Effect of citric acid addition on the morphology and activity of Ni2P supported on mesoporous zeolite ZSM-5 for the hydrogenation of 4,6-DMDBT and phenanthrene

Article abstract of DOI:10.1016/j.jcat.2016.11.019

Preparing small, highly dispersed Ni₂P particles is important for improving the hydrogenation ability of Ni₂P. Here, Ni₂P nanoparticles (approximately 4.3 nm) on mesoporous zeolite ZSM-5 (Ni₂P/MZSM-5-CA) were prepared using citric acid (CA) as an assistant agent. The formation mechanism of small Ni₂P particles when CA was added was investigated by combining UV–vis diffuse reflectance spectroscopy, Fourier transform infrared spectroscopy, and temperature-programmed reduction with a transmission electron microscope and CO chemisorption. The results indicated that the formed CA–Ni complex with high viscosity favors the Ni precursor dispersed on the dried catalyst. After calcination, the released Ni species strongly interacted with surface acidic hydroxyl groups on MZSM-5, leading to the formation of Ni₂P particles with small sizes and good dispersion under a reducing atmosphere. The reaction rate constants and TOFs over Ni₂P/MZSM-5-CA (16.2 × 10^?2μmol g^?1s^?1and 9.7 × 10^?4s^?1) are much higher than over Ni₂P/MZSM-5 (8.2 × 10^?2μmol g^?1s^?1and 8.3 × 10^?4s^?1) in 4,6-dimethyldibenzothiophene hydrodesulfurization. In addition, Ni₂P/MZSM-5-CA catalyst shows higher activity than Ni₂P catalyst without CA in phenanthrene hydrogenation.

Full text of DOI:10.1016/j.jcat.2016.11.019

Journal of Catalysis 345 (2017) 295–307
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Effect of citric acid addition on the morphology and activity of Ni2P
supported on mesoporous zeolite ZSM-5 for the hydrogenation
of 4,6-DMDBT and phenanthrene
a
a,
Lei Zhang a, Wenqian Fu b, Quanyong Yu b, Tiandi Tang b,c, , Yicheng Zhao , Yongdan Li
⇑
⇑
a Collaborative Innovation Center of Chemical Science and Engineering, State Key Laboratory of Chemical Engineering, School of Chemical Engineering,
Tianjin University, Tianjin 300072, People’s Republic of China
b School of Petrochemical Engineering, Changzhou University, Changzhou, Jiangsu 213164, People’s Republic of China
c 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:
Preparing small, highly dispersed Ni2P particles is important for improving the hydrogenation ability of
Ni2P. Here, Ni2P nanoparticles (approximately 4.3 nm) on mesoporous zeolite ZSM-5 (Ni2P/MZSM-5-CA)
were prepared using citric acid (CA) as an assistant agent. The formation mechanism of small Ni2P par-
ticles when CA was added was investigated by combining UV–vis diffuse reflectance spectroscopy,
Fourier transform infrared spectroscopy, and temperature-programmed reduction with a transmission
electron microscope and CO chemisorption. The results indicated that the formed CA–Ni complex with
high viscosity favors the Ni precursor dispersed on the dried catalyst. After calcination, the released Ni
species strongly interacted with surface acidic hydroxyl groups on MZSM-5, leading to the formation
of Ni2P particles with small sizes and good dispersion under a reducing atmosphere. The reaction rate
Received 11 July 2016
Revised 9 November 2016
Accepted 12 November 2016
Keywords:
Ni2P nanoparticles
CA assistant
Hydrogenation
4,6-Dimethyldibenzothiophene
Phenanthrene
constants and TOFs over Ni2P/MZSM-5-CA (16.2 ꢀ 10ꢁ2
l
mol gꢁ1 sꢁ1 and 9.7 ꢀ 10ꢁ4 sꢁ1) are much higher
than over Ni2P/MZSM-5 (8.2 ꢀ 10ꢁ2
l
mol gꢁ1 sꢁ1 and 8.3 ꢀ 10ꢁ4 sꢁ1) in 4,6-dimethyldibenzothiophene
hydrodesulfurization. In addition, Ni2P/MZSM-5-CA catalyst shows higher activity than Ni2P catalyst
without CA in phenanthrene hydrogenation.
Ó 2016 Elsevier Inc. All rights reserved.
1. Introduction
ration of polyaromatics at low temperature and pressure [7–9].
One solution for this problem is to develop a highly active
In recent years, with the reduction of petroleum reservoirs and
excessive exploitation, the polyaromatic hydrocarbons and bulky
organic sulfur compound content in crude oil are increasing
[1–3]. To meet increasingly stringent environmental legislation,
hydrotreatment catalyst that could achieve deep hydrodesulfuriza-
tion (HDS) of 4,6-DMDBT [10–12] as well as deep saturation of pol-
yaromatics [13].
Noble metal catalysts, such as Pd [14,15], Pt [16], and Pd–Pt
[17,18], exhibit high hydrogenation ability, but their sensitivity
to sulfur and high cost limit their industrial application. Numerous
studies have shown that nickel phosphide (Ni2P) catalysts present
high intrinsic hydrogenation activity in HDS reactions [19–21]. In
addition, the catalytic performance of the Ni2P catalyst is closely
connected with its active phase morphology [22–24]. It is reported
that there are tetrahedral Ni(1) sites and square pyramidal Ni(2)
sites in the Ni2P active phase [25–27]. The Ni(1) and Ni(2) sites
are responsible for desulfurization and hydrogenation, respec-
tively, and the number of Ni(2) sites increases with the decrease
of the Ni2P particle size. Therefore, a decrease in the Ni2P particle
size should increase the hydrogenation active sites and improve
the hydrogenation activity.
the production of ultraclean fuel through
a hydrotreatment
technique over conventional metal sulfide catalysts must be
carried out at a high temperature and high pressure [4–6], which
greatly increases the production cost of fuel. This is because metal
sulfide catalysts with low hydrogenation activity have difficulty
removing refractory sulfur-containing compounds, such as 4,6-
dimethyldibenzothiophene (4,6-DMDBT), and realizing deep satu-
⇑
Corresponding authors at: College of Chemistry and Materials Engineering,
Wenzhou University, Wenzhou 325035, People’s Republic of China. Fax: +86
57786689300 (T. Tang); Collaborative Innovation Center of Chemical Science and
Engineering, State Key Laboratory of Chemical Engineering, School of Chemical
Engineering, Tianjin University, Tianjin 300072, People’s Republic of China. Fax: +86
22 27405243 (Y. Li).
E-mail addresses: tangtiandi@wzu.edu.cn (T. Tang), ydli@tju.edu.cn (Y. Li).
http://dx.doi.org/10.1016/j.jcat.2016.11.019
0021-9517/Ó 2016 Elsevier Inc. All rights reserved.

296
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
An alternative approach to obtaining the Ni2P active phase with
(k = 0.154059 nm). The crystallite size of the passivated catalyst
was determined by the line broadening of the most intense peaks
at 2h = 40.8°, using the Scherrer equation, Dc = Kk/b cos(h), where K
is a constant taken as 0.9, k is the wavelength of the X-ray radia-
tion, b is the width of the peak at half maximum, and 2h is the
Bragg angle.
a small size is to increase the dispersion of Ni2P particles using a
high-surface-area support [22,27–29]. Oyama et al. reported that,
relative to a low-surface-area SiO2-supported Ni2P catalyst, smaller
Ni2P crystallites were formed on high-surface-area SiO2 (SiO2–H)
and had higher activity and better stability in the 4,6-DMDBT
HDS reaction [22]. Furthermore, compared with SiO2–H, much
smaller Ni2P particles were obtained by using higher-surface-
area ordered molecular sieves (MCM-41 or SBA-15) as supports
[25,29].
Another recent approach to prepare small crystalline Ni2P par-
ticles is to strengthen the metal–support interaction by choosing
supports with unique surface properties [23,30,31]. For instance,
Ni2P nanoparticles with very small and good dispersion were
formed on mordenite nanosheets (NS-HMOR) or mesoporous zeo-
lite ZSM-5. This is because the large number of acidic hydroxyl
groups on zeolites can strongly interact with Ni species, resulting
in highly dispersed Ni2P nanoparticles, which showed high activity
in the HDS of 4,6-DMDBT.
Associated studies have shown that the particle size of the bulk
Ni2P catalyst can be decreased by adding citric acid (CA) to the
preparation solution containing nickel nitrate and diammonium
hydrogen phosphate [32–34]. The obtained small Ni2P particles
with a relatively high external surface area shows better activity
in the HDS of the 4,6-DMDBT than the Ni2P particles prepared in
the absence of CA [33]. Possibly, the formed Ni(II)–CA complexes
are able to effectively inhibit metal aggregation [33–35], resulting
in a bulk Ni2P catalyst with small particle size, exposing more
active sites. Thus, it should be practicable to improve the Ni2P cat-
alyst with small particles by taking advantage of the special surface
properties of mesoporous zeolite and the nature of Ni(II)–CA
complexes.
We synthesized mesoporous zeolite ZSM-5 (MZSM-5), which is
impregnated with a solution containing nickel nitrate, diammo-
nium hydrogen phosphate, and CA to prepare the Ni2P catalyst
(Ni2P/MZSM-5). After a dried sample was calcined and reduced,
small Ni2P nanoparticles formed on MZSM-5. The formation mech-
anism of small Ni2P particles by adding CA as well as the effect of
the amount of CA on the Ni2P nanoparticle size in the catalyst
preparation was investigated. In addition, the activity of the
Ni2P/MZSM-5 catalysts was examined with respect to 4,6-
DMDBT HDS and phenanthrene hydrogenation.
The nitrogen adsorption isotherm was measured at the temper-
ature of liquid nitrogen (ꢁ196 °C) on a Micromeritics ASAP 2020 M
instrument. Approximately 0.15 g of sample was degassed at
200 °C for 10 h before the measurement. The specific surface area
was calculated according to the Brunauer–Emmett–Teller (BET)
equation. The mesopore size distribution was obtained by the Bar
rett–Joyner–Halenda (BJH) model using the desorption branch.
Transmission electron microscope (TEM) images were taken on
a JEM-2100 and FEI Tecnai G2 F30 instruments, operating at 200
and 300 kV, respectively. Before the measurement, the sample
was cut into thin slices and placed onto a Cu grid coated with a car-
bon membrane. The size distributions of the Ni2P phases were
obtained by statistical analyses carried out for at least 10 images
of 200–300 Ni2P particles from different regions of each catalyst.
The average Ni2P particle size (Daver) was calculated according to
the equation
P
n
i¼1di
Da
¼
;
v
er
ni
where ni is the number of Ni2P particles and di is the size of each
Ni2P particle.
The ultraviolet–visible (UV–vis) absorption was obtained on a
Shimadzu UV-3600 spectrometer. The UV–vis spectrum of the
nickel (II) species in the impregnation solution was recorded in
the wavelength range 300–1000 nm using 1 cm quartz cells. The
UV–vis diffuse reflectance spectrum (DRS) in the range 240–
800 nm was also obtained, using BaSO4 as a reference. Fourier
transform infrared spectroscopy (FT-IR) was performed on a Nico-
let iS50 instrument. The data were acquired using 16 scans and a
resolution of 4 cmꢁ1. The speciation diagrams of the Ni(II)–CA
species in solutions were predicted and plotted using Medusa
software [36].
The amount of residual carbon on the calcined catalysts was
measured by a carbon/sulfur analyzer (LECO CS844). Before mea-
surement, the carbon/sulfur analyzer was calibrated by a standard
sample. A temperature-programmed reduction (TPR) experiment
was carried out on a Micromeritics AutoChem II 2920 automatic
analyzer equipped with a thermal conductivity detector (TCD).
The calcined sample (60 mg) was reduced under a stream of H2–
Ar gas mixture (10 vol.% H2 at a flow rate of 50 mL/min), with a
heating rate of 15 °C/min up to 930 °C. To examine the acidity of
the support and catalysts, ammonia temperature-programmed
desorption (NH3-TPD) was performed on the same instrument, as
described elsewhere [31].
Dynamic CO pulsed chemisorption was also performed on a
Micromeritics ASAP 2920 automatic analyzer equipped with a
TCD. The passivated catalyst sample (0.06 g, maintained at room
temperature for 3 h in 0.5 wt.% O2/N2) was loaded into a quartz
reactor and reduced in H2 flow (99.999%, 100 mL/min STP) at
500 °C for 2 h with a ramp of 10 °C/min. Then the sample was
purged with He (99.999%, 50 mL/min STP) at 500 °C for 2 h and
cooled to 35 °C under a He flow. The CO pulse was repeatedly
injected until the response from the detector showed no further
CO uptake after consecutive injections [31].
2. Experimental
2.1. Catalyst
The mesoporous zeolite ZSM-5 (MZSM-5) was synthesized
hydrothermally as described in a previous study [13]. The Ni2P cat-
alysts were prepared by incipient wetness impregnation of the
MZSM-5 support using an aqueous solution containing the
required amounts of diammonium hydrogen phosphate ((NH4)2-
HPO4), nickel nitrate (Ni(NO3)2ꢂ6H2O), and citric acid (CA). The
molar ratio of Ni/P/CA was 1:2:X (X = 0, 1, 2, and 3), and the Ni
loading was 6.0 wt.%. After impregnation, the sample was dried
at ambient temperature for 12 h in air and then further dried at
100 °C for 12 h. Subsequently, the dried sample was extruded,
crushed and sieved to 20–40 mesh, and calcined at 500 °C for 3 h.
The catalysts with different amounts of CA were denoted as Ni2P/
MZSM-5-CA-X (X = 0, 1, 2, and 3), where X is the CA/Ni mole ratio.
The X-ray photoelectron spectroscopic (XPS) experiment was
carried out on an ESCALAB MK II instrument. Before XPS character-
ization, the freshly reduced Ni2P catalyst under a H2 stream was
transferred into a bottle filled with cyclohexane at room tempera-
ture. The cyclohexane was removed, and the residual catalyst was
2.2. Characterization
X-ray diffraction (XRD) measurements were carried out on a
RIGAKU UltimalV diffractometer using CuK
a
radiation

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
297
quickly moved to the sample holder before being loaded into the
XPS chamber.
35000
30000
25000
20000
15000
10000
5000
0
(a)
2.3. Activity test
The synthesized Ni2P catalyst was tested in a stainless steel
fixed-bed continuous-flow reactor. For the HDS of 4,6-DMDBT,
the calcined and screened 0.05 g catalyst (20–40 mesh) was fully
mixed with silicon carbide (20–40 mesh, 2.0 g) and loaded into
the fixed-bed reactor. Then the precursor of the catalyst was con-
verted to the active Ni2P phase by the TPR method in H2 at a flow
rate of 160 mL/min. The detailed activation process was as follows:
the catalyst bed was heated to 120 °C at 10 °C/min and maintained
for 1 h at this temperature, and then further heated to 400 °C at
5 °C/min and from 400 to 550 °C at 2.5 °C/min; then the tempera-
ture was maintained at 550 °C for 3 h. Finally, the HDS experiments
were performed at a temperature of 300 °C, a pressure of 5.0 MPa, a
weight hourly space velocity (WHSV) of 75.0 hꢁ1, a H2 flow of
60 mL/min, and a liquid feed of 0.5 wt.% 4,6-DMDBT in decalin.
The intrinsic HDS activity of the catalyst was also investigated after
the effects of mass transfer were eliminated by the Weisz–Prater
(CWP) and Mears criteria (CM) [37]. The detailed experimental pro-
cess and reaction conditions, as well as the calculations, are pre-
sented in Supplementary Material. The catalytic HDS rate
constants (kHDS) and HDS turnover frequencies of 4,6-DMDBT
(TOF, sꢁ1) were obtained by the two equations [6,38]
MZSM-5
5
10
15
20
25
30
35
40
45
50
2 Τheta (deg.)
0.7
350
300
250
200
150
100
21.7 nm
(b)
0.6
0.5
0.4
0.3
0.2
0.1
0.0
1
10
Pore Diameter (nm)
ꢀ
ꢁ
F
W
1
1 ꢁ x
kHDS
¼
ln
;
F ꢀ x
TOF ¼
0.0
0.2
0.4
0.6
0.8
1.0
W ꢀ M
Relative Pressure (P/Po)
where x is the conversion of the 4,6-DMDBT (%), F is the molar feed
rate of the 4,6-DMDBT (
mol sꢁ1), W is the catalyst mass (g), kHDS is
the HDS rate constant (
mol gꢁ1 sꢁ1), and M is the number of mole
Fig. 1. (a) XRD pattern and (b) N2 adsorption isotherms of the MZSM-5 sample
(inset, mesoporous size distribution of the sample).
l
l
of loaded sites, which was determined by the CO uptake.
In addition, a phenanthrene (PHE) hydrogenation experiment
was performed as follows: The calcined and screened catalyst
(0.2 g, 20–40 mesh) was diluted with silicon carbide (1.5 g, 20–
40 mesh) before being loaded into the fixed-bed reactor. The acti-
vation of the catalyst was the same as described above. The hydro-
genation reaction conditions were as follows: a temperature of
260 °C, a total pressure of 5.0 MPa, a H2 flow of 60 mL/min, a WHSV
of 21.0 hꢁ1, and 1.0 wt.% PHE in decalin. The reaction product was
collected and analyzed using an offline Agilent 7890B GC installed
with a flame ionization detector.
3. Results and discussions
3.1. General characterization
Fig. 1a gives the XRD pattern of the synthesized MZSM-5, which
displays typical diffraction peaks in the range from 5° to 50°, con-
sistent with the MFI structure. The N2 sorption isotherm of MZSM-
5 in powder form, as shown in Fig. 1b, shows a hysteresis loop at a
relative pressure of 0.45–0.96 due to the introduction of a meso-
porous structure. Correspondingly, the mesopore size distribution
of the MZSM-5 is mainly centered at 21 nm (inset in Fig. 1b). The
mesopore size is in good agreement with the TEM results of the
ultrathin section MZSM-5, providing direct evidence for the suc-
cessful introduction of a sufficient number of hierarchical meso-
pores into the crystals (Fig. 2). These nanoporous channels can
reduce the diffusion limitation on the mass transfer of the bulky
molecules. The detailed textural properties of the supports and
corresponding catalysts are listed in Table 1. The synthesized
Fig. 2. TEM image of the sectioned thin MZSM-5 sample.
MZSM-5 specific surface area is 429 m2/g, the external surface area
is 193 m2/g, and the mesoporous volume is 0.43 m3/g. After load-
ing of Ni2P precursors and carrying out the reduction, the MZSM-
5-supported Ni2P catalysts show a decrease in textural parameters
compared with the MZSM-5 support. It is worth noting that there
are no significant differences in the textural properties for the

298
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
Table 1
The textural and acidic properties of the samples.
Sample
SBET (m2/g)a
Vmicro (cm3/g)b
Vmeso (cm3/g)c
Smic (m2/g)d
Sext (m2/g)e
Acidic amounts (lmol/g)
MZSM-5f
MZSM-5
Ni2P/MZSM-5-CA-0
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-3
429
365
261
258
262
259
0.10
0.08
0.05
0.05
0.05
0.05
0.43
0.27
0.36
0.39
0.37
0.39
236
193
113
109
115
112
193
172
148
148
146
147
176
–
160
143
120
114
a
b
c
d
e
f
BET surface area.
Microporous pore volume, obtained from the t-plot method.
Mesoporous pore volume obtained from the t-plot method.
Microporous surface area.
External surface area obtained from the t-plot method.
The sample was in powder form, and other samples were formed into tablets at a pressure of 10 MPa.
MZSM-5-supported Ni2P catalyst with different CA/Ni ratios, sug-
gesting that the CA addition and its amount have little influence
on the textural properties of the catalyst.
structures are observed (Fig. S1 in the Supplementary Material).
The EDX elemental mapping analysis suggests that some Ni and
P species are present besides the Si, Al, and O in the selected zone
(Fig. S2), and the Ni and P content is 0.33 and 0.08 wt.% (Table S1),
respectively. These results demonstrate that Ni and P species
existed in the micropore channels of ZSM-5, indicating that Ni spe-
cies could be inside MZSM-5 micropores, and not all Ni species are
Ni2P on the external surface of zeolite crystals. Meanwhile, the
micropore volume of the Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3)
catalysts is also decreased relative to that of the MZSM-5 support,
suggesting that some Ni and P species can migrate into the MZSM-
5 micropore channels during the catalyst preparation. The high-
resolution TEM image of Ni2P/MZSM-5-CA-2 shows that the Ni2P
particles with a visible lattice spacing of approximately 0.22 nm
are consistent with the {111} crystallographic planes of Ni2P,
which have a d-spacing value of 0.221 nm (Fig. S3). The size distri-
bution of the Ni2P particles was also analyzed by statistics method,
and the results are shown in Fig. 5. The Ni2P particle size on the
Ni2P/MZSM-5-CA-0 catalyst ranges from 3 to 18 nm (Fig. 5a), while
that on Ni2P/MZSM-5-CA-X (X = 2, 3) samples are mainly 2–7 nm
(Fig. 5c and d). The calculated average sizes of the Ni2P particle
on these samples decrease as follows: Ni2P/MZSM-5-CA-0
The XRD results of the passivated Ni2P/MZSM-5-CA-X (X = 0, 1,
2, and 3) samples are shown in Fig. 3. A typical peak at 40.7 asso-
ciated with the Ni2P phase (PDF: 03-0953) was detected, demon-
strating that the Ni2P active phases were formed on Ni2P/MZSM-
5-CA-X (X = 0, 1, 2, and 3) catalysts. This peak at 40.7 is broadened
with increasing CA/Ni mole ratios, suggesting that relatively smal-
ler Ni2P particles could be formed with an increase in the CA/Ni
mole ratio. Analysis of the Ni2P phase size calculated by the Scher-
rer equation suggests that the particle size of the Ni2P active phase
is decreased with an increasing CA/Ni ratio, from 9.0 to 8.2, to 6.9,
and then to 6.8 nm (Table 2). There is no significant decrease in the
size of Ni2P particles as the CA/Ni mole ratio is up to 3 (6.8 nm).
Therefore, it can be concluded that the CA addition in the impreg-
nation solution plays an essential role in the formation of smaller
Ni2P nanoparticles.
Fig. 4 shows the TEM images of the reduced Ni2P/MZSM-5-CA-X
(X = 0, 1, 2, and 3) samples, giving information about the particle
size and dispersion of the Ni2P active phase. Compared with
the Ni2P/MZSM-5-CA-0 sample (Fig. 4a), much smaller Ni2P
particles were formed on Ni2P/MZSM-5-CA-X (X = 2, 3) samples
(Fig. 4c and d), suggesting that adding the CA to the impregnation
solution resulted in the formation of Ni2P particles with small
sizes. To insight into the fact that whether the nickel species are
inside of the micropores of MZSM-5, the mesopore-free zeolite
ZSM-5-supported Ni2P (Ni2P/ZSM-5) catalyst in the presence of
CA was prepared and characterized by TEM imaging and energy-
dispersive X-ray spectroscopy (EDX). It is clear that micropore
(8.3 nm) > Ni2P/MZSM-5-CA-1
(7.1 nm) > Ni2P/MZSM-5-CA-2
(4.5 nm) ꢃ Ni2P/MZSM-5-CA-3 (4.3 nm), which is in good agree-
ment with the XRD analysis results.
Generally, small Ni2P particles with good dispersion could pro-
vide a large number of active sites in the hydrotreatment reaction,
which can be determined by the CO chemisorption method [39,40].
Table 2 presents the CO chemisorption results of the reduced Ni2P/
MZSM-5-CA-X (X = 0, 1, 2, 3) samples. Clearly, Ni2P/MZSM-5-CA-2
(146
Ni2P particle sizes, have much higher CO uptake values than those
of Ni2P/MZSM-5-CA-0 (93 mol/g) and Ni2P/MZSM-5-CA-1
(118 mol/g). This result further confirms that adding CA to the
lmol/g) and Ni2P/MZSM-5-CA-3 (147 lmol/g), with similar
l
l
impregnation solution for the catalyst preparation can favor the
formation of small Ni2P particles, leading to the exposure of more
active sites.
The acidity of the MZSM-5 support and reduced Ni2P/MZSM-5-
CA-X (X = 0, 1, 2 and 3) samples was measured by NH3-TPD exper-
iments, and these results are exhibited in Fig. 6 and Table 1. The
NH3 desorption profiles in the range of 300–450 °C were observed
for MZSM-5 and its supported catalysts, which correspond to the
desorbed NH3 in interaction with the strong acidic sites [41]. After
the introduction of Ni and P species and reduction, the number of
e
d
c
b
a
acidic sites of 176 lmol/g for MZSM-5 was decreased to 160 lmol/
g for the Ni2P/MZSM-5-CA-0, which was due to the Ni species
interacting with the acidic hydroxyl groups on MZSM-5; this is dis-
cussed in detail in our previous work [31]. When Ni and P species
are introduced onto MZSM-5 by adding CA, the number of acidic
sites on the reduced catalysts was further decreased. The numbers
10
20
30
40
50
2 Theta (deg.)
Fig. 3. XRD patterns of the (a) Ni2P (PDF: 03-0953), (b) Ni2P/MZSM-5-CA-0, (c)
Ni2P/MZSM-5-CA-1, (d) Ni2P/MZSM-5-CA-2, and (e) Ni2P/MZSM-5-CA-3 samples.

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
299
Table 2
The Ni2P particle size, CO uptake, and carbon content of the samples.
Sample
Dc (nm)a
Daver. (nm)b
CO uptake (
lmol/g)
Residual carbon (%)c
Ni2P/MZSM-5-CA-0
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-3
9.0
8.2
6.9
6.8
8.3
7.1
4.5
4.3
93
–
118
146
147
0.171
0.196
0.253
a
b
c
Determined by the Scherrer equation.
Obtained by statistical analyses of the TEM images.
Calcined supported Ni2P catalysts.
Fig. 4. TEM images of the reduced (a) Ni2P/MZSM-5-CA-0, (b) Ni2P/MZSM-5-CA-1, (c) Ni2P/MZSM-5-CA-2, and (d) Ni2P/MZSM-5-CA-3 samples.
of acidic sites on MZSM-5, Ni2P/MZSM-5-CA-0, Ni2P/MZSM-5-CA-
1, Ni2P/MZSM-5-CA-2, and Ni2P/MZSM-5-CA-3 are 176, 160, 143,
120, and 114 lmol/g, respectively. These results indicate that add-
ing CA to the catalyst impregnation solution can favor the Ni spe-
cies interaction with MZSM-5 through the surface acidic hydroxyl
groups.
The state of the Ni and P species on the freshly reduced Ni2P cat-
alysts was analyzed by the XPS technique, and the spectrum and
the surface atom ratio on the catalyst are shown in Fig. 7 and
Table 3, respectively. As reported in the literature [28,38,42], Ni
and P species in the Ni2P phase can exist as Nid+ (0 < d < 1) and
Pdꢁ (0 < d < 1), which correspond to binding energies of 852.5 to
853.4 eV and 129.2 to 129.5 eV, respectively. In our case, the
Ni2p peak with a binding energy of 852.4 eV and a P2p peak with
binding energy of 129.2 eV suggested that the Ni2P active phase
was formed on both the reduced Ni2P/MZSM-5-CA-0 and Ni2P/
MZSM-5-CA-2 catalysts. In addition, a small peak with binding
energy of 855.5 eV on both the samples was corresponding to the
Ni2+ species, along with satellite peaks at approximately 6.0 eV
above the Ni2+ species, which may be due to the oxidation in the
process of sample transfer [27]. Meanwhile, broad peaks were
detected for the P2p XPS spectra on the two samples (Fig. 7), which
could involve the contributions of both PO34ꢁ and H2PO3ꢁ species
[27,38], respectively. The surface atom ratio on both of the cata-
lysts shows a high P/Ni ratio on reduced Ni2P/MZSM-5-CA-0
(4.03) and Ni2P/MZSM-5-CA-2 (3.70) catalysts, which means that
most P species were enriched on the catalyst surface. Ni2P/
MZSM-5-CA-2 shows a higher Ni (or P)/(Si + Al) ratio (0.041 or
0.16) compared with the Ni2P/MZSM-5-CA-0 (0.032 or 0.13), which
may be because more of the Ni2P active phase was dispersed on the
external surface as CA was added.
3.2. Ni(II)–CA structure identification in the impregnation solution
Generally, the complex structure of Ni(II)–CA depends on the
parameters of Ni species concentration, CA concentration, and

300
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
0.4
0.3
0.2
0.1
0.4
(a)
(b)
0.3
0.2
0.1
0.0
0.0
0.4
0.3
0.2
0.1
0.0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18
Ni2P Particle Size (nm)
Ni2P Particle Size (nm)
0.4
0.3
0.2
0.1
0.0
(c)
(d)
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18
Ni2P Particle Size (nm)
Ni2P Particle Size (nm)
Fig. 5. The size distributions of Ni2P particles on reduced (a) Ni2P/MZSM-5-CA-0, (b) Ni2P/MZSM-5-CA-1, (c) Ni2P/MZSM-5-CA-2, and (d) Ni2P/MZSM-5-CA-3 samples.
speciation diagram of the Ni(II)–CA species in solutions was pre-
dicted using Medusa software, and the results are shown in
374
(e)
Fig. 8. When the CA/Ni ratio in the impregnation solution was 1,
the Ni(II)–CA structure was Ni(cit)ꢁ; when the CA/Ni ratio in the
(d)
impregnation solution was 2 or 3, the Ni(II)–CA structures were
the same as each other, Ni(Hcit)(cit)3ꢁ
.
(c)
(b)
(a)
To confirm the prediction of Ni(II)–CA complex structure,
impregnation solutions with different CA/Ni ratios were analyzed
by UV–vis and FT-IR spectroscopy. Fig. 9 shows the UV–vis spectra
of the impregnation solution with CA/Ni ratios of 0, 1, 2, and 3.
Generally, the absorption bands at 390–400 and 650–750 nm, cor-
364
350
responding to the two spin-allowed 3A2g ? 3T1g (P) and 3A2g
?
3T1g (F) transitions, are assigned to the octahedral nickel species
in aqueous solution [43]. In this work, the impregnation solution
without CA shows two absorption bands at 397 and 650–750 nm
(Fig. 9a), typical of Ni (II) ions in an octahedral coordination of
[Ni(H2O)6]2+. When the impregnation solution contains CA, a
blue-shift phenomenon is observed, indicating that the Ni(II)–CA
complex is formed in the aqueous solution (Fig. 9b–d) [5,43]. For
example, when the CA/Ni ratio in solution is 1, 2, and 3, the absorp-
tion band at 397 nm shifts to 392, 390, and 390 nm, respectively,
which could be due to the replacement of the H2O in [Ni
(H2O)6]2+ by CA. These results suggest that the structure of the
Ni(II)–CA complex is associated with the CA/Ni ratio in the
250
300
350
400
450
500
550
600
Temperature (oC)
Fig. 6. NH3-TPD profiles of the (a) MZSM-5 support and the reduced (b) Ni2P/
MZSM-5-CA-0, (c) Ni2P/MZSM-5-CA-1, (d) Ni2P/MZSM-5-CA-2, and (e) Ni2P/MZSM-
5-CA-3 samples.
the pH value of the solution [36,4]. In the present work, the Ni spe-
cies concentration is 0.716 M, and the pH value is approximately
3.6–4.1 for the employed impregnation solution. In this case, the

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
301
Based on the UV–vis and FT-IR results, as well as the prediction
from Fig. 8, the proposed structure of Ni(II)–CA complexes in the
impregnation solution are shown in Fig. 11. At a CA/Ni ratio of 1,
the structure of the Ni(cit)ꢁ complex shows that Ni2+ ions are com-
plexed with CA via three carboxyl groups and one hydroxyl group,
as well as two H2O molecules (Fig. 11b). At CA/Ni ratios of 2 or 3,
four carboxyl groups and two hydroxyl groups of the two CA mole-
cules involve with the coordination of Ni2+ species to form Ni(Hcit)
(cit)3ꢁ complex, remaining uncoordinated carboxyl groups
(Fig. 11c).
Ni 2p
Niδ+
Ni2+
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-0
Satellite
3.3. Ni(II)–CA structure identification on the dried catalyst
After impregnating the MZSM-5 support with a solution con-
taining Ni, P, and CA, the dried samples were also analyzed by
FT-IR and UV–vis spectroscopy and the results are shown in
Fig. 12. In contrast to the FT-IR spectra of impregnation solutions
with different CA/Ni ratios, similar results are obtained on these
dried Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3) catalyst samples
(Fig. 12a). The absorption bands at 1634, 1402, and 1383 cmꢁ1
for all the samples are assigned to the vibration of H2O, NOꢁ3 , and
PO34ꢁ, respectively [44–46]. After CA is introduced, two absorption
bands at 1726 and 1575 cmꢁ1 appear, which correspond to the
vibration of the nondissociated carboxyl group in CA and the sym-
metric vibration of COOꢁ in the Ni(II)–CA complex, respectively
885
880
875
870
865
860
855
850
845
Binding Energy (eV)
H PΟ−
P 2p
PΟ34−
2
3
Pδ-
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-0
[45]. It is notable that
a very low absorption intensity at
1726 cmꢁ1 in the dried Ni2P/MZSM-5-CA-1 sample indicates that
most carboxyl groups in CA were involved with the formation of
the Ni(II)–CA complex. In contrast, a relatively high absorption
intensity at 1726 cmꢁ1 in the Ni2P/MZSM-5-CA-2 and Ni2P/
MZSM-5-CA-3 samples suggests that besides the coordinated car-
boxyl groups, many residual carboxyl groups are nondissociated
from the Ni(II)–CA complex or CA. These results indicate that Ni
(II)–CA complexes with different structures may be formed on
the MZSM-5 surface due to the change in CA/Ni ratio in impregna-
tion solutions. This suggestion is further supported by the analysis
of UV–vis spectra for the dried Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and
3) samples (Fig. 12b). The band at 409 nm for the dried Ni2P/
MZSM-5-CA-0 sample shifts to low wavelengths at 403, 397, and
397 nm for the dried Ni2P/MZSM-5-CA-X (X = 1, 2, and 3) samples,
respectively. Combined the FT-IR and UV–vis results, it could be
suggested that the Ni(II)-CA complex on the dried Ni2P/MZSM-5-
CA-1 sample is mainly in the form of Ni(cit)ꢁ, while on the dried
Ni2P/MZSM-5-CA-2 and Ni2P/MZSM-5-CA-3, samples are domi-
142 140 138 136 134 132 130 128 126 124
Binding Energy (eV)
Fig. 7. The XPS spectra of (a) Ni 2p and (b) P 2p in the reduced Ni2P/MZSM-5-CA-0
and Ni2P/MZSM-5-CA-2 samples.
impregnation solution. In addition, the same structured Ni(II)–CA
complex should be formed at CA/Ni ratios of 2 and 3 in the solu-
tion. The suggestion is further confirmed by the FT-IR spectra of
the impregnation solution with different CA/Ni ratios, and the
results are shown in Fig. 10. There are three absorption bands at
1636, 1400, and 1385 cmꢁ1 observed on all samples, which corre-
sponding to the vibration of H2O, NOꢁ3 , and PO43ꢁ, respectively [44–
46]. When CA was added into the impregnation solution, two other
absorption bands at 1730 and 1571 cmꢁ1 appear, which are
assigned to the vibration of the nondissociated carboxyl group in
CA and the symmetric vibration of COOꢁ in the Ni(II)–CA complex,
respectively [45]. It is found that the absorption intensity at
1730 cmꢁ1 is much stronger for the impregnation solution with a
CA/Ni ratio of 2 or 3 than for that with a CA/Ni ratio of 1. These
results suggest that the formed Ni(II)–CA complexes at CA/Ni ratios
of 2 and 3 have more residual nondissociated carboxyl groups than
those with a CA/Ni ratio of 1, indicating that the CA/Ni ratio in
impregnation affects the Ni(II)–CA complex structure.
nantly in the form of Ni(Hcit)(cit)3ꢁ
.
3.4. Discussion of small Ni2P particle formation
3.4.1. Facilitation of the formation of small Ni2P particles by CA
addition
CA addition plays an important role in the preparation of small
Ni2P particles on the MZSM-5 support. It has been reported that
when aqueous chelated metal complexes are utilized to prepare
supported metal catalysts, a metal complex film with a high viscos-
ity can be formed on the support surface, which can greatly restrict
the aggregation of metal species [33,34]. In this work, the formed
Table 3
The XPS analysis results of the reduced Ni2P/MZSM-5-CA-0 and Ni2P/MZSM-5-CA-2 samples.
Catalysts
Ni2p3/2 (binding energy)a
P2p3/2 (binding energy)a
Superficial atomic ratio
Ni2+
Ni2P
PO34ꢁ
H2POꢁ3
Ni2P
P/Ni
Ni/(Si + Al)
P/(Si + Al)
Ni2P/MZSM-5-CA-0
Ni2P/MZSM-5-CA-2
855.3 (33.8)
855.7 (39.4)
852.4 (66.2)
852.3 (60.6)
134.5 (47.8)
134.7 (48.5)
133.4 (45.0)
133.7 (42.8)
129.2 (7.2)
129.2 (8.7)
4.03
3.67
0.032
0.041
0.13
0.16
a
The data in parentheses are atom percentages.

302
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
CA/Ni=3
c
b
CA/Ni=2
CA/Ni=1
CA/Ni=0
a
1571
1636
1730
1385
1400
2400
2200
2000
1800
1600
1400
1200
-1
Wavenumber (cm )
Fig. 10. FT-IR spectra of the impregnation solution with different CA/Ni molar
ratios.
pH
Fig. 8. Predominance area diagrams for aqueous Ni(II)–CA solution. [Ni(II)]
aq = 0.716 M. Dotted lines correspond to CA concentrations for (a) CA/Ni(II) = 1,
(b) CA/Ni(II) = 2, and (c) CA/Ni(II) = 3 used in the present work.
peaks at 669 °C and 785 °C are assigned to the reduction of Ni and P
species, respectively [31]. The temperature peak of the Ni species
shifted from 669 °C to 705, 719, and 719 °C, respectively, for cal-
cined Ni2P/MZSM-5-CA-X (X = 1, 2, 3) samples. These results prove
that the Ni–support interaction became stronger with the increase
in the amount of CA addition to benefit the dispersion of the Ni2P
active phase on MZSM-5, which is supported by the XRD, TEM, and
CO uptake results as discussed above. In addition, the temperature
interval of the reduction peaks of the Ni and P species decreased
from 116 °C for Ni2P/MZSM-5-0 to 62 °C for Ni2P/MZSM-5-CA-1
and disappeared for Ni2P/MZSM-5-CA-X (X = 2, 3). The transforma-
tion of the oxide precursors into phosphides begins with the reduc-
tion of nickel oxide reduction; then the formed Ni assists the
production of volatile phosphorus species and the formation of
the nickel phosphide phases [47]. Therefore, the CA-added Ni2P
catalyst with a small reduction-peak interval could enable the
reaction between the Ni species and the reduced P species in a
timely manner to form relatively small Ni2P nanoparticles. The
strengthened interaction of Ni with MZSM-5 inducing the highly
dispersed small Ni2P particles is in agreement with the XRD anal-
ysis results and the TEM observations, as discussed above. The Ni2P
particle sizes are 9.0, 8.2, 6.9, and 6.8 nm on Ni2P/MZSM-5-CA-X
(X = 0, 1, 2, and 3) by XRD analysis, and the values for those sam-
ples are 8.3, 7.1, 4.5, and 4.3 nm, when statistically calculated
based on the TEM images.
390
390
CA/Ni=3
392
CA/Ni=2
397
CA/Ni=1
730
656
CA/Ni=0
400
500
600
700
800
900
Wavelength (nm)
Fig. 9. UV–vis spectra of the impregnation solution with different CA/Ni molar
ratios.
Ni(II)–CA complex could favor the Ni species being highly dis-
persed on the mesopore surface and on the outer surface of
MZSM-5, preventing the aggregation of Ni species compared with
that without CA addition during the catalyst preparation process.
During the calcination of the dried catalyst, the Ni(II)–CA com-
plexes on the MZSM-5 surface are gradually decomposed (Fig. S4 in
the Supplementary Material), and the released Ni species may be
able to interact with acidic hydroxyl groups on the zeolite surface
[31]. As discussed in the NH3-TPD characterization, the acidic sites
on the reduced Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3) catalysts are
decreased compared with those on the MZSM-5 support. These
results indicate that CA addition favors the Ni species interacting
with acidic hydroxyl groups on the MZSM-5, and the formed
Ni2P active phase can cover the acidic sites.
3.4.2. Effect of the different Ni(II)–CA structures on the Ni2P particle
size
As discussed above, the different Ni(II)–CA structures, Ni(cit)ꢁ
and Ni(Hcit)(cit)3ꢁ, can be formed by changing the CA/Ni ratios
and the pH values in the impregnation solution, which may change
the interaction between Ni species and MZSM-5 support. In con-
trast to Ni(cit)ꢁ, the Ni(Hcit)(cit)3ꢁ complex has remaining car-
boxyl groups, which benefit the formation of a relatively strong
interaction between surface hydroxyl groups of zeolite and car-
boxyl groups in Ni(Hcit)(cit)3ꢁ via hydrogen bond (Fig. S5 in the
Supplementary Material) [48]. In this manner, relatively small
Ni2P particles could be formed after the catalyst calcination and
reduction. For instance, when the CA/Ni ratio is 1, the interaction
of the Ni(cit)ꢁ complex with the zeolite is relatively weak, resulting
in a Ni2P active phase with average particle size 7.1 nm on the
Ni2P/MZSM-5-CA-1 catalyst. In contrast, when the CA/Ni ratio is
2 or 3, the interaction of the Ni(Hcit)(cit)3ꢁ complex with the zeo-
The TPR results confirm the strong interactions between Ni spe-
cies and the MZSM-5. Fig. 13 gives the TPR curves of the calcined
Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3) samples. For the reduction
profile of the calcined Ni2P/MZSM-5-CA-0 sample, the temperature

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
303
Fig. 11. Molecule structure of the (a) citric acid and the Ni(II)–CA complexes of (b) Ni(cit)ꢁ and (c) Ni(Hcit)(cit)3ꢁ. The hydrogen atom was hidden.
719
Ni2P/MZSM-5-CA-3
(a)
(d)
Ni2P/MZSM-5-CA-2
719
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-0
(c)
708
770
785
1726
1575
(b)
(a)
669
1634
1402
1383
2400
2200
2000
1800
1600
1400
100 200 300 400 500 600 700 800 900
Temperature (oC)
-1
Wavenumber (cm )
1.2
Fig. 13. H2 TPR profiles of the calcined (a) Ni2P/MZSM-5-CA-0, (b) Ni2P/MZSM-5-
CA-1, (c) Ni2P/MZSM-5-CA-2, and (d) Ni2P/MZSM-5-CA-3 samples.
397
(b)
1.0
0.8
0.6
0.4
0.2
0.0
composed carbonaceous residuals could exist in the catalyst pre-
cursor, and could act as a structure promoter to inhibit the
aggregation of active phases [33,34,49]. In our case, the residual
carbon in the calcined Ni2P catalyst increased with the increase
in the amount of CA in the catalyst preparation (Table 2), which
is able to inhibit the growth of nickel oxide particles during the cal-
cination process, leading to the formation of small Ni2P particles.
To verify this judgment, an additional carbon-containing Ni2P/
MZSM-5 catalyst (Ni2P/MZSM-5-C) was prepared as follows: First,
the MZSM-5 support was impregnated with the CA solution. After
drying, the sample was calcined at 350 °C for 4 h under N2. Then
the obtained C-containing MZSM-5 was impregnated with the
solution containing Ni and P precursors. The dried sample was cal-
cined at 500 °C for 3 h in O2–Ar mixed gas (5 mol.% O2 in Ar). The Ni
loading was 6.0 wt.%, the Ni/P molar ratio was 1:2, and the residual
carbon in this catalyst is 0.153 wt.%. The Ni2P particle size on the
Ni2P/MZSM-5-C catalyst is 8.5 nm by XRD data, a little smaller than
that on Ni2P/MZSM-5-CA-0 catalyst (9.0 nm). In addition, the CO
chemisorption results show that the CO uptake over Ni2P/MZSM-
Ni2P/MZSM-5-CA-3
397
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-0
403
409
400
500
600
700
800
Wavelength (nm)
Fig. 12. The FT-IR (a) and UV–vis (b) spectra of the dried samples of Ni2P/MZSM-5-
CA-X (X = 0, 1, 2, and 3).
lite is relatively strong, resulting in a Ni2P active phase with aver-
age particle size 4.3–4.5 nm on the Ni2P/MZSM-5-CA-2 and Ni2P/
MZSM-5-CA-3 catalysts. It is mentioned above that the reduction
temperature of Ni species on the Ni2P/MZSM-5-CA-1 sample is
705 °C, lower than those of 719 and 719 °C on Ni2P/MZSM-5-CA-
2 and Ni2P/MZSM-5-CA-3 (Fig. 13), suggesting that a relatively
strong Ni–support interaction exists in Ni2P/MZSM-5-CA-2 and
Ni2P/MZSM-5-CA-3.
5-C (99
lmol/g) is higher than that over Ni2P/MZSM-5-CA-0
(93 mol/g, Table 2). These results indicate that the residual car-
l
bon in the catalyst promotes the formation of small Ni2P particles.
It is found that the reduction temperature of Ni2P/MZSM-5-C
(708 °C) is higher than that of Ni2P/MZSM-5-CA-0 (669 °C, Fig. S6
in the Supplementary Material), which implies that the residual
carbon strengthened the interaction between the Ni species and
MZSM-5 for the Ni2P/MZSM-5-C catalyst. The strengthened inter-
action should facilitate the formation of a small Ni2P active phase
on the Ni2P/MZSM-5-C catalyst. Therefore, it could be concluded
3.4.3. Effect of the residual carbon in the catalyst on the formation of
Ni2P particles
It is reported that the when CA was used as a chelating agent to
prepare the bulk metal phosphide (Ni2P, MoP) catalyst, the unde-

304
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
Table 4
that the residual carbon in the calcined Ni2P/MZSM-5-CA-X (X = 1,
2, and 3) catalysts also plays a role in inhibiting the growth of the
nickel oxide particles and benefiting the formation of the small
Ni2P particles on the MZSM-5.
Intrinsic activity of the Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3) catalysts in the 4,6-
DMDBT HDS.a
b
Catalyst
kHDS (10ꢁ2
l
mol gꢁ1 sꢁ1
)
TOF (10ꢁ4 sꢁ1
)
Ni2P/MZSM-5-CA-0
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-2
Ni2P/MZSM-5-CA-3
8.2 0.15
11.5 0.39
16.2 0.43
16.1 0.69
8.3 0.17
8.9 0.23
9.7 0.19
9.6 0.34
3.5. Activity of the Ni2P/MZSM-5-CA
The catalytic performance of the CA-promoted Ni2P catalysts
supported on MZSM-5 was first examined in the HDS of the 4,6-
DMDBT. Fig. 14 shows the dependence of 4,6-DMDBT conversion
over Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3) catalysts on reaction
time. Clearly, the Ni2P/MZSM-5-CA-0 catalyst exhibits the lowest
catalytic activity (conversion of 35.1% at 15 h). In contrast, the cat-
alytic activity over the CA-added Ni2P catalysts is gradually
increased with increasing CA amount. For example, the Ni2P/
MZSM-5-CA-2 catalyst has a much higher catalytic activity (con-
version of 41.3% at 15 h) than the Ni2P/MZSM-5-CA-1 catalyst
(conversion of 38.3% at 15 h). These results suggest that the cat-
alytic activity of the MZSM-5-supported Ni2P catalysts was obvi-
ously increased when the Ni2P particle size was decreased from
8.3 nm on Ni2P/MZSM-5-CA-0 to 7.1 nm on Ni2P/MZSM-5-CA-1
and 4.5 nm on Ni2P/MZSM-5-CA-2. However, by further increasing
the amount of CA in the catalyst preparation, the obtained Ni2P/
MZSM-5-CA-3 catalyst exhibits comparable 4,6-DMDBT conver-
sion compared with Ni2P/MZSM-5-CA-2 catalyst (41.3% vs 41.2%
at 15 h). This may be due to the fact that the formed Ni2P phases
on these two catalysts have similar particle size and dispersion,
which was confirmed by the XRD, TEM, and CO uptake.
To compare the intrinsic HDS catalytic activity, the reaction rate
constants (kHDS) and TOFs for the HDS of the 4,6-DMDBT over these
supported Ni2P catalysts were also obtained in the absence of mass
transfer limitations, and these results are listed in Table 4 (the fit-
ting curves are given in Fig. S7 in the Supplementary Material).
Clearly, it is found that kHDS increased with an increase in the
amount of the CA addition: Ni2P/MZSM-5-CA-0 < Ni2P/MZSM-5-
CA-1 < Ni2P/MZSM-5-CA-2. This can be explained by the fact that
the addition of different amounts of CA could change the CA–Ni
(II) complex structure and the strength of the interaction between
the Ni species and MZSM-5 support, altering the particle size of the
formed Ni2P phase. The Ni2P particle size on the Ni2P/MZSM-5-CA-
2 and Ni2P/MZSM-5-CA-3 (3–6 nm) is smaller than on Ni2P/MZSM-
5-CA-0 (4–12 nm) and Ni2P/MZSM-5-CA-1 (4–10 nm, Fig. 4), and
a
Reaction conditions: 0.05 g catalyst, total pressure 5.0 MPa, temperature 300°C,
H2 flow rate 60 mL minꢁ1, 0.5 wt.% 4,6-DMDBT in decalin. The 4,6-DMDBT con-
version is controlled by changing the WHSV.
b
Adj. R2 > 95%.
the average Ni2P particle sizes on Ni2P/MZSM-5-CA-0, Ni2P/
MZSM-5-CA-1, Ni2P/MZSM-5-CA-2, and Ni2P/MZSM-5-CA-3 are
8.3, 7.1, 4.5, and 4.3 nm (Table 2), respectively. The smaller Ni2P
particles could expose more active sites, supported by the increas-
ing CO uptakes (Table 2), which improves the catalytic activity.
Because the Ni2P/MZSM-5-CA-2 and Ni2P/MZSM-5-CA-3 catalysts
have analogous Ni2P particle size distributions and CO uptakes,
the reaction rate constants on the two catalysts are at similar
levels.
The catalytic activity of these catalysts was also compared using
the TOF based on the active site, which was measured by CO
chemisorption (Table 4). It is found that the TOF value is gradually
increased with decreasing particle size of the Ni2P phase, although
there was measurement error in the TOF values (Fig. S8 in the Sup-
plementary Material). Oyama and co-workers reported that the
4,6-DMDBT HDS reaction is structure-sensitive [22], and the Ni2P
particle size can be decreased by using a high-surface-area sup-
port, leading to the a higher TOF in the HDS of the 4,6-DMDBT
[25]. Similar results were also reported by Cecilia et al. [38], and
the TOF is increased with decreasing Ni2P particle size that was
obtained by changing metal loadings. However, for the bulk metal
phosphide catalysts, Smith et al. [50] reported the TOF is increased
with increasing particle size of CoxNi2P catalysts in the HDS of the
4,6-DMDBT. They explain that the reason for this phenomenon is
related to the surface acidity induced by the incomplete reduction
of phosphate species [50]. In a subsequent study for the bulk Nix-
MoP catalyst, the TOF is increased with an increase in the particle
size of NixMoP [33]. The authors explain that the electron density
of the Mo in NixMoP is increased with increasing Ni content and
facilitates the dissociation of the 4,6-DMDBT molecule, leading to
the high HDS activity [47]. These results imply that the TOF in
the HDS of the 4,6-DMDBT for the metal phosphide catalysts
should be affected by many factors, such as the particle size, the
metal electron density, and the surface properties.
The product selectivity in the HDS of the 4,6-DMDBT reaction
on these Ni2P/MZSM-5-CA-X (X = 0, 1, and 2) catalysts was com-
pared at the low and similar conversion level (about 20%, Table 5).
In general, the 4,6-DMDBT molecule undergoes HDS via direct
desulfurization (DDS) and hydrogenation (HYD) by two parallel
reaction pathways [51–53]. From Fig. S9 in the Supplementary
Material, DM-BP is 3,30-dimethylbiphenyl from the HDS of 4,6-
DMDBT by the direct desulfurization (DDS) pathway; DM-CHB
and DM-BCH are 3,30-dimethylcycohexylbenzene and 3,30-
dimethyl bicyclohexyl, respectively, by the hydrogenation (HYD)
pathway. Some hydrogenated sulfur-containing intermediates,
such as 4,6-dimethyltetrahydrodibenzothiophene (DM-TH-DBT),
4,6-dimethylhexahydrodibenzothiophene (DM-HH-DBT), and 4,6-
dimethylperhydrodibenzothiophene (DM-PH-DBT), can be formed
in the HYD pathway [51]. As can be seen in Table 5, DM-CHB and
DM-BCH are produced in greater proportions than DM-BP, indicat-
ing that the HYD pathway is preferred for all employed catalysts.
This is in line with the results presented by Prins [14] and Oyama
50
45
40
35
30
25
20
0
2
4
6
8
10
12
14
16
Time on Stream (h)
Fig. 14. 4,6-DMDBT conversion as a function of reaction time for the (j) Ni2P/
MZSM-5-CA-0, (d) Ni2P/MZSM-5-CA-1, (N) Ni2P/MZSM-5-CA-2, and (.) Ni2P/
MZSM-5-CA-3 catalysts (reaction conditions: 0.05 g catalyst, total pressure 5.0 MPa,
temperature of 300 °C, H2 flow of 60 mL minꢁ1, WHSV of 75.0 hꢁ1, 0.5 wt.% 4,6-
DMDBT in decalin).

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
305
Table 5
Product selectivity of 4, 6-DMDBT over Ni2P/MZSM-5-CA-X (X = 0, 1, and 2) catalysts.
Conv. (%)a
Product selectivity (%)
b
DMBCHþDMCHI
DMBCHþDMCHI
Catalyst
DMBP
SHYD
DMBP
DM THDBT
DM HHDBT
DM PHDBT
DM BCH
DM CHB
Ni2P/MZSM-5-CA-0
Ni2P/MZSM-5-CA-1
Ni2P/MZSM-5-CA-2
20.1
20.7
19.4
4.2
3.6
3.4
41.1
44.0
43.7
9.5
7.1
5.7
9.5
8.3
8.2
10.5
8.3
9.0
25.3
28.6
29.9
8.5
10.3
11.4
0.59
0.62
0.68
a
4,6-DMDBT conversion.
b
The SHYD is the total selectivity of DM-THDBT, DM-HHDBT, and DM-PHDBT.
[25], who reported that 4,6-DMDBT desulfurization mainly takes
place through the HYD route due to the steric hindrance induced
by the methyl groups at the 4 and 6 positions.
100
80
60
40
20
0
Moreover, the ratio of HYD to DDS ((DMBCH + DMCHB)/DMBP)
is increased following the order: Ni2P/MZSM-5-CA-0 (8.5) < Ni2P/
MZSM-5-CA-1 (10.3) < Ni2P/MZSM-5-CA-2 (11.4), which should
be associated with the difference in Ni2P particle size. The Ni2P par-
ticles with small size benefit the HYD pathway in the HDS of the
4,6-DMDBT. Oyama et al. have proposed that two types of Ni sites,
tetrahedral Ni(1) and square pyramidal Ni (2) sites, exist in the
Ni2P particles, and the Ni(1) sites are responsible for the direct
desulfurization, while the Ni(2) sites are highly active for the
hydrogenation [25]. As the Ni2P crystallite size goes down, the
number of Ni(2) sites is increased, resulting in high HYD selectivity
of the Ni2P catalyst with small size [25]. In this work, the particle
size of the prepared Ni2P active phase on Ni2P/MZSM-5-CA-0
(8.3 nm), Ni2P/MZSM-5-CA-1 (7.1 nm), and Ni2P/MZSM-5-CA-2
(4.5 nm) gradually decreased. Thus, the Ni2P particles with small
size facilitate HYD.
Furthermore, the (DMBCH + DMCHB)/SHYD ratio was increased
with the decreased Ni2P particle size (Table 5), suggesting that C–S
bond breaking occurred more easily on the small Ni2P particles. In
addition, small Ni2P particles may lead to the adsorption of 4,6-
DMDBT with less steric hindrance [22]. In this manner, the
adsorbed molecules could swing around to cause interaction of
the phenyl rings with the surface, resulting in hydrogenation and
subsequent desulfurization [22]. On the other hand, Smith and
co-workers reported that when the bulk Ni2P catalyst was pre-
pared in the presence of CA, the DDS pathway was accelerated with
the decrease in the Ni2P particle size [33,34]. They explained that
the Brønsted acidity from the incomplete reduction of phosphate
salts can promote the migration of methyl groups to give isomer-
ized products such as 2,8-DMDBT. The isomerized products with
less steric hindrance could easily adsorb onto the small Ni2P parti-
4
8
12 16 20 24 28 32 36 40 44 48 52 56
Time on Stream (h)
Fig. 15. Phenanthrene conversions as a function of reaction time for (j) Ni2P/
MZSM-5-CA-0, (d) Ni2P/MZSM-5-CA-1, (N) Ni2P/MZSM-5-CA-2, and (.) Ni2P/
MZSM-5-CA-3 catalysts.
the phenanthrene can first be hydrogenated to dihydrophenan-
threne (DHP) and 1,2,3,4-tetrahydrophenanthrene (THP), and fur-
ther deeply hydrogenated to octahydrophenanthrene (OHP) with
symmetric and asymmetric and perhydrophenanthrene (PHP,
Fig. S10). OHP and PHP are usually considered to be deep hydro-
genation products (DHPS). The DHPS selectivity follows the order
Ni2P/MZSM-5-CA-1 (29.9%) > Ni2P/MZSM-5-CA-0 (28.8%) > Ni2P/
MZSM-5-CA-2 (26.1%). The DHPS selectivity may depend not only
on the Ni2P particle size but also on the acidity of the catalyst. The
Ni2P/MZSM-5-CA-0, Ni2P/MZSM-5-CA-1, and Ni2P/MZSM-5-CA-2
catalysts have Ni2P particles sizes 8.3, 7.1, and 4.5 nm, respectively,
while their acidic site concentrations are 160, 143, and 120 lmol/g.
cle via
r adsorption, leading to the facilitation of the DDS pathway.
As a result, both the Ni2P particle size and the surface properties
affect the contribution of the HYD and DDS pathways to 4,6-
DMDBT conversion.
4. Conclusions
To further investigate the hydrogenation ability of the Ni2P/
MZSM-5-CA-X (X = 1, 2 and 3) catalysts, phenanthrene hydrogena-
tion was performed. Fig. 15 shows the dependence of phenan-
threne conversion over Ni2P/MZSM-5-CA-X (X = 0, 1, 2, and 3)
catalysts on the reaction time. It is clear that the phenanthrene
conversion is increased with the increase in the amount of CA for
these catalysts. For example, the phenanthrene conversion over
Ni2P/MZSM-5-CA-0 is 63.2% at 54 h, much lower than that over
Ni2P/MZSM-5-CA-1, Ni2P/MZSM-5-CA-2, and Ni2P/MZSM-5-CA-3
(83.6, 90.1, and 90.4%). These results indicate that increasing the
CA content in the catalyst impregnation solution could improve
the prepared Ni2P active phase with high dispersion and small par-
ticles size, leading to the Ni2P particle exposing more hydrogena-
tion active sites, as discussed above. The product selectivity of
phenanthrene hydrogenation was analyzed at a similar conversion
level, and the results are shown in Table S2 in the Supplementary
Material. According to the literature [54–56], the aromatic ring of
Small and highly dispersed Ni2P nanoparticles supported on
mesoporous zeolite ZSM-5 (MZSM-5) are successfully synthesized
by a CA-assisted incipient wetness impregnation method. Ni can be
coordinated with CA in the impregnation solution to form Ni(II)–
CA complexes with different structure, Ni(cit)ꢁ and Ni(Hcit)(cit)3ꢁ
,
by changing the CA/Ni ratios. After impregnation and drying, the
highly viscous Ni(II)–CA complex may favor the good dispersion
of the Ni precursor on the MZSM-5 surface. In addition, the differ-
ent Ni(II)–CA structures, Ni(cit)ꢁ and Ni(Hcit)(cit)3ꢁ, led to differ-
ent interaction between Ni species and the MZSM-5 support. In
the calcination process, the released Ni species can interact
strongly with MZSM-5 through surface acidic hydroxyl groups,
leading to the formation of small Ni2P particles with good disper-
sion under the reduction atmosphere. Compared with the Ni2P cat-
alyst prepared without CA, CA-assisted Ni2P catalysts show high
activity in the 4,6-DMDBT HDS and phenanthrene hydrogenation
reactions.

306
L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
[22] Y. Shu, Y.K. Lee, S.T. Oyama, Structure-sensitivity of hydrodesulfurization of
Acknowledgments
4,6-dimethyldibenzothiophene over silica-supported nickel phosphide
catalysts, J. Catal. 236 (2005) 112.
This work was supported by the National Natural Science Foun-
dation of China (21336008 and U1463203) and the Natural Science
Foundation of Zhejiang Province of China (LZ14B060003). The
authors thank the anonymous reviewers for the invaluable advices.
[23] Y.S. Kim, G.N. Yun, Y.K. Lee, Novel Ni2P/zeolite catalysts for naphthalene
hydrocracking to BTX, Catal. Commun. 45 (2014) 133.
[24] S. Tian, X. Li, A.J. Wang, R. Prins, Y.Y. Chen, Y.K. Hu, Facile preparation of Ni2P
with
a sulfur-containing surface layer by low-temperature reduction of
Ni2P2S6, Angew. Chem. Int. Ed. 55 (2016) 4030.
[25] S.T. Oyama, Y.K. Lee, The active site of nickel phosphide catalysts for the
hydrodesulfurization of 4,6-DMDBT, J. Catal. 258 (2008) 393.
[26] A. Cho, A. Takagaki, R. Kikuchi, S.T. Oyama, Active sites in Ni2P/USY catalysts
for the hydrodeoxygenation of 2-methyltetrahydrofuran, Top. Catal. 58 (2015)
219.
[27] H. Song, J. Wang, Z.D. Wang, H.L. Song, F. Li, Z.S. Jin, Effect of titanium content
on dibenzothiophene HDS performance over Ni2P/Ti-MCM-41 catalyst, J. Catal.
311 (2014) 257.
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.2016.11.019.
[28] A.I. d’Aquino, S.J. Danforth, T.R. Clinkingbeard, B. Ilic, L. Pullan, M.A. Reynolds,
B.D. Murray, M.E. Bussell, Highly-active nickel phosphide hydrotreating
catalysts prepared in situ using nickel hypophosphite precursors, J. Catal.
335 (2016) 204.
[29] T.I. Korányi, Z. Vít, D.G. Poduval, R. Ryoo, H.S. Kim, Emiel J.M. Hensen, SBA-
15-supported nickel phosphide hydrotreating catalysts, J. Catal. 253 (2008)
119.
[30] K.K. Bando, Y. Koike, T. Kawai, G. Tateno, S.T. Oyama, Y. Inada, M. Nomura, K.
Asakura, Quick X-ray absorption fine structure studies on the activation
process of Ni2P supported on K-USY, J. Phys. Chem. C 115 (2011) 7466.
[31] L. Zhang, W.Q. Fu, Q.Y. Yu, T.D. Tang, Y.C. Zhao, H.W. Zhao, Y.D. Li, Ni2P clusters
on zeolite nanosheet assemblies with high activity and good stability in the
hydrodesulfurization of 4,6-dimethyldibenzothiophene, J. Catal. 338 (2016)
210.
[32] N. Jiang, X.W. Xu, H.L. Song, H. Song, F.Y. Zhang, Effect of citric acid on the
hydrodesulfurization performance of unsupported nickel phosphide, Ind. Eng.
Chem. Res. 55 (2016) 555.
References
[1] R.G. Tailleur, Deep hydrotreating diesel fractions for low-sulfur fuel production
using gas-phase and trickle bed reactors systems operating in parallel, Catal.
Today 220–222 (2014) 133.
[2] N. Bui, C. Geantet, G. Berhault, Maleic acid, an efficient additive for the
activation of regenerated CoMo/Al2O3 hydrotreating catalysts, J. Catal. 330
(2015) 374.
[3] J.N. Díaz de León, T.A. Zepeda, G. Alonso-Nuñez, D.H. Galván, B. Pawelec, S.
Fuentes, Insight of 1D
ultra-deep hydrodesulfurization catalyst, J. Catal. 321 (2015) 51.
[4] T.E. Klimova, D. Valencia, J.A. Mendoza-Nieto, P. Hernández-Hipólito, Behavior
of NiMo/SBA-15 catalysts prepared with citric acid in simultaneous
hydrodesulfurization
dimethyldibenzothiophene, J. Catal. 304 (2013) 29.
[5] P.A. Nikulshin, V.A. Salnikov, A.V. Mozhaev, P.P. Minaev, V.M. Kogan, A.A.
Pimerzin, Relationship between active phase morphology and catalytic
properties of the carbon-alumina-supported Co(Ni)Mo catalysts in HDS and
HYD reactions, J. Catal. 309 (2014) 386.
c-Al2O3 nanorods decoration by NiWS nanoslabs in
of
dibenzothiophene
and
4,6-
[33] R. Wang, K.J. Smith, Hydrodesulfurization of 4,6-dimethyldibenzothiophene
over high surface area metal phosphides, Appl. Catal. A Gen. 361 (2009)
18.
[6] A. López-Benítez, G. Berhault, A. Guevara-Lara, Addition of manganese to
alumina and its influence on the formation of supported NiMo catalysts for
dibenzothiophene hydrodesulfurization application, J. Catal. 344 (2016) 59.
[7] S.F. Shan, P. Yuan, W. Han, G. Shi, X.J. Bao, Supported NiW catalysts with
tunable size and morphology of active phases for highly selective
hydrodesulfurization of fluid catalytic cracking naphtha, J. Catal. 330 (2015)
288.
[8] L. Zhang, W.Q. Fu, M. Xiang, W.C. Wang, M.Y. He, T.D. Tang, MgO nanosheet
assemblies supported CoMo catalyst with high activity in hydrodesulfurization
of dibenzothiophene, Ind. Eng. Chem. Res. 54 (2015) 5580.
[34] R. Wang, K.J. Smith, The effect of preparation conditions on the properties of
high-surface area Ni2P catalysts, Appl. Catal. A Gen. 380 (2010) 149.
[35] R. Takahashi, S. Sato, T. Sodesawa, M. Suzuki, N. Ichikuni, Ni/SiO2 prepared by
sol-gel process using citric acid, Micropor. Mesopor. Mater. 66 (2003) 197.
[36] I. Puigdomenech, Medusa Software, Chemical Equilibrium Diagrams, 2010.
(http://www.inorg.kth.se/medusa).
[37] H.S. Fogler, Elements of Chemical Reaction Engineering, third ed., Prentice Hall,
Michigan, 2004.
[38] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, A
novel method for preparing an active nickel phosphide catalyst for HDS of
dibenzothiophene, J. Catal. 263 (2009) 4.
[39] Y.K. Lee, S.T. Oyama, Bifunctional nature of a SiO2-supported Ni2P catalyst for
hydrotreating: EXAFS and FTIR studies, J. Catal. 239 (2006) 376.
[40] T.I. Korányi, É. Pfeifer, J. Mihály, K. Föttinger, Infrared spectroscopic
investigation of CO adsorption on SBA-15- and KIT-6-supported nickel
phosphide hydrotreating catalysts, J. Phys. Chem. A 112 (2008) 5126.
[41] K. Cheng, L. Zhang, J. Kang, X. Peng, Q. Zhang, Y. Wang, Selective
transformation of syngas into gasoline-range hydrocarbons over mesoporous
H-ZSM-5-supported cobalt nanoparticles, Chem. Eur. J. 21 (2015) 1928.
[42] S.J. Sawhill, K.A. Layman, D.R.V. Wyk, M.H. Engelhard, C. Wang, M.E. Bussell,
Thiophene hydrodesulfurization over nickel phosphide catalysts: effect of the
precursor composition and support, J. Catal. 231 (2005) 300.
[43] H.D. Wu, A.J. Duan, Z. Zhao, D.H. Qi, J.M. Li, B. Liu, G.Y. Jiang, J. Liu, Y.C. Wei, X.
Zhang, Preparation of NiMo/KIT-6 hydrodesulfurization catalysts with tunable
sulfidation and dispersion degrees of active phase by addition of citric acid as
chelating agent, Fuel 130 (2014) 203.
[44] A. Jentys, G. Warecka, M. Derewinski, J.A. Lercher, Adsorption of water on ZSM-
5 zeolites, J. Phys. Chem. 93 (1989) 4837.
[45] O.V. Klimov, A.V. Pashigreva, M.A. Fedotov, D.I. Kochubey, Y.A. Chesalov, G.A.
Bukhtiyarova, A.S. Noskov, Co-Mo catalysts for ultra-deep HDS of diesel fuels
prepared via synthesis of bimetallic surface compounds, J. Mol. Catal. A Chem.
322 (2010) 80.
[46] E. Metwalli, M. Karabulut, D.L. Sidebottom, M.M. Morsi, R.K. Brow, Properties
and structure of copper ultraphosphate glasses, J. Non-Cryst. Solids 344 (2004)
128.
[47] J.A. Rodriguez, J.Y. Kim, J.C. Hanson, S.J. Sawhill, M.E. Bussell, Physical and
chemical properties of MoP, Ni2P, and MoNiP hydrodesulfurization catalysts:
time-resolved X-ray diffraction, density functional, and hydrodesulfurization
activity studies, J. Phys. Chem. B 107 (2003) 6276.
[48] S. Badoga, K.C. Mouli, K.K. Soni, A.K. Dalai, J. Adjaye, Beneficial influence of
EDTA on the structure and catalytic properties of sulfided NiMo/SBA-15
catalysts for hydrotreating of light gas oil, Appl. Catal. B Environ. 125 (2012)
67.
[49] V.M.L. Whiffen, K.J. Smith, S.K. Straus, The influence of citric acid on the
synthesis and activity of high surface area MoP for the hydrodeoxygenation of
4-methylphenol, Appl. Catal. A Gen. 419–420 (2012) 111.
[50] I.I. Abu, K.J. Smith, The effect of cobalt addition to bulk MoP and Ni2P catalysts
for the hydrodesulfurization of 4,6-dimethyldibenzothiophene, J. Catal. 241
(2006) 356.
[9] X. Li, Y. Chai, B. Liu, H. Liu, J. Li, R. Zhao, C. Liu, Hydrodesulfurization of 4,6-
dimethyldibenzothiophene over CoMo catalysts supported on
different morphology, Ind. Eng. Chem. Res. 53 (2014) 9665.
c-Alumina with
[10] 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,
Extraordinarily high activity in the hydrodesulfurization of 4,6-
dimethyldibenzothiophene over Pd supported on mesoporous zeolite Y, J.
Am. Chem. Soc. 133 (2011) 15346.
[11] Y.K. Lee, Y.Y. Shu, S.T. Oyama, Active phase of a nickel phosphide (Ni2P)
catalyst supported on KUSY zeolite for the hydrodesulfurization of 4,6-
DMDBT, Appl. Catal. A: Gen. 322 (2007) 191.
[12] L. Zhang, W.Q. Fu, Q.P. Ke, S. Zhang, H.L. Jin, J.B. Hu, S. Wang, T.D. Tang, Study of
hydrodesulfurization of 4,6-DM-DBT over Pd supported on mesoporous USY
zeolite, Appl. Catal. A Gen. 433–434 (2012) 251.
[13] W.Q. Fu, L. Zhang, D.F. Wu, M. Xiang, Q. Zhuo, K. Huang, Z.D. Tao, T.D. Tang,
Mesoporous zeolite-supported metal sulfide catalysts with high activities in
the deep hydrogenation of phenanthrene, J. Catal. 330 (2015) 423.
[14] Y.Y. Sun, R. Prins, Hydrodesulfurization of 4,6-dimethyldibenzothiophene over
noble metals supported on mesoporous zeolites, Angew. Chem. Int. Ed. 47
(2008) 8478.
[15] T.D. Tang, C.Y. Yin, L.F. Wang, Y. Ji, F.S. Xiao, Good sulfur tolerance of a
mesoporous Beta zeolite-supported palladium catalyst in the deep
hydrogenation of aromatics, J. Catal. 257 (2008) 125.
[16] A. Niquille-Röthlisberger, R. Prins, Hydrodesulfurization of 4,6-
dimethyldibenzothiophene and dibenzothiophene over alumina-supported
Pt, Pd, and Pt-Pd catalysts, J. Catal. 242 (2006) 207.
ˇ
ˇ
[17] Z. Vít, D. Gulková, L. Kaluza, J. Kupcík, Pd-Pt catalysts on mesoporous SiO2-
Al2O3 with superior activity for HDS of 4,6-dimethyldibenzothiophene: effect
of metal loading and support composition, Appl. Catal. B Environ. 179 (2015)
44.
[18] Y. Sun, H. Wang, R. Prins, Hydrodesulfurization with classic Co-MoS2 and Ni-
MoS2/c-Al2O3 and new Pt-Pd on mesoporous zeolite catalysts, Catal. Today
150 (2010) 213.
[19] S.T. Oyama, X. Wang, F.G. Requejo, T. Sato, Y. Yoshimura, Hydrodesulfurization
of petroleum feedstocks with a new type of nonsulfide hydrotreating catalyst,
J. Catal. 209 (2002) 1.
[20] R. Prins, M.E. Bussell, Metal phosphides: preparation, characterization and
catalytic reactivity, Catal. Lett. 142 (2012) 1413.
[21] N. Koike, S. Hosokai, A. Takagaki, S. Nishimura, R. Kikuchi, K. Ebitani, Y. Suzuki,
S.T. Oyama, Upgrading of pyrolysis bio-oil using nickel phosphide catalysts, J.
Catal. 333 (2016) 115.

L. Zhang et al. / Journal of Catalysis 345 (2017) 295–307
307
[51] T.D. Tang, L. Zhang, W.Q. Fu, Y.L. Ma, J. Xu, J. Jiang, G.Y. Fang, F.S. Xiao, Design
and synthesis of metal sulfide catalysts supported on zeolite nanofiber
bundles with unprecedented hydrodesulfurization activities, J. Am. Chem.
Soc. 135 (2013) 11437.
[52] H.D. Wu, A.J. Duan, Z. Zhao, T.S. Li, R. Prins, X.F. Zhou, Synthesis of NiMo
hydrodesulfurization catalyst supported on a composite of nano-sized ZSM-5
zeolite enwrapped with mesoporous KIT-6 material and its high isomerization
selectivity, J. Catal. 317 (2014) 303.
[54] S.C. Korre, M.T. Klein, R.J. Quann, Polynuclear aromatic hydrocarbons
hydrogenation. 1. Experimental reaction pathways and kinetics, Ind. Eng.
Chem. Res. 34 (1995) 101.
[55] H.B. Yang, Y.C. Wang, H.B. Jiang, H.X. Weng, F. Liu, M.F. Li, Kinetics of
phenanthrene hydrogenation system over CoMo/Al2O3 catalyst, Ind. Eng.
Chem. Res. 53 (2014) 12264.
[56] W.Q. Fu, L. Zhang, D.F. Wu, Q.Y. Yu, T. Tang, T.D. Tang, Mesoporous zeolite
ZSM-5 supported Ni2P catalysts with high activity in the hydrogenation of
phenanthrene and 4,6-Dimethyldibenzothiophene, Ind. Eng. Chem. Res. 55
(2016) 7085.
[53] M. Lewandowski, Hydrotreating activity of bulk NiB alloy in model reaction of
hydrodesulfurization 4,6-dimethyldibenzothiophene, Appl. Catal. B: Environ.
160–161 (2014) 10.