Behavior of NiMo/SBA-15 catalysts prepared with citric acid in simultaneous hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene

Journal of Catalysis 304 (2013) 29–46

Contents lists available at SciVerse ScienceDirect

Journal of Catalysis

journal homepage: w ww.elsevier.com/locate/jcat

Behavior of NiMo/SBA-15 catalysts prepared with citric acid in simultaneous

hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene

⇑

Tatiana E. Klimova , Diego Valencia, Juan Arturo Mendoza-Nieto, Patricia Hernández-Hipólito

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

a r t i c l e i n f o

a b s t r a c t

Article history:

In the present work, NiMo catalysts supported on SBA-15 were prepared using citric acid (CA) during the

synthesis. The objective of this work was to realize a comparative study of NiMoCA/SBA-15 catalysts pre-

pared under different conditions in order to get a deeper insight into the effect of the thermal treatment

and pH of the impregnation solution used on the catalytic behavior in deep hydrodesulfurization (HDS).

Catalysts were prepared by simultaneous impregnation of Ni and Mo species and CA, using impregnation

solutions of acidic or basic pH values (pH = 1 or 9, respectively). The speciation diagrams of Ni(II) and

Mo(VI) species in aqueous solution as a function of pH were established. Nicit⁴₂^ꢀcomplex was formed

in aqueous solution at pH = 9. After the impregnation, NiMoCA/SBA-15 catalysts were dried at 100 °C

and some of them were calcined at 500 °C in air atmosphere. Prepared catalysts were characterized by

thermogravimetric analysis (TGA/DTG), nitrogen physisorption, powder X-ray diffraction (XRD), UV–

Vis diffuse reﬂectance spectroscopy (UV–Vis DRS), temperature-programmed reduction (TPR), and

high-resolution transmission electron microscopy (HRTEM) and tested in simultaneous HDS of dibenzo-

thiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) in a batch reactor at 300 °C for 8 h.

XRD characterization showed that Ni and Mo oxide species were well dispersed in all catalysts prepared

with CA. In contrast, reference NiMo/SBA-15 catalysts prepared by co-impregnation of Ni and Mo species,

without the addition of CA, showed signals of crystalline phases: (NH₄)₄[Ni(OH)₆Mo₆O₁₈] 4H₂O after dry-

ing at 100 °C and NiMoO₄after calcination at 500 °C. HDS of DBT showed differences in activity and selec-

tivity of the catalysts depending on the pH of the impregnation solutions and the temperature at which

the catalysts were treated: NiMoCA/SBA-15 catalysts prepared from acidic impregnation solutions were

more active for HDS of DBT than those prepared using basic ones. Both dried and calcined catalysts pre-

pared at pH = 1 were selective toward the hydrogenation (HYD) route of hydrodesulfurization. However,

the selectivity of the catalysts prepared from basic solutions (pH = 9) was strongly affected by the ther-

mal treatment: dried catalyst was highly selective for the direct desulfurization (DDS) of DBT, whereas

the calcined one for the HYD route. NiMoCA/SBA-15 catalysts with high hydrogenation ability showed

high activity in hydrodesulfurization of 4,6-DMDBT.

Received 5 March 2013

Revised 12 March 2013

Accepted 13 March 2013

Keywords:

Deep hydrodesulfurization

NiMo catalysts

Citric acid

SBA-15

Dibenzothiophene

4,6-Dimethyldibenzothiophene

1. Introduction

HDS catalysts are the active phase (transition metal sulﬁde, gener-

ally MoS₂or WS₂), promoter (NiS or CoS), and support (c-Al₂O₃)

Nowadays, the demand of high-quality, low-sulfur transporta-

tion fuels is growing because of the necessity to solve environmen-

tal problems induced by SO_xemissions and improving air quality

[1,2]. Sulfur removal is also strongly desirable in order to increase

the durability of reﬁnery equipment and to minimize the poisoning

of the DeNO_xand combustor catalysts that use noble metals [2].

The main industrial process to accomplish this purpose is hydrode-

sulfurization (HDS), in which S-containing molecules react with

hydrogen in the presence of a heterogeneous catalyst at high tem-

peratures and pressures. The main components of the traditional

[3,4]. In order to attend the demand of production of low-sulfur-

containing fuels, many efforts are aimed to design novel hydro-

treating catalysts, highly active and selective especially for HDS

of the refractory polyaromatic sulfur compounds.

In general, there are different possibilities that allow to modify

the characteristics of HDS catalysts and as a consequence, their

behavior. For instance, using new supports (MCM-41, SBA-15, KIT-

6, etc.) with appropriate physicochemical properties such as large

surface area, good thermal and mechanical resistance, and the abil-

ity to disperse well NiMo and CoMo HDS active phases leads to

attractive HDS catalysts [5–7]. Some mesostructured silica supports

have been modiﬁed with different heteroatoms (P, Al, Ti, Zr, W, etc.)

to increase the activity and selectivity of NiMo and CoMo HDS

⇑

Corresponding author. Fax: +52 55 5622 5371.

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

http://dx.doi.org/10.1016/j.jcat.2013.03.027

30

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

catalysts [8–17]. Another route to enhance the catalytic activity re-

lies on correctly selecting metal precursors and solvents; the use of

Keggin-type heteropolyacids (H₃PMo₁₂O₄₀and H₃PW₁₂O₄₀) for

example, as distinct precursors for NiMo and NiW catalysts sup-

ported on SBA-15 has also led to good results [18]. A third option

consists in using chelating agents and organic ligands (L) in the

preparation of HDS catalysts [19]. It is well known that, under the

appropriate experimental conditions (solution pH, metal and ligand

concentrations, etc. [20]), chelating agents are able to coordinate

metals and thus increase solubility and stability of the metal precur-

sors used for the preparation of supported catalysts. CoMo and NiW

although no selectivity modiﬁcations were observed for any molar

ratio.

In summary, the results described above are sometimes contra-

dictory. This means that the behavior of HDS catalysts prepared

with ligands strongly varies with a number of factors: preparation

conditions and thermal treatment, nature of the support as well as

active phase and promoter used, model molecule used for catalytic

tests, etc.

In our group, a series of NiMo catalysts supported on SBA-15

were prepared with citric acid and tested in hydrodesulfurization

of DBT [35,36]. We selected this mesoporous molecular material

as a support because of its high hydrothermal stability, attractive

textural properties (high surface area and pore volume), and pore

size appropriate for the transformation of large organic molecules

such as dibenzothiophene-type compounds [18]. In work [35],

NiMo/SBA-15 catalysts were prepared varying the pH of the

impregnation solution and conditions of the thermal treatment

after the impregnation step. We observed that these two factors

modiﬁed activity and selectivity of the catalysts in HDS of DBT. A

detailed kinetic study undertaken in this work conﬁrmed this re-

sult. However, no clear explanation could be drawn for the changes

observed in catalytic behavior. In the following work [36], NiMo/

SBA-15 catalysts were prepared using basic impregnation solutions

containing different amounts of citric acid. These catalysts were

sulﬁded without previous calcination and showed very high selec-

tivity for the direct desulfurization of DBT. In both works, we found

that all catalysts prepared with citric acid contained a small

amount of carbon, which increased along with the amount of citric

acid used in the catalyst preparation. In the present work, we tried

to achieve a greater insight into the effect of citric acid on the char-

acteristics and behavior of NiMo/SBA-15 catalysts in deep HDS. For

this, NiMo(CA)/SBA-15 catalysts were prepared from impregnation

solutions of two pH values (pH = 1 or 9) and subjected to different

thermal treatments. We were interested in following the evolution

in the characteristics of Ni and Mo species upon the addition of cit-

ric acid to the solutions used for catalyst preparation, impregnation

of these solutions to the SBA-15 support and drying or calcination

treatment, up to the sulﬁdation of the catalysts. Our interest was to

look for the relationship between the characteristics of Ni and Mo

species at different steps of the catalyst preparation and their

activity and selectivity in simultaneous HDS of two model com-

pounds, dibenzothiophene (DBT), and 4,6-dimethyldibenzothioph-

ene (4,6-DMDBT).

catalysts supported on

c-Al₂O₃prepared with nitriloacetic acid

(NTA), ethylenediaminetetraacetic acid (EDTA) and cyclohexanedi-

aminetetraacetic acid (CyDTA) showed higher HDS and HYD activi-

ties than those prepared without the chelating agents [21]. In the

same work, however, the activity of NiMo/c-Al₂O₃catalysts almost

did not change when ligands were used. In work [22], high activity

in thiophene HDS was obtained for NiWS catalysts supported on sil-

ica prepared with the same ligands (EDTA, NTA, and CyDTA). It was

found that the addition of chelating agents results in a decrease in

the temperature of reduction of the active phase and an increase

in the temperature of sulﬁdation of Ni species. When Ni–L com-

plexes were used, both Ni and W were converted into sulﬁdes at

similar temperatures, leading to the formation of an active NiWS

phase [22]. CoMo catalysts supported on alumina and silica–alu-

mina prepared with chelating agents (EDTA, NTA) had better cata-

lytic performance when they were sulﬁded without previous

calcination [23] which was attributed to a better promotion of the

CoMoS phase [24,25]. Most of the above studies used HDS of thio-

phene or hydrogenation reactions of alkenes as catalytic tests. How-

ever, when CoMo-NTA catalyst supported on c-Al₂O₃was tested in

HDS of dibenzothiophene (DBT), one of the refractory model mole-

cules of gasoline and diesel fuels, no changes in activity or selectiv-

ity were detected [26].

Citric acid (CA) was employed to synthesize Co₂[Mo₄(C₆H₅O₇)₂

O₁₁] complex in aqueous solution at acidic pH to be used as a pre-

cursor of the CoMoS phase [27]. CoMo/c-Al₂O₃catalysts prepared

from this complex showed improved activity for deep HDS. Some

changes in the behavior of these materials were observed when

they were treated at different temperatures (in the range of 100–

500 °C) [28]. This was ascribed to the calcination of the complex

and formation of different Co–Mo–CA species adsorbed on

c-

Al₂O₃surface before sulﬁdation [29]. Catalysts supported on metal

oxides other than Al₂O₃or alumina modiﬁed with zeolites were

found to be active in hydrodesulfurization of refractory compounds

such as DBT and 4,6-dimethyldibenzothiophene (4,6-DMDBT). For

example, attractive results were obtained with CoMoP/HY-Al₂O₃

catalyst prepared with CA [30]. This catalyst diminished the sulfur

content below 16 ppm of S and had some advantages in resistance

to the presence of N-containing compounds reaching elimination

of N below 5 ppm at 350 °C. Addition of CA to this catalyst resulted

in a considerable increase in activity when commercial diesel was

used. This was ascribed to a better dispersion of the active phase

and the prevention of the sulﬁdation of Co at low temperatures

[31]. The optimal temperature for the thermal treatment of this

catalyst before sulﬁdation stage was 100 °C. At higher tempera-

tures (between 300 and 500 °C), the rate constants decreased

2. Experimental

2.1. Catalyst preparation

SBA-15 silica with hexagonal p6mm structure was prepared

according to a reported procedure [37,38] using the triblock

copolymer Pluronic P123 (M_av= 5800, EO₂₀PO₇₀EO₂₀, Aldrich) as

the structure-directing agent and tetraethyl orthosilicate (TEOS,

Aldrich, 99.999%) as the silica source. P123 copolymer (4 g) was

dissolved in water (30 g) and 2 M HCl (120 g) solution at 35 °C.

Then, TEOS (8.5 g) was added into the solution. The mixture was

stirred at 35 °C for 20 h and then aged at 80 °C for 48 h without

stirring. The solid product was recovered by ﬁltration, washed with

deionized water, and air-dried at room temperature. Calcination

was carried out in static air at 550 °C for 6 h.

NiMo catalysts supported on SBA-15 were prepared by a stan-

dard incipient wetness co-impregnation technique reported else-

where [39]. The calcined support was co-impregnated using

aqueous solutions of ammonium heptamolybdate, (NH₄)₆Mo₇O₂₄

ꢁ4H₂O (Merck, 99%), nickel nitrate, Ni(NO₃)₂ꢁ6H₂O (Baker), and

citric acid, C₆H₈O₇ꢁH₂O (Merck, 99.5%) at pH = 1 and 9. Acidic pH

[32]. Citric acid was also used in the preparation of CoMo/c-

Al₂O₃catalysts modiﬁed with boron. Addition of CA increased the

dispersion of Co and Mo species, resulting in a larger amount of

Co–Mo–S phase after sulﬁdation without active-site blocking by

cobalt sulﬁde clusters [33]. NiMo catalysts supported on ZrO₂–

TiO₂were prepared at acidic pH with the addition of EDTA or CA,

and they were sulﬁded without previous calcination [34]. For both

chelating agents, rate constants of DBT HDS increased. Better activ-

ity results were obtained when the Ni/CA molar ratio was 1:2,

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

31

was obtained just by dissolving the above compounds in deionized

water, whereas ammonium hydroxide solution was used to adjust

pH of basic impregnation solution to 9. In both cases, the molar ra-

tio Ni/Mo/CA = 0.5:1:1.3 was maintained. In addition, a reference

NiMo/SBA-15 catalyst was prepared by co-impregnation of nickel

nitrate and ammonium heptamolybdate, without CA addition

and adjusting pH. After co-impregnation, all catalysts were dried

at 100 °C for 12 h in static air atmosphere, and some of them were

calcined in air at 500 °C for 4 h. The nominal composition of the

catalysts was 12 wt.% of MoO₃and 3 wt.% of NiO. As all catalysts

were supported on SBA-15, hereinafter, we will denote them just

as NiMo and NiMoCA with following speciﬁcations: ‘‘d’’, for dried

catalysts; ‘‘c’’, for dried and then calcined catalysts; ‘‘1’’, for cata-

lysts prepared from solutions with acidic pH value (pH = 1); and

‘‘9’’, for catalysts prepared from solutions with basic pH value

(pH = 9).

divergence and scattering slits of 0.05°. UV–Vis electronic spectra

of the samples were recorded in the wavelength range of 200–

800 nm using a Varian Cary 100 conc spectrophotometer equipped

with a diffuse reﬂectance attachment. Polytetraﬂuoroethylene was

used as a reference. TPR experiments were carried out in a Microm-

eritics AutoChem II 2920 automatic analyzer equipped with a TC

detector. Before TPR experiments, the samples were pre-treated

in situ at 120 °C (for dried catalysts) and 400 °C (for calcined cata-

lysts) for 3 h under air ﬂow and cooled in an Ar stream. The reduc-

tion step was performed under a stream of an Ar/H₂mixture (90/

10 mol/mol and 50 ml/min), with a heating rate of 10 °C/min up

to 1000 °C. Carbon and sulfur contents in selected catalysts were

determined using automatic Perkin–Elmer 2400 Series II CHNS/O

analyzer with TC detector. HRTEM studies of sulﬁded catalysts were

performed using a Jeol 2010 microscope (resolving power 1.9 Å).

The solids were ultrasonically dispersed in n-heptane, and the sus-

pension was collected on carbon-coated grids. Slab length and layer

stacking distributions of MoS₂crystallites in each sample were

established from the measurement of at least 300 crystallites de-

tected on several HRTEM micrographs taken from different parts

of the same sample dispersed on the microscope grid.

2.2. Characterization of Ni(II) and Mo(VI) species in impregnation

solutions

Aqueous speciation diagrams were plotted using Medusa soft-

ware [40] for pure Ni(II) and Mo(VI) species and for the same species

in presence of CA in amount used in the present work. For all spe-

cies, molar concentration was used. The Ni(II) and Mo(VI) two

dimensional diagrams (predominance area) were plotted for a large

concentration range of each metal (M) to observe different species

in aqueous solutions. For M–CA diagrams, [Ni(II)]_aq= 0.167 M and

[Mo(VI)]_aq= 0.340 M were used. For the two dimensional diagrams,

CA concentrations were varied around experimental used value. In

addition, metal species in impregnation solutions were character-

ized by UV–Vis spectroscopy. The spectra were recorded in the

wavelength range 200–800 nm using a Varian Cary 100 conc spec-

trophotometer using 1-cm quartz cells.

All electron DFT calculations were employed to obtain the

structure of Nicit⁴₂^ꢀcomplex with the Perdew and Wang’s ex-

change and correlation functional (PW91) [41] with the TZP basis

set. Geometry optimization was performed with solvation COSMO

model, and water was used in DFT calculation. The calculations

were performed in the Amsterdam Density Functional (ADF)

2008 program [42].

2.4. Catalytic activity

The HDS activity tests were performed in a batch reactor at

300 °C and 7.3 MPa total pressure for 8 h with constant stirring.

Prior to the catalytic activity evaluation, the catalysts were sulﬁded

ex situ in a tubular reactor at 400 °C for 4 h in a stream of 15 vol.%

of H₂S in H₂under atmospheric pressure. The sulﬁded catalysts

(0.15 g) were transferred in an inert atmosphere (Ar) to a batch

reactor (Parr) with 40 mL of n-hexadecane solution of DBT (Aldrich,

1300 ppm of S) and 4,6-DMDBT (Aldrich, 500 ppm of S). The course

of the reaction was followed by withdrawing aliquots each hour

and analyzing them in an Agilent 6890A chromatograph. To cor-

roborate product identiﬁcation, the product mixture was analyzed

in a Hewlett Packard GC–MS instrument.

3. Results

3.1. Complexes in aqueous solution

2.3. Catalyst characterization

Figs. 1 and 2 show speciation diagrams for Ni(II) and Mo(VI) spe-

cies present in aqueous solutions at different pH values, in absence

and presence of citric acid. It is well known that citric acid is a polyp-

SBA-15 support and NiMo(CA)/SBA-15 catalysts were character-

ized by thermogravimetric analysis (TGA/DTG), N₂physisorption,

powder X-ray diffraction (XRD), UV–Vis diffuse reﬂectance spec-

troscopy (DRS), temperature-programmed reduction (TPR), and

high-resolution transmission electron microscopy (HRTEM). TGA

experiments were performed with dried samples (100 °C/12 h) on

a Mettler-Toledo TGA/SDTA 851^e, under static air atmosphere using

a heating rate of 10 °C/min from 40 to 1000 °C. N₂adsorption–

desorption isotherms were measured with a Micromeritics ASAP

2020 automatic analyzer at liquid N₂temperature. Prior to the

experiments, the samples were degassed (p < 10^ꢀ1Pa) at 100 °C

and 270 °C (for catalysts prepared at 100 and 500 °C, respectively)

for 6 h. Speciﬁc surface areas (S_BET) were calculated by the BET

method, the total pore volume (V_p) was determined by nitrogen

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

were obtained from the desorption isotherms by the BJH method.

The mesopore diameter (D_P) corresponds to the maximum of the

rotic acid with the negative logarithms of ionization constants (pK_a1

,

pK_a2, and pK_a3) equal to 3.13, 4.76 and 6.40, respectively. Fig. 1A

shows Ni(II) species present in solutions of different Ni(II) concen-

trations at different pH values. In the present work, we used impreg-

nation solutions with nickel concentration of 0.167 M, which

corresponds to the log[Ni(II)] = ꢀ0.77. It can be clearly seen in this

ﬁgure that for this concentration, Ni(II) is soluble only at acidic pH

values (pH < 5.5). At higher pH values, nickel species precipitate

forming nickel hydroxide Ni(OH)₂. When CA is added to 0.167 M

nickel nitrate solution, solubility of Ni(II) species increases due to

the formation of different Ni(II)–CA complexes (Fig. 1B). Structure

and composition of these complexes depend on the solution pH

and CA concentration [43]. In our impregnation solutions, concen-

tration of CA was 0.437 M, which corresponds to log[CA] = ꢀ0.36

(shown as a dotted line, Fig. 1B). For this concentration of CA, four

complexes can be formed depending on the pH. Thus, solvated

Ni²⁺is present at pH between 0 and 1.8, NiH₂cit⁺, NiHcit and

Ni(Hcit)(cit)^3ꢀcomplexes are formed at intermediate pH (between

1.8 and 6.5), and Nicit⁴₂^ꢀis formed at pH interval between 6.5 and 9.5

(Fig. 1B). In the present work, we used impregnation solutions with

acidic and basic pH values (1 and 9, respectively). According to the

pore size distribution. The micropore area (S ) was estimated using

l

the correlation of t-Harkins and Jura (t-plot method). XRD patterns

were recorded in the 3° 6 2

fractometer, using Cu K radiation (k = 1.5406 Å) and a goniometer

speed of 1°(2 Small-angle XRD (2 0.5–10°) was

) min^ꢀ1

performed on a Bruker D8 Advance diffractometer using small

H 6 70° range on a Siemens D5000 dif-

a

H

.

H

32

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

1.0

0

-2

-4

-6

(A)

0.5

0.0

- 0.5

-1.0

-8

-10

2

4

6

8

10

12

2

4

6

8

10

12

pH

1.0

0.5

0.0

1.0

0.5

(B)

0.0

-0.5

-1.0

- 0.5

- 1.0

-1.5

-2.0

2

4

6

8

10

12

2

4

6

8

10

12

pH

Fig. 1. Predominance area diagrams for aqueous solution of (A) Ni(II) and (B) Ni(II)–

CA. [Ni(II)]_aq= 0.167 M, [CA]_aq= 0.437 M. Dotted lines correspond to Ni(II) and CA

concentrations used in the present work.

Fig. 2. Predominance area diagrams for aqueous solution of (A) Mo(VI) and (B)

Mo(VI)–CA. [Mo(VI)]_aq= 0.340 M, [CA]_aq= 0.437 M. Dotted lines correspond to

Mo(VI) and CA concentrations used in the present work.

diagram shown in Fig. 1B, no Ni(II)–CA complex should be formed in

our acidic impregnation solution, whereas in the basic impregna-

tion solution, the Nicit⁴₂^ꢀcomplex should be present. It seems that

the principal advantage of using CA during catalyst preparation is

stabilization of the Ni(II) impregnation solutions in a wide range

of pH due to formation of Ni(II)–CA complexes and prevention of

Ni(OH)₂precipitation at pH values above 6.

Mo(VI) diagrams (Fig. 2) also exhibit changes in the type and

dispersion of Mo species present in aqueous solutions of different

Mo(VI) concentrations and pH values. At acidic pH (pH < 4), molyb-

denum is present in solution as hydrated MoO₃, heptamolybdates

are predominant at pH between 4 and 8, and molybdate is formed

at pH above 8 (Fig. 2A). This is well in line with previously pub-

lished results [44,45]. When CA is added to a Mo(VI) aqueous solu-

tion, there are also some changes in the type of molybdenum

species at intermediate pH values. Thus, MoO₃(H₂cit)^ꢀ, MoO₃

(Hcit)^2ꢀ, and MoO₃(cit)^3ꢀcomplexes are formed at high CA concen-

tration (Fig. 2B). In our work, the concentration of citric acid used

corresponds to log[CA] = ꢀ0.36. For this concentration of CA, there

is no formation of complexes between CA and Mo(VI) at pH values

equal to 1 and 9. At acidic pH (between 0 and ꢂ2.5), molybdenum

should be present in impregnation solutions as hydrated MoO₃,

whereas in basic conditions (at pH > 8), MoO²₄^ꢀshould be a

predominant species (Fig. 2B). Resuming the above information

from speciation diagrams for Ni(II) and Mo(VI) species (Figs. 1

and 2), we were expecting that under our experimental conditions,

at an acidic pH, in both absence and presence of citric acid, there

should be the same Ni(II) and Mo(VI) species and that no com-

plexes should form between Ni and Mo cations and citric acid.

On the other hand, at basic pH of the impregnation solution

(pH = 9), in absence of CA, nickel species should precipitate form-

ing nickel hydroxide Ni(OH)₂, whereas in presence of CA solubility

of Ni(II) should increase due to the formation of a Nicit⁴₂^ꢀcomplex.

In order to conﬁrm these predictions, UV–Vis absorption spec-

tra of nickel nitrate and ammonium heptamolybdate aqueous solu-

tions were recorded. These spectra are very sensitive to the Ni and

Mo coordination, and the replacement of water by other ligands

(complex formation) can be clearly distinguished. The concentra-

tions of Ni and Mo species, as well as of CA, were the same as used

in the preparation of our catalysts, and pH values of solutions were

1 or 9. Nickel nitrate spectra showed three main absorption bands

(Fig. 3): two sharp bands at 300 and around 375–400 nm and one

broad band in the 550–800 nm region. The absorption band at

300 nm is common for all electronic spectra shown in Fig. 3A,

and it is due to the

terion. Two other bands are assigned to the two spin-allowed

p–

p^ꢃtransition of uncomplexed nitrate coun-

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

33

1.2

1.0

0.8

0.6

0.4

0.2

0.0

A

300

(a)

(b)

375

(d)

398

629

650

730

200

300

400

500

600

700

800

(nm)

Fig. 4. Structure of Nicit⁴₂^ꢀcomplex calculated by DFT.

5

4

3

2

1

0

B

absorption band at 230–260 nm is assigned to the isolated molyb-

date species in tetrahedral coordination (Mo (Td)), whereas the

band at 270–330 nm corresponds to polymeric octahedral Mo(VI)

species (Mo (Oh)). In addition, both types of Mo(VI) species show

the second absorption band at about 210–220 nm. In the case of

our impregnation solutions, it can be clearly seen in Fig. 3B that

at acidic pH (curves a and b), octahedral Mo(VI) species are present

with the absorption edge energy (E_g) ꢂ 3.6 eV, which can be hy-

drated MoO₃species. The signals observed for the solutions at

pH = 9, with and without CA, point out the presence of tetrahedral

molybdate species (MoO²₄^ꢀ) with E_gvalue near to 4.4 eV. For both

pH values studied, there was almost no difference between the

spectra of solutions with and without CA. Therefore, no Mo(VI)–

CA complexes were formed in basic or acidic solutions. It should

be mentioned that the above results are in good agreement with

predictions made on the basis of predominance area diagrams

(Fig. 2).

(c)

(a)

(d)

(b)

200

250

300

350

(nm)

400

450

500

Fig. 3. UV–Vis spectra of the solutions of (A) nickel nitrate and (B) ammonium

heptamolybdate: (a) solution at pH = 1 without citric acid, (b) solution at pH = 1

with citric acid, (c) solution at pH = 9 without citric acid, and (d) solution at pH = 9

with citric acid.

The tentative structure of the Nicit⁴₂^ꢀcomplex, obtained by

quantum chemical methods using a solvation model with water,

is reported in Fig. 4. The Ni(II) atom is octahedrally coordinated

to six O atoms of two citrate anions. This result conﬁrms observa-

tions made by UV–Vis spectroscopy for basic Ni(II)–CA solutions

with molar ratio Ni/CA = 1:2.6 (Fig. 3A) regarding an octahedral

symmetry around the metal atom. Nicit⁴₂^ꢀcomplex is more stable

in basic aqueous solutions than its free Ni²⁺counterpart due to

the chelating effect of citrate.

3

d–d transitions, A_2g! T_1gðPÞ and A_2g! T_1gðFÞ, of octahedrally

coordinated Ni²⁺ions (d⁸) [46,47]. In both UV–Vis spectra of nickel

nitrate solutions at acidic pH (curves a and b, Fig. 3A), with and

without citric acid, the same bands corresponding to octahedral

nickel species [Ni(H₂O)₆]²⁺can be observed: a sharp band at

398 nm and a broad band with two maxima at about 650 and

730 nm. This indicates that no Ni(II)–CA complex is formed in

the aqueous solution at pH = 1. However, when the pH of the solu-

tion of nickel nitrate and citric acid was adjusted to pH = 9, remark-

able changes in bands’ positions and intensity were observed.

Thus, the sharp band shifted from 398 to 375 nm, and the broad

band transformed into a more symmetrical band with just one

maximum at 629 nm. A blue shift of the above electronic transi-

tions in the spectrum of Ni(II)–CA solution at basic pH is consistent

with the formation of a Nicit₂^4ꢀcomplex and with the fact that cit-

rate is a stronger ligand than H₂O.

The UV–Vis spectra of ammonium heptamolybdate solutions

(Fig. 3B) clearly demonstrate that the characteristics of Mo(VI) spe-

cies in aqueous solution strongly depend on pH. Since Mo⁶⁺has a

d⁰electronic conﬁguration, only the absorption corresponding to

ligand-to-metal charge transfer (LMCT) O^2ꢀ? Mo⁶⁺is observed

in the 200–400 nm range. The exact position of these bands reﬂects

the local symmetry around the Mo(VI) species depending on their

coordination and aggregation state [48]. As reported elsewhere, the

3.2. Catalyst characterization

TGA/DTG results obtained for uncalcined NiMo and NiMoCA

catalysts supported on SBA-15 are shown in Fig. 5. NiMo catalyst

prepared without citric acid (NiMo(d) sample) lost about 7% of

its weight at the temperature interval 25–500 °C. Weight loss in

the low-temperature range (25–120 °C) can be attributed to the

elimination of physisorbed water and ammonia from uncalcined

samples, whereas at higher temperatures, it is due to decomposi-

tion of deposited metal salts (nickel nitrate and ammonium hepta-

molybdate). Both catalysts prepared with CA at acidic or basic pH

showed similar TGA/DTG proﬁles with total weight loss of around

24%. This weight loss was much higher than that for the NiMo(d)

sample. In the TGA/DTG proﬁles of the catalysts prepared with cit-

ric acid, two weight loss steps can be clearly seen (Fig. 5); the ﬁrst

one (from room temperature to ꢂ150 °C) is due to evaporation of

34

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

0

-5

0.000

-0.001

-0.002

-0.003

-0.004

-0.005

-0.006

of the catalysts at 500 °C in an air atmosphere. To conﬁrm this,

chemical analysis of the calcined NiMoCA catalysts was performed

with a CHNS/O analyzer. Only small amounts of residual carbon

were found in the catalysts after calcination: 0.59 wt.% of carbon

in NiMoCA(1-c) sample and 0.77 wt.% in NiMoCA(9-c). This result

shows that during calcination, most of the citric acid used in the

preparation of the catalysts was eliminated by combustion, and

only small amounts of carbonaceous material, probably coke, were

left.

(a)

-10

-15

-20

-25

The textural properties of NiMo(CA) catalysts are given in Table

1. The SBA-15 support had a BET surface area equal to 850 m²/g,

micropore area of 140 m²/g, total pore volume of 1.09 cm³/g, and

pore diameter of 6.5 nm [35]. All NiMo(CA) catalysts showed lower

speciﬁc textural characteristics. Thus, dried and calcined NiMo cat-

alysts prepared without citric acid had surface areas of around

600 m²/g. This decrement can be related to an increase in the sam-

ple’s density after the incorporation of Ni and Mo precursors in the

SBA-15 support. After the calcination of the NiMo(d) sample, one

could expect an increase in the textural characteristics due to

dehydration and decomposition of precursor salts. However, this

did not occur. After calcination of the NiMo(d) sample, a slight de-

crease was observed in BET and micropore areas and pore volume.

So, some agglomeration of Ni and Mo oxide species and pore block-

ing was expected in the calcined NiMo(c) sample. Regarding dried

NiMoCA catalysts, their S_BETand other speciﬁc textural characteris-

tics were much lower than those of the NiMo(d) reference, which

can be attributed to a larger increase in the samples’ density due

to the addition of CA during catalyst preparation. Calcination of

both dried NiMoCA catalysts at 500 °C resulted in an increase in

the surface area and porosity. These results seem to be reasonable

considering that CA is eliminated by combustion during calcination

of the NiMoCA samples in an air atmosphere, increasing pore vol-

ume and available surface area. Both dried and calcined NiMoCA

catalysts prepared using acidic impregnation solutions showed lar-

ger surface areas than those of the corresponding samples pre-

pared from basic impregnation solutions. This result cannot be

ascribed to the difference in the samples’ density, since their

behavior in TGA/DTG was quite similar. Therefore, this must be

due to some differences in the characteristics of Ni and Mo oxide

species in the catalysts prepared from acidic and basic impregna-

tion solutions.

(b)

(c)

200

400

600

800

1000

Temperature (°C)

Fig. 5. Thermogravimetric analyses of uncalcined NiMo(CA) catalysts supported on

SBA-15 (air atmosphere, TGA – solid lines, DTG – dotted lines): (a) NiMo(d); (b)

NiMoCA(1-d); and (c) NiMoCA(9-d).

physisorbed water and ammonia, and the second one (between

150 and 500 °C) can be ascribed to decomposition of Ni and Mo

precursors and elimination of CA by decomposition and combus-

tion in air. This assignment of the peaks was conﬁrmed by DTA re-

sults (not shown), in which an endothermic signal was observed

for low temperatures (below 150 °C) and an exothermic one for

the temperature interval between 150 and 500 °C. The difference

in weight loss between the reference NiMo catalyst prepared with-

out citric acid and those prepared with CA is about 17%, which cor-

responds to the weight of CA incorporated in the catalysts. So TGA

results indicate that decomposition of citric acid starts at 150 °C

and ends at about 500 °C, and that almost all CA added to the cat-

alysts is eliminated in an air atmosphere at temperatures up to

500 °C. It is also worth mentioning that there are almost no differ-

ences among the TGA proﬁles of NiMoCA catalysts prepared using

basic or acidic impregnation solutions, although in one of the cases

a Nicit⁴₂^ꢀcomplex is formed, and not in the other one. Therefore,

the formation of the Ni(II)–CA complex did not affect the temper-

ature of elimination of citric acid. Previously, it has been reported

that thermal treatment is very important for the supported NiMo

and CoMo catalysts prepared with ligands (EDTA, NTA, CA): It

was observed that catalysts sulﬁded without previous calcination

resulted in better HDS activity [23,28]. According to the above

TGA results, in the present work, we selected two temperatures

(100 and 500 °C) for thermal treatment of the prepared catalysts.

We were expecting that all the CA added during the preparation

of the catalysts will be maintained in the samples after drying at

100 °C (in the form of a Nicit⁴₂^ꢀcomplex or not, depending on

pH), and almost all the CA will be decomposed after calcination

The nitrogen adsorption–desorption isotherm of the SBA-15

support is shown in Fig. 6 (black squares). It corresponds to a type

IV isotherm characteristic for mesoporous materials, with a H1

hysteresis loop characteristic of a well-formed SBA-15 silica

[37,38]. The shape of the isotherm and of the hysteresis loop did

not suffer important changes after the incorporation of NiMo(CA)

species (Fig. 6). Isotherms of all catalysts, dried and calcined,

shown in Fig. 6 are of type IV with a H1 hysteresis loop. This indi-

Table 1

Textural characteristics of NiMo(CA) catalysts and their reduction behavior.

a

b

Sample

S_BET(m²/g)

S

(m²/g)

V_p(cm³/g)

D_p^c(nm)

TPR results

l

H₂consumption (mmol/g)

a^d

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

613

598

393

597

340

501

108

83

60

73

64

55

0.84

0.78

0.67

0.84

0.62

0.85

6.0

5.8

6.0

6.1

6.0

1.71

1.86

2.79

2.69

2.64

2.37

0.59

0.64

0.96

0.93

0.91

0.82

a

b

c

Speciﬁc surface area calculated by BET method.

Micropore area estimated using the correlation of t-Harkins and Jura (t-plot method).

Pore diameter corresponding to the maximum of the pore size distribution calculated from the desorption isotherm by the BJH method.

d

a

, degree of the reduction of Ni and Mo oxide species determined from H₂consumption measured for each sample and the theoretical value corresponding to their

complete reduction (2.90 mmol/g).

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

35

800

700

600

500

400

300

200

100

0

(100)

(A)

(B)

(A)

(110)

(200)

(a)

(b)

(c)

(d)

(b)

(c)

(d)

1.0 1.5 2.0 2.5 3.0

2

(°)

2

(°)

0.0

0.2

0.4

0.6

P/P⁰

0.8

1.0

Fig. 7. Small-angle XRD patterns of (a) SBA-15 support and dried (A) and calcined

(B) catalysts: (b) NiMo, (c) NiMoCA(1), and (d) NiMoCA(9).

800

33-0948). Similar results have already been reported for the

NiMo/SBA-15 catalysts prepared by co-impregnation of ammonium

heptamolybdate and nickel nitrate from aqueous solutions [49].

Formation of the above or any other crystalline phases was not ob-

served in the NiMo catalysts prepared with citric acid. Ni and Mo

species were well-dispersed in both dried and calcined catalysts

prepared from acidic or basic impregnation solutions containing

CA. This illustrates the advantage of using citric acid during catalyst

preparation by incipient wetness co-impregnation technique.

In order to obtain more information regarding the dispersion of

Ni and Mo oxide species in the prepared catalysts, UV–Vis diffuse

reﬂectance spectroscopy (DRS) and temperature-programmed

reduction (TPR) were used. UV–Vis DR spectra for NiMo(CA)/SBA-

15 catalysts are shown in Fig. 9. In these spectra, absorption bands

corresponding to ligand-to-metal charge transfer (LMCT)

O^2ꢀ? Mo⁶⁺can be observed in the 200–400 nm region [50]. Posi-

tions of the bands corresponding to octahedral (Oh) and tetrahedral

(Td) Mo(VI) species have already been described previously. It can

be seen in Fig. 9A that in the spectrum of the dried NiMo catalyst

prepared without CA (curve (a)), signals corresponding to different

types of Mo species are present. Thus, the absorption band with the

maximum at 250 nm evidences the presence of tetrahedral Mo spe-

cies, whereas the signal at 300–320 nm is characteristic for octahe-

drally coordinated polymolybdate species. The latter ones showed

UV absorption edge energy (E_g) of 3.3 eV, which is close to the pre-

viously reported value for ammonium heptamolybdate [50]. A mix-

ture of different Mo species is also present in the calcined NiMo

catalyst (curve (a), Fig. 9B). However, the proportion of tetrahedral

Mo species, Mo (Td), increases in comparison with agglomerated

octahedral ones, Mo (Oh). This change is well in line with powder

XRD results, which showed the formation of a (NH₄)₄[Ni(OH)₆Mo_6-

O₁₈]ꢁ4H₂O crystalline phase in the dried NiMo/SBA-15 catalyst and

its transformation into NiMoO₄in the calcined one (Fig. 8). It can

be mentioned that in the ﬁrst compound, Mo(VI) is coordinated

octahedrally, whereas in the second one, its coordination changes

to a tetrahedral one. In addition, some well-dispersed octahedral

Mo species which absorb in the range of 260–300 nm are present

in the calcined NiMo/SBA-15 catalyst. The spectra of dried NiMoCA

catalysts prepared from acidic or basic impregnation solutions sur-

prisingly resulted to be very similar between each other (Fig. 9A),

although the metal species in the solutions used for catalyst prepa-

ration were quite different (Fig. 3). In both spectra (curves (b and c),

Fig. 9A), two overlapping absorption bands can be observed; the

ﬁrst one with the maximum at 210–220 nm, and the second one

in the 230–330 nm range. These signals can be ascribed to the

presence of tetrahedral or well-dispersed octahedral Mo species.

(B)

700

600

500

400

300

200

100

0

0.0

0.2

0.4

0.6

P/P⁰

0.8

1.0

Fig. 6. Nitrogen adsorption–desorption isotherms of (j) SBA-15 support, (h) dried,

and (s) calcined NiMo catalysts: (A) catalysts prepared without CA and (B)

NiMoCA(1) catalysts.

cates that the characteristics of the original pore structure of the

SBA-15 support were conserved in the catalysts. However, the ef-

fect of calcination was different for the NiMo sample prepared with

and without citric acid. The amount of N₂adsorbed by calcined

NiMoCA catalysts was signiﬁcantly larger than that adsorbed by

their dried counterparts, which is in line with textural character-

ization results from Table 1.

Fig. 7 shows small-angle XRD patterns of the SBA-15 support

and NiMo(CA) catalysts. All of them exhibit three well-resolved

peaks corresponding to (100), (110), and (200) reﬂections associ-

ated with p6mm hexagonal symmetry of mesopores, characteristic

of SBA-15 materials. Preservation of the same three signals in the

XRD patterns of dried and calcined NiMo(CA) catalysts points out

that long-range pore arrangement of the SBA-15 support was

maintained in all catalysts after deposition of Ni and Mo species

and citric acid.

Fig. 8 shows powder XRD patterns of the SBA-15 support and

NiMo(CA) catalysts. It can be observed that crystalline phases are

present in both dried and calcined NiMo catalysts prepared by co-

impregnation of Ni and Mo precursors without addition of citric

acid. Thus, an Anderson-type heteropolymolybdate with formula

(NH₄)₄[Ni(OH)₆Mo₆O₁₈]ꢁ4H₂O (JCPDS card 01-070-9672) was

formed in the NiMo(d) catalyst (Fig. 8A, curve b). After calcination

at 500 °C for 4 h, this phase was transformed into NiMoO₄(JCPDS

36

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

40

(A)

(c)

35

(b)

30

25

20

15

(a)

10

5

0

200

250

300

350

400

450

500

(nm)

50

45

40

35

30

25

20

15

10

5

(B)

(b)

(B)

(c)

(a)

0

200

250

300

350

400

450

500

(nm)

Fig. 8. XRD patterns of (a) SBA-15 support and dried (A) and calcined (B) catalysts:

(b) NiMo, (c) NiMoCA(1), and (d) NiMoCA(9).

(JCPDS card 01-070-9672) and d NiMoO₄(JCPDS 33-0948).

Fig. 9. UV–Vis DR spectra of dried (A) and calcined (B) catalysts: (a) NiMo, (b)

NiMoCA(1), and (c) NiMoCA(9).

j

(NH₄)₄[Ni(OH)₆Mo₆O₁₈]ꢁ4H₂O

and 663 °C. The low-temperature reduction peak (360 °C) can be

attributed to the ﬁrst step of reduction (Mo⁶⁺+ 2e^ꢀ? Mo⁴⁺) of

octahedral Mo species well-dispersed on the SBA-15 surface [51–

53]. The second peak at 467 °C can be associated with the same

step of reduction of more agglomerated octahedral Mo⁶⁺species,

probably of molybdenum species which are forming an Ander-

son-type heteropolymolybdate, (NH₄)₄[Ni(OH)₆Mo₆O₁₈]ꢁ4H₂O, de-

tected by powder XRD. A symmetric and well-deﬁned form of

this peak supports this assumption [52]. Finally, a high-tempera-

ture reduction peak (663 °C) can be ascribed to the second step

of reduction of the octahedral Mo species of different degrees of

agglomeration (Mo⁴⁺+ 4e^ꢀ? Mo⁰) and to the ﬁrst step of reduc-

tion of isolated tetrahedral Mo⁶⁺species in weak interaction with

the support [51]. So, the above reduction proﬁle of NiMo(d) cata-

lyst supported on SBA-15 indicates the presence of different types

of Mo oxide species, tetrahedrally and octahedrally coordinated

ones with different degrees of agglomeration, which is well in line

with the results of XRD and DRS characterization of this sample.

Reduction proﬁles obtained for NiMoCA(1-d) and NiMoCA(9-d)

showed fewer signals: only one well-deﬁned peak with a maxi-

mum at a low temperature (333 and 368 °C, respectively) and a

shoulder at a higher temperature. The high-temperature reduction

peak observed for the NiMo(d) catalyst at 663 °C was not detected

in the TPR proﬁles of the catalysts prepared with citric acid. This

result allowed us to assume that when CA was added in the

E_gvalues determined from curves (b and c) were about 3.8 eV,

which points out to a better dispersion of octahedral Mo species

in the catalysts prepared with citric acid in comparison with the ref-

erence dried NiMo/SBA-15 sample. After the calcination of NiMoCA

catalysts, they also showed similar electronic spectra (curves (b and

c), Fig. 9B), from which the presence of Mo (Td) and highly dispersed

Mo (Oh) can be inferred. The only difference was that in the spec-

trum of the calcined catalyst prepared from a basic pH impregnation

solution, a signal of tetrahedral Mo(VI) (ꢂ250 nm) was slightly bet-

ter deﬁned. The above DRS results make evident that the character-

istics of the metal species deposited on the SBA-15 surface change

when citric acid is added in the impregnation solution.

TPR characterization results for NiMo(CA) catalysts supported

on SBA-15 are shown in Fig. 10. Results obtained for catalysts dried

at 100 °C (chart A, Fig. 10) are difﬁcult to interpret exactly, since

decomposition of deposited metal salts (ammonium heptamolyb-

date and nickel nitrate) as well as of citric acid can take place dur-

ing the TPR experiment. Therefore, not all TCD signals observed in

these TPR proﬁles can be attributed to hydrogen consumption due

to reduction of Ni and Mo oxide species. Nevertheless, some differ-

ences can be seen among the catalysts prepared with and without

CA. Thus, dried NiMo/SBA-15 catalyst prepared without CA shows

hydrogen consumption in a broad temperature interval (between

300 and 700 °C) with three main reduction peaks at 360, 467,

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

37

addition of citric acid resulted in a more homogeneous distribution

of octahedral Mo oxide species.

333

368

(A)

Results from the quantiﬁcation of the total amount of hydrogen

consumed during the reduction of the catalysts are shown in Table

1. It can be seen that the reduction of the samples prepared with

citric acid proceeds with higher consumption of hydrogen than

of the reference NiMo(d) and NiMo(c) catalysts. Degree of reduc-

368

(c)

tion (a) of NiMoCA samples was in the range of 0.82–0.96, which

indicates that the majority of Ni and Mo oxide species were re-

duced in the TPR experiments. For comparison, NiMo(d) and Ni-

470

467

Mo(c) catalysts showed

a values of about 0.6. The above results

(b)

(a)

point out that the addition of citric acid during catalyst preparation

not only results in a decrease in the temperature of reduction of Ni

and Mo oxide species (Fig. 10), but also makes their reduction more

complete. This fact can be explained by a better dispersion of Ni

and Mo oxide species in the catalysts prepared with citric acid

and an increase in the proportion of Mo⁶⁺ions in octahedral coor-

dination, which can be reduced much easier than the correspond-

ing tetrahedral ones.

360

663

200

400

600

800

1000

Temperature (^oC)

HRTEM characterization of sulﬁded NiMo(CA)/SBA-15 catalysts

was performed to obtain more information about the dispersion of

MoS₂crystallites in the catalysts prepared with and without citric

acid. Representative micrographs selected for each one of the cat-

alysts are shown in Fig. 11. On the left side of this ﬁgure, NiMo(CA)

catalysts are shown, which were dried at 100 °C for 12 h and then

sulﬁded. On the right side of the ﬁgure, similar catalysts are shown,

which were at ﬁrst dried in the same conditions and then calcined

at 500 °C for 4 h prior to sulﬁdation. The typical fringes due to

MoS₂crystallites with 0.61 nm interplanar distances located in

and out of SBA-15 mesopore canals were observed in micrographs

of all sulﬁded catalysts. Reference NiMo(d) and NiMo(c) catalysts

prepared without citric acid showed MoS₂crystallites of different

lengths (between 2 and 10 nm) and with variable stacking degrees

(the number of stacking layers changed between 2 and 8). Addition

of citric acid during catalyst preparation, in general, resulted in a

better dispersion of MoS₂particles. As it can be seen in Fig. 11c–

f, all NiMoCA catalysts prepared from impregnation solutions of

different pH values, dried or calcined before sulﬁdation, had smal-

ler and less-stacked MoS₂particles. Thus, short (2–4 nm) MoS₂

crystallites with 1–3 layers can be observed after sulﬁdation of

dried NiMoCA catalysts prepared from acidic or basic impregnation

solutions (Fig. 11c and e, respectively). When NiMoCA catalysts

were calcined at 500 °C before sulﬁdation, MoS₂particles of similar

length were obtained (micrographs (d) and (f), Fig. 11); however,

the number of stacked layers in them slightly increased. This qual-

itative observation was conﬁrmed by calculation of length and

stacking degree distributions for NiMo(CA) catalysts (Fig. 12) and

by the determination on their basis of average morphology (aver-

age length and number of layers) of MoS₂crystallites in different

catalysts (Table 2). Fig. 12 clearly shows a decrease in MoS₂slabs

length and stacking for all catalysts prepared with CA, indepen-

dently on the pH value (1 or 9) of the impregnation solution used

and thermal treatment (drying, or drying and calcination). In addi-

tion, length and stacking distributions for all catalysts prepared

with citric acid were narrower than those for the reference

NiMo/SBA-15 samples prepared without CA. NiMoCA(9-d) and

NiMoCA(1-d) catalysts showed MoS₂particles of similar lengths

(2.8–2.9 nm). However, a stacking degree of the sulﬁded Mo spe-

cies was slightly lower in the catalyst prepared using a basic

impregnation solution than in the NiMoCA(1-d) counterpart pre-

369

(B)

(c)

365

(b)

(a)

390

360

420

620

600

200

400

800

1000

Temperature (^o

C)

Fig. 10. TPR proﬁles of (A) dried and (B) calcined catalysts: (a) NiMo, (b)

NiMoCA(9), and (c) NiMoCA(1).

impregnation solutions, only octahedral Mo species easy to reduce

were formed in the catalysts.

The TPR results for calcined NiMo(CA)/SBA-15 catalysts are

shown in Fig. 10B. It can be appreciated that after calcinations,

NiMo catalyst prepared without CA (curve a) continues showing

a complex TPR proﬁle. A low-temperature reduction peak with a

maximum at 390 °C showed two shoulders: at about 360 and

420 °C, which points out the coexistence in this sample of octahe-

dral Mo species of different degrees of agglomeration [51–53]. The

signal at 620 °C can be attributed to the second step of reduction of

octahedral Mo species and to the ﬁrst step of reduction of small

MoO₃clusters. Finally, low-intensity hydrogen consumption in

the 700–1000 °C region corresponds to the reduction of tetrahedral

Mo species detected by XRD as a NiMoO₄crystalline phase. After

calcination of the NiMoCA catalysts in an air atmosphere and elim-

ination of citric acid, only one well-deﬁned peak with the maxi-

mum at 365–369 °C was observed in the TPR proﬁles of both

NiMoCA(1-c) and NiMoCA(9-c). The position of this signal corre-

sponds to the ﬁrst step of reduction of dispersed octahedral Mo

species. It is interesting to notice again that it seems that the char-

acteristics of Mo species in calcined catalysts prepared from acidic

and basic impregnation solutions are similar, independently of the

pH of the impregnation solution used. In both cases (pH = 1 and 9),

pared from the acidic impregnation solution (Table

2 and

Fig. 12). Comparison of the average morphology of MoS₂active

phase in the NiMoCA catalysts dried just at 100 °C before sulﬁda-

tion and in the similar catalysts dried and then calcined at 500 °C

before sulﬁdation (Table 2) points out to a slight increase in the

number of stacked layers in MoS₂particles after calcination (from

38

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

(a)

(b)

(d)

(f)

(c)

(e)

Fig. 11. HRTEM micrographs of sulﬁded catalysts supported on SBA-15: (a) NiMo(d), (b) NiMo(c), (c) NiMoCA(1-d), (d) NiMoCA(1-c), (e) NiMoCA(9-d), and (f) NiMoCA(9-c).

1.8–2.1 layers in dried catalysts to 2.7 layers in the calcined ones,

Table 2). No signiﬁcant differences were observed in the average

length and stacking values obtained for both previously calcined

and then sulﬁded NiMoCA catalysts prepared from impregnation

solutions of pH = 1 or 9. Only a small difference in stacking degree

distributions can be seen for these catalysts in Fig. 12, namely, the

sulﬁded NiMoCA(1-c) sample showed 2–4-layered MoS₂slabs,

whereas in the NiMoCA(9-c) sample, mostly 2- and 3-layered slabs

were observed. The above HRTEM observations seem to be in line

with previously described results of characterization of dried and

calcined precursors of sulﬁded NiMoCA catalysts. Thus, it was de-

scribed above that the addition of citric acid improves dispersion

of Ni and Mo precursor species in dried and calcined NiMoCA cat-

alysts, and in accordance with this, sulﬁded NiMoCA catalysts also

showed a better dispersion on the SBA-15 surface than the refer-

ence samples prepared without CA (Table 2).

on the edge surface of MoS₂crystallites (f_Mo, Table 2), which can be

considered as an indicator of the amount of sulﬁded Mo species

located on the catalytically active surface of MoS₂particles [54].

The f_Mofraction was calculated using equations reported in [17],

assuming that the MoS₂crystallites are perfect hexagons [54].

The f_Movalues reported for NiMo(CA) catalysts in Table 2 show a

noticeable increase (ꢂ40%) in the proportion of Mo species on

the catalytically active surface of sulﬁded catalysts prepared using

citric acid. Thus, reference NiMo catalysts prepared without CA had

lower fractions of Mo atoms on the edge surface (f_Mo= 0.26 for

NiMo(d) and 0.29 for NiM(c), respectively) than those for the

NiMoCA counterparts (f_Mobetween 0.37 and 0.40).

In order to determine the existence of some carbonaceous

material resulting from citric acid after sulﬁdation, freshly suﬁded

catalysts were analyzed on a Perkin–Elmer CHNS/O analyzer.

Results shown in Table 2 indicate that all catalysts prepared with

CA contain a small amount of carbon (ꢂ0.5–2.0 wt.%). Carbon con-

tent in the NiMoCA(1-c) and NiMoCA(9-c) catalysts slightly

Results from the HRTEM characterization of sulﬁded NiMo(CA)

catalysts were used to determine the average fraction of Mo atoms

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

39

90

80

70

60

50

40

30

20

10

0

55

(A)

50

45

40

35

NiMo(d)

NiMoCA(1-d)

NiMoCA(9-d)

30

NiMoCA(1-d)

NiMoCA(9-d)

25

20

15

10

5

0

< 2

2-4

4-6

6-8

> 8

1

2

3

4

5

6

7

8

Length (nm)

Number of layers

90

80

70

60

50

40

30

20

10

0

50

45

40

35

30

25

20

15

10

5

(B)

NiMo(c)

NiMoCA(1-c)

NiMoCA(9-c)

NiMoCA(1-c)

NiMoCA(9-c)

0

< 2

2-4

4-6

6-8

> 8

1

2

3

4

5

6

7

8

Length (nm)

Number of layers

Fig. 12. Length and stacking distributions of MoS₂crystallites in sulﬁded NiMo(CA) catalysts supported on SBA-15. (A) Catalysts dried at 100 °C and (B) catalysts calcined at

500 °C.

Table 2

ever, some remaining carbon was also found in these catalysts

after sulﬁdation. In addition, it can be noted from the above results

that the amount of remaining carbon was higher for the samples

prepared from basic impregnation solutions, in which a Nicit₂^4ꢀ

complex was formed, than for similar samples prepared from

acidic impregnation solutions containing citric acid. Regarding sul-

fur content in the sulﬁded catalysts tested, in all samples prepared

with citric acid, it was around 6.2–6.6 wt.% (Table 2). This is very

close to the theoretically estimated amount of sulfur (6.62 wt.%)

required for complete transformation of impregnated molybde-

num and nickel species into MoS₂and NiS, respectively. Sulfur con-

tent in the reference NiMo/SBA-15 catalysts prepared without CA

was lower (ꢂ4.8–5.6 wt.%). This result shows that the degree of

sulﬁdation of Ni and Mo species was noticeable higher in the NiMo

catalysts prepared with the addition of citric acid, probably, be-

cause of a better dispersion of Ni and Mo oxide species, making

their reduction and sulﬁdation easier (Tables 1 and 2).

Results from characterization of sulﬁded NiMo(CA) catalysts supported on SBA-15 by

HRTEM and chemical analysis: average length (L) and stacking degree (N) of the MoS₂

particles, estimated fraction of Mo atoms on the edge surface of MoS₂crystallites

(f_Mo), and carbon and sulfur contents.

Catalyst

HRTEM

Chemical analysis

L (nm)

N

f_Mo

Carbon (%)

Sulfur (%)

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

4.5

4.0

2.9

3.0

2.8

3.1

3.9

3.7

2.1

2.7

1.8

2.7

0.26

0.29

0.39

0.38

0.40

0.37

n. d.

0.97

0.53

1.99

0.58

4.79

5.57

6.61

6.49

6.63

6.24

decreased after their sulﬁdation (from 0.59 to 0.53 and from 0.77

to 0.58 wt.%, respectively). This indicates that the carbonaceous

residues, remaining in the catalysts after calcination at 500 °C in

an air atmosphere, continue being present in the sulﬁded catalysts.

Catalysts, which were sulﬁded after drying at 100 °C (without pre-

vious calcination in air atmosphere), had a larger amount of resid-

ual carbon (around 1 and 2 wt.% of C for NiMoCA(1-d) and

NiMoCA(9-d), respectively). In this case, most of the citric acid con-

tained in the dried NiMoCA(1-d) and NiMoCA(9-d) samples was

decomposed into gaseous species and lost during sulﬁdation. How-

3.3. Catalytic activity

Catalytic behavior of the synthesized and characterized NiMo

(CA) catalysts supported on SBA-15 was evaluated in simultaneous

hydrodesulfurization (HDS) of two model compounds, dibenzo-

thiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT).

40

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

Fig. 13. Reaction network for hydrodesulfurization of DBT-type compounds.

These S-containing aromatic molecules were selected because they

are representative for diesel fuel and, in addition, have quite differ-

ent reactivity toward two possible routes of HDS. It is well known

that HDS of dibenzothiophene-type compounds can occur by the

two parallel reaction routes (Fig. 13): in the ﬁrst one, called the di-

rect desulfurization route (DDS), sulfur elimination takes place via

hydrogenolysis of C–S bonds, yielding biphenyl-type compounds;

whereas in the second one, the hydrogenation pathway (HYD),

desulfurization occurs after hydrogenation of one of the aromatic

rings of DBT molecule, yielding ﬁrst tetrahydro- and hexahydroge-

nated intermediates, and then cyclohexylbenzene- and dicyclo-

hexyl-type products. Nowadays, it is well documented that

unsubstituted DBT prefers the DDS route, although it can also react

through the HYD route over catalysts with high hydrogenation

ability or upon appropriate experimental conditions [55]. On the

contrary, the HYD pathway is the principal route for the HDS of

dibenzothiophenes with alkyl substituents in positions 4 and 6,

such as 4,6-DMDBT [56].

The DBT conversions obtained over NiMo(CA) catalysts at 2, 4,

and 8 h reaction time are shown in Table 3. The lowest conversions

(around 70% at 8 h reaction time) were obtained with the NiMo(d)

and NiMo(c) catalysts prepared without citric acid. All NiMoCA/

SBA-15 catalysts prepared with citric acid showed higher catalytic

activity than the reference NiMo(d) and NiMo(c) samples, espe-

cially, the catalysts prepared from impregnation solutions with

acidic pH. This can be clearly observed from DBT conversions ob-

tained at 2- and 4-h reaction times (Table 3). Thus, 98–99% of

DBT conversion was obtained at 8 h with these catalysts. NiMoC-

A/SBA-15 catalysts prepared from basic impregnation solutions

(pH = 9) showed slightly lower DBT conversions (81–85% at 8 h)

than those prepared from solutions with pH = 1. Calcination of

the NiMoCA(1) and (9) catalysts before sulﬁdation almost did not

change DBT conversions in comparison with the counterparts that

were just dried before sulﬁdation. The above observations can also

be conﬁrmed by comparing the initial reaction rates and TOF val-

ues reported in Table 3. Initial reaction rates obtained over NiMoC-

A(1-d) and (1-c) catalysts were ꢂ30–40% higher than those of the

NiMoCA(9) samples and twice higher than those of the reference

NiMo/SBA-15 samples prepared without citric acid. All the above

results show that the addition of citric acid in the impregnation

solutions used in the preparation of HDS catalysts had a beneﬁcial

effect on their activity. An interesting fact is that CA addition

caused a larger increase in the catalytic activity of the NiMo/

SBA-15 catalysts when they are prepared at pH 1 than at pH 9. This

is ever more surprising taking into account that at acidic pH values

of impregnation solutions, no Ni(II)–CA or Mo(VI)–CA complexes

were formed, whereas at basic conditions, Nicit⁴₂^ꢀwas detected.

In addition, average morphology of sulﬁded MoS₂phase was found

to be very similar for the catalysts prepared from acidic and basic

impregnation solutions (Table 2), with no signiﬁcant difference in

the fraction of Mo atoms located on the catalytically active surface

of MoS₂crystallites (f_Mo).

Table 4 shows the reaction product distributions obtained with

different NiMo(CA) catalysts at 50% of DBT conversion. As it was

expected, biphenyl (BP) and cyclohexylbenzene (CHB) were the

principal reaction products accompanied with small amounts of

tetrahydrodibenzothiophene (4HDBT) and dicyclohexyl (DCH).

The NiMoCA(9-d) catalyst supported on SBA-15 the only one

among the catalysts tested which showed a very high proportion

of BP in the reaction products. The amount of BP obtained with this

catalyst at 50% of DBT conversion was ﬁve times higher than that of

CHB (Table 4). All other NiMo(CA) catalysts showed signiﬁcantly

higher hydrogenation ability leading to the formation of compara-

ble amounts of BP and CHB (Table 4). For NiMo(d) and NiMoCA(1-

d) catalysts, CHB/BP ratios obtained at 50% of DBT conversion were

below 1 (ꢂ0.6–0.8), indicating a slight preference for the formation

of BP through the direct desulfurization route of the reaction. After

calcination in air atmosphere, all NiMo(CA) catalysts showed an in-

crease in their hydrogenation ability. Thus, NiMo(c) catalyst gave a

CHB/BP ratio of nearly one, whereas NiMoCA/SBA-15 catalysts pre-

pared with citric acid showed higher proportion of CHB in compar-

ison with BP (CHB/BP ratios ꢂ1.2–1.5, Table 4). This result seems to

be reasonable and can be attributed to the changes in the average

morphology of the catalysts after calcination determined by

HRTEM (Table 2). Fig. 14 shows reaction product distributions ob-

tained in DBT HDS with different catalysts of the NiMoCA series in

function of DBT conversion. Again, it can be seen from this ﬁgure

that the principal product obtained with the NiMoCA(9-d) catalyst

was BP for all the range of DBT conversions studied (up to 85%),

whereas the NiMoCA(1-d) catalyst desulfurized DBT through both

the DDS and HYD routes giving similar amounts of BP and CHB up

to 80% of DBT conversion. Calcined NiMoCA catalysts showed a

similar behavior, being CHB the main reaction product at conver-

sions higher than 30–40%. Summarizing the results shown in Table

4 and Fig. 14, it can be concluded that the pH value of the impreg-

nation solution used in the preparation of the NiMoCA catalysts

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

41

Table 3

Conversions of DBT obtained over NiMo(CA)/SBA-15 catalysts at different reaction times, initial reaction rate (r_o), and turnover frequency (TOF).^a

À

Á

TOF ꢄ 10³(s^ꢀ1

)

r_oꢄ 10⁷mol g s^ꢀ1

ꢀ1

Catalyst

DBT conversion (%)

2 h

cat

4 h

8 h

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

22

21

41

39

26

25

45

44

78

55

50

74

71

98

99

85

81

2.4

2.3

4.7

3.1

3.0

1.1

1.0

1.5

0.9

1.0

a

TOF was calculated as a number of DBT molecules reacted per second per one Mo atom located on the edge surface. The number of Mo atoms located on the edge surface

was estimated using the total amount of Mo in the catalysts and the fraction of Mo atoms on the edge surface of MoS₂particles (f_Mo) determined for each catalyst by HRTEM.

Table 4

Composition of products obtained in DBT hydrodesulfurization over NiMo(CA)/SBA-15 catalysts (at 50% of DBT conversion).

Product^b(%)

BP

Product ratios

a

Catalyst

t₅₀(h)

4HDBT

CHB

DCH

BP

4HDBT

CHB

BP

4HDBT

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

4.5

4.6

2.3

2.4

3.6

4.0

57.2

46.7

52.3

41.5

80.6

36.3

5.5

5.7

4.1

4.6

4.1

4.9

35.5

44.7

40.7

50.4

15.2

54.9

1.8

2.9

3.5

0.1

3.9

10.4

8.2

12.8

9.0

19.7

7.4

0.15

0.13

0.10

0.09

0.27

0.09

0.62

0.96

0.78

1.21

0.19

1.51

a

Time required to reach 50% of DBT conversion.

BP = biphenyl; 4HDBT = tetrahydrodibenzothiophene; CHB = cyclohexylbenzene; DCH = dicyclohexyl.

b

60

(A)

60

50

40

30

20

10

0

(B)

50

40

30

20

10

0

20

40

60

80

100

0

20

40

60

80

100

DBT conversion (%)

60

50

40

30

50

40

30

20

10

(C)

(D)

20

10

0

20

40

60

80

0

20

40

60

80

DBT conversion (%)

Fig. 14. Product selectivity for hydrodesulfurization of DBT over (A) NiMoCA(1-d), (B) NiMoCA(1-c), (C) NiMoCA(9-d), and (D) NiMoCA(9-c). Detected products: (d) biphenyl

(BP), (s) cyclohexylbenzene (CHB), (j) dicyclohexyl (DCH), and (h) tetrahydrodibenzothiophene (4H-DBT).

42

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

has a strong effect on the selectivity of dried and sulﬁded catalysts,

whereas this effect was minimal when the catalysts were calcined

at 500 °C before sulﬁdation. In addition, it can be noted that ther-

mal treatment is very important for the catalysts prepared from

basic impregnation solutions (pH = 9) in which a Nicit⁴₂^ꢀcomplex

was formed. Selectivity of these catalysts changed in almost eight

times after calcination: from CHB/BP ratio of 0.19 for the NiMoC-

A(9-d) to 1.51 for the NiMoCA(9-c) sample.

all catalysts tested (Table 6). The highest proportion of DMBP

(11.3%) was obtained with the NiMoCA(9-d) catalyst. It seems that

low overall activity of this catalyst in HDS of 4,6-DMDBT is due to

its low hydrogenation ability. Both calcined NiMoCA(1-c) and (9-c)

catalysts showed higher hydrogenation ability than other tested

samples (see MCHT/DMBP ratios, Table 6). Among them, NiMoC-

A(1-c) showed slightly higher catalytic activity than the NiMoC-

A(9-c) one as well as higher hydrogenation ability (Tables 5 and

6). Comparison of dried and calcined NiMoCA catalysts also

showed an increase in the hydrogenation ability of the catalysts

after calcination, which is in line with observations made in HDS

of DBT. Regarding catalytic behavior of the reference NiMo(d)

and NiMo(c) samples prepared without using citric acid, we ob-

served that these catalysts showed MCHT/DMBP ratios similar to

those obtained with NiMoCA catalysts (Table 6). However, in the

reaction products obtained with them, an increase in the amount

of pre-hydrogenated intermediates (4H-4,6-DMDBT and 6H-4,6-

DMDBT) can be clearly seen. It is evidenced that these catalysts

have high hydrogenation ability and easily transform 4,6-DMDBT

into tetrahydro- and hexahydro-products, of which further conver-

sion into methylcyclohexyltoluene is retarded due to a low number

of active sites able to eliminate sulfur through hydrogenolysis of C–

S bond. Lower activity of the NiMo catalysts in comparison with

the catalysts of the NiMoCA series and their lower ability for C–S

bond cleavage seems to be in agreement with HRTEM results for

these catalysts (Table 2), where it was found that the active

MoS₂particles in NiMo(d) and NiMo(c) in general are larger

(ꢂ4.0–4.5 nm) and more stacked (ꢂ4 layers) than in other sulﬁded

samples.

Table 5 shows the results obtained in hydrodesulfurization of

4,6-DMDBT over prepared NiMo(CA)/SBA-15 catalysts. It is well

known that 4,6-DMDBT, one of the most refractory sulfur com-

pounds in diesel fuels [57,58], reacts preferentially through the

HYD route which, in general, is attributed to the steric hindrance

because of the presence of two alkyl substituents in positions 4

and 6 of this molecule [56]. Because of this, we expected that cat-

alysts that showed high HYD ability would result appropriate for

the elimination of sulfur from this refractory aromatic compound.

In addition, two NiMo/SBA-15 catalysts prepared without citric

acid were used for comparison purposes. It can be seen in Table

5 that there is a signiﬁcant difference in the catalytic activity of

the tested samples in HDS of 4,6-DMDBT. As expected, NiMoC-

A(1-c) and NiMoCA(9-c) catalysts showed good performance in

the hydrodesulfurization of 4,6-DMDBT (95 and 90% of conversion

at 8 h, respectively), followed by the NiMo(1-d) catalyst (88% of

conversion, Table 5). For the same reaction time, conversions ob-

tained with reference NiMo(d) and NiMo(c) catalysts were signiﬁ-

cantly lower (37% and 46%). NiMoCA(9-d) catalyst gave the lowest

conversion of 4,6-DMDBT among all catalysts tested (30% at 8 h).

This catalyst resulted to be four times less active than the corre-

sponding calcined NiMoCA(9-c) sample. The above results can also

be conﬁrmed by the values of initial reaction rate (r_o) and TOF

listed in Table 5. It should be mentioned that all kinetic parameters

shown in Table 5 for the reaction of 4,6-DMDBT are smaller than

the corresponding ones reported in Table 3 for HDS of DBT, which

can be attributed to lower reactivity of 4,6-DMDBT in comparison

with unsubstituted DBT. However, it can be seen that the decrease

in the activity with the change of the model sulfur-containing

compound was less pronounced for NiMoCA(1-c) and NiMoCA

(9-c) catalysts than for the NiMoCA(9-d) sample. It seems that this

is related to the HYD ability of the tested catalysts.

4. Discussion

It is well known that behavior of Ni(Co)Mo(W) HDS catalysts is

in relation to the characteristics of the active MoS₂or WS₂phase,

namely, with its dispersion and stacking degree. Different ap-

proaches have been used previously in order to tune the above

characteristics and to modify by this way activity and selectivity

of HDS catalysts. In our introduction, some of them were men-

tioned, for instance, change of support, modiﬁcation of precursor

metal species, or the use of chelating agents. In general, all these

options lead to a modiﬁcation in the characteristics of the sup-

ported Ni and Mo oxide species and, as a consequence, in the mor-

phology of the obtained sulﬁded catalysts and their performance.

In the present work, NiMo catalysts supported on high-surface-

area mesoporous silica of SBA-15 type were prepared using citric

acid as an additive and adjusting the pH of the impregnation solu-

tions to acidic or basic values. Our principal interest was to inves-

tigate what is the effect that the addition of citric acid has on these

catalysts. We tried to follow the relationship between the charac-

teristics of the Ni and Mo metal species present in solutions used

for the catalysts’ preparation, their characteristics in supported

Product distributions obtained for different catalysts at 30% of

4,6-DMDBT conversion are shown in Table 6. The preferential

pathway of HDS, as expected, was found to be the HYD route.

Methylcyclohexyltoluene (MCHT) was the main desulfurized

product obtained with all catalysts tested. Other products of

the HYD route: pre-hydrogenated S-containing intermediates

(tetrahydro-4,6-dimethyldibenzothiophene, 4H-4,6-DMDBT, and

hexahydro-4,6-dimethyldibenzothiophene, 6H-4,6-DMDBT) and

dimethylbicyclohexyl (DMBCH) were also obtained in smaller

proportions. The amount of dimethylbiphenyl (DMBP), the desul-

furized product of the DDS route, was between 7% and 11.3% for

Table 5

Conversions of 4,6-DMDBT obtained over NiMo(CA)/SBA-15 catalysts at different reaction times, initial reaction rate (r_o), and turnover frequency (TOF).^a

À

Á

TOF ꢄ 10³(s^ꢀ1

)

r_oꢄ 10⁷mol g s^ꢀ1

ꢀ1

Catalyst

4,6-DMDBT conversion (%)

cat

2 h

4 h

8 h

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

10

12

23

30

7

19

24

47

59

14

55

37

46

88

95

30

90

0.4

0.5

1.1

1.3

0.3

1.2

0.2

0.3

0.4

0.1

0.4

27

a

TOF was calculated as a number of 4,6-DMDBT molecules reacted per second per one Mo atom located on the edge surface. The number of Mo atoms located on the edge

surface was estimated using the total amount of Mo in the catalysts and the fraction of Mo atoms on the edge surface of MoS₂particles (f_Mo) determined for each catalyst by

HRTEM.

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

43

Table 6

Composition of products obtained in 4,6-DMDBT hydrodesulfurization over NiMo(CA)/SBA-15 catalysts (at 30% of 4,6-DMDBT conversion).

a

Catalyst

t₃₀(h)

Product^b(%)

4H

Product ratios

4H 6H/MCHT

6H

DMBP

MCHT

DMBCH

MCHT/DMBP

NiMo(d)

NiMo(c)

NiMoCA(1-d)

NiMoCA(1-c)

NiMoCA(9-d)

NiMoCA(9-c)

6.3

5.0

2.6

2.0

8.0

3.5

26.1

27.1

21.7

19.2

20.5

24.3

17.5

18.2

12.0

10.3

9.7

8.8

7.0

8.5

8.2

11.3

7.6

38.7

37.9

46.8

49.0

48.4

44.8

8.9

9.8

11.0

13.3

9.5

1.13

1.20

0.72

0.60

0.64

0.76

4.40

5.41

5.51

5.98

4.28

5.90

13.6

a

Time required to reach 30% of 4,6-DMDBT conversion.

4H = tetrahydro-4,6-dimethyldibenzothiophene (4H-4,6-DMDBT); 6H = hexahydro-4,6-dimethyldibenzothiophene (6H-4,6-DMDBT); DMBP = dimethylbiphenyl;

b

MCHT = methylcyclohexyltoluene; DMBCH = dimethylbicyclohexyl.

state after drying and calcination of the catalysts in an air atmo-

sphere and after sulﬁdation and the catalytic performance of the

prepared formulations, which was tested in simultaneous HDS of

two model compounds, DBT and 4,6-DMDBT.

the formation of an Anderson-type heteropolymolybdate (NH₄)₄

[Ni(OH)₆Mo₆O₁₈] 4H₂O in dried, and NiMoO₄in calcined NiMo cat-

alysts supported on SBA-15 (Fig. 8). No crystalline phases were de-

tected in the catalysts of the NiMoCA series independently of the

pH of the impregnation solution used. DRS and TPR characteriza-

tions (Figs. 9 and 10) gave more detailed information regarding

Ni(II) and Mo(VI) species present in the NiMoCA samples. It was

found that the characteristics of supported Mo(VI) species in dried

and calcined NiMoCA catalysts were similar for the samples pre-

pared from acidic and basic impregnation solutions containing cit-

ric acid. Mostly, well-dispersed octahedral Mo(VI) species, which

are easy to be reduced, were present in both NiMoCA(1) and NiM-

oCA(9) catalysts. We did not expect this result, since metal species

obtained in the solutions used for catalyst preparation were quite

different at basic and acidic pH values: at pH = 9, dispersed MoO₄^2ꢀ

and Nicit⁴₂^ꢀcomplex were present in the impregnation solution,

whereas at pH = 1, polymeric octahedral Mo(VI) species and sol-

vated Ni²⁺were observed. A comparison of the UV–Vis spectra of

acidic and basic solutions with those of the supported catalysts

(Figs. 3 and 9) shows that for both pH values, the type of Mo(VI)

species changed after their impregnation on the SBA-15 surface

and drying. Thus, after impregnation from the basic impregnation

solution, highly dispersed tetrahedral molybdate species trans-

formed into slightly agglomerated octahedral Mo ones. On the con-

trary, the dispersion of polymeric octahedral Mo species present in

acidic impregnation solution increased after their incorporation on

the SBA-15 surface. To explain this result, it is necessary to take

into account that SBA-15 silica is a support with low metal–sup-

port interaction, and its isoelectric point (IEP) is equal to 2.1 [59].

So, when the impregnation solution is put in contact with the

SBA-15 support, its pH is not the same as that of the starting

impregnation solution due to the small amount of solution used

for catalyst preparation. This change in the solution pH during

impregnation could be the ﬁrst reason for the change in the char-

acteristics of the impregnated metal species. In addition, the SBA-

15 surface charge is different when the support is in contact with

impregnation solutions of different pH values (1 and 9), and there-

fore, its interaction with soluble cationic and anionic species

should also be different. At pH = 1, SBA-15 surface is positively

charged, which facilitates the interaction with negatively charged

agglomerated octahedral Mo(VI) species from the impregnation

solution and promotes an increase in their dispersion in supported

form. On the other hand, at basic pH (pH = 9), SBA-15 surface has a

strong negative charge impeding the interaction with molybdate

anions from the solution. Probably, this lack of metal–support

interaction provokes some agglomeration of Mo(VI) species during

drying and evaporation of the solvent.

Initially, speciation diagrams for Ni(II) and Mo(VI) were ob-

tained which describe metal species present in aqueous solutions

at different pH values, in absence and presence of citric acid (Figs.

1 and 2). Then, prepared solutions were characterized by UV–Vis

absorption spectroscopy (Fig. 3) to conﬁrm predictions from speci-

ation diagrams. It was found that characteristics of both Ni(II) and

Mo(VI) species in impregnation solutions at pH 1 and 9 were differ-

ent. Regarding Ni(II), solvated Ni²⁺was present in aqueous solution

at acidic pH and Nicit⁴₂^ꢀcomplex was formed in basic solution. The

structure of the Nicit⁴₂^ꢀcomplex, obtained by quantum chemical

methods (Fig. 4), conﬁrmed that Ni(II) atom is octahedrally coordi-

nated with six O atoms of two citrate anions, which was in line

with UV–Vis spectral data. In addition, it can be mentioned that

formation of the Nicit⁴₂^ꢀcomplex allowed to maintain Ni(II) species

soluble in basic aqueous solutions and avoid precipitation of

Ni(OH)₂as it happens in absence of citric acid. For Mo(VI) species,

it was conﬁrmed that the type of species present in aqueous solu-

tions strongly depends on pH: at acidic pH, octahedrally coordi-

nated polymolybdates, probably hydrated MoO₃species, were

present, whereas highly dispersed tetrahedral molybdate species

(MoO²₄^ꢀ) existed in solutions at pH = 9. UV–Vis results, in agree-

ment with predictions made on the basis of the predominance area

diagrams, showed that for both pH studied, no Mo(VI)–CA com-

plexes were formed. Resuming the above results, it was shown that

different kinds of Ni(II) and Mo(VI) species were present in solu-

tions used for the preparation of the catalysts of the NiMoCA series

supported on SBA-15.

Characterization of prepared NiMoCA catalysts by TGA (Fig. 5)

showed that decomposition of Ni and Mo precursors and elimina-

tion of CA by decomposition and combustion in air takes place at

the temperature interval 150–500 °C. Almost all CA added to the

catalysts (around 17 wt.%) was eliminated by combustion in air

atmosphere at temperatures up to 500 °C. In addition, no notice-

able differences were observed in TGA proﬁles of the NiMoCA cat-

alysts prepared using basic or acidic impregnation solutions. We

did not observe by TGA any effect of the formation of the Nicit₂^4ꢀ

complex on the temperature of elimination of citric acid or on its

amount. However, chemical analysis of the NiMoCA catalysts cal-

cined at 500 °C for 4 h in an air atmosphere showed that elimina-

tion of CA was not complete after the above thermal treatment. A

small amount of carbonaceous material (0.59–0.77 wt.%) was still

remaining in the catalysts after calcination. This amount was high-

er for the NiMoCA(9-c) catalyst prepared from basic impregnation

solution, in which the Nicit⁴₂^ꢀcomplex was formed.

In addition, use of citric acid resulted in a more homogeneous

distribution of well-dispersed metal species in the SBA-15-sup-

ported catalysts in comparison with the reference NiMo/SBA-15

prepared without CA (XRD, DRS, TPR). An explanation for this can-

not be related to the interaction of Mo(VI) species with citric acid,

Further characterization of prepared dried and calcined NiMoCA

catalysts by XRD showed that the addition of citric acid to the

impregnation solutions containing Ni(II) and Mo(VI) precursors

avoided interaction between these two metal species leading to

44

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

as it was observed previously in works [27–29], in which bimetallic

Co₂[Mo₄(C₆H₅O₇)₂O₁₁] complex was formed. In our experimental

conditions, no Mo(VI)–CA complexes were detected. Previously, it

has been reported that homogeneous distributions and high dis-

persions of the supported metal oxide species can be obtained,

when aqueous solutions containing chelating agents (citrate, ace-

tate, EDTA, etc.) are used in the synthesis [60]. However, this was

not attributed to a formation of complexes, but to a steep increase

in viscosity upon solvent evaporation, which inhibits redistribution

of impregnated solution upon drying of the impregnated catalysts.

In addition, a gel-like phase can be formed that favors high disper-

sion of the active phase even after full drying and calcination. Also,

it was observed that because of the relatively weak interaction

with the support, many of the supported metal oxides prepared

with chelating agents can be reduced into the metal phases at

low temperatures in a hydrogen ﬂow, without signiﬁcant loss of

dispersion. It seems that the above explanation can be completely

applied for the results obtained in the present work for the NiMoC-

A catalysts supported on SBA-15. It can be concluded from our re-

sults that drying is an important step in the preparation of these

catalysts, and that not only the solution pH, but also the presence

of citric acid in the impregnation solutions affect the characteris-

tics of the metal species in dried and calcined catalysts.

Morphology of the sulﬁded NiMo(CA)/SBA-15 catalysts was

investigated by HRTEM. Obtained results (Table 2 and Figs. 11

and 12) were well in line with the characterization of dried and

calcined catalysts described above. Thus, better and more homoge-

neous dispersions of MoS₂particles were observed for the catalysts

prepared with the addition of citric acid than for the reference

NiMo/SBA-15 samples prepared without CA. Sulﬁded NiMoCA cat-

alysts had higher fraction of Mo atoms located of the catalytically

active surface (f_Mo) than corresponding NiMo/SBA-15 counterparts

(Table 2). In addition, chemical analysis of the freshly sulﬁded sam-

ples showed that sulfur content and, therefore, degree of sulﬁda-

tion, were much higher in the NiMoCA catalysts than in the

reference NiMo/SBA-15 ones. The above results are in good agree-

ment with previous publications [61–63]. Thus, in work [61], the

addition of citric acid in the impregnation solution resulted in an

increased Mo dispersion and enhanced activity in the hydrotreat-

ment of a light gas–oil fraction. In addition, XPS characterization

showed that the sulﬁdation degree of Mo was improved by the

addition of citric acid [62,63]. According to our results, the above

increase in the dispersion of the deposited Ni and Mo oxide species,

as well as in the sulﬁdation degree of the catalysts when prepared

with the addition of citric acid to the impregnation solution, was

observed for all the catalysts prepared with CA independently of

the pH of the impregnation solution and subsequent thermal treat-

ment (drying or calcination). So, it seems that these effects of citric

acid are not directly related to the formation of any speciﬁc Ni–CA

or Mo–CA chelating complexes. After calcination of the NiMoCA

catalysts prepared at distinct pH values, more stacked sulﬁded

Mo species were observed in comparison with sulﬁded catalysts

prepared from dried, but not calcined, NiMoCA precursors. Average

morphology of the calcined and then sulﬁded NiMoCA(1-c) and

NiMoCA(9-c) catalysts again was very similar (Table 2), which is

in agreement with characterization results of the corresponding

calcined precursors. Other important observation was the presence

of small amounts of carbonaceous residues in all sulﬁded catalysts

prepared with CA. Most, but not all, citric acid was eliminated dur-

ing calcination of the NiMoCA catalysts in an air atmosphere or

during sulﬁdation step. Amount of residual carbon was slightly

higher (ꢂ2 wt.%) in the NiMoCA(9-d) catalyst prepared from basic

impregnation solution in which Nicit⁴₂^ꢀcomplex was formed.

Catalytic activity results, in general, were in line with the re-

sults from the characterization of the sulﬁded NiMo(CA) catalysts

by HRTEM and by sulfur content determination. Thus, NiMoCA

catalysts prepared with citric acid showed higher catalytic activi-

ties in HDS of DBT (Table 3) than the reference NiMo/SBA-15 sam-

ples prepared without CA. This can be attributed to a better

dispersion of the MoS₂active phase and higher degree of sulﬁda-

tion in the catalysts prepared with citric acid. Calcination of the

NiMoCA(1) and (9) catalysts before sulﬁdation did not have a

noticeable impact on their overall catalytic activity in HDS of

DBT (Table 3), which is in accordance with similar f_Mofractions

for dried and calcined catalysts prepared from the same impregna-

tion solutions (Table 2). However, it was found that two NiMoCA

catalysts prepared using acidic impregnation solution (pH = 1)

showed signiﬁcantly higher DBT conversions and other kinetic

parameters, including TOF number, than those prepared at pH = 9

(Table 3). This result is difﬁcult to explain only on the basis of

the corresponding f_Mofractions, which are similar for the samples

prepared at basic and acidic conditions (Table 2). Other interesting

observation was made when we compared selectivities of the dif-

ferent catalysts in DBT HDS (Table 4). Thus, calcination of the NiM-

oCA(1) catalyst resulted in an increase in the hydrogenation ability

and in the proportion of the products of the HYD route, which, in

agreement with literature [17,54], can be attributed to an increase

in the stacking degree of the MoS₂phase (Table 2). Selectivity of

the NiMoCA(9) catalyst after calcination also changed in the same

direction, as well as the average stacking of the MoS₂phase in it.

However, this change was much stronger for the NiMoCA(9) cata-

lyst than for the NiMoCA(1). Thus, after calcination of the NiMoC-

A(9) catalyst, CHB/BP ratio changed in almost eight times: from

0.19 to 1.51, whereas for the NiMoCA(1) sample, the change was

much smaller: from 0.78 to 1.21 (Table 4). Above results show that

after calcination, both catalysts prepared at pH 1 and 9 have sim-

ilar selectivities (CHB/BP ratios ꢂ1.2–1.5), although product ratios

obtained with dried and then sulﬁded counterparts signiﬁcantly

differed, namely, the NiMoCA(9-d) catalyst was the only one sam-

ple showed a clear preference toward the direct desulfurization of

DBT. We think that this particular behavior of the catalyst prepared

with CA at basic pH can be attributed, but only partially, to the

morphological characteristics of the MoS₂active phase, in particu-

lar to its lowest stacking degree (1.8 layers in average) among all

prepared catalysts (Table 2). It seems that the difference in the

stacking degree between NiMoCA(9-d) and NiMoCA(1-d) samples

is not so strong (2.1 and 1.8 layers, respectively) for it to be the un-

ique reason, which could explain the difference in four times in the

CHB/BP ratios observed for these two catalysts (Table 4). Therefore,

we think that the explanation for such high hydrogenolysis ability

of the NiMoCA(9-d) catalyst should also be related to the formation

of a Nicit⁴₂^ꢀcomplex in the impregnation solution at basic pH and

the presence of this complex on the catalyst surface up to the sulf-

idation step. Upon calcination in air at 500 °C, this complex suffers

thermal decomposition, and most of the citric acid used in the cat-

alyst preparation is eliminated by combustion leaving only a small

amount of residual carbon on the support surface. However, when

NiMoCA(9-d) catalyst is subjected to sulﬁdation after drying (with-

out previous calcination), it can be supposed that the formation of

a Nicit⁴₂^ꢀcomplex produces a drastic increase in the temperature of

sulﬁdation of Ni leading to a larger amount of Ni–Mo–S phase and

an increase in the coverage of Ni on the edges of MoS₂particles. A

similar effect of a better promotion of the active MoS₂or WS₂

phases by Ni(Co) was reported previously for various HDS catalysts

prepared with chelating agents, such as EDTA, NTA and citric acid,

forming complexes with Ni(II) or Co(II) [22–25,33,64–66]. Product

distributions obtained for the HDS of DBT (Table 4) point out that

the catalytically active Ni–Mo–S phase obtained in the sulﬁded

NiMoCA(9-d) catalyst has a very large proportion of the active sites

responsible for the hydrogenolysis of C–S bond in DBT. However,

such active sites are not able to desulfurize a more refractory com-

pound as 4,6-DMDBT, which prefers the HYD route of the reaction.

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

45

This explains low activity of the NiMoCA(9-d) catalyst in HDS of

4,6-DMDBT.

obtained for the catalysts prepared from acidic impregnation solu-

tions and for the NiMoCA(9-c) sample, it was concluded that this

effect is due to the presence of citric acid in the impregnation solu-

tions used and not to the formation of any metal–CA complex. The

more likely explanation for this result seems to be an increase in

viscosity upon solvent evaporation, which inhibits redistribution

of impregnated solution upon drying of the impregnated catalysts.

Another similarity in all NiMo catalysts prepared with citric acid

was an increase in the degree of reduction of Ni and Mo oxide spe-

cies and in the degree of their sulﬁdation. This could also have par-

ticipated in the higher catalytic activity of these catalysts in

comparison with the reference NiMo/SBA-15 samples. Another fac-

tor that could have inﬂuenced catalytic behavior of the samples

prepared with citric acid was the presence of some residual carbon

after sulﬁdation of the catalysts.

The only catalyst NiMoCA(9-d), which was prepared from the

impregnation solution where a Nicit⁴₂^ꢀcomplex was formed,

showed a catalytic behavior signiﬁcantly different from the ob-

served with the other tested samples. This catalyst showed high

activity for DBT HDS with a clear preference for the direct desulfur-

ization route of the reaction. This high selectivity for the DDS route

limited the ability of this catalyst for the elimination of 4,6-

DMDBT, the molecule which is transformed preferentially through

the hydrogenation pathway of HDS. The particular catalytic behav-

ior of this sample was attributed to the presence of the Nicit₂^4ꢀ

complex in the catalyst, which drastically increases the tempera-

ture of Ni sulﬁdation leading to an increase in the Ni coverage on

the edges of MoS₂active particles. The formed Ni–Mo–S active

sites were responsible for the elimination of sulfur from DBT

through the direct desulfurization route, and their formation was

attributed to the speciﬁc interaction of Ni(II)–CA which took place

in the case of this catalyst.

Finally, we observed a presence of a small amount of residual

carbon (ꢂ0.5–2 wt.%) in all catalysts prepared with citric acid.

Probably, the presence of such carbon also plays some role in the

catalytic behavior of the NiMoCA catalysts. Results obtained in

our work are in line with recent reports regarding NiW, NiMo,

and CoMo HDS catalysts supported on alumina modiﬁed with a

small amount of carbon [67,68]. In work [67], alumina modiﬁed

with 1.3 wt.% of carbon was prepared by impregnation of alumina

with an aqueous solution of citric acid, followed by drying and cal-

cination under N₂ﬂow. It was observed that Ni and W species sup-

ported on C-modiﬁed alumina were easier to be reduced than

those on alumina. The addition of citric acid decreased the average

WS₂slab length but increased stacking number and facilitated the

formation of Ni–W–S active phase resulting in a higher activity for

HDS of 4,6-DMDBT. In work [68], NiMo and CoMo catalysts were

supported on alumina coated with different amounts of carbon

(1.2, 2.3, and 3.8 wt.%). It was found that speciﬁc catalytic activity

of NiMo and CoMo grew proportionally to the carbon content on

the catalysts. However, maximal total catalytic activity in HDS

and HYD was reached for carbon content of 1–2 wt.%. It was also

observed that carbon covering Al₂O₃carrier leads to an increase

in the average stacking number in the MoS₂slabs. In other work

[69], Co(Ni)Mo/c-Al₂O₃catalysts were prepared from tetraalkyl-

ammonium thiomolybdates. These catalysts showed the superior

hydrotreating activity when compared with the conventional

NiMo and CoMo supported on

c-Al₂O₃. The improved HDS and

HDN activity was attributed to the presence of carbon (0.4–

0.6 wt.%) in sulﬁded catalysts and to the formation of active ‘‘car-

bo-sulﬁde’’ phases. In our work, we do not have any evidence of

the formation of such a molybdenum carbide phase. However,

presence of a small amount of residual carbon in all catalysts pre-

pared using citric acid can explain observed changes in MoS₂mor-

phology and better catalytic performance than of the reference

NiMo/SBA-15 catalysts prepared without citric acid.

Finally, it can be concluded that the use of citric acid during the

preparation of NiMo catalysts supported on SBA-15 silica allows to

prepare in one-step active HDS catalysts, and that the correct

selection of pH of the impregnation solution used and of the fol-

lowing thermal treatment is an easy way to tune activity and selec-

tivity of these catalysts in the desired direction.

5. Conclusion

Acknowledgments

In the present work, NiMo catalysts supported on SBA-15 were

prepared by co-impregnation of Ni and Mo species with addition of

citric acid (CA) in the impregnation solutions of two pH values

(1 and 9). The objective was to realize a comparative study of

NiMoCA/SBA-15 catalysts prepared under different conditions in

order to get a deeper insight into the effect of the thermal treat-

ment and pH of the impregnation solution used on the behavior

of NiMoCA/SBA-15 catalysts in deep hydrodesulfurization (HDS).

Reference NiMo/SBA-15 catalysts prepared without CA showed

low activity in HDS of both DBT and 4,6-DMDBT, which was attrib-

uted to the formation of mixed Ni- and Mo-containing crystalline

phases: an Anderson-type heteropolymolybdate (NH₄)₄[Ni(OH)₆

Mo₆O₁₈] 4H₂O after drying at 100 °C and NiMoO₄after calcination

at 500 °C. The addition of citric acid to the impregnation solutions

allowed to stabilize them in a wide range of pH and to avoid

precipitation of Ni(OH)₂at basic pH. In basic impregnation solu-

tion, the Nicit⁴₂^ꢀcomplex was formed. The structure of this com-

plex was optimized by all electron DFT calculations. However, no

Mo–CA complexes were formed in impregnation solutions used

for the catalyst preparation.

Financial support by CONACyT-Mexico (Grant 100945) is grate-

fully acknowledged. D.V. acknowledges CONACyT for his scholar-

ship (207815). The authors thank Departamento de

Supercómputo de la Dirección General de Cómputo y Tecnologías

de la Información y Comunicación (DGTIC) UNAM for CPU time.

The authors are grateful to M. Portilla, E. Reynoso, M. Aguilar, C.

Salcedo Luna, and I. Puente Lee for technical assistance with TGA,

small-angle and powder XRD, and HRTEM characterizations,

respectively.

References

[1] http://www.dieselnet.com/standards/fuels.html.

[2] A. Stanislaus, A. Maraﬁ, M.S. Rana, Catal. Today 153 (2010) 1.

[3] T.A. Pecoraro, R.R. Chianelli, J. Catal. 67 (1981) 430.

[4] H. Topsøe, B.S. Clausen, F.E. Massot, Hydrotreating Catalysis, Springer-Verlag,

1996.

[5] U.T. Turaga, C. Song, Catal. Today 86 (2003) 129.

[6] K. Soni, K.C. Mouli, A.K. Dalai, J. Adjaye, Catal. Lett. 136 (2010) 116.

[7] A. Wang, Y. Wang, T. Kabe, Y. Chen, A. Ishihara, W. Qiany, J. Catal. 199 (2001)

19.

[8] J.M. Herrera, J. Reyes, P. Roquero, T. Klimova, Micropor. Mesopor. Mater. 83

(2005) 283.

[9] V. Sundaramurthya, I. Eswaramoorthia, A.K. Dalaia, J. Adjaye, Micropor.

Mesopor. Mater. 111 (2008) 560.

Three NiMoCA/SBA-15 catalysts prepared with citric acid,

NiMoCA(1-d), NiMoCA(1-c) and NiMoCA(9-c) showed high

activities in HDS of DBT and 4,6-DMDBT, higher than those of the

reference NiMo/SBA-15 samples. This can be ascribed to a better

and more homogeneous dispersion of oxidic and sulﬁded Ni and

Mo in the catalysts prepared with CA. Since similar results were

[10] O.Y. Gutiérrez, G.A. Fuentes, C. Salcedo, T. Klimova, Catal. Today 116 (2006)

485.

[11] O.Y. Gutiérrez, D. Valencia, G.A. Fuentes, T. Klimova, J. Catal. 249 (2007) 140.

46

T.E. Klimova et al. / Journal of Catalysis 304 (2013) 29–46

[12] T.A. Zepeda, J.L.G. Fierro, B. Pawelec, R. Nava, T. Klimova, G.A. Fuentes, T.

Halachev, Chem. Mater. 17 (2005) 4062.

[13] R. Nava, R.A. Ortega, G. Alonso, C. Ornelas, B. Pawelec, J.L.G. Fierro, Catal. Today

127 (2007) 70.

[14] D. Valencia, T. Klimova, Catal. Today 166 (2011) 91.

[15] S. Garg, K. Soni, G.M. Kumaran, M. Kumar, J.K. Gupta, L.D. Sharma, G.M. Dhar,

Catal. Today 130 (2008) 302.

[39] D. Valencia, I. García-Cruz, T. Klimova, Stud. Surf. Sci. Catal. 175 (2010) 529.

[40] I. Puigdomenech, Medusa Software, Chemical Equilibrium Diagrams, 2010.

(http://www.kemi.kth.se/medusa).

[41] P. Perdew, Y. Wang, Phys. Rev. B 45 (1992) 13244.

[42] N. Louwen, C. Pye, E.v. Lenthe, ADF2008.01 COSMO-RS, SCM, Theoretical

Chemistry: Vrije Universiteit: Amsterdam, The Netherlands, 2008. <http://

www.scm.com>.

[16] L. Dimitrov, R. Palcheva, A. Sponjakina, K. Jiratova, J. Porous Mater. 18 (2011)

425.

[43] O.Yu. Zelenin, Russ. J. Coord. Chem. 33 (2007) 346.

[44] D.S. Kim, Y. Kurusu, I.E. Wachs, F.D. Hardcastle, K. Segawa, J. Catal. 120 (1989)

325.

[45] L. Wang, W.K. Hall, J. Catal. 77 (1982) 232.

[46] I. Bonneviot, O. Legendre, M. Kermarec, D. Olivier, M. Che, J. Colloid Interface

Sci. 134 (1990) 534.

[47] K.-Q. Sun, E. Marceau, M. Che, Phys. Chem. Chem. Phys. 8 (2006) 1731.

[48] M. Fournier, C. Louis, M. Che, P. Chaquin, D. Masure, J. Catal. 119 (1989) 400.

[49] A. Sampieri, S. Pronier, S. Brunet, X. Carrier, C. Louis, J. Blanchard, K. Fajerwerg,

M. Breysse, Micropor. Mesopor. Mater. 130 (2010) 130.

[50] R.S. Weber, J. Catal. 151 (1995) 470.

[17] O.Y. Gutiérrez, T. Klimova, J. Catal. 281 (2011) 50–62.

[18] L. Lizama, T. Klimova, Appl. Catal. B: Environ. 82 (2008) 139.

[19] R. Cattaneo, T. Shido, R. Prins, J. Catal. 185 (1999) 199.

[20] M. Che, O. Clause, Ch. Marcilly, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),

Preparation of Solid Catalysts, Wiley-VCH, 1999, pp. 315–340 (Chapter 4.1.1).

[21] T. Shimizu, K. Hiroshima, T. Honma, T. Mochizuki, M. Yamada, Catal. Today 45

(1998) 271.

[22] G. Kishan, L. Coulier, V.H.J. de Beer, J.A.R. van Veen, J.W. Niemantsverdriet, J.

Catal. 196 (2000) 180.

[23] K. Al-Dalama, A. Stanislaus, Energy Fuels 20 (2006) 1777.

[24] M.S. Rana, J. Ramírez, A. Gutiérrez-Alejandre, J. Ancheyta, L. Cedeño, S.K. Maity,

J. Catal. 246 (2007) 100.

[25] M.A. Lélias, E. Le Guludec, L. Mariey, J. van Gestel, A. Travert, L. Oliviero, F.

Maugé, Catal. Today 150 (2010) 179.

[26] M.A. Lélias, P.J. Kooyman, L. Mariey, L. Oliviero, A. Travert, J. van Gestel, J.A.R.

van Veen, F. Maugé, J. Catal. 267 (2009) 14.

[27] O.V. Klimov, A.V. Pashigreva, M.A. Fedotov, D.I. Kochubey, Y.A. Chesalov, G.A.

Bukhtiyarova, A.S. Noskov, J. Mol. Catal. A: Mol. 322 (2010) 80.

[28] A.V. Pashigreva, G.A. Bukhtiyarova, O.V. Klimov, Y.A. Chesalov, G.S. Litvak, A.S.

Noskov, Catal. Today 149 (2010) 19.

[51] R. López Cordero, A. López Agudo, Appl. Catal. A: Gen. 175 (2000) 23.

[52] R. López Cordero, F.J. Gil Llambias, A. López Agudo, Appl. Catal. 74 (1991) 125.

[53] R. Thomas, E.M. van Oers, V.H.J. de Beer, J.A. Moulijn, J. Catal. 84 (1983) 275.

[54] E.J.M. Hensen, P.J. Kooyman, Y. van der Meer, A.M. van der Kraan, V.H.J. de

Beer, J.A.R. van Veen, R.A. van Santen, J. Catal. 199 (2001) 224.

[55] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345.

[56] C. Song, X. Ma, Appl. Catal. B: Environ. 41 (2003) 207.

[57] K.G. Knudsen, B.H. Cooper, H. Topsoe, Appl. Catal. A: Gen. 189 (1999) 205.

[58] B.C. Gates, H. Topsoe, Polyhedron 16 (1997) 3213.

[59] L.Y. Lizama, T.E. Klimova, J. Mater. Sci. 44 (2009) 6617.

[60] A. Jos van Dillen, R.J.A.M. Terörde, D.J. Lensveld, J.W. Geus, K.P. de Jong, J. Catal.

216 (2003) 257.

[29] G.A. Bukhtiyarova, O.V. Klimov, D.I. Kochubey, A.S. Noskov, A.V. Pashigreva,

Nucl. Instrum. Methods Phys. Res. Sect. A 603 (2009) 119.

[30] T. Fujikawa, Top. Catal. 52 (2009) 872.

[61] J.A. Bergwerff, M. Jansen, B.R.G. Leliveld, T. Visser, K.P. de Jong, B.M.

Weckhuysen, J. Catal. 243 (2006) 292.

[31] T. Fujikawa, M. Kato, T. Ebihara, K. Hagiwara, T. Kubota, Y. Okamoto, J. Jpn.

Petrol. Inst. 48 (2005) 114.

[62] T. Fujikawa, H. Kimura, K. Kiriyama, K. Hagiwara, Catal. Today 111 (2006)

188.

[32] T. Fujikawa, M. Kato, T. Ebihara, K. Hagiwara, T. Kubota, Y. Okamoto, J. Jpn.

Petrol. Inst. 48 (2005) 106.

[63] N. Rinaldi, Usman, K. Al-Dalama, T. Kubota, Y. Okamoto, Appl. Catal. A: Gen.

360 (2009) 130.

[33] N. Rinaldi, T. Kubota, Y. Okamoto, Ind. Eng. Chem. Res. 48 (2009) 10414.

[34] J. Escobar, M.C. Barrera, J.A. de los Reyes, J.A. Toledo, V. Santes, J.A. Colín, J. Mol.

Catal. A: Mol. 287 (2008) 33.

[64] R. Cattaneo, T. Weber, T. Shido, R. Prins, J. Catal. 191 (2000) 225.

[65] L. Coulier, V.H.J. de Beer, J.A.R. van Veen, J.W. Niemantsverdriet, J. Catal. 197

(2001) 26.

[35] D. Valencia, T. Klimova, Catal. Commun. 21 (2012) 77.

[36] D. Valencia, T. Klimova, Appl. Catal. B: Environ. 129 (2013) 137.

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

Science 279 (1998) 548.

[38] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120 (1998)

6024.

[66] T. Kubota, N. Rinaldi, K. Okumura, T. Honma, S. Hirayama, Y. Okamoto, Appl.

Catal. A: Gen. 373 (2010) 214.

[67] H. Li, M. Li, Y. Chu, F. Liu, H. Nie, Appl. Catal. A: Gen. 403 (2011) 75.

[68] P.A. Nikulshin, N.N. Tomina, A.A. Pimerzin, A.V. Kucherov, V.M. Kogan, Catal.

Today 149 (2010) 82.

[69] V. Sundaramurthy, A.K. Dalai, J. Adjaye, J. Mol. Catal. A: Chem. 294 (2008) 20.

Article Doi

Behavior of NiMo/SBA-15 catalysts prepared with citric acid in simultaneous hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene

DOI: 10.1016/j.jcat.2013.03.027

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.jcat.2013.03.027

Full text of DOI:10.1016/j.jcat.2013.03.027

Products guided by the article

R&D Labs maybe for 103272-52-2

Relevant to this article

Hot Product