Spindler, J. Volter, J. Catal. 81 (1983) 8–16.
[57] A. Galadima, B.M. Ibrahim, Afr. Sci. 11 (2010) 115–121.
[58] C. Moran, E. Gonzalez, J. Sanchez, R. Solano, G. Carruyo, A. Moronta, J. Colloids
Interface Sci. 315 (2007) 164–169.
[24] G. Leclercq, A. El Gharbi, L. Gengembre, T. Romero, L. Leclercq, S. Pietrzyk, J.
Catal. 148 (1994) 550–561.
[25] G. Leclercq, T. Romero, S. Pietrzyk, J. Grimblot, L. Leclercq, J. Mol. Catal. 25
(1984) 67–86.
[26] L.I. Ali, A.-G.A. Ali, S.M. Aboul-Fotouh, A.K. Aboul-Gheit, Appl. Catal. A: Gen.
177 (1999) 99–100.
[27] T.M. Tri, J. Massardier, P. Gallezot, B. Imelik, J. Catal. 85 (1984) 244–252.
[28] N. Kariya, A. Fukuoka, M. Ichikawa, Appl. Catal. A: Gen. 233 (2002) 91–102.
[29] N. Kariya, A. Fukuoka, T. Utagawa, M. Sakuramoto, Y. Goto, M. Ichikawa, Appl.
Catal. A: Gen. 247 (2003) 247–259.
[59] J.R. Regalbuto, J.W. Ha, Catal. Lett. 29 (1994) 189–207.
[60] R.L. Cordero, F.J.G. Llambias, A.L. Agudo, Appl. Catal. 74 (1991) 125–136.
[61] H. Jiang, L. Wang, W. Cui, Y. Xu, Catal. Lett. 57 (1999) 95–102.
[62] T. Kataoka, J.A. Dumesic, J. Catal. 112 (1988) 66–79.
[63] L. Vordonis, P.G. Koutsoukos, A. Lycourghiotis, J. Catal. 98 (1986) 296–307.
[64] L. Vordonis, P.G. Koutsoukos, A. Lycourghiotis, Colloids Surf. 50 (1990)
353–361.
[65] Y.V. Plyuto, I.V. Babich, I.V. Plyuto, A.D. Van Langeveld, J.A. Moulijn, Appl. Surf.
Sci. 119 (1997) 11–18.
[66] S.O. Grim, L.J. Matienzo, Inorg. Chem. 14 (1975) 1014–1018.
[67] T.A. Paterson, J.C. Carver, D.E. Leyden, D.M. Hercules, J. Phys. 80 (1976)
1700–1708.
[30] K. Jaroszewska, A. Masalska, D. Marek, J.R. Grzechowiak, A. Zemska, Catal.
Today 223 (2014) 76–86.
[31] C. Petit, A. Katrib, P. Girard, F. Garin, G. Maire, J. Mol. Catal. 85 (1993) 75–95.
[32] G.C. Bond, Fundamental and Applied Catalysis (Series), Dehydrogenation of
Alkanes, Springer, US, 2005, pp. 501–523.
[33] T. Matsuda, A. Hanai, F. Uchijima, H. Sakagami, N. Takahashi, Microporous
Mesoporous Mater. 51 (2002) 155–164.
[68] D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, J. Catal. 207 (2002) 266–273.
[69] K.S. Kim, N. Winograd, R.E. Davis, J. Am. Chem. Soc. 93 (1971) 6296–6297.
[70] Y. Okada, E. Sasaki, E. Watanabe, S. Hyodo, H. Nishijima, Int. J. Hydrogen
Energy 31 (2006) 1348–1356.
[71] J.K.A. Clarke, Chem. Rev. 75 (1975) 291–305.
[34] H. Belatel, H. Al-Kandari, F. Al-Khorafi, A. Katrib, F. Garin, Appl. Catal. A: Gen.
275 (2004) 141–147.
[72] H. Al-Kandari, A.M. Mohamed, F. Al-Kharafi, M.I. Zaki, A. Katrib, Appl. Catal. A:
Gen. 417–418 (2012) 298–305.
[35] Yu. I. Yermakov, B.N. Kuznetsov, Yu A. Ryndin, J. Catal. 42 (1976) 73–78.
[36] H. Al-Kandari, A.M. Mohamed, F. Al-Kharafi, A. Katrib, Catal. Commun. 12
(2011) 1188–1192.
[37] R. Navarro, B. Pawelec, J.L.G. Fierro, P.T. Vasudevan, Appl. Catal. A: Gen. 148
(1996) 23–40.
[73] I.L.K. Herz, W.D. Gillespie, E.E. Petersen, G.A. Somorjai, J. Catal. 67 (1981)
371–386.
[74] J. Gao, H. Zhao, X. Yang, B.E. Koel, S.G. Podkolzin, Angew. Chem. Int. Ed. 53 (9)
(2014) 3641–3644.
[75] J. Gao, H. Zhao, X. Yang, B.E. Koel, S.G. Podkolzin, ACS Catal. 3 (6) (2013)
1149–1153.
[38] G. MuraliDhar, F.E. Massoth, J. Shabtai, J. Catal. 85 (1984) 44–52.