Potential of templated mesoporous aluminas as supports for HDS CoMo catalysts

Article abstract of DOI:10.1039/c5nj02865e

Aluminas, exhibiting different textural properties, were prepared using a template assisted sol-gel method. In the presence of a triblock copolymer as a porogen, and owing to the control of the synthesis parameters, alumina textural properties can be adjusted up to values higher than those of conventional alumina. Among obtained supports, some were selected in order to cover surface area values from 350 to 450 m² g^-1, and pore diameters from 5.4 to 9.0 nm. CoMo derived catalysts, prepared by impregnation with a (Co₂Mo₁₀O₃₈H₄)Co₃ solution, with a concentration corresponding to a molybdenum density of 4 atMo nm^-2 (or 3 when not possible to achieve 4), exhibit decreased textural parameters. However, despite the changes induced by the preparation step, the final catalysts always display attractive textural properties. Activities in thiophene hydrodesulfurization (HDS) can be increased by a factor three, depending on the textural properties of the supports used and possibly to some extend on the residual carbon content in the material. The performance of the most efficient solid in the HDS of thiophene was evaluated in the HDS of a straight run gas oil (SRGO) and was found higher than that of a commercial alumina based catalyst and close to that of a commercial catalyst modified by a chelating agent, evaluated under the same conditions.

Full text of DOI:10.1039/c5nj02865e

NJC
View Article Online
View Journal
PAPER
Potential of templated mesoporous aluminas as
supports for HDS CoMo catalysts
Cite this:DOI: 10.1039/c5nj02865e
Andres Mino,a Christine Lancelot,†*a Pascal Blanchard,a Carole Lamonier,a
˜
Loıc Rouleau,b Magalie Roy-Auberger,b Sebastien Royer*ac and Edmond Payena
´
¨
Aluminas, exhibiting different textural properties, were prepared using a template assisted sol–gel
method. In the presence of a triblock copolymer as a porogen, and owing to the control of the synthesis
parameters, alumina textural properties can be adjusted up to values higher than those of conventional
alumina. Among obtained supports, some were selected in order to cover surface area values from 350
to 450 m2 ꢀ1, and pore diameters from 5.4 to 9.0 nm. CoMo derived catalysts, prepared by impregna-
g
tion with a (Co2Mo10O38H4)Co3 solution, with a concentration corresponding to a molybdenum density
of 4 atMo nmꢀ2 (or 3 when not possible to achieve 4), exhibit decreased textural parameters. However,
despite the changes induced by the preparation step, the final catalysts always display attractive textural
properties. Activities in thiophene hydrodesulfurization (HDS) can be increased by a factor three,
depending on the textural properties of the supports used and possibly to some extend on the residual
carbon content in the material. The performance of the most efficient solid in the HDS of thiophene
was evaluated in the HDS of a straight run gas oil (SRGO) and was found higher than that of a commer-
cial alumina based catalyst and close to that of a commercial catalyst modified by a chelating agent,
evaluated under the same conditions.
Received (in Nottingham, UK)
15th October 2015,
Accepted 20th January 2016
DOI: 10.1039/c5nj02865e
www.rsc.org/njc
large pore volume allows the impregnation of high metal
1. Introduction
loadings.6,7 In addition, large pore diameters (Dp) ensure an
The constant evolution of sulfur content regulations in trans- efficient diffusion of the reactants to the active sites, an issue of
portation fuels, toward lower concentrations, imposes a drastic particular importance for the treatment of heavy feedstocks.8
improvement of the hydrodesulfurization (HDS) of petroleum
In this context, the development of alumina supports with
feedstock which can be achieved through the development improved textural properties is of prime interest. In line with the
of more efficient catalytic formulations.1 Conventional HDS first reports on the synthesis of mesostructured silica,9 many
catalysts are composed of nanometric MoS2 crystallites, promoted templating strategies were proposed for alumina.10–13 The first
by cobalt atoms, dispersed on alumina (CoMo-S phase).2–4 This preparation of templated alumina, in 1996, was proposed using
active phase is obtained by the sulfidation of an oxidic precursor a non-ionic template assisted sol–gel approach.10 Using a
generally prepared by incipient wetness impregnation of the Tergicol-type template, alumina with a surface area higher than
precursors (e.g. cobalt nitrate and ammonium heptamolybdate) 500 m2 gꢀ1 can be produced with the pore size in the mesopore
on a porous support.5 Textural parameters of the support, i.e. domain. At the same time, Vaudry et al. proposed the synthesis
specific surface area (SSA) and pore volume (Vp), are parameters of alumina in the presence of carboxylic acids.11 Parameters
conditioning dispersion of the active precursors. A high surface such as the amount of water, the type of solvent, the calcination
area is needed to achieve high active phase dispersion while a temperature and the nature of the structuring agent (from differ-
ent carboxylic acids) significantly impact the textural properties
of the final material. After calcination at moderate temperature,
a
´
Univ. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UMR 8181-UCCS-Unite de
e.g. 400 1C, alumina exhibiting a surface area up to 700 m2 gꢀ1
with narrow pore size distribution in the low mesopore domain
size (B2 nm) was obtained.
Catalyse et Chimie du Solide, F-59000 Lille, France.
E-mail: christine.lancelot@univ-lille1.fr
b IFP Energies Nouvelles, Rond-Point de l’Echangeur de Solaize, BP3, 69360 Solaize,
France
Until now, non-ionic templating agents have been mostly
used in order to produce porous aluminas exhibiting improved
surface areas and relatively large pores (Table 1). Prepared
using a template assisted sol–gel synthesis route, alumina with
surface areas higher than 400 m2 gꢀ1 can be easily obtained
c
´
Universite de Poitiers, CNRS IC2MP UMR 7285, 4 Rue Michel Brunet, 86022
Poitiers Cedex, France. E-mail: sebastien.royer@univ-poitiers.fr
´
´
† Unite de Catalyse et de Chimie de Solide, UMR CNRS 8181, Universite
ˆ
Lille 1, batiment C3, 59655 Villeneuve d’Ascq, France. Fax: (+33) 3 20 43 65 61;
Tel: (+33) 3 20 43 49 50
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
after calcination at a temperature compatible with the hydro- in the presence of P123. For their selected application, that
treatment process (that is between 400 and 600 1C). Even at are HDS and HDN (hydrodenitrogenation) of heavy gas oil,
such high surfaces, the mesopore size could be stabilized, from they related the improved performance measured for these
2–3 nm (small mesopores) to 10 nm and higher (large mesopores materials to a higher reducibility of molybdenum when supported
adequate for catalytic treatment of heavy charges). Ordered alumina over these materials. The limited amounts of the available litera-
is however rarely obtained, except using an Evaporation Induced ture however show the real potential of mesostructured alumina
Self-Assembly [EISA] process14–16 and most of the materials as support for HDS catalysts, especially when the petroleum
have vermicular porous structures, demonstrating a lower level cuts to be treated correspond to the medium to high distillation
of organization compared to that obtained with silica.
While mesostructured silica-based materials were extensively
temperature ranges.
In this work, starting from the approach proposed by Zhang
studied as HDS catalyst supports,22–32 the literature concerning and Pinnavaia17 consisting of a copolymer triblock templating
the use of non-siliceous mesostructured oxides for HDS reactions sol–gel derived synthesis, aluminas having different textural
is in contrast relatively limited. A high capacity for molybdenum properties were prepared. The study of the synthesis parameters
33
ˇ
dispersion was however reported by Cejka et al. and Kaluˇza allowed us to adjust the final textural properties of the supports,
et al.34 over these mesoporous aluminas. Adequately dispersed which were further used for the dispersion of cobalt and
molybdenum species were identified on these supports, even at molybdenum precursors. Catalysts (and supports when relevant)
MoO3 loading as high as 30 wt%, while the formation of bulk were characterized to evaluate the impact of the support char-
phases is evidenced from 16 wt% MoO3 loading over classical acteristics on the thiophene HDS performance of the final CoMo
aluminas.35 On mesoporous supported CoMo-S catalysts, Bejenaru based materials.
et al.36 showed that the material HDS activity (as evaluated for the
thiophene desulfurization) varies over a wide range, depending
on the initial support morphology and its stability upon the
impregnation–calcination–sulfidation steps. The best performances
were obtained using alumina support presenting a fine fibrillary
2. Experimental
2.1. Alumina synthesis
type morphology with large mesopores. HDS performances of Alumina preparation presented in this study consists of a
the catalytic materials also depend on the nature of the molyb- template assisted sol–gel synthesis, as initially proposed by
date species present in the oxidic form on the support surface. Zhang and Pinnavaia (Table 1).17 Chemicals for the synthesis
Low activity was paralleled to the formation of oxidic mono- are used as received: Pluronic P123 (Aldrich, poly(ethylene
molybdate species, whereas polymolybdate species, identified glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol),
before sulfidation, ensured higher activity for the final materials. EO20PO70EO20), aluminum-sec-butoxide (Aldrich, 97%) as the
Badoga et al.37 prepared NiMo catalysts supported on mesoporous aluminum source, and 1-butanol (Aldrich, 99.5%+) as the
alumina obtained by the hydrolysis of aluminum isopropoxide solvent. The molar composition of the initial solution is
Table 1 Selected reference studies for the production of high surface area mesoporous alumina supports
Thermal
treatment
Synthesis approach
Main characteristics of final materials
Zhang and
Pinnavaia17
MSU-X aluminas prepared using a template (Tergicol From 500 1C Amorphous materials, with SSA from 350 to
and Pluronic) derived sol–gel route, with the insertion to 600 1C
530 m2
g
ꢀ1, Dp from 5.0 to 10.8 nm and Vp from
of Ce/La for thermal stability improvement
0.3 to 1.3 cm3 gꢀ1 – wormhole-like porosity
Pinnavaia et al.18 Review the different approach for the preparation of amorphous/crystalline mesoporous alumina, using various surfactants.
Propose the preparation of mesoporous g-type alumina, from the boehmite precursor. A crystalline AlOOH precursor can be
obtained by the hydrothermal treatment of the amorphous precursor in the presence of the template.
19
ˇ
Cejka et al.
Preparation of mesoporous alumina, using stearic
From 420 1C Amorphous–crystalline materials (depends on
acid as a surfactant, with intermediate hydrothermal to 1000 1C
treatment being applied. Thermal treatment at
intermediate to high temperatures to evaluate textural
stabilities
calcination T), with SSA from 480 to 140 m2 gꢀ1
Dp from 3.4 to 6.0 nm and Vp from 0.53 to
0.28 cm3 gꢀ1 – wormhole-like porosity
,
Niesz et al.20
Evaporation induced self-assembly of a hybrid pre-
cursor, starting from an organic aluminum precursor
in HCl acidified ethanol solution and the Pluronic type
template. Calcination at intermediate temperature to
ensure the combustion of the copolymer template
400 1C
540 1C
SSA from 206 to 410 m2 gꢀ1, Dp from 3.9 to 6.7 nm and
Vp from 0.44 to 0.80 cm3 gꢀ1 – from the hexagonal-type
pore structure (low H2O content) to wormhole-like
porosity (larger H2O content)
Caragheorgheopol Two approaches for the preparation of Pluronic and
SSA from 270 to 450 m2
g
ꢀ1, Dp from 4.3 to 8.9 nm
et al.21
Tergicol templated aluminas: (i) precipitation by
ammonia of an organic aluminum precursor and
(ii) hydrolysis of an organic aluminum precursor
in anhydrous alcohol
and Vp from 0.53 to 1.36 cm3 gꢀ1
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
1 Al:xP123:15ButOH:yH2O (where x = 0, 0.2, 0.3, 0.6 mol, and X-ray source. The binding energies were referred to the Al2p
y = 2.9, 4.5, 5.9). An adequate mass of P123 is initially dissolved photopeak of the support at 74.6 eV. The surface atomic ratios
in 1-butanol at a temperature of 35 1C under mechanical were calculated using the CASA XPS software after subtracting
stirring at 400 rpm. After complete dissolution of the template, the nonlinear Shirley background.
the hydrolysis temperature is adjusted to z 1C (z = 15, 25, 50 1C).
Dispersion of the oxomolybdate phase over the support was
The aluminum precursor is then added dropwise, and the evaluated using Raman spectroscopy. Spectra were recorded at
mixture is stirred at 400 rpm for 1 h. The hydrolysis step is room temperature using a Raman microprobe Infinity from
carried out by the dropwise addition of a butanol/water Jobin-Yvon, equipped with a N2 cooled CCD detector. The exciting
solution (VButOH/VH O = 1) at the desired z 1C temperature. laser source was the 532 nm line of a Nd-YAG laser.
2
The solution is aged under stirring for 4 h. The gel is thereafter
2.4. Thiophene catalytic test
transferred into a Teflon-lined autoclave and thermally treated
at w 1C (w = 15, 45, 65, 85 1C) for 40 h under static conditions.
After thermal treatment, the product is recovered by filtration,
washed with butanol, and then dried at 60 1C for 24 h. Finally,
the dried solid is calcined under air at 600 1C for 4 h (temperature
increase rate = 1 1C minꢀ1), a temperature sufficient to ensure the
removal of organic matter from the solid. The nomenclature of
the alumina support is as follows: Al–Sx–Wy–Thz–HTw, where Sx
refers to the surfactant content, Wy to the water content, Thz to
the temperature of hydrolysis and HTw to the hydrothermal
treatment temperature.
HDS of thiophene is commonly admitted as a reference HDS
catalytic test, as thiophene is the simplest molecule represen-
tative of the aromatic sulfur compounds present in the feeds to
be treated.39 Also, HDS of thiophene was selected since this
reaction is known to efficiently probe the dispersion of the active
phase of HDS catalysts, and will thus allow a first comparison
between the solids. The tests were performed at atmospheric
pressure in a flow-type reactor packed with 0.2 g of catalyst.
Before reaction, the oxidic catalyst was sulfided at atmospheric
pressure under a flow of 10 vol% of H2S in H2 (100 mL minꢀ1), at
400 1C for 2 h (temperature increase rate of 6 1C minꢀ1). The
reactor was then naturally cooled down to the reaction tempera-
ture of 300 1C under sulfidation flow. Thiophene, previously
purified by two successive vacuum distillations, was introduced
in the reactor at a constant pressure of 6.65 kPa in a flow
of hydrogen (total flow rate = 10 mL minꢀ1). Reactants and
products (butane, but-1-ene, trans-but-2-ene, and cis-but-2-ene)
were analyzed using a gas chromatograph equipped with a flame
ionization detector and a Plot-alumina column. Taking into account
the relative response factor for each compound, the conversion
a was calculated using eqn (1).
2.2. Catalyst preparation
Supported CoMo oxidic precursors were prepared by inci-
pient wetness impregnation, with an aqueous solution of
Co2Mo10O38H4Co3 salt synthetized according to ref. 38, which
will be named Co2Mo10 hereafter. Concentration is calculated
taking into account a final molybdenum density of 4 atMo nmꢀ2
on the final materials, except otherwise specified. The Co/Mo
molar ratio, imposed by the stoichiometry of the heteropoly-
compound, is 0.5. Impregnation is performed by the dropwise
addition of the solution, followed by 2 h of maturation under a
wet atmosphere in order to ensure a complete filling of the pores
by the impregnation solution. Thereafter, the solids are dried
overnight at 100 1C, before being calcined at 500 1C under oxygen.
Catalysts are named CoMox/Al-y, with x being the molybdenum
content in wt% of MoO3, and y the selected alumina.
4
P
ai
4
i¼1
a ¼
(1)
4
P
athio
ai
þ
3:4 i¼1 4
2.3. Physical characterization
where ai is the peak area corresponding to the i product and
athio is the area of the peak of thiophene.
Values of conversion are reported after four hours of catalyst
activity stabilization.
N2 adsorption–desorption isotherms were recorded at –196 1C
using an automated ASAP2010 instrument from MICROMERITICS.
Before each run, a known mass of the sample (around 0.200 g)
was heated at 350 1C under vacuum for 3 h. Specific surface
areas were calculated from the linear part of the Brunauer–
2.5. Straight run gas oil (SRGO) HDS activity
Emmett–Teller line. Pore size distributions were obtained by The efficiency of the catalysts towards more refractory molecules
applying the Barrett–Joyner–Halenda (BJH) equation to the was evaluated through the HDS of a gas oil, a real feedstock,
desorption branch of the isotherm. The total pore volume instead of using model molecules such as those classically used
was estimated from the N2 uptake at a P/P0 value corresponding (like HDS of DBT). Two catalysts were tested using a high
to the top of the condensation capillary step: 0.99 for solids pressure up-flow microreactor. Prior to testing the solids were
presenting type IV isotherms, 0.80 for Al-A solids and 0.95 for shaped in pellets (size between 0.3 and 0.5 mm). For each test
solids Al-E and Al-F.
10 mL of catalyst pellets diluted with 15 mL of carborundum
Catalysts in their oxidic form were characterized using X-ray (0.25 mm) were loaded in the isothermal zone of the reactor. The
photoelectron spectroscopy (XPS). The powdered samples were catalysts were sulfided using a mixture of dimethyldisulfide
pressed into an indium foil attached to the sample holder. XPS (DMDS) and SRGO containing 1 wt% S (2 wt% DMDS + SRGO)
spectra were recorded using a VG ESCALAB 220 XL spectro- and H2. After sulfidation, the feed was switched to the test feed
meter equipped with a monochromatic Al Ka (E = 1486.6 eV) (the same SRGO as for sulfidation), the operating conditions
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
being as follows: pressure = 35 bar; H2/oil = 250 NL/L; LHSV = 2 hꢀ1
NJC
.
gel. When comparing with solid Al-A prepared without porogen,
During each run, the temperature was changed in the sequence the addition of 0.2 mole of P123 allows obtaining a solid with
350–360–370 1C. The total sulfur contents of feed and product higher SSA (+24% for Al-B). An increase in pore volume and in
samples were determined using an Antek 9000S sulfur analyzer pore diameter is also observed, +25% and +11%, respectively.
(Ultra-Violet fluorescence).
When still increasing the porogen content, SSA decreases from
445 m2 gꢀ1 for the Al-B solid to 337 m2 gꢀ1 for the Al-D solid,
while the largest values of 0.98 cm3 gꢀ1 for the pore volume and
7.4 nm for the pore diameter are obtained at an intermediate
value of P123 content, that is 0.3 mole of P123 added per mol of
Al3+ (Al-C solid). These results suggest that to preserve a surface
area larger than 400 m2 gꢀ1, the structuring agent content has
to be lower than 0.3 mole per mol of Al3+ in the studied range.
The isotherm obtained for the solid prepared without poro-
gen Al-A is intermediate between type IV and type II, whereas a
pure type IV isotherm is obtained (shown for Al-B, Fig. 1a) for
solids prepared with P123. This suggests more homogeneous
aggregation of particles in the case of the solid prepared with
P123. Pore size distribution also appears larger in the presence
of porogen (Al-B) than without (Al-A) (Fig. 1b), while in the
literature, in the case of organized mesoporous structures, the
presence of the structuring agent usually leads to narrower pore
size distributions. For the solids synthesized in this work,
porosity seems to be generated by the arrangement of elemen-
tary crystallites, the repartition of the pore size being more
related to their size and arrangement. No modification of the
isotherm shape is further driven by par the porogen content
increase. Then, in contrast to what is observed for ordered
materials, no formation of channel-type porosity (as in silica
materials, e.g. SBA-15, KIT-6, MCM-41) is awaited since porosity
originates from elementary particle arrangements in the aggre-
gates formed.
3. Results and discussion
Synthesis parameters were varied during the preparation of the
alumina supports, such as the quantity of porogen, the quantity
of water for the hydrolysis step, the hydrolysis temperature and
the autoclaving temperature (solids Al-A to Al-K). Among the
prepared supports, 6 aluminas were then selected to be used
as HDS CoMo catalyst supports, on the basis of their textural
parameters. Surface areas, porous volumes and pore diameters
of the alumina and of the corresponding catalysts are listed in
Table 2, as well as the maximum molybdenum densities imposed by
the water pore volume of the solids (corresponding to the amount
of water necessary to wet the support without excess). Among
the selected supports, two cannot achieve the desired dMo of
4 atMo nmꢀ2 due to their low porous volume, but were however
chosen due to their high SSA (Al-B and Al-I). The corresponding cata-
lysts were then prepared at a molybdenum density of 3 atMo nmꢀ2
.
The dMo achieved on the final solid (dMoreal) is calculated taking
into account the variation of the surface area of the support during
the impregnation–calcination process. For comparison purposes,
a commercial alumina, the Pural SB3 noted Al-P, was also used
for catalyst preparation, with a dMo of 4 atMo nmꢀ2
.
3.1. Characterization of aluminas synthetized by the sol–gel
method in the presence of P123
Effect of the quantity of water during the hydrolysis step,
Effect of structuring agents, A–B–C–D aluminas in Table 2. B–E–F aluminas in Table 2. The quantity of water added for the
In samples Al-A, Al-B, Al-C and Al-D, porogen content was hydrolysis step increased from 3.0 to 5.9 mol of H2O per mol of
varied from 0 to 0.6 mole per mole of Al3+ in the synthesis Al3+: Al-B (3 mol), Al-E (4.5 mol) and Al-F (5.9 mol). A decrease in
Table 2 Textural properties of alumina supports and derived catalysts
a
a
SBET
Vp
Dp
dMomax
SBET
Vp
Dp
dMoreal
DSBET DDp DVp
Sample
(m2 gꢀ1
)
(cm3 gꢀ1
)
(nm) (at per nm2) (m2 gꢀ1
)
(cm3 gꢀ1
)
(nm) (at per nm2) (%)
(%)
(%)
Al-A: Al-S0-W3-Th15-HT45
CoMo24.0/Al-A
Al-B: Al-S0.2-W3-Th15-HT45
CoMo22.8/Al-B
Al-C: Al-S0.3-W3-Th15-HT45
Al-D: Al-S0.6-W3-Th15-HT45
Al-E: Al-S0.2-W4.5-Th15-HT45 438
CoMo27.4/Al-E
Al-F: Al-S0.2-W5.9-Th15-HT45 402
CoMo25.9/Al-F
Al-G: Al-S0.2-W3-Th25-HT45
Al-H: Al-S0.2-W3-Th50-HT45
CoMo26.1/Al-H
Al-I: Al-S0.2-W3-Th15-HT15
CoMo21.1/Al-I
Al-J: Al-S0.2-W3-Th15-HT65
Al-K: Al-S0.2-W3-Th15-HT85
Al-P: Pural
359
445
0.62
0.78
5.4
6.0
4.1
3.1
—
331
—
303
—
—
—
375
—
266
—
—
354
—
318
—
—
—
—
0.47
—
0.55
—
—
—
0.81
—
0.64
—
—
0.76
—
0.61
—
—
—
—
4.6
—
6.0
—
—
—
7.3
—
7.2
—
—
7.4
—
6.0
—
—
—
—
4.3
—
4.4
—
—
—
4.7
—
6.0
—
—
4.6
—
3.8
—
—
—
—
—
—
ꢀ8
ꢀ15 ꢀ24
—
—
—
ꢀ32
0
ꢀ29
412
337
0.98
0.70
1.26
7.4
6.8
9.0
—
—
5.3
—
—
—
—
—
—
—
—
—
ꢀ15
ꢀ19 ꢀ36
1.31
8.9
6.3
—
—
ꢀ34
—
—
ꢀ19 ꢀ51
429
407
1.04
1.00
7.5
7.6
—
4.6
—
—
—
—
ꢀ13
ꢀ3 ꢀ24
402
0.81
6.1
3.3
—
—
—
ꢀ21
—
—
ꢀ2 ꢀ25
409
407
179
0.97
1.03
0.46
6.5
7.5
7.3
—
—
4.8
—
—
—
—
—
—
—
CoMo14.1/Al-P
196
0.45
7.3
4.0
ꢀ9
0
ꢀ2
a
Values corrected by removing the contribution of the weight gain consecutive to the introduction of the active phase (expressed per gram of support).
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Fig. 1 Effect of adding porogen on the N2 adsorption–desorption isotherm shape (a) and BJH pore size distributions (b); comparison between Al-A
(without porogen) and Al-B.
SSA is observed, from 445 m2 gꢀ1 for sample Al-B to 402 m2 gꢀ1 hydrolysis temperature, the surface area decreases but remains
for sample Al-F, but the three materials present values of SSA higher than 400 m2 ꢀ1. The decrease in surface area is also
g
above 400 m2
and pore diameter is observed with the water content increase, volume (Table 2).
with maximum values of pore volume and pore diameter being Effect of the hydrothermal treatment temperature, I–B–J–K
g
ꢀ1. A significant increase in pore volume accompanied by a slight increase in pore volume and pore
1.31 cm3 gꢀ1 and B9.0 nm, respectively, obtained for Al-F aluminas in Table 2. The temperature applied for the hydro-
prepared with the highest water content. A significant modifica- thermal treatment (15 1C (Al-I) to 85 1C (Al-K)) also impacts the
tion of the isotherm shape is however observed with the increase textural properties of the final alumina. All the materials from
in H2O content used for the hydrolysis step: the isotherm is this series present SSA above 400 m2 gꢀ1, the highest value being
type IV with 3 moles of H2O and is shifting to type-IV and II obtained at 45 1C (Al-B with 445 m2 gꢀ1). While an increase in
complex isotherms at higher water content (Fig. 2). At the temperature from 15 1C to 45 1C does not significantly impact
highest water content, the contribution of the type II is becoming the pore volume and the pore size, both parameters are observed
predominant.
to progressively increase with the hydrothermal treatment from
Hydrolysis temperature effect, B–G–H aluminas in Table 2. 45 1C to 85 1C. At 85 1C, the pore volume reaches 1.03 cm3 gꢀ1
The increase in the hydrolysis temperature, from 15 1C (Al-B) to with a mean pore diameter of 7.5 nm. The increase in hydro-
45 1C (Al-H), induces the modification of the textural parameters of thermal temperature does not result in a modification of the
the final alumina. However, the isotherm shape remains isotherm shape that remains of pure type IV whatever the
unchanged, pure type IV shape, whatever the hydrolysis tempera- autoclave temperature (type IV, Fig. 3).
ture in the range considered. The main impact of the hydrolysis
From this series of materials, aluminas with different textural
temperature is on the surface area that decreases with the increase properties were selected according to (i) decreasing SSA: Al-B 4
in hydrolysis temperature. Consequently, decreasing the hydrolysis Al-E 4 Al-F, Al-H, Al-I 4 Al-A c Al-P, and (ii) decreasing pore
temperature down to ambient temperature (15 1C, for Al-B) volume: Al-F 4 Al-E 4 Al-H 4 Al-B, Al-I 4 Al-A c Al-P, with
allows obtaining a surface area close to 450 m2 gꢀ1. At higher Al-F, Al-H, and Al-I at the almost iso-surface area.
Fig. 2 N2 adsorption–desorption isotherms (a) and BJH pore size distributions (b) of aluminas Al-B, Al-E and Al-F (n(H2O) = 2.9, 4.5, 5.9).
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
decrease in pore volume values is observed for all these
materials, larger than 20%, independent of the evolution of
the pore diameter and/or surface area. Then, a significant
rearrangement of the particles in the aggregates is awaited
upon active phase impregnation. The oxidic precursors can be
however classified according to the evolution of their textural
properties upon impregnation, knowing that the porous
volume is always decreasing.
ꢁ CoMo26.1/Al-H is the most stable support, with SSA and
pore diameter with respective variations of 13% of SSA and 3%
of pore diameter.
ꢁ CoMo22.8/Al-B and CoMo21.1/Al-I underwent a significant
decrease in SSA (32 and 21%) while they retain their pore
diameter (variation of 0 and 2%, respectively).
Fig. 3 N2 adsorption–desorption isotherms obtained for aluminas pre-
pared at different hydrothermal treatment temperatures. Al-I, 15 1C; Al-B,
45 1C; Al-J, 65 1C, and Al-K, 85 1C.
ꢁ CoMo27.4/Al-E and CoMo25.9/Al-F exhibit a significant
decrease in both SSA and pore diameter, with 14 and 34%,
respectively, for the SSA and 19% for the pore diameter.
ꢁ CoMo24.0/Al-A underwent an important decrease in the
pore diameter (15%) with almost no variation of its SSA (8%).
Most of the solids present molybdenum density close to that
expected, except for solids CoMo22.8/Al-B and CoMo25.9/Al-F for
which the dMo on the final solid is much higher, in agreement
with the significant evolution of the support SSA upon impreg-
nation and calcination steps (above 30%).
The fragility of the sol–gel derived aluminas towards impreg-
nation is then clearly evidenced, especially when compared to
the textural properties of a commercial Pural alumina that
remain unchanged. The sensitivity of mesoporous alumina
supports obtained from the sol–gel route for catalyst prepara-
tion conditions was already highlighted in previous studies. On
alumina obtained by the hydrolysis of aluminum sec-butoxide,
in the presence of P123, Bejenaru et al.36 measured a decrease
of 19% for the surface area and of 47% for the pore volume.
Pore size distribution was also observed to significantly evolve,
from a bimodal one for alumina to a monomodal one, with a
narrower width of pore distribution, for the oxidic precursor.
Kaluˇza et al.,34 on alumina prepared by the hydrolysis of
aluminum sec-butoxide in the presence of a stearic acid type
template, also observed significant evolution of the textural
properties of calcined alumina upon catalyst preparation.
3.2. CoMo oxidic precursor properties
3.2.1 Stability of supports upon impregnation-calcination
processes. CoMo oxidic precursors were characterized after
impregnation and calcination. All the samples retain isotherms
similar to those obtained before impregnation. Indeed, CoMo-
containing materials derived from Al-B, Al-H and Al-I present
the isotherm of type IV (as shown for catalysts prepared on Al-I
in Fig. 4a), whereas Al-A, Al-E and Al-F based ones exhibit a
mixed II/IV isotherm shape (as illustrated in Fig. 5a for catalyst
CoMo27.4/Al-E).
Significant variations of the textural parameters are observed
upon catalyst preparation (Table 2). For CoMo22.8/Al-B, CoMo21.1/
Al-I and CoMo25.9/Al-F, a significant decrease in SSA is noticed
(larger than 20% of SSA decrease), while for CoMo26.1/Al-H,
CoMo27.4/Al-E and CoMo24.0/Al-A, a smaller SSA reduction is
measured, with DSBET always inferior to 15%. For CoMo22.8/Al-B,
CoMo21.1/Al-I and CoMo26.1/Al-H, the pore sizes and pore size
distributions remain of comparable value to that of the bare
corresponding alumina supports (Fig. 4b and Table 2, decrease
of value below 4%). Larger evolutions are observed for catalysts
CoMo27.4/Al-E, CoMo25.9/Al-F and CoMo24.0/Al-A, with the
decrease in pore size being between 15 and 34%. A significant
Fig. 4 N2 adsorption–desorption iostherms (a) and BJH pore size distributions (b) obtained for Al-I and CoMo21.1/Al-I.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Fig. 5 N2 adsorption–desorption iostherms (a) and BJH pore size distributions (b) obtained for Al-E and CoMo27.4/Al-E.
The authors attributed these textural property evolutions to a
Over conventional alumina, Mazurelle et al., with the same
sintering of particles during the thermal treatment step. However, impregnating precursor, observed the formation of MoO3 and
despite significant textural property modifications, final catalysts still CoMoO4 for a loading largely lower than the one reported in
present adequate properties for use in the heterogeneous catalysis the present study (B16 wt% MoO3).35 The limited pore volume
process. In addition, the use of these mesoporous supports allows (0.8 cm3 gꢀ1) displayed by the g-alumina extrudates used is at
increasing the loading in the active phase (up to 27 wt% MoO3, the origin of the poor Mo phase dispersion (at a surface density
Table 2), while the loading is limited when classical support, like of 4.3 atMo nmꢀ2). Indeed at this Mo loading, considering the
commercial Pural support (14.1 wt%, Table 2), is used.
3.2.2 Dispersion of cobalt and molybdenum species in the leading to precipitation during the impregnation/drying step.
oxidic precursors. Fig. 6 shows the Raman spectra of the CoMo The formation of Co3O4 is usually not observed when using the
support pore volume, the impregnation solution is saturated,
catalysts. Materials can be classified according to the species Co2Mo10 heteropolycompound as a precursor of molybdenum and
identified. CoMo22.8/Al-B, CoMo21.1/Al-I and CoMo26.1/Al-H cobalt, and whatever the support nature (zirconium oxide, titanium
present a broad line at 952 cmꢀ1, corresponding to polymolyb- oxide or g-alumina).35,43–47 However, commercial supports, classi-
date species well dispersed on the support surface (Fig. 6a), cally used for hydrotreatment studies, are highly stable toward the
similar to what is observed by impregnation over classical impregnation process.43,44
alumina.40,41 Lines at 690, 480 and 516 cmꢀ1, attributed to
All sol–gel aluminas of our series underwent a decrease in
Co3O4, are also identified.42 While the intensity of the lines surface area and pore volume during the impregnation/calcination
associated with polymolybdate species remains unchanged, the process. For the three solids CoMo27.4/Al-E, CoMo25.9/Al-F and
intensity of the line at 690 cmꢀ1 is low in the Raman spectra CoMo24.0/Al-A with CoMoO4, a severe decrease in pore diameter
of CoMo21.1/Al-I and CoMo26.1/Al-H catalysts compared to (between 15 and 19%) is observed in addition to the decrease in pore
CoMo22.8/Al-B, indicating that Co3O4 is probably minor on these volume. In our case the concentration of the impregnating solutions
two solids. On CoMo27.4/Al-E, CoMo25.9/Al-F and CoMo24.0/Al-A is calculated to obtain a dMo inferior to the maximum dMo imposed
(Fig. 6b), lines identified at 820 and 940 cmꢀ1 evidence the presence by the initial water pore volume of the support. Thus the solutions
of b-CoMoO4, characteristic of a poor dispersion of molybdenum are not saturated and no CoMoO4 is expected after calcination. But
species in the porous network of the support.43 However considering the modification of the porous network during the impregnation
the very high diffusion cross-section of CoMoO4 compared to that of step induces an excess of solution outside the pores and subsequent
polymolybdate species, the presence of well-dispersed polymolybdate precipitation during the drying step of CoMo species leading to the
species can also be assumed.
formation of CoMoO4 during calcination.
Fig. 6 Raman spectra obtained for the CoMo oxidic precursors.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
On the other hand, solids with Co3O4 present a stable pore
NJC
Table 3 nMo/nAl, nCo/nAl and nCo/nMo atomic ratios; (th) theoretical values
calculated from material composition; (XPS) experimental values obtained
from XPS analysis
diameter upon impregnation (with limited variations between 0
and 4%). Co2Mo10 solution, used for impregnation, exhibits a
slightly acidic pH (close to 4). Under these conditions, the
alumina surface is positively charged (the alumina zero point
charge is B8).48 [Co2Mo10O38H4]6ꢀ anions are expected to interact
with the positively charged surface of alumina. However, free Co2+
counter cations in solution will not interact with the surface
support. It seems that in the case of the decrease in porous volume
while maintaining the pore diameter, a significant amount of Co2+
containing solution can be expulsed outside the pores and
precipitates as Co hydroxide out of the internal porosity during
the solvent evaporation step, leading to the formation of Co3O4
upon thermal treatment. In any case, molybdenum species are
well dispersed in these solids, demonstrating the ability of
these sol–gel supports to maintain high molybdenum disper-
sion at high Mo content (up to 26.1 wt% MoO3).
3.2.3 Surface analysis of the oxidic precursors. XPS spectra
of Al2p, O1s, Mo3d, Co2p and C1s levels were recorded. Fig. 7
shows the typical spectra of Mo3d and Co2p levels, recorded for
CoMo26.1/Al-H. For all materials, the Mo3d5/2 and Co2p3/2
photopeaks are located at binding energies of 233.1 ꢂ 0.2 eV
and 781.5 ꢂ 0.2 eV, respectively. These positions are character-
istic of Mo6+ and Co2+ species in oxidic environments.
nMo/nAl nMo/nAl nCo/nAl nCo/nAl nCo/nMo C1s
Sample
(th)
(XPS)
(th)
(XPS)
(XPS)
(%)
CoMo24.0/Al-A 0.12
CoMo22.8/Al-B 0.11
CoMo27.4/Al-E 0.15
CoMo25.9/Al-F 0.14
CoMo26.1/Al-H 0.14
0.16
0.13
0.17
0.19
0.15
0.14
0.061
0.057
0.074
0.068
0.069
0.051
0.050
0.054
0.051
0.050
0.054
0.057
0.31
0.42
0.30
0.27
0.37
0.40
0.8
11.1
18.3
19.6
17.4
8.5
CoMo21.1/Al-I
0.10
The values of nCo/nMo ratios are all lower than the theoretical
value of 0.5. The highest ratios are obtained for the Co3O4
containing solids, with values around 0.4, close to the optimum
Co/Mo ratio of 0.5 as reported in the literature.41
Finally, all materials exhibit a high surface carbon content,
ranging between 8 and 19% (Table 3), except for catalyst
CoMo24.0/Al-A, prepared without a porogen, presenting a very
low percentage (0.8%). This indicates that the carbon deposition
observed in the other solids is mainly related as expected to the
carbon residues issued from the calcination of the structuring
agent.
3.3. Catalytic performances in the HDS of thiophene
Table 3 presents atomic surface ratios, as determined by XPS,
as well as theoretical bulk ratios. For all solids, nMo/nAl surface Catalytic performance of the solids was evaluated in the HDS of
ratios are higher than the bulk ones, the differences being thiophene at 300 1C (Table 4). Reaction rates are found to vary
independent of the presence or not of CoMoO4. Indeed, the over a wide range, from 79.3 to 209.4 L hꢀ1 kgꢀ1, depending on
CoMoO4 phase should lead to lower surface ratios than bulk the support used. CoMo24.0/Al-A presents the lowest reaction
ratios, as only a fraction of molybdenum is detected in these rate, of only 79.3 L hꢀ1 kgꢀ1, and CoMo26.1/Al-H is the
bulk particles.49 This difference can be due to the presence in most active material from the series, with a reaction rate of
our case of particles of CoMoO4 small enough to ensure the 209.4 L hꢀ1 kgꢀ1. Meanwhile, CoMo22.8/Al-B, CoMo21.1/Al-I,
complete detection of molybdenum by XPS, and/or to large CoMo27.4/Al-E and CoMo25.9/Al-F present intermediate reaction
alumina particles with consequently not all Al atoms detected rates, in a narrow range between 133.1 and 147.1 L hꢀ1 kgꢀ1. For
by XPS. Concerning nCo/nAl surface ratios, values appear to be these materials, the molybdenum loading varies between 21 and
lower than bulk ratios, except for sample CoMo21.1/Al-I (increase 27% and despite these significant differences in Mo loading,
of 12%) that contains a low quantity of Co3O4 compared to they present similar activities. Under comparable reaction
other samples. Obtaining a lower nCo/nAl surface ratio could conditions, Bejenaru et al.36 reported variable activities,
thus be due to the presence of Co3O4 or CoMoO4. These lower between 55.1 and 161.1 L hꢀ1 kgꢀ1, for CoMo-S materials with
ratios could also be related to the migration of cobalt atoms in 10 wt% Mo, issued from the impregnation of mesostructured
the tetrahedral sites of alumina particles that would be too aluminas with different morphologies with conventional pre-
large to be completely analyzed by XPS.
cursors i.e. ammonium heptamolybdate and cobalt nitrate.
Fig. 7 XPS spectra of Mo3d (a) and Co2p (b) levels recorded for CoMo26.1/Al-H.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Table 4 Catalytic properties of the synthesized materials in the thiophene
HDS reaction at 300 1C. Reaction rates are expressed per kg of catalyst or
normalized per mole of molybdenum
residual carbon has been reported to be beneficial by several
authors that suggest that carbon may induce different sulfida-
tion degrees of Co and Mo, and different structures and/or
morphologies of the generated active phase.53,54 Different sulfida-
tion degrees of Mo and Co may obviously lead to more or less
promoted and dispersed active phases.55 The role of carbonaceous
residues of the support surface in the sulfided phase properties is
multiple. Heat generated by the sulfidation reaction can be
partially absorbed by these residues, resulting in crystallite growth
different from that over carbon free support.50 Glasson et al.50
observed positive effects of the presence of carbonaceous species
on the CoMo-S/Al2O3 activity for the HDS of thiophene and
Reaction rate
Normalized reaction rate
ꢀ1
Sample
(L hꢀ1 kgꢀ1
)
(L hꢀ1 molMo
)
CoMo24.0/Al-A
CoMo22.8/Al-B
CoMo27.4/Al-E
CoMo25.9/Al-F
CoMo26.1/Al-H
CoMo21.1/Al-I
CoMo14.1/Al-P
79.3
145.9
133.1
147.1
209.4
135.3
153.5
47.6
92.1
70.0
81.8
115.5
92.3
156.1
Activities reported here are significantly higher, but the materials petroleum fractions. These authors also stated that carbon restricts
prepared in this study present higher Mo loadings (Table 2). The or inhibits the migration and/or growth of the sulfided phase,
CoMo catalyst prepared over commercial Pural alumina also isolating active phase crystallites (geometric order effect). Also,
shows an intermediate activity of 153.5 L hꢀ1 kgꢀ1, an activity Co(Ni)Mo-S supported over activated carbon showed higher HDS
coherent with those reported in the literature for commercial activity than the active phase supported on alumina.56–59 The
alumina supported CoMo catalysts. Indeed, Glasson et al.50 formation of an active carbide phase, ‘‘CoMoC’’, could also
measured the reaction rate in thiophene HDS at 300 1C around explain the better performance when the active phase is
141 L hꢀ1 kgꢀ1 for a CoMo catalyst supported on commercial supported over carbon.60 Consequently, such a phenomenon,
alumina (12.3% Mo). Materials presented here appear therefore modifying the intrinsic activities of the CoMo-S phase, could
to be particularly efficient, especially considering CoMo26.1/ not be excluded over these carbon containing supports.
Al-H with activity at least 20% higher than those reported in
the literature.
On CoMo27.4/Al-E and CoMꢀo125.9/Al-F, intermediate activities
at around 70–80 L hꢀ1 molMo are obtained. These supports
In order to compare the performance of solids at different Mo present low stability upon impregnation, with a significant
loadings, reaction rates were normalized per mole of molybdenum. decrease in both SSA (14 and 34%) and pore diameter (19%).
Normalized activity is found to vary also by a factor 3, between Raman spectroscopy evidenced the formation of CoMoO4 species,
47.6 and 115.5 L hꢀ1 molMoꢀ1. The most efficient solid, which can explain the lower performances of these solids. Con-
CoMo26.1/Al-H, is the material prepared with the support sequently, if the supports were stabilized before impregnation,
presenting the best stability toward impregnation: the modifi- significantly higher activity could be achieved since the pore
cation of SSA of 13% and the pore diameter of only 3%. network reorganization is expected to be at the origin of the
CoMo22.8/Al-B and CoMo21.1/Al-I present a lower and compar- formation of this detrimental CoMoO4 phase. Finally, the lowest
able normalized reaction rate (92 L hꢀ1 molMoꢀ1) for similar performance is obtained for CoMo24.0/Al-A (47.6 L hꢀ1 molMoꢀ1),
final SSA (B300 m2 gꢀ1). However, both supports suffered from which can be related to both the detrimental presence of
the significant porous volume decrease upon impregnation CoMoO4 species and to the lowest carbon content (0.8%) since
while the pore diameter remained unchanged. These three the parent support is prepared without an organic porogen. The
solids, CoMo22.8/Al-B, CoMo21.1/Al-I and CoMo26.1/Al-H, pre- catalyst CoMo14.1/Al-P prepared on the commercial alumina
sent similar pore characteristics (type IV isotherms) and well- presents the highest conversion per mole of molybdenum
dispersed polymolybdate species as evidenced by Raman (156.1 L hꢀ1 molMoꢀ1). However higher conversion per mass
spectroscopy. The presence of Co3O4 is evidenced in these three of catalyst could be achieved for derived sol–gel alumina based
solids, suggesting that the loss of a fraction of the initial cobalt solids since their textural properties allow higher molybdenum
in the form of a poorly dispersed bulk oxide phase (that will not loading, while preserving satisfying molybdenum dispersion.
incorporate the CoMoS active phase) does not directly influence
the material catalytic performance. This result can be explained
on the basis of literature data analysis. Indeed, theoretical
3.4. Catalytic performances in the HDS of SRGO
calculations, based on the density functional theory (DFT), The most efficient catalyst in the HDS of thiophene CoMo28.7/
showed that a ratio Co/Mo of 0.3 is enough to promote in Al-H was selected for further evaluation of its performance in
an optimal way the MoS2 slabs.51,52 Consequently, the loss of the HDS of a SRGO, to be compared to that of the commercial
part of cobalt in the form of bulk cobalt oxide, in the case of the Pural based solid. For both solids, the residual sulfur content in
Al-B, Al-H and Al-I supports, does not significantly impact the the hydrotreated product is plotted as a function of the reactor
activity of the catalysts since sufficient well dispersed cobalt temperature, as shown in Fig. 8. The mesoporous sol–gel based
atoms remain for promoting the MoS2 slabs, these three solids catalyst appears to be significantly more efficient than the
presenting the highest Co/Mo XPS atomic ratio (around 0.4). commercial supported one (per volume of catalyst and per mass
It can also be noted that for these three solids, the amount of of catalyst, taking into account their similar apparent densities),
residual carbon content is relatively high (around 18%) and with 74 versus 415 ppmS at 350 1C and 10 versus 100 ppmS at
could thus impact catalytic activity. Indeed the presence of 370 1C. Moreover, these performances are close to that of a
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
than a catalyst prepared over commercial alumina, when comparing
the activity normalized per molybdenum atom. However, the use of
these supports allows us to reach higher activities per gram of
catalyst, due to the possibility of significantly increasing the Mo
loading without the formation of detrimental phases (CoMoO4).
The most efficient catalyst in the HDS of thiophene,
CoMo28.7/Al-H, was evaluated in the HDS of a SRGO. Its
performance appears to be higher than that of the commercial
Pural based catalyst and close to that of a commercial catalyst
modified by a chelating agent, evaluated under the same conditions.
These results clearly evidence the potential of the sol–gel
aluminas developed in this study as support for a highly loaded
HDS catalyst, not only in the conversion of model molecules
but also more interestingly in the desulfurization of real
feedstocks, with the challenge now to stabilize the porosity of
these supports.
Fig. 8 Catalytic performance of the CoMo28.7/Al-H and CoMo14.1/Al-P
catalysts in the HDS of a SRGO: residual sulfur content in the desulfurized
feed versus the temperature of the reactor.
commercial catalyst modified by a chelating agent, evaluated Acknowledgements
under the same conditions.61
The Equatorian Government (for the attribution of a PhD grant to
These results clearly evidence the potential of the sol–gel
aluminas developed in this study as support for the HDS catalyst,
not only in the conversion of model molecules but more inter-
estingly also in the desulfurization of real feedstocks. From our
knowledge, this is the first time a mesoporous sol–gel alumina
based CoMo HDS catalyst was evaluated in HDS of SRGO under
experimental conditions representative of industrial ones.
˜
A. Mino) and IFPEN (for financial support) are acknowledged.
`
´
Chevreul Institute (FR 2638), Ministere de l’Enseignement Superieur
´
et de la Recherche, Region Nord – Pas de Calais and FEDER
are acknowledged for supporting and funding this work.
M. Trentesaux is acknowledged for XPS analysis support.
References
4. Conclusions
1 Directive 2009/30/EC of the European Parliament.
2 H. Topsøe, B. S. Clausen and F. E. Massoth, Hydrotreating
Catalysis, Springer, Berlin, 1996.
3 R. Prins, V. H. J. De Beer and G. A. Somorjai, Catal. Rev.: Sci.
Eng., 1989, 32, 1.
4 E. J. M. Hensen, Y. van der Meer, J. A. R. van Veen and
J. W. Niemantsverdriet, Appl. Catal., A, 2007, 322, 16.
5 J. F. Lepage, Catalyse de Contact: conception, preparation et
Alumina supports were prepared using a template assisted sol–
gel method. The adjustment of alumina textural properties was
performed through the control of synthesis parameters, such as
the quantity of porogen, the quantity of water, the temperature
of the hydrolysis step, and the temperature of the hydrothermal
treatment. High surface areas, close to 400 m2 gꢀ1 for most of the
aluminas, were obtained with pore diameters up to 9 nm. Among
the prepared materials, six supports, with varying textural properties,
were selected for the preparation of CoMo-S HDS catalysts.
´
mise en œuvre des catalyseurs industriels, TECHNIP, Paris,
1978.
`
Characterization revealed a fragility of the support texture
toward the impregnation. A significant decrease in surface area,
pore volume and pore diameter is observed depending on the
materials. However, despite these significant alterations, the final
values of textural parameters remain attractive for HDS catalyst
preparation. Characterization of the oxidic precursors evidenced
two groups of solids, one where molybdenum species are well
dispersed on the support (over the more stable supports), and the
other where the presence of b-CoMoO4 is observed (the more
6 S. Dzwigaj, C. Louis, M. Breysse, M. Cattenot, V. Belliere,
C. Geantet, M. Vrinat, P. Blanchard, E. Payen, S. Inoue,
H. Kudo and Y. Yoshimura, Appl. Catal., B, 2003, 41, 181.
7 C. Lamonier, D. Soogund, J. Mazurelle, P. Blanchard,
D. Guillaume and E. Payen, Stud. Surf. Sci. Catal., 2006, 162, 713.
8 I. Merdrignac and D. Espinat, Oil Gas Sci. Technol., 2007,
62, 7.
9 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and
J. S. Beck, Nature, 1992, 359, 710.
sensible supports, the formation of this phase being related to 10 S. A. Bagshaw and T. J. Pinnavaia, Angew. Chem., 1996,
the pore diameter decrease during the impregnation process). 35, 1102.
The performance in thiophene HDS at 300 1C varies by a factor 11 F. Vaudry, S. Khodabandeh and M. E. Davis, Chem. Mater.,
3 depending on the support used. Low activity is measured for 1996, 8, 1451.
materials showing the formation of a CoMoO4 phase before 12 S. Valange, J.-L. Guth, F. Kolenda, S. Lacombe and
sulfidation. Higher activity is measured over the more stable Z. Gabelica, Microporous Mesoporous Mater., 2000, 35–36, 597.
supports, the most efficient solid of the series showing the best 13 C. Marquez-Alvarez, N. Zilkova, J. Perez-Pariente and
ˇ
´
´
´
ˇ
stability toward impregnation. These materials are less active
J. Cejka, Catal. Rev.: Sci. Eng., 2008, 50, 222.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
14 K. Niesz, P. Yang and G. A. Somorjai, Chem. Commun., 2005, 37 S. Badoga, R. V. Sharma, A. K. Dalai and J. Adjaye, Appl.
1986. Catal., A, 2015, 489, 86.
15 Q. Yuan, A.-X. Yin, C. Luo, L.-D. Sun, Y.-W. Zhang, W.-T. Duan, 38 C. Martin, C. Lamonier, M. Fournier, O. Mentre, V. Harle,
H.-C. Liu and C.-H. Yan, J. Am. Chem. Soc., 2008, 130, 3465. D. Guillaume and E. Payen, Chem. Mater., 2005, 17, 4438.
16 J.-P. Dacquin, J. Dhainaut, D. Duprez, S. Royer, A. F. Lee and 39 M. J. Ledoux, C. P. Huu and F. Luck, J. Catal., 1990, 121, 70.
´
´
´
K. Wilson, J. Am. Chem. Soc., 2009, 131, 12896.
17 W. Zhang and T. J. Pinnavaia, Chem. Commun., 1998, 1185.
18 T. J. Pinnavaia, Z. Zhang and R. W. Hicks, Stud. Surf. Sci.
Catal., 2005, 156, 1.
40 E. Payen, J. Grimblot, J. C. Lavalley, M. Daturi and F. Mauge,
Handbook of Vibrational Spectroscopy, John Wiley & Sons,
2006.
41 P. Blanchard, C. Lamonier, A. Griboval and E. Payen,
Appl. Catal., A, 2007, 322, 33.
42 C.-W. Tang, C.-B. Wang and S.-H. Chien, Thermochim. Acta,
2008, 473, 68.
43 C. Lamonier, C. Martin, J. Mazurelle, V. Harle, D. Guillaume
ˇ
´
19 J. Cejka, P. J. Kooyman, L. Vesela, J. Rathousk´y and A. Zukal,
Phys. Chem. Chem. Phys., 2002, 4, 4823.
20 K. Niesz, P. Yang and G. A. Somorjai, Chem. Commun., 2005,
1986.
´
21 A. Caragheorgheopol, H. Caldararu, G. Ionita, F. Savonea,
and E. Payen, Appl. Catal., B, 2007, 70, 548.
ˇ
ˇ
´
N. Zilkova, A. Zukal and J. Cejka, Langmuir, 2005, 44 C. I. Cabello, F. M. Cabrerizo, A. Alvarez and H. J. Thomas,
21, 2591. J. Mol. Catal. A: Chem., 2002, 166, 89.
22 L. Vradman, M. V. Landau, M. Herskowitz, V. Ezersky, 45 P. A. Nikulshin, D. I. Ishutenko, A. A. Mozhaev,
M. Taliander, S. Nikitenko, Y. Koltypin and A. Gedanken,
J. Catal., 2003, 213, 163.
23 L. Lizama and T. Klimova, Appl. Catal., B, 2008, 82, 139.
K. I. Maslakov and A. A. Pimerzin, J. Catal., 2014, 312, 152.
46 P. A. Nikulshin, A. V. Mozhaev, K. I. Maslakov, A. A. Pimerzin
and V. M. Kogan, Appl. Catal., B, 2014, 158–159, 161.
24 R. Huirache-Acuna, B. Pawelec, E. Rivera-Munoz, R. Nava, 47 Al. A. Pimerzin, P. A. Nikulshin, A. V. Mozhaev, A. A. Pimerzin
J. Espino and J. L. G. Fierro, Appl. Catal., B, 2009, 92, 168. and A. I. Lyashenko, Appl. Catal., B, 2015, 168–169, 396.
25 V. La Parola, B. Dragoi, A. Ungureanu, E. Dumitriu and 48 P. Euzen, P. Raybaud, X. Krokidis, H. Toulhoat, J.-L. Le Loarer,
A. M. Vanezia, Appl. Catal., A, 2010, 386, 43.
26 P. E. Boahene, K. K. Soni, A. K. Dalai and J. Adjaye,
Appl. Catal., A, 2011, 402, 31.
27 T. Chiranjeevi, G. Muthu Kumaran and G. Murali Dhar,
Pet. Sci. Technol., 2008, 26, 690.
28 M. Hussain, S.-K. Song, J.-H. Lee and S.-K. Ihm, Ind. Eng.
Chem. Res., 2006, 45, 536.
29 R. Nava, A. Infantes-Molina, P. Castao, R. Guil-Lopez and
B. Pawelec, Fuel, 2011, 90, 2726.
J. P. Jolivet and C. Froidefond, Alumina. In Handbook of Porous
Solids, Wiley-VCH, Weinheim, 2002, p. 1591.
49 P. Blanchard, E. Payen, J. Grimblot, O. Poulet and
R. Loutaty, Stud. Surf. Sci. Catal., 1997, 106, 211.
50 C. Glasson, C. Geantet, M. Lacroix, F. Labruyere and
P. Dufresne, J. Catal., 2002, 212, 76.
51 P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan and
H. Toulhoat, J. Catal., 2000, 190, 128.
52 A. D. Gandubert, E. Krebs, C. Legens, D. Costa, D. Guillaume
and P. Raybaud, Catal. Today, 2008, 130, 149.
30 T. A. Zepeda, T. Halachev, B. Pawelec, R. Nava, T. Klimova,
`
G. A. Fuentes and J. L. G. Fierro, Catal. Commun., 2006, 53 F. Labruyere, P. Dufresne, M. Lacroix and M. Breysse, Catal.
7, 33. Today, 1998, 43, 111.
31 K. Soni, K. C. Mouli, A. K. Dalai and J. Adjaye, Catal. Lett., 54 J. Vangestel, J. Leglise and J. C. Duchet, J. Catal., 1994,
2010, 136, 116. 145, 429.
32 M. T. Nguyen Dinh, P. Rajbhandari, C. Lancelot, P. Blanchard, 55 L. Coulier, V. H. J. De Beer, J. A. R. Van Veen and
C. Lamonier, M. Bonne, S. Royer, F. Dumeignil and E. Payen,
ChemCatChem, 2014, 6, 328.
J. W. Niemantsverdriet, J. Catal., 2001, 197, 26.
56 M. Kouzu, Y. Kuriki, F. Hamdy, K. Sakanishi, Y. Sugimoto
and I. Saito, Appl. Catal., A, 2004, 265, 61.
ˇ
´
33 J. Cejka, N. Zilkova, L. Kaluˇza and M. Zdraˇzil, Stud. Surf. Sci.
Catal., 2002, 141, 243.
57 L. Kaluˇza and M. Zdraˇzil, Carbon, 2001, 39, 2023.
ˇ
ˇ
´
34 L. Kaluˇza, M. Zdraˇzil, N. Zilkova and J. Cejka, Catal. Commun., 58 S. J. Moon and S. K. Ihm, Korean J. Chem. Eng., 1994, 11, 111.
2002, 3, 151.
59 E. J. M. Hensen, M. J. Vissenberg, V. H. J. de Beer, J. A. R. van
Veen and R. A. van Santen, J. Catal., 1996, 163, 429.
60 R. R. Chianelli and G. Berhault, Catal. Today, 1999, 53, 357.
35 J. Mazurelle, C. Lamonier, C. Lancelot, E. Payen, C. Pichon
and D. Guillaume, Catal. Today, 2008, 130, 41.
36 N. Bejenaru, C. Lancelot, P. Blanchard, C. Lamonier, 61 N. Frizi, P. Blanchard, E. Payen, P. Baranek, C. Lancelot,
L. Rouleau, E. Payen, F. Dumeignil and S. Royer, Chem.
Mater., 2009, 21, 522.
M. Rebeilleau, C. Dupuy and J.-P. Dath, Catal. Today, 2008,
130, 32.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.