Mesoporous zeolite-supported metal sulfide catalysts with high activities in the deep hydrogenation of phenanthrene

Article abstract of DOI:10.1016/j.jcat.2015.07.026

Developing highly active hydrogenation catalysts for deep aromatics saturation is of great importance in the production of ultraclean diesel fuel at a low cost. Toward this goal, we synthesized a mesoporous zeolite ZSM-5 (MZSM-5) that was cost-effective and available on a large scale, and used it as a support for the preparation of highly efficient metal sulfide catalysts (NiMoS/MZSM-5 and CoMoS/MZSM-5) for the deep hydrogenation of phenanthrene. The intrinsic activity of the NiMoS/MZSM-5 catalyst (7.4 × 10^-4 mol kg^-1 s^-1) was much higher than that of the alumina-supported NiMo catalyst (NiMoS/γ-Al₂O₃, 4.8 × 10^-4 mol kg^-1 s^-1), and the selectivity of the deep hydrogenation products over NiMoS/MZSM-5 (20.9%) was higher than for NiMoS/γ-Al₂O₃ (15.2%). Compared with γ-Al₂O₃, the relatively weak metal-support interaction could facilitate the formation of polymolybdates on MZSM-5. After sulfidation, the more multistacked MoS₂ active phases were formed on the MZSM-5, enhancing the hydrogenation activity of the NiMoS/MZSM-5 catalyst.

Full text of DOI:10.1016/j.jcat.2015.07.026

Journal of Catalysis 330 (2015) 423–433
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Mesoporous zeolite-supported metal sulfide catalysts with high
activities in the deep hydrogenation of phenanthrene
b
a
a
a
Wenqian Fu a, Lei Zhang b, Dongfang Wu a, , Mei Xiang , Qian Zhuo , Kai Huang , Zhongdong Tao ,
⇑
Tiandi Tang b,
⇑
a School of Chemistry and Chemical Engineering, Southeast University, Nanjing 211189, People’s Republic of China
b College of Chemistry and Materials Engineering, Wenzhou University, Wenzhou 325035, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
Developing highly active hydrogenation catalysts for deep aromatics saturation is of great importance in
the production of ultraclean diesel fuel at a low cost. Toward this goal, we synthesized a mesoporous
zeolite ZSM-5 (MZSM-5) that was cost-effective and available on a large scale, and used it as a support
for the preparation of highly efficient metal sulfide catalysts (NiMoS/MZSM-5 and CoMoS/MZSM-5) for
the deep hydrogenation of phenanthrene. The intrinsic activity of the NiMoS/MZSM-5 catalyst
Received 7 May 2015
Revised 2 July 2015
Accepted 29 July 2015
(7.4 ꢀ 10ꢁ4 mol kgꢁ1 sꢁ1
) was much higher than that of the alumina-supported NiMo catalyst
Keywords:
Mesoporous zeolite ZSM-5
NiMo catalyst
Phenanthrene
Deep hydrogenation
(NiMoS/
c
-Al2O3, 4.8 ꢀ 10ꢁ4 mol kgꢁ1 sꢁ1), and the selectivity of the deep hydrogenation products over
NiMoS/MZSM-5 (20.9%) was higher than for NiMoS/c-Al2O3 (15.2%). Compared with c-Al2O3, the
relatively weak metal–support interaction could facilitate the formation of polymolybdates on
MZSM-5. After sulfidation, the more multistacked MoS2 active phases were formed on the MZSM-5,
enhancing the hydrogenation activity of the NiMoS/MZSM-5 catalyst.
Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction
[19–23], surface area [24–26], and pore structure [27–29]. For
instance, the activity of a metal sulfide supported on c-Al2O3 can
The high aromatic content in diesel fuel lowers the fuel quality
and contributes to the formation of undesirable emissions in
exhaust gases [1–5]. A solution to this problem is to saturate the
be enhanced through addition of additives (F, P, and B as well as
acidic zeolite) to the support by modifying the acidity of the
catalyst or increasing the metal dispersion [19–23]. However, the
deep hydrogenation of aromatics remains a challenge. Ordered
mesoporous molecular sieves with high surface area and mesopore
volume, such as MCM-41 and SBA-15, were used as supports of
metal sulfide catalysts to increasing the metal dispersion and mass
transfer, resulting in high catalytic activity [26,30,31]. However,
the relatively lower hydrothermal stability of this type of material
severely limits its industrial applications [13,30]. Crystalline
zeolites with unique porous structures and modifiable surface
properties are widely used in oil and gas processing [32]. For example,
metal sulfide clusters (MoSx) dispersed inside the micropores of
zeolite Y and ZSM-5 have high activity in toluene hydrogenation
and hydrodesulfurization of dibenzothiophene [33,34]. Neverthe-
less, conventional zeolites cannot accomplish deep hydrogenation
of the polyaromatic hydrocarbons due to their pore size limitation
[35–37]. Recently, mesoporous zeolites with excellent catalytic
activity for the conversion of bulky molecules have been success-
fully synthesized [38–46]. In particular, metal sulfide catalysts
(CoMoS2 and NiMoS2) supported on mesostructured zeolite
nanofiber assemblies show unprecedented hydrodesulfurization
activity in the deep hydrodesulfurization (HDS) of the bulky
aromatics by hydrogenation over conventional
c-Al2O3-supported
metal sulfide catalysts (CoMoS/ -Al2O3 and NiMoS/
c
c
-Al2O3)
[5–10]. However, the deep saturation of aromatics over these types
of catalysts is difficult because of their low catalytic activities
[2,11]. Because aromatic hydrogenation is an exothermic and
reversible reaction, catalysts with high activities at low tempera-
ture and pressure could benefit deep hydrogenation. Extensive
studies show that noble metal catalysts such as Pd, Pt, and Pd–Pt
have high activities [12–14], but they are fairly sensitive to sulfur
and nitrogen compounds, severely limiting their practical applica-
tions [15–18]. Therefore, developing highly active catalysts with
good sulfur resistance for the deep hydrogenation of aromatics at
low temperature and pressure is of great importance for the
cost-effective production of clean diesel fuel.
The hydrogenation activity of supported metal catalysts is
greatly influenced by the nature of the support, such as its acidity
⇑
Corresponding authors. Fax: +86 025 52090618 (D. Wu), +86 577 86689300
(T. Tang).
E-mail addresses: dfwu@seu.edu.cn (D. Wu), tangtiandi@wzu.edu.cn (T. Tang).
http://dx.doi.org/10.1016/j.jcat.2015.07.026
0021-9517/Ó 2015 Elsevier Inc. All rights reserved.

424
W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
4,6-dimethyldibenzothiophene [47]. As an extension of this work,
we report here a facile method for synthesizing mesoporous
ZSM-5 (MZSM-5) at low cost at a large scale. After loading nickel
(or cobalt) and molybdenum, followed by sulfidation, the
MZSM-5-supported metal sulfide catalysts (NiMoS/MZSM-5 and
CoMoS/MZSM-5) exhibit high activity in the deep hydrogenation
of bulky aromatic phenanthrene under mild conditions (280 °C
CoMoS/NB-MOR, and CoMoS/
EDTA was 1:2:1 and the Mo loading was 12.0 wt.%. MZSM-5- and
-Al2O3-supported Ni (3.7 wt.%) and Mo (12.0 wt.%) catalysts were
c-Al2O3. The molar ratio of Co/Mo/
c
prepared in the same way.
MZSM-5-supported Pd catalyst (Pd/MZSM-5) was prepared by
incipient wetness impregnation of MZSM-5 with an aqueous solu-
tion of [Pd(NH3)4]Cl2. After impregnation, the sample was dried in
air at ambient temperature for 20 h and further dried at 120 °C for
12 h. The dried sample was finally calcined at 450 °C for 4 h. The Pd
loading of this catalyst was 3.0 wt.%.
and 5 MPa), as compared with conventional NiMoS/c-Al2O3 and
Pd catalysts supported on MZSM-5 (Pd/MZSM-5). This work pro-
vides a new opportunity for the preparation of series of highly
active hydrogenation catalysts for the deep saturation of polyaro-
matics. This is of great significance for increasing fuel quality and
controlling undesirable emissions in exhaust gases.
2.3. Characterization
An X-ray diffraction (XRD) pattern was obtained with a RIGAKU
UltimalV diffractometer using Cu
Ka radiation. Nitrogen
2. Experimental
physisorption was conducted at ꢁ196 °C on a Micromeritics ASAP
2020 M apparatus. The sample was degassed for 8 h at 300 °C
before the measurement. Specific surface area was calculated from
the adsorption data using the Brunauer–Emmett–Teller (BET)
equation. The pore size distribution was calculated according to
the Barrett–Joyner–Halenda (BJH) model using adsorption data.
Fourier transform infrared spectroscopy (FT-IR) of the MZSM-5
sample was performed on a Bruker TENSOR 27 equipped with a
reactor cell. To investigate the surface properties of the MZSM-5,
the sample was outgassed overnight at 450 °C and 7.5 Pa before
IR measurement. Thermogravimetric (TG) analyses of the fresh
and spent catalysts were carried out from 40 to 900 °C at
10 °C minꢁ1 under flowing air (20 mL minꢁ1) on a STA6000 instru-
ment. For the spent catalysts, the sample was thoroughly washed
with ethanol and dried at 60 °C for 12 h before TG analysis. The
acidity of the sulfided catalysts was measured using stepwise
temperature-programmed desorption of ammonia (NH3-STPD) on
a Micromeritics ASAP2920 instrument. A 200-mg sample was
placed in a quartz tube and pretreated in a helium stream at
450 °C for 2 h. After the sample was cooled to 120 °C, NH3–He
mixed gas (10 vol.% NH3) was passed over the sample for 30 min.
After removal of the physically adsorbed NH3 by flowing helium
for 2 h at 120 °C, the sample was treated as follows: (1) increasing
the temperature from 120 to 180 °C at a rate of 10 °C minꢁ1 and
holding at 180 °C for 30 min, (2) increasing the temperature from
180 to 250 °C at a rate of 10 °C minꢁ1 and holding at 250 °C for
30 min, (3) increasing the temperature from 250 to 300 °C at a rate
of 10 °C minꢁ1 and holding at 350 °C for 30 min, and (4) increasing
the temperature from 350 to 800 °C at a rate of 10 °C minꢁ1 and
holding at 800 °C for 30 min. The desorbed NH3 was collected in
dilute hydrochloric acid and titrated with a dilute sodium hydrox-
ide solution to determine the acidic site density of the samples.
Temperature-programmed reduction (TPR) of the dried catalysts
was also conducted on a Micromeritics ASAP2920 instrument
equipped with a cold trap (ꢁ80 °C, filled with a mixture of iso-
propanol and liquid nitrogen) that was installed in front of the
TCD entrance. Thus, the products caused by decomposition of
the EDTA and metal precursor in the sample could be trapped in the
cold trap (discussed in the Supporting Information; see Fig. S1). A
dried sample (50 mg) was placed in a quartz tube and heated to
1000 °C at 15 °C minꢁ1 in a H2–Ar (10 vol.% H2) gas mixture stream.
The UV–vis diffuse reflectance spectra (UV–vis DRS) were
obtained on a Perkin-Elmer Lambda25 spectrometer with an
integration sphere. The Raman spectra were measured with a Ren-
ishaw Raman 2000 microprobe (532 nm laser excitation). X-ray
photoelectron spectroscopy (XPS) was performed on an ESCALAB
MK II system. Before analysis, the dried catalyst was sulfided in
mixed H2S–H2 gas with 10% H2S (60 mL minꢁ1, STP) at 400 °C for
180 min. After sulfidation, the catalyst was purged by He
(99.999%, 80 mL minꢁ1) at 400 °C for 60 min. After cooling to room
temperature, the sulfided catalyst was transferred under a helium
2.1. Material synthesis
MZSM-5 was synthesized hydrothermally from an aluminosili-
cate gel with a molar composition of Al2O3/50SiO2/16.2Na2O/
0.015RCC/1490H2O, where RCC was a random cationic copolymer
that contained quaternary ammonium groups and was synthesized
using cheap starting materials [47]. In a typical run, 20.4 L of water
glass was mixed with 23.2 L NaOH aqueous solution (2.1 wt.%) fol-
lowed by 3.0 L RCC. After stirring at room temperature for 2 h,
22.2 L acidic Al2(SO4)3 aqueous solution (3.7 wt.%) was added.
The mixture was further stirred for 2 h to yield an aluminosilicate
gel. The gel was transferred into a stainless steel autoclave (100 L)
for dynamic crystallization at 170 °C for 3 days. After filtration and
washing, the sample was dried at 120 °C overnight and calcined in
air at 550 °C for 5 h. Mesopore-free ZSM-5 was synthesized by the
same procedure except for the absence of RCC. Nanofiber bundles
of mordenite (NB-MOR) with mesoporous structure were synthe-
sized according to a previous work [47].
2.2. Catalyst preparation and pretreatment
The NiMo catalysts were prepared by an incipient wetness
impregnation method using an ammonia solution containing required
amounts of ammonium heptamolybdate ((NH4)6Mo7O24ꢂ4H2O),
nickel nitrate (Ni(NO3)2ꢂ6H2O), and ethylenediaminetetraacetic
acid (EDTA). The molar ratio of Ni/Mo/EDTA was 1:2:1 and the
Mo loading was 12.0 wt.%. The pH value of the impregnation
solution was about 11. MZSM-5, ZSM-5, NB-MOR, and c-Al2O3 in
powder form were used as supports to prepare the catalysts.
After impregnation, the sample was dried under ambient condi-
tions for 20 h and subsequently dried at 120 °C for 12 h without
calcination. The dried powder catalyst sample was tableted
(15 MPa) and crushed and sieved to form 40–60 mesh particles.
The dried particle sample was directly presulfided in a gas mixture
of H2–H2S (10 vol.% H2S) from room temperature to 400 °C at a
heating rate of 2 °C minꢁ1 and kept for 3 h at 400 °C before cat-
alytic testing. The dried catalysts with MZSM-5, ZSM-5, NB-MOR
and
NiMo/ZSM-5, NiMo/NB-MOR, and NiMo/
respondingly, the sulfided catalysts were designated as NiMoS/
MZSM-5, NiMoS/ZSM-5, NiMoS/NB-MOR, and NiMoS/ -Al2O3,
c-Al2O3 supports were designated as NiMo/MZSM-5,
c
-Al2O3, respectively. Cor-
c
respectively. To investigate the relationship between the active
phase morphology and the hydrogenation activity on the
MZSM-5-supported NiMo catalysts, the catalysts with different
Mo loadings (3.0 and 6.0 wt.%) were prepared by the same
procedure, and the sulfided catalysts were designated as
NiMoS/MZSM-5-L and NiMoS/MZSM-5-I. MZSM-5-, NB-MOR-,
and
c-Al2O3-supported CoMo catalysts were prepared similarly.
The sulfided catalysts were designated as CoMoS/MZSM-5,

W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
425
stream into a bottle filled with absolute cyclohexane. For XPS
experiments, the cyclohexane in the bottle was removed and the
residual sulfided catalyst was quickly moved to a sample holder
and then transferred into an analysis chamber in the XPS instru-
ment. For comparison, the spent catalysts were also analyzed by
XPS.
external diffusion limitations were checked by the Weisz–Prater
criterion (CWP) and the Mears criterion (CM) [51], respectively:
robsq
cR2P
Deff CS
CWP
¼
;
The morphology of the MZSM-5 sample was observed with a
field emission scanning electron microscope (SEM) on a SUPRA55
apparatus operated at an accelerating voltage of 5 kV. The sample
was coated with gold to create contrast. The transmission electron
microscopy (TEM) image was obtained on a JEM-2100F microscope
with a limited line resolution capacity of 1.4 Å at 200 kV. Before
characterization by TEM, the samples were cut into thin slices
and dropped onto a Cu grid coated with carbon membrane. Typi-
cally, at least 40 TEM micrographs were measured for the sulfided
robs
kcCAb
q
bRPn
CM
¼
;
where robs is the observed reaction rate, mol kgꢁ1 sꢁ1; n is the reac-
tion order; RP is the catalyst particle radius, m; c is the bulk density
q
of the catalyst bed, kg mꢁ3
bed, kg mꢁ3
b = (1 ꢁ
;
qb is the bulk density of the catalyst
e)qc; e is porosity; CS is the bulk concentra-
,
q
tion of phenanthrene at the external surface of the catalyst,
mol mꢁ3; CAb is the bulk concentration of phenanthrene, mol mꢁ3
and kc is the external mass transfer coefficient, m sꢁ1
The heat transfer resistance was checked by the Mears criterion
[51],
;
.
NiMo catalysts supported on MZSM-5 and c-Al2O3, including 400–
500 slabs taken from different parts of each catalyst. To measure
the MoS2 dispersion (fMo), which is defined as the average fraction
of Mo atoms at the MoS2 edge surface, we assumed that the MoS2
slabs are present as perfect hexagons. The fMo was calculated by
dividing the total number of Mo atoms at the edge surface by the
total number of Mo atoms using the slab sizes determined from
the TEM micrographs [48],
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢁ
D
HrobsqbR E
p
CMH
¼
;
ꢀ
ꢀ
ꢀ
ꢀ
hpT2Rg
where 4H is the heat of reaction, kJ molꢁ1; E is the activation
energy, kJ molꢁ1; h is the heat transfer coefficient between gas
Ri¼1;...;tð6ni ꢁ 6Þ=Ri¼1;...;tð3n2i ꢁ 3ni þ 1Þ
and pellet, kJ mꢁ2 sꢁ1 kꢁ1; Rg is the gas constant, kJ molꢁ1 Kꢁ1
and T is the reaction temperature, K.
;
fMo
¼
where ni is the number of Mo atoms along one edge of a MoS2 slab
determined from its length (L = 3.2(2ni ꢁ 1) Å), and t is the total
number of slabs shown in the TEM micrographs. The average slab
The values for CWP, CM and CMH were 0.001, 0.0036, and
6.5 ꢀ 10ꢁ7 under experimental conditions (see the Supporting
Information). In general, the internal and external diffusion limita-
tions and the heat transfer effect can be neglected during the
kinetic experiments when the values of CWP, CM and CMH are below
1, 0.15, and 0.15, respectively. Thus, in our case, the internal and
external diffusion limitations and the heat transfer effect could
be neglected during the kinetic experiments. In this case, the
observed reaction rate (robs) can be used to assess the intrinsic
hydrogenation activity of the catalysts. robs was calculated as [52]
length ðLÞ and stacking layer number ðNÞ were calculated according
to the equations [49,50]
P
n
i¼1xiLi
L ¼ P
n
i¼1xi
P
n
i¼1yiNi
N ¼
P
n
i¼1yi
dx
dðW=FÞ
robs
¼
;
where Li is the length of slab i, Ni is the number of layers in slab i, xi
is the number of slabs with length Li, and yi is the number of slabs
with Ni layers.
where F is the reactant molar flow (mol sꢁ1), x is the phenanthrene
conversion (%), and W is the catalyst mass (kg).
The phenanthrene hydrogenation turnover frequencies (TOF,
sꢁ1) were calculated according to the equation [48]
2.4. Activity tests
The phenanthrene (PHE) hydrogenation was carried out in a
fixed-bed continuous-flow stainless steel reactor (internal diame-
ter 6.5 mm and length 500 mm). A sample of 0.2 g presulfided cat-
alyst was diluted with 1.5 g SiC before being loaded into the
reactor. Both ends of the catalyst bed were filled with extra 0.3 g
SiC. The calcined Pd catalyst was pretreated in situ with flowing
H2 (40 mL minꢁ1) at 300 °C for 2 h before catalytic testing. The
operating conditions were as follows: total pressure of 5.0 MPa,
temperature of 280 °C, 0.6 wt.% phenanthrene, and 0.1 wt.%
dimethyldisulfide (795 ppm of sulfur) in decalin. The weight
hourly space velocity of the feed solution was 14.0 hꢁ1, and the
H2/oil ratio was 1125 Nm3/m3. The products were collected and
analyzed using an Agilent 7890A GC installed with a FID detector.
The products were further analyzed with a mass spectrometer
(SHIMADZU GCMS-QP2010).
F ꢀ x
TOF ¼
;
NMo ꢀ fMo
where x is the phenanthrene conversion (%), NMo is the number of
Mo atoms in the catalyst (mol), and fMo is the MoS2 dispersion.
3. Results and discussion
3.1. Characterization
Fig. 1 shows the XRD patterns of the supports and catalysts. The
MZSM-5 sample exhibits typical peaks in the range of 5–50° asso-
ciated with the MFI structure. After impregnation of the MZSM-5
with nickel and molybdenum, followed by drying and sulfidation,
the intensity of the diffraction peaks in the samples is slightly
weakened, whereas the MFI structure in the samples is main-
tained. No other crystalline phases (MoO3 or b-NiMoO4) are
detected in the NiMo/MZSM-5 and NiMoS/MZSM-5 catalysts, indi-
cating that the Ni and Mo species are well dispersed in both dried
and sulfided catalysts. Notably, there is no change in the peak posi-
tion for the spent NiMoS/MZSM-5 catalyst, demonstrating that the
MFI structure in the catalyst is retained after hydrotreating.
To investigate the intrinsic hydrogenation activity of the cata-
lysts, the mass transfer and heat transfer effect can be ruled out
using the Weisz–Prater criterion and a Mears criterion [51]. The
experimental test used a desired amount presulfided catalyst
(60–80 mesh, containing 62.5 lmol Mo) diluted with 1.8 g SiC
(40–60 mesh) that was loaded into the reactor. After the reaction
was stable, the phenanthrene conversion was kept a low level by
changing the weight hourly space velocity. The internal and

426
W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
400
2.7 nm
0.6
0.4
0.2
0.0
b
d
c
b
a
300
200
100
16 nm
a
10
100
Pore Diameter (nm)
b
a
10
20
30
40
50
0.0
0.2
0.4
0.6
0.8
1.0
2 Theta (deg.)
Relative Pressure (P/P )
0
Fig. 1. XRD patterns of (a) MZSM-5, (b) NiMo/MZSM-5, (c) NiMoS/MZSM-5, and (d)
spent NiMoS/MZSM-5 samples.
Fig. 3. Nitrogen adsorption–desorption isotherms and pore size distributions of (a)
MZSM-5 and (b) c-Al2O3 in fashioned form. (The isotherm of c-Al2O3 has been offset
by 70 cm3 gꢁ1 along the vertical axis for clarity. The pore size distribution of
has been offset by 0.24 cm3 gꢁ1along the vertical axis for clarity.).
c-Al2O3
8.1 nm
0.6
400
b
0.4
b
22 nm
300
0.2
2.5 nm
2.1 nm
0.36
0.24
0.12
0.00
300
a
100
0.0
b
a
10
250
200
150
100
50
Pore Diameter (nm)
200
100
a
b
a
10
100
Pore Diameter (nm)
0.0
0.2
0.4
0.6
0.8
)
1.0
Relative Pressure (P/P0
Fig. 2. Nitrogen adsorption–desorption isotherms and pore size distributions of the
(a) MZSM-5 and (b) -Al2O3 samples in powder form. (The isotherm of -Al2O3 has
been offset by 60 cm3 gꢁ1 along the vertical axis for clarity. The pore size
c
c
0.0
0.2
0.4
0.6
0.8
1.0
Relative Pressure (P/P0 )
distribution of
clarity.).
c
-Al2O3 has been offset by 0.24 cm3 gꢁ1along the vertical axis for
Fig. 4. Nitrogen adsorption–desorption isotherms and pore size distributions of (a)
NiMoS/MZSM-5 and (b) NiMoS/ -Al2O3 catalysts. (The isotherm of NiMoS/ -Al2O3
has been offset by 50 cm3 gꢁ1 along the vertical axis for clarity. The pore size
distribution of NiMoS/
for clarity.).
c
c
Fig. 2 shows the N2 adsorption–desorption isotherms and pore
size distributions of the MZSM-5 and -Al2O3 samples in powder
form. The isotherms present a hysteresis loop, indicating that the
MZSM-5 and -Al2O3 samples contain a mesoporous structure.
c
-Al2O3 has been offset by 0.1 cm3 gꢁ1 along the vertical axis
c
c
Correspondingly, the pore sizes of the two samples are centered
at 22 and 8.1 nm. Because the catalyst was fashioned at a high
pressure of 15 MPa before activity testing, to compare the change
in the texture parameters of the support with the fashioned cata-
lyst in the same way, N2 adsorption of the fashioned supports
and catalysts was performed, and the results are shown in Figs. 3
and 4, respectively. Sample texture parameters of the supports
and corresponding catalysts are presented in Table 1. Clearly, when
the powder supports were transformed into fashioned forms, the
mesopore volume and pore size decreased due to the extruding
effect on the powder sample. After loading of metals and sulfida-
tion, the mesopores were still presented in the catalyst samples
(Fig. 4), but the BET surface area, mesopore volume, and pore size
dimensions decreased further (inset in Fig. 4 and Table 1) due to
the metal filling in the pores.
Table 1
Texture parameters of the supports and catalysts.
Sample
SBET
Vmicro
Vmeso
Smic
Sext
a
b
c
d
e
(m2 gꢁ1
)
(cm3 gꢁ1
)
(cm3 gꢁ1
)
(m2 gꢁ1
)
(m2 gꢁ1
)
MZSM-5f
360
322
315
0.08
–
0.10
0.08
–
0.10
0.05
–
0.30
0.45
0.06
0.21
0.35
0.06
0.14
0.25
0.04
185
–
245
192
–
250
127
16
175
322
67
155
303
59
109
213
40
f
c
-Al2O3
ZSM-5f
Tableted MZSM-5g 347
g
Tableted
c
-Al2O3
303
309
236
230
217
Tableted ZSM-5g
NiMoS/MZSM-5
NiMoS/c-Al2O3
NiMoS/ZSM-5
0.07
177
a
b
c
BET surface area.
The SEM image reveals that the MZSM-5 particles have an
Microporous pore volume, obtained from t-plot method.
Mesoporous pore volume, obtained from BJH adsorption cumulative volume of
‘‘olive’’ morphology with particle sizes of 0.5 ꢀ 1.5
lm (Fig. 5a).
pores between 1.7 and 300 nm diameter.
The TEM image of the thin-sectioned MZSM-5 gives direct evi-
dence for the presence of abundant hierarchical mesopores in the
crystals (Fig. 5b, light areas). The size range of the mesopores is
in good agreement with that obtained from N2 sorption (Fig. 2a,
insert).
d
Micropore surface area.
External surface area, obtained from t-plot method.
The sample was in powder form, and
e
f
c-Al2O3 was purchased from Shanghai
HENGYE Chemical Industry Co., Ltd.
g
The sample was tableted under a pressure of 15 MPa before N2 physisorption.

W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
427
Table 2
Total acidity and acidic site distribution of the NiMoS/MZSM-5 and NiMoS/
c-Al2O3
catalysts.
Catalyst
Acidic amounts of the catalysts (
120–180 °C 180–250 °C 250–350 °C >350 °C Total
l )
mol gꢁ1
NiMoS/MZSM-5 90
NiMoS/ -Al2O3 111
110
140
120
145
47
–
367
396
c
a
b
200
250
300
350
400
450
500
Wavelength (nm)
Fig. 7. UV–vis DR spectra of (a) NiMo/MZSM-5 and (b) NiMo/c-Al2O3 samples.
These results indicate that the area of the desorption profile in
the high-temperature region is not completely induced by the
adsorbed NH3. The acid–base titration results also demonstrate
that the strongly acidic sites on the NiMoS/MZSM-5 catalyst are
Fig. 5. (a) SEM image of the MZSM-5 sample and (b) TEM image of the thin-
sectioned MZSM-5 sample.
only 47
NiMoS/
l
mol gꢁ1, while there are no strongly acidic sites on
c
-Al2O3 (Table 2).
UV–vis analysis was performed to obtain the coordination state
of the Mo species on the dried catalysts (Fig. 7). Generally, the
absorption band at 230–260 nm is attributed to the isolated
molybdate species in tetrahedral coordination, whereas the band
at 270–330 nm is characteristic of polymeric octahedral Mo
species [31,53]. In our case, the absorption band around at
800
600
250 nm for NiMo/
NiMo/MZSM-5 are observed, which indicate that the Mo species
on -Al2O3 are mainly present in the form of tetrahedral Mo, while
the polymeric octahedral Mo species are prominent on MZSM-5.
These results suggest that the proportion of octahedral Mo species
c-Al2O3 and the band around at 300 nm for
400
c
b
200
a
on MZSM-5 is higher than that on c-Al2O3 support. The octahedral
Mo species could interact weakly with the MZSM-5 support.
This makes them easily reduced to more multistacked MoS2
active phases, enhancing the hydrogenation activity of the
NiMoS/MZSM-5.
0
50
100
Time (min)
150
200
Fig. 6. NH3-STPD profiles of (a) NiMoS/MZSM-5 and (b) NiMoS/c-Al2O3 catalysts.
The Raman analysis experiments were performed to further
investigate the differences in the state of the Mo species on the
MZSM-5 and c-Al2O3 supports. This could be beneficial for a better
Fig. 6 shows the NH3-STPD profiles of the NiMoS/MZSM-5 and
NiMoS/ -Al2O3 catalysts. Generally, the acidity of the catalyst is
understanding of the metal–support interaction. The band cen-
tered at approximately 950 cmꢁ1 for the two catalysts is ascribed
to the symmetric stretching of the Mo@O bond in the polymolyb-
c
classified as weak acidic (120–180 °C), medium weak acidic
(180–250 °C), medium strong acidic (250–350 °C), and strong
acidic (>350 °C) sites according to the desorption temperature of
the adsorbed ammonia [29]. The acid–base titration results are
presented in Table 2. It can be seen that the acidic site density
6ꢁ
4ꢁ
24
dates (Mo7O and Mo8O ) [54–56] (Fig. 8). The polymolybdates
24
are considered to interact weakly with the support, resulting in
higher reducibility and activity during the hydrogenation reaction
[54,56]. However, a shoulder band is observed at 895 cmꢁ1 on the
on the NiMoS/
NiMoS/MZSM-5 across all studied desorption temperatures
(Fig. 6). To further investigate the acidity of the NiMoS/ -Al2O3
c-Al2O3 sample is slightly higher than that on the
NiMo/c-Al2O3 that can be attributed to the Mo@O stretching mode
of the isolated MoO24ꢁ with distorted tetrahedral symmetry
[55,57]. This Mo species present relatively low reducibility during
the catalyst sulfidation due to the strong interaction of the Mo with
c
and NiMoS/MZSM-5 catalysts in the high desorption temperature
range of 350–800 °C, He-STPD of the two catalyst samples was per-
formed (Fig. S2 in the Supporting Information). Clearly, the profiles
of the two catalysts in the range of 350–800 °C are still present.
the
c-Al2O3 via MoAOAAl bridges [50,58]. In contrast, the band at
895 cmꢁ1 is not observed on the NiMo/MZSM-5 sample, while the

428
W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
370 °C
147
281
561
578
°C
818 861
335
378
948
192 °C
c
993
445
405
°C
°
C
a
b
°
C
515 °C
950
1046
690 °C
367 °C
214
895
210 °C
a
475
575 720
600
b
200
400
600
800
1000
200
400
800
1000
1200
Temperature (oC)
Raman Shift (cm-1)
Fig. 9. H2-TPR profiles of (a) Ni/MZSM-5, (b) Mo/MZSM-5, and (c) NiMo/MZSM-5
samples.
Fig. 8. Raman spectra of (a) NiMo/MZSM-5 and (b) NiMo/c-Al2O3 samples.
bands at 993, 818, 335, and 281 cmꢁ1 are detected for the
NiMo/MZSM-5 sample, which could be induced by ‘‘free’’ MoO3
crystallites composed of octahedral MoO6 [59]. It has been
reported that the ‘‘free’’ MoO3 crystallites can be formed on the
silica [60], which is due to the weaker interaction between the
metal and surface hydroxyl on the support [57,60–62]. In our case,
IR analysis shows that the acidic hydroxyl and silica hydroxyl
groups exist on the MZSM-5 [63] (Fig. S3 in the Supporting
Information), which results in relatively weak metal–support
interaction, leading to formation of easily reducible ‘‘free’’ MoO3
and polymolybdates in the NiMo/MZSM-5 catalyst.
450 °C
625 °C
800 °C
325 °C
c
371 °C
584 °C
b
600 °C
340 °C
a
To investigate the reduction behavior of Mo and Ni species in
the catalyst precursors, TPR of the Ni, Mo, and NiMo catalysts sup-
ported on MZSM-5 and c-Al2O3 was carried out, and the results are
200
400
600
800
1000
Temperature (o C)
shown in Figs. 9 and 10. For the Ni/MZSM-5 sample (Fig. 9a), the
overlapping profiles in the temperature range of 367–690 °C can
be assigned to the reduction of Ni2+ and Ni octahedral species
(Ni2oc+t) [64,65]. The very small profiles with peak at 210 °C may
be associated with the superficial reduction of nickel oxide [66].
For the Mo/MZSM-5 sample (Fig. 9b), the reduction profile with
peak at 445 °C is assigned to the reduction of the octahedral Mo
Fig. 10. H2-TPR profiles of (a) Ni/c-Al2O3, (b) Mo/c-Al2O3, and (c) NiMo/c-Al2O3
samples.
slabs; and the Type II phase is fully sulfurized, has higher S coordi-
nation of Mo and Ni, and has less dispersion of the MoS2, which are
mostly multilayer slabs. In addition, the Type II phase has higher
intrinsic activity than the Type I phase in the hydrogenation reac-
tion [50,54,69]. Representative TEM images of the NiMoS/MZSM-5
species (Mo6+ + 2eꢁ ? Mo4+
MZSM-5 support; the profile with peak at 578 °C is assigned to
the further reduction of the octahedral Mo species (Mo4+ + 4eꢁ
) in weak interaction with the
?
Mo0) [53,67]. Compared with the reduction behavior of the
Mo/MZSM-5 and Ni/MZSM-5 samples, the profile with a peak at
370 °C should involve the reduction of octahedral Mo and Ni2+ spe-
cies, and the profile with a peak at 561 °C should be attributed to
the further reduction of the octahedral Mo species and Ni2oc+t species
for the NiMo/MZSM-5 sample (Fig. 9c) [53,64,65]. Similarly, the
reduction profile in the relatively low temperature range of 325–
and NiMoS/
the less stacked MoS2 crystallites with shorter length (black
fringes) are highly dispersed on the surface of the -Al2O3, while
c-Al2O3 catalysts are shown in Fig. 11. It is clear that
c
the relatively multistacked MoS2 crystallites with longer length
are primarily located in the mesopore channels (Fig. 11a and
insert) of the MZSM-5. The weak metal–MZSM-5 support interac-
tion, discussed earlier, leads to the formation of MoS2 crystallites
with high stacking number and extended average length during
sufidation [70]. The length and stacked layer number of the MoS2
crystallites in the two sulfided catalysts was further statistically
analyzed, and the results are shown in Fig. 12 and Table 3. The
length and stacking of the MoS2 slabs on the NiMoS/MZSM-5 cat-
alyst are mainly 4–6 nm and 2–4 layers, respectively, greater than
625 °C is also observed on the NiMo/c-Al2O3 sample, which should
be attributed to the reduction of the octahedral Mo species and all
Ni2+ species (Fig. 10c). Additionally, the reduction profile of the Mo
species at 325–450 °C and 450–625 °C could be attributed to the
reduction of the octahedral Mo species in well-dispersed poly-
molybdates and in different degrees of polymolybdates [53,67].
Meanwhile, a broad profile centered at 800 °C represents the deep
reduction of all Mo4+ species and the reduction of isolated tetrahe-
those on the NiMoS/c-Al2O3 catalyst (2–4 nm, 1–2 layers; Fig. 12).
The average stacking and length of the MoS2 slabs on the
dral Mo species in strong interaction with the
These results indicate that the Mo and Ni species on MZSM-5 are
more easily reduced than those on -Al2O3.
To obtain information on metal sulfide active phases on the sur-
faces of MZSM-5 and -Al2O3, TEM was carried out. Generally, the
NiMoS2 active phases in metal sulfide catalysts can be subdivided
into NiMoS2-I (Type I) and NiMoS2-II (Type II) [50,54,68]. The typ-
ical properties of the Type I phase are lower S coordination of Mo
and Ni, and higher dispersion of the MoS2, which are mostly single
c-Al2O3 support [67].
NiMoS/MZSM-5 are 3.3 and 5.9 nm, greater than those on
NiMoS/c-Al2O3 (1.8 and 3.5 nm; Table 3). The MoS2 slabs with
more stacks and greater length would hamper the dispersion of
the MoS2 active phase [56]. Thus, the MoS2 dispersion over
c
c
NiMoS/
Table 3). Compared with NiMoS/MZSM-5, the higher MoS2 disper-
sion on NiMoS/ -Al2O3 could be due to a change in the support
c-Al2O3 (0.29) is higher than over NiMoS/MZSM-5 (0.16;
c
properties. The silanol groups and acidic hydroxyl groups on
MZSM-5 could interact weakly with the Mo species to benefit

W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
429
40
NiMoS/γ-Al2O3
a
NiMoS/MZSM-5
30
20
10
0
3
4
6
>8
0
1
2
7
5
Length (nm)
NiMoS/γ-Al2O3
NiMoS/MZSM-5
50
40
30
20
10
0
b
0
1
2
3
4
5
6
7
Number of Layers
Fig. 11. TEM images of (a) thin-sectioned NiMoS/MZSM-5 and (b) NiMoS/c-Al2O3
catalysts.
Fig. 12. Length (a) and layer stacking (b) distributions of MoS2 slabs in NiMoS/
MZSM-5 and NiMoS/ -Al2O3 catalysts.
c
the formation of polymolybdates and ‘‘free’’ MoO3 species. They
are easily transformed into multistacked MoS2 active phases dur-
ing sulifidation, resulting in lower fMo on NiMoS/MZSM-5 [54,56].
be attributed to Ni2+ oxide species (including NiAl2O4 and NiOx),
NiMoS phase, and Ni sulfides (NiSx) [71], respectively. The relative
quantities of the Ni species (NiMoS, NiSx, and Ni2+oxides) are
shown in Table 4. The proportion of the NiMoS phase on the
NiMoS/MZSM-5 catalyst (50.8%) is higher than that on the
In contrast, the Mo species strongly interact with
c-Al2O3 via
MoAOAAl bridges, leading to the Mo species being difficult to sul-
fide into the MoS2 active phase [50], and obtaining a high fMo on
NiMoS/c-Al2O3 catalyst (25.6%), indicating that using mesoporous
NiMoS/
phases are more easily formed on the surface of the MZSM-5 rela-
tive to -Al2O3.
c
-Al2O3. Therefore, the multistacked MoS2 (Type II) active
zeolites as a support increases the proportion of Ni–Mo interaction
and leads to the formation of more NiMoS phases. Notably, the pro-
portion of Mo4+ and NiMoS phase on the spent NiMoS/MZSM-5 cat-
c
XPS experiments over sulfided catalysts were also performed to
study the sulfidation degree of Mo and Ni species. The decomposi-
tion of the Mo3d spectra is shown in Fig. 13a. Molybdenum can
alyst is also higher than on the spent NiMoS/c-Al2O3 catalyst
(Fig. S4 in the Supporting Information and Table 4). Nevertheless,
the Ni content in the NiMoS phase on the NiMoS/MZSM-5 catalyst
decreases after catalyst testing in phenanthrene hydrogenation.
This is because the Ni species segregated from the MoS2 edges
and then aggregated into Ni2+ clusters. Such a phenomenon was
also reported by Guichard and co-workers [72], and was explained
in more detail by the dynamic model proposed by Nikulshin and
colleagues [73,74]. In this model, the sulfur and promoters (Ni or
Co) repeatedly migrated between the adjacent multilayered MoS2
crystallites in the course of reduction–sulfidation processes under
hydrogen, and the coke formed on the MoS2 edge can block the
promoters in the structure of neighboring layers. This led to the
aggregation of promoter atoms. In our case, the TG analysis data
show that the weight loss occurred on the spent NiMoS/MZSM-5
catalyst (Fig. S5 in the Supporting Information), which could be
induced by coke formation. Coke formation could hinder the
migration of the Ni atoms between the adjacent multilayered
MoS2 slabs and form separated Ni2+ clusters.
exist as disulfide MoS2 (Mo4+
)
located at 229.0 0.1 eV
(Mo4+3d5/2) and 232.1 0.1 eV (Mo4+3d3/2), as a Mo oxide phase
(MoOx, Mo6+
)
located at 232.2 0.1 eV (Mo6+3d5/2
)
and
235.3 0.1 eV (Mo6+3d3/2), and as an intermediate state Mo oxysul-
fide (MoOxSy, Mo5+) at 230.2 0.1 eV (Mo5+3d5/2) and 233.4 0.1 eV
(Mo5+3d3/2) [71]. The higher atomic percentage of Mo4+ obtained for
the NiMoS/MZSM-5 catalyst (Table 4) demonstrates that the sulfi-
dation degree of the Mo species, defined as the ratio of Mo4+ to
the sum of Mo4+
,
Mo5+
,
and Mo6+ [21], is higher on
-Al2O3. In fact, the Mo4+ content
NiMoS/MZSM-5 than on NiMoS/
c
depends inversely on the metal–support interaction [21]. The
higher Mo4+ content of NiMoS/MZSM-5 could be attributed to
the weaker molybdenum–support interaction [21], which favors the
formation of octahedral Mo species on the MZSM-5, in agreement
with the UV–vis and Raman results. The octahedral Mo species
are easy to sulfide, leading to the formation of more MoS2 active
phases responsible for higher hydrogenation activity [54,56].
The Ni2p XPS spectra of the catalysts are shown in Fig. 13b. The
binding energy of the Ni2p peaks at 856.1, 853.7, and 853.1 eV can
The atomic ratio of Mo/(Al + (Si)) in NiMoS/MZSM-5 is slightly
lower than over NiMoS/c-Al2O3 (Table S1), indicating that
relatively larger Mo species with lower dispersion are formed on

430
W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
MZSM-5. There is an obvious decrease in the Ni/(Ni + Mo) atomic
ratio on the NiMoS/MZSM-5 catalyst due to more effective promo-
tion of Ni species to MoS2. Furthermore, the atomic ratio of S/Mo is
that of NiMoS/c-Al2O3. The higher sulfidation degree can be attrib-
uted to the relatively weak interaction between molybdenum spe-
cies and MZSM-5 support.
much higher on NiMoS/MZSM-5 than on NiMoS/c-Al2O3, suggest-
ing that the sulfidation degree of NiMoS/MZSM-5 is higher than
3.2. Phenanthrene hydrogenation
Phenanthrene hydrogenation is a consecutive reaction that first
creates dihydrophenanthrene (DHP) and 1,2,3,4-tetrahydro-
phenanthrene (THP), followed by the formation of symmetric
and asymmetric octahydrophenanthrene (OHP) as well as
perhydrophenanthrene (PHP) [75–77]. The reaction network of
phenanthrene hydrogenation is shown in Fig. S6 in the Supporting
Information. Fig. 14a shows the time dependence of phenanthrene
conversion over various catalysts in the presence of 795 ppm sulfur
in the form of dimethyldisulfide. Obviously, phenanthrene conver-
sion over NiMoS/MZSM-5 catalyst is much higher than over
a
Mo4+3d5/2
Mo4+3d3/2
NiMoS/γ-Al2O3
Mo6+3d5/2
Mo5+3d5/2
Mo5+3d3/2
Mo6+3d3/2
S2s
Mo4+3d5/2
Mo4+3d3/2
Mo6+3d5/2
NiMoS/MZSM-5
NiMoS/ZSM-5, NiMoS/c-Al2O3, and Pd/MZSM-5 catalysts. Gener-
Mo5+3d5/2
Mo4+3d3/2
Mo6+3d3/2
ally, the catalytic performance of supported metal sulfide catalysts
is influenced by the support parameters, including the surface area,
pore structure, and acidity. NiMoS/MZSM-5 has higher external
surface area (109 m2 gꢁ1) and mesopore volume (0.14 cm3 gꢁ1),
whereas NiMoS/ZSM-5 has lower external surface area
(40 m2 gꢁ1, Table 1). Because the hydrogenation of the bulky
phenanthrene molecule mainly occurs on the external surface
and in the mesopores, the large mesopores in the
NiMoS/MZSM-5 catalyst could favor mass transfer and result in
higher phenanthrene conversion than on NiMoS/ZSM-5. In
addition, the crystal MoS2 phases could only form on the external
surface (only 67 m2 gꢁ1) of the ZSM-5, and not in the micropores
(0.54 ꢀ 0.56 nm), due to the relatively large crystal dimension of
S2s
240
237
234
231
228
225
222
Binding Energy (eV)
b
Ni2+
shake up of Ni(II)
NiMoS/γ-Al2O3
NiMoS
NiSx
NiMoS
NiSx
100
Ni2+
shake up of Ni(II)
NiMoS/MZSM-5
880
a
80
60
40
20
0
890
870
860
850
Binding Energy (eV)
Fig. 13. Decomposition of XPS Mo3d (a) and Ni2p (b) spectra of NiMoS/MZSM-5 and
NiMoS/ -Al2O3 catalysts.
c
Table 3
Average length and number of the stacking layers of MoS2 crystallites.
0
10
20
30
40
50
60
Time on Stream (h)
Catalyst
Average length (nm)
Average stacking
fMo
(number of layers)
100
80
60
40
20
0
NiMoS/MZSM-5
NiMoS/c-Al2O3
NiMoS/MZSM-5-I
NiMoS/MZSM-5-L
5.9
3.5
4.6
3.7
3.3
1.8
3.0
2.4
0.16
0.29
0.20
0.22
b
Table 4
XPS results of the different contributions Mo3d and Ni2p3/2 obtained for sulfided
catalysts.
Catalyst
Mo distribution (at.%)
Ni distribution (at.%)
Mo4+
Mo5+
Mo6+
NiMoS
NiSx
Ni2+
NiMoS/MZSM-5
81.1
74.8
73.5
68.9
78.9
78.5
45.4
10.5
8.8
7.4
9.0
9.1
8.5
16.4
19.1
22.1
12.0
10.0
5.8
50.8
25.6
39.7
25.1
57.8
66.2
45.9
24.4
16.2
15.4
20.1
14.5
9.3
24.8
58.2
44.9
54.7
27.7
24.5
30.7
NiMoS/
Spent NiMoS/MZSM-5
Spent NiMoS/ -Al2O3
c-Al2O3
0
10
20
30
40
50
60
Time on stream (h)
c
NiMoS/MZSM-5-I
NiMoS/MZSM-5-L
NiMoS/ZSM-5
Fig. 14. Dependence of the (a) phenanthrene conversion and (b) DHPS selectivity
on reaction time over (j) NiMoS/MZSM-5, ( ) NiMoS/ -Al2O3, ( ) Pd/MZSM-5, and
) NiMoS/ZSM-5 catalysts.
11.5
48.8
c
23.4
(

W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
431
MoS2 phases [54]. In contrast, MZSM-5 provides a large external
surface area (175 m2 gꢁ1), which could benefit the dispersion of
the metal precursor and lead to the formation of more MoS2 phases
in the mesopore channels. This is supported by XPS analysis data
(Table 4). The percentages of the Mo4+ species and NiMoS on the
NiMoS/MZSM-5 (81.1% and 50.8%) are much higher than those
on NiMoS/ZSM-5 (45.4% and 45.9%). These results indicate that
more MoS2 phases could be formed in the mesopores and provide
more accessible active sites, further enhancing the catalytic activ-
ity of NiMoS/MZSM-5.
Although noble metal catalysts have higher hydrogenation
activity than metal sulfide catalysts [2,4], the phenanthrene con-
version over Pd/MZSM-5 is much lower than that over
NiMoS/MZSM-5 under the same reaction conditions. This is
because the Pd/MZSM-5 catalyst is easily poisoned by sulfur com-
pounds, resulting in the deactivation of the catalyst [2,16]. As
observed in the initial reaction stage, the phenanthrene conversion
on Pd/MZSM-5 is rapidly reduced (Fig. 14a). In contrast, the
NiMoS/MZSM-5 catalyst shows high initial activity and the conver-
sion is not decreased with prolonged reaction time (Fig. 14a).
has a hydrogenation capability superior to that of the conventional
NiMoS/ -Al2O3 catalyst. The higher activity of the NiMoS/MZSM-5
catalyst could be assigned to the difference in surface properties of
the -Al2O3 and MZSM-5 supports. Acidic hydroxyl and silica
hydroxyl groups existed on the MZSM-5, which interacted weakly
with the Mo species to benefit the formation of polymolybdates
and ‘‘free’’ MoO3 species on the MZSM-5 during the catalyst pre-
treatment process. After sulfidation, the polymolybdates and ‘‘free’’
MoO3 species can easily be transformed into more multistacked
MoS2 active phases that favor the adsorption of aromatics and
therefore enhance the hydrogenation activity of the catalyst [69].
c
c
In contrast, basic hydroxyl groups existed on the c-Al2O3, which
interacted strongly with the Mo species, facilitating the formation
of isolated MoO24ꢁ species that are difficult to sulfide into the MoS2
active phases [50,58]. The details were discussed in the
characterizations.
It is important that the NiMoS/MZSM-5 catalyst shows a long
catalyst life (Fig. S7 in the Supporting Information). There is no
activity loss in phenanthrene hydrogenation. The good catalyst life
of the NiMoS/MZSM-5 catalyst is one of the key factors for practical
application in industry. It is worth mentioning that the
NiMoS/MZSM-5 catalyst has relatively low acidity, slightly lower
Conventional NiMoS/c-Al2O3 catalyst has higher external surface
area and mesopore volume (213 m2 gꢁ1and 0.25 m3 gꢁ1) than
NiMoS/MZSM-5 (109 m2 gꢁ1 and 0.14 m3 gꢁ1), but it is still difficult
to obtain higher phenanthrene conversion. For example, the con-
than that of the NiMoS/
compared with NiMoS/
c
-Al2O3 catalyst (Fig. 6 and Table 2). Thus,
c-Al2O3 catalyst, the higher activity of the
version over NiMoS/
c
-Al2O3 (80.5%) catalyst is much lower than that
NiMoS/MZSM-5 catalyst is not caused by the acidity. Not only the
NiMoS/MZSM-5 but also the CoMoS/MZSM-5 catalyst has excellent
deep hydrogenation activity in phenanthrene hydrogenation
(Fig. S8 in the Supporting Information). The phenanthrene conver-
sion and DHPS selectivity over CoMoS/MZSM-5 catalyst (88.1%
over NiMoS/MZSM-5 (98.7%) catalyst at a reaction time of 40 h
(Table 5). To further compare the capability of deep hydrogenation
of phenanthrene over various catalysts, the dependence of
the sum of OHP and PHP (products of deep hydrogenation of
phenanthrene, DHPS) with reaction time is shown in Fig. 14b. Very
interestingly, the DHPS selectivity over NiMoS/MZSM-5 is much
and 95.4%) are higher than those over CoMoS/c-Al2O3 (80.8% and
47.0%) catalyst (Table S2). In addition, the mesostructured
mordenite-supported NiMo and CoMo catalysts also show high
conversion and deep hydrogenation ability in phenanthrene hydro-
higher than over NiMoS/
over NiMoS/MZSM-5 is 98.6%, and those over NiMoS/
c-Al2O3. For instance, the DHPS selectivity
c-Al2O3 and
Pd/MZSM-5 are 51.6% and 3.2% (Table 5), respectively. These
results indicate that the mesoporous zeolites are better supports
for preparing metal sulfide catalysts with high hydrogenation
activity. In addition, compared with NiMoS/MZSM-5 catalyst, the
genation in comparison with c-Al2O3 supported ones (Fig. S8).
3.3. Relationship between morphology of active phase and catalytic
performance
activity of NiMoS/c-Al2O3 decreased rapidly in the initial stage.
In general, the deactivation of the metal sulfide catalyst is mainly
induced by coke formation and the reduction of the active phases
[72]. In the present work, the TG analysis shows that the weight
Generally, the stacking and slab length of the MoS2 crystallites
strongly influence the intrinsic activity of the catalyst [56,65,78].
Because the stacking and slab length of the MoS2 crystallites could
be controlled by varying the Mo loadings on the support, it is pos-
sible to establish the relationship between the slab length and
stacking of the MoS2 and hydrogenation activity by using different
Mo loading catalysts. Fig. S9 in the Supporting Information shows
representative TEM images of MZSM-5-supported NiMo catalysts
with different metal loadings (3.0, 6.0, and 12.0 wt.% Mo).
Fig. S10 in the Supporting Information shows the distribution of
the lengths and stacking of MoS2 slabs on the catalysts. The length
and stacking of the MoS2 slabs clearly increase with Mo loading
(Figs. S9 and S10). The average length and stacking of the MoS2
slabs also increase with Mo loading (Table 3).
loss of the spent NiMoS/c-Al2O3 catalyst is significant compared
with that of the spent NiMoS/MZSM-5 catalyst (Fig. S5 in the
Supporting Information). The coke hydrocarbons might adsorb
onto the active sites and lead to deactivation of the catalysts [72].
In the absence of mass and heat transfer effects, the observed
reaction rate (robs) and TOF of phenanthrene hydrogenation over
catalysts are presented in Table 6. The robs and TOF for
NiMoS/MZSM-5 (7.4 ꢀ 10ꢁ4 mol kgꢁ1 sꢁ1 and 15.0 ꢀ 10ꢁ4 sꢁ1) cat-
alyst are much higher than for NiMoS/c-Al2O3 catalyst
(4.8 ꢀ 10ꢁ4 mol kgꢁ1 sꢁ1 and 8.9 ꢀ 10ꢁ4 sꢁ1). In addition, the selec-
tivity of the deep hydrogenation product (OHP) over
NiMoS/MZSM-5 is much higher than that over NiMoS/
(Table 6). These results suggest that the NiMoS/MZSM-5 catalyst
c
-Al2O3
From Tables 3 and 6, it is clear that the TOF value increases with a
decrease in the average length and average stacking of the MoS2
phase on the MZSM-5 support. This result suggests that the intrinsic
activities of the metal sulfide catalysts could correlate with the
morphology (length and stacking number) of MoS2 crystallites.
Nikulshin and co-workers also reported that the hydrogenation
Table 5
Product selectivity in phenanthrene hydrogenation over various catalysts at reaction
time of 40 h.
activity of the NiW and Co(Ni)Mo catalysts supported on c-Al2O3
Run Catalyst
PHE conv. (%) Product selectivity (%)
DHP THP OHP PHP DHPSa
1.4 80.7 17.9 98.6
increased with the decrease in slab length and Ni/W(Mo)edge ratio of
the active phase [65,78]. In addition, the MoS2 particles with shorter
slab length and relatively less stacking could provide more exposure
of the edge of MoS2 slabs, leading to an increase in the available sites
for Ni promotion [65]. Therefore, compared with NiMoS/MZSM-5
and NiMoS/MZSM-5-I catalysts, the NiMoS/MZSM-5-L catalyst pre-
sents the highest dispersion of MoS2 particles with short slab length
1
2
3
4
NiMoS/MZSM-5 98.7
–
NiMoS/
c
-Al2O3
80.5
21.2
38.3
34.6 13.8 51.6
51.4 40.5 8.1
77.8 19.0 3.2
–
–
–
51.6
8.1
3.2
NiMoS/ZSM-5
Pd/MZSM-5
a
The sum of OHP and PHP (deep hydrogenation products, DHPS).

432
W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
Table 6
Phenanthrene hydrogenation performance over various catalysts.
b
robs ꢀ 104 (mol kgꢁ1 sꢁ1
)
TOF ꢀ 104 (sꢁ1
)
Product selectivityd (%)
ro0 bs ꢀ 104 (mol kgꢁ1 sꢁ1
)
TOFMo4þ ꢀ 104 (sꢁ1
)
a
c
Catalyst
DHP
THP
OHP
PHP
NiMoS/MZSM-5
NiMoS/c-Al2O3
NiMoS/MZSM-5-I
NiMoS/MZSM-5-L
7.4
4.8
–
6.7
4.3
7.1
7.7
15.0
8.9
18
18.0
12.0
–
31.3
38.7
31.9
34.5
47.7
46.1
48
20.9
15.2
20.1
18.5
–
–
–
–
–
20
–
47
a
b
c
dx
The observed reaction rate calculated per Mo atom in the sulfide catalyst, ro0 bs
¼
.
dðNMo =FÞ
The number of reacted PHE molecules per second and per Mo atom at the edge surface.
TOF transformed per mol of Mo4+ sites determined from XPS, TOFMo4þ
¼
.
Fꢀx
NMo4þ ꢀfMo
d
Determined at about 20% of total phenanthrene conversion by changing the weight hourly space velocity.
[10] J.N. Díaz de León, T.A. Zepeda, G. Alonso-Nuñez, D.H. Galván, B. Pawelec, S.
Fuentes, J. Catal. 321 (2015) 51.
[11] Y.-K. Lee, S.T. Oyama, J. Catal. 239 (2006) 376.
and less stacking, which favors the formation of more Ni-promoted
MoS2 active sites and NiMoS2 Type II phases. XPS analysis also
shows that the Ni content in the NiMoS phase is 66.2, 57.8,
and 50.8% in the NiMoS/MZSM-5-L, NiMoS/MZSM-5-I, and
NiMoS/MZSM-5 catalysts (Table 4). Based on these results, it can
be concluded that the catalytic activity depends on the number of
NiMoS phases as well as the shape of the active phase crystallites.
In addition, although the highest TOF number is obtained over the
NiMoS/MZSM-5-L catalyst with the lowest metal loading, an opti-
mal balance could be achieved between the catalytic activity and
catalyst loading in the feedstock.
[12] A. Corma, A. Martínez, V. Martínez-Soria, J. Catal. 169 (1997) 480.
[13] T.D. Tang, C.Y. Yin, L.F. Wang, Y.Y. Ji, F.-S. Xiao, J. Catal. 249 (2007) 111.
[14] X. Li, F. Zhou, A.J. Wang, L.Y. Wang, Y. Wang, Energy Fuels 26 (2012) 4671.
[15] C.S. Song, Catal. Today 86 (2003) 211.
[16] T.D. Tang, C.Y. Yin, L.F. Wang, Y.Y. Ji, F.-S. Xiao, J. Catal. 257 (2008) 125.
[17] T. He, Y.X. Wang, P.J. Miao, J.Q. Li, J.H. Wu, Y.M. Fang, Fuel 106 (2013) 365.
[18] J.G. Téllez-Romero, R. Cuevas-García, J. Ramírez, P. Castillo-Villalón, R.
Contreras-Bárbara, M.C. Salcedo-Luna, R.I. Puente-Lee, Catal. Today 250
(2015) 12.
[19] D. Li, A. Nishijima, D.E. Morris, G.D. Guthrie, J. Catal. 188 (1999) 111.
[20] M.Y. Sun, D. Nicosia, R. Prins, Catal. Today 86 (2003) 173.
[21] Y. Fan, H. Xiao, G. Shi, H.Y. Liu, Y. Qian, T.H. Wang, G.B. Gong, X.J. Bao, J. Catal.
279 (2011) 27.
[22] W.B. Chen, F. Maugé, J. van Gestel, H. Nie, D.D. Li, X.Y. Long, J. Catal. 304 (2013)
47.
4. Conclusions
[23] F. Rashidi, T. Sasaki, A.M. Rashidi, A.N. Kharat, K.J. Jozani, J. Catal. 299 (2013)
321.
[24] J. Ren, A.J. Wang, J. Tan, G. Cao, C. Liu, Y. Li, M. Lu, Y. Hu, Stud. Surf. Sci. Catal.
165 (2007) 721.
Mesoporous ZSM-5 was successfully synthesized using a water–
glass-containing cationic copolymer with quaternary ammonium
groups as a template, and applied for preparation of highly active
[25] S.T. Oyama, Y.-K. Lee, J. Catal. 258 (2008) 393.
metal sulfide catalysts. Compared with
c-Al2O3-supported NiMo
[26] Y. Li, D.H. Pan, C.Z. Yu, Y. Fan, X.J. Bao, J. Catal. 286 (2012) 124.
[27] Y.Y. Shu, Y.-K. Lee, S.T. Oyama, J. Catal. 236 (2005) 112.
[28] Y.Y. Sun, R. Prins, Angew. Chem. Int. Ed. 47 (2008) 8478.
[29] L. Zhang, W.Q. Fu, Q.P. Ke, S. Zhang, H.J. Jin, J.B. Hu, S. Wang, T.D. Tang, Appl.
Catal. A 573–434 (2012) 251.
(or CoMo) catalysts, the MZSM-5-supported NiMo (or CoMo) cata-
lysts show very high activity in the deep hydrogenation of phenan-
threne. This feature is attributed to the fact that more multistacked
MoS2 active phases are formed on the MZSM-5 due to the weaker
[30] A. Corma, Chem. Rev. 97 (1997) 2373.
[31] L. Peña, D. Valencia, T. Klimova, Appl. Catal. B 147 (2014) 879.
[32] A. Corma, M.J. Díaz-Cabañas, J. Martínez-Triguero, F. Rey, J. Rius, Nature 418
(2002) 514.
metal–support interaction relative to
c-Al2O3. In addition, the
catalytic activity depends on the number of NiMoS phases as well
[33] D. Li, H.F. Xu, G.D. Guthrie Jr., J. Catal. 189 (2000) 281.
[34] P. Leyrit, T. Cseri, N. Marchal, J. Lynch, S. Kasztelan, Catal. Today 65 (2001) 249.
[35] S. van Donk, A.H. Janssen, J.H. Bitter, K.P. de Jong, Catal. Rev. 45 (2003) 297.
as the shape of the active phase crystallites.
ˇ
Acknowledgements
[36] J. Cejka, G. Centi, J. Perez-Pariente, W.J. Roth, Catal. Today 179 (2012) 2.
[37] H.D. Wu, A.J. Duan, Z. Zhao, T.S. Li, R. Prins, X.F. Zhou, J. Catal. 317 (2014) 303.
[38] F.-S. Xiao, L.F. Wang, C.Y. Yin, K.F. Lin, Y. Di, J.X. Li, R.R. Xu, D.S. Su, R. Schlögl, T.
Yokoi, T. Tatsumi, Angew. Chem. Int. Ed. 45 (2006) 3090.
[39] K. Na, C. Jo, J. Kim, K. Cho, J. Jung, Y. Seo, R.J. Messinger, B.F. Chmelka, R. Ryoo,
Science 333 (2011) 328.
[40] W.Q. Fu, L. Zhang, T.D. Tang, Q.P. Ke, S. Wang, J.B. Hu, G.Y. Fang, J.X. Li, F.-S.
Xiao, J. Am. Chem. Soc. 133 (2011) 15346.
[41] Z.X. Qin, B.J. Shen, X.H. Gao, F. Lin, B.J. Wang, C.M. Xu, J. Catal. 278 (2011) 266.
This work was supported by the National Science Foundation of
China (21376050 and U1463203) and the Natural Science Founda-
tion of Zhejiang Province of China (Z14B060012). The authors are
grateful to Prof. Dr. S. Ted Oyama for his kind help and guidance
on this research. We also thank the anonymous reviewers for their
invaluable advice.
ˇ
´
[42] J. Zecevic, C.J. Gommes, H. Friedrich, P.E. de Jongh, K.P. de Jong, Angew. Chem.
Int. Ed. 51 (2012) 4213.
[43] Z.X. Qin, B.J. Shen, Z.W. Yu, F. Deng, L. Zhao, S.G. Zhou, D.L. Yuan, X.H. Gao, B.J.
Wang, H.J. Zhao, H.J. Liu, J. Catal. 298 (2013) 102.
[44] Z.Z. Kang, G.Y. Fang, Q.P. Ke, J.B. Hu, T.D. Tang, ChemCatChem 5 (2013) 2191.
[45] Y.L. Ding, Q.P. Ke, T.T. Liu, W.C. Wang, M.Y. He, K.Q. Yang, H.L. Jin, S. Wang, T.D.
Tang, Ind. Eng. Chem. Res. 53 (2014) 13903.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2015.07.026.
[46] W.Q. Fu, T.T. Liu, Z.X. Fang, Y.L. Ma, X. Zheng, W.C. Wang, X.J. Ni, M.L. Hu, T.D.
Tang, Chem. Commun. 51 (2015) 5890.
[47] T.D. Tang, L. Zhang, W.Q. Fu, Y.L. Ma, J. Xu, J. Jiang, G.Y. Fang, F.-S. Xiao, J. Am.
Chem. Soc. 135 (2013) 11437.
[48] G. Berhault, M.P. De la Rosa, A. Mehta, M.J. Yácaman, R.R. Chianelli, Appl. Catal.
A 345 (2008) 80.
[49] H. Wang, Y. Fan, G. Shi, H.Y. Liu, X.J. Bao, J. Catal. 260 (2008) 119.
[50] W.K. Lai, W.J. Song, L.Q. Pang, Z.F. Wu, N. Zheng, J.J. Li, J.B. Zheng, X.D. Yi, W.Q.
Fang, J. Catal. 303 (2013) 80.
[51] H.S. Fogler, Elements of Chemical Reaction Engineering, third ed., Michigan,
2004.
References
[1] A. Stanislaus, B.H. Copper, Catal. Rev. Sci. Eng. 36 (1994) 75.
[2] C.S. Song, X.L. Ma, Appl. Catal. B 41 (2003) 207.
[3] H.J. Zhang, X.C. Meng, Y.D. Li, Y.S. Lin, Ind. Eng. Chem. Res. 46 (2007) 4186.
[4] H.J. Kim, C.S. Song, Energy Fuels 28 (2014) 6788.
[5] M. Gupta, J. He, T. Nguyen, F. Petzold, D. Fonseca, J. Jasinski, M.K. Sunkara, Appl.
Catal. B. 180 (2016) 246.
[52] G.Q. Wu, Y.Q. Wang, L.N. Wang, W.P. Feng, H.N. Shi, Y. Lin, T. Zhang, X. Jin, S.H.
Wang, X.X. Wu, P.X. Yao, Chem. Eng. J. 215–216 (2013) 306.
[53] T.E. Klimova, D. Valencia, J.A. Mendoza-Nieto, P. Hernández-Hipólito, J. Catal.
304 (2013) 29.
[54] S. Badoga, K.C. Moulia, K.K. Soni, A.K. Dalaia, J. Adjaye, Appl. Catal. B 125 (2012)
67.
[6] M. Egorova, R. Prins, J. Catal. 224 (2004) 278.
[7] S.K. Bej, S.K. Maity, U.T. Turaga, Energy Fuels 18 (2004) 1227.
[8] D. Laurenti, Bo Phung-Ngoc, C. Roukoss, E. Devers, K. Marchand, L. Massin, L.
Lemaitre, C. Legens, A.-A. Quoineaud, M. Vrinat, J. Catal. 297 (2013) 165.
[9] J.C. Mogica-Betancourt, A. López-Benítez, J.R. Montiel-López, L. Massin, M.
Aouine, M. Vrinat, G. Berhault, A. Guevara-Lara, J. Catal. 313 (2014) 9.

W. Fu et al. / Journal of Catalysis 330 (2015) 423–433
433
[55] J.A. Mendoza-Nieto, O. Vera-Vallejo, L. Escobar-Alarcón, D. Solís-Casados, T.
[67] K. Al-Dalama, A. Stanislaus, Thermochim. Acta 520 (2011) 67.
Klimova, Fuel 110 (2013) 268.
[68] H. Topsøe, B.S. Clausen, Appl. Catal. 25 (1986) 273.
[56] P.A. Nikulshin, D.I. Ishutenko, A.A. Mozhaev, K.I. Maslakov, A.A. Pimerzin, J.
Catal. 312 (2014) 152.
[69] E.J.M. Hensen, P.J. Kooyman, Y. van der Meer, A.M. van der Kraan, V.H.J. de
Beer, J.A.R. van Veen, R.A. van Santen, J. Catal. 199 (2001) 224.
[70] P.A. Nikulshin, V.A. Salnikov, A.V. Mozhaev, P.P. Minaev, V.M. Kogan, A.A.
Pimerzin, J. Catal. 309 (2014) 386.
[71] T.K.T. Ninh, L. Massin, D. Laurenti, M. Vrinat, Appl. Catal. A 407 (2011) 29.
[72] B. Guichard, M. Roy-Auberger, E. Devers, C. Legens, P. Raybaud, Catal. Today
130 (2008) 97.
[57] H.C. Hu, I.E. Wachs, S.R. Bare, J. Phys. Chem. 99 (1995) 10897.
[58] M.J. Vissenberg, L.J.M. Joosten, M.M.E.H. Heffels, A.J. van Welsenes, V.H.J. (San)
de Beer, R.A. van Santen, J.A. Rob van Veen, J. Phys. Chem. B 104 (2000) 8456.
[59] J. Medema, C. van Stam, V.H.J. de Beer, A.J.A. Konings, D.C. Koningsberger, J.
Catal. 53 (1978) 386.
[60] H. Jeziorowski, H. Knöringer, P. Grange, P. Gajardo, J. Phys. Chem. 84 (1980)
1825.
[61] I.E. Wachs, Catal. Today 27 (1996) 437.
[73] V.M. Kogan, P.A. Nikulshin, N.N. Rozhdestvenskaya, Fuel 100 (2012) 2.
[74] P.A. Nikulshina, A.V. Mozhaeva, K.I. Maslakov, A.A. Pimerzina, V.M. Kogan,
Appl. Catal. B 158–159 (2014) 161.
[62] L. Medici, R. Prins, J. Catal. 163 (1996) 28.
[75] S.C. Korre, M.T. Klein, R.J. Quann, Ind. Eng. Chem. Res. 34 (1995) 101.
[76] A.R. Beltramone, D.E. Resasco, W.E. Alvarez, T.V. Choudhary, Ind. Eng. Chem.
Res. 47 (2008) 7161.
[77] H.B. Yang, Y.C. Wang, H.B. Jiang, H.X. Weng, F. Liu, M.F. Li, Ind. Eng. Chem. Res.
53 (2014) 12264.
[63] M. Mihaylov, K. Hadjiivanov, Langmuir 18 (2002) 4376.
[64] T. Borowiecki, W. Gac, A. Denis, Appl. Catal. A 270 (2004) 27.
[65] P.A. Nikulshin, P.P. Minaev, A.V. Mozhaev, K.I. Maslakov, M.S. Kulikova, A.A.
Pimerzin, Appl. Catal. B 176–177 (2015) 374.
[66] S.-C. Qi, X.-Y. Wei, Z.-M. Zong, J. Hayashi, X.-H. Yuan, L.-B. Sun, ChemCatChem
5 (2013) 3543.
[78] P.P. Minaev, P.A. Nikulshin, M.S. Kulikova, A.A. Pimerzin, V.M. Kogan, Appl.
Catal. A. http://dx.doi.org/10.1016/j.apcata.2015.05.012.