Simultaneous promotion of hydrogenation and direct desulfurization routes in hydrodesulfurization of 4,6-dimethyldibenzothiophene over NiW catalyst by use of SiO2-Al2O3 support in combination with trans-1,2-diaminocyclohex

Article abstract of DOI:10.1016/j.apcata.2010.05.026

Effect of novel SiO₂-Al₂O₃ support on surface structure and hydrodesulfurization (HDS) activity of the NiW catalyst prepared using trans-1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid (CyDTA) was investigated by combina

Full text of DOI:10.1016/j.apcata.2010.05.026

Applied Catalysis A: General 383 (2010) 79–88
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Simultaneous promotion of hydrogenation and direct desulfurization routes in
hydrodesulfurization of 4,6-dimethyldibenzothiophene over NiW catalyst by use
of SiO -Al O support in combination with
2
2
3
ꢀ
ꢀ
trans-1,2-diaminocyclohexane-N,N,N ,N -tetraacetic acid
a,∗
a
a
b
Naoto Koizumi , Yusuke Hamabe , Shohei Yoshida , Muneyoshi Yamada
a
Department of Applied Chemistry, Graduate School of Engineering, Tohoku University, Aoba 6-6-07, Aramaki, Aoba-ku, Sendai 980-8579, Japan
Akita National College of Technology, 1-1, Iijima-Bunkyo-cho, Akita 011-8511, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Effect of novel SiO -Al O support on surface structure and hydrodesulfurization (HDS) activity of the NiW
2
2
3
Received 9 December 2009
Received in revised form 18 May 2010
Accepted 19 May 2010
ꢀ
ꢀ
catalyst prepared using trans-1,2-diaminocyclohexane-N,N,N ,N -tetraacetic acid (CyDTA) was investi-
gated by combination of quasi in situ TEM, in situ diffuse reflectance FTIR spectroscopy (DRIFTS) coupled
with NO adsorption, and 4,6-dimethyldibenzothiophene (4,6-DMDBT) HDS activity measurement. When
the catalysts were prepared without using CyDTA, quasi in situ TEM observation showed that WS2-like
slabs in the SiO2-Al2O3 supported W and NiW catalysts had higher stacking degree, but smaller basal
plane sizes compared with the Al2O3 supported counterparts. In situ DRIFTS coupled with NO adsorption
further revealed that the formation of the Ni–W–S phase was suppressed by use of SiO2-Al2O3 support.
The SiO2-Al2O3 supported NiW catalyst showed higher activity for hydrogenation (HYD) of 4,6-DMDBT
into tetrahydro-DMDBT (THDMDBT) than the Al2O3 supported counterpart. However, this catalyst failed
to promote the HYD route because of lower activity for C–S bond cleavage of THDMDBT, and showed
lower HDS activity than the Al O supported NiW catalyst. On the other hand, the formation of the
Available online 26 May 2010
Keywords:
Hydrodesulfurization (HDS)
4
,6-Dimethyldibenzothiophene
(
4,6-DMDBT)
NiW catalyst
SiO2-Al2O3
trans-1,2-diaminocyclohexane-N,N,N ,N -
ꢀ
ꢀ
2
3
tetraacetic acid
Ni–W–S phase was promoted when the SiO2-Al2O3 supported NiW catalyst was prepared using CyDTA,
which was accompanied with further increase of stacking degree and decrease of the basal plane size of
the WS2-like slabs. Double-layered slabs were predominant in the SiO2-Al2O3 supported NiW catalyst
prepared with using CyDTA. This catalyst showed higher activity for both HYD of 4,6-DMDBT, and C–S
bond cleavage of 4,6-DMDBT and THDMDBT. The HYD route and direct desulfurization (DDS) route were
promoted simultaneously over this catalyst, leading to three times higher HDS activity than the Al2O3
supported NiW catalyst. These results suggest that the formation of the multi-layered Ni–W–S phase
with the smaller basal plane size is crucial for promoting the HYD route in HDS of 4,6-DMDBT.
(
CyDTA)
©
2010 Elsevier B.V. All rights reserved.
1
. Introduction
studies, it has been suggested that HDS of 4,6-DMDBT proceeds in
two parallel reaction routes, i.e. the direct desulfurization (DDS)
and hydrogenation (HYD) routes [2–9]. Because the HYD route is
much faster than the DDS route [6], promotion of the HYD route is
a good strategy for effective HDS of 4,6-DMDBT [10].
Because of severe regulations for reducing emission from
diesel vehicles, the production of ultra low sulfur diesel (ULSD)
fuel has become one of most important issues at refiner-
ies. Further reducing sulfur content of ULSD fuel would be
indispensable in near future to achieve near-zero emission of
diesel vehicle [1]. Therefore, fundamental studies on effective
hydrodesulfurization (HDS) of refractory sulfur compounds such as
NiW-based catalysts are well known to show higher activity for
HYD of aromatic hydrocarbon than CoMo and NiMo-based catalysts
[11]. Improvement in aromatic HYD activity of the NiW catalyst
thus may be one of promising ways to achieve effective HDS of 4,6-
DMDBT. However, much less attention has been paid to this catalyst
compared with the CoMo and NiMo-based catalysts. In our previous
studies, it was found that use of trans-1,2-diaminocyclohexane-
4
,6-dimethyldibenzothiophene (4,6-DMDBT) are still important.
So far, extensive efforts have been made to improve understand-
ing of fundamental aspects of HDS of 4,6-DMDBT. Based on these
ꢀ
ꢀ
N,N,N ,N -tetraacetic acid (CyDTA) for preparation of the Al O
2
3
supported NiW catalyst improved activity for C–S bond cleav-
age of 4,6-DMDBT and tetrahydro-DMDBT (THDMDBT), leading
to improvement in total HDS activity by a factor of 1.6 [12,13].
∗ Corresponding author. Tel.: +81 022 795 7215; fax: +81 022 795 7215.
E-mail address: koizumi@erec.che.tohoku.ac.jp (N. Koizumi).
0
926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.05.026

8
0
N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
Table 1
Chemical composition of the catalysts prepared in the present study.
Catalyst
Composition
WO3 (mass%)
NiO (mass%)
Ni/W (mol mol−1)
CyDTA/W (mol mol
−1
)
−1
CyDTA/Ni (mol mol )
W/Al2O3
Ni/W/Al2O3
CyDTA-Ni/W/Al2O3
W/SiO2-Al2O3
Ni/W/SiO2-Al2O3
CyDTA-Ni/W/SiO2-Al2O3
15.0
15.0
15.0
15.0
15.0
15.0
0.0
1.6
1.6
0.0
1.6
1.6
0.0
0.0
0.0
0.6
0.0
0.0
0.6
–
0.32
0.32
0.0
0.32
0.32
0.00
1.88
–
0.00
1.88
In other words, preparation with using CyDTA promoted the DDS
route. Quasi in situ XPS and in situ W LIII-edge EXAFS measurement
suggested that use of CyDTA retarded the sulfidation of Ni species
through the formation of a stable complex with Ni2 , which facili-
obtain the SiO -Al O supported NiW catalysts. For preparation
2 2 3
of the SiO -Al O supported NiW catalyst using CyDTA, aque-
2
2
3
ous solution containing both Ni(NO ) ·6H O and CyDTA (Dojindo
3
2
2
+
Chemicals, >99%) was used instead of aqueous Ni(NO ) ·6H O
3
2
2
tated coordination of Ni atoms on edge sites of WS -like slabs, i.e.
solution. Hereafter, the catalyst was denoted by, for example,
CyDTA-Ni/W/SiO -Al O , which means that the NiW catalyst was
2
the formation of the Ni–W–S phase [14,15].
2
2
3
On the other hand, use of Al O3 support modified with boric
prepared by use of SiO -Al O support in combination with
2
2 2 3
acid (B-Al O ) for the W and NiW catalysts improved activity for
CyDTA. The Al O3 supported catalysts were also prepared using
2
2
3
2
−1
HYD of 4,6-DMDBT [12,13] into THDMDBT. Although this cata-
lyst failed to promote the HYD route because of lower activity
for C–S bond cleavage of THDMDBT, it was then found that use
of B-Al O support in combination with CyDTA improved activ-
␥-Al O3 (NK-30129, specific surface area = 314 m g ) powder as
2
references. Chemical compositions of the prepared catalysts are
summarized in Table 1.
2
3
ity of the NiW catalyst not only for HYD of 4,6-DMDBT, but also
for C–S bond cleavage of 4,6-DMDBT and THDMDBT. As a result,
the HYD and DDS routes were promoted simultaneously. This cat-
alyst showed 2.4 times higher activity for HDS of 4,6-DMDBT than
2
.2. Activity test
HDS of 4,6-DMDBT was carried out using a conventional high-
pressure flow reactor [25–27]. The reactor consisted of a stainless
steel tube with internal diameter (Ø) of 8 mm in a four-zone
electronic furnace. Four sets of temperature controllers and ther-
mocouples regulated temperature in the catalyst bed within ± 1 K.
The prepared catalyst was charged in the stainless steel reactor
the conventional Al O supported NiW catalyst. Diffuse reflectance
2
3
FTIR spectroscopy (DRIFTS) showed that strong basic Al–OH species
on Al O3 surface was depleted by modification with boric acid,
2
suggesting that use of B-Al O support for the NiW catalyst in com-
2
3
bination with CyDTA affected morphology of the Ni–W–S phase
that was effectively formed by the complex formation between
and then sulfided in a stream of 5% H S/H2 (>99.99995%) at 673 K
2
and 1.1 MPa. After sulfiding pretreatment, temperature was cooled
down to 573 K. A feed consisted of 0.3 mass% 4,6-DMDBT (Aldrich,
97%) diluted with decalin (Wako Pure Chemical Industry Ltd., 99%)
was introduced into the reactor with high-pressure hydrogen flow.
HDS of 4,6-DMDBT was performed at 573 K and 5.1 MPa (H2 par-
tial pressure). H2 flow rate and H2 to feed ratio were 300 mL
2+
Ni and CyDTA. In previous studies, several authors suggested
importance of morphology of MoS2 [16–20] and/or WS2 [21]-
like slabs for promotion of the HYD route. However, these still
exists apparent discrepancy among these studies. Furthermore,
little information was obtained about structure of the promoter
atoms, such as the formation of the coordinatively unsaturated sites
−1 −1
STP) min and 1000 mL mL , respectively. The time factor (W/F)
(
(
CUS) of the Co(Ni)–Mo–S and Ni–W–S phases in these previous
studies.
Based on these studies, effect of novel SiO -Al O support on
was varied by changing catalyst weight charged.
Liquid products were collected from a product stream every
1 h, and analyzed with gas chromatographs (GC) equipped with
flame ionization detector (Shimadzu, GC-17A), mass spectrometer
2
2
3
surface structure and HDS activity of the NiW catalyst prepared
using CyDTA was investigated in this work by combination of
quasi in situ TEM, DRIFTS coupled with NO adsorption and 4,6-
(Shimadzu, QP5000) and/or an atomic emission detector (Agilent
Technology, G2350A). Temperature-programmed GC analysis with
a high-resolution capillary column (Supelco, Petrocol DH, length:
1
DMDBT HDS activity measurements. SiO -Al O support used in
2
2
3
the present study had unique surface acidic property, on which
Al O surface was covered with small SiO particles. Our attentions
00 m, internal diameter: 0.25 mm) was used for separation of the
2
3
2
product mixture.
were paid to clarify surface structure of the catalyst that effectively
promotes the HYD route in HDS of 4,6-DMDBT.
2
.3. XRD measurement
XRD pattern of ␥-Al O and SiO -Al O support used for catalyst
2
. Experimental
2
3
2
2
3
preparation was measured on a RINT diffractometer (Rigaku). Cu
K␣ radiation (ꢀ = 1.54056 Å) was used as an X-ray source with an
X-ray tube operating at 40 kV and 200 mA. Diffraction intensities
2.1. Catalyst preparation
◦
−1
were recorded at 0.02 s scan speed. Observed diffraction peaks
were assigned by referring to JCPDS data.
Preparation of the catalysts was carried out in accordance
with the preparation method for the ␥-Al O3 supported cata-
2
lysts reported previously [15,22–24]. Briefly, aqueous solution
containing (NH ) H W12O40·H O (Aldrich, >99.99%) was firstly
2.4. XPS measurement
4
6
2
2
impregnated onto sieved SiO -Al O (NK-31378, specific surface
2
2
3
2
−1
area = 320 m g ) powder followed by drying (383 K, 16 h) and cal-
XP spectra of support materials were measured with ESCA200
spectrometer (SIENTA). Monochromatized Mg K␣ line was used as
an excited source. Charge shift was corrected using BN powder as
an internal standard.
cination (673 K, 12 h). Aqueous solution containing Ni(NO ) ·6H O
3
2
2
(
Wako Pure Chemicals, purity > 99.9%) was then impregnated onto
the thus prepared W catalysts followed by drying (383 K, 16 h) to

N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
81
2.5. Quasi in situ TEM measurement
Effect of SiO -Al O support on morphology of WS -like slabs
2
2
3
2
(
or the Ni–W–S phase) was investigated by quasi in situ TEM
measurement. Before measurement, the catalyst was sulfided in
a stream of 5% H S/H2 (>99.99995%) at 673 K and 1.1 MPa using a
2
stainless steel reactor (Ø 4 mm). After sulfiding pretreatment, the
catalyst was cooled down to room temperature in the 5% H S/H2
2
stream. The reactor was then flushed with a H2 (>99.999%) stream
to remove residual H S/H , and transferred to a grove box filled
2
2
with N2 (O2 content ≤50 ppm) without contacting with the air in
the grove box. The catalyst was taken from the reactor, crashed
into fine powder using a motor and pestle, and then suspended
into ethanol solution (> 99.5%) in the grove box. A few droplets of
this suspension were dropped onto carbon-coated Cu grid followed
by drying at ambient temperature. The sample was transferred
into a vacuum chamber, and subjected to TEM measurment using
an HF-2000 (Hitachi, Ltd.) with 200 kV accelerate voltage. 40–50
Fig. 1. XRD pattern of Al2O3 (a) SiO2-Al2O3 (b) powder used for catalyst preparation.
micrographs were taken for each sample. 730–830 WS -like slabs
2
(
400–460 WS -like clusters) were analyzed for each sample to cal-
2
Selector diffuse reflectance attachment (Specac) with 4 cm 1 spec-
tral resolution. The obtained spectrum was converted into a
Kubelka–Munk function with a KBr spectrum recorded at room
temperature followed by normalization to the spectrum of the cat-
alyst before NO adsorption (also recorded at room temperature).
−
culate distribution of slab length and stacking number. Average slab
length (L) and stacking number (N) were calculated using following
equations.
¯
¯
ꢀ
n log L
i
i
¯
ꢀ
log L =
(1)
(2)
ni
ꢀ
3. Results
n N
i
i
N¯ =
ꢀ
ni
3.1. Bulk and surface properties of SiO -Al O support
3
2
2
2
.6. NH -TPD measurement
Specific surface area and pore volume of SiO -Al O powder
2 2 3
3
used in this work were almost the same as those of reference ␥-
Acidity of support was determined by temperature-
programmed desorption of preadsorbed NH (NH -TPD). NH -TPD
Al O powder. However, other physical and chemical properties of
2
3
3
3
3
SiO -Al O powder were not published by the manufacturer. Bulk
2
2
3
profile of SiO -Al O powder was compared with those of
and surface properties of this support material were thus investi-
gated for better understanding of its support effect.
2
2
3
reference ␥-Al O , SiO -Al O , conventional (amorphous) SiO -
2
3
2
2
3
2
−
1
Al O (Fuji Silysia, Al O /SiO = 8.3/91.7 mass mass
,
surface
Fig. 1 depicts XRD patterns of SiO -Al O and reference ␥-
2
3
2
3
2
2 2 3
2
−1
2
−1
)
area = 401 m g ) and SiO2 (Fuji Silysia, surface area = 230 m g
powder. The sample was pretreated at 773 K for 1 h in 60 mL min
Al O powder. Broad peaks of the ␥-Al O phase were observed
2 3 2 3
−
1
◦
◦
◦
at 2ꢁ = 37.7 , 46.2 and 67.1 in the diffraction pattern of ␥-Al O3
2
Ar flow, and then NH3 was adsorbed at ambient temperature
powder. These diffraction peaks were observed in the pattern of
−1
for 30 min in 60 mL min
NH3 adsorption, the sample was purged in 60 mL min Ar flow.
10%NH /90%He. After saturation of
SiO -Al O powder as well, whereas they were evidently weaker
3
2 2 3
−
1
and broader. Furthermore, a much weaker and broader peak was
observed in the diffraction pattern of SiO -Al O powder in a range
NH -TPD profiles were obtained by analyzing the desorbed species
3
2
2
3
◦
using quadrupole mass spectrometer (Q-MS) with heating from
of 25–30 . This broader peak was related to the SiO2 phase, sug-
gesting that the SiO2 phase was in a form of small particles or had
amorphous nature. XPS measurement was also performed to calcu-
late Si:Al atomic ratio of SiO -Al O powder in sub-surface region.
−
1
3
73 K (or room temperature) to 973 K at a rate of 10 K min . The
obtained profile was fitted with Gauss–Lorentz functions (Wave
Metrics, Igor).
2
2
3
The thus calculated Si:Al atomic ratio was 20:80.
2.7. DRIFTS coupled with NO adsorption
3.2. Acidity of support
DRIFTS coupled with NO adsorption was used to investigate
effect of support on distribution of the CUS on W and Ni promoter.
Apparatus used for measurement was reported previously [28–30].
Before NO adsorption, the catalyst was sulfided in following man-
ners. The sample holder was filled with a finely powdered catalyst,
and put in a high-pressure DRIFT chamber. The chamber was then
connected with a high-pressure flow apparatus equipped with
flow controllers and pressure regulators. 5% H S/H (>99.99995%)
NH -TPD and in situ DRIFT measurements were then performed
3
to characterize acidity of these support materials. Fig. 2 summarizes
NH -TPD profile of ␥-Al O (a), SiO -Al O (b), conventional SiO2-
3
2
3
2
2
3
Al O (c) and SiO (d) powder. Five desorption peaks were observed
2
3
2
in NH -TPD profile of ␥-Al O powder. According to the previous
3
2
3
study [31], NH -TPD profiles were arbitrarily divided into three
3
◦
regions (100–200, 200–350, 350–700 C) to provide a measure
2
1
2
−
stream was fed into the chamber (30 mL min ) at the pressure of
of weak, medium and strong acid sites, respectively. In NH -TPD
3
◦
1
1
.1 MPa, followed by heating the sample up to 673 K at a rate of
0 K min . After reaching 673 K, temperature was kept constant
profile of SiO2 powder, a peak was observed only below 100 C,
−1
indicating the formation of physisorbed NH . On the other hand,
3
◦
for 2 h. The catalyst was cooled down to room temperature followed
four desorption peaks were observed at 161, 213, 292 and 681 C
by purging residual H S/H2 in the cell with He flow (>99.99995%).
in NH -TPD profile of SiO -Al O powder, showing the presence of
2
3 2 2 3
After sulfidation pretreatment, NO adsorption was carried out
by the pulse method. DRIFT spectrum of adsorbed NO was recorded
on an FTS 575C FTIR spectrometer (Varian, Inc.) equipped with
weak, medium and strong acid sites in this support material. Com-
parison of NH -TPD profile of SiO -Al O powder with profile of
3
2
2
3
␥-Al O3 powder indicated that the medium acid site increased at
2

8
2
N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
Fig. 2. NH3-TPD profiles of Al2O3 (a), SiO2-Al2O3 (b), amorphous SiO2-Al2O3 (c), and SiO2 (d) powder. Heating rate: 10 K min−1
.
the expense of the weak acid site in SiO -Al O powder. The strong
at ambient temperature after drying pretreatment in a He stream
2
2
3
acid site was comparable between these materials. On the con-
trary, only two peaks related to the medium and strong acid sites
were observed in profile of conventional SiO -Al O powder, show-
at 773 K and 1 h. In the spectrum of ␥-Al O powder, broad IR bands
were observed at 3764, 3726, 3675 and 3578 cm . All these bands
were related to O–H stretching vibration of isolated and associ-
2
3
−
1
2
2
3
ing that SiO -Al O powder has different acidity than conventional
ated hydroxyl groups on the ␥-Al O3 surface [34]. On the other
2
2
3
2
−
1
(
amorphous) SiO -Al O powder.
hand, an IR band was observed at 3743 cm in the spectrum of
SiO2 powder, showing the presence of the isolated hydroxyl group
on the surface. In the spectrum of SiO -Al O powder, IR bands
2
2
3
Fig. 3 depicts in situ DRIFT spectrum of SiO -Al O powder
2
2
3
in the OH region in comparison with reference ␥-Al O , conven-
2
3
2
2
3
−
1
tional SiO -Al O and SiO2 powder. These spectra were recorded
were observed at 3738 and 3721 cm . Comparison with the ref-
erence spectra mentioned above indicated that the former band
was related to stretching vibration of the isolated Si–OH group,
whereas the latter band indicated the presence of some isolated
2
2
3
−
1
Al–OH group. No IR bands were observed above 3750 cm in the
spectrum of SiO -Al O powder. The spectrum of conventional
2
2
3
−
1
SiO -Al O powder provided the sharp band at 3743 cm simi-
2
2
3
lar to SiO2 powder. Broad spectral feature was also observed in
−
1
3
700–3500 cm region.
Based on these characterization results of the support mate-
rials, it was suggested that, on the SiO –Al O surface, small
2
2
3
SiO2 particles blocked up most of isolated and associated Al–OH
species on the Al O surface. Especially, strong basic Al–OH species
2
3
−
3764 cm ) [32] was totally absent from the SiO –Al O surface.
2 2 3
1
(
This will weaken W-support interaction by preventing the forma-
tion of W–O–Al bonds, and affect morphology of sulfide clusters
and the formation of the CUS. This point was then investigated by
quasi in situ TEM and DRIFTS coupled with NO adsorption.
3
.3. Morphology of WS -like slabs on the Al O and SiO -Al O
2 2 3 2 2 3
supported catalysts
3.3.1. W catalyst
Quasi in situ TEM measurements were performed after sul-
fiding pretreatment at 673 K and 1.1 MPa. In these photographs,
single-layered, ca. 4 nm-sized WS -like slabs are mainly observed,
2
and distributed homogeneously on both the catalysts. 730–830
WS2 slabs were analyzed for each catalyst to calculate dis-
tribution of their slab length (Fig. 4) and stacking number
Fig. 3. In situ DRIFT spectra of Al2O3 (a), SiO2-Al2O3 (b), amorphous SiO2-Al2O3 (c),
and SiO2 (d) powder in their OH region. The spectra were recorded in helium stream
at ambient temperature after pretreatment in helium stream at 773 K (1 h).

N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
83
Table 2
Average slab length ( L¯ ) and stacking number ( N¯ ) of WS2-like slabs on the Al2O3 and SiO2-Al2O3 supported catalysts calculated from Eqs. (1) and (2).
W/Al2O3
Ni/W/Al2O3
CyDTA-Ni/W/Al2O3
W/SiO2-Al2O3
Ni/W/SiO2-Al2O3
CyDTA-Ni/W/SiO2-Al2O3
L¯ (nm)
N¯
4.6
1.4
3.9
1.6
3.7
1.6
4.0
1.5
3.3
1.7
3.3
1.8
average slab length was 4.6 nm as tabulated in Table 2. Further-
more, the major fraction was present as single-layered slab on this
catalyst (ca. 60%). The fraction of the slabs decreased exponentially
with increasing stacking degree on this catalyst. Compared with the
Al O supported catalyst, the fraction of the WS -like slabs with 5-
2
3
2
6
nm length was slightly lower on the W/SiO -Al O catalyst. The
2 2 3
average slab length of the WS -like slabs on the W/SiO -Al O cat-
2
2
2
3
alyst was also slightly shorter compared with the Al O3 supported
2
counterpart (4.0 vs. 4.6 nm, Table 2). More clear difference was
observed in distribution of stacking degree of the WS -like slabs
2
between these catalysts (Fig. 5). The fraction of the single-layered
slab was above 55% on the Al O3 supported catalyst. This fraction
2
decreased to 45% on the SiO -Al O supported catalyst, whereas
2
2
3
the fraction of the double-layered slabs increased to around 40%
instead. Thus use of SiO -Al O support slightly increased stacking
2
2
3
degree of the WS -like slabs.
2
3.3.2. NiW catalyst
The NiW catalysts prepared with or without using CyDTA were
also subjected to quasi in situ TEM measurement. Distribution of
the WS slab length and stacking number are shown in Figs. 4 and 5.
2
The average slab length and stacking number are tabulated in
Table 2.
Fig. 4. Distribution of the length of WS2-like slabs on the Al2O3 (a) SiO2-Al2O3 (b)
supported catalysts calculated from quasi in situ TEM micrographs shown in Fig. 3
Distribution of the slab length (Fig. 4(a)) showed that the frac-
tion of the WS -like slabs with 5–6 nm length was slightly smaller
(
730–830 WS2 slabs were analyzed for each catalyst to obtain distribution).
2
on the Ni/W/Al O3 and CyDTA-Ni/W/Al O3 catalysts compared
2
2
with the W/Al O3 catalyst. The average length of the WS -like
slabs on the former catalysts was also 0.7–0.9 nm shorter than
2
2
(
Fig. 5). Average slab length and stacking degree are tabulated in
Table 2.
The length of WS -like slabs on the W/Al O catalyst was dis-
those on the W/Al O3 catalyst (Table 2). Besides, the maxi-
2
2
2
3
mum shifted toward a shorter length in distribution of the slab
tributed from 1 to 15 nm with a maximum at around 4 nm. The
length for the CyDTA-Ni/W/Al O3 catalyst compared with the
2
Ni/W/Al O3 catalyst, showing that preparation of the NiW/Al O3
2
2
catalyst using CyDTA decreased the length of the WS -like slabs.
2
Distribution of stacking degree of the slabs was slightly differ-
ent among these catalysts as well (Fig. 5(a)). The addition of Ni
to the W/Al O3 catalyst increased the fraction of the double-
2
layered slabs at the expense of the single-layered slab. Preparation
using CyDTA enhanced this trend. However, the single-layered
slabs were still major slabs in the CyDTA-Ni/W/Al O3 cata-
2
lyst.
The length and stacking degree of the WS -like slabs changed
2
in similar ways on the SiO -Al O supported catalysts. The aver-
2
2
3
age length of the WS -like slabs was 0.7-0.9 nm shorter on the
2
Ni/W/SiO -Al O and CyDTA-Ni/W/SiO -Al O catalysts compared
2
2
3
2
2
3
with the W/SiO -Al O catalyst. The maximum shifted toward a
2
2
3
shorter length in distribution of the slab length for the CyDTA-
Ni/W/SiO -Al O catalyst compared with the Ni/W/SiO -Al O
2
2
3
2
2
3
catalyst. This indicated that preparation of the NiW/SiO -Al O cat-
2
2
3
alyst with using CyDTA decreased the length of the slabs as well,
although the average lengths of the slabs on the CyDTA-Ni/W/SiO2-
Al O and Ni/W/SiO -Al O catalysts were identical. The fraction
2
3
2
2
3
of the single-layered slabs decreased in following order; W/SiO2-
Al O > Ni/W/SiO -Al O > CyDTA-Ni/W/SiO -Al O .
2
3
2
2
3
2
2
3
Compared with the Al O3 supported catalysts, the length of the
2
slabs was shorter for the SiO -Al O supported catalysts. Further-
2
2
3
more, the fraction of the single-layered slab was always smaller
on the SiO -Al O supported catalysts. Most important difference
was that the double-layered slab was major species in the CyDTA-
Ni/W/SiO2-Al2O3 catalyst, whereas the major fraction is present
2
2
3
Fig. 5. Distribution of the stacking number of WS2-like slabs on the Al2O3 (a) SiO2-
Al2O3 (b) supported catalysts calculated from quasi in situ TEM micrographs shown
in Fig. 3 (730–830 WS2 slabs were analyzed for each catalyst to obtain distribution).

8
4
N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
Fig. 7. Promoting effect of SiO2-Al2O3 support on 4,6-DMDBT HDS activity of the
NiW catalyst prepared using CyDTA. Sulfiding pretreatment: 5% H2S/H2, 673 K, 1.1
MPa (2 h) HDS of 4,6-DMDBT: 0.3 mass% 4,6-DMDBT/decalin, 573 K, 5.1 MPa (H2),
−
1
W/F = 1.3 gcat h g4,6-DMDBT
.
alyst. This suggests the formation of the Ni–W–S phase in which Ni
atoms block up the W atoms at the edge sites of the WS -like slabs.
2
The Ni/W/SiO -Al O catalyst showed a similar DRIFT spectrum.
2
2
3
−
1
However, the band at around 1760 cm was stronger compared
with the Ni/W/Al O3 catalyst, suggesting that the formation of the
2
Ni–W–S phase was suppressed on SiO -Al O support compared
2
2
3
Fig. 6. Effect of SiO2-Al2O3 support on distribution of the CUS on the NiW catalyst
prepared using CyDTA investigated by in situ DRIFTS coupled with NO adsorption;
CyDTA-Ni/W/Al2O3 (a), Ni/W/Al2O3 (b), CyDTA-Ni/W/SiO2-Al2O3 (c), Ni/W/SiO2-
Al2O3 (d), W/Al2O3 (e), Ni/Al2O3 (f) Sulfiding pretreatment: 5% H2S/H2, 673 K,
with Al O3 counterpart.
2
When the Ni/W/Al O3 catalysts was prepared using CyDTA, the
2
−
1
bands at around 1760 cm
almost disappeared. This suggested
that the Ni–W–S phase was effectively formed on this catalyst as
suggested by quasi in situ XPS and in situ W LIII-edge EXAFS mea-
surement on the catalyst sulfided at high-pressure in our previous
study [15]. CyDTA has a similar effect on the DRIFT spectra of NO
adsorbed on the SiO -Al O supported counterpart. Therefore, it
1
.1 MPa (2 h) NO pulse adsorption: 10% NO/He, ambient temperature.
as the single-layered slab in the Al O3 supported counterparts
2
(
Fig. 5(b)).
2
2
3
was suggested that use of CyDTA promoted the formation of the
Ni–W–S phase in the Ni/W/SiO -Al O catalyst as well.
3
.4. Formation of CUS on the Al O3 and SiO -Al O supported
2 2 2 3
2
2
3
catalysts
Distribution of the CUS on the Ni and W catalysts was then
investigated by DRIFTS coupled with NO adsorption. Fig. 6 shows
3.5. Effect of support on HDS activity and selectivity
DRIFT spectra of NO adsorbed on the Ni/W/Al O3 and CyDTA-
3.5.1. 4,6-DMDBT HDS activity
2
Ni/W/Al O catalysts, in comparison with those of NO adsorbed
In HDS of 4,6-DMDBT with the Al O and SiO -Al O supported
2
3
2
3
2
2
3
on the W/Al O and Ni/Al O3 catalysts. This figure also includes
catalysts, dimethylbiphenyl (DMBPh), dimethylcyclohexylbenzene
(DMCHB), and dimethyl-bicyclohexyl (DMBCH) were obtained
as HDS products. HYD products, tetrahydrodimethyldibenzoth-
iophene (THDMDBT) and hexahydrodimethyldibenzo-thiophene
(HHDMDBT), were also detected in their products. Fig. 7 com-
pares the yield of these products normalized to total metal (Ni + W)
content of the catalysts (W/F = 1.3 gcat h g4,6-DMDBT ). HDS (HYD)
activity was defined as the total yield of the HDS (HYD) products
normalized to total metal content of the catalysts.
2
3
2
spectra of the SiO -Al O supported catalysts. In the spectrum of
2
2
3
NO adsorbed on W/Al O catalyst, doublet bands were observed
2
1
3
−
at 1761 and 1694 cm . These doublet bands were similar to those
reported for Mo/Al O catalysts [33–36], and thus assigned to sym-
2
3
metric and anti-symmetric vibration of dnitrosyl species adsorbed
on the CUS of W sulfide species. In contrast, an asymmetric band
−
1
−1
was observed at 1842 cm
in the spectrum of NO adsorbed on
the Ni/Al O catalyst. This band could be assigned to symmetric
2
3
vibration of dinitrosyl species adsorbed on the CUS of Ni species,
with anti-symmetric vibration of this species overlapped in a lower
frequency region [34]. Similar band was observed in the spectrum
Effect of SiO -Al O support on HDS and HYD activity of the
2
2
3
monometal sulfide catalysts were investigated first. In HDS of 4,6-
DMDBT with the W/Al O catalyst, THDMDBT and HHDMDBT were
2
3
of NO on the Ni/W/Al O3 catalyst, whereas the intensity of this
main products. Only small amount of DMCHB was obtained with
this catalyst, whereas DMBPh was not detected in the product. On
the other hand, the DMCHB yield obtained with the W/SiO -Al O
2
band is much weaker compared with Ni/Al O3 catalyst. Further-
2
−
1
more, a weak shoulder band was observed at around 1760 cm
2
2
3
in this spectrum. By comparison with the spectrum of NO on the
catalyst was 10 times greater than that with the Al O3 supported
2
W/Al O3 catalyst, this shoulder band was assigned to symmetric
vibration of dinitrosyl species on the CUS of W sulfide species.
In our previous paper [37], the DRIFT spectrum of NO adsorbed
counterpart. Small amount of DMBPh was also detected in the prod-
uct obtained with this catalyst. Furthermore, the THDMDBT and
HHDMDBT yields were also higher for the W/SiO -Al O catalyst.
2
2
2
3
on physical mixture of the W/Al O3 and Ni/Al O3 catalysts was
Therefore, the W/SiO -Al O catalyst showed higher HDS and HYD
2
2
2 2 3
reported. Compared with the spectrum of NO on this physical mix-
ture, the relative intensity of the band of symmetric vibration of
dinitrosyl species on Ni species was much stronger for the spec-
activity than the Al O3 supported counterpart. For the Ni catalyst,
2
higher HDS and HYD activity was obtained with the SiO -Al O
2
2
3
supported catalyst as well.
trum of NO on the Ni/W/Al O3 catalyst. In other words, the CUS
In HDS of 4,6-DMDBT with the Ni/W/Al O3 catalyst, the HDS
2
2
was more selectively formed on Ni species on the Ni/W/Al O3 cat-
products (DMBPh, DMCHB and DMBCH) were main products. Small
2

N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
85
mentioned below included selectivity for the corresponding iso-
mers. It is also worthy to note that material (carbon) balance was
7
0% or less for the Ni/W/SiO -Al O catalyst when the conversion
2 2 3
was higher than 60 mol%. No cracking products were detected in
the products. Besides, carbon balance was always 90% or higher
irrespective of the conversion for the Al O3 supported catalysts. It
2
was suggested that carbon deposition occurred on acidic sites of
the SiO –Al O surface during HDS of 4,6-DMDBT.
2
2
3
Fig. 9 compares DMBPh, DMCHB and THDMDBT selectivity of the
Ni/W/SiO -Al O and Ni/W/Al O3 catalysts. DMBPh selectivity of
2
2
3
2
these catalysts was almost the same in the whole conversion range.
On the other hand, the Ni/W/SiO -Al O catalyst showed lower
2
2
3
selectivity for DMCHB than the Ni/W/Al O catalyst. In stead, THD-
2
3
MDBT selectivity of the former catalyst was higher. These results
suggested that use of SiO -Al O support promoted HYD of 4,6-
2
2
3
DMDBT into THDMDBT and/or HHDMDBT, whereas it suppressed
C–S bond cleavage of THDMDBT. Such the effect of SiO -Al O sup-
2
2
3
port will be related with the differences in morphology of WS slabs
and distribution of the CUS as described previously.
2
Fig. 8. Product selectivity over the Al2O3 supported NiW catalyst as a function of
the 4,6-DMDBT conversion; DMBPh (ꢁ), DMCHB (ꢀ), DMBCH (♦), THDMDBT (ꢂ),
HHDMDBT (×).
3
3
.6. Effect of CyDTA on HDS activity and selectivity
amounts of the HYD products (THDMDBT and HHDMDBT) were
also detected in the products. The HDS product yield obtained
with the Ni/W/SiO -Al O catalyst was only 60% of that with the
.6.1. 4,6-DMDBT HDS activity
Fig. 7 also depicts effect of CyDTA on the yield of the
2
2
3
products obtained with the Ni/W/Al O and Ni/W/SiO -Al O cata-
2
3
2
2
3
Ni/W/Al O catalyst. On the other hand, the HYD product yield with
−1
2
3
lysts (W/F = 1.3 gcat h g4,6-DMDBT ). As reported previously [12,13],
higher yields of DMBPh and DMCHB were obtained with the CyDTA-
Ni/W/Al O catalyst compared with the Ni/W/Al O catalyst. Use of
the former catalyst was evidently higher compared with the lat-
ter. Therefore, the Ni/W/SiO -Al O catalyst showed lower HDS
2
2
3
2
3
2
3
activity, but higher HYD activity than the Al O3 supported coun-
2
CyDTA improved HDS activity of the Ni/W/Al O3 catalyst by a fac-
2
terpart. As a result, total activity (total product yield normalized to
tor of 1.6. On the other hand, use of CyDTA for the Ni/W/SiO -Al O
2
2
3
total metal content) was comparable between the Ni/W/Al O3 and
2
catalyst improved the yields of DMBPh and DMCHB by factors of 3
and 5, respectively. DMBCH was also detected in the products with
the CyDTA-Ni/W/SiO -Al O catalyst. This catalyst showed three
Ni/W/SiO -Al O catalysts. Fig. 7 also shows that the addition of Ni
2
2
3
to the W/Al O catalyst improved HDS activity by a factor of 69. This
2
3
2
2
3
large activity enhancement is well known as the promoting effect
of Ni. However, when Ni was added to the W/SiO -Al O catalyst,
times higher HDS activity than the Ni/W/Al O3 catalyst. It should
2
2
2
3
be also noted that HDS activity of the CyDTA-Ni/W/SiO -Al O cata-
2
2
3
HDS activity enhancement was no more than 6. In other words,
the promoting effect of Ni was significantly suppressed by use of
SiO -Al O support. This is consistent with the results from DRIFTS
lyst was still higher than that of the B-Al O supported NiW catalyst
2
3
prepared using CyDTA reported previously [12,13].
2
2
3
coupled with NO adsorption which showed that the formation of
the Ni–W–S phase was suppressed on SiO -Al O support.
3
.6.2. Product selectivity
To investigate effect of CyDTA on product selectivity, HDS
2
2
3
of 4,6-DMDBT was conducted with the Ni/W/Al O and CyDTA-
2
3
3
.5.2. Selectivity
Fig. 7 suggested that use of SiO -Al O support instead of Al O
3
Ni/W/SiO2-Al2O3 catalysts at various W/F. DMBPh, DMCHB and
THDMDBT selectivity of these catalysts was depicted in Fig. 10
as a function of the conversion of 4,6-DMDBT. Similar to the
reaction with the Ni/W/SiO2-Al2O3 catalyst, DMCHB, DMBPh and
DMBCH isomers were detected in the products with the CyDTA-
Ni/W/SiO2-Al2O3 catalyst when the conversion was higher than
65 mol%. DMCHB, DMBPh and DMBCH selectivity depicted in Fig. 10
included the corresponding isomers. From this figure, one can easily
understand that product selectivity of the Ni/W/Al2O3 and CyDTA-
Ni/W/SiO2-Al2O3 catalysts are almost identical. On the other hand,
the yields of the HDS products, DMBPh, DMCHB and DMBCH,
obtained with the CyDTA-Ni/W/SiO2-Al2O3 catalyst were evidently
higher than those with the Ni/W/Al2O3 catalyst. These results sug-
gest that both the DDS and HYD routes are promoted on the
CyDTA-Ni/W/SiO2-Al2O3 catalyst at the same rate compared with
the Ni/W/Al2O3 catalyst. Such simultaneous promotion is a reason
for higher 4,6-DMDBT HDS activity of the CyDTA-Ni/W/SiO2-Al2O3
catalyst.
2
2
3
2
support affected HDS and HYD selectivity of the NiW catalyst. To
investigate this point in detail, HDS of 4,6-DMDBT was conducted
with the Ni/W/Al O and Ni/W/SiO -Al O catalysts at various W/F.
Product selectivity of these catalysts was compared in wide conver-
sion range.
2
3
2
2
3
Fig. 8 depicts selectivity for each product obtained with
the Ni/W/Al O catalyst as an example. At the low conversion
2
3
(
10 mol%), THDMDBT was a main product. Small amounts of
DMBPh, DMCHB and HHDMDBT were also detected at this conver-
sion. Selectivity for DMBPh, THDMDBT and HHDMDBT decreased
with the conversion of 4,6-DMDBT, whereas DMCHB and DMBCH
selectivity increased. DMCHB was a main product at a higher con-
version range (≤80 mol%). This conversion-selectivity profile was
well consistent with previous results which suggested that HDS
of 4,6-DMDBT proceeds in two parallel routes, i.e. direct desul-
furization (DDS) and HYD routes [38–40]. In addition to these
products, small amounts of DMCHB, DMBPh and DMBCH isomers
were detected in the products obtained with the Ni/W/SiO -Al O
4. Discussion
2
2
3
catalyst when the conversion was higher than 40 mol%. Because
these isomers were not observed at lower conversions, it is reason-
able to assume that they are formed from DMCHB, DMBPh and/or
DMBCH, but not from 4,6-DMDBT isomers. Therefore, selectivity
In the present study, novel SiO -Al O support with the unique
acidic property was used for preparation of the NiW catalyst in
combination with CyDTA. In this section, obtained results were dis-
2
2
3
of the Ni/W/SiO -Al O catalyst for DMCHB, DMBPh and DMBCH
cussed in terms of effect of SiO -Al O support on morphology of
2
2
3
2 2 3

8
6
N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
Fig. 9. Effect of SiO2-Al2O3 support on DMBPh, DMCHB and THDMDBT selectivity over the NiW catalyst; Ni/W/SiO2-Al2O3 (closed symbol), Ni/W/Al2O3 (open symbol).
Fig. 10. Effect of SiO2-Al2O3 support in combination with CyDTA on DMBPh, DMCHB THDMDBT selectivity over the NiW catalyst; CyDTA-Ni/W/SiO2-Al2O3 (closed symbol),
Ni/W/Al2O3 (open symbol) catalyst.

N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
87
the WS -like slabs and the formation of the CUS on the Ni promoter.
Relationship with 4,6-DMDBT HDS activity and selectivity was also
discussed here.
addition through the effective formation of the Ni–W–S phase as
suggested by DRIFTS coupled with NO adsorption.
2
4
.2. Effect of SiO -Al O support on the formation of the CUS
2 2 3
4
.1. Effect of SiO -Al O support on morphology of WS -like slabs
2 2 3 2
In situ DRIFT coupled with NO adsorption further revealed that
use of SiO -Al O support for preparation of the NiW catalyst also
Effect of support on morphology of WS -like slabs has been
2
2
2
3
investigated by several authors. Sun et al. [21] investigated effect of
modification of Al O3 support with ammonium fluoride (F-Al O )
affected distribution of the CUS; the CUS was less selectively formed
on Ni in the Ni/W/SiO -Al O catalyst compared with the Al O3
2
2
3
2
2
3
2
on morphology of the WS -like slabs in the W and NiW catalysts
supported counterpart. This suggested that the formation of the
Ni–W–S phase was suppressed on the SiO –Al O surface. The rea-
2
prepared using ammonium metatungstate (AMT) and ammonium
tetrathiotungstate (ATT) by means of high-resolution TEM. They
found that, irrespective of the type of W source, modification with
2
2
3
son for this is not so clear. Ni2+ may be trapped by some kinds of
cation-exchange sites on the SiO –Al O surface during the catalyst
2
2
3
ammoniumfluorideincreasedstackingdegree ofthe WS -like slabs
in the W and NiW catalysts. Major fractions were present as double-
preparation steps like zeolite [42], which could not be involved in
the formation of the Ni–W–S phase during sulfiding pretreatment.
2
layered slabs on the F-Al O3 supported catalysts prepared using
When the Ni/W/SiO -Al O catalyst is prepared using CyDTA, the
2
2 2 3
2+
ATT. On the other hand, modification with ammonium fluoride
hardly affected the basal plane sizes of the slabs on the ATT derived
catalysts. ver Meer et al. [41] reported inhomogeneous distribu-
formation of such Ni species will be prevented because CyDTA
forms the stable complex with Ni2+. Thus, the formation of the
Ni–W–S phase was promoted by use of CyDTA on SiO –Al O
2
2
3
tion of the WS -like slabs on the amorphous silica-alumina (ASA)
surface as well as on Al O3 surface. Combining with the results
2
2
supported NiW catalyst. The slabs were preferentially distributed
over the alumina part of ASA support. Single-layered slabs predom-
inated in these catalysts.
from quasi in situ TEM measurement, it was suggested that the
multi-layered Ni–W–S phase with the smaller basal plane size was
effectively formed by use of SiO -Al O support in combination
2
2
3
Quasi in situ TEM observation in the present study showed that
the WS -like slabs were homogeneously distributed on SiO -Al O
with CyDTA.
2
2
2
3
support as well as on Al O3 support. Stacking degree of the WS2-
4.3. 4,6-DMDBT HDS activity and selectivity
2
like slabs in the W/SiO -Al O and Ni/W/SiO -Al O catalysts was
2
2
3
2
2
3
higher compared with the Al O3 supported counterparts. Further-
So far, extensive efforts have been made to improve under-
standing of fundamental aspects of HDS of 4,6-DMDBT. Based on
these studies, it has been suggested that HDS of 4,6-DMDBT pro-
ceeds in the DDS HYD routes [2–9]. Because the HYD route is much
faster than the DDS route [6], promotion of the HYD route is a good
strategy for effective HDS of 4,6-DMDBT.
2
more, the WS -like slabs on the SiO -Al O supported catalysts had
2
2
2
3
smaller basal plane sizes compared with the Al O supported coun-
2
3
terparts. Distribution of stacking number of the WS -like slabs in
2
these SiO -Al O supported catalysts was similar to those on the
2
2
3
F-Al O supported catalysts prepared using ATT [21], whereas it
2
3
was reported that use of F-Al O3 support hardly affected the basal
From this point of view, several authors investigated effects of
additives such as boric acid and ammonium fluoride on the reac-
tion routes in HDS of DBT and 4,6-DMDBT on NiMo and CoMo
2
plane sizes of the slabs. DRIFT measurement on SiO -Al O sup-
2
2
3
port after drying in He showed that strong basic Al–OH species was
absent from the surface of this support material, indicating that
catalysts. Among these studies, the addition of boric acid to Al O3
2
weak W-support interaction leads to the formation of the WS -like
slabs with higher stacking degree, but smaller basal plane sizes. Dif-
support was reported to promote the HYD route in HDS of DBT
and 4,6-DMDBT on NiMo catalyst [16–19]. Based on Mo K-edge
EXAFS and XPS measurement, Li et al. [16] suggested that the for-
2
ferent effect of SiO -Al O and F-Al O3 support on the basal plane
2
2
3
2
sizes of the WS -like slabs suggests that W-support interaction is
mation of single-layered MoS -like slabs was responsible for the
2
2
weaker for SiO -Al O support.
promotion of the HYD route. On the contrary, Kagami et al. [19]
2
2
3
Use of SiO -Al O support for the NiW catalyst in combination
found that preparation of NiMo/Al O3 using NTA improved 4,6-
2
2
3
2
with CyDTA further increased stacking degree of the WS -like slabs.
DMDBT HDS activity by a factor of 2. They suggested that the
2
The double-layered WS -like slabs became predominant over this
formation of MoS -like slabs with higher stacking degree was effec-
2
2
catalyst. It also decreased the basal plane size of the slabs. CyDTA
tive for promoting the HYD route, leading to higher HDS activity.
showed similar effect for the Al O3 supported catalyst. Obviously,
For CoMo/Al O catalyst, Kwak et al. [20] reported that the addition
2
2 3
CyDTA enhanced the morphological changes of the WS -like slabs
of ammonium fluoride to Al O3 support promoted the HDY route,
2
2
induced by Ni addition. It is then stressed that these morphologi-
cal changes induced by CyDTA and/or Ni addition were similar to
those observed for the SiO -Al O supported catalysts in compari-
leading to improvement in 4,6-DMDBT HDS activity by a factor
of 1.6. Based on temperature-programmed sulfidation measure-
ment, they suggested that ammonium fluoride addition improved
the dispersion of the MoS2 slabs, which was responsible for the
promotion of the HYD route. All these studies suggested the impor-
2
2
3
son with the Al O3 supported counterparts. Thus their effect could
2
be explained in terms of W-support interaction; In our previous
study, in situ W LIII-edge EXAFS measurement of the W/Al O3 and
tance of morphology of the MoS -like slabs for promotion of the
2
2
Ni/W/Al O catalysts sulfided at high-pressure showed that the
HYD route, although apparent discrepancies exist among them.
One should also remind that different types of catalytic functions,
i.e. HYD of the aromatic ring and subsequent C–S bond cleavage
of hydrogenated 4,6-DMDBT, are required to promote the HYD
route. The higher ability for C–S bond cleavage of 4,6-DMDBT and
hydrogenated DMDBT is suggested being brought by Co and/or Ni
addition [8,43,40]. Therefore, surface structure of these promoters
should be also taken into consideration to establish the structure-
activity relationship for the HYD route in HDS of 4,6-DMDBT.
It was found in the present study that the CyDTA-Ni/W/SiO2-
2
3
addition of the Ni promoter to the W/Al O3 catalyst increased sul-
2
fiding degree of the WS -like slabs [15]. Higher sulfiding degree
2
of the WS -like slabs on the promoted catalyst would be brought
2
by breaking of some W–O–Al bonds during sulfiding pretreatment,
due to higher ability for H2 dissociation of the Ni atoms. Breaking
of some W–O–Al bonds results in weaker W-support interaction,
leading to the formation of the WS -like slabs with higher stacking
2
degree, but smaller basal plane sizes after sulfiding pretreatment.
This mechanism would effectively work when the Ni atoms are
located near WS -like slabs, i.e. during the formation of the Ni-
Al O catalyst showed three times higher 4,6-DMDBT HDS activity
2 3
2
Mo-S phase. Thus, CyDTA enhanced morphological changes by Ni
than the Ni/W/Al O3 catalyst. This activity enhancement was
2

8
8
N. Koizumi et al. / Applied Catalysis A: General 383 (2010) 79–88
greater than the difference in NO adsorption capability of these two
catalysts. It is also noted that 4,6-DMDBT HDS activity of the CyDTA-
Ni/W/SiO -Al O catalyst was still higher than that of the B-Al O3
(4) The formation of the Ni–W–S phase was effectively promoted
when the SiO -Al O supported NiW catalyst was prepared
2
2
3
using CyDTA, which was accompanied with further increase
of stacking degree and decreased the basal plane size of the
2
2
3
2
supported NiW catalyst prepared using CyDTA reported previously
12,13]. The activity enhancement by use of SiO -Al O support in
[
WS -like slabs.
2
2
3
2
combination with CyDTA was greater than those reported previ-
ously (also see above).
(5) This catalyst showed higher activity for both HYD of 4,6-
DMDBT, and C–S bond cleavage of 4,6-DMDBT and THDMDBT.
The HYD and DDS routes in HDS of 4,6-DMDBT were pro-
moted simultaneously on this catalyst, leading to three times
It was also found that use of SiO -Al O support promoted HYD
2
2
3
of 4,6-DMDBT to THDMDBT and/or HHDMDBT over the W cata-
lyst. Taking the quasi in situ TEM results into consideration, it was
higher HDS activity than the Al O3 supported NiW catalyst.
2
suggested that the formation of the WS -like slabs with higher
These results suggested that the formation of the multi-layered
Ni–W–S phase with smaller basal plane sizes is crucial for pro-
moting the HYD route in HDS of 4,6-DMDBT.
2
stacking degree, but the smaller basal plane size was effective for
promoting HYD of 4,6-DMDBT. Higher stacking degree of the WS2-
like slabs is effective for HYD of 4,6-DMDBT, probably because the
slabs with higher stacking degree will reduce steric hindrance for
References
␲
-adsorption of 4,6-DMDBT onto their coordinatively unsaturated
[
[
1] C. Song, Catal. Today 86 (2003) 211–263.
2] D.R. Kilanowski, H. Teeuwen, V.H. de Beer, B.C. Gates, G.C.A. Schuit, H. Kwart, J.
Catal. 55 (1978) 129–137.
edge and corner sites. On the other hand, the smaller basal plane
size increases the fraction of the corner sites, which will also reduce
steric hindrance for ␲-adsorption of 4,6-DMDBT. It is noted, how-
ever, that the W/SiO -Al O catalyst showed higher HDS activity
[3] M. Houalla, D.H. Broderick, A.V. Sapre, N.K. Nag, V.H. de Beer, B.C. Gates, H.
Kwart, J. Catal. 61 (1980) 523–527.
[
[
2
2
3
4] T. Kabe, A. Ishihara, Q. Zhang, Appl. Catal. A 97 (1993) L1–L9.
5] T. Isoda, M. Xiaoliang, I. Mochida, Sekiyu Gakkaishi 37 (1994) 368–375.
than the W/Al O catalyst. Thus, the WS2-like slabs with the above
2
3
morphology also have weak activity for C–S bond cleavage of 4,6-
DMDBT and/or THDMDBT. This structure-activity relationship will
also hold true for higher activity for HYD of 4,6-DMDBT over the
[6] P. Michaud, J.L. Lemberton, G. Pérot, Appl. Catal. A 169 (1998) 343–353.
[7] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29–36.
[
8] F. Bataille, J.L. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire, E. Schulz,
M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409–422.
Ni/W/SiO -Al O catalyst
2
2
3
[9] S. Bendezú, R. Cid, J.L.G. Fierro, A.L. Agudo, Appl. Catal. A 197 (2000) 47–60.
[10] K. Segawa, Shokubai 48 (2006) 228–233.
11] H. Yasuda, M. Higo, S. Yoshitomi, T. Sato, M. Imamura, H. Matsubayashi, H.
Shimada, A. Nishijima, Y. Yoshimura, Catal. Today 39 (1997) 77–87.
12] N. Koizumi, M. Shingu, K. Hata, Y. Murata, H. Itou, M. Yamada, Prep. Am. Chem.
Soc., Div. Pet. Chem. 49 (2004) 292–296.
13] K. Hata, Master Thesis, Department of Applied Chemistry, Graduate School of
Engineering, Tohoku University, 2004.
[14] Y. Ohta, Master Thesis, Department of Applied Chemistry, Graduate School of
Engineering, Tohoku University, 2000.
On the other hand, activity for C–S breaking of 4,6-DMDBT and
THDMDBT was suppressed by use of SiO -Al O support for the
[
[
[
2
2
3
NiW catalyst. This was apparently related with the fact that the
formation of the Ni–W–S phase was suppressed on the SiO –Al O
2
2
3
surface as suggested by DRIFTS coupled with NO adsorption. When
the SiO -Al O supported NiW catalyst was prepared using CyDTA,
2
2
3
the formation of the Ni–W–S phase was promoted, which was
accompanied with further increase of stacking degree and decrease
[
[
[
[
15] H. Itoh, N. Koizumi, N. Sakamoto, T. Honma, M. Shingu, M. Yamada, J. Jpn. Pet.
Inst. 47 (2004) 249–257.
16] D. Li, T. Sato, M. Imamura, H. Shimada, A. Nishijima, J. Catal. 170 (1997) 357–
365.
of the basal plane size of the WS -like slabs. Such sulfide structure
2
is effective for both HYD of 4,6-DMDBT, and C–S bond cleavage of
17] D. Li, T. Sato, M. Imamura, H. Shimada, A. Nishijima, Appl. Catal. B 16 (1998)
4
,6-DMDBT and THDMDBT. The HYD and DDS routes were thus
255–260.
promoted simultaneously on this catalyst, leading to higher HDS
activity. Our results suggest that the formation of the multi-layered
Ni–W–S phase with the smaller basal plane size is crucial for pro-
moting the HYD route in HDS of 4,6-DMDBT.
18] P. Torres-Mancera, J. Ramírez, R. Cuevas, A. Gutiérrez-Alejandre, F. Murrieta, R.
Luna, Catal. Today 107 (2005) 551–558.
[19] N. Kagami, B.M. Vogelaar, A.D. van Langeveld, J.A. Moulijn, Appl. Catal. A 293
2005) 11–23.
(
[
[
[
20] C. Kwak, J.J. Lee, J.S. Bae, K. Choi, S.H. Moon, Appl. Catal. A 200 (2000) 233–242.
21] M. Sun, P.J. Kooyman, R. Prins, J. Catal. 206 (2002) 368–375.
22] Y. Ohta, T. Shimizu, T. Honma, M. Yamada, Stud. Surf. Sci. Catal. 127 (1999)
5
. Conclusions
1
61–168.
23] T. Shimizu, K. Hiroshima, T. Honma, T. Mochizuki, M. Yamada, Catal. Today 45
1998) 271–276.
24] N. Koizumi, Y. Murata, N. Sakamoto, H. Itoh, M. Yamada, J. Jpn. Pet. Inst. 48
2005) 344–350.
[25] S. Kasahara, N. Koizumi, J. Iwahashi, M. Yamada, Sekiyu Gakkaishi 38 (1995)
45–352.
[
[
(
Effect of novel SiO -Al O support on surface structure and HDS
2
2
3
activity of the NiW catalyst prepared using CyDTA was investi-
gated by combination of quasi in situ TEM, DRIFTS coupled with NO
adsorption and 4,6-DMDBT HDS activity measurements. Important
results obtained are summarized as follows.
(
3
[
[
[
26] S. Kasahara, T. Shimizu, M. Yamada, Catal. Today 35 (1997) 59–64.
27] S. Kasahara, Y. Udagawa, M. Yamada, Appl. Catal. B 12 (1997) 225–235.
28] N. Koizumi, K. Takahashi, M. Yamazaki, M. Yamada, Catal. Today 45 (1998)
3
13–318.
(
1) In situ DRIFT measurement on SiO -Al O support showed that
2 2 3
[29] M. Yamada, N. Koizumi, M. Yamazaki, Catal. Today 50 (1999) 3–8.
[30] M. Yamada, Catal. Surv. Jpn. 3 (1999) 3–15.
strong Al–OH species was absent from the surface of this sup-
port material.
2) Without using the chelating agent, quasi in situ TEM mea-
[
31] S. Rajagopal, T.L. Grimm, D.J. Collins, R. Miranda, J. Catal. 137 (1992) 453–
61.
32] H. Knözinger, P. Ratnasamy, Catal. Rev. 17 (1978) 31–70.
4
(
[
surement showed that the WS -like slabs on the SiO -Al O
[33] W.S. Millman, W.K. Hall, J. Phys. Chem. 83 (1979) 427–428.
34] N.-Y. Topsøe, H. Topsøe, J. Catal. 84 (1983) 386–401.
35] Y. Okamoto, Y. Katoh, Y. Mori, T. Imanaka, S. Teranishi, J. Catal. 70 (1981)
2
2
2
3
[
[
supported W and NiW catalysts had higher stacking degree, but
smaller basal plane sizes compared with the Al O3 supported
2
445–448.
counterparts. On the other hand, the formation of the Ni–W–S
phase was suppressed when SiO -Al O support was used for
[36] Y. Okamoto, A. Maezawa, T. Imanaka, J. Catal. 120 (1989) 29–45.
[37] N. Koizumi, M. Yamazaki, M. Iijima, M. Yamada, Appl. Surf. Sci. 121–122 (1997)
2
2
3
429–432.
the NiW catalyst instead of Al O3 support.
2
[
38] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345–471.
(
3) The SiO -Al O supported NiW catalyst showed higher activity
2 2 3
[39] M. Egorova, R. Prins, J. Catal. 224 (2004) 278–287.
[40] M. Egorova, R. Prins, J. Catal. 225 (2004) 417–427.
for HYD of 4,6-DMDBT into THDMDBT than the Al O supported
2
3
[
41] Y. ver Meer, E.J.M. Hensen, J.A.R. van Veen, A.M. ver Kraan, J. Catal. 228 (2004)
33–446.
42] H. Itoh, Doctor thesis, 2004, Tohoku University, Japan.
counterpart. However, because of lower activity for C–S bond
cleavage of THDMDBT, this catalyst failed to promote the HYD
route, and showed lower HDS activity than the Al O supported
4
[
[43] N. Koizumi, K. Hata, A. Nishina, M. Yoshida, M. Yamada, Prep. Am. Chem. Soc.,
Div. Pet. Chem. 50 (2005) 382–385.
2
3
NiW catalyst.