Mo and NiMo catalysts supported on SBA-15 modified by grafted ZrO2 species: Synthesis, characterization and evaluation in 4,6-dimethyldibenzothiophene hydrodesulfurization

Journal of Catalysis 249 (2007) 140–153

www.elsevier.com/locate/jcat

Mo and NiMo catalysts supported on SBA-15 modiﬁed by grafted ZrO₂

species: Synthesis, characterization and evaluation in

4,6-dimethyldibenzothiophene hydrodesulfurization

a

b

a,∗

Oliver Y. Gutiérrez , Diego Valencia , Gustavo A. Fuentes , Tatiana Klimova

a

Facultad de Química, Universidad Nacional Autónoma de México (UNAM), Cd. Universitaria, Coyoacán, México D.F., 04510, Mexico

b

Área de Ingeniería Química, Universidad Autónoma Metropolitana-Iztapalapa (UAM-I), Av. Michoacán y Purísima, Iztapalapa, México D.F., 09340, Mexico

Received 2 February 2007; revised 18 April 2007; accepted 21 April 2007

Available online 1 June 2007

Abstract

A series of ZrO -containing mesoporous SBA-15 supports and their respective Mo and NiMo catalysts were prepared to study the effect of

2

zirconia loading on the characteristics of Ni and Mo species and their catalytic activity in 4,6-dimethyldibenzothiophene hydrodesulfurization

(

HDS). ZrO -containing SBA-15 solids with different metal loadings (up to 23 wt% of ZrO ) were prepared by chemical grafting at room

2

temperature. Supports and catalysts were characterized by N physisorption, XRD, UV–vis DRS, TPR, chemical analysis, and HRTEM. The

2

incorporation of zirconia into the SBA-15 support provides better dispersion to the deposited molybdenum species, increasing the effective

surface of the MoS phase. Mo and NiMo catalysts supported on SBA-15 materials showed an increase in catalytic activity in 4,6-dimethyldi-

2

benzothiophene HDS with zirconia loading in the support. Unpromoted Mo catalysts were active in the formation of hydrogenated intermediates

of 4,6-DMDBT, namely tetrahydro- and hexahydrodimethyldibenzothiophenes; however, they were not able to realize efﬁcient sulfur elimination

from these intermediates with the formation of desulfurized products. Addition of the Ni promoter resulted in a further increase in Mo dispersion,

as well as in the acceleration of C–S bond cleavage in hydrogenated intermediates of 4,6-DMDBT, improving the overall kinetics of the HYD

pathway of HDS.

©

Keywords: SBA-15; Mesoporous molecular sieves; Zirconia; Chemical grafting; Mo and NiMo catalysts; Deep hydrodesulfurization;

4,6-Dimethyldibenzothiophene

1

. Introduction

Many different approaches have been taken to achieve this

goal, including varying the catalyst support. Different materi-

als have been assayed as new supports for HDS catalysts [3,4].

TiO and ZrO have attracted attention due to the greater (by

three to ﬁve times) intrinsic HDS activity demonstrated by

Mo catalysts supported on these oxides compared with their

The demand for high-quality, low-sulfur transportation fu-

els (gasoline, diesel) is growing with increasing concerns re-

lated to public health [1]. Traditionally, the removal of sul-

fur from petroleum-derived feedstocks has been realized by

a hydrodesulfurization (HDS) process using alumina-supported

CoMoS and NiMoS catalysts. However, to satisfy new, more

stringent environmental restrictions on the sulfur content of fu-

els, more active HDS catalysts are being requested from indus-

try [2].

2

alumina-supported counterparts [3–7]. TiO supports have been

2

widely studied, being the subject of numerous investigations

of the support effect in hydrotreating catalysts [3–5]. ZrO2-

supported catalytic systems have been much less widely studied

up to now. This neglect is indeed surprising because, as reported

previously, Ni-promoted MoS2 and WS2 catalysts supported

*

on ZrO show higher hydrodesulfurization and hydrogenation

Corresponding author. Fax: +52 55 56225371.

2

E-mail address: klimova@servidor.unam.mx (T. Klimova).

(HYD) activities than titania-supported ones [3,8,9].

doi:10.1016/j.jcat.2007.04.014

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

141

In general, the high activity observed for the zirconia-

supported catalysts is related to the morphology of the sulﬁded

phase. MoS2 crystallites on zirconia have been found to be

shorter and to have less stacking of the layers compared with

ﬁnding into account, we continue our research of novel HDS

catalysts supported on ZrO2-modiﬁed SBA-15 materials.

In the present work, we report some new results on the

Zr-SBA-15 support preparation and characterization. To fur-

ther investigate the cause of the high catalytic activity of the

NiMo/Zr-SBA-15 formulations, we attempted to analyze sepa-

rately the contributions of the ZrO2-containing support and of

the Ni promoter in the catalytic behavior of these systems. For

this purpose, a series of SBA-15 materials modiﬁed with dif-

ferent amounts of zirconia was prepared by chemical grafting

and used to support unpromoted Mo catalysts and Ni promoted

ones, both of which were evaluated in the 4,6-dimethyldibenzo-

thiophene (4,6-DMDBT) HDS reaction.

those on alumina [7,10]. Smaller MoS particles should have

2

a greater number of active sites through increased exposure

of MoS2 crystallite edges. This supposition was conﬁrmed by

comparing the amount of Mo active sites in Mo catalysts sup-

ported on alumina and zirconia determined by dynamic NO

adsorption; ﬁve times more NO-adsorbing sites were found

in the zirconia-supported catalyst compared with the Al O -

2

3

supported counterpart [11]. In addition, it was suggested that

zirconia promotes the sulﬁdation of oxomolybdate entities and

promotes the creation of the promoted “NiMoS” phase, which

seems to be more accessible to the reactant molecules than in

alumina-based catalysts [7,12]. But despite promising results

2

. Experimental

2

.1. Support and catalyst preparation

obtained with pure ZrO support, its possible application in

2

catalysts is restricted due to the low surface area (<100 m /g)

The siliceous SBA-15 material used in this work has been

and porosity of commercially available zirconia and structural

instability during high-temperature treatment. Because of zir-

conia’s low surface area and despite its high intrinsic activity

synthesized as described elsewhere [22,23]. ZrO2-modiﬁed

SBA-15 supports were prepared by chemical grafting [24].

Zirconium(IV) propoxide (Zr(n-PrO)4, 70 wt% solution in 1-

propanol, Aldrich) was used as the zirconia source, and absolute

ethanol was used as the solvent (EtOH, Aldrich, 99.999%). In

the grafting procedure, calcined SBA-15 was slurried in EtOH

containing Zr(n-PrO)4 for 8 h at room temperature. To elimi-

nate excess Zr(n-PrO)4, the ﬁltered material was washed with

dry EtOH. The solid was then dried in air at room tempera-

(

per Mo atom), the speciﬁc activity (per g of catalyst) of ZrO₂-

supported catalysts is lower than that of alumina-supported

catalysts [7].

Recently, different zirconia-containing materials suitable

for hydrotreating catalysts have been synthesized and tested

in HDS reactions. Promising results have been obtained with

ZrO -containing mixed oxide supports (Al O –ZrO [11,13,

2

3

2

◦

ture and calcined in static air at 550 C for 5 h. Hereinafter,

1

4], ZrO2–TiO2 [15], and ZrO2–SiO2 [13,16]). Attempts also

ZrO2-containing SBA-15 materials are designated Zr(x)SBA-

have been made to prepare pure zirconia supports with ordered

1

5 samples, where x represents the wt% of ZrO2 in the sample.

mesopore structure or to introduce Zr⁴species in siliceous

mesoporous molecular sieves of different types (e.g., MCM-41,

HMS, SBA-15). Zr-HMS materials with different Zr/Si ratios

have been prepared and used as supports for Mo HDS cata-

lysts, and showed higher catalytic activity than the reference

+

To assess thermal stability, the Zr(23)SBA-15 material was

◦

calcined at different temperatures (700, 800, 900, or 1000 C)

for 1 h. For the hydrothermal stability studies, the freshly syn-

thesized material [SBA-15 or Zr(23)SBA-15] was reﬂuxed in

distilled water (at a water-to-sample ratio of 1 L/g) for 24 h

Mo/SiO material in thiophene HDS [17]. It was found that

◦

2

and dried at 100 C for 24 h before further characterization.

zirconium improved the hydrogenation activity of the sam-

ples and facilitated reduction of the supported molybdenum

species. However, MCM-41 and HMS-type materials have poor

mechanical and hydrothermal stability [18,19] and relatively

small pores (about 20–35 Å diameter), which can be easily

blocked by HDS active phases [17,20,21]. Mesoporous silicas

of the SBA family seem to be more suited for modiﬁcation

Mo- and NiMo/SBA-15 catalysts were prepared by a stan-

dard incipient wetness technique. The calcined supports were

impregnated successively using aqueous solutions of ammo-

nium heptamolybdate, (NH4)6Mo7O24·4H2O (Aldrich), and

nickel nitrate, Ni(NO3)2·6H2O (Aldrich). Mo was impregnated

◦

ﬁrst. After each impregnation, the catalysts were dried (100 C,

◦

2

4 h) and calcined (500 C, 4 h). In addition, NiMo catalyst

with ZrO because of their larger pores, thicker pore walls,

2

supported on γ -alumina was prepared for comparison purposes.

The nominal composition of the catalysts was 12 wt% of MoO3

and 3 wt% of NiO. Hereinafter, the catalysts are designated as

Mo or NiMo/corresponding support.

and higher hydrothermal stability compared with MCM-41

[

20–23]. Recently, our group prepared a series of TiO - and

2

ZrO2-containing SBA-15 materials and the respective NiMo

catalysts and tested them in 4,6-dimethyldibenzothiophene

(

4,6-DMDBT) HDS [24]. We found that incorporating titania or

2

.2. Zirconium(IV) deposition isotherm

zirconia in the support provided better dispersion to deposited

Ni and Mo active species and increased their catalytic activity

in HDS of 4,6-DMDBT, one of the most refractory sulfur com-

Equilibrium adsorption experiments were performed at 25±

◦

0.1 C. Siliceous SBA-15 parent material, calcined previously

◦

pounds. The catalytic activity obtained with ZrO -containing

and dried at 120 C for 15 h, was used as an adsorbent. In each

2

supports was greater than that obtained with TiO2-containing

supports, and substantially higher (almost double) than ob-

tained with the reference NiMo/γ -Al2O3 catalyst. Taking this

experiment, 1.0 g of the SBA-15 was suspended in 200 mL

of the Zr(n-PrO) ethanol solution of various concentrations

4

−

4

−1

of Zr(IV) ranging between 0.8 × 10 and 2.3 × 10 mol/L.

142

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

This suspension was kept in inert atmosphere (Ar) under con-

tinuous stirring for 8 h, allowing sufﬁcient time to reach the

equilibrium of Zr(IV) adsorption. Subsequently, the solid was

separated from the liquid by ﬁltration, dried, and calcined as

described above. The surface concentration of Zr(IV) atoms in

microscope (resolving power 1.9 Å). The solids were ultrason-

ically dispersed in heptane, and the suspension was collected

on carbon coated grids. Slab length and layer stacking distribu-

tions of MoS2 crystallites in each sample were established from

the measurement of at least 300 crystallites detected on several

HRTEM pictures taken from different parts of the same sample

dispersed on the microscope grid.

2

the calcined solid (Γ , atoms Zr/nm ) was calculated from the

results of chemical analysis. Zr(IV) loadings obtained from the

analysis of the ZrO2-grafted solids were found to be in good

agreement with those calculated on the basis of the difference

between zirconium(IV) concentration in the solution before and

after deposition using the relationship

2.4. Catalytic activity

The 4,6-DMDBT HDS activity tests were performed in a

◦

ꢀ

ꢁ

batch reactor at 300 C and 7.3 MPa total pressure for 8 h. Be-

Γ = V (C0 − Ceq) /WS,

where C0, Ceq, V , W, and S are the Zr(IV) concentration be-

fore and after adsorption (mol/L), suspension volume (0.2 L),

(1)

fore the catalytic activity evaluation, the catalysts were sulﬁded

◦

ex situ in a tubular reactor at 400 C for 4 h in a stream of

1

5 vol% of H2S in H2 under atmospheric pressure. The sulﬁded

2

catalysts (0.15 g) were transferred in an inert atmosphere (Ar)

to a batch reactor (Parr) containing 40 mL of 4,6-DMDBT solu-

tion in n-hexadecane (0.024 mol/L). The course of the reaction

was followed by withdrawing aliquots each hour and analyzing

them on an HP-6890 chromatograph. To corroborate product

identiﬁcation, the product mixture was analyzed on a Hewlett

Packard GC-MS instrument.

weight (g), and speciﬁc surface area (m /g) of the support, re-

spectively.

2

.3. Support and catalyst characterization

The supports and catalysts were characterized by N2

physisorption, X-ray diffraction (XRD), UV–vis diffuse reﬂec-

tance spectroscopy (DRS), temperature-programmed reduction

3

. Results and discussion

(

TPR), SEM-EDX, and high-resolution transmission electron

microscopy (HRTEM). N2 adsorption/desorption isotherms

were measured with a Micromeritics ASAP 2000 automatic

analyzer at liquid N2 temperature. Before the experiments,

the samples were degassed (p < 10 Pa) at 270 C for 6 h.

Speciﬁc surface areas were calculated by the BET method

3

.1. Supports

−

1

◦

Fig. 1A illustrates the deposition isotherm of Zr(IV) species

on the surface of the parent SBA-15 obtained experimentally

at room temperature (25 C). The initial solution concentra-

◦

(

SBET), the total pore volume (Vp) was determined by nitro-

tion of zirconium(IV) propoxide was changed to reach different

metal loadings in Zr(IV)-grafted solids. A plateau observed in

the isotherm corresponds to the monolayer surface coverage.

The experimental points are well ﬁtted with a Langmuir-type

isotherm,

gen adsorption at a relative pressure of 0.98, and pore size

distributions were determined from the desorption isotherms

by the BJH method. The mesopore diameter (DP) corresponds

to the maximum of the pore size distribution. The micropore

area (Sμ) was estimated using the correlation of t-Harkins

and Jura (t-plot method). XRD patterns were recorded in the

Γ/Γm = (KCeq)/(1 + KCeq).

(2)

◦

3

ꢀ 2Θ ꢀ 90 range on a Siemens D5000 diffractometer, us-

ing CuKα radiation (λ = 1.5406 Å) and a goniometer speed

This type of isotherm implies localized Langmuir-type adsorp-

tion of the zirconium species on SBA-15 surface with too

weak (if any) lateral interactions. Therefore, most of the ad-

sorbed zirconia species should exist in isolated form, and zir-

conia agglomerates, if exist, should be present in the samples

only as a minority fraction. This result points out high disper-

sion of grafted Zr(IV) species and conﬁrms the conclusion of

our previous work in which only isolated Zr(IV) species in

tetrahedral coordination were detected on SBA-15 surface by

UV–visible diffuse reﬂectance spectroscopy regardless of ZrO2

loading [24]. The monolayer capacity (Γm) found from lin-

earized Langmuir plot (Fig. 1B),

◦

−1

◦

of 1 (2Θ)min . Small-angle XRD (2Θ = 1 –10 ) was per-

formed on a Bruker D8 Advance diffractometer using small

◦

divergence and scattering slits of 0.05 . The a0 unit-cell para-

meter was estimated from the position of the (100) diffraction

√

line (a0 = d100 × 2/ 3) [25]. Pore wall thickness, δ, was

assessed by subtracting DP from the a0 unit-cell parameter

corresponding to the distance between the centers of adjacent

mesopores. UV–vis-NIR electronic spectra of the samples were

recorded in the wavelength range 200–2500 nm using a Cary

|

5E| spectrophotometer equipped with a diffuse reﬂectance at-

tachment. BaSO4 was used as reference. TPR experiments were

carried out in an automated ISRI-RIG-100 characterization sys-

tem equipped with a thermal conductivity detector. In the TPR

Ceq/Γ = 1/(Γm × K) + Ceq/Γm,

(3)

◦

2

experiments, the samples were pretreated in situ at 500 C for

was 2.15 Zr(IV) atoms per nm . This value is slightly below

2

h under air ﬂow and cooled in an Ar stream. The reduc-

the silanol surface density of ∼3.7 OH/nm that has been re-

tion step was performed with an Ar/H2 mixture, with a heating

ported previously for SBA-15 material [26]. From the above

results, it can be inferred that in the monolayer-like distribu-

tion of metal species reached at room temperature, each zirco-

nium atom is retained on the surface by one or two Zr–O–Si

◦

rate of 10 C/min, up to 1000 C. The chemical analysis of

the Zr-containing SBA-15 supports was performed by Desert

Analytics. HRTEM studies were performed using a Jeol 2010

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

143

linkages. In addition, the existence of some isolated hydrox-

yls inert to the interaction with zirconium n-propoxide at room

temperature also can be expected on the surface of the parent

SBA-15 material. To prove this hypothesis, an additional ex-

periment was performed in which Zr(IV) grafting was realized

from dry n-propanol solution at its boiling temperature (97 C)

for 8 h. In this case, zirconium(IV) loading (24.6 wt% of ZrO2)

was only slightly higher than that obtained at room tempera-

ture (23.4 wt% of ZrO2). Therefore, it can be assumed that

most surface hydroxyl groups of SBA-15 available for inter-

action with zirconium(IV) n-propoxide react with it at room

temperature. Previously it was suggested that SBA-15 materi-

als exhibit a “rough” inner surface containing different types

of silanol groups: “isolated” silanol groups, “interacting” SiOH

groups (silanols that are close enough to form H-bonds), “gemi-

nal” OH groups associated with surface defects (two OH groups

◦

2

attached to the same Q silicon atom), and “hidden” OH groups

located inside of small pores [26]. Because of space require-

ments of a Zr(IV) n-propoxide, accessible OH groups (groups

located on the inner surface of SBA-15 mesopores) can be ex-

pected to participate principally in the interaction with Zr(IV)

species. The probability of the reaction between Zr(n-PrO)₄

and “hidden” hydroxyls seems to be much lower. Therefore,

it can be assumed that a small amount of unreacted ≡Si–OH

groups can still be present inside SBA-15 micropores in the

samples covered with Zr(IV) monolayer at room temperature

and higher.

The textural and structural characteristics of Zr(x)SBA-15

materials are given in Table 1. The parent SBA-15 mater-

2

ial shows a surface area of 860 m /g, a micropore area of

2

3

1

39 m /g, and a total pore volume of 1.163 cm /g. Zirco-

nia grafting on the SBA-15 surface results in a decrease of

area and pore volume values. Thus, BET surface area decreases

2

to 544 m /g and micropore area decreases to 102 m /g after

Zr(IV) incorporation. Total pore volume and micropore volume

follow similar tendencies after incorporation of Zr(IV) into the

SBA-15 support. This decrease can be due to an increase in

the density of Zr(x)SBA-15 materials with ZrO wt loading.

2

In a previous report [24], we showed that Zr(IV) grafting does

not produce changes in the characteristic shape of the SBA-15

isotherm (type IV isotherm with an H1 hysteresis loop), with

the exception of some decrease in the amount of adsorbed N₂,

indicating that the original pore structure of the parent SBA-15

material is maintained after Zr(IV) incorporation. In addition,

small-angle XRD patterns of Zr(x)SBA-15 samples were found

to be similar to those of parent siliceous SBA-15 [24]. The po-

sitions of the reﬂections characteristic of the p6mm hexagonal

symmetry of SBA-15 and their intensities did not change after

Zr(IV) grafting, indicating that the long-range periodicity order

Fig. 1. (A) Variation of the surface concentration of Zr(IV) atoms on SBA-15

with equilibrium concentration of zirconium(IV) n-propoxide in dry ethanol at

◦

2

5 C. (B) Plot of Ceq/Γ against Ceq (linearized Langmuir plot).

Table 1

Textural and structural characteristics and chemical composition of Zr(x)SBA-15 materials

Sample

ZrO

Zr

SBET

(m /g)

Sμ

VP

(cm /g)

Vμ

DP

(Å)

a

δ

2

0

a

2

3

(wt%)

(atoms/nm )

(m /g)

(cm /g)

(Å)

SBA-15

SBA-15(HT)

Zr(6)SBA-15

Zr(13)SBA-15

Zr(17)SBA-15

Zr(21)SBA-15

Zr(22)SBA-15

Zr(23)SBA-15

Zr(23)SBA-15(HT)

0

6.2

13.4

17.1

21.1

22.5

23.4

n.d.

0

860

488

751

624

605

583

548

544

540

139

65

1.163

1.352

1.026

0.887

0.813

0.769

0.736

0.708

0.788

0.054

0.024

0.050

0.048

0.042

0.041

0.040

0.019

56

78

56

57

102

104

n.d.

102

103

46

26

n.d.

45

46

b

c

0.4

1.1

1.4

1.8

2.0

2.1

n.d.

125

110

108

105

102

54

b

a

Determined by chemical analysis.

Samples after hydrothermal treatment in boiling water for 24 h.

Not determined.

b

c

144

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

Fig. 3. Powder XRD patterns of Zr(23)SBA-15 sample calcined at different

temperatures for 1 h: (a) 550; (b) 700; (c) 800; (d) 900; and (e) 1000 C.

Fig. 2. Transmission electron microscopy image of the Zr(23)SBA-15 support.

◦

*

Tetragonal ZrO (JCPDS card 17-0923).

2

of the SBA-15 sample was preserved intact and that its unit cell

parameter and pore wall thickness did not change signiﬁcantly

nitrate solution followed by drying leads to the loss of long-

range ordering and a substantial decrease in the surface area and

pore volume [20]. The better hydrothermal stability of SBA-15

silica was attributed to its much thicker pore walls in com-

parison with MCM-41 [22,23]. However, siliceous SBA-15,

as well as other amorphous silicas, can suffer some structural

destruction of mesoporous framework in boiling water due to

the interaction of surface silanol groups with water molecules.

Figs. 4 and 5 show nitrogen adsorption–desorption isotherms

and small-angle XRD patterns of the SBA-15 and Zr(23)SBA-

15 samples before and after hydrothermal treatment in boiling

water for 24 h. In general, the N2 sorption isotherm of siliceous

SBA-15 maintains its shape after treatment in boiling water.

However, the amount of N2 adsorbed at relative pressures be-

low 0.6 decreases to almost half that of the untreated SBA-15.

In line with this ﬁnding, signiﬁcant decreases in both BET and

micropore surface areas are seen (Table 1). These decreases are

accompanied by a signiﬁcant increase in pore diameter from 56

to 78 Å for the sample reﬂuxed for 24 h. Nevertheless, small-

angle XRD results (Fig. 5) demonstrate that the thick-wall

SBA-15 material retains its structural integrity and long-order

pore arrangement after being reﬂuxed. No signiﬁcant change

was found in the unit-cell parameter value of SBA-15 after

reﬂuxing. The foregoing results demonstrate that a signiﬁcant

decrease in pore wall thickness (from 46 to 26 Å) occurs during

hydrothermal treatment.

(Table 1).

Fig. 2 shows a TEM micrograph of the Zr(23)SBA-15 sup-

port. A well-ordered hexagonal array of mesopores can be seen

for this sample when the electron beam is parallel to the main

axis of the cylindrical pores. When the electron beam is perpen-

dicular to the main axis, the cylindrical pores are viewed from

the side as a striped image. The distance between the centers

of two adjacent pores is about 100 Å, and the pore diameter is

about 60 Å. The image of the SBA-15 material modiﬁed with

a monolayer of grafted ZrO species is similar to that reported

2

previously for the parent SBA-15 material [24]. Formation of

ZrO agglomerates on the external surface or inside the meso-

2

pores of SBA-15 particles was not detected.

To assess the thermal stability of Zr(IV) species grafted on

the SBA-15 surface, the Zr(23)SBA-15 support was calcined at

◦

different temperatures (between 700 and 1000 C). The results

of powder XRD of the thermally treated samples, displayed

in Fig. 3, show that dispersed Zr(IV) species remain immobi-

◦

lized in SBA-15 at calcination temperatures below 800 C. The

formation of tetragonal ZrO2 crystalline phase (JCPDS card

1

7-0923) can be clearly seen in samples calcined at 800, 900,

◦

and 1000 C. According to the literature, the tetragonal zirconia

phase is metastable and easily transforms into the monoclinic

one, which is stable below 1170 C. Our results demonstrate

◦

that the immobilization of zirconia species by grafting on the

SBA-15 surface delays the crystallization of ZrO and leads to

stabilization of the tetragonal phase of zirconia at room tem-

perature. Previously, it was reported that such an effect can be

Incorporation of zirconia onto the SBA-15 surface by graft-

ing resulted in a signiﬁcant improvement in hydrothermal sta-

bility of the support, as is evident from Figs. 4 and 5. Only

a decrease in micropore area and a slight increase in the to-

tal pore volume were detected for the Zr(23)SBA-15 sample

after reﬂuxing (Table 1). In this case, the BET surface area

of the sample was maintained almost the same as before the

hydrothermal treatment, indicating that an increase in meso-

pore area at the expense of micropore area. It seems that only

some micropores are sensible for the presence of water mole-

cules. The enhanced stability of SBA-15 due to Zr(IV) grafting

can be attributed to the disappearance of most surface ≡Si–OH

groups, which can be easily attacked by water molecules, and

2

related to a high dispersion of ZrO particles [11,27].

2

The hydrothermal behavior of Zr(IV)-grafted SBA-15 ma-

terials was also characterized because of its importance for

the possible use of these materials in applications associated

with aqueous solutions. Previously it was reported that the hy-

drothermal stability of siliceous SBA-15 is much better than

that of MCM-41 which, when subjected to reﬂuxing in wa-

ter for short periods, readily loses its hexagonal structure and

is rendered amorphous [28,29]. Moreover, it was found that

even aqueous impregnation of MCM-41 material with cobalt

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

145

Fig. 4. Nitrogen adsorption–desorption isotherms of siliceous SBA-15 and Zr(23)SBA-15 supports before (a) and after treatment in boiling water for 24 h (b).

Fig. 5. Small-angle XRD patterns of siliceous SBA-15 and Zr(23)SBA-15 supports before (a) and after treatment in boiling water for 24 h (b).

the appearance of new ≡Zr–OH groups resistant to hydroly-

Table 2

Textural and structural characteristics of Mo and NiMo catalysts supported on

Zr(x)SBA-15 materials

sis with water. Similar stabilization of MCM-41 by introducing

3+

Al species onto its surface has been reported previously and

attributed to the creation of “protective surface composite lay-

er” preventing destruction of the silica wall [30].

Sample

SBET

Sμ

VP

Vμ

DP a0

δ

2

3

(m /g) (m /g) (cm /g) (cm /g) (Å) (Å) (Å)

Mo/SBA-15

601

460

416

394

550

82

60

53

49

71

53

44

39

0.851

0.642

0.592

0.596

0.784

0.646

0.552

0.537

0.031

0.022

0.019

0.018

0.026

0.019

0.016

0.014

55 102 47

56 102 46

54 102 48

55 102 47

54 100 46

54 102 46

Returning to the above results, it can be stated that the chem-

ical grafting procedure used in this work allows preparation

of Zr-containing SBA-15 materials with high loading (up to

Mo/Zr(17)SBA-15

Mo/Zr(21)SBA-15

Mo/Zr(23)SBA-15

NiMo/SBA-15

NiMo/Zr(17)SBA-15 450

NiMo/Zr(21)SBA-15 378

NiMo/Zr(23)SBA-15 367

2

3 wt%) of dispersed ZrO species with no signiﬁcant dam-

2

age to the initial SBA-15 structure. The presence of Zr(IV)

species grafted on SBA-15 surface signiﬁcantly improves the

hydrothermal stability of the mesoporous molecular sieve, mak-

ing it attractive for different applications in catalysis.

siliceous SBA-15 (with SBET decreasing to 30% for the Mo

catalyst and 36% for the NiMo catalyst) than for the catalysts

supported on ZrO2-containing materials. No changes were de-

tected in the form of N2 adsorption–desorption isotherms or in

the shape of the hysteresis loop after successive incorporation

of Mo and Ni species into the SBA-15 supports (Fig. 6).

In line with this, (100), (110), and (200) reﬂections of the

p6mm hexagonal structure of SBA-15 are still observed in the

3

.2. Mo and NiMo catalyst characterization

The textural characteristics of Mo and NiMo catalysts sup-

ported on SBA-15 molecular sieves are given in Table 2. A sig-

niﬁcant decrease in BET surface area and total pore volume

occurs after incorporation of Mo and Ni into the supports. This

decrease is more pronounced for the catalysts supported on

146

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

small-angle XRD patterns of Mo and NiMo catalysts (Fig. 7).

The intensity of these reﬂections decreases after Ni and Mo

incorporation (especially for pure silica SBA-15 support), prob-

ably because Mo is a strong X-ray absorber [21] or due to some

loss of long-range order of the support mesostructure induced

by a partial blocking of its pore channels by Ni and Mo oxidic

species. A low dispersion of molybdate species has been de-

tected for mesostructured silica-supported catalysts [2,21,31].

In line with this, powder XRD patterns of Mo and NiMo cat-

alysts supported on siliceous SBA-15 [Fig. 8, curves (a)] re-

veal the formation of crystalline MoO3 in orthorhombic phase

Incorporation of Zr(IV) oxide into the surface of SBA-15

seems to improve the dispersion of Mo and Ni species. A de-

crease in textural characteristics after incorporation of Ni and

Mo into Zr-containing SBA-15 supports was not as strong as

that for the siliceous SBA-15 (Table 2). Comparing the intensi-

ties of the small-angle XRD signals of Zr(23)SBA-15 support

with those of corresponding Mo and NiMo catalysts reveals that

the effect of Ni and Mo incorporation is much smaller in this

case than for the pure silica SBA-15 sample (Fig. 7). The fore-

going results allow us to suppose that Ni and Mo oxidic species

should be better dispersed on Zr(x)SBA-15 materials than on

the parent SBA-15. Results from powder XRD characterization

of Mo and NiMo catalysts (Fig. 8) conﬁrm this supposition.

The diffractograms of unpromoted Mo catalysts clearly show

(

JCPDS card 35-609). The intensity of the characteristic XRD

signals of MoO3 decreases with Ni addition. The average size

of the MoO3 crystals, determined using Scherrer’s equation, de-

creases from 961 Å for unpromoted Mo/SBA-15 catalyst to

an increase in MoO dispersion with zirconia loading in the

3

support. Low-intensity signals of the orthorhombic MoO3 crys-

talline phase can still be observed for the Mo catalyst supported

5

19 Å for the Ni-promoted one. The size of MoO3 crystals

compared with the pore diameter values of the used SBA-15

material clearly indicates that they should be located on the ex-

ternal surface of the support particles.

on SBA-15 with 6 wt% of grafted ZrO ; however, these reﬂec-

2

tions are not detected for unpromoted Mo catalysts supported

on SBA-15 materials with 13 wt% or higher ZrO loadings.

2

Adding nickel to the Mo/Zr(6)SBA-15 catalyst resulted in a fur-

ther increase in MoO dispersion. The foregoing results from

3

nitrogen physisorption and XRD characterization of Mo and

NiMo catalysts clearly show that both zirconium and nickel ox-

ides similarly improve the dispersion of Mo oxidic species on

the SBA-15 surface.

UV–vis DRS spectra of Mo, Ni, and NiMo catalysts (Fig. 9)

were recorded to obtain more information about Ni and Mo

6+

oxidic species. For Mo species, the absorption due to the

2

−

6+

O

4

→ Mo charge transfer transition occurs in the 200–

00 nm range. The exact position of this band depends on the

state of molybdenum coordination and aggregation [32–34];

the isolated (tetrahedral) molybdate species show absorption

at 250–280 nm, whereas polymolybdate (octahedral) species

absorb at 300–330 nm. In addition, both types of Mo species

exhibit a second strong absorption band at about 220–230 nm.

The DRS spectra of unpromoted Mo catalysts (Fig. 9) show the

Fig. 6. Nitrogen adsorption–desorption isotherms of: Zr(23)SBA-15 (a);

Mo/Zr(23)SBA-15 (b); and NiMo/Zr(23)SBA-15 (c).

Fig. 7. Small-angle XRD patterns of supports (a); Mo (b); and NiMo (c) catalysts supported on SBA-15 and Zr(23)SBA-15.

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

147

Fig. 8. Powder X-ray diffraction patterns of Mo and NiMo catalysts supported on SBA-15 (a); Zr(6)SBA-15 (b); and Zr(13)SBA-15 (c). * Orthorhombic MoO

3

(JCPDS card 35-609).

Fig. 9. UV–vis diffuse reﬂectance spectra of Mo, Ni, and NiMo catalysts supported on SBA-15 (a) and Zr(23)SBA-15 (b). The spectrum of each support was

subtracted from the spectrum of corresponding catalyst.

presence of a mixture of Mo⁶oxidic species in tetrahedral (Td )

and octahedral (Oh) coordinations. However, the proportions of

different types of Mo species change with the composition of

the support. Thus, in the spectrum of the catalyst supported on

+

increased dispersion of Mo species in all catalysts. This is in-

dicated by an increase in Eg values after the incorporation of

nickel to 3.5 eV for the NiMo/SBA-15 catalyst and 3.7 eV for

the catalyst supported on Zr(23)SBA-15. In addition to a blue

shift of the band-gap absorption edge, Ni incorporation also

produced a signiﬁcant change in the shape of the DRS spec-

trum of the Mo catalyst supported on siliceous SBA-15, namely

6+

siliceous SBA-15, absorption bands at 300–330 nm (O Mo

)

h

and 250 nm (Td Mo⁶) can be clearly observed. In the spectrum

of Mo/Zr(23)SBA-15 catalyst these signals are less deﬁned,

whereas absorption in 260–300 nm region increases its inten-

sity.

+

absorption bands at 300 nm (agglomerated polymeric O Mo

h

species) and 250 nm (tetrahedral Mo⁶species) become less

+

These changes conﬁrm that the addition of zirconium to the

SBA-15 support leads to an increase in the dispersion of O_h

molybdenum species and to a decrease in the proportion of the

Td species. The absorption edge energies (Eg) calculated for

Mo catalysts are also well in line with the above conclusion.

deﬁned, and absorption at 260–280 nm increases. The changes

in E value and the spectrum shape induced by Ni addition

g

were not so marked when ZrO2-containing support was used.

It seems that the effect of Ni addition on the characteristics of

the deposited Mo oxidic species is stronger for the pure silica

The E value of 3.2 eV found for Mo/SBA-15 catalyst is char-

SBA-15 support than for the ZrO -containing one.

g

2

acteristic of the crystalline MoO3 phase. Meanwhile, a higher

Eg value (3.6 eV) of Mo/Zr(23)SBA-15 catalyst indicates an in-

creased dispersion of polymolybdate species. The DRS spectra

of Ni/SBA-15 and Ni/Zr(23)SBA-15 catalysts are also shown in

Fig. 9. Only the presence of octahedral Ni species in the form

of NiO (bands at 415, 650, and 730 nm [35]) can be observed

for all catalysts, regardless of the support used. Comparing

the DRS spectra of unpromoted Mo catalysts and Ni-promoted

ones (Fig. 9) demonstrates that adding a Ni promoter leads to

The TPR results for Ni, Mo, and NiMo catalysts are shown

in Fig. 10. The TPR proﬁle of the Ni/SBA-15 sample exhibits a

broad low-temperature reduction region (300–550 C) that can

◦

be clearly ascribed to the reduction of octahedrally coordinated

2+

Ni species weakly bound to the silica surface [31]. These

nickel oxide species are reduced almost completely (98%) to

metallic nickel (Ni²→ Ni ) in a single reduction step. The

+

0

reduction proﬁle of the Mo/SBA-15 catalyst shows two main

reduction peaks at 572 and 740 C and a shoulder at about

◦

148

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

Fig. 10. TPR proﬁles for Ni (a), Mo, (b) and NiMo (c) catalysts supported on SBA-15 and Zr(23)SBA-15.

◦

6

14 C. The low temperature peak (572 C) is generally as-

of Mo catalyst supported on SBA-15 and Zr(23)SBA-15 shows

that zirconium(IV) grafting onto the SBA-15 surface leads to

a slight decrease in the temperature of reduction of octahedral

molybdenum species (the ﬁrst step in their reduction occurs at

6+

4+

sociated with the reduction of Mo to Mo of polymeric

octahedral Mo species weakly bound to silica surface (prob-

ably small clusters of MoO3) [36,37]. Comparing the present

TPR pattern with that of bulk MoO3 (main peak at 600–630 C

37]), hydrogen consumption near 614 C can be assigned to

◦

540 C; the second step, at 700 C). In addition, the incorpora-

tion of Zr(IV) into SBA-15 results in disappearance of a peak

◦

[

◦

the reduction of crystalline MoO3 phase detected by XRD. The

peak at about 740 C in the case of SBA-15-supported Mo cat-

with a maximum at 740 C observed in the case of Mo/SBA-15

◦

catalyst and ascribed to the ﬁrst step in the reduction of tetrahe-

dral Mo⁶species. Adding nickel to the Mo/Zr(23)SBA-15 cat-

alyst leads to a further decrease in the temperature of reduction

of Mo oxidic species, indicating that the effect of Ni promoter

is similar in the siliceous and zirconia-containing SBA-15 sup-

ports. It appears that in both cases, this is due to the better

dispersion of Mo species in presence of nickel.

+

alyst can be ascribed to the second step in the reduction of the

polymeric octahedral Mo species (from Mo⁴to Mo ) or to

+

0

6+

the ﬁrst step of reduction of isolated tetrahedral Mo species

in strong interaction with the support [37,38]. It should be men-

tioned that the intensity of this high-temperature reduction peak

in our Mo/SBA-15 catalyst is unusually high compared with

the TPR proﬁles of similar silica-supported catalysts prepared

by impregnation of ammonium heptamolybdate solutions (e.g.,

Mo/SiO2 [39], Mo/MCM-41 [31]). This can be attributed to

the presence of a high proportion of isolated Td molybdenum

species in strong interaction with the SBA-15 surface, probably

because of the particular characteristics of this material, such

as a high degree of roughness of its inner surface and a great

number of surface defects [26]. In the TPR proﬁle of NiMo

catalyst supported on SBA-15 the reduction of Mo species oc-

HRTEM characterization of sulﬁded Mo and NiMo catalysts

was performed to gain more insight into the dispersion of cat-

alytically active MoS2 species. Fig. 11 shows HRTEM micro-

graphs of Mo and NiMo catalysts supported on parent SBA-15

and Zr(23)SBA-15. The typical fringes due to MoS2 crystallites

with 6.1 Å interplanar distances were observed on micrographs

of all sulﬁded catalysts. Inhomogeneous dispersion of the MoS2

phase can be seen in the unpromoted Mo catalyst supported

on siliceous SBA-15 (micrograph (a), Fig. 11). MoS2 crystal-

lites with lengths between 25 and 80 Å and stacking from two

to seven layers are formed. In addition, some very large MoS2

crystals (600–2000 Å) were also found on the external surface

of SBA-15 particles. Nickel addition to this catalyst resulted in

better dispersion and more homogeneous distribution of MoS2

crystals (Fig. 11b). Particles with lengths between 40 and 60 Å

and stacking from 2 to 5 layers are predominant in this case.

Zirconium incorporation in the support leads, as expected, to

better dispersion of MoS2 active phase (Figs. 11c and 11d).

Figs. 12 and 13 show the slab length and stacking degree distri-

butions for Mo and NiMo catalysts supported on SBA-15 and

Zr(23)SBA-15. It can be clearly observed that ZrO2 addition in

the support leads to a decrease in the stacking degree and length

of MoS2 particles of unpromoted Mo catalysts. Short MoS2

particles (20–40 Å) with small number of layers (2 or 3) are

formed predominantly in the sulﬁded Mo/Zr(23)SBA-15 cata-

lyst. When Ni is added to the Mo/SBA-15 catalyst, morphology

of the sulﬁded catalyst (stacking degree and slab length) is also

changed signiﬁcantly. However, when Zr(23)SBA-15 is used

◦

curs at lower temperatures (450 and 714 C for the ﬁrst and

second steps of reduction of Oh Mo⁶species) than in the

+

corresponding unpromoted Mo/SBA-15 catalyst. In addition,

hydrogen consumption at 580–600 C becomes less intense, in

line with the increased dispersion of MoO3 species in the pres-

ence of the Ni promoter as detected by XRD.

◦

The TPR proﬁle of Ni catalyst supported on Zr-containing

SBA-15 is more complex than that of the Ni/SBA-15 sam-

ple (Fig. 10). Two poorly resolved broad reduction peaks can

be seen in the TPR of this sample, indicating the presence

of at least of two different Ni²surface species. The low-

+

◦

temperature H2 consumption region (maximum at 470 C) can

be assigned to the reduction of a NiO-like phase in weak in-

teraction with the support. The intensity of this peak decreases

after the incorporation of zirconia onto the SBA-15 surface. At

the same time, a new high-temperature reduction peak (maxi-

mum at 585 C) appears in the TPR proﬁle of Ni/Zr(23)SBA-15

sample as a result of the stronger interaction of nickel species

with zirconia-containing support. Comparing the TPR patterns

◦

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

149

Fig. 11. HRTEM micrographs of sulﬁded catalysts: Mo/SBA-15 (a); NiMo/SBA-15 (b); Mo/Zr(23)SBA-15 (c); and NiMo/Zr(23)SBA-15 (d).

(A)

(B)

Fig. 12. Layer stacking distribution of MoS crystallites in sulﬁded Mo (A) and NiMo (B) catalysts supported on SBA-15 (a); and Zr(23)SBA-15 (b).

2

as the support, the addition of Ni has only a slight effect on

the morphology of MoS2 particles. These results conﬁrm that

incorporation of zirconium into the support affects the disper-

sion of molybdenum active phase similar to that of a nickel

promoter. In both cases, a more homogeneous distribution of

smaller MoS2 particles is obtained. It seems that ZrO2 added

in the SBA-15 support acts as a textural promoter for Mo cata-

lysts.

3.3. Catalytic activity

In the present study, 4,6-dimethyldibenzothiophene, one of

the most refractory sulfur compounds in gasoil [40], was se-

lected as a model molecule to examine the catalytic activity of

Mo and NiMo catalysts in deep HDS. The HDS of methyl-

substituted DBT derivatives is known to occur through two

parallel reactions (Fig. 14): (i) direct desulfurization (DDS),

150

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

(A)

(B)

Fig. 13. Length distribution of MoS crystallites in sulﬁded Mo (A) and NiMo (B) catalysts supported on SBA-15 (a); and Zr(23)SBA-15 (b).

2

Fig. 14. Reaction scheme for the hydrodesulfurization of 4,6-DMDBT.

yielding the corresponding substituted biphenyl products, and

ii) hydrogenation with subsequent desulfurization (HYD),

creasing ZrO2 loading in the support, reaching a maximum with

the sample prepared using the SBA-15 covered with a mono-

layer of Zr(IV) species [Zr(23)SBA-15]. The activity trend with

(

yielding ﬁrst substituted tetrahydrodibenzothiophenes, then the

corresponding hexahydro- derivatives and, ﬁnally, cyclohexyl-

benzene-type compounds [41,42]. It has been shown that under

HDS conditions (i.e., in presence of an organic sulfur com-

pound), biphenyl-type products do not readily hydrogenate into

cyclohexylbenzene-type ones [43].

The conversions of 4,6-DMDBT obtained over Mo and

NiMo catalysts supported on SBA-15 and Zr(x)SBA-15 mate-

rials are shown in Fig. 15. It can be observed that 4,6-DMDBT

conversion reached at 8 h changes in a wide range. The low-

est conversion (25%) was obtained with unpromoted Mo cata-

lyst supported on siliceous SBA-15. The catalytic activities of

all Mo catalysts supported on Zr(x)SBA-15 were signiﬁcantly

greater than that of the reference Mo/SBA-15 catalyst. Thus,

the change in the ZrO content in the support seems to be the

2

result of the changes in the dispersion of Mo active species. As

reported above, the incorporation of zirconia onto the SBA-15

surface provided better dispersion to oxidic and sulﬁded molyb-

denum phases. In addition, an increase in the proportion of eas-

ily reduced octahedrally coordinated Mo⁶species was found

in Mo/SBA-15 catalysts after zirconia incorporation. Therefore,

an increase in the total amount of HDS active sites can be ex-

+

pected in Mo catalysts supported on ZrO -modiﬁed SBA-15

2

materials compared with the corresponding pure silica sup-

port. This explains the increase in 4,6-DMDBT conversion with

ZrO loading seen in the Mo/Zr(x)SBA-15 catalysts.

2

Further increases in catalytic activity were seen after the

incorporation of the Ni promoter into the Mo catalysts. 4,6-

DMDBT conversions of 57–94% were obtained at 8 h of re-

4

,6-DMDBT conversions of 40–58% were obtained at 8 h of

reaction time. Moreover, catalytic activity increased with in-

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

151

reaction product distributions at the same total 4,6-DMDBT

conversion (30%) were compared for different catalysts (Ta-

ble 3). The ratios of the principal products of HYD and

DDS routes (methylcyclohexyltoluene [MCHT] and dimethyl-

biphenyl [DMBP], respectively) from Table 3 show that for all

catalysts studied, the preferential pathway is HYD. The pro-

portion of MCHT increases with the incorporation of ZrO2

into the support for both Mo and NiMo catalysts. The maxi-

mum MCHT/DMBP ratios (4.82 for the Mo catalyst and 7.68

for the NiMo catalyst) were reached with the catalysts sup-

ported on SBA-15 with the highest ZrO2 loading (23 wt%).

These catalysts were the most active in the Mo and NiMo

series, respectively. Therefore, our results show that incorpo-

ration of ZrO2 into the SBA-15 support leads to increased

hydrogenation activity of Mo and NiMo catalysts, which pro-

motes the HYD route of HDS of 4,6-DMDBT. This result is

in line with a previous study [17] that found similar enhance-

ment of HYD activity of NiMo catalysts when zirconium was

added into mesoporous HMS silica support. In general, un-

promoted Mo catalysts supported on ZrO2-modiﬁed SBA-15

materials were found to be very active for the hydrogenation of

the initial 4,6-dimethyldibenzothiophene; however, they were

not able to efﬁciently eliminate sulfur from the hydrogenated

intermediates (THDMDBT and HHDMDBT) obtained in the

HYD route. Adding nickel to Mo catalysts resulted in a fur-

ther increase in the MCHT/DMBP ratio, reﬂecting the change

in the contributions of the DDS and the HYD pathways into the

formation of the corresponding desulfurized products. In ad-

dition, incorporation of nickel into Mo catalysts resulted in a

signiﬁcant change in the product distribution of the HYD path-

way (Table 3). In particular, the amounts of tetrahydro- and

hexahydrodimethyldibenzothiophenes were much smaller (2–3

times smaller) with Ni-promoted catalysts than with the corre-

sponding unpromoted ones. In line with this, the (THDMDBT

+ HHDMDBT)/MCHT ratio decreased signiﬁcantly after Ni

addition. Therefore, Ni increases the rate of sulfur elimination

from hydrogenated intermediates of the HYD route. Previously

it was shown that in the HDS of unsubstituted DBT, Ni en-

hances the rate of the direct desulfurization (DDS) pathway

involving cleavage of the C–S bond in the starting reactant

molecule [42]. Because the 4,6-DMDBT used in our study has

steric constraints that hinder the DDS pathway, the increase in

C–S bond cleavage was much more pronounced for the hy-

(

A)

(

B)

Fig. 15. 4,6-DMDBT conversions obtained over Mo (A) and NiMo (B)

catalysts supported on SBA-15 (a); Zr(6)SBA-15 (b); Zr(13)SBA-15 (c);

Zr(17)SBA-15 (d); and Zr(23)SBA-15 (e).

action time. The activity of NiMo catalysts followed the same

trend with zirconia loading in the support as described above for

unpromoted Mo catalysts; that is, the lowest 4,6-DMDBT con-

version was obtained with the catalyst supported on siliceous

SBA-15, and catalytic activity was increased with increasing

ZrO loading in the support. The 4,6-DMDBT conversions ob-

tained with all NiMo/Zr(x)SBA-15 catalysts were signiﬁcantly

higher than that obtained with the reference NiMo/γ -Al O₃

2

catalyst (61% at 8 h).

To elucidate the effect of Ni and ZrO2 addition in Mo/SBA-

5 catalysts on the reaction pathways of 4,6-DMDBT, the

1

Table 3

Reaction product compositions obtained over Mo and NiMo catalysts at 30% of total 4,6-DMDBT conversion

a

b

Catalyst

t

Product composition (%)

Product ratio

30

(

h)

TH

HH

DMBP

MCHT

DMBCH

(TH + HH)/MCHT

MCHT/DMBP

Mo/SBA-15

10

40.9

39.1

38.3

26.8

19.1

16.2

16.8

16.2

15.1

9.1

6.1

5.6

7.9

7.1

6.9

10.1

8.3

7.7

29.8

31.7

33.0

46.6

55.9

59.2

4.6

5.9

6.7

7.4

10.6

11.3

1.94

1.75

1.62

0.77

0.45

0.37

3.78

4.49

4.82

4.63

6.77

7.68

Mo/Zr(13)SBA-15

Mo/Zr(23)SBA-15

NiMo/SBA-15

NiMo/Zr(13)SBA-15

NiMo/Zr(23)SBA-15

5.7

4.4

4.5

2.5

1.8

a

Time in which 30% of 4,6-DMDBT conversion is reached.

b

TH (THDMDBT), tetrahydrodimethyldibenzothiophene; HH (HHDMDBT), hexahydrodimethyldibenzothiophene; DMBP, dimethylbiphenyl; MCHT, methyl-

cyclohexyltoluene; DMBCH, dimethylbicyclohexyl.

152

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

drogenated intermediates of 4,6-DMDBT, products with no

such steric hindrance as the initial 4,6-DMDBT. As a result,

Ni addition strongly improved the overall kinetics of the HYD

pathway of 4,6-DMDBT and produced only a slight increase in

the DDS route. The general ability of Ni promoter to enhance

the hydrogenolysis activity of Mo-based catalysts was previ-

ously attributed to different factors. The addition of nickel was

supposed to decrease the strength of the bond between molyb-

denum and the sulfur atoms resulting from decomposition of

the organic molecules or the metal–sulfur bond in the sulﬁde

itself ([42] and references therein]). Another possible factor is

based on the increased electronic density and thus the basic-

dimethyldibenzo thiophene. The activity trends with zirconia

loading in the support were similar for unpromoted and Ni-

promoted Mo catalysts. In both cases, increased overall catalyst

activity was observed with zirconia loading, which can be at-

tributed to better dispersion of the MoS active phase with the

2

corresponding increase in its effective surface. In addition, a

detailed analysis of product distributions showed that adding

ZrO to the SBA-15 support also led to an increase in the

2

catalysts’ hydrogenation ability. Nevertheless, all unpromoted

Mo catalysts, independent of the support used, were not able

to efﬁciently realize the C–S bond cleavage, leading to sulfur

elimination. Adding the nickel promoter resulted in a further

increase in the dispersion of Mo oxidic species, making their

reduction easier. However, the most important result of adding

nickel to the Mo/SBA-15 catalysts was electronic modiﬁcation

ity of the particular S²centers involved in the attack on the

hydrogen atom (in β position relative to the sulfur atom in the

organic molecule) leading to the C–S bond cleavage. It seems

that our results conﬁrm the conclusion of Bataille et al. [42]

that a typical C–S bond cleavage center on a promoted cata-

lyst should contain at least a promoter atom in the vicinity of

a sulﬁde anion.

Returning to the previous observations, it can be concluded

that zirconium(IV) incorporation into SBA-15 supports resulted

in two principal effects: (i) increased overall catalyst activity

due to better dispersion of Mo active species and (ii) enhanced

hydrogenation ability of the catalysts. However, unpromoted

Mo catalysts, even when supported on Zr-containing SBA-15

solids, are not able to efﬁciently realize the C–S bond cleav-

age, leading to sulfur elimination. The Ni promoter leads to

further increase in MoS2 dispersion, but the most important

effect of nickel is the creation of active sites responsible for

the C–S bond cleavage principally in prehydrogenated inter-

mediate species obtained in the HYD pathway of the reaction

−

of the active MoS phase, resulting in the creation of active

2

sites responsible for C–S bond hydrogenolysis principally in

pre-hydrogenated intermediates obtained in the HYD pathway

of the reaction. Finally, the greatest catalytic activity was ob-

tained with the NiMo/Zr(23)SBA-15 catalyst, which showed

the greatest hydrogenation ability induced by the incorporation

of zirconia into the SBA-15 support in combination with the

highest C–S bond cleavage activity enhanced by the addition of

nickel.

Acknowledgments

Financial support was provided by CONACYT-Mexico

(Grant 46354-Y). The authors thank M. Aguilar Franco, C. Sal-

cedo Luna, and I. Puente Lee for technical assistance with the

small-angle XRD, powder XRD, and HRTEM characteriza-

tions, respectively.

(Fig. 14). Finally, the excellent activity of NiMo/Zr(23)SBA-

1

5 catalyst for 4,6-DMDBT HDS can be attributed to its good

hydrogenation ability induced by the incorporation of zirconia

into SBA-15 support in combination with the high C–S bond

cleavage activity enhanced by the addition of nickel.

References

[1] K.-H. Choi, N. Kunisada, Y. Korai, I. Mochida, K. Nakano, Catal. To-

day 86 (2003) 277.

[

2] G. Murali Dhar, G. Muthu Kumaran, M. Kumar, K.S. Rawat, L.D. Sharma,

B. David Raju, K.S. Rama Rao, Catal. Today 99 (2005) 309.

3] M. Breysse, J.L. Portefaix, M. Vrinat, Catal. Today 10 (1991) 489.

4] M. Breysse, P. Afanasiev, C. Geantet, M. Vrinat, Catal. Today 86 (2003)

5.

4

. Conclusion

[

Zirconia-containing SBA-15 materials with different ZrO2

loadings (up to 23 wt%) were prepared by chemical grafting at

room temperature. No considerable damage of the initial SBA-

1

species were highly dispersed on the SBA-15 surface at both

low and high, ZrO2 loadings. The Zr–O–Si linkages formed

during the grafting procedure were strong enough to resist ther-

mal treatment at temperatures up to 800 C. The presence of

ZrO2 species on SBA-15 surface signiﬁcantly improved the hy-

drothermal stability of the material.

Results from XRD, DRS, TPR and HRTEM characteriza-

tions conﬁrm that incorporation of zirconia into the SBA-15

support improved the dispersion of oxidic and sulﬁded Mo

species, making their distribution more homogeneous. This

ﬁnding demonstrates that zirconia acts as a textural promoter

for Mo catalysts.

[5] H. Shimada, T. Sato, Y. Yoshimura, J. Hiraishi, A. Nishijima, J. Catal. 110

1988) 275–284.

(

5 structure was observed after zirconia incorporation. Zirconia

[6] J.C. Duchet, M.J. Tilliette, D. Cornet, L. Vivier, G. Perot, L. Bekakra,

C. Moreau, G. Szabo, Catal. Today 10 (1991) 579.

[7] M. Vrinat, D. Hamon, M. Breysse, B. Durand, T. des Courieres, Catal.

Today 20 (1994) 273.

◦

[8] M. Ji, P. Afanasiev, M. Vrinat, W. Li, C. Li, Appl. Catal. A 257 (2004)

57.

9] M. Jia, P. Afanasiev, M. Vrinat, W. Li, H. Xu, Q. Ge, Petrochem. Tech. 33

2004) 202.

10] M. Vrinat, M. Breysse, C. Geantet, J. Ramirez, F. Massoth, Catal. Lett. 26

1994) 25.

1

[

(

[

(

[11] J. Ramírez, R. Cuevas, P. Castillo, M.L. Rojas, T. Klimova, Bulg. Chem.

Commun. 30 (1998) 207.

12] F. Maugé, J.C. Duchet, J.C. Lavalley, S. Houssenbay, E. Payen, J. Grim-

[

blot, S. Kasztelan, Catal. Today 10 (1991) 561.

[

13] S. Damyanova, L. Petrov, P. Grange, Appl. Catal. A 239 (2003) 241.

14] E. Lecrenay, K. Sakanishi, I. Mochida, T. Suzuka, Appl. Catal. A 175

(1998) 237.

NiMo catalysts supported on ZrO2-containing SBA-15

molecular sieves showed high performance in the HDS of 4,6-

O.Y. Gutiérrez et al. / Journal of Catalysis 249 (2007) 140–153

153

[

15] M.C. Barrera, M. Viniegra, J. Escobar, M. Vrinat, J.A. de los Reyes,

[27] T. Klimova, M.L. Rojas, P. Castillo, R. Cuevas, J. Ramirez, Microporous

Mesoporous Mater. 20 (1998) 293.

[28] R. Mokaya, J. Phys. Chem. B 103 (1999) 10204.

[29] Q.-H. Xia, K. Hidajat, S. Kawi, Mater. Lett. 42 (2000) 102.

F. Murrieta, J. García, Catal. Today 98 (2004) 131.

16] M.S. Rana, S.K. Maity, J. Ancheyta, G. Murali Dhar, T.S.R. Prasada Rao,

Appl. Catal. A 268 (2004) 89.

17] N.G. Kostova, A.A. Spojakina, K. Jiratova, O. Solcova, L.D. Dimitrov,

L.A. Petrov, Catal. Today 65 (2001) 217.

18] P.J. Kooyman, P. Waller, A.D. van Langeveld, C. Song, K.M. Reddy,

J.A.R. van Veen, Catal. Lett. 90 (2003) 131.

[

30] S. Kawi, S.-C. Shen, Mater. Lett. 42 (2000) 108.

31] T. Klimova, M. Calderón, J. Ramírez, Appl. Catal. A 240 (2003) 29.

32] C.C. Williams, J.G. Ekerdt, J.-M. Jehng, F.D. Hardcastle, A.M. Turek, I.E.

Wachs, J. Phys. Chem. 95 (1991) 8781.

[

33] C.C. Williams, J.G. Ekerdt, J.-M. Jehng, F.D. Hardcastle, I.E. Wachs,

J. Phys. Chem. 95 (1991) 8791.

34] M. Henker, K.-P. Wendlandt, J. Valyon, P. Bornmann, Appl. Catal. 69

[

19] Z. Li, L. Gao, S. Zheng, Appl. Catal. A 236 (2002) 163.

20] A.Y. Khodakov, V.L. Zholobenko, R. Bechara, D. Durand, Microporous

Mesoporous Mater. 79 (2005) 29.

(1991) 205.

[

21] A. Sampieri, S. Pronier, J. Blanchard, M. Breysse, S. Brunet, K. Fajerw-

erg, C. Louis, G. Pérot, Catal. Today 107–108 (2005) 537.

22] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka,

G.D. Stucky, Science 279 (1998) 548.

23] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem.

Soc. 120 (1998) 6024.

[35] B. Scheffer, J.J. Heijeinga, J.A. Moulijn, J. Phys. Chem. 91 (1987) 4752.

[36] R. López Cordero, A. López Agudo, Appl. Catal. A 202 (2000) 23.

[37] R. López Cordero, F.J. Gil Llambias, A. López Agudo, Appl. Catal. 74

(1991) 125.

[38] S. Damyanova, A. Spojakina, K. Jiratova, Appl. Catal. A 125 (1995) 257.

[39] C. Louis, M. Che, M. Anpo, J. Catal. 141 (1993) 453.

[40] B.C. Gates, H. Topsøe, Polyhedron 16 (1997) 3213.

[41] C. Pophal, F. Kameda, K. Hoshino, S. Yoshinaka, K. Segawa, Catal. To-

day 39 (1997) 21.

24] O.Y. Gutiérrez, G.A. Fuentes, C. Salcedo, T. Klimova, Catal. Today 116

(2006) 485.

[

25] T. Yamada, H. Zhou, K. Asai, I. Honma, Mater. Lett. 56 (2002) 93.

26] I.G. Shenderovich, G. Buntkowsky, A. Schreiber, E. Gedat, S. Sharif,

J. Albrecht, N.S. Golubev, G.H. Findenegg, H.-H. Limbach, J. Phys.

Chem. B 107 (2003) 11924.

[42] F. Bataille, J.L. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire,

E. Schulz, M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409.

[43] P. Michaud, J.L. Lemberton, G. Pérot, Appl. Catal. A 169 (1998) 343.

Article Doi

Mo and NiMo catalysts supported on SBA-15 modified by grafted ZrO2 species: Synthesis, characterization and evaluation in 4,6-dimethyldibenzothiophene hydrodesulfurization

DOI: 10.1016/j.jcat.2007.04.014

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.jcat.2007.04.014

Full text of DOI:10.1016/j.jcat.2007.04.014

Products guided by the article

R&D Labs maybe for 103272-52-2

Relevant to this article

Hot Product