Influence of the activation atmosphere on the hydrodesulfurization of Co-Mo/SBA-15 catalysts prepared from sulfur-containing precursors

Article abstract of DOI:10.1016/j.apcata.2012.01.015

Mesoporous SBA-15 material was used as support of binary Co-Mo hydrodesulfurization (HDS) catalysts prepared using a novel approach based on the use of already sulfided precursors (ammonium tetrathiomolybdate and cobalt diethyldithiocarbamate). The effects of atmosphere and activation temperature were studied to optimize the preparation of highly active CoMo/SBA-15 hydrodesulfurization catalysts. Two sets of catalysts were synthesized using either a N₂/H₂ (10% H₂) or a H ₂/H₂S (15% H₂S) atmosphere at three different temperatures of activation (723, 773 and 823 K). The catalysts were tested in the HDS of dibenzothiophene (DBT) and the catalysts were characterized by N ₂ physisorption, X-ray diffraction (XRD), scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM). The use of already sulfided precursors leads to a homogeneous dispersion of the active phase inside the SBA-15 channels. Moreover, the N₂/H ₂ activation procedure at 723 K allows obtaining optimized HDS active catalysts. Finally, a confinement effect of MoS₂ slabs inside the SBA-15 channels leads to a high selectivity along the direct desulfurization pathway.

Full text of DOI:10.1016/j.apcata.2012.01.015

Applied Catalysis A: General 419–420 (2012) 95–101
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Influence of the activation atmosphere on the hydrodesulfurization of
Co-Mo/SBA-15 catalysts prepared from sulfur-containing precursors
a,∗
a
b
c
d
G. Alonso-Nú n˜ ez , J. Bocarando , R. Huirache-Acu n˜ a , L. Álvarez-Contreras , Z.-D. Huang ,
d
e
f
a
a
W. Bensch , G. Berhault , J. Cruz , T.A. Zepeda , S. Fuentes
Universidad Nacional Autónoma de México, Centro de Nanociencias y Nanotecnología, Ensenada, Baja California, C.P. 22860, Mexico
Universidad Michoacana de San Nicolás de Hidalgo. Morelia, Michoacán, Mexico
a
b
c
d
e
f
Centro de Investigación en Materiales Avanzados S. C., Chihuahua, Chih., C.P. 31109, Mexico
Institut für Anorganische Chemie, Christian-Albrechts University of Kiel, Max-Eyth-Str. 2, D-24118 Kiel, Germany
Institut de Recherches sur la Catalyse et l’Environnement de Lyon, CNRS Université Lyon I, 02 Av. A. Einstein, 69100 Villeurbanne, France
Facultad de Ciencias Químicas e Ingeniería, Universidad Autónoma de Baja California, Tijuana, B. C., Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
Mesoporous SBA-15 material was used as support of binary Co-Mo hydrodesulfurization (HDS) cat-
alysts prepared using a novel approach based on the use of already sulfided precursors (ammonium
tetrathiomolybdate and cobalt diethyldithiocarbamate). The effects of atmosphere and activation tem-
perature were studied to optimize the preparation of highly active CoMo/SBA-15 hydrodesulfurization
catalysts. Two sets of catalysts were synthesized using either a N2/H2 (10% H2) or a H2/H2S (15% H2S) atmo-
sphere at three different temperatures of activation (723, 773 and 823 K). The catalysts were tested in the
HDS of dibenzothiophene (DBT) and the catalysts were characterized by N2 physisorption, X-ray diffrac-
tion (XRD), scanning electron microscopy (SEM) and high-resolution transmission electron microscopy
Received 4 November 2011
Received in revised form 13 January 2012
Accepted 15 January 2012
Available online 25 January 2012
Keywords:
Hydrodesulfurization
SBA-15
(HRTEM). The use of already sulfided precursors leads to a homogeneous dispersion of the active phase
Activation
ATM
inside the SBA-15 channels. Moreover, the N2/H2 activation procedure at 723 K allows obtaining opti-
mized HDS active catalysts. Finally, a confinement effect of MoS2 slabs inside the SBA-15 channels leads
to a high selectivity along the direct desulfurization pathway.
Dithiocarbamate
Confinement effect
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
the development of highly dispersed active phase on zeolites or
mesoporous materials has been very intense in order to respect the
stringent regulations concerning the maximum amount of sulfur
admitted in fuels. Some results have been presented before in pub-
lications concerning the effect of support in hydrotreating catalysts
for ultra clean fuels [12,14].
Nowadays, the processing of “more dirty” feeds containing
larger amounts of sulfur has become more and more urgently
needed and a new generation of transition metal sulfide-based
catalysts with higher activities, greater selectivity, and bet-
ter resistance to poisoning is required. Using high-performance
hydrodesulfurization (HDS) catalysts is then necessary to achieve
lower sulfur concentration levels satisfying environmental restric-
tions [1].
Typically, hydrotreatment (HDT) reactions are catalyzed by
sulfided Co(Ni)Mo(W) catalysts supported on alumina [2]. The ori-
gin of the almost exclusive use of alumina as support has been
ascribed to its outstanding textural and mechanical properties and
its relatively low cost [3]. However, the presence of undesirable
strong metal–support interaction when using alumina has trig-
gered research devoted to the development of new supports for
HDT applications [4–13]. Thus, more recently, research concerning
On the other hand, the catalyst activation is also a crucial step
for the physicochemical properties (nature, composition and dis-
persion of the active phase) and consequently for activity and
selectivity properties. In this respect, the temperature of treatment
used during this step can affect the degree of reduction and sulfida-
tion as well as the dispersion of the active phase. The influence of
activation over the HDS of dibenzothiophene (DBT) on CoMo/Al O ,
2
3
NiMo/Al O , Ru/MgF materials has been investigated extensively
2
3
2
[15–23].
The influence of atmosphere during the decomposition of
ammonium tetrathiomolybdate for the activation of unpromoted
and cobalt-promoted alumina-supported MoS2 catalysts was also
investigated by Pawelec et al. [24]. This study showed a strong
influence of the activation procedure. For the unpromoted cata-
lyst, the most efficient treatment was related to the use of a 10%
∗ Corresponding author. Tel.: +52 646 174 46 02.
E-mail address: galonso@cnyn.unam.mx (G. Alonso-Nú n˜ ez).
H S/H2 atmosphere while the cobalt-promoted catalyst needed
2
0
926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2012.01.015

9
6
G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
Table 1
a more complex activation combining first a calcination treat-
List of catalysts synthesized with different activation atmospheres.
ment followed by successive H S/H2 and H2 treatments. However,
2
in the case of CoMo/SBA-15 catalytic systems, the influence of
atmosphere and temperature used during the activation of this
type of catalysts has not been reported in the literature. More-
over, the control of the dispersion of the active phase inside the
SBA-15 channels remains challenging [25]. The classical wetness
impregnation based on the use of ammonium heptamolybdate and
cobalt nitrate was not found to be successful in order to locate
MoS2 slabs inside the SBA-15 mesopores [26,27]. An alternative
solution was then envisaged by Vradman et al. [28] leading to
a good dispersion of WS2 slabs but using a complicated proce-
Atmosphere
Temperature (K)
723
773
823
Set 1
Set 2
N2/H2 (H2 = 10%)
H2S/H2 (H2S = 10%)
S1A
S2A
S1B
S2B
S1C
S2C
2
.2. Activation of Co-Mo/SBA-15
Activation was performed by heating ex situ the as-obtained
solids under a N /H2 (10% (v/v) H ) mixture (set 1) or under
2
2
dure based on the sonication of a W(CO) -sulfur-diphenylmethane
6
◦
a H S/H (15% (v/v) H S) mixture (set 2) from 298 K to T sulf
2
2
2
solution with SBA-15. Another interesting approach is based on
the use of already-sulfided precursors. This kind of precursors
is well-known and was envisaged extensively in previous stud-
ies [29–36]. Recent results have also shown that the use of an
already-sulfided precursor, ammonium tetrathiomolybdate (ATM),
was an interesting simple solution to achieve a better dispersion
of MoS2 slabs inside the SBA-15 channels [10] even if progress
is still needed in this respect. Indeed, in this study, ATM was
stirred in aqueous solution with CoSx/SBA-15 solids obtained
by impregnation of cobalt acetate on the SBA-15 support fol-
lowed by sulfidation. This approach led to a limited intimate
mixing between cobalt and molybdenum and therefore to a
non-ideal homogeneous dispersion inside the mesopores. One
possible solution is the use of both already sulfided precursors
of cobalt and molybdenum through a successive mixing of the
SBA-15 support with two different solutions of ATM and cobalt
diethyldithiocarbamate.
◦
(
T sulf = 723, 773 or 823 K) at a heating rate of 4 K/min. Once the
◦
desired value of T sulf was reached, the temperature was held con-
stant for 4 h under the flow of the activating gas mixture. After this
period, the gas was switched from the H S/H2 (15% (v/v) H S) or
2
2
N /H (10% (v/v) H ) mixture to pure N2 flow while the tempera-
2
2
2
ture was decreased to 298 K. The as-obtained catalysts were then
kept under nitrogen atmosphere to prevent oxidation before being
transferred to the batch reactor for HDS reaction. The catalysts are
labeled according to Table 1.
2.3. Sample characterization
Nitrogen adsorption–desorption measurements were carried
out at 77 K on a Quantachrom Autosorb. Samples were degassed
under flowing argon at 573 K for 4 h before nitrogen adsorp-
tion. The surface area measurements were performed according
to the Brunauer–Emmett–Teller (BET) method. XRD patterns were
collected on a Philips X Pert MPD diffractometer at room temper-
The objective of the present study will then be to determine
the optimized conditions for the activation of CoMo/SBA-15 cat-
alysts prepared through a novel approach based on the use of
already-sulfided Co and Mo precursors. After contacting SBA-15
with cobalt diethyldithiocarbamate and ATM, two sets of catalysts
were then activated varying atmosphere and temperature. These
ature using monochromatized Cu K␣ radiation (ꢀ = 1.54056
A˚ ) in
◦
a 10–70 2ꢁ range. FESEM was used to observe any morphologi-
cal variation in catalysts due to the applied gas flow used during
the activation procedure. Scanning electron micrographs were
obtained with a JSM-7401F microscope (cold FE, 30 kV). The trans-
mission electron microscopy (TEM) used was a JEOL JEM 2200FS
microscope (200 kV). The samples for FESEM and TEM analysis were
prepared by dispersing the powder products as a slurry in acetone
before drying it on a Cu grid.
catalysts were characterized by N -physisorption, X-ray Diffraction
2
(
XRD), scanning electron microscopy with field emission (FESEM)
and high-resolution transmission electron microscopy (HRTEM).
the sulfide catalysts were tested in the hydrodesulfurization of
DBT. This reaction was selected for catalyst screening purposes
because (1) DBT is a typical sulfur-containing hydrocarbon present
in the petroleum fraction of high-boiling oil or coal derived liquids
2
.4. Catalytic activity and selectivity
[
2] and (2) because of the absence of any diffusion limitation for
DBT molecule to entry into the porous structure of the SBA-15-
supported catalysts [28].
HDS of DBT was carried out in a Parr model 4520 high-pressure
batch reactor. 1.0 g of the catalyst and 150 mL of the freshly pre-
pared solution of DBT in decalin (5%, w/w) were introduced into
the reactor. The reactor was then purged and pressurized to 3.4 MPa
2
. Experimental
(
490 psi) with hydrogen and then heated up to 623 K with contin-
uous stirring at 600 rpm. The reaction was followed during 5 h by
analyzing samples every 30 min using a PerkinElmer XL gas chro-
matograph equipped with an OV-17 packed column.
2
.1. Support and catalyst preparation
The SBA-15 support material was prepared following the
The main reaction products for the HDS of DBT are biphenyl
synthesis route reported in reference [18]. The metal based com-
position for each catalyst was set to 12 wt% of Mo and 3 wt% of
Co. 2 g of SBA-15 powder (sieved between 80 and 120 ␮m) was
first stirred for 4 h in 100 mL of a saturated chloroform solution of
cobalt diethyldithiocarbamate, CoS CN(C H5) (CoDETC). The sol-
vent was then removed by filtration and the obtained product was
dried under vacuum overnight.
The dried product was then stirred in 150 mL of a saturated
aqueous solution of ammonium tetrathiomolybdate at 298 K. After
(
BP) formed by direct C S bond cleavage of DBT (the so-called
direct desulfurization pathway, DDS) and cyclohexylbenzene (CHB)
formed by an initial hydrogenation of one of the aromatic rings of
tetrahydrodibenzothiophene (THDBT) followed by C S bond rup-
ture (the hydrogenating pathway, HYD). The HYD/DDS selectivity
ratio is based on the product concentration ratio according to Eq.
2
2
2
(
1).
HYD/DDS = (THDBT + CHB)/BP
(1)
4
h of stirring, the product was filtered without washing and then
The rate constant was determined from the DBT conversion as
function of time assuming DBT conversion as a pseudo zero-order
reaction [37].
activated. The precursor ATM was synthesized according to the syn-
thesis method reported in [10] and the CoDETC was obtained from
Aldrich.

G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
97
4
3
2
1
00
00
00
00
8
6
4
2
00
00
00
00
0
SBA-15
S1A
S1B
S2C
S2B
S2A
S1C
0
.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
0.4
0.6
0.8
1.0
P/Po
P/Po
Fig. 1. Nitrogen adsorption–desorption isotherms of the SBA-15 support and of N2/H2 activated CoMo/SBA-15 catalysts (set 1) (a) and of H2/H2S activated CoMo/SBA-15
catalysts (set 2) (b).
3
3
3
. Results and discussion
The surface area of the SBA-15 support reaches a very high
value of 992 m /g. After introduction of Co and Mo, the surface
2
.1. Structural characterization of the CoMo/SBA-15 catalysts
area decreases showing that porosity was partly blocked by the
filling of the mesopores by Co/MoS2 particles. However, specific
2
.1.1. Textural properties
surface areas remain high ranging between 400 and 470 m /g. This
Nitrogen adsorption–desorption isotherms of the SBA-15 sup-
effect can also be more clearly observed if considering surface areas
per gram of support. In this case, values ranging between 540 and
624 m /gsupport are obtained showing a decrease of about 40–45%
port and of the two sets of catalysts activated under N /H2 (set 1)
2
2
or H /H S (set 2) are shown in Fig. 1. All samples exhibit the typical
2
2
type-IV isotherms for mesoporous materials with cylindrical pores.
The SBA-15 support shows a typical type H1 adsorption–desorption
hysteresis loop characteristic for cylindrical pore channels. The loop
seems particularly developed in this case indicating the presence
of a large amount of mesopores [38].
of the surface area of the support due to pore filling. The decrease
in surface area also seems slightly more important for H /H S
activated samples leading to lower normalized values. This was
confirmed when analyzing the evolution of pore volumes and pore
diameters. While the SBA-15 support presents average pore diam-
eter of about 7 nm, a shift to smaller pore diameters was observed
2
2
The N2 adsorption–desorption isotherms of the CoMo/SBA-15
catalysts are substantially different from the ones observed for the
SBA-15 support alone even if type IV isotherms are still obtained
showing that the mesoporosity was preserved after introduction of
Co and Mo. This was observed for both N /H and H /H S activated
for both H /N2 or H /H S activated samples with similar values
2
2
2
around 4 nm. This result confirms the location of Co/MoS particles
2
inside the mesopores of the SBA-15 support in both cases. Simi-
larly, the use of the N /H2 atmosphere leads to slightly higher pore
2
2
2
2
2
samples. In both cases, the shape of the hysteresis loop changes and
corresponds to the formation of ink-bottle pores. This effect is char-
acteristic of a percolation phenomenon caused by small Co/MoS2
particles settling within the mesopores [39]. Moreover, a shift to
smaller P/P0 values for the onset of the hysteresis loop is observed
indicating the presence of smaller pore diameters reflecting the
filling of pores by sulfide particles. Table 2 reports the specific sur-
face areas, pore volumes and pore diameter values of the SBA-15
support alone and of the different activated samples. Surface areas
normalized per weight of silica and the normalized pore values are
also reported in Table 2.
volumes for a given temperature of treatment. Considering these
data, it can be observed that by using the N /H2 atmosphere dur-
2
ing activation, similar or even slightly better textural properties of
CoMo/SBA-15 catalysts could be obtained than using the traditional
H /H S atmosphere.
2
2
3.1.2. X-ray diffraction
Fig. 2 shows the XRD patterns for set 1 (Fig. 2A) and for set 2
(Fig. 2B). All the XRD patterns present broad peaks corresponding
to the poorly crystalline 2H-MoS phase with signals at 2ꢁ values
2
◦
◦
◦
◦
of 14 , 33 , 40 and 58 corresponding to the (0 0 2), (1 0 0), (1 0 3)
and (1 1 0) diffraction peaks. The amorphous silica leads to a broad
peak at 2ꢁ = 22 . The different sets of catalysts exhibit relatively
◦
Table 2
similar crystallinity whatever the atmosphere used during activa-
tion. Only a moderate increase in peak intensities can be observed
with increasing temperature of treatment. However, this increase
seems similar either after activation under H /H S or under N /H .
This point was confirmed through the determination of the stack-
ing degree of MoS2 slabs using the Scherrer equation applied to
the (0 0 2) diffraction peak. Results are reported in Table 3. Stack-
ing degree values vary in a narrow range between 4.6 and 5.1
for all the studied samples. A very moderate increase of the slab
stacking with the temperature of treatment is observed whatever
the mode of activation. However, this increase is more marked for
Specific surface areas (SSA), pore volumes (Vp) and pore diameter (D) values of the
SBA-15 support and of N2/H2 activated (set 1) and H2/H2S activated (set 2) SBA-15-
supported CoMo catalysts. The values given into parentheses for the surface areas
are the values per gram of support instead of per gram of catalyst.
2
2
2
2
3
˚
Catalyst
SSA
Vp (cm /g)
D (A)
2
3
(
m /g)
Normalized
(cm /g)
Normalized
SBA-15
S1A
S1B
S1C
S2A
992
1.0
1.2
1.0
71
37
42
37
40
38
40
432 (573)
472 (624)
457 (604)
408 (540)
466 (616)
405 (536)
0.58
0.63
0.61
0.54
0.62
0.54
0.52
0.54
0.52
0.46
0.49
0.47
0.57
0.60
0.57
0.51
0.54
0.52
S2B
S2C
N /H activated samples leading to slightly higher stacking degrees
2
2
after treatment at 773 and 823 K. It should also be noted that this

9
8
G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
A
S1A
*
2H-MoS2
C: S1A
o Co S
9 8
*
*
2H-MoS2
(
002)
*
*
ο
D
ο
*
*
S1C
(100)
*
*
(
103)
*
A
After HDS
(
110)
*
S1B
S1A
Before HDS
1
0
20
30
40
50
60
70
20
30
40
50
60
70
2−
Theta (°)
2−
Theta (°)
D: S2A
B
o
9
Co S8
*
2H-MoS2
*
2H-MoS2
*
*
ο
ο
(
002)
*
(
100)
*
*
ο
(103)
ο
S2C
S2B
*
*
(
110)
*
After HDS
S2A
Before HDS
1
0
20
30
40
50
60
70
15
20
25
30
35
40
45
50
55
60
65
70
75
2−
Theta (°)
2−Theta (°)
Fig. 2. XRD patterns (acquired before HDS) of the CoMo/SBA-15 samples obtained after activation under N2/H2 (set 1) (A) or under H2/H2S (set 2) (B). Comparison of the XRD
patterns before and after HDS for the S1A sample activated under N2/H2 at 723 K (C) and for the S2A sample activated under H2/H2S at 723 K (D).
Table 3
cases, diffraction peaks of the Co9S8 phase can be detected after
HDS reaction. However, the crystallinity of the cobalt sulfide phase
appears much more marked for the S2A sample activated under
Average stacking degrees of MoS2 slabs (determined from XRD patterns) and slab
length (determined by TEM) of the N2/H2 activated (set 1) and the H2/H2S activated
(
set 2) SBA-15-supported CoMo catalysts. The average stacking is determined by
applying the Scherrer equation to the (0 0 2) MoS2 diffraction peak and considering
a value of 6.17 A˚ for the interlayer distance.
H2/H2S. This suggests that a higher proportion of cobalt was not
initially in interaction with the edges of MoS2 slabs leading after
HDS to a higher tendency to sintering. Therefore, the activation
under H /H S seems to lead to CoMo catalytic systems less stable
Catalyst
Average stacking degree (XRD)
Average slab length (nm, TEM)
2
2
S1A
S1B
S1C
S2A
S2B
S2C
4.6
4.9
5.1
4.7
4.7
4.9
5.2
10.3
20.1
9.5
10.0
7.3
under HDS reaction conditions.
3.1.3. Scanning electron microscopy
The morphology of the activated catalysts was also character-
ized by FE-SEM. The morphology of the CoMo/SBA-15 catalysts
appears agglomerated (Fig. 3). Compared to the SBA-15 support
alone, this could be in line with the decrease of surface area
observed by N2 physisorption measurements (Table 2). However,
some differences can also be noticed when comparing the 723 K-
increase in stacking remains weak taking into account the rela-
tively high temperature of treatment used herein. This would be
in agreement with a localization of the MoS2 slabs inside the SBA-
2
activated S1A and S2A samples obtained respectively from a N /H2
1
5 mesoporous material restraining their growth along the [0 0 1]
or a H /H S treatment. For the S1A sample (Fig. 3A), the charac-
direction. Finally, the absence of diffraction peaks due to cobalt
sulfide suggests that the promoter atoms are well dispersed for the
different sets of catalysts.
2
2
teristic well-ordered hexagonal arrays of honeycomb-like SBA-15
mesopores can be clearly observed. On the opposite, a denser mor-
phology was found for the S2A sample (Fig. 3B) while the support
porosity is not detected anymore. This result suggests that the
Fig. 2C and D shows respectively a comparison of the XRD
patterns for CoMo/SBA-15 catalysts activated at the lowest tem-
perature (723 K) using either the N /H (S1A; Fig. 2C) or the H /H S
N /H treatment was less harmful for preserving an ordered poros-
2
2
2
2
2
2
ity of the CoMo/SBA-15 catalysts than the H /H S activation.
atmosphere (S2A, Fig. 2D) before and after HDS reaction. In both
2
2

G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
99
Fig. 3. Scanning electron micrographs of (A) the S1A sample activated under N2/H2 at 723 K and (B) the S2A sample activated under H2/H2S at 723 K.
Table 4
3.1.4. Transmission electron microscopy
Activity, selectivity (HYD/DDS ratios), and turnover frequency values in the
hydrodesulfurization of dibenzothiophene for the N2/H2 activated (set 1) and the
H /H S activated (set 2) SBA-15-supported CoMo catalysts.
The CoMo/SBA-15 catalysts activated using either N /H2 or
2
H /H S atmosphere have also been characterized by transmission
2
2
2
2
electron microscopy. Fig. 4a and b presents TEM pictures acquired
respectively for the 723 K activated samples obtained under N /H2
−
−1
)
Catalyst
k (specific)
HYD/DDS ratio
TOF (10 2
s
2
−7
−1 −1
g )
(
1 × 10 mol s
(
S1A) or under H /H S (S2A). The SBA-15 channels with a diam-
2 2
S1A
S1B
S1C
S2A
S2B
S2C
19.2
16.8
15.0
12.3
13.9
16.4
0.25
0.25
0.23
0.18
0.21
0.23
2.6
4.2
7.1
2.8
3.4
3.0
eter of about 5 nm can be clearly discerned on TEM pictures. The
distance between two consecutive centers of hexagonal opening
of channels, as estimated from TEM images, is close to the average
pore size obtained by N adsorption measurements. The MoS slabs
2
2
are in their vast majority located in an homogeneous way inside
the channels of the SBA-15 support whatever the post-treatment
atmosphere suggesting that the use of successive impregnations
of the cobalt diethyldithiocarbamate precursor and of ATM are an
efficient way to distribute the Co/MoS2 particles inside the SBA-15
mesoporosity. Only few slabs can be detected outside the SBA-15
channels (Fig. 4a).
smaller average slab length than those obtained under H /H S
2
2
(
5.2 nm vs 9.5 nm) (Table 3). After activation at 773 K, similar slab
lengths are obtained whatever the atmosphere used to activate
the CoMo/SBA-15 catalysts. While after activation under N /H ,
the size doubled when increasing the temperature of treatment by
0 K for 723–773 K, the slab dimensions of H /H S activated sam-
ples remain relatively constant. Similar conclusions can be reached
after activation at 823 K. While the slab length of the CoMo/SBA-
5 catalyst activated under N /H2 still doubled compared to the
similar solid but activated at 773 K, the slab length of the H /H S
activated samples even slightly decreased. However, it should be
kept in mind that, except for the S1C sample (N /H ; 823 K), the
slab lengths measured in the present study appear smaller than
those reported previously using ATM but with cobalt acetate [10].
A propensity to a better dispersion (particularly if activated under
2
2
This result is interesting since it shows the ability to obtain
homogeneously cobalt-promoted MoS2 nanoparticles inside the
SBA-15 channels. Compared to previous studies [10,25], this shows
the interest of using ATM solution to disperse homogeneously
MoS2 nanoparticles inside SBA-15 channels if an appropriate
post-treatment is used contrary to the classical heptamolybdate
impregnation approach [26]. This also represents an improvement
compared to our previous study [10] for which cobalt acetate was
used leading to an inhomogeneous dispersion inside this meso-
porous material. Apparently, the use of an already-sulfided cobalt
precursor is able to enhance the final dispersion of the active phase.
Statistics about the slab length were carried out for each
CoMo/SBA-15 catalyst activated at 723 K, 773 K or 823 K using
either a N /H or H /H S atmosphere. It should be noted here that
5
2
2
1
2
2
2
2
2
H /H S) is therefore observed if using the already sulfided cobalt
2
2
diethyldithiocarbamate. When activated under N /H , the sinter-
ing of the MoS slabs then occurs only along the slab direction while
preceding XRD results did not show any significant increase of the
stacking degree. This 1D sintering process indirectly confirms the
localization of the MoS2 slabs inside the SBA-15 channels.
2
2
2
2
2
2
2
only slab length estimation was performed directly on TEM pictures
and not the evaluation of the stacking degree. Indeed, in previous
studies [6,40], synchrotron XRD (SR-XRD) techniques were found
by some of us to be the most efficient method to determine the com-
plete morphology (slab length, stacking degree) of supported MoS2
nanoparticles. Moreover, comparison performed between SR-XRD
results and estimation of particle size from TEM images led us to
conclude that direct measurements on micrographs are a reliable
approach for evaluating slab length along the basal direction. How-
ever, this comparison also emphasized a strong overestimation of
the stacking degree when using only TEM statistics. This is related
to the ability of TEM to detect about only 10% of the MoS2 particles
present on a support leading to a selective detection of the most
stacked slabs [41].
3.2. Catalytic activity in HDS reaction
Table 4 reports the DBT HDS activity of the different CoMo/SBA-
15 catalysts obtained by activation under N /H2 or H /H S at the
2
2
2
different temperatures of treatment (723, 773, and 823 K). The
N /H activated catalysts exhibit higher activities than their H /H S
2
2
2
2
activated counterparts at low temperatures of treatment (723 and
773 K). This is particularly the case for the S1A sample obtained
−
7
−1 −1
g ) with
by activation under N /H2 at 723 K (19.2 × 10 mol s
2
an activity 56% higher than for the same catalyst but activated
under H /H S. Comparison to XRD results suggests that this higher
Results show that after activation at 723 K, the Co/MoS2
2
2
nanoparticles obtained after treatment under N /H2 present
activity is partly due to a higher stability of the cobalt-promoted
2

1
00
G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
Fig. 4. Transmission electron micrographs of (A) the S1A sample activated under N2/H2 at 723 K and (B) the S2A sample activated under H2/H2S at 723 K.
phase under HDS conditions after the N /H treatment leading to a
smaller segregation effect. However, with the increase of the tem-
perature of treatment, the activity tends to decrease moderately for
to the formation of Co/MoS2 particles outside the SBA-15 channels
or to a poor dispersion of the active phase due to a restrained filling
of the mesopores when using Mo oxide precursors and/or cobalt
nitrate. The replacement of ammonium heptamolybdate by ATM
led to the synthesis of CoMo/SBA-15 catalysts better dispersed into
the SBA-15 channels but still inhomogeneously [10]. As a result,
the activity of CoMo/SBA-15 catalysts (with 10 wt% Mo and 2 wt%
2
2
the N /H activated solids. On the opposite, a 30% increase in activ-
2
2
ity is observed for the H /H S activated catalysts when increasing
2
2
the temperature from 723 K to 823 K. Therefore, after activation at
73 K, the N /H activated solid is only 20% more active than the
7
2
2
−
7
−1 −1
g using SBA-15
H /H S activated catalyst while after activation at 823 K, the H /H S
Co) increases up to 13.1 and 13.7 × 10 mol s
2
2
2
2
activated CoMo/SBA-15 becomes more active.
with respectively pore diameters of 6 and 9 nm, therefore present-
ing activities similar to the H /H S activated S2A and S2B samples.
With the increasing temperature of treatment, the decrease in
2
2
activity of the N /H2 activated CoMo/SBA-15 catalysts reflects the
Combination of two already sulfided Mo and Co precursors (ATM
and cobalt diethyldithiocarbamate) in the present study led to a
substantially higher HDS activity if an optimized N /H activation
procedure is used. This could be related to the better dispersion of
the active phase inside SBA-15 channels using these sulfided pre-
cursors allowing to take benefit in a larger extent of the high surface
mesoporous SBA-15 material.
2
progressive formation of bigger MoS2 slabs resulting in a lower
proportion of Mo edge active sites. Similarly, the higher activ-
ity of the S2C sample activated at 823 K in comparison with the
same H /H S activated catalysts but treated at lower tempera-
2
2
2
2
tures mainly reflects a smaller average slab length and therefore
a higher percentage of edge active sites. This point can be more
clearly emphasized if calculating turnover frequency values tak-
ing into account the number of Mo edge active sites as determined
The analysis of selectivity results expressed as HYD/DDS ratios
shows relatively similar values whatever the samples ranging
between 0.18 and 0.25 (Table 4). These results are in agreement
with those obtained by Huang et al. [10] on CoMo/SBA-15 cata-
lysts obtained by using ATM and cobalt acetate but differ from
those acquired by Nava et al. [26] on similar CoMo/SBA-15 cata-
lysts but prepared using ammonium heptamolybdate and cobalt
nitrate (HYD/DDS = 0.40). Therefore, in the present case, the as-
obtained CoMo/SBA-15 catalysts exhibit a depleted HYD activity.
These results could be first analyzed taking into account the Rim-
Edge model [43]. According to this model, the DBT HDS selectivity
along the DDS route leading to biphenyl or along the HYD route
leading to cyclohexylbenzene is controlled by the stacking degree
of the MoS2 nanoparticles, the more the nanoparticles are stacked,
the less the resulting catalyst will be hydrogenating. The reason
for such a selectivity variation would be related to the accessi-
bility of edge sites to the DBT molecule. The pre-requisite for the
hydrogenation of the phenyl ring of the DBT molecule would be
a sterically demanding ␩6 flat adsorption mode possible only on
external layers of MoS2 particles (called rim sites) while the steric
hindrance of internal layers would lead to an absence of hydro-
genation on these sites (called edge sites). On the opposite, the
from the MoS hexagonal model developed by Kasztelan et al. [42].
2
In this case, relatively similar TOF values were obtained for the set
of H /H S activated CoMo/SBA-15 catalysts. This would be in agree-
2
2
ment with the fact that the increase in activity only results from a
higher number of edge active sites and not to any improvement of
the intrinsic activity per site, i.e. the quality of these sites. Interest-
ingly, the determination of the TOF values for the N /H2 activated
2
catalysts shows an opposite evolution to the H /H S activation case.
2
2
A strong increase in TOF values can be observed when increasing
the temperature of treatment from 723 K to 773 K and finally to
−
2
−1
8
23 K with values respectively of 2.6, 4.2, and 7.1 × 10
s . There-
fore, while active sites are not more intrinsically active for the
N /H activated catalyst after treatment at 723 K compared to the
2
2
H /H S activation case, the use of an N /H atmosphere at higher
2
2
2
2
temperatures of treatment to activate our CoMo/SBA-15 catalysts
can substantially improve the intrinsic activity per site allowing to
partly compensate the lower percentage of edge active sites due
to sintering of the active phase. These combined effects lead to a
lower loss in activity than expected if considering only the increase
of the MoS2 particle size.
1
These results are also interesting if compared to our previous
attempts to prepare CoMo/SBA-15 catalysts. In this respect, Nava
et al. [26] tried initially to prepare CoMo/SBA-15 starting from
ammonium heptamolybdate and cobalt nitrate. The DBT HDS activ-
pre-requisite for the C S bond rupture step would be the ␩ (S)
adsorption mode possible on both rim and edge sites.
In the present case, the similar HYD/DDS ratios obtained what-
ever the samples would be in agreement with the absence of
significant variation of the stacking degree. However, a deeper anal-
ysis of these results can also be made. Indeed, whatever the sample
considered, a depleted HYD selectivity is obtained which appeared
−
7
−1 −1
g ) (similar Co and Mo loadings)
ity obtained (10.8 × 10 mol s
appears significantly lower than in the present case even if com-
pared to the less active S2A catalyst. One possible reason was due

G. Alonso-Nú ˜n ez et al. / Applied Catalysis A: General 419–420 (2012) 95–101
101
particularly marked if compared to results previously acquired on
industrial CoMo/Al O catalysts (0.49) [26]. As already observed by
Huang et al. [10] and Nava et al. [26], this depleted HYD character
is in fact related to a confinement effect of Co/MoS2 nanoparti-
cles inside the mesoporosity of the SBA-15 support. Indeed, the
localization of the particles inside the mesopores hinders the pos-
[2] H. Topsøe, B.S. Clausen, F.E. Massoth, in: J.R. Anderson, M. Boudart (Eds.),
Hydrotreating Catalysis and Science and Technology, vol. 11, Springer-Verlag,
Berlin/Heidelberg, 1996.
2
3
[
3] C.N. Satterfield, Heterogeneous Catalysis in Industrial Practice: Catalyst
Preparation and Manufacture, second edition, Krieger Publishing Co., 1996,
pp.87–129.
[
4] H. Shimada, Catal. Today 86 (2003) 17–29.
[
5] Y. Sakashita, Y. Araki, H. Shimada, Appl. Catal. A: Gen. 215 (2001) 101–110.
sibility for the DBT molecule to adsorb in a sterically demanding
[6] G. Berhault, M. Perez De la Rosa, A. Mehta, M. José-Yacaman, R.R. Chianelli,
Appl. Catal. A: Gen. 345 (2008) 80–88.
6
␩
flat adsorption and therefore to be hydrogenated leading to an
[
7] M. Breysse, G. Berhault, S. Kasztelan, M. Lacroix, F. Maugé, G. Pérot, Catal. Today
6 (2001) 15–22.
enhancement of the DDS selectivity. This effect observed for SBA-
5 supported catalysts was also observed systematically for MoS2
44,45], Co/MoS2 [36,46] and Ni/MoS2 [47] nanoparticles loaded in
carbon mesopores. The development of a carbon mesoporous orga-
6
1
[
[8] Y.V. Joshi, P. Ghosh, M. Daage, W.N. Delgass, J. Catal. 257 (2008) 71–80.
[9] I. Eswaramoorthi, V. Sundaramurthy, N. Das, A.K. Dalai, J. Adjaye, Appl. Catal.
A: Gen. 339 (2008) 187–195.
10] Z.-D. Huang, W. Bensch, L. Kienle, S. Fuentes, G. Alonso, C. Ornelas, Catal. Lett.
122 (2008) 57–67.
[
nization led to depleted HYD selectivity whatever the MoS -based
2
catalysts to be formed.
[11] Y. Saih, K. Segawa, Appl. Catal. A: Gen. 353 (2009) 258–265.
[
[
12] M. Breysse, P. Afanasiev, C. Geantet, M. Vrinat, Catal. Today 86 (2003) 5–16.
13] T. Chiranjeevi, P. Kumar, S.K. Maity, M.S. Rana, G. Murali Dhar, T.S.R. Prasad Rao,
Microporous Mesoporous Mater. 44 (2001) 547–556.
In this respect, it is interesting to observe that similar HYD/DDS
ratios were obtained by Huang et al. [10] for which cobalt-
promoted MoS2 nanoparticles were mainly located inside the
SBA-15 channels even if inhomogeneously dispersed. Similarly, the
presence of Co-promoted MoS2 nanoparticles outside the SBA-15
channels [26] could explain the higher hydrogenating character
observed for CoMo/SBA-15 catalysts prepared using Mo oxide pre-
cursors and cobalt nitrate.
[14] G. Pérot, Catal. Today 86 (2003) 111–128.
[
[
15] R. Prada Silvy, F. Delannay, P. Grange, B. Delmon, Polyhedron 5 (1986) 195–198.
16] M. Wojciechowska, M. Pietrowski, B. Czajka, S. Łomnicki, Catal. Lett. 87 (2003)
153–157.
[17] R.D. Mi c´ i c´ , R.P. Marinkovi c´ -Nedu cˇ in, Z. Schay, I. Nagy, M. Hadnadev, E.E. Kiss,
Rev. Roum. Chim. 53 (2008) 629–634.
[
[
18] Z.D. Huang, W. Bensch, A. Lotnyk, L. Kienle, S. Fuentes, J. Bocarando, G. Alonso,
C. Ornelas, J. Mol. Catal. A: Chem. 323 (2010) 45–51.
19] S. Texier, G. Berhault, G. Pérot, V. Harlé, F. Diehl, J. Catal. 223 (2004)
4
04–418.
4
. Conclusions
[
[
20] S. Texier, G. Berhault, G. Pérot, F. Diehl, Appl. Catal. A: Gen. 293 (2005) 105–119.
21] J. Van Gestel, J. Léglise, J.C. Duchet, J. Catal. 145 (1994) 429–436.
[22] Y. Okamoto, K. Hioka, K. Arakawa, T. Fujikawa, T. Ebihara, T. Kubota, J. Catal.
268 (2009) 49–59.
23] H. Farag, A.N.A. El-Hendawy, K. Sakanishi, M. Kishida, I. Mochida, Appl. Catal.
B: Environ. 91 (2009) 189–197.
[24] B. Pawelec, R. Navarro, J.L.G. Fierro, P.T. Vasudevan, Appl. Catal. A: Gen. 168
(1998) 205–217.
25] E.J.M. Hensen, P.J. Kooyman, Y. Van der Meer, A.M. van der Kraan, V.H.J. de Beer,
J.A.R. van Veen, R.A. van Santen, J. Catal. 199 (2001) 224–235.
26] R. Nava, R.A. Ortega, G. Alonso, C. Ornelas, B. Pawelec, J.L.G. Fierro, Catal. Today
127 (2007) 70–84.
27] D. Valencia, T. Klimova, Catal. Today 166 (2011) 91–101.
28] L. Vradman, M.V. Landau, M. Herskowitz, V. Ezersky, M. Talianker, S. Nikitenko,
Y. Koltypin, A. Gedanken, J. Catal. 213 (2003) 163–175.
The synthesis of CoMo/SBA-15 catalysts was herein per-
formed using
a novel approach using already-sulfided Mo
[
and Co precursors, ammonium tetrathiomolybdate and cobalt
diethyldithiocarbamate allowing to homogeneously disperse a
high proportion of cobalt-promoted MoS2 nanoparticles inside
SBA-15 channels. Optimization of the activation parameters (tem-
perature of treatment, atmosphere) was subsequently carried out
in order to enhance catalytic properties. Compared to the classical
[
[
[
[
H /H S atmosphere, the use of a N /H atmosphere allows to obtain
2
2
2
2
better stabilized Co/MoS2 nanoparticles for which cobalt segrega-
tion under hydrodesulfurization conditions is limited. Moreover,
depending on the atmosphere used, the temperature of activation
differently influences the morphology of the active phase leading
to a rapid monodimensional sintering at high temperature under
N /H while its influence is hardly noticed under H /H S. The
combination of using already sulfided Co and Mo precursors with
the optimization of the activation procedure led to highly active
CoMo/SBA-15 catalysts. Finally, the localization of cobalt-promoted
[29] K. Ramanathan, S.W. Weller, J. Catal. 95 (1985) 249–259.
30] P.T. Vasudevan, S.W. Weller, J. Catal. 99 (1986) 235–238.
31] S. Fuentes, G. Diaz, F. Pedraza, H. Rojas, N. Rosas, J. Catal. 113 (1988)
[
[
5
35–539.
[32] J. Cruz-Reyes, M. Avalos-Borja, M.H. Farias, S. Fuentes, J. Catal. 137 (1992)
32–242.
2
2
2
2
2
[
[
33] P.T. Vasudevan, F. Zhang, Appl. Catal. A: Gen. 112 (1994) 161–173.
34] K. Wilkinson, M.D. Merchán, P.T. Vasudevan, J. Catal. 171 (1997) 325–328.
[35] Y. Yoneyama, C. Song, Catal. Today 50 (1999) 19–27.
[36] H. Nava, C. Ornelas, A. Aguilar, G. Berhault, S. Fuentes, G. Alonso, Catal. Lett. 86
(
2003) 257–265.
MoS nanoparticles inside the mesopores of the SBA-15 support led
to confinement effects and to more direct desulfurization selective
catalytic systems.
2
[
37] R. Huirache-Acu n˜ a, M.A. Albiter, C. Ornelas, F. Paraguay-Delgado, R. Martínez-
Sánchez, G. Alonso-Nu n˜ ez, Appl. Catal. A: Gen. 308 (2006) 134–142.
[38] F.J. Brieler, P. Grundmann, M. Fröba, L.M. Chen, P.J. Klar, W. Heimbrodt, H.A.K.
von Nidda, T. Kurz, A. Loidl, J. Am. Chem. Soc. 126 (2004) 797–807.
[
39] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199–218.
Acknowledgments
[40] M. Perez De la Rosa, S. Texier, G. Berhault, A. Camacho, M.J. Yacaman, A. Mehta,
S. Fuentes, J.A. Montoya, F. Murrieta, R.R. Chianelli, J. Catal. 225 (2004) 288–299.
[
41] R.R. Chianelli, A.F. Ruppert, M.J. Yacaman, A. Vazquez-Zavala, Catal. Today 23
1995) 269–281.
42] S. Kasztelan, A. Wambeke, L. Jalowiecki, J. Grimblot, J.P. Bonnelle, J. Catal. 124
(1990) 12–21.
The authors acknowledge CONACyT for financial support
Project 26067), and PAPIIT project IN102509-3, the National Nano-
technology Laboratory (NANOTECH), Chihuahua, México and Red
de Nanociencias y Nanotecnología, the valuable technical assis-
tance from W. Antunez, C. Ornelas, and E. Torres and E. Aparicio.
(
(
[
[
[
43] M. Daage, R.R. Chianelli, J. Catal. 149 (1994) 414–427.
44] G. Alonso, G. Berhault, A. Aguilar, V. Collins, C. Ornelas, S. Fuentes, R.R. Chianelli,
J. Catal. 208 (2002) 359–369.
[
45] G. Alonso, M.H. Siadati, G. Berhault, A. Aguilar, R.R. Chianelli, Appl. Catal. A:
Gen. 263 (2004) 109–117.
References
[
46] L. Alvarez, J. Espino, C. Ornelas, J.L. Rico, M.T. Cortez, G. Berhault, G. Alonso, J.
Mol. Catal. A: Chem. 210 (2004) 105–117.
◦
[47] L. Alvarez, G. Berhault, G. Alonso, Catal. Lett. 125 (2008) 35–45.
[1] CE bulletin n 715/2007 of the European parliament and council, 20/06/07.