Effects of preparation conditions in hydrothermal synthesis of highly active unsupported NiMo sulfide catalysts for simultaneous hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene

Article abstract of DOI:10.1016/j.cattod.2009.05.001

Unsupported NiMo sulfide catalysts were prepared from ammonium tetrathiomolybdate (ATTM) and nickel nitrate by using a hydrothermal synthesis method involving water, organic solvent and hydrogen. The activity of these catalysts in the simultaneous hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) was much higher than that of the commercial NiMo/Al₂O₃ sulfide catalysts. Interestingly, the unsupported NiMo sulfide catalysts showed higher activity for hydrogenation (HYD) pathway than the direct desulfurization (DDS) pathway in the HDS of DBT. The same trends were observed for the HDS of 4,6-DMDBT. Morphology, surface area, pore volume and the HDS activity of unsupported NiMo sulfide catalyst depended on the catalyst preparation conditions. Higher temperature and higher H₂ pressure and addition of an organic solvent were found to increase the HDS activity of unsupported NiMo sulfide catalysts for both DBT and 4,6-DMDBT HDS. Higher preparation temperature increased HYD selectivity but decreased DDS selectivity. High-resolution TEM images revealed that unsupported NiMo sulfide prepared at 375 °C shows lower number of layers in the stacks of catalyst with more curvature and shorter length of slabs compared to that prepared at 300 °C. On the other hand, higher preparation pressure increased DDS selectivity but decreased HYD selectivity for HDS of 4,6-DMDBT. HRTEM images showed higher number of layers in the stack for the NiMo sulfide prepared under an initial H₂ pressure of 3.4 MPa compared to that under 2.1 MPa. The optimal Ni/(Mo + Ni) ratio for the NiMo sulfide catalyst was 0.5, higher than that for the conventional Al₂O₃-supported NiMo sulfide catalysts. This was attributed to the high dispersion of the active species and more active NiMoS generated. The present study also provides new insight for controlling the catalyst selectivity as well as activity by tailoring the hydrothermal preparation conditions.

Full text of DOI:10.1016/j.cattod.2009.05.001

Catalysis Today 149 (2010) 52–61
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Effects of preparation conditions in hydrothermal synthesis of highly active
unsupported NiMo sulfide catalysts for simultaneous hydrodesulfurization of
dibenzothiophene and 4,6-dimethyldibenzothiophene
a,b,1
a,
a
b
Boonyawan Yoosuk
, Chunshan Song *, Jae Hyung Kim , Chawalit Ngamcharussrivichai ,
b
Pattarapan Prasassarakich
a
Clean Fuels and Catalysis Program, EMS Energy Institute and Department of Energy and Mineral Engineering, Pennsylvania State University, University Park, PA 16802, USA
b
Department of Chemical Technology, Faculty of Science, Chulalongkorn University, Bangkok 10330, Thailand
A R T I C L E I N F O
A B S T R A C T
Article history:
Unsupported NiMo sulfide catalysts were prepared from ammonium tetrathiomolybdate (ATTM) and
nickel nitrate by using a hydrothermal synthesis method involving water, organic solvent and hydrogen.
The activity of these catalysts in the simultaneous hydrodesulfurization (HDS) of dibenzothiophene
(DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) was much higher than that of the commercial
Available online 11 June 2009
Keywords:
Hydrothermal synthesis
Unsupported catalyst
Hydrodesulfurization
2 3
NiMo/Al O sulfide catalysts. Interestingly, the unsupported NiMo sulfide catalysts showed higher
activity for hydrogenation (HYD) pathway than the direct desulfurization (DDS) pathway in the HDS of
DBT. The same trends were observed for the HDS of 4,6-DMDBT. Morphology, surface area, pore volume
and the HDS activity of unsupported NiMo sulfide catalyst depended on the catalyst preparation
MoS
2
NiMo sulfide
ATTM
2
conditions. Higher temperature and higher H pressure and addition of an organic solvent were found to
increase the HDS activity of unsupported NiMo sulfide catalysts for both DBT and 4,6-DMDBT HDS.
Higher preparation temperature increased HYD selectivity but decreased DDS selectivity. High-
resolution TEM images revealed that unsupported NiMo sulfide prepared at 375 8C shows lower number
of layers in the stacks of catalyst with more curvature and shorter length of slabs compared to that
prepared at 300 8C. On the other hand, higher preparation pressure increased DDS selectivity but
decreased HYD selectivity for HDS of 4,6-DMDBT. HRTEM images showed higher number of layers in the
Dibenzothiophene
4
,6-Dimethyldibenzothiophene
Diesel
Catalyst preparation
stack for the NiMo sulfide prepared under an initial H
.1 MPa. The optimal Ni/(Mo + Ni) ratio for the NiMo sulfide catalyst was 0.5, higher than that for the
conventional Al -supported NiMo sulfide catalysts. This was attributed to the high dispersion of the
2
pressure of 3.4 MPa compared to that under
2
2 3
O
active species and more active NiMoS generated. The present study also provides new insight for
controlling the catalyst selectivity as well as activity by tailoring the hydrothermal preparation
conditions.
ß 2009 Elsevier B.V. All rights reserved.
1
. Introduction
The challenge of fulfilling the world’s growing transportation
of new refining processes and the increasing use of new forms of
energy production, e.g. fuel cells, exemplify operational issues.
Together, these trends are driving the need for deep desulfuriza-
tion catalysts. Such a demand for more active hydrodesulfurization
(HDS) catalysts has triggered a significant increase in research
activity on HDS catalyst development [1–3]. These studies have
energy needs is no longer a simple issue of producing enough
liquid hydrocarbon fuels. This challenge is accentuated by a
complex interplay of environmental and operational issues.
Environmental issues include societal demands that liquid
hydrocarbon fuels are clean and least polluting. The emergence
2
shown that one way to improve the performance of MoS -based
catalysts could be to increase the active phase loading or use bulk
sulfides. The emergence of highly loaded sulfide catalysts
demonstrates that sulfide-based systems still have great potential
for improvement.
Conventional catalysts used in the hydrotreating process
include mainly Ni or Co promoted Mo- or W-based catalysts
*
Corresponding author. Tel.: +1 814 863 4466; fax: +1 814 865 3573.
E-mail address: csong@psu.edu (C. Song).
1
Current address: National Metal and Materials Technology Center, Pathumthani
1
0
2120, Thailand.
2 3
supported by g-Al O [4]. They are active in converting thiophene
920-5861/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2009.05.001

B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
53
and benzothiophenes, but not active enough to desulfurize
efficiently the most refractory sulfur-containing polyaromatic
compounds, e.g. dibenzothiophene (DBT) and its alkyl-substituted
derivatives which are the major portion of sulfur in the high-
boiling fraction of crude oil and middle distillates of refinery
streams such as light cycle oil. In 1997, Gates and Topsoe [5]
pointed out that 4-MDBT (4-methyldibenzothiophene) and 4,6-
DMDBT are the most appropriate compounds for investigation of
candidate catalysts and reaction mechanism.
It is well established that Mo-based catalysts used for HDS
reactions are promoted by the addition of cobalt or nickel in much
more than trace amounts. The synergetic effects of promoter on the
catalytic activity of the Mo sulfides have been reported by many
researchers in the literatures [6–8]. The effect of promoter in Mo
sulfide catalysts on HDS reactions has been attributed to the
amount of promoter atoms that can be accommodated on the
catalysts from different bottles of ATTM reagent. In this work, a
bottle of newly purchased ATTM was used in synthesis, and the
ATTM reagent was stored in a refrigerator in order to minimize
oxidative degradation.
2.2. Catalyst synthesis and characterization
Unsupported NiMo sulfide catalysts were synthesized by using
the hydrothermal method involving water, an organic (hydro-
carbon) solvent and hydrogen. The catalyst synthesis was carried
out in a 25 mL batch reactor. Typically, ATTM and Ni(NO
3
)
2
2
ꢀ6H O
were dissolved in 10 g of de-ionized water. The atomic ratio of Ni/
(Mo + Ni) was varied in the range of 0–0.56. Then, an organic
solvent decalin was added. The reactor was sealed and purged with
hydrogen before being pressurized to a desired initial pressure and
placed in a preheated fluidized sand bath for 2 h under vertical
agitation. Subsequently, the reactor was removed from the sand
bath and immediately quenched in water bath. The unsupported
catalysts synthesized were separated and immerged under an
organic solvent before use.
2
edges of MoS layers and also to the electronic transfer that the
promoter atom induces on Mo atoms located at these sites [9,10].
Generally, Mo sulfide catalysts are prepared by an impregnation
of aqueous solution of oxide precursors on Al
by drying, calcination and presulfidation with either H
2
O
3
support, followed
/H S or an
2
2
The commercial-supported catalyst NiMo/Al
2
O
3
(Cr424)
organic sulfur compound such as dimethyldisulfide (DMDS) [11]. An
alternative to prepare sulfide catalysts is to use precursor
compounds already sulfided-like alkyl ammonium thiometalates
obtained from the Criterion Catalyst Company was used to
compare the activity with the synthesized unsupported Mo-based
sulfide catalysts for the HDS of 4,6-DMDBT and DBT. The catalyst
was crushed to a particle size of <1 mm and presulfided at 350 8C
4 2 4
precursors (NR ) MeS (R: alkyl or H; Me: Mo or W). Some
researchers have proposed effective preparation of Mo sulfide
catalysts with precursors such as ammonium tetrathiomolybdate
2 2
for 4 h in a flow of 5 vol.% H S/H at flow rate of 150 ml/min. The
sulfided catalyst was subsequently stored in decalin to minimize
2
(ATTM) [7,8,12–15]. These Mo sulfide catalysts can be synthesized
oxidation. The specific surface area is 163 m /g.
directly via a thermal decomposition of ATTM and do not need
further presulfidation. Devers et al. [16] observed the relatively high
activity of Mo sulfide catalysts prepared via hydrothermal method
compared with the Mo sulfide catalysts prepared by thermal
decomposition. Recently, Song et al. [17,18] reported a new method
for preparing the highly dispersed unsupported Mo sulfide catalyst
from aqueous ATTM solution mixed with an organic solvent via
hydrothermal synthesis under hydrogen. The use of organic solvent
helps to improve the dispersion of the precursor molecules. While
unsupported Mo sulfide catalysts have been often used as model
catalysts for hydrotreating reactions, their high activities could
become more attractive for commercial applications.
Our previous work [19] showed that hydrothermal synthesis
using water and organic solvent under hydrogen could produce
highly active Mo-based sulfide nano-size particles. Promoter (Ni
and Co) affects both the number and the activity of active site of Mo
sulfides. Temperature programmed reduction showed the close
interaction between Ni or Co and Mo in the bimetallic sulfide
catalysts resulting in shift of TPR peaks to lower temperatures
2
The N adsorption and desorption isotherms were measured on
a Micromeritics ASAP 2020 instrument. Pore size distributions of
the samples were determined from the isotherms by the Barrett–
Joyner–Hallenda (BJH) method. Fresh catalyst samples were
vacuum-dried before the adsorption measurement. The X-ray
diffraction (XRD) patterns were obtained on a Scintag Powder
Diffractometer with Cu Ka emission, 30 mA 35 kV with a scanning
speed of 28/min. The diffractograms were analyzed using the
standard JCPDS files. The high-resolution transmission electron
microscopy (HRTEM) was performed on a JEOL JEM-2010F
transmission electron microscope. A small amount of sample
was ground with a mortar and pestle. The sample was then
suspended in ethanol and sonicated. A drop of the suspension was
put on a lacey carbon film supported by a Cu grid.
2.3. Hydrodesulfurization reaction
Catalytic activity and selectivity of metal sulfides were
evaluated with a simultaneous HDS of DBT and 4,6-DMDBT in a
25 mL horizontal tubing micro-reactor. The reactor was loaded
with 0.023 g synthesized catalyst and 4 g reactant mixture (0.4%
mole 4,6-DMDBT and 0.4% mole DBT in decalin). The sealed reactor
was purged with hydrogen then pressurized up to 2.8 MPa and
placed in a fluidized sand bath which was preheated up to 350 8C
for 30 min. The reactor was vertically agitated at a rate of
2
upon Ni or Co addition to MoS . In the present study, unsupported
NiMo sulfide catalysts were prepared using ATTM by hydrothermal
method. This paper focuses on the effects of preparation conditions
on the activity and selectivity of unsupported NiMo sulfide
catalysts for the simultaneous hydrodesulfurization of DBT and
4
,6-DMDBT.
200 strokes/min; the reaction was not diffusion controlled. After
2
. Experimental
the reaction, the reactor was removed from the sand bath and
immediately quenched in a cold-water bath. Finally, liquid
products and the catalysts were collected. The liquid products
were identified by GC/MS (Shimadzu GC12A/QP-500) and quanti-
tatively analyzed with a GC-FID (HP5890) with a XTI-5 column
(Restek). Both GC/MS and GC-FID were programmed from 55 to
240 8C at heating rate of 5 8C/min. The DBTs conversion is defined
as {100 ꢁ [initial DBTs ꢂ final DBTs]/initial DBTs}. Duplicate or
triplicate experiments were performed and the average of
duplicated tests is reported here. The errors for conversion values
were typically within plus/minus 1.5 wt.%. The HYD/DDS ratio is
defined as [selectivity to (THDBT + CHB + BCH)/selectivity to BP].
2.1. Materials
4 2 4
Ammonium tetrathiomolybdate ((NH ) MoS ), nickel nitrate
hexahydrate (Ni(NO
3
)
2
ꢀ6H
2
O), dibenzothiophene (DBT), 4,6-
dimethyldibenzothiophene (4,6-DMDBT) and decahydronaphtha-
lene (decalin, used as solvent) were purchased from Aldrich
Chemical Company and were used without further purification. It
was noted that the long-time storage of ATTM in closed vials that
was exposed to air may lead to degradation of the reagent resulting
in deviations in the observed activity of in-situ generated MoS
2

5
4
B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
Fig. 1. Reaction pathways of desulfurization of polyaromatic sulfur compounds via a ring hydrogenation network (denoted by the rate constant k
hydrogenolysis network (denoted by the rate constant k ) (adapted from [45]).
1
) and a direct C–S bond
2
3
. Results and discussion
Hydrodesulfurization of DBT-type sulfur compounds generally
sulfide catalysts prepared at various temperatures. The selectivity
to biphenyl (BP) and cyclohexylbenzene (CHB) which were the
major products of DBT HDS showed different trends. The CHB
selectivity (from HYD pathway) increased while the BP selectivity
(from DDS pathway) decreased when the catalyst preparation
temperature was increased. However, the selectivity of tetra-
hydrodibenzothiophene (THDBT) and bicyclohexane (BCH), the
minor products, did not change significantly.
proceeds through two pathways as shown in Fig. 1. The direct-
desulfurization (DDS) or hydrogenolysis path involves chemisorp-
tion of the sulfur atom in the reactant molecule on an exposed Mo
ion at a sulfur vacancy, which is followed by hydrogen transfer and
sulfur elimination to complete desulfurization [20]. The HDS of
DBT occurs predominantly via the DDS pathway. However,
hydrogenolysis does not saturate the benzene rings in 4,6-DMDBT.
Therefore, HDS catalysts promoting the hydrogen-efficient hydro-
genolysis route to HDS are generally considered to be more
desirable. The second path is called hydrogenation (HYD) involving
The product selectivity of 4,6-DMDBT HDS is different from that
in DBT HDS. Tetra-hydrodimethyldibenzothiophene (THDMDBT)
was the major product from hydrogenation along with methylcy-
0
clohexyltoluene (MCHT) from HYD path, together with 3,3 -
0
dimethylbiphenyl (3,3 -DMBP). Similar to the DBT HDS product
0
the adsorption of sulfur compounds on the MoS
2
stack through the
distribution, 3,3 -DMBP, a product from DDS pathway, decreased
p
-electrons on the aromatic rings. This is followed by hydrogena-
when catalyst preparation temperature was increased, while the
products from the HYD pathway (THDMDBT and MCHT) increased.
tion of one of the aromatic rings and then carbon–sulfur bond
cleavage to complete desulfurization. 4,6-DMDBT is generally
desulfurized through the HYD pathway because the direct
adsorption of 4,6-DMDBT on the active sites is hindered by the
two-methyl groups at the 4- and 6-positions. If 4,6-DMDBT is
hydrogenated to tetrahydro- or hexahydro-DMDBT, this steric
hindrance is reduced and the electron density on the S atom is
increased, rendering the hydrogenated 4,6-DMDBT desulfurized
more easily [21–23].
3.1. Effect of catalyst preparation conditions on HDS activity of NiMo
sulfide catalysts
3.1.1. Catalyst preparation temperature
Fig. 2 shows the effects of catalyst preparation temperature on
the activity of unsupported NiMo sulfide catalysts for the DBT
and 4,6-DMDBT HDS. When the preparation temperature
was increased from 300 to 375 8C, the conversion of DBT and
4
4
,6-DMDBT increased gradually from 50.3% to 63.9% and from
1.6% to 51.9%, respectively. The DBT conversion was approxi-
Fig. 2. Effect of preparation temperature on simultaneous HDS of DBT and 4,6-
DMDBT over unsupported NiMo sulfide catalysts (other preparation parameters:
pressure = 2.8 MPa, solvent amount = 1 g, Ni/(Mo + Ni) = 0.43).
mately 1.2 times the 4,6-DMDBT conversion. Table 1 presents the
product distribution in the DBT and 4,6-DMDBT HDS over NiMo

B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
55
Table 1
Effect of preparation temperatures and pressures in catalyst preparation on simultaneous HDS of 4,6-DMDBT and DBT over unsupported NiMo sulfide catalysts at 2.8 MPa and
50 8C for 30 min (other catalyst preparation parameters: solvent amount = 1 g and Ni/(Mo + Ni) = 0.43).
3
Preparation temperature (8C) 300 325 350
Preparation H init. pressure (MPa) 2.8 2.8 2.8
375
2.8
350
1.4
350
2.1
350
3.4
2
DBT conversion (wt.%)
50.3
53.8
58.5
63.9
39.8
54.2
63.4
Selectivity (%)
THDBT
BP
5.2
46.5
39.6
8.7
5.4
43.4
42.0
9.2
6.2
41.1
42.6
10.1
4.9
38.7
47.1
9.3
10.9
51.4
31.6
6.1
9.5
37.4
43.1
10.0
5.6
45.0
40.8
8.6
CHB
BCH
a
HYD/DDS
1.2
1.3
1.4
1.6
0.9
1.7
1.2
4
,6-DMDBT conversion (wt.%)
41.6
44.8
47.3
51.9
29.4
40.4
50.5
Selectivity (%)
THDMDBT
33.7
40.6
24.0
1.7
35.6
37.1
25.8
1.5
37.8
33.2
27.0
2.0
41.8
25.4
31.0
1.8
45.6
29.4
21.7
3.3
41.9
31.4
24.5
2.2
35.3
32.8
28.9
3.0
0
3
,3 -DMBP
MCHT
DMBCH
b
HYD/DDS
1.5
1.2
1.7
1.2
2.0
1.2
2.9
1.2
2.4
1.4
2.2
1.3
2.0
1.3
DBT/4,6-DMDBT conv ratio
a
HYD/DDS = [selectivity to (THDBT + CHB + BCH)/selectivity to BP].
b
0
HYD/DDS = [selectivity to (THDMDBT + MCHT + DMBCH)/selectivity to 3,3 -DMBP].
0
The selectivity to 3,3 -DMBCH, the minor product, did not change
significantly. Consequently, the HYD/DDS ratio increased from 1.2
to 1.6 for the DBT HDS and from 1.5 to 2.9 for the 4,6-DMDBT HDS
with increasing preparation temperature from 300 to 375 8C.
The specific surface areas of the unsupported NiMo sulfide
catalysts prepared at various temperatures are shown in Table 2. In
general, increasing the preparation temperature increased both the
surface area and pore volume of the resulting catalysts which are
catalyst depends on the coexistence of two different sites that
promote HDS through the DDS pathway and the HYD pathway. A
very useful ‘‘Rim-Edge’’ model for MoS
proposed in 1994 by Daage and Chianelli [24]. In this model, the
MoS catalyst is described as stacks of several discs; the top and
2
stacks (slabs) has been
2
bottom discs are associated with rim sites (HYD sites) that
catalyze predominantly hydrogenation of large molecules such
as DBT; the discs ‘‘sandwitched’’ between the top and bottom
discs are associated with edge sites and sulfur removal from
large molecules is catalyzed mainly by edge sites (DDS). How the
rim sites catalyze hydrogenation of DBT and DMDBT was not
clear from the Rim-Edge model. In another earlier study, Tanaka
2
largely mesoporous. The highest surface area (249 m /g) and pore
volume (0.39 cm /g) were observed when catalyst was prepared at
temperature of 375 8C while the catalyst with lower surface area
121 m /g) and pore volume (0.19 cm /g) was obtained when a
3
2
3
(
lower preparation temperature (300 8C) was used.
reported that by cutting perpendicularly to basal plane of MoS
2
Fig. 3 presents the HRTEM images of unsupported NiMo sulfide
catalysts prepared at 300 and 375 8C, respectively. The black
single crystal into several pieces the catalytic selectivity of the
reaction could be modified [25]. This single crystal cutting
increased the area of the edge planes while the area of the basal
planes stayed constant. This cutting increases hydrogenation
activity of catalyst. More recent studies by Topsoe, Besenbacher,
Nørskov and their coworkers using advanced in-situ spectro-
scopic techniques coupled with experimental and computational
correlations have revealed that there exist metallic-like brim
sites which are located adjacent to the edge of Mo sulfide slabs
and such brim sites are active for hydrogenation [26–28]. The
brim sites location is different from that in the previously
proposed rim sites or corner sites. It is important to note that
only the top layer of multi-stacks of CoMo and NiMo sulfides will
expose the brim sites which play a key role for hydrogenation
reactions [26]. These new findings allow us to better understand
the importance of the slab length and slab stacking layers of Mo
sulfide catalysts.
According to the above-mentioned advances in structure-
activity correlations, catalysts with a predominance of single
unstacked layers will have a greater selectivity to hydrogenated
products than those with a predominance of stacked layers. In the
same context, catalysts with lower number of layers in stacks and/
or with shorter slabs will have higher hydrogenation activity than
the catalysts with higher number of layers in stacks and or with
longer slabs.
thread-like fringes in Fig. 3 correspond to the MoS
fringes observed in the images had a spacing of about 0.65 nm that
was the characteristic of (0 0 2) basal planes of crystalline MoS
2
slabs. The
2
.
The HRTEM images clearly showed that the average number of
layers in the stacks of catalyst prepared at 300 8C was higher (6–10
layers, Fig. 3A) than that of the catalyst prepared at 375 8C (2–4
layers, Fig. 3B). Furthermore, the catalyst prepared at lower
temperature had long slabs. The curvature of slabs increased while
the slab length decreased upon the increasing in the preparation
temperature.
The above results can be better rationalized by considering
that the catalytic activity and selectivity of Mo-based sulfide
Table 2
Surface area and pore volume of fresh NiMo sulfide catalysts prepared at various
preparation temperatures and pressures (other catalyst preparation parameters:
solvent amount = 1 g and Ni/(Mo + Ni) = 0.43).
Preparation
Preparation
Surface area
Pore volume
2
3
temperature (8C)
pressure (MPa)
(m /g)
(cm /g)
300
325
350
375
350
350
350
2.8
2.8
2.8
2.8
1.4
2.1
3.4
121
158
201
249
83
0.19
0.21
0.28
0.39
0.17
0.22
0.37
From the present HDS and HRTEM results on the effect of
preparation temperature, it appears that there are relatively more
HYD sites than the DDS sites on the unsupported NiMo sulfide
161
231

5
6
B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
Fig. 3. High-resolution TEM images of unsupported NiMo sulfide catalysts prepared at difference temperatures: (A) 300 8C; (B) 375 8C.
catalysts prepared at higher temperature. The catalyst prepared at
lower temperature had higher number of layers and long slabs and
it showed lower hydrogenation selectivity. On the other hand, the
catalyst prepared at higher temperature had lower number of
layers and short slabs and it showed higher hydrogenation
selectivity. The HRTEM images are consistent with the observed
results on the ratio of HYD/DDS in the product selectivity.
Based on the above results and discussion, it is highly likely that
the NiMo sulfide catalysts prepared at higher temperatures under
the hydrothermal conditions in this work possess larger numbers
of the brim sites (adjacent to the edge on the top layer of Mo-based
sulfides) that have been recently discovered and shown to be
active for hydrogenation by Topsoe, Besenbacher, Nørskov and
their coworkers [26–28].
Product distributions for the DBT and 4,6-DMDBT HDS are
presented in Table 1. For the DBT HDS, generally, BP was a major
product from the DDS pathway. However, it is very interesting to
note that the major product from HYD pathway was CHB. The
THDBT and BCH amount was detected around 6–10%. The HYD/
DDS ratio was over 1.0 for all catalysts except for the NiMo sulfide
prepared at 1.4 MPa, which had the HYD/DDS ratio of 0.9. In the
pressure range of 1.4–2.1 MPa, the selectivity to CHB increased
from 31.6% to 43.1% while the selectivity to BP decreased from
51.4% to 37.4%. Increasing pressure further to 3.4 MPa, the reverse
trends were observed. The results showed that there are mixed
effects of preparation pressure on hydrogenation and direct
desulfurization pathway of DBT HDS. Initially, it favors hydro-
genation pathway, but above 2.1 MPa the direct removal of sulfur
is more favorable.
0
3.1.2. Catalyst preparation pressure
For the 4,6-DMDBT HDS, THDMDBT, 3,3 -DMBP and MCHT were
Fig. 4 shows the effect of catalyst preparation pressure on the
subsequent hydrodesulfurization. Similarly to the catalyst pre-
paration temperature effect, a gradual increase in the conversion of
the major products while DMBCH amount were detected around
2.0% to 3.3%. Unlike the preparation temperature effect, the 3,3 -
0
DMBP selectivity from DDS pathway did not show a considerable
change but the selectivity to THDMDBT from HYD pathway
decreased significantly from 45.6% to 35.3% when the catalyst
preparation pressure was increased from 1.4 to 3.4 MPa. The HYD/
DDS ratio of 4,6-DMDBT HDS also decreased from 2.4 to 2.0 when
the preparation pressure was increased. The results showed that
4
,6-DMDBT and DBT was observed when H
preparation was increased. In the pressure range of 1.4–3.4 MPa,
pressure in catalyst preparation showed a remarkably strong
2
pressure in catalyst
H
2
effect on the HDS activity of resulting catalyst. The DBT and 4,6-
DMDBT conversion increased from 39.8% to 63.4% and from 29.4%
to 50.5%, respectively.
2
unsupported NiMo sulfide catalysts prepared at higher H pressure
became less selective for HYD pathway, but more selective for DDS
pathway for HDS of 4,6-DMDBT.
2
These results indicate that the H pressure for catalyst
preparation significantly affects the HDS activity of the resulting
NiMo sulfide catalysts. In the decomposition of ATTM, hydrogen is
necessary for forming the catalytically active bimetallic sulfide.
More detail on formation of the bimetallic sulfide in the presence of
hydrogen is described in the Section 3.3.2. The higher H
increased the activity of NiMo sulfide catalysts for both DBT and
,6-DMDBT HDS. The low H pressure might not provide enough
hydrogen for the decomposition and the reaction of ATTM.
Possibly, the higher H pressure assisted the reaction and
2
pressure
4
2
2
conversion of ATTM and Ni precursor to NiMo sulfide with a
more porous and more active phase for DBT and 4,6-DMDBT HDS.
The specific surface area and pore volume of catalysts prepared
2
at difference H pressure are shown in Table 2. Similar to the
preparation temperature effect, higher preparation pressure leads
to higher surface area and larger pore volume. The unsupported
2
Fig. 4. Effect of preparation pressure of H on simultaneous HDS of DBT and 4,6-
DMDBT over unsupported NiMo sulfide catalysts (other preparation parameters:
NiMo sulfide catalyst prepared at an initial H
has a lower surface area of 83 m /g, that prepared at 2.1 MPa has a
2
pressure of 1.4 MPa
2
temperature = 350 8C, solvent amount = 1 g, Ni/(Mo + Ni) = 0.43).

B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
57
2
Fig. 5. High-resolution TEM images of unsupported NiMo sulfide catalysts prepared at difference H pressure: (A) 2.1 MPa; (B) 3.4 MPa.
2
medium surface area of 161 m /g, while the catalyst prepared at
3
the increase in the MCHT selectivity was more pronounced than
that in the 3,3 -DMBP selectivity.
2
0
.4 MPa has a much higher surface area (231 m /g). Thus, the NiMo
sulfide catalyst with higher specific surface area and larger pore
volume could be synthesized at higher preparation pressure.
Fig. 5 presents the HRTEM images of unsupported NiMo sulfide
Table 4 presents the specific surface area and pore volume of
unsupported NiMo sulfide catalysts prepared with and without
catalysts prepared at 350 8C under initial H
.4 MPa. Similar to Fig. 3, the fringes observed in the photographs
are characteristic of slabs of crystalline MoS . According to the
images, the relatively lower preparation pressure (2.1 MPa) leads
to catalyst with lower average number of layers (2–6 layers,
Fig. 5A) in the stacks, while the higher preparation pressure
2
pressure of 2.1 and
Table 3
3
Effect of organic solvent amount in catalyst preparation on simultaneous HDS of
2
4,6-DMDBT and DBT over unsupported NiMo sulfide catalysts at 2.8 MPa and 350 8C
for 30 min (other catalyst preparation parameters: temperature = 350 8C, initial H
pressure = 2.8 MPa and Ni/(Mo + Ni) = 0.43).
2
Solvent amount(g)
0
1
3
(
(
3.4 MPa) leads to catalyst with higher average number of layers
4–14 layers, Fig. 5B) in the stacks. However, unlike the case of
DBT conversion (wt.%)
52.4
58.5
73.5
temperature effects (Fig. 3) discussed above, the change in the slab
length was not apparent when preparation pressure was increased
in this range. As shown in Fig. 5, the catalysts prepared at 2.1 and
Selectivity (%)
THDBT
BP
7.7
46.7
38.4
7.2
6.2
41.1
42.6
10.1
3.4
40.7
46.2
9.7
CHB
3
.4 MPa appear to show similar slab length of basal planes.
The product selectivity and HYD/DDS ratio results as shown in
BCH
a
HYD/DDS
1.1
1.4
1.5
Table 1 are consistent with HRTEM results. In the pressure range of
.1–3.4 MPa, the higher H pressure can induce an increase of the
2
2
4,6-DMDBT conversion (wt.%)
35.8
47.3
59.3
number of stacked layers and resulted in lower hydrogenation
selectivity, while lower average number of layers in the stacks can
be induced by lower preparation pressure and leads to high
hydrogenation selectivity. Therefore, it is likely that there exist
more brim sites that are active for hydrogenation on the NiMo
sulfide prepared at relatively lower pressure under the conditions
employed.
Selectivity (%)
THDMDBT
46.2
31.2
20.9
1.7
37.8
33.2
27.0
2.0
31.0
36.0
31.1
1.9
0
3,3 DMBP
MCHT
DMBCH
b
HYD/DDS
2.2
1.5
2.0
1.2
1.8
1.2
DBT/4,6-DMDBT conv ratio
a
HYD/DDS = [selectivity to (THDBT + CHB + BCH)/selectivity to BP].
b
0
3.2. Effects of organic solvent in catalyst preparation on HDS activity
HYD/DDS = [selectivity to (THDMDBT + MCHT + DMBCH)/selectivity to 3,3 -
DMBP].
of NiMo sulfide catalysts
Table 4
An organic solvent (decalin) was added to the aqueous solution
of ATTM and Ni precursor and the effect of solvent amounts was
investigated. Table 3 illustrates the effect of solvent amount used
on the activity of resulting NiMo sulfide catalysts for the HDS of
DBT and 4,6-DMDBT. Both the DBT and 4,6-DMDBT conversions
increased significantly when the solvent added during the catalyst
preparation was increased from 0 to 3 g. The NiMo sulfide catalyst
prepared with 3 g organic solvent showed much higher activity for
the HDS of DBT (73.5% conversion) and 4,6-DMDBT (59.3%
conversion). For the DBT HDS, BP was predominant over the
NiMo catalyst prepared without the solvent. With increasing the
amount of solvent, however, the BP selectivity slightly decreased.
Moreover, the selectivity of CHB became more significant than the
BP selectivity. The similar trends were observed in the 4,6-DMDBT
HDS. When the catalyst was prepared with higher solvent amount,
Surface area and pore volume of fresh catalysts prepared at various solvent amount
and Ni/(Mo + Ni) ratios (other catalyst preparation parameters: tempera-
2
ture = 350 8C and initial H pressure = 2.8 MPa).
Solvent
Ni/(Mo + Ni)
Surface area
Pore volume
2
3
amount (g)
(m /g)
(cm /g)
0
1
3
1
1
1
1
1
1
0.43
0.43
0.43
75
201
258
320
245
207
187
159
4
0.16
0.28
0.43
0.72
0.39
0.28
0.25
0.21
0.01
a
0
0.20
0.33
0.50
0.56
b
1.0
a
MoS
2
alone without a promoter.
b
NiS alone without Mo.

5
8
B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
adding solvent. The results of BET analysis show that the NiMo
sulfide catalyst prepared in the presence of solvent has much
2
2
higher surface area (258 m /g) than that without solvent (75 m /
g). The pore volume of catalyst also increased with increasing
solvent amount. Therefore, it is clear that organic solvent addition
in the preparation led to a highly active unsupported NiMo sulfide
catalyst with high-surface area. Compared with the catalyst
prepared without solvent addition, the catalysts prepared in the
presence of organic solvent have larger surface area for gas–liquid–
solid contact, and therefore more active hydrogen species (such as
H atom) could be produced for the HDS reaction. This may partially
rationalize much higher activity of the catalyst prepared with the
organic solvent addition than the one prepared in the absence of
the solvent.
The beneficial effect of organic solvent in this study is
consistent with the results reported by Yoneyama and Song
[
18]. They reported that the addition of solvent in the preparation
of MoS from ATTM provided the high selectivity for cleavage of C–
2
C bond and C–O bond and hydrogenation of naphthalene, and the
addition of water led to much higher catalytic activity. The
presence of the organic solvent helps to disperse ATTM-containing
water droplet during preparation with vigorous agitation. This
results a fine molecular dispersion of precursor molecules in the
aqueous solution isolated by organic solvent. Afanasiev [29] also
Fig. 6. Pore size distribution of unsupported Mo and NiMo sulfide catalysts. CoMo
sulfide is also shown for comparison. Preparation parameters: temperature
=
2
350 8C, initial H pressure = 2.8 MPa, solvent amount = 1 g (Me/(Mo + Me)
= 0.43, Me = Ni or Co for bimetallic sulfide samples).
2
reported that the presence of organic matter decreased MoS layer
stacking and textural stabilization. In the present preparation
procedure, the presence of organic solvent inhibits the crystal-
lization of the sulfide. The role of the organic (hydrocarbon) solvent
seems to be isolating and dispersing the droplets of aqueous
solution containing precursor molecules dissolved in water before
decomposition, facilitating the generation of catalytically active
sulfide particles, and preventing the particle growth by reducing
the stacking of layered sulfide slabs, as also suggested by XRD
patterns of the catalysts (see later section).
HDS data suggests that apart from the changes in number of active
sites in the NiMo sulfide catalysts, the effectiveness of the active
site may be enhanced.
The XRD patterns of the prepared NiMo sulfide catalyst series
are shown in Fig. 7. Compared with a commercial MoS
unsupported Mo-based sulfide catalysts exhibited broad diffrac-
tion peaks, indicating a very poorly crystallized MoS structure,
2
powder, all
2
particularly when the promoter was presented. The catalyst with
the Ni/(Mo + Ni) ratio of 0.33 showed the diffraction peaks of
2 2
poorly crystalline MoS , indicating that the MoS maintains its
3
.3. Effect of Ni/(Mo + Ni) ratio on HDS activity of NiMo sulfide
structure in the presence of amorphous Ni, as reported earlier [32].
For the catalysts with the Ni/(Mo + Ni) ratio above 0.43, the
diffraction peaks of the second metal sulfide appeared progres-
catalyst
Based on the above results on effects of preparation conditions,
3 4
sively. Ni S was detected in the catalysts with the Ni/(Mo + Ni)
a series of unsupported NiMo catalysts were prepared with
different Ni/(Mo + Ni) mole ratios at the same conditions of 350 8C
and 2.8 MPa H pressure with 1 g of organic solvent.
2
ratio of 0.43, 0.50 and 0.56. Moreover, NiS was detected in the
catalysts with the Ni/(Mo + Ni) ratio of 0.50 and 0.56. In most cases,
the ternary Mo–Ni–S phases did not appear clearly. It is probably
due to the fact that there is overlapping of diffraction peaks from
3.3.1. Specific surface area and structural properties
2
MoS and Mo–Ni–S phase. Another reason is that the active
The BET specific surface areas of the unsupported NiMo sulfide
structures (Mo–Ni–S phase) are possibly presented as very small
nano-crystallites, which cannot be characterized by X-ray diffrac-
tion method [33].
catalyst are summarized in Table 4. Pore size distributions of Mo
and NiMo sulfide catalysts are shown in Fig. 6. The surface area and
pore volume of all catalysts were measured before the HDS
reaction. The results showed that both Mo and NiMo sulfides are
largely mesoporous with a bimodal pore structure. For compar-
ison, the pore size distribution of CoMo sulfide prepared under the
same condition is also shown in Fig. 6. The unsupported Mo sulfide
catalyst prepared by hydrothermal method has a higher surface
area and larger volume, with bimodal pore size distribution (two
peak pore diameters around 3.8 and 6.2 nm). The surface area and
pore volume of catalysts were reduced after addition of the
promoter, but the bimodal pore size distribution is also apparent
with the bimetallic sulfide catalysts with the two peak pore
diameters (around 3.9 and 6.6 nm), as shown in Fig. 6. The decrease
in surface area in sulfide catalyst after promoter addition for
bimetallic sulfide catalysts was also reported by other researches
3.3.2. Catalytic activity and selectivity
Table 5 shows the effects of the Ni/(Mo + Ni) mole ratio on the
HDS specific activity of unsupported NiMo sulfide catalysts. The
Mo sulfide catalyst was prepared by the same procedure as NiMo
sulfide catalysts, but without the Ni precursor. For the Ni sulfide
catalyst, the same procedure was followed without the addition of
2
ATTM, and CS was used as sulfur source for sulfiding nickel [34].
A main purpose is to prepare catalysts with high activity in
converting the refractory sulfur compounds such as 4,6-DMDBT
that the conventional catalysts are not active enough to desulfurize
efficiently. The catalytic activity results in Table 5 show clearly that
the unsupported NiMo sulfide catalysts had a much higher HDS
activity than the commercial alumina-supported NiMo catalyst
(Cr424). Moreover, all unsupported NiMo sulfides and even the
[
1
7,30]. They found that the surface area was decreased from 50 to
2
5–25 m /g after addition of promoter to Mo sulfide catalyst [31].
unpromoted MoS
than the commercial NiMo/Al O
2
catalysts show higher 4,6-DMDBT conversion
catalyst.
The surface area of unsupported Mo sulfide catalysts was affected
by the Ni loading, which influenced the morphology and/or the
degree of aggregation of the Ni–Mo phase. This results along with
2 3
Surprisingly, 4,6-DMDBT is slightly more reactive than DBT
over the unsupported Mo sulfide prepared using ATTM without

B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
59
2 2
Fig. 7. XRD patterns of unsupported NiMo sulfide catalysts with various Ni/(Mo + Ni) ratios (MoS (C) represents Aldrich MoS reagent).
Table 5
Effect of Ni/(Mo + Ni) mole ratio on simultaneous HDS of 4,6-DMDBT and DBT over unsupported NiMo sulfide catalysts at 2.8 MPa and 350 8C for 30 min (other catalyst
preparation parameters: temperature = 350 8C, initial H pressure = 2.8 MPa, solvent amount = 1 g).
2
a
Catalysts
MoS
0
2
NiMoS
0.20
2
NiS
1
Cr424
Ni/(Mo + Ni)
0.33
51.3
0.43
58.5
0.50
67.8
0.56
57.1
DBT conversion (wt.%)
27.7
46.3
11.2
53.6
Selectivity (%)
THDBT
BP
44.5
29.2
19.7
6.6
10.1
46.6
36.4
6.9
7.0
44.9
39.9
8.2
6.2
41.1
42.6
10.1
4.2
39.5
45.6
10.7
6.4
44.8
40.5
8.3
26.5
56.7
7.9
4.0
64.3
28.6
3.1
CHB
BCH
8.9
b
HYD/DDS
2.4
1.1
1.2
1.4
1.5
1.2
0.8
7.5
0.6
4
,6-DMDBT conversion (wt.%)
31.8
34.8
40.0
47.3
54.0
44.4
26.0
Selectivity (%)
THDMDBT
87.1
7.8
45.0
26.6
26.8
1.6
45.4
26.3
26.6
1.7
37.8
33.2
27.0
2.0
33.9
33.4
30.6
2.1
42.1
31.2
25.2
1.4
84.6
10.7
4.3
40.6
32.6
26.6
0.2
0
3
,3 -DMBP
MCHT
3.8
DMBCH
1.3
0.4
c
HYD/DDS
11.8
0.9
2.8
1.3
2.8
1.3
2.0
1.2
2.0
1.3
2.2
1.3
8.4
1.5
2.1
2.1
DBT/4,6-DMDBT conv ratio
a
Commercial alumina-supported NiMo sulfide catalyst.
b
c
HYD/DDS = [selectivity to (THDBT + CHB + BCH)/selectivity to BP].
0
HYD/DDS = [selectivity to (THDMDBT + MCHT + DMBCH)/selectivity to 3,3 -DMBP].
adding Ni (31.8% 4,6-DMDBT conversion vs. 27.7% DBT conver-
sion). This was not expected initially but we have confirmed this
trend by three duplicated experiments. This is mainly due to a
HDS of DBT and 4,6-DMDBT over unsupported Mo and NiMo
sulfide catalysts mainly take the route of hydrogenation. The
reactivity of 4,6-DMDBT relative to that of DBT is much higher on
the unsupported Mo and NiMo sulfide catalysts compared to that
over conventional alumina-supported NiMo sulfide catalysts, as
can be seen from the ratio of DBT/4,6-DMDBT conversions shown
in Table 5.
The HYD/DDS ratio decreased significantly when Ni was added
into Mo sulfide catalyst. For example, for 4,6-DMDBT HDS, the
HYD/DDS ratio is 11.8 with the Mo sulfide catalyst and 2.0 for the
catalyst with the Ni/(Mo + Ni) ratio of 0.5. These results suggest
that the promoting effect was largely due to the enhancement of
DDS pathway, although HYD pathway is still a major pathway.
The remarkable increase in the catalytic activity by the addition
of Co or Ni to Mo sulfide catalyst has been reported by many
higher activity of this unsupported and unpromoted MoS
HYD pathway which is the prominent for both sulfur compounds
on this catalyst. As shown in Table 5, the HYD/DDS ratio of MoS for
both DBT and 4,6-DMDBT HDS was very high (2.4 for DBT HDS and
1.8 for 4,6-DMDBT HDS). The conversion of DBT and 4,6-DMDBT
2
for the
2
1
increased and reached maximum when Ni addition was increased
up to the Ni/(Mo + Ni) ratio of 0.5. However, the conversion
decreased when the ratio was further increased. These results
indicated that there is a significant synergetic effect of Ni on the Mo
catalysts for both DBT and 4,6-DMDBT HDS. The HYD/DDS ratio
was above 1 for all catalysts even for the HDS of DBT for which the
DDS pathway is generally predominant. It also indicated that the

6
0
B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
research groups. Upon increasing the concentration of the
promoter atoms, the HDS activity may increase significantly.
The other promoting effects may be highly dependent on the
catalysts preparation procedure. The variation in the optimum Me/
indeed catalytically active, they might help adsorb and dissociate
hydrogen molecule. The resulting H species could attack the MoS
particles and create coordinatively unsaturated site at the edges
[43,44]. However, it should note that excess Ni above the desired
ratio to Mo decreased the catalytic activity significantly. Under
2
(Me + Mo) (Me = Co, Ni) mole ratios reported in the literatures may
be explained. In the case of the unsupported catalyst, the optimum
Me/(Me + Mo) atomic ratios are in the range of 0.3–0.54 for CoMo
sulfide catalysts, [30,35,36] and from 0.4 to 0.55 for Ni–Mo sulfide
catalysts [37,38]. In the case of supported catalyst, the optimum
Me/(Me + Mo) atomic ratios is generally in the range from 0.2 to 0.4
for CoMo or NiMo sulfide catalysts [6,39]. In this study, the activity
of the catalyst with the Ni/(Mo + Ni) ratio of 0.2 and 0.33 was much
2
elevated H pressure, sulfur vacancy could be produced during the
NiMo sulfide catalyst preparation stage and maintained during the
hydrodesulfurization reaction. Therefore, the present results
clearly demonstrate that the proper combination of Ni and Mo,
not separate Ni sulfide, is responsible for higher catalytic activity.
4. Conclusions
2
higher than MoS and NiS sulfide catalysts. The NiMo sulfide
catalyst with the Ni/(Mo + Ni) ratio of 0.50 showed the highest
activity among all catalysts tested. This optimum Ni/(Mo + Ni)
ratio agrees very well with the previous results reported for the
unsupported NiMo sulfide catalysts.
The study of unsupported sulfide catalyst is a promising route
for developing a better understanding and contributes to the
development of better hydrotreating catalysts. Hydrothermal
preparation of transition-metal sulfides is particularly interesting,
because this method provides a highly active catalyst which does
not require the sulfidation step. The unsupported NiMo sulfide
catalysts synthesized from ATTM by the hydrothermal method
exhibited the excellent catalytic activity for the simultaneous HDS
of DBT and 4,6-DMDBT.
For the unsupported NiMo sulfide catalysts, the HYD pathway
was predominant as compared with the DDS pathway in the HDS
of DBT as well as the HDS of 4,6-DMDBT. The unsupported NiMo
sulfide catalysts exhibit strong synergy between Mo and Ni for the
HDS reaction. The maximum synergy in the HDS of DBT and 4,6-
DMDBT was observed over catalyst with the Ni/(Mo + Ni) mole
ratio close to 0.5.
During the catalyst preparation, the reaction between ATTM
and Ni(NO
3 2
) leading to the bimetallic sulfide precursor was
suggested as follows [7]:
2
ðNH
4
Þ
2
MoS
4
þ NiðNO
3
Þ
2
! ðNH
4
Þ
2
fNiðMoS
4
Þ
2
4 3
g þ 2NH NO
(
1)
In the presence of hydrogen gas, the bimetallic sulfide is formed
according to the reactions:
NiðMoS
4
3
Þ
Þ
2
2
þ H
þ H
2
2
! NiðMoS
3
2
Þ
Þ
2
2
þ H
þ H
2
2
S
S
(2)
(3)
NiðMoS
! NiðMoS
2
The preparation conditions (temperature, H pressure, and
The chemical interaction between Ni and Mo atoms desired for
effective NiMo catalysts implies that a good dispersion of the Ni
addition of an organic solvent to aqueous solution of Mo and Ni
precursors) have a significant effect on the HDS activity and
selectivity of unsupported NiMo sulfide catalysts. Both higher
2
promoter on MoS is necessary for the formation of a large number
of active NiMoS sites. Scanning tunneling microscope study by
Lauritsen et al. revealed that the presence of Ni promoter atoms
2
temperature and higher H pressure in preparation led to higher
surface area and larger pore volume and higher activity of the
resulting NiMo sulfide catalysts for simultaneous HDS of DBT and
4,6-DMDBT.
However, the selectivity of NiMo sulfide catalysts was affected
in different way. The NiMo catalyst prepared at higher preparation
temperature became more selective for the HYD pathway but less
selective for the DDS pathway for both DBT and 4,6-DMDBT HDS,
as reflected by the increasing HYD/DDS ratio from 1.2 to 1.6 for DBT
and from 1.5 to 2.9 for 4,6-DMDBT with increasing preparation
temperature from 300 to 375 8C.
causes the change of the shape of the MoS
triangular to hexagonally truncated structure [40]. Ni atoms may
be located at the edge of MoS crystallites to form Ni–Mo–S
structure which is considered to be most active sites for HDS
6,41,42]. The hydrothermal method developed in our laboratory
2
nanoclusters from
2
[
leads to the formation of very small size of NiMo sulfide cluster in
the unsupported catalyst as shown by TEM in our present and
previous work [19].
It is postulated that the organic solvent help to disperse and
isolate the droplets of aqueous solution containing ATTM and Ni
precursor under agitation which is heated rapidly to elevated
temperatures of 300–375 8C (close to sub- and super-critical fluid
Increasing the initial H
2
pressure in catalyst preparation
increased the activity of unsupported NiMo sulfide catalysts for
both DBT and 4,6-DMDBT conversions, but the there are mixed
effects on the selectivity. Initial increase in preparation pressure
from 1.4 to 2.1 MPa enhanced HYD selectivity (increased HYD/DDS
ratio); further increasing pressure to 2.8 and 3.4 MPa increased
DDS selectivity but decreased HYD selectivity (decreased HYD/DDS
ratio) for DBT HDS. For 4,6-DMDBT HDS, HYD/DDS ratio generally
decreased with increasing preparation pressure from 1.4 to
3.4 MPa.
HRTEM revealed that the NiMo sulfide catalyst prepared at
lower temperature (300 8C) had higher number of layers (6–10
layers) and long slabs and it showed lower hydrogenation activity.
On the other hand, the NiMo sulfide catalyst prepared at higher
temperature (375 8C) had lower number of layers (2–4 layers) and
short slabs and it showed higher hydrogenation selectivity. For the
unsupported NiMo sulfide catalysts prepared at 350 8C under
2
conditions) under H pressure, where ATTM reacts with the Ni
precursor and decomposes to bimetallic sulfide species. Subse-
quently, very fine particles (nano-size) of NiMo sulfide catalyst are
generated. Since Mo is surrounded by Ni in solution before and
2
during hydrothermal decomposition under H , more Ni atoms
would be incorporated into smaller Mo sulfide nano-crystallites to
form more particles with NiMoS phases. From Table 5, the
conversion of both sulfur compounds increased when the Ni/
(Mo + Ni) atomic ratio was increased up to 0.5. This likely because,
more active catalyst particles were formed with more Ni atoms,
incorporated into the small nano-sized crystallites of Mo sulfide.
However, for the catalyst with Ni/(Mo + Ni) atomic ratio above 0.5,
excess Ni atoms may partly block the more active phase of NiMoS
leading to decrease in the conversion of sulfur compounds.
Although the separate phases of the Ni
for the catalysts with Ni/(Mo + Ni) ratio between 0.43 and 0.5
Fig. 7), there was no apparent detrimental effect on the conversion
of DBT and 4,6-DMDBT. What role does this separate Ni sulfide play
is a matter of debate. If these metal (Ni and Co) sulfide particles are
3
S
4
and NiS were detected
2
initial H pressure of 2.1 and 3.4 MPa, HRTEM revealed that the
lower preparation pressure leads to catalyst with lower average
number of layers (2–6 layers) in the stacks, while the higher
preparation pressure leads to catalyst with higher average number
of layers (4–14 layers) in the stacks. It is possible that the NiMo
(

B. Yoosuk et al. / Catalysis Today 149 (2010) 52–61
61
[
9] R. Candia, B.S. Clausen, H. Topsoe, J. Catal. 77 (1982) 564.
sulfide catalysts prepared at higher temperatures and/or at
relatively lower pressures under the hydrothermal conditions in
this work possess larger numbers of the brim sites (adjacent to the
edge on the top layer of Mo-based sulfides) that have been recently
discovered and shown to be active for hydrogenation by Topsoe,
Besenbacher, Nørskov and their coworkers [26–28].
An organic solvent (decalin) addition in the catalyst preparation
resulted in the highly active catalysts with higher surface area and
larger pore volume. This suggests that the use of a hydrocarbon
solvent help to isolate and disperse the droplets of aqueous
[
[
10] S. Harris, R.R. Chianelli, J. Catal. 98 (1986) 17.
11] T. Kabe, A. Ishihara, W. Qian, Hydrodesulfurization and Hydrodenitrogenation:
Chemistry and Engineering, Wiley-VCH, Weinhein, 1999, p. 14.
12] H. Nava, F. Pedraza, G. Alonso, Catal. Lett. 99 (2005) 65.
13] H. Nava, C. Ornelas, A. Aguilar, G. Berhault, S. Fuentes, G. Alonso, Catal. Lett. 86
[
[
(2003) 257.
[14] H. Farag, K. Sakanishi, M. Kouzu, A. Matsumura, Y. Sugimoto, I. Saito, J. Mol. Catal.
A-Chem. 206 (2003) 399.
15] Y. Iwata, K. Sato, T. Yoneda, Y. Miki, Y. Sugimoto, A. Nishijima, H. Shimada, Catal.
Today 45 (1998) 353.
16] E. Devers, P. Afanasiev, B. Jouguest, M. Vrinat, Catal. Lett. 82 (2002) 13.
[17] C.S. Song, Y. Yoneyama, M.R. Kondam, US Patent No. 6,451,729, Method for
[
[
preparing a highly active, unsupported high surface area MoS
18] (a) Y. Yoneyama, C.S. Song, Catal. Today 50 (1999) 19;
2
catalyst (2002).
3 2
solution containing ATTM and Ni(NO ) under agitation. This
[
results a fine molecular dispersion of precursor molecules in
aqueous solution which leads to the formation of nano-sized,
highly active NiMoS catalyst particles. The hydrothermal method
with suitable conditions leads to formation of nano-sized Mo
sulfide clusters and more Ni atoms could be incorporated into
smaller Mo sulfide nano-crystallites to form more active NiMoS
catalysts.
(
b) Y. Yoneyama, C.S. Song, Energy Fuels 16 (2002) 767.
[19] B. Yoosuk, J.H. Kim, C.S. Song, C. Ngamcharussrivichai, P. Prasassarakich, Catal.
Today 130 (2008) 14.
20] R. Prins, in: G. Ertl, H. Knozinger, J. Weitkamp (Eds.), Handbook of Heterogeneous
[
Catalysis, 4, Wiley-VCH, Berlin, 1997, p. 1909.
[21] X. Ma, J.H. Kim, C.S. Song, Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem. 48 (2003)
35.
1
[
[
22] X. Ma, H.H. Schobert, Prepr. Pap. Am. Chem. Soc., Div. Fuel Chem. 42 (1997) 657.
23] R. Shafi, G. Hutchings, Catal. Today 59 (2000) 423.
The present study provides new insight into structure–activity–
selectivity relationship and allows for controlling the catalyst
selectivity as well as activity of unsupported NiMo sulfide catalysts
by tailoring the hydrothermal preparation conditions.
[24] M. Daage, R.R. Chianelli, J. Catal. 149 (1994) 414.
[
[
[
25] K.I. Tanaka, Adv. Catal. 33 (1985) 99.
26] H. Topsoe, Appl. Catal. A: Gen. 322 (2007) 3.
27] J.V. Lauritsen, M. Nyberg, J.K. Norskov, B.S. Clausen, H. Topsoe, E. Laegsgaard, F.
Besenbacher, J. Catal. 224 (2004) 94.
[28] A. Logadottir, P.G. Moses, B. Hinnemann, N.Y. Topsoe, K.G. Knudsen, H. Topsoe, J.K.
Norskov, Catal. Today 111 (2006) 44.
29] P. Afanasiev, G. Xia, G. Berhault, B. Jouguet, M. Lacroix, Chem. Mater. 11 (1999) 3216.
30] K. Inamura, R. Prins, J. Catal. 147 (1994) 515.
Acknowledgments
[
[
We gratefully acknowledge the support of US Department of
Energy, National Energy Technology Laboratory through the
Refinery Integration Program, US Environmental Protection
Agency through NSF/EPA TSE program and the Thailand Research
Fund through the Royal Golden Jubilee Program.
[31] O. Weisser, S. Labda, Surface Catalysts: Their Properties and Application, Perga-
mon, Oxford, 1973.
[
[
32] R. Candia, B.S. Clausen, H. Topsoe, Bull. Soc. Chim. Belg. 90 (1981) 1225.
33] L.S. Byskov, B. Hammer, J.K. Norskov, B.S. Clausen, H. Topsoe, Catal. Lett. 47 (1997)
177.
[
[
[
[
[
34] K.M. Reddy, B. Wei, C.S. Song, Catal. Today 43 (1998) 261.
35] R.C. Hoodless, R.B. Moyes, P.B. Wells, Bull. Soc. Chim. Belg. 93 (1984) 673.
36] M.L. Vrinat, L. Demourgues, Appl. Catal. A: Gen. 5 (1983) 43.
37] K.C. Pratt, J.V. Sanders, N. Tamp, J. Catal. 66 (1980) 82.
38] C. Gachet, R. Paulus, L. Demourgues, C. Durand, H. Toulhoat, Bull. Soc. Chim. Belg.
93 (1984) 681.
References
[
[
[
[
1] C.S. Song, Catal. Today 86 (2003) 211.
2] C.S. Song, X.L. Ma, Appl. Catal. B-Environ. 41 (2003) 207.
3] S. Brunet, D. Mey, G. Perot, C. Bouchy, F. Diehl, Appl. Catal. A: Gen. 278 (2005) 143.
4] J.G. Speight, B. Ozum, Petroleum Refining Processes, Marcel-Dekker, New York,
[39] T. Isoda, S. Nagao, X. Ma, Y. Korai, I. Mochida, Appl. Catal. A: Gen. 150 (1997) 11.
[40] J.V. Lauritsen, S. Helveg, E. Laegsgaard, I. Stensgaard, B.S. Clausen, H. Topsoe, E.
Besenbacher, J. Catal. 197 (2001) 1.
2
002, p. 455.
[41] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345.
[42] M. Breysse, G. Djega-Mariadassou, S. Pessayre, C. Geantet, M. Vrinat, G. Perot, M.
Lemaire, Catal. Today 84 (2003) 129.
[43] B. Delmon, Surf. Rev. Lett. 2 (1995) 25.
[44] P. Grange, X. Vanhaeren, Catal. Today 36 (1997) 375.
[45] J.H. Kim, X. Ma, C.S. Song, Y. Lee, S.T. Oyama, Energy Fuel 19 (2005) 353.
[
[
5] B.C. Gates, H. Topsoe, Polyhedron 16 (1997) 3213.
6] H. Topsoe, B.S. Clausen, F.E. Massoth, in: J.R. Anderson, M. Boudart (Eds.), Catalysis
Science and Technology, 11, Springer, Berlin, 1996, p. 162.
7] F. Pedraza, S. Fuentes, Catal. Lett. 65 (2000) 107.
8] L. Alvarez, J. Espino, C. Ornelas, J.L. Rico, M.T. Cortez, G. Berhault, G. Alonso, J. Mol.
Catal. A-Chem. 210 (2004) 105.
[
[