Effect of the support on the high activity of the (Ni)Mo/ZrO 2-SBA-15 catalyst in the simultaneous hydrodesulfurization of DBT and 4,6-DMDBT

Article abstract of DOI:10.1016/j.jcat.2011.04.001

Series of Mo- and NiMo-catalysts were supported on ZrO₂, Al ₂O₃, SBA-15, and ZrO₂-modified SBA-15 and tested in the simultaneous hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothioph

Full text of DOI:10.1016/j.jcat.2011.04.001

Journal of Catalysis 281 (2011) 50–62
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Effect of the support on the high activity of the (Ni)Mo/ZrO –SBA-15 catalyst
2
in the simultaneous hydrodesulfurization of DBT and 4,6-DMDBT
1
⇑
Oliver Y. Gutiérrez , Tatiana Klimova
Facultad de Química, Departamento de Ingeniería Química, Universidad Nacional Autónoma de México, Cd. Universitaria, Coyoacán, México D.F., C.P. 04510, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
Series of Mo- and NiMo-catalysts were supported on ZrO , Al O , SBA-15, and ZrO -modified SBA-15 and
2
2
3
2
Received 26 January 2011
Revised 30 March 2011
Accepted 2 April 2011
Available online 6 May 2011
tested in the simultaneous hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-
dimethyldibenzothiophene (4,6-DMDBT). The rate constants of the main steps in the HDS reaction net-
work of both molecules were calculated, and the materials were characterized by N physisorption,
2
X-ray diffraction, UV–vis diffuse reflectance spectroscopy, temperature-programmed reduction and sulf-
idation, NO adsorption and high-resolution transmission electron microscopy. Tetrahedral and octahe-
Keywords:
SBA-15
ZrO –SBA-15
2
Deep hydrodesulfurization
NiMoS
dral Mo species in the oxide precursors were related to the monolayers and multilayered MoS
2
structures, respectively, in the sulfide catalysts. The morphology of the active phase and the formation
of the NiMoS phase were the most important factors during the HDS process of DBT-type compounds
given that the turnover frequency values did not depend on the support composition or the morphology
MoS
Dibenzothiophene
,6-Dimethyldibenzothiophene
2
2
of the active phase. The monolayers of MoS had low activity for the HDS of both molecules on the unpro-
moted catalysts, whereas on the NiMo catalysts, DBT reacted on monolayers and stacked NiMoS clusters
but 4,6-DMDBT was converted only on the later structures. The optimum active phase–support interac-
4
tion strength on ZrO
total active surface and the active surface not sterically hindered by the support, i.e., stacked MoS
clusters with short lengths. Thus, the NiMo/ZrO –SBA-15 catalyst was able to convert DBT-type com-
pounds with typical and low reactivity on the NiMo/Al catalyst.
Ó 2011 Elsevier Inc. All rights reserved.
2
–SBA-15 materials led to the characteristics of the active phase that maximized the
2
-like
2
2 3
O
1
. Introduction
Thus, the fundamental study of the interaction between the sul-
fur containing molecules and the catalysts has been the aim of
many groups in the last years. The understanding of this issue
would allow designing catalysts able to remove sulfur from
refractory and not refractory compounds in a single reaction
stage.
It is possible to use novel supports, active phases, promoters, or
additives to upgrade the performance of the HDS catalyst, typically
2
MoS supported on alumina and promoted with Co or Ni [2–5]. The
use of an alternative material as carrier is a promising option be-
cause the support has a key effect in the catalytic performance. A
diversity of materials has been tested as HDS supports, i.e., pure
oxides, mixed oxides [6–12], carbon [13], and mesostructured sili-
cas such as MCM-41 and SBA-15 [14–19].
The application of mesostructured silicas as HDS supports is of
special interest due to their outstanding textural characteristics,
e.g., pore volume and surface area [20]. Pure siliceous mesostruc-
tured materials were successfully applied as supports for W and
Mo catalysts in the HDS of dibenzothiophene (DBT) and thiophene
and hydrogenation of cyclohexene [18,21]. It was shown afterward
that adding metal oxides to the mesostructured silica led to more
active catalysts for the same reactions [22]. The use of pure sili-
ceous and modified SBA-15 materials as HDS supports has been
The requirements that a fuel must meet in transport applica-
tions have become more stringent in the last decade. The con-
centration of sulfur, a major impurity in crude oil, has been
limited to few parts per million in many countries. However,
removing sulfur from diesel by the standard hydrodesulfuriza-
tion (HDS) process is a major challenge because of the presence
of low reactivity compounds so-called refractory. The response of
refineries to the demand of high-quality diesel has been to in-
crease the severity of the process conditions or to implement
several reaction stages [1]. Sulfur is removed from the most
reactive species in the first reactor or catalytic bed, whereas
the refractory compounds react in subsequent stages on a sec-
ond catalyst or at different conditions. However, the use of cat-
alysts active in the HDS of refractory compounds is a more
convenient option to improve the entire HDS process because
it minimizes the investment needed to upgrade the product.
⇑
Corresponding author. Fax: +52 55 56225371.
E-mail address: klimova@servidor.unam.mx (T. Klimova).
Present address: Lehrstuhl II für Technische Chemie, Department Chemie,
1
Technische Universität München, Garching, Germany.
0
021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.04.001

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
51
extensively studied by our group. Mo and NiMo catalysts sup-
2.2. Support and catalyst characterization
ported on SBA-15 materials modified with Ti, Zr or Al showed high
activity in the HDS of 4,6-dimethyldibenzothiophene (4,6-DMDBT)
N
2
adsorption–desorption isotherms were measured with a
[
23–26]. NiMo catalysts supported on SBA-15 modified with differ-
ent metal oxides (MgO, CaO, BaO, TiO , or ZrO ) were tested in the
HDS of DBT [27]. The common observation in previous works was
that adding ZrO to the SBA-15 led to the most remarkable
Micromeritics ASAP 2000 automatic analyzer at liquid N temper-
ature. The samples were degassed (p < 10 Pa) at 543 K for 6 h
prior to the experiments. Specific surface areas were calculated
l
by the BET method (SBET), and the micropore area (S ) was esti-
2
ꢁ
1
2
2
2
improvement in the activity of the catalysts in the HDS of both
DBT-type compounds, which was correlated with the morphology
of the active phase.
It was interesting that the same system presented the highest
activity for the HDS of DBT and 4,6-DMDBT because some studies
suggested that improving the activity of the Mo-based catalysts for
the HDS of DBT decreases the activity in the HDS of 4,6-DMDBT
mated using the correlation of t-plot method. XRD patterns of
the oxide precursors were recorded in a Siemens D5000 diffrac-
tometer, using Cu Ka radiation (k = 0.15406 nm) and a goniometer
ꢁ
1
speed of 1°(2
H
) min . UV–vis electronic spectra of the samples
were recorded using
a
Varian Cary 100 spectrophotometer
equipped with a diffuse reflectance attachment. Polytetrafluoro-
ethylene was used as reference. TPR experiments were carried
out in a Micromeritics AutoChem II 2920 automatic analyzer
equipped with a TCD detector. For the TPR experiments, the sam-
ples were pretreated in situ at 673 K for 3 h under synthetic air flow
and cooled down to room temperature under Ar stream. The reduc-
2
and vice versa [28]. For instance, incorporating SiO to the alumina
support decreased the activity in the HDS of DBT but led to higher
activity in the HDS of 4,6-DMDBT [29,30]. Thus, the aim of this
work was to understand the high HDS activity of the (Ni)Mo/
ZrO
2
–SBA-15 system by combining detailed physicochemical char-
2
tion step was performed with an H /Ar mixture with a heating rate
of 10 K min from room temperature to 1273 K.
ꢁ1
acterization of the catalysts with the kinetics of the simultaneous
HDS of DBT and 4,6-DMDBT. Mo- and NiMo catalysts supported
on Al
goal.
For the TPS experiments, a sample of the oxide precursor was
ꢁ
1
2
O
2
2
, ZrO , and SBA-15 were also investigated to achieve the
loaded in a quartz reactor and heated at 5 K min to 673 K in a
mixture of 10 vol.% H S/H . The gases evolved were monitored
2
2
The materials were characterized by N
2
physisorption, X-ray
with a mass spectrometer (QME 200, Pfeiffer Vacuum). After the
TPS experiments, the corresponding XRD patterns of the freshly
sulfided catalysts were recorded in a Philips X’Pert Pro System
diffraction (XRD), UV–vis diffuse reflectance spectroscopy (DRS),
temperature-programmed reduction and sulfidation (TPR and
TPS), high-resolution transmission electron microscopy (HRTEM),
and adsorption of NO. The catalytic performance of the sulfide
catalyst was evaluated in the simultaneous HDS of DBT and 4,
(Cu Ka1-radiation, 0.154056 nm) using a step size of 0.017°(2H)
and 115 s as count time per step.
HRTEM studies of the sulfided catalysts were carried out using a
Jeol 2010 microscope (resolving power 1.9 Å). The solids were
ultrasonically dispersed in heptane, and the resulting suspensions
were collected on carbon-coated grids. The slab length and layer-
6
-DMDBT. The results showed a qualitatively clear relationship be-
tween the oxide and sulfide form of the catalysts explained in
terms of support-active-phase interaction strength. Furthermore,
the analysis provided quantitative evidence of the key role of the
2
stacking distributions of MoS clusters in each sample were deter-
MoS
2
phase morphology in the performance of the catalysts. The
–SBA-15 system is postulated as an alternative
mined from the measurement of at least 300 crystallites detected
on several HRTEM pictures taken from different places of the
microscope grids. The concentration of adsorbed NO on the sulfide
catalysts was determined in a flow apparatus equipped with a
quartz reactor and a mass spectrometer as detector (QME 200,
Pfeiffer Vacuum). Samples of 0.1 g of each oxide precursor were
(
Ni)Mo/ZrO
2
catalyst to remove sulfur from refractory and non-refractory com-
pounds in a single reaction stage.
2
. Experimental
2 2
activated in situ in 10 vol.% H S/H at 673 K for 3 h. After flushing
the reactor with He for 5 h at 300 K, pulses of 10 vol.% of NO/He
were introduced periodically. The concentration of gas adsorbed
was calculated as the sum of the uptakes per pulse.
2.1. Support and catalyst preparation
SBA-15 was synthesized according to the procedure reported in
literature [31]. The triblock copolymer Pluronic P123 (Mav = 5800,
EO20PO70EO20) and tetraethyl orthosilicate (P99%, Aldrich) were
used as structure-directing agent and silica source, respectively.
2.3. Catalytic activity
The catalysts were sulfided ex situ in a tubular reactor at 673 K
ZrO
impregnation of the siliceous SBA-15 with solutions of zirconium
IV) propoxide (70 wt.% in 1-propanol, Aldrich) in dry ethanol (Al-
2
-modified supports were prepared by incipient wetness
for 4 h in a stream of 15 vol.% of H
sure prior to the activity evaluation. The particle size of the sam-
ples was 105–149 m to avoid effects of internal diffusion. The
2 2
S in H under atmospheric pres-
(
l
drich, 99.999%). The solids were dried at room temperature and
treated in static air at 823 K for 5 h after impregnation. In the fol-
simultaneous HDS of DBT and 4,6-DMDBT was performed in a
batch reactor (Parr) at 573 K and 7.3 MPa total pressure for 8 h.
The reactant solution consisted of DBT and 4,6-DMDBT in hexadec-
ane, with sulfur concentration of 1300 and 500 ppm in DBT and
4,6-DMDBT, respectively. The reactor was charged with 40 ml of
the reactant solution and 0.15 g of freshly sulfided catalyst trans-
ferred under inert atmosphere to the reactor in each run. The
course of the reaction was followed by withdrawing aliquots peri-
odically and analyzing them on an HP-6890 chromatograph with a
HP-1 column.
lowing, modified materials are referred as ZrO
the weight% loading of ZrO in the parent SBA-15. Pure ZrO
prepared using ZrOCl (99%, Aldrich) and ammonium
ꢀ8H
hydroxide solution (28% in H O, Sigma–Aldrich) as reported by
Jia et al. [7]. Alumina ( -Al ) was prepared by thermal treatment
2
(x)SBA where x is
2
2
was
2
2
O
2
c
2 3
O
of pseudo-boehmite Catapal B in air at 973 K for 4 h.
Mo and NiMo oxide precursors were prepared by successive
impregnation of aqueous solutions of ammonium heptamolybdate
(
99.98%, Sigma–Aldrich) and nickel nitrate (P98.5%, Sigma–Al-
drich) on the supports. The materials were treated in air at 373 K
for 24 h and then at 773 K for 4 h after each impregnation. The
2.4. Kinetic analysis
nominal compositions were 12.4 wt.% of MoO
and 12 wt.% of MoO and 3 wt.% of NiO for NiMo catalysts. Hereaf-
ter, the catalysts are denoted as (Ni)Mo/support.
3
for Mo catalysts
The HDS of DBT-type compounds occurs through the two paral-
lel reactions illustrated in Fig. 1: direct desulfurization (DDS) yield-
ing the corresponding substituted biphenyl products (species ‘‘b’’)
3

5
2
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
Table 1
(a)
(c)
Textural characteristics of supports and oxides precursors. Surface area (SBET) and
micropore area (S
) in m2
g
ꢁ1
.
l
kHYD
Support
Mo
NiMo
S
BET
S
l
S
BET
S
l
S
BET
S
l
S
S
SBA-15
826
581
505
391
58
135
105
75
65
5.9
602
400
346
283
54
75
55
35
32
2.3
0
543
374
317
244
58
65
32
25
15
1
R
R
R
R
R
R
R
R
ZrO (25)SBA
2
ZrO
ZrO
ZrO
2
2
2
(37)SBA
(50)SBA
kDDS
kDS
c
2
-Al O
3
199
0
191
185
0
○
●
●
●
●
●
●
Mo/ZrO (25)SBA
2
(
b)
(d)
●
●
●
NiMo/SBA-15
●
Fig. 1. Reaction network for the HDS of DBT-type compounds. R@H, for DBT;
R@CH for 4,6-DMDBT.
3
●
○
Mo/SBA-15
and hydrogenation (HYD) with subsequent desulfurization (DS),
yielding first hydrogenated intermediates (species ‘‘c’’) and then
cyclohexylbenzene-type compounds (species ‘‘d’’) [32,33]. Since
the early works of Houalla et al. [34], the assumption of first-order
kinetics has been especially useful in the study of HDS reactions.
Indeed, the first-order expressions fit experimental results better
than the Langmuir–Hinshelwood equations in the HDS of model
molecules (DBT) [35]. Thus, in this work, the results of the activity
tests are analyzed by applying a pseudo-first-order kinetics ap-
proach to the reaction network presented in Fig. 1. Reversibility
in the HYD route and the hydrogenation of the biphenyl com-
pounds are not considered because they are negligible at low con-
versions and high rates of sulfur extraction from the hydrogenated
intermediates [36–39].
○
○
○
□
□ ○
○
○
Mo/ZrO2
Mo/Al O
▲
▲
▲
▲
2
3
10
20
30
40
50
60
70
2
Θ (degree)
Fig. 2. XRD patterns of Mo/Al
ZrO (25)SBA. (N) -alumina (JCPDS card 29-0063); (s) tetragonal ZrO
17-0923); (h) monoclinic ZrO2 (JCPDS card 37-1484); (d) orthorhombic phase of
MoO (JCPDS card 35-609).
2
O
3
, Mo/ZrO
2
, Mo/SBA-15, NiMo/SBA-15, and Mo/
2
c
2
(JCPDS card
3
The equation system corresponding to the reaction network of
Fig. 1 is presented in Eqs. (1)–(4), where r
i
i
and C are the reaction
rate and the molar concentration of species i; k
n
is the pseudo-first
supported on ZrO
loadings <25 wt.%) were discussed in [41,42]. The application of
impregnated ZrO –SBA-15 as HDS support was studied only for
NiMo catalysts in [25]. Here, we report the characterization results
that complement previous collaborations and allow us to describe
the Mo and NiMo materials in the oxide state (denoted as oxide
precursors).
2 2
–SBA-15 modified by chemical grafting (ZrO
reaction rate constant of the step n according with Fig. 1. This
equation system has one set of kinetic constants that fit the exper-
imental data and offers the possibility of analyzing the three most
important steps in the HDS of DBT-type compounds, i.e., direct
desulfurization (kDDS) or hydrogenation (kHYD) of the starting com-
pound and the sulfur removal from the hydrogenated intermedi-
ates (kDS). The experimental data from the simultaneous HDS of
DBT and 4,6-DMDBT are used to numerically fit the reaction rate
constants; only the results at conversions below 35% of DBT or
2
2 2 3 2
Siliceous SBA-15, ZrO , Al O , and SBA-15 modified with ZrO
loadings of 25, 37, and 50 wt.%, were synthesized and subsequently
utilized as supports for Mo and NiMo catalysts. The main textural
properties of supports and the corresponding oxide precursors are
shown in Table 1. In the following, only the results for the oxide
4
,6-DMDBT are considered.
r
r
r
r
a
¼ ꢁðkHYD þ kDDSÞC
a
ð1Þ
ð2Þ
ð3Þ
ð4Þ
precursors supported on ZrO
of the (Ni)Mo/ZrO (x)SBA materials.
The powder XRD patterns of Mo oxide precursors supported on
ZrO , Al or ZrO (x)SBA materials did not show the presence of
any crystalline Mo phase (Fig. 2). The peaks in Mo/ZrO belonged
to a mixture of monoclinic (JCPDS card 37-1484) and tetragonal
ZrO (JCPDS card 17-0923). The peaks corresponded to the ex-
2
(25)SBA are given as representative
b
c
¼ kDDS
¼ kHYD
C
C
c
a
2
ꢁ kDS
C
c
2
2
O
3
2
a
2
d
¼ kDS
C
2
pected
Mo/ZrO
c
-Al
2
O
3
structure in Mo/Al
2
O
3
, (JCPDS card 29-0063). In
) was
),
and others not visible in Fig. 2 due to the scale were assigned to
tetragonal ZrO . The Mo precursor supported on SBA-15 exhibited
diffraction peaks corresponding to orthorhombic MoO (JCPDS card
35-609). The intensity of these signals diminished after the
addition of Ni (see Fig. 2). The average size of the MoO crystals,
3
. Results
.1. Characterization of oxide precursors
The effect of the ZrO loading into the structure of SBA-15 by
2
(25)SBA, the broad peak between 15° and 35°(2H
attributed to amorphous silica whereas the signal at 31°(2
H
3
2
2
3
incipient wetness impregnation has been described in previous
works [25,40]. The characteristics of Mo and NiMo catalyst
3

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
53
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
increasing agglomeration, e.g., bulk MoO
3
absorbs at 300–330 nm.
Td, Oh
Oh
T
d
h
and O Mo species showed a second absorption band around
2
30 nm.
(d)
The DRS spectra of Mo precursors (Fig. 3) showed that mixtures
of T and O Mo species were present in all cases, although the pro-
portion depended on the support. In the spectra of the Mo precur-
sor supported on SBA-15, ZrO or Al , absorption in the region
00–330 nm produced bands that were not present in the spec-
d
h
(
b)
2
2 3
O
3
(
c)
2
trum of ZrO (25)SBA. In the later material, the absorption band
(
a)
was well defined and had a maximum at 230 nm. Hence, heteroge-
neous distributions of O Mo species in several degrees of agglom-
eration were obtained on pure oxides, whereas supporting Mo on
ZrO (25)SBA increased the dispersion of O molybdenum species.
The absorption edges (E ) calculated for Mo/ZrO and Mo/Al
were 3.4 and 3.5 eV, respectively. The absorption edges were
around 4.1 on ZrO (x)SBA. These values confirmed that
h
2
h
g
2
2 3
O
(
a)
(
d)
2
a
significantly better dispersion of Mo species took place on the Zr-
modified SBA-15 materials.
(
a)
2
00
250
300
350
400
450
500
The addition of Ni to the materials had appreciable influence
only on SBA-15 as shown in Fig. 3 (for clarity, spectra of other
NiMo precursors were omitted). The absorption edge of the
Mo/SBA-15 catalyst was 3.7 eV. After Ni addition, a shift of the
band-gap absorption edge to 4.2 eV was observed pointing to a
significant effect of Ni on the dispersion of Mo precursor sup-
ported on SBA-15.
Wavelength (nm)
Fig. 3. DRS UV–vis spectra of Mo precursor supported on: SBA-15 (a); Al
ZrO (c); and ZrO (25)SBA (d). The dotted line gives the spectra of the NiMo/SBA-15
catalyst.
2 3
O (b);
2
2
determined using Scherrer equation, was 95 nm for Mo/SBA-15
and 52 nm for the corresponding Ni-promoted catalyst. XRD pat-
terns of others NiMo precursors were omitted because they were
identical to the patterns of the corresponding Mo precursors.
UV–vis DRS spectra of Mo and NiMo oxide precursors were re-
corded to obtain further information concerning the dispersion of
Mo oxide species on the supports. Fig. 3 shows the spectra of se-
lected catalysts after subtracting the spectra of the corresponding
support. Hence, the adsorption bands observed in the spectra were
The TPR profiles for Mo oxide precursors are shown in Fig. 4. On
siliceous SBA-15, the low-temperature reduction peak observed at
6
+
4+
763 K is ascribed to the first reduction step (Mo ? Mo ) of poly-
meric O Mo species [45,46]. The H consumption is constant from
823 K to 990 K and then declines gradually to the base line at
1273 K. This broad H consumption region can be produced by
the overlapping of the reduction of crystalline MoO species
(873–900 K), the second step of reduction of the polymeric O
h
2
2
3
h
4
+
0
Mo species (Mo ? Mo ) and the first step of reduction of T
d
Mo
produced only by the ligand-to-metal charge transfer O2 ? Mo
ꢁ
6+
.
species [43,44]. Reduction of Mo species supported on ZrO (x)SBA
2
The position of the absorption bands was determined by the coor-
dination and agglomeration degree of the Mo species in the sam-
ple. The bands in the range 260–280 and 300–320 nm were
materials takes place in a narrow temperature region and pro-
duced better defined signals. Fig. 4 shows the reduction profile
for Mo/ZrO
2
(x)SBA, and the low temperature peak (first reduction
assigned to the isolated molybdate tetrahedral (T
lybdate octahedral (O ) Mo species, respectively [43–45]. The
absorption bands of O
d
) and polymo-
step of O
h
Mo species) is observed at 733 K. The second reduction
species and the reduction of T Mo species are limited to
d
step of O
h
h
h
species shifted to higher wavelengths with
Mo
NiMo
(
(
a)
b)
(
a)
(b)
(
(
c)
d)
(
(
c)
d)
(
(
e)
f)
(
(
e)
f)
470
670
870
1070
1270
470
670
870
1070
1270
Temperature (K)
Temperature (K)
2 2 2 2 2 3
Fig. 4. TPR profiles for Mo and NiMo oxide precursors supported on: SBA-15 (a); ZrO (25)SBA (b); ZrO (37)SBA (c); ZrO (50)SBA (d); ZrO (e); and Al O (f).

5
4
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
o
o
o
o
Mo
(a)
Mo/ZrO (25)SBA
2
*
*
*
*
*
*
*
*
NiMo/SBA-15
Mo/SBA-15
NiMo
Mo
(b)
o
o
NiMo
Mo
(c)
o
x
Mo/ZrO2
Δ
NiMo
NiMo
Δ
Δ
Mo
580
Δ
Δ
(
d)
Mo/Al O3
2
380
430
480
530
630
680 25 50
75 100
10
20
30
40
50
60
70
Temperature (K)
Time (min)
2Θ (degree)
Fig. 5. H
2
S profiles observed in the TPS experiments of Mo and NiMo catalysts
(25)SBA (b); ZrO (c); and Al (d).
Fig. 6. XRD patterns of sulfided Mo/Al
Mo/ZrO (25)SBA. ( -alumina (JCPDS card 29-0063); (s) tetragonal ZrO
card 17-0923); (ꢂ) monoclinic ZrO (JCPDS card 37-1484); (ꢃ) MoS (JCPDS card 37–
492).
2
O
3
, Mo/ZrO
2
, Mo/SBA-15, NiMo/SBA-15 and
supported on the following: SBA-15 (a); ZrO
2
2
2
O
3
2
D
)
c
2
(JCPDS
2
2
1
temperatures below 1073 K where a defined maximum is observed
at 923 K.
mainly to O–S exchange on Mo4+. The different steps of this sulfi-
dation process could overlap on zirconia leading to the observed
2
Increasing loadings of ZrO in the support slightly shift the po-
sition of the first reduction peak to lower temperatures and in-
crease the intensity of the second reduction peak. The same
continuous H S consumption. Unfortunately, we were not able to
2
determine quantitatively the H S consumption in the TPS experi-
2
temperature regions can be distinguished on ZrO
intensity of the reduction signal at high temperature, however,
points to a higher proportion of T Mo species than on the SBA-type
materials. The shoulders at 593 K on zirconia and 853 K on alumina
2
and Al
2
O
3
; the
ments. However, some qualitative observations can be mentioned,
e.g., the net H S uptake on the Mo catalyst supported on SBA-15
2
d
and ZrO (25)SBA was 15% and 10% higher than on Al O or ZrO2.
2
2 3
It can also be safely asserted that the NiMo oxide precursors con-
can be ascribed to agglomerated O
UV–vis DRS characterization.
h
Mo species as inferred from the
sume more H S than the respective Mo materials (Fig. 5). This sug-
gests higher sulfidation degrees for the Ni-promoted catalysts than
2
The thermograms of NiMo precursors are given in Fig. 4. Two
pronounced effects are observed after adding Ni, i.e., the shift of
the first reduction peak to lower temperatures compared with
for the unpromoted catalysts. Furthermore, a significant shift of the
H S-release peak to lower temperatures was observed on the
2
NiMo/Al O catalyst indicating that Ni has a stronger effect on
2 3
the Mo counterparts and the significantly higher H
between 573 and 773 K. These changes indicate the increase in
the proportion of O Mo species and the promotion of the reduc-
2
consumption
the catalyst supported on alumina easing the sulfidation.
h
tion of the same species. These general effects are much smaller
on alumina where the second reduction peak remains as the most
intense. Note that on all materials, the intensity of the reduction
peaks at high temperatures decreases but their positions do not
shift compared with the Mo counterparts. Hence, the proportion
of Mo species that are reduced at high temperatures decreases
but the reduction temperature of such species is not affected.
3.2. Characterization of Mo- and NiMo-sulfide catalysts
The powder XRD patterns of selected sulfided catalysts are pre-
sented in Fig. 6. The catalysts supported on ZrO
2 2 3
, Al O , or
ZrO (25)SBA did not show the presence of any Mo phase. The only
2
present crystalline phases were those observed in the respective
oxide precursors, i.e., monoclinic (JCPDS card 37-1484) and tetrag-
Although the presence of T
d
Mo species in the oxide precursor can-
onal ZrO
29-0063) in Mo/Al
The Mo and NiMo catalysts supported on SBA-15 exhibited diffrac-
tion peaks corresponding to MoS (JCPDS card 37-1492). The inten-
sity of these peaks was higher on the Mo/SBA-15 unpromoted
catalyst pointing to higher dispersion on the promoted counter-
part. The NiMo catalysts supported on others materials exhibited
the same XRD patterns (not shown) observed for the correspond-
ing Mo catalysts.
2
(JCPDS card 17-0923) in Mo/ZrO
2
,
c
-Al
2
O
3
(JCPDS card
not be unambiguously asserted only on the basis of TPR results, the
UV–vis spectra of the precursor and observations of tetrahedral or
monomolybdate species on alumina, zirconia, or silica surfaces
support our interpretations [7,43,47,48].
O
2 3
and tetragonal ZrO
2
in Mo/ZrO
2
(25)SBA.
2
The H
precursors are shown in Fig. 5. In general, H
temperatures about 500 K. Then, the samples release H
ing maxima between 500 and 570 K. Finally, continuous H
sumption is observed before returning to the zero line at 680 K.
Only the catalysts supported on ZrO exhibited a continuous H
2
S profiles observed in the TPS experiments of the oxide
S is consumed below
S produc-
S con-
2
2
2
The HRTEM study was focused on catalysts supported on SBA-
15, Al O , ZrO , and ZrO (25)SBA. Typical MoS fringes with 6.2 Å
2 3 2 2 2
2
2
S
consumption. According to Arnoldy and coworkers [49], the TPS
process can be described as follows. The sulfidation below 500 K
interplanar distance were observed in all the micrographs of the
sulfided catalysts. Representative micrographs of the NiMo catalyst
6+
occurs through substitution by S of the O bound to Mo and the
reduction of these cations to Mo4 through rupture of Mo–S bonds
and formation of elemental S. The S deposited on the support is re-
supported on SBA-15 and ZrO
The length and number of layers distributions of MoS
Mo catalysts are presented in Fig. 8. More than 50% of the observed
crystals of MoS have monolayers, and more than 35% have two
layers on alumina or zirconia. The layer-stacking distribution is
2
(25)SBA are presented in Fig. 7.
+
2
for the
duced with H
2
2
to H S at 500–570 K leading to a H
2
S production sig-
2
nal. The H S consumption at temperatures above 570 K is ascribed
2

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
55
particles were obtained on pure SBA-15 (n = 2.6) among Mo cata-
lysts. The stacking degree slightly decreased with the ZrO loading
to 2.3. On alumina and zirconia, the MoS had the lowest stacking
degree (1.4–1.5). On the other hand, the longest crystals are ob-
tained on the pure Al (59 Å) followed by SBA-15 and ZrO with
52 and 49 Å, respectively. The ZrO (25)SBA material propitiates the
shorter MoS particles with an average size of 36 Å. On the NiMo
2
A
2
2
O
3
2
MoS2
MoS2
2
2
catalysts, the biggest crystals are formed on SBA-15 with an aver-
age size of 47 Å and 3.6 layers. The number of layers is around 2 on
ZrO
MoS
2
and Al
phase supported on ZrO
2
O
3
with crystal size of 46 and 36 Å, respectively. The
(25)SBA has the medium stacking
2
2
degree of 2.9 and the shortest length (31 Å).
The average fraction of Mo atoms on the edge surface of the
MoS2
MoS
tive surface of the crystal [50]. The fMo value is determined from
Eq. (5), assuming that the MoS crystals are perfect hexagons
51]. In Eq. (5), the numerator is the number of atoms in the active
surface (edges) and the denominator is the total number of Mo
2
crystals denoted as ‘‘fMo’’ is a dispersion indicator of the ac-
20 nm
2
[
20nm
B
atoms in the crystal; ‘‘t’’ is the number of layers in the MoS
cles and ‘‘n
2
parti-
i
’’ is the number of Mo atoms in one edge, determined
from the length L (HRTEM) as shown in Eq. (6).
MoS2
P
ð6n
i
ꢁ 6Þ
i¼1;...;t
f
Mo ¼ P
ð5Þ
ð6Þ
2
MoS2
ð3n ꢁ 3n
i
þ 1Þ
i¼1;...;t
i
MoS 2
L
n
i
¼ 6 þ 0:5
:4
We propose in this work a second structure factor, i.e., the average
fraction of Mo atoms in the edge surface of the MoS
2
crystal out of
0
the layer attached to the support ‘‘f ’’. The aim of this fraction is
Mo
Fig. 7. Micrograph of NiMo/SBA-15 (A) and NiMo/ZrO
MoS particles are highlighted.
2
(25)SBA-15 (B) catalysts. The
exploring the relation between the morphology of MoS
HDS of DBT-type compounds. This fraction is obtained from Eq.
7) where the denominator is the total number of Mo atoms in
the crystal and the numerator is the number of atoms in the
MoS edges without considering one layer. Note that the difference
between Eqs. (5) and (7) is the term ‘‘t ꢁ 1’’ in the numerator of Eq.
7). Aim of such modification is to obtain the average fraction of ac-
2
and the
2
(
quite broad on SBA-15; for instance, 10% of the crystals have five or
more layers, and the percentage of monolayers is 9%. In contrast,
2
more than 80% of the crystals have 2 or 3 layers on ZrO
It is clear in Fig. 8-b that the length distribution is also highly
homogenous on ZrO (25)SBA; the length of more than 60% of the
MoS particles is limited in the 21- to 40-Å range. Crystals longer
2
(25)SBA.
(
2
tive surface that is not too close to the support, where stronger ste-
ric hindrance is expected [52].
2
than 60 Å are not detected on this material. The distributions are
less homogeneous on pure oxides, even though most of the crystal
lengths are also in the range of 21–60 Å.
P
ð6n ꢁ 6Þ
i
0
i¼1;...;tꢁ1
F
¼
P
ð7Þ
Mo
2
ð3n ꢁ 3n
i
þ 1Þ
The length and layer-stacking distributions of MoS
2
-type struc-
i
i¼1;...;t
tures on the NiMo catalysts preserve the trend observed for Mo
catalysts as shown in Fig. 9. The presence of Ni reduces the crystal
length but increases the stacking degree in the promoted catalysts
compared with the corresponding Mo counterparts.
0
The values for fMo and f were determined using the information
Mo
obtained from HRTEM (see Table 3). For the fMo values, the observed
trends in terms of the support were
c
-Al O < SBA-15 < ZrO <
2
3
2
The average length and number of layers of the MoS
in each catalyst are presented in Table 2. The most stacked MoS
2
particles
ZrO (25)SBA for the Mo catalysts and SBA-15 < ZrO <
c
-Al O <
2
2
2 3
0
2
ZrO (25)SBA for the NiMo catalysts. In the case of f values, the
2
Mo
60
50
40
30
20
10
0
70
60
50
40
(
a)
1
2
3
4
5
6
(b)
0 - 20
21 - 40
41 - 60
61 - 80
81 - 100
>100
3
2
1
0
0
0
0
SBA-15 ZrO (25)SBA ZrO2
γ-Al O
SBA-15 ZrO (25)SBA ZrO2
γ-Al O
2
2
3
2
2
3
2
Fig. 8. Number of layers (a) and length in Å (b) of the MoS particles in the Mo catalysts supported on selected materials.

5
6
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
60
50
40
30
20
10
0
80
0
-20
(
a)
1
2
3
4
5
6
(b)
7
0
0
21-40
4
1-60
6
>61
50
40
3
2
1
0
0
0
0
SBA-15 ZrO (25)SBA ZrO2
γ-Al O
SBA-15 ZrO (25)SBA ZrO2
γ-Al O
2 3
2
2
3
2
Fig. 9. Number of layers (a) and length in Å (b) of the MoS
2
-like particles in the NiMo catalysts supported on selected materials.
Table 2
100
Average length (L) and number of layers (t) of the MoS
catalysts supported on selected materials.
2
phase in Mo and NiMo
NiMo/
γ
-Al
2
2
O
3
9
8
7
6
5
0
0
0
0
0
DBT
2
NiMo/ZrO (25)SBA
NiMo/ZrO
Catalyst
t
L (Å)
NiMo/SBA-15
Mo/ZrO
(25)SBA
Mo/SBA-15
Mo/ZrO
Mo/ZrO
2.6
2.3
1.4
1.5
3.6
2.9
2.1
1.8
52
36
49
59
47
31
46
36
2
(25)SBA
2
2
2 3
Mo/c-Al O
NiMo/SBA-15
NiMo/ZrO
NiMo/ZrO
40
Mo/SBA-15
Mo/ZrO2
2
(25)SBA
30
2
NiMo/
c
-Al
2
O
3
20
2 3
Mo/γ-Al O
10
0
0
1
2
3
4
5
6
7
8
Table 3
Average fraction of Mo atoms in the edge surface of the MoS
Reaction time (h)
2
crystals (fMo) and
1
00
average fraction of Mo atoms in the edge surface of the MoS
2
-type crystals out of the
0
2
NiMo/ZrO (25)SBA
layer attached to the support (f ) in Mo and NiMo catalysts; concentration of
Mo
90
ꢁ
1
adsorbed NO (
l
mol g
).
catalyst
8
7
6
5
0
0
0
0
4,6-DMDBT
Mo/ZrO
2
(25)SBA
Support
Mo
NiMo
0
0
f
Mo
f
NOads.
f
Mo
f
NOads.
Mo
Mo
NiMo/SBA-15
SBA-15
0.23
0.32
0.24
0.21
0.14
0.18
0.07
0.07
160.7
240.1
175.9
147.8
0.25
0.37
0.26
0.32
0.18
0.24
0.14
0.14
170.3
265.9
190.1
196.9
NiMo/
NiMo/ZrO
Mo/SBA-15
2 3
γ-Al O
ZrO
ZrO
2
(25)SBA
2
40
2
c
2 3
-Al O
3
2
1
0
0
0
0
Mo/ZrO
2
Mo/ -Al
γ
2
O
3
trends obtained for Mo and NiMo catalysts were ZrO
2
(25)SBA >
SBA-15 > ZrO
2
=
c
-Al
2
O
3
. The addition of Ni had a more pronounced
0
0
1
2
3
4
5
6
7
8
effect on fMo and f , when added to the systems supported on alu-
Mo
Reaction time (h)
mina and zirconia. Note that the unpromoted active phase sup-
ported on alumina showed the lowest dispersion (fMo) whereas on
the alumina-supported NiMo catalyst the active phase was the most
dispersed among those supported on pure oxides. However, the ac-
Fig. 10. Conversions of DBT and 4,6-DMDBT at different reaction times over Mo
dashed lines) and NiMo (continuous lines) catalysts supported on selected
materials (573 K, 7.3 MPa H ).
(
2
tive phase was always more disperse on the ZrO
pure oxides. In general, Ni increases the stacking degree and short-
ens the length of the MoS clusters (Table 2), presumably forming
2
(25)SBA than on
adsorbed NO increases around 33%. On the other supports, the
adsorption of NO on the sample increases around 10% after Ni load-
ing. Thus, it is confirmed that the fMo values correlate well with the
dispersion of the active surface on the sulfide catalysts.
2
the NiMoS phase as discussed below. In other words, promotion in-
creases the total dispersion of active surface fMo and the number of
0
slabs that are not in contact with the support f
.
Mo
NO adsorbs on the exposed cations of promoted or unpromoted
MoS ; thus, the concentration of adsorbed NO is correlated with
2
3.3. Catalytic activity
the edge dispersion of the catalyst [53]. It is observed in Table 3
that the total uptake of NO on the sulfide catalysts follows the
same increasing trend than that observed for the fMo value, i.e.,
The activity of Mo and NiMo catalysts series supported on
ZrO (x)SBA-15 and pure oxides was determined in the simulta-
2
Mo/Al
2
O
3
< Mo/ZrO
2
< Mo/SBA-15 < Mo/ZrO
2
(25)SBA for Mo cata-
< Mo/ZrO (25)
neous HDS of DBT and 4,6-DMDBT. The conversion of both reac-
tants is presented in Fig. 10. The results obtained over the
lysts and NiMo/SBA-15 < NiMo/ZrO
2
< NiMo/Al
2
O
3
2
SBA for promoted catalysts. Furthermore, the effect of the Ni
promotion was stronger on alumina where the concentration of
catalysts supported on ZrO (25)SBA were plotted as a representa-
2
tive of the catalyst supported on ZrO -modified SBA-15. The trends
2

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
57
3
2
2
1
1
0
5
0
5
0
5
0
70
(
a) HDS of DBT
60
(b) HDS of DBT
Mo/ZrO (25)SBA
NiMo/ZrO (25)SBA
2
2
50
40
30
20
10
0
0
10
20
30
40
50
0
10 20 30 40 50 60 70 80 90 100
DBT Conversion (%)
DBT Conversion (%)
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
70
60
50
40
30
20
10
0
(d) HDS of 4,6-DMDBT
(
c) HDS of 4,6-DMDBT
Mo/ZrO (25)SBA
NiMo/ZrO (25)SBA
2
2
0
10
20
30
40
50
60
70
0
10 20 30 40 50 60 70 80 90 100
4,6-DMDBT Conversion (%)
4,6-DMDBT Conversion (%)
Fig. 11. Product yield along with the conversion from the HDS of DBT (a and b) and 4,6-DMDBT (c and d) detected on Mo/ZrO
d). In (a and b): tetrahydrodibenzothiophene (h), biphenyl (s), cyclohexylbenzene (
2
(25)SBA (a and c) and NiMo/ZrO
), and dicyclohexyl (e). In (c and d): tetrahydrodimethyldibenzothiophene and
) and dimethyldicyclohexyl (}).
2
(25)SBA (b and
D
hexahydrodimethyldibenzothiophene (h), dimethylbiphenyl (s), methylcyclohexyltoluene (
D
depended on the support and promotion. The addition of Ni in-
creased the conversion of DBT and 4,6-DMDBT on all the catalysts;
sults obtained on the other systems are described below as pseu-
do-first-order constants. Some general comments apply for all
the systems. As described in Fig. 1, the hydrogenated dibenzothi-
ophenic and the biphenylic compounds are primary products,
whereas the cyclohexylbenzene- and dicyclohexyl-type com-
pounds are secondary products. The yield of BP and CHB is similar
at low DBT conversion on the Mo and NiMo catalysts; the yield of
DMBP, however, remains very low on both catalysts at any 4,6-
DMDBT conversion. The presence of Ni promotes the conversion
of the hydrogenated intermediates whose yield increases steadily
on Mo catalysts. Indeed, THDMDBT and HHDMDBT (lumped to-
gether in Fig. 11) are by far the most abundant products in the
HDS of 4,6-DMDBT on Mo catalysts. The concentration of the fully
hydrogenated and desulfurized compounds, DCH and DMDCH, is
significant only at conversions above 35%; thus, the pseudo-first-
order constants determined from Eqs. (1)–(4) are valid to discuss
the main steps in the reaction pathway of DBT and 4,6-DMDBT.
Further description of the activity results is based on the first-
order constants of the main reaction steps (Tables 4 and 5). The
sum of the individual rate constants of each reaction pathway
(kHYD + kDDS) is the overall rate constant, i.e., the net consumption
rate of DBT or 4,6-DMDBT. It is observed that the overall reaction
rate constants followed the trends above described in terms of con-
version for Mo and NiMo catalysts. Furthermore, the activities of
supporting the catalyst on ZrO
tive formulations. The conversion of both reactants over Mo cata-
lysts followed the same increasing trend: Mo/ -Al < Mo/
ZrO < Mo/SBA-15 < Mo/ZrO (25)SBA. The conversion of DBT on
the NiMo catalysts increased in the order: NiMo/SBA-15 < NiMo/
ZrO < NiMo/ -Al < NiMo/ZrO (25)SBA. The conversion of 4,6-
DMDBT on NiMo catalysts increased as follows: NiMo/ZrO < Ni-
Mo/ -Al < NiMo/SBA-15 < NiMo/ZrO (25)SBA.
2
(x)SBA, however, led to the most ac-
c
2 3
O
2
2
2
c
2
O
3
2
2
c
2
O
3
2
The main detected products in the HDS of DBT were tetra-
hydrodibenzothiophene (THDBT), biphenyl (BP), cyclohexylben-
zene (CHB), and dicyclohexyl (DCH). The products from the
HDS of 4,6-DMDBT were tetrahydrodimethyldibenzothiophene
(
THDMDBT), hexahydrodimethyldibenzothiophene (HHDMDBT),
dimethylbiphenyl (DMBP), methylcyclohexyltoluene (MCHT), and
dimethyldicyclohexyl (DMDCH). Only at DBT and 4,6-DMDBT con-
version above 70% on the NiMo catalysts, small amounts of ben-
zene, toluene, cyclohexane and methylcyclohexane were
detected. The small concentration of those products and the ab-
sence of isomers of 4,6-DMDBT indicated a very low acidity of
the materials here were studied. On the contrary, high concentra-
tion of isomers and products from C–C bond cleavage reactions has
3
+
been observed on Al -modified SBA-15 due to the contribution of
acid sites in the silica support doped with trivalent ions [24].
For the sake of brevity, only the product distributions observed
on (Ni)Mo/Zr(25)SBA are presented here (Fig. 11), the activity re-
the NiMo catalyst for both reactants increased with the ZrO
ing in the SBA-15 up to 37 wt.% of ZrO and decreased with 50 wt.%
of ZrO in the support.
2
load-
2
2

5
8
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
Table 4
Table 5
ꢁ
1
ꢁ6
Pseudo-first-order constants (L(s gcatalyst
HDS of DBT and 4,6-DMDBT over Mo catalysts supported on selected materials.
)
ꢂ 10 ) for the main reaction steps of the
ꢁ1
ꢁ6
Pseudo-first-order constants (L(s gcatalyst
HDS of DBT and 4,6-DMDBT over NiMo catalysts supported on selected materials.
)
ꢂ 10 ) for the main reaction steps of the
Support
SBA-15
Reactant
DBT
k
DDS
k
HYD
k
DS
k
DDS + kHYD
kDDS/kHYD
Support
SBA-15
Reactant
DBT
k
DDS
k
HYD
k
DS
k
DDS + kHYD
kDDS/kHYD
1.6
0.2
2.2
3.7
12
10.2
3.8
3.9
0.7
0.06
5.1
0.73
6.9
6.8
256 11.9
71 7.6
0.74
0.1
4,6-DMDBT
4,6-DMDBT
ZrO
ZrO
2
(25)SBA
DBT
2.9
0.2
2.3
6.9
10
8.0
5.2
7.1
1.3
0.03
ZrO
ZrO
ZrO
ZrO
2
2
2
2
(25)SBA DBT
10.1
0.88 13.7
9.4
387 19.5
133 14.6
1.1
0.06
4,6-DMDBT
4,6-DMDBT
2
DBT
1.7
0.2
0.5
3.1
17
4.7
2.2
3.3
3.4
0.06
(37)SBA DBT
11.3
0.98 16.4
10.2
477 21.5
159 17.3
1.1
0.06
4,6-DMDBT
4,6-DMDBT
c
-Al
2
O
3
DBT
0.9
0.1
0.3
1.1
14
2.4
1.2
1.2
3
0.01
(50)SBA DBT
10.6
1.3
8.7
14.3
342 19.3
102 15.6
1.2
0.09
4,6-DMDBT
4,6-DMDBT
DBT
12.3
4.6
3.6
440 16.5
2.7
0.1
4,6-DMDBT
0.35
11.4
0.54
131
246 16.8
136 6.4
3.9
c
-Al
2
O
3
DBT
5.1
5.9
2.2
0.1
The kDDS/kHYD ratio is presented in Tables 4 and 5 as an indicator
of the preferred desulfurization route. The 4,6-DMDBT followed
preferentially the hydrogenation pathway (kDDS/kHYD < 1), regard-
less of the support and the presence of promoter. The preferred
route for the DBT was strongly dependent on the support. The di-
rect desulfurization was clearly followed on zirconia and alumina
4,6-DMDBT
tween the catalysts used in the different studies. The similarity be-
tween reaction rate constants determined in the single and
simultaneous HDS indicates that the effect of competitive adsorp-
tion is negligible for the determination of kinetic parameters.
(kDDS/kHYD P 3 for Mo catalysts) although the addition of Ni seems
to promote the HYD pathway to the larger extent (kDDS/kHYD de-
creased to an average of 2.5). On pure SBA-15, the hydrogenation
of DBT was favored over the direct sulfur removal (kDDS/kHYD is
0
.7 and 0.74 on Mo and NiMo catalysts, respectively). The Mo cat-
alysts supported on ZrO (x)SBA showed the kDDS/kHYD ratio of 1.3;
thus, the DDS pathway was only slightly preferred. On NiMo cata-
4. Discussion
2
4.1. Effect of the oxide precursor on the sulfide catalysts
2
lysts supported on ZrO (x)SBA-15, the kDDS/kHYD ratio slightly de-
creased to around 1.1.
Qualitative relations between the oxide precursors and the cor-
The rate constant kDS shows the ability of the catalyst to remove
the sulfur from the hydrogenated intermediates. The values indi-
cate that the hydrogenated intermediates are more reactive than
the starting DBT-type compounds. The kDS values are one order
of magnitude higher than the kDDS values for the same reactant
on Mo catalysts. The promotion with Ni had a dramatic effect on
the desulfurization of hydrogenated intermediates as the values
of kDS for both molecules over NiMo catalysts (Table 5) are two
or three orders of magnitude higher than the values observed using
the corresponding Mo catalysts (Table 4).
responding sulfide catalysts can be stated from our results. T and
Oh Mo species were observed on the siliceous SBA-15, the later
d
being the most abundant species; indeed, a fraction of O Mo spe-
h
cies was agglomerated to MoO (XRD). In line with this heteroge-
3
neous distribution of oxide species, the sulfide catalyst supported
on SBA-15 exhibited some monolayers of MoS as well as the lon-
2
gest and most stacked MoS crystals. Some MoS particles on SBA-
2
2
15 were large enough to be detected by XRD. The major proportion
of T Mo species on alumina and zirconia led mainly to MoS mon-
d
2
olayers in the sulfided catalysts. The predominance of Oh Mo spe-
These results were in accordance with previous reports where
the activity of promoted and unpromoted Mo catalysts supported
on ZrO -modified SBA-15 were higher than the counterparts sup-
2
cies on ZrO (x)SBA with high dispersion led to MoS clusters with
2
2
medium stacking degree (2–3 layers) and short length. The higher
proportion of O Mo species led to MoS crystals with more layers
h
2
ported on other materials for the HDS of DBT and 4,6-DMDBT in
separate reaction steps [23,25–27,40–42]. In Table 6, the pseudo-
first-order constants for selected catalysts determined here are
compared with the constants determined in the HDS of the single
model compounds as determined in previous studies. Clearly, the
values for a given compound and catalytic system are very similar;
in the cases that the reaction constants differ significantly, trends
are not observed. Thus, the apparent variations in the pseudo-
first-order constant can be attributed to the slight differences be-
on the NiMo catalysts than on the unpromoted counterparts.
The relation between oxide precursors and Mo catalysts is ex-
plained in terms of the strength of interaction. On ZrO2 and
Al O , that have strong interactions with Mo species, the reduction
of an important fraction of Mo species is possible only at high tem-
perature, i.e., the Td Mo species are reduced above 870 K (Fig. 4).
The low reducibility points to low sulfidation degree as confirmed
2
3
by the larger H S consumption of the oxide precursors supported
2
on SBA-15 materials than those on Al O or ZrO . Thus, not all
2
3
2
Table 6
ꢁ
1
ꢂ 10ꢁ6) calculated on selected catalysts for the HDS of a single model molecule and the simultaneous HDS as determined in
Overall pseudo-first-order constants (L(s gcatalyst
this work.
)
a
DBT
4,6-DMDBT
Catalyst ?
Support ;
SBA-15
NiMo
Mo
NiMo
Ref. [27]
This study
Ref. [41]
2.9b
This study
Ref. [42]
This study
b
11.4
21.4
12.5
11.9
19.5
16.8
3.9
7.1
1.2
5
15
5.2
7.6
14.6
5.2
ZrO
-Al
2
–SBA-15c
b
9.2
b
b
c
2
O
3
-
a
b
c
The metal loading and the reaction conditions (573 K and 7.3 MPa H
Calculated from the corresponding conversion–time plots reported in the reference.
The preparation procedure of this material differs; however, the ZrO loading in the SBA-15 is around 23 wt.%.
2
pressure) are the same in all cases.
2

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
59
0
Mo is bonded with S, and O linkages between MoS
2
clusters and
k4,6-DMDBT) on Mo catalysts were plotted against the fMo and f val-
Mo
the support exist after the sulfidation process. Conversely, on
SBA-15, the weak interaction of pure silica with Mo oxide allows
ues. The plot corresponding to the HDS on NiMo catalysts is pre-
sented in Fig. 13.
the development and agglomeration of O
of this interaction propitiates multilayered MoS
sulfidation, whereas incompletely sulfided MoS
have a monolayer structure [54,55]. The medium interaction
strength led to O Mo species that do not agglomerate on
ZrO (25)SBA; thus, after sulfidation, the sulfide species are orga-
h
Mo species. The nature
structures after
clusters typically
Fig. 12 shows the expected general trend of higher HDS activity
0
2
at increasing f or f values. However, the black symbols, corre-
Mo
2
sponding to the reaction constants of both compounds correlated
0
with the f
numbers, exhibit the most proportional behavior,
Mo
0
h
i.e., increasing f values lead always to higher rate constants. With
the assumption that the f values reflect the proportion of Mo
Mo
0
2
Mo
nized in dispersed MoS
monolayers.
The addition of Ni weakens the interaction of Mo species with
the support (the proportion of O Mo species increases) and eases
its reduction and sulfidation (TPR and TPS). In line with this rea-
soning, the effect of promotion is higher on the active sulfide phase
2 3 2
supported on Al O and ZrO , where the interaction is stronger
2
slabs that do not have tendency to form
2
atoms on the edges of MoS -type structures out of the bottom
layer, the most feasible correlation in the case of Mo catalysts be-
0
tween kDBT, k4,6-DMDBT and f values suggests a very low activity of
Mo
h
monolayers. Our discussion on the transition from Mo oxide pre-
cursors to unpromoted catalysts implies that the monolayers in
these catalysts are only partially sulfided and have poor HDS activ-
ity. Thus, it is likely that most of the catalytic activity is due to
than on SBA-type materials. Table 3 shows that fMo increases 52–
2
MoS with more than one layer where less O-bonded Mo is ex-
0
5
8% and f increases 100% after the promotion on alumina or zir-
pected. The incomplete sulfidation of Mo species on unpromoted
catalysts could also led to unclear correlations between morphol-
ogy of the active phase and the catalytic performance.
Mo
0
conia. In contrast, the increments of fMo and f are 15% and 33% on
Mo
ZrO
respectively. The increase in adsorbed NO after Ni promotion was
also the highest on Al
2
(x)SBA, whereas on pure SBA-15, they are only 8% and 28%,
Fig. 13 shows a linear correlation between fMo and the activity
0
2
O
3
.
for the HDS of DBT as well as between f and the HDS activity
Mo
of 4,6-DMDBT. This linear trend clearly shows that the activity of
the NiMo catalyst for the HDS of DBT depends on the dispersion
4
.2. Effect of the morphology of MoS
2
and NiMoS phases on the
of the active phase (fMo). Hence, MoS
alyst are active in the HDS of DBT as well as stacks of MoS
agreement with the literature [58]. The MoS stacks in the NiMo
2
monolayers in the NiMo cat-
catalytic activity
2
in
2
At this point, it is stated that the physical implication of the Ni
promotion is the appearance of the NiMoS phase where promoter
catalyst can be easily assigned to the NiMoS phase type II that
tends to form stacked structures with sparse oxygen. It can also
be assumed that the monolayers in the NiMo catalysts correspond
to type II NiMoS monolayers with minimal interaction with the
support and high activity in the HDS of not-hindered DBT-type
compounds [57]. Interestingly, the activity of the promoted cata-
atoms replace Mo at positions in the edges of the MoS
2
cluster
according to the ‘‘Co–Mo–S’’ model [53]. The NiMoS phase is
claimed to be intrinsically much more active that the pure MoS
phase mainly because of the enhancing of the activation of H
2
2
0
[
56]. Although the NiMoS phase cannot be distinguished from pure
lysts in the HDS of 4,6-DMDBT correlates with the f values (frac-
Mo
MoS
2
in the micrographs or XRD patterns, the much higher intrin-
tion of edges that are not deposited directly on the supports).
sic activity of the NiMo catalysts allows us to assume that the
MoS -like structures observed in the HRTEM images correspond
mainly to the NiMoS phase. Furthermore, the presence of NiS , ex-
2
Consequently, the monolayers of MoS -type structures and the
2
slab that is attached to the support in a stack have limited or no
interaction with 4,6-DMDBT regardless the sulfidation degree.
x
0
pected when the Ni is not completely incorporated in the NiMoS
phase [57], was neither observed by XRD nor HRTEM.
A qualitative relation arises between f for Mo and NiMo cata-
Mo
lysts and the reaction rate constant of the hydrogenation route
0
0
The values of fMo and f were compared with the activity of Mo
(kHYD). The value of f increases, depending on the support, in
Mo
Mo
and NiMo catalysts. In Fig. 12, the values of the overall rate con-
2 3 2 2
the order Al O < ZrO < SBA-15 < ZrO (x)SBA. The values of kHYD
stants for the HDS of DBT and 4,6-DMDBT (denoted as kDBT and
for DBT and 4,6-DMDBT follow the same trend. This is well in line
8
f‘Mo - kDBT
f‘Mo - k4,6-DMDBT
f‘Mo - kDBT
f‘Mo - k4,6-DMDBT
2
1
1
0
5
0
5
0
7
6
5
4
3
2
1
0
fMo - kDBT
fMo - k4,6-DMDBT
fMo - kDBT
fMo - k4,6-DMDBT
0
0.1
0.2
0.3
0
0.1
0.2
0.3
0.4
f’Mo or fMo
f’Mo or fMo
Fig. 12. Overall reaction rate constants of the HDS of DBT and 4,6-DMDBT on Mo
Fig. 13. Overall reaction rate constants of the HDS of DBT and 4,6-DMDBT on NiMo
0
0
catalysts correlated with the fMo and f values.
catalysts correlated with the fMo and f values.
Mo
Mo

6
0
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
with previous observations that the hydrogenation ability of the
catalyst is favored by multilayered MoS clusters [23,49,59].
The formation of NiMoS does change not only the morphology
but also the intrinsic activity of the MoS -like phase. This is evident
sequent removal of sulfur from the hydrogenated intermediates,
e.g., HHDMDBT to DMDCH in Fig. 14, is considerably faster on
2
the promoted phase than on the MoS
2
.
2
in the dramatic increase in the rate constant of desulfurization of
4.3. On the active sites in the MoS and NiMoS phases
2
hydrogenated intermediates (kDS) after Ni promotion, which can-
not be explained by the changes in MoS
2
morphology. It seems that
The nature of the active sites in MoS
versial. It is accepted that coordinatively unsaturated sites located
on the edge of the MoS slabs are the active sites for hydrogenoly-
sis reactions [33]. Adsorption and hydrogenation of aromatic rings
have been attributed to adjacent unsaturated Mo atoms with spe-
cific chemical environment [33], whereas the adsorption and
hydrogenation of thiophenic rings have been claimed to occur on
2
or NiMoS is still contro-
the NiMoS phase fastens the regeneration of active sites or creates
new centers especially active for sulfur removal from hydroge-
nated intermediates. This is well in line with other reports that
claim that the promoter enhances the rate of C–S bond breaking
2
[
33]. As a consequence of this ‘‘electronic’’ influence of Ni, the re-
sults from Mo and NiMo catalysts (Figs. 12 and 13) cannot be com-
pared in terms of morphology of the active phase.
2
the fully sulfide edges of MoS that exhibit metallic characters
The above discussed correlations between the structure factors
(brim sites) [62]. All postulated theories imply, however, that the
active phase of the HDS catalyst has two kinds of active sites for
the HDS of DBT-type compounds; one catalyzes the direct removal
of S (hydrogenolysis site) and the other is active for hydrogenation
steps. The performance of the catalyst is determined by the balance
between hydrogenation and desulfurization sites. Note that the
values of the reaction rate constants determined in this work are
proportional to the concentration of active sites because the con-
centration of products and reactants adsorbed on the surface can
be considered constant at low conversions for all the experiments.
Thus, the rate constants can be used to discuss the catalytic perfor-
mance in terms of concentration of active sites.
0
(f
Mo and f ) and the activity suggest that any possible influence of
Mo
the support during reaction has a minor importance in the HDS
process. The effect of the studied supports on the HDS of DBT-type
compounds seems to be limited to the relation between interaction
strength and MoS
emphasize the characteristics of the active phase in the (Ni)Mo/
ZrO (x)SBA catalysts. The length and layer-stacking degree of the
MoS -type particles are very homogenous on ZrO (25)SBA (Figs. 8
and 9). Most of the clusters have relatively narrow range of length,
whereas between 70% and 80% of MoS crystals were composed by
two or three layers. This average characteristics of the active phase
on ZrO (x)SBA would lead to the morphology that maximizes the
2
or NiMoS morphology. Thus, it is worthing to
2
2
2
2
2
The reaction rates of the HDS of 4,6-DMDBT are comparable
with those of DBT on Mo catalysts, which indicates that 4,6-
DMDBT is not intrinsically less reactive. The hydrogenation of
4,6-DMDBT is ten times faster than its direct desulfurization
whereas the kDDS/kHYD ratio for the HDS of DBT is close to one on
SBA-type materials and around 2 on zirconia and alumina. This
clear difference in the preferred reaction pathway for each reactant
is not dependent on the promotion in a large extent. Thus, we infer
that the difference in selectivity depends more on the geometry of
adsorption of the reactant on the active sites than on the morphol-
ogy of the active phase. A flat adsorption mode (the aromatic rings
approach parallel to the surface) is required to hydrogenate the
reactant molecule either on the vacancies [63] or on brim sites
[62,64] whereas a perpendicular absorption via S-metal bonding
leads to hydrogenolysis [65]. The methyl groups in the 4 and 6
positions in 4,6-DMDBT do not inhibit the flat adsorption mode
but they hamper the perpendicular adsorption [66]. It is likely that
both kinds of active sites are accessible for DBT but 4,6-DMDBT re-
acts mainly on hydrogenation sites.
amount of accessible active sites. Low deactivation has been ob-
served in long-term reactions in trickle bed reactor using SBA-
1
2
5-supported HDS catalysts [60,61]. Thus, the NiMo/ZrO –SBA-15
system is a good candidate to be scaled up; however, this issue de-
pends on many other factors that are not in the scope of this study.
Fig. 14 shows an schematic representation of the interaction of
DBT and 4,6-DMDBT with the supported Mo and NiMo catalysts.
2
From our results, we conclude that pure MoS is present in the for-
mer case, whereas NiMoS is the representative phase in the pro-
moted catalysts. Note that both phases show the same structure
in HRTEM. The morphology–activity relations for unpromoted cat-
2
alysts suggest that the MoS layer attached to the support has low
reactivity in the HDS of both molecules. This is possibly due to the
presence of oxygen bonds to the support that lessen the activity of
the slabs. The formation of the NiMoS phase on the promoted cat-
alysts decreases the interaction with the support and maximizes
the activity and accessibility of the active surface. All the slabs
are accessible for DBT while 4,6-DMDBT has restricted interaction
with the slab adjacent to the support because of steric hindrance.
The selectivity is ruled in first instance by the morphology of the
dibenzothiophenic compound and in a minor extent in the stacking
degree. This is that hydrogenation is always the preferred pathway
in the HDS of 4,6-DMDBT, whereas DBT reacts by DDS or HYD
With the aim of clarifying the discussion, the kDDS/kHYD ratio
(Tables 4 and 5) of the HDS of DBT and 4,6-DMDBT on Mo and
NiMo catalysts are plotted in Fig. 15. Regardless of the promoter,
the ratios are around 1 for the HDS of DBT on Mo and NiMo catalyst
supported on SBA-type materials. Under the assumption that DBT
reacts on both kinds of sites, the ratio close to 1 suggests a bal-
anced proportion of sites. Conversely, the kDDS/kHYD ratio is clearly
without any intrinsic preference. On highly stacked MoS
2
-like clus-
ters, however, the HYD activity of the catalysts increases. The sub-
NiMoS
S
S
S
S
NiMoS
S
MoS
2
S
S
S
NiMoS
Support
Support
2
Fig. 14. Schematic representation of the interaction of DBT and 4,6-DMDBT with supported MoS and NiMoS phase.

O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
61
1
0
Table 7
ꢁ
1
Turnover frequency values (h ) on Mo and NiMo catalysts determined from the
overall pseudo-first-order constants and the corresponding
Turnover frequency values (s mol ) determined from the overall pseudo-first-
order constants and the corresponding concentration of adsorbed NO.
0
f
Mo and
f
values.
Mo
ꢁ
1
ꢁ1
DBT
Support
Reactant
Mo
NiMo
1
0
0
f
Mo
f
NOads.
f
Mo
f
NOads.
Mo
Mo
SBA-15
DBT
0.22
0.23
0.37
0.38
0.59
0.61
0.65
0.41
0.9
0.58
1.76
1.11
4,6-DMDBT
Zr(25)SBA
DBT
0.22
0.3
0.39
0.51
0.54
0.74
0.71
0.53
1.1
0.82
1.83
1.37
0.1
4,6-DMDBT
4
,6-DMDBT
ZrO
2
DBT
0.12
0.19
0.42
0.62
0.31
0.47
0.86
0.21
1.63
0.4
2.22
0.52
4,6-DMDBT
c
-Al
2
O
3
DBT
0.08
0.08
0.23
0.23
0.2
0.2
0.7
0.27
1.59
0.62
2.09
0.82
4,6-DMDBT
0.01
0
10 20 30 40 50 ZrO2 Al O
2
3
ZrO (x)SBA-15
2
dispersion given by fMo or the concentration of NO adsorbed in
the sample (other parameters must be included to obtain units
of frequency). Understanding the physical meaning of those
TOF values, however, is not straightforward because not all the
active sites would be accessible for the dibenzothiophenic com-
pounds, specially for 4,6-DMDBT. Furthermore, calculating TOF
values for each pathway is not possible because the fraction of
active surface (or active sites) where HYD and DDS occur re-
mains unknown. The calculation of TOF values for each pathway
from the same dispersion parameter (NO adsorbed or structure
factor) would lead to artifacts, e.g., the HYD sites are much more
active to convert 4,6-DMDBT than DBT. In contrast, evaluating
overall values of TOF from (kDDS + kHYD) is important to evidence
the structure–activity correlations proposed in this study.
Fig. 15. kDDS/kHYD observed in the HDS of DBT and 4,6-DMDBT over Mo (white
symbols) and NiMo (black symbols) catalysts supported on the materials shown in
the x axis.
2 3 2
higher than 1 on Al O and ZrO , confirming the general agreement
that desulfurization sites are much more abundant than hydroge-
nation sites on alumina and zirconia. The kDDS/kHYD ratio for the
HDS of 4,6-DMDBT below 0.1 reflects the difficulty of this molecule
to adsorb on desulfurization sites. Note in Fig. 15 that the kDDS/kHYD
ratio for the HDS of 4,6-DMDBT is higher on promoted catalysts
than on Mo catalysts. This points to the enhancement of direct
desulfurization sites or the creation of new active sites in the Ni-
MoS phase that are able to directly desulfurize the 4,6-DMDBT
after promotion, which are related to the dramatic enhancement
of sulfur removal from hydrogenated intermediates (kDS).
The TOF values determined from both structure factors (fMo
0
and f ) as well as from the concentration of adsorbed NO for
Mo
Mo and NiMo catalysts are presented in Table 7. Clear discussion
cannot be made from the TOF values of the unpromoted cata-
The morphology of MoS
2
phase has been quantitatively related
lysts probably because of the not fully sulfidation of the MoS
2
to the overall HDS activity of DBT and 4,6-DMDBT. It is also noticed
that the multilayered structures favor the hydrogenation pathway
of both molecules. The kDDS/kHYD ratio, however, cannot be corre-
phase, which is more aggravating on alumina. However, it is
important to note that the TOF values on SBA-type materials
are very similar, which suggest that the improving of the activity
of the catalysts is not due a higher intrinsic activity of the active
sites.
0
lated with fMo or f . This indicates that the hydrogenolysis activity
Mo
2
is not determined by the number of layers of the MoS crystal. Fur-
thermore, the hydrogenation activity depends on the structure but
the hydrogenation sites are not limited to the ‘‘top’’ slabs as the
adsorption on ‘‘brim sites’’ implies. It is clear that less stacked clus-
The case of TOF values on NiMo is much more conclusive.
There is not trend among the values determined from any param-
eter. However, comparing the values for DBT calculated from fMo
0
ters of MoS
monolayers of MoS
2
should have higher proportion of ‘‘top’’ slabs, i.e., the
or NiMoS should have high hydrogenation
with the 4,6-DMDBT TOF values determined from f (bold num-
Mo
bers in Table 7), it is clear that they do not differ significantly.
Furthermore, the TOF values of the HDS of DBT as determined
from the concentration of adsorbed NO are quite similar as well
(italic bold numbers in Table 7). This has two main implications:
(i) the above-mentioned assignments of active sites, i.e., all sites
are accessible for DBT but only a fraction for 4,6-DMDBT, is cor-
rect; (ii) the intrinsic activity of the active sites do not depend
on the support or on the structure of the molecule in a large ex-
tent. The rest of the TOF values of the NiMo catalysts (not high-
lighted in Table 7) are too high or too low because the fraction
of accessible active sites is underestimated for DBT and overesti-
mated for 4,6-DMDBT.
2
ability because the flat adsorption on the basal plane is not con-
strained. This is contrary to our results and to the behavior gener-
2
ally accepted of MoS monolayers. Thus, the adsorption and
subsequent hydrogenolysis or hydrogenation of DBT-type com-
pounds can be ascribed to coordinatively unsaturated sites at the
perimeter of the (1 0 1 0) plane in all the MoS slabs. The planar
2
adsorption of a DBT molecule would require an approximate area
of 9.5 ꢂ 5 Å, whereas the distance of an exposed Mo cation in the
first slab to the support should be smaller than 6.1 Å (the distance
between Mo atom layers in two stacked MoS
absorption of DBT or 4,6-DMDBT on active sites located on the
1 0 1 0) planes would strongly be affected by the support. For
stacked MoS particles, the increasing distance to the support leads
2
slabs). Thus, the flat
0
(
The average TOF values on Mo catalysts determined from f is
Mo
ꢁ
1
0.35 and 0.44 h for DBT and 4,6-DMDBT. The average TOF values
for DBT (from fMo) and
0.6 h for 4,6-DMDBT (from f ). On the other hand, the TOF val-
2
ꢁ
1
more space to the reactants for the adsorption required for hydro-
genation. The metal–sulfur perpendicular adsorption is possible on
monolayers because the molecule adsorbs perpendicularly to the
active phase, parallel to the support.
It is possible to determine turnover frequency (TOF) values by
dividing the pseudo-first-order constants by either the overall
on NiMo catalysts increase to 0.74 h
ꢁ1
0
Mo
ues determined from the adsorption of NO increase up to ten times
after the addition of Ni. These variations substantiate the promoter
effect of Ni, i.e., the higher HDS performance of the NiMoS phase
2
than that of the pure MoS .

6
2
O.Y. Gutiérrez, T. Klimova / Journal of Catalysis 281 (2011) 50–62
[
[
16] U. Turaga, C. Song, Catal. Today 86 (2003) 129.
17] A. Wang, Y. Wang, T. Kabe, Y. Chen, A. Ishihara, W. Qian, P. Yao, J. Catal. 210
2002) 319.
5
. Conclusions
(
According to the above-mentioned results, the following con-
[18] A. Sampieri, S. Pronier, J. Blanchard, M. Breysse, S. Brunet, K. Fajerwerg, C.
Louis, G. Pérot, Catal. Today 107/108 (2005) 537.
clusions can be stated.
[
19] L. Vradman, M. Landau, M. Herskowitz, V. Ezersky, M. Talianker, S. Nikitenko,
Y. Koltypin, A. Gedanken, J. Catal. 213 (2003) 163.
The characteristics of the oxide Mo species depend on the inter-
action strength with the support. The increasing proportion of T
Mo species indicates stronger interactions. The highest proportion
of O Mo species with the maximum dispersion was found among
the studied catalysts on ZrO (x)SBA supports.
The characteristics of the oxide species on ZrO
to MoS clusters with short length (36 and 31 Å on Mo and NiMo
catalysts respectively) and 2–3 layers in average. Such morphology
maximizes the active surface and the proportion of slabs sterically
unhindered for the proximity of the support. Consequently, the
d
[
20] A. Taguchi, F. Schuth, Micropor. Mesopor. Mater. 77 (2005) 1.
[21] G. Murali Dhar, G. Muthu Kumaran, M. Kumar, K. Rawat, L. Sharma, B. Raju, K.
Rama Rao, Catal. Today 99 (2005) 309.
[
h
22] G. Muthu Kumaran, S. Garg, K. Soni, M. Kumar, L. Sharma, G. Murali Dhar, K.
Rama Rao, Appl. Catal. A 305 (2006) 123.
2
[23] O. Gutiérrez, G. Fuentes, C. Salcedo, T. Klimova, Catal. Today 116 (2006) 485.
[
[
2
(x)SBA conduct
24] T. Klimova, J. Reyes, O. Gutiérrez, L. Lizama, Appl. Catal. A 335 (2008) 159.
25] O. Gutiérrez, K. Romero, G. Fuentes, T. Klimova, Stud. Surf. Sci. Catal. 162
2
(
2006) 355.
[26] O. Gutiérrez, F. Pérez, C. Salcedo, G. Fuentes, X. Bokhimi, T. Klimova, Stud. Surf.
Sci. Catal. 165 (2007) 803.
[
27] T. Klimova, L. Peña, L. Lizama, C. Salcedo, O. Gutierrez, Ind. Eng. Chem. Res. 48
2009) 1126.
2
catalytic system NiMo/ZrO (x)SBA-15 is an outstanding catalyst
(
because it can convert and extract sulfur from two kinds of DBT-
type compounds with different behaviors in HDS.
[28] F. Dumeignil, K. Sato, M. Imamura, N. Matsubayashi, E. Payen, H. Shimada,
Appl. Catal. A 315 (2006) 18.
[
29] A. Dugulan, High-Pressure Sulfidation of Hydrotreating Catalysts, Genesis and
properties of the Active Phase, IOS Press, Amsterdam, 2008. p. 176.
It is shown that the intrinsic HDS activity of the catalysts does
not depend importantly on the composition of the support. In con-
trast, the formation of the NiMoS phase and the morphology of the
active phase are the key factors in the HDS of DBT-type
compounds.
[30] F. Sánchez, J. Ramírez, R. Cuevas, A. Gutiérrez, C. Fernández, Ind. Eng. Chem.
Res. 48 (2009) 1178.
[
[
[
[
31] D. Zhao, Q. Huo, J. Feng, B. Chmelka, G. Stucky, J. Am. Chem. Soc. 120 (1998)
024.
32] C. Pophal, F. Kameda, K. Hoshino, S. Yoshinaka, K. Segawa, Catal. Today 39
(1997) 21.
33] F. Bataille, J. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire, E. Schulz,
M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409.
34] M. Houalla, N. Nag, A. Sapre, D. Broderick, B. Gates, AIChE J. 24 (1978) 1015.
6
The MoS
ity in the HDS of both DBT-type compounds. In promoted catalysts
NiMoS phase), the sulfide monolayers or the layer attached to the
2
monolayers in unpromoted catalysts have low activ-
(
support in a multilayered cluster is active for the HDS of DBT, but
has little activity for the HDS of 4,6-DMDBT. This is not related to a
different location of hydrogenation and hydrogenolysis sites in the
MoS cluster but to the structure of the reactant molecule and the
2
space available for the interaction with the active sites. In the
proximity of the support, the flat adsorption is hindered and thus,
[35] Y. Wang, Z. Sun, A. Wang, L. Ruan, M. Lu, J. Ren, X. Li, C. Li, Y. Hu, P. Yao, Ind.
Eng. Chem. Res. 43 (2004) 2324.
[
[
36] J. Kim, X. Ma, C. Song, Energy Fuels 19 (2005) 353.
37] N. Kagami, B. Vogelaar, A. van Langeveld, J. Moulijn, Appl. Catal. A 293 (2005)
11.
[
[
38] J. Chen, H. Yang, Z. Ring, Catal. Today 98 (2004) 227.
39] P. Da Costa, C. Potvin, J. Manoli, J. Lemberton, G. Pérot, G. Djéga-Mariadassou, J.
Mol. Catal. A 184 (2002) 323.
4
,6-DMDBT that adsorbs preferably in this mode does not react on
[40] O. Gutiérrez, E. Ayala, I. Puente, T. Klimova, Chem. Eng. Commun. 196 (2009)
163.
1
sulfide monolayers.
[
[
41] O. Gutiérrez, D. Valencia, G. Fuentes, T. Klimova, J. Catal 249 (2007) 140.
42] O. Gutiérrez, F. Pérez, G. Fuentes, X. Bokhimi, T. Klimova, Catal. Today 130
(2008) 292.
Acknowledgments
[
[
43] Z. Liu, Y. Chen, J. Catal. 177 (1998) 314.
44] C. Williams, J. Ekerdt, J. Jehng, F. Hardcastle, A. Turek, I. Wachs, J. Phys. Chem.
Financial support by CONACYT-Mexico (Grant 100945) is
gratefully acknowledged. The authors thank M. Aguilar Franco, C.
Salcedo Luna and I. Puente Lee and for their technical assistance
with XRD and HRTEM characterizations. The authors also thank
Professor J.A. Lercher for providing laboratory facilities for the
characterization of sulfided materials at the Catalysis Research
Center in the Technische Universität München.
95 (1991) 8781.
[
45] C. Williams, J. Ekerdt, J. Jehng, F. Hardcastle, I. Wachs, J. Phys. Chem. 95 (1991)
8791.
[
[
46] R. López-Cordero, F. Gil-Llambias, A. López-Agudo, Appl. Catal. 74 (1991) 125.
47] J. Ramírez, R. Cuevas, P. Castillo, M. Rojas, T. Klimova, Bulg. Chem. Commun. 30
(
1998) 207.
[48] C. Louis, M. Che, M. Anpo, J. Catal. 141 (1993) 453.
[
[
49] P. Arnoldy, J. van den Heijkant, G. de Bok, L. Moulijn, J. Catal. 92 (1985) 35.
50] E. Hensen, P. Kooyman, Y. van der Meer, A. van der Kraan, V. de Beer, J. van
Veen, R. van Santen, J. Catal. 199 (2001) 224.
References
[51] S. Kasztelan, H. Toulhoat, J. Grimblot, J. Bonnelle, Appl. Catal. 13 (1984) 127.
[
[
[
[
52] H. Shimada, Catal. Today 86 (2003) 17.
53] N. Topsøe, H. Topsøe, J. Catal. 84 (1983) 386.
54] H. Topsøe, B. Clausen, Appl. Catal. 25 (1986) 273.
55] S. Bouwens, F. van Zon, M. van Dijk, A. van der Kraan, V. de Beer, J. van Veen, D.
Koningsberger, J. Catal. 146 (1994) 375.
[
[
[
1] C. Hsu, P. Robinson, Practical Advances in Petroleum Processing, Springer, USA,
006. p. 866.
2] A. Ishihara, F. Dumeignil, J. Lee, K. Mitsuhashi, E. Qian, T. Kabe, Appl. Catal. A
89 (2005) 163.
3] F. Dumeignil, K. Sato, M. Imamura, N. Matsubayashi, E. Payen, H. Shimada,
Appl. Catal. 315 (2006) 18.
2
2
[
[
56] P. Moses, B. Hinnemann, H. Topsøe, J. Nørskov, J. Catal. 268 (2009) 201.
57] S. Eijsbouts, L. van den Oetelaar, R. van Puijenbroek, J. Catal. 229 (2005)
3
52.
58] G. Berhault, M. De la Rosa, A. Mehta, M. Yácaman, R. Chianelli, Appl. Catal. A
45 (2008) 80.
[
[
[
4] M. Nagai, T. Fukiage, S. Kurata, Catal. Today 106 (2005) 201.
5] E. Kraleva, A. Spojakina, K. Jiratova, L. Petrov, Catal. Lett. 112 (2006) 203.
6] H. Shimada, T. Sato, Y. Yoshimura, J. Hiraishi, A. Nishijima, J. Catal. 110 (1988)
[
3
[
[
[
59] P. Da Silva, N. Marchal, S. Kasztelan, Stud. Surf. Sci. 106 (1997) 353.
60] K. Soni, K. Mouli, A. Dalai, Catal. Lett. 136 (2010) 116.
61] V. Sundaramurthy, I. Eswaramoorthi, A. Dalai, J. Adjaye, Micropor. Mesopor.
Mater. 111 (2008) 560.
2
75.
[
[
7] M. Jia, P. Afanasiev, M. Vrinat, Appl. Catal. A 278 (2005) 213.
8] M. Jia, P. Afanasiev, M. Vrinat, W. Li, H. Xu, Q. Ge, Petrochem. Technol. 33
(
2004) 202.
[
62] J. Lauritsen, M. Nyberg, J. Nørskov, B. Clausen, H. Topsøe, E. Lægsgaard, F.
Besenbacher, J. Catal. 224 (2004) 94.
[
9] J. Weissman, E. DeCanio, J. Edwards, Catal. Lett. 24 (1994) 113.
[
[
10] S. Damyanova, L. Petrov, P. Grange, Appl. Catal. A 239 (2003) 241.
11] M. Rana, S. Maity, J. Ancheyta, G. Murali Dhar, T. Prasada Rao, Appl. Catal. A
[
[
63] C. Song, X. Ma, Appl. Catal. B 41 (2003) 207.
64] J. Lauritsen, J. Kibsgaard, G. Olesen, P. Moses, B. Hinnemann, S. Helveg, J.
Nørskov, B. Clausen, H. Topsøe, E. Lægsgaard, F. Besenbacher, J. Catal. 249
2
68 (2004) 89.
[
12] M. Barrera, M. Viniegra, J. Escobar, M. Vrinat, J. de los Reyes, F. Murrieta, J.
García, Catal. Today 98 (2004) 131.
(
2007) 220.
65] S. Cristol, J. Paul, E. Payen, D. Bougeard, F. Hutschka, S. Clémendot, J. Catal. 224
2004) 138.
66] V. Vanrysselberghe, G. Froment, Ind. Eng. Chem. Res. 37 (1998) 4231.
[
[
[
[
[
13] J. Lee, S. Han, H. Kim, J. Koh, T. Hyeon, S. Moon, Catal. Today 86 (2003) 141.
14] A. Wang, Y. Wang, T. Kabe, Y. Chen, A. Ishihara, W. Qian, J. Catal. 199 (2001) 19.
15] T. Klimova, M. Calderón, J. Ramírez, Appl. Catal. A 240 (2003) 29.
(