A feasible approach to the synthesis of nickel phosphide for hydrodesulfurization

Article abstract of DOI:10.1016/j.jcat.2012.11.008

In this paper, we propose a simple and feasible method for synthesizing bulk and supported nickel phosphides from oxide precursors. The new approach uses a low hydrogen flow speed and is not affected by the heating rate. The results indicate that Ni₂P can be synthesized at 600 °C from its oxide precursors with a mole ratio of Ni/P = 2/1. The hydrodesulfurization activity results indicate that the direct-reduction method shows excellent performance in the synthesis of supported catalysts.

Full text of DOI:10.1016/j.jcat.2012.11.008

Journal of Catalysis 299 (2013) 1–9
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
A feasible approach to the synthesis of nickel phosphide for hydrodesulfurization
⇑
Qingxin Guan, Xun Cheng, Rongguan Li, Wei Li
Key Laboratory of Advanced Energy Materials Chemistry (Ministry of Education), College of Chemistry, Nankai University, Tianjin 300071, China
a r t i c l e i n f o
a b s t r a c t
Article history:
In this paper, we propose a simple and feasible method for synthesizing bulk and supported nickel phos-
phides from oxide precursors. The new approach uses a low hydrogen flow speed and is not affected by
the heating rate. The results indicate that Ni2P can be synthesized at 600 °C from its oxide precursors with
a mole ratio of Ni/P = 2/1. The hydrodesulfurization activity results indicate that the direct-reduction
method shows excellent performance in the synthesis of supported catalysts.
Received 30 July 2012
Revised 2 November 2012
Accepted 7 November 2012
Ó 2012 Elsevier Inc. All rights reserved.
Keywords:
Desulfurization
Hydrogen
Metal phosphide
Nickel
Phosphate
1. Introduction
aided by high temperature and low water vapor pressure [19].
Thus, Ni2P can only be obtained with a low heating rate, excess
Noble metals (e.g., Pt, Pd, Rh, Au) are very useful in many scien-
tific fields because of their excellent performance. However, the
low availability of noble metals severely restricts their wider appli-
cation. Interstitial compounds have attracted considerable atten-
tion in recent decades and have emerged in more and more areas
(e.g., dye-sensitized solar cells, fuel cells, catalysis) as alternative
materials to noble metals [1,2]. Among the various types of inter-
stitial compounds, metal phosphides have definitely emerged as
superstars because of their unexpected properties in catalysis,
magnetic applications, and applications in telecommunications,
electronic and optoelectronic devices, lithium batteries, and solar
cells [1,3]. However, the difficulty of preparing metal phosphides
inhibits their use in large-scale applications.
Metal phosphides were reported as early as the 1930s [4], and
their synthesis and applications underwent tremendous develop-
ment in the following decades. A variety of methods for synthesiz-
ing metal phosphides have been reported [3,5–18]. Generally, the
solvothermal method is not feasible for the deposition of metal
phosphides on supports. The temperature-programmed reduction
(TPR) method has proved to be an effective and general way to pre-
pare bulk and supported metal phosphides [5]. However, the syn-
thesis requires strict temperature-programmed steps and
relatively high temperatures. The conversion of oxide precursors
to phosphides by H2 reduction is neither thermodynamically nor
kinetically favorable [12]. Hence, the forward reaction has to be
phosphorus, and a high H2 flow speed, as has been recognized by
most researchers in recent decades. However, based on the above
analysis, we could not reach the conclusion that the strict temper-
ature-programmed steps are indispensable. In our recent studies,
we performed a new examination of the reduction of metal phos-
phate precursors and obtained some unexpected results.
2. Experimental methods
Siliceous MCM-41 was purchased from Tianjin Chemist Scien-
tific Ltd., China. It had a specific surface area of 956 m2 gꢀ1, a pore
volume of 1.1 cm3 gꢀ1, and a BJH average pore size of 3.32 nm.
Dibenzothiophene (DBT) was of analytical pure grade and pur-
chased from Alfa Aesar. Other reagents were analytical pure grade
and purchased from Tianjin Guangfu Fine Chemical Research Insti-
tute, China.
2.1. Synthesis of bulk and supported nickel phosphides
The synthesis of nickel phosphide involves two main steps: (1)
the oxidic precursor is obtained by coimpregnation with an
aqueous solution of diammonium hydrogen phosphate
((NH4)2HPO4) and nickel(II) nitrate (Ni(NO3)2ꢁ6H2O), followed by
drying and calcination; (2) the precursor is converted to nickel
phosphide in flowing H2. In a typical experiment, bulk nickel phos-
phide was obtained as follows. Initially, 14.84 g of Ni(NO3)2ꢁ6H2O
and 3.35 g of (NH4)2HPO4 were dissolved in 27 mL of deionized
water and stirred for 1 h. Then, the slurry was evaporated at
⇑
Corresponding author. Fax: +86 22 23508662.
E-mail address: weili@nankai.edu.cn (W. Li).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.11.008

2
Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
FAo
W COuptake
XA
120 °C for 3 h and calcined at 500 °C for 2 h to obtain the precursor.
Subsequently, 0.6 g of precursor and 10 mL of quartz sand were
loaded into the quartz tube reactor (as a preliminary heating zone).
The precursor materials were then reduced to phosphides in H2
flow (60 mL minꢀ1) at 600 °C for 1 h. Finally, the product was
cooled to ambient temperature under flowing H2 and was passiv-
ated for 1 h under flowing 1% O2/N2. Similarly to the synthesis of
bulk nickel phosphide, supported nickel phosphide was prepared
by reducing the supported oxidic precursor in flowing H2.
TOF ¼
;
ð1Þ
where FAo is the molar rate of reactant fed into the reactor
mol sꢀ1), W is the catalyst weight (g), COuptake is the uptake of
chemisorbed CO (
mol gꢀ1), and XA is the reactant conversion (%).
(l
l
3. Results and discussion
For comparison, bulk and MCM-41-supported Ni2P were also
successfully prepared using the TPR method [20]. The temperature
program was as follows. The flow rate of H2 in the whole procedure
was 250 mL minꢀ1. A heating rate of 5 °C minꢀ1 was used from
room temperature to 400 °C, and temperature was maintained at
400 °C for 1 h. Then, a heating rate of 2 °C minꢀ1 was used from
400 to 600 °C, and temperature was maintained at 600 °C for 3 h.
Subsequently, the product was cooled to ambient temperature un-
der flowing H2 and was passivated for 1 h under flowing 1% O2/N2.
3.1. Synthesis of bulk and supported nickel phosphides
The oxide precursor of 4NiOꢁP2O5 was obtained by calcination of
nickel(II) phosphate at 600 °C. As shown in Fig. 1, the H2 TPR
results indicate that the oxide precursor of 4NiOꢁP2O5 is not a
simple mixture of NiO and P2O5, which is an amorphous com-
pound, and no diffraction peaks were detected in the XRD patterns.
The high temperature in the H2 TPR experiment also indicated that
this reaction is thermodynamically unfavorable and a high temper-
ature is required. However, these data do not demonstrate that a
low heating rate and high H2 flow speed are indispensable.
2.2. Characterization
In this paper, a simple improvement in the reaction process
completely broke the requirement for the strict temperature-pro-
grammed steps, namely the use of a preheating zone and low H2
flow speed. To confirm the method, many experiments were per-
formed using Ni2P as a model compound. As shown in Fig. 2, the
synthesis of Ni2P was not affected by heating rates, and Ni2P could
Powder X-ray diffraction (XRD) was performed on a Bruker D8
focus diffractometer with Cu K
a radiation at 40 kV and 40 mA.
The compositions of the samples were measured using atomic
absorption spectrophotometry (AAS). The transmission electron
microscopy (TEM) images were acquired using a Philips Tecnai
G2 F-20 field emission gun transmission electron microscope. Scan-
ning electron microscopy (SEM) images were acquired using a Tes-
can Vega3 SBH scanning electron microscope. Nitrogen adsorption
was measured with a BEL-Mini adsorption analyzer. The CO chemi-
sorption was performed with Micromeritics Chemisorb 2750 gas-
adsorption equipment. The sample was loaded into a quartz
reactor and pretreated in 10% H2/Ar at 450 °C for 3 h. After cooling
in He, pulses of 10% CO/He in a He carrier (25 mL (NTP) minꢀ1) were
injected at 30 °C through a loop tube. The H2 TPR of samples was
performed with Micromeritics Chemisorb 2750 gas-adsorption
equipment. The sample was loaded into a quartz reactor and pre-
treated in He at 600 °C for 3 h. After cooling in He, the sample
was reduced using 10% H2/Ar as the reducing gas at a heating rate
of 10 °C minꢀ1 up to 950 °C. The gas flow rate was 25 mL minꢀ1
X-ray photoelectron spectra (XPS) were recorded using a Kratos
Axis Ultra DLD spectrometer employing a monochromated Al K
X-ray source (h = 1486.6 eV), hybrid (magnetic/electrostatic)
optics, and a multichannel plate and delay line detector. All XPS
m,
survey spectra were recorded with a pass energy of 80 eV, and
high-resolution spectra with a pass energy of 40 eV. To subtract
the surface charging effect, the C1s peak has been fixed at a binding
energy of 284.6 eV.
.
a
m
Fig. 1. The H2 TPR patterns of different oxide precursors at a heating rate of
spectra were recorded using an aperture slot of 300 ꢂ 700
l
10 °C minꢀ1
.
2.3. Catalytic activity test
The HDS catalytic activities were evaluated using 3000 ppm
DBT in decalin. The HDS reaction was carried out in a fixed-bed
microreactor. The catalyst was pelleted, crushed, and sieved with
20–40 mesh. One gram of the catalyst was diluted with SiO2 to a
volume of 5.0 mL in the reactor. Prior to the reaction, catalysts
were pretreated in situ with flowing H2 (60 mL minꢀ1) for 3 h.
The testing conditions for the HDS reaction were 3 MPa, weight
hourly space velocity (WHSV) = 6 hꢀ1, and H2/oil = 720. Liquid
products were collected every hour after a stabilization period of
6 h. Both feed and products were analyzed with a FuLi 9790 gas
chromatograph equipped with a flame ionization detector and an
OV-101 column. The DBT conversion and turnover frequency
(TOF) were used to evaluate the HDS activity. The TOF was calcu-
lated using the equation
Fig. 2. The XRD patterns of bulk Ni2P prepared from oxide precursors (in a mole
ratio of Ni/P = 2/1) using different heating rates: (a) TPR method steps, (b)
5 °C minꢀ1, (c) 20 °C minꢀ1, (d) was reduced directly at 600 °C.

Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
3
Fig. 3. The XRD patterns of bulk nickel phosphides prepared from oxide precursors in the same quartz tubes: (a) different calcination temperatures, (b) different mole ratios
of Ni/P, (c) different loadings, and (d) different H2 flow speeds.
be synthesized at different heating rates; it was even reduced
directly at 600 °C for 1 h. Furthermore, as shown in Fig. 3, the
impacts of the calcination temperature, mole ratios of Ni/P, load-
ings, and H2 flow speed were studied in the same quartz tube.
The results indicate that Ni2P can be synthesized at 600 °C from
the oxide precursors with a mole ratio of Ni/P = 2/1, and the syn-
thesis conditions (e.g., loadings and H2 flow speed) were very
broad. The calcination temperature and mole ratio of Ni/P are the
key factors in the synthesis. Decreasing the mole ratio of Ni/P leads
to the generation of phosphorus-rich products (e.g., Ni5P4).
Decreasing the calcination temperature leads to the generation of
Ni12P5 from oxide precursors at
a mole ratio of Ni/P = 2/1
(Fig. 3a). According to density functional theory, the larger the
number of d orbitals of Ni that are not bonded to p orbitals of P,
the less stable will be the surface, because the dangling bonds
formed by unbonded d orbitals increase the surface energy [21].
As shown in Fig. 4, the compositions of terminated layers of
Ni2P and Ni12P5 were Ni3P and Ni4P5, respectively. The Ni4P5 struc-
ture has a lower surface energy than the Ni3P structure, which is
more stable under the same conditions. As shown in Table 1
[22], Ni2P has a lower atomic density than Ni12P5, which is more
stable at higher temperatures in the mole ratio Ni/P = 2/1. Hence,
the appearance of Ni12P5 at low temperatures might be determined
by its low surface energy, while the appearance of Ni2P at high
temperatures might be determined by its lower atomic density.
To further study the surfaces of bulk Ni2P and Ni12P5 samples,
the XPS technique was employed. The results show that the super-
ficial atomic ratios of Ni/P are 0.41 and 0.57 for Ni2P and Ni12P5,
respectively. Fig. 5 shows the Ni2p and P2p core level spectra of
bulk Ni2P and Ni12P5. The Ni2p core level spectrum of Ni2P involves
Fig. 4. The structures of Ni2P and Ni12P5 phases show the alternation of differently
terminated layers.
two peaks (Fig. 5a). The first one is assigned to Nid+ of Ni2P and cen-
tered at 853.1 eV, and the second one is at 856.4 eV, corresponding
to Ni2+ ions interacting possibly with phosphate ions as a conse-
quence of a superficial passivation [6]. Meanwhile, the P2p binding
energy of Ni2P was observed at 129.7 eV for Pdꢀ of Ni2P and
133.7 eV for surface nickel phosphate species [6], due to the

4
Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
Table 1
The lattice parameters of Ni2P and Ni12P5.
Metal phosphides
a (Å)
b (Å)
c (Å)
a
(°)
b (°)
c
(°)
Atom number
Unit cell volume (Å3)
Atomic density (Åꢀ3
)
Ni2P
Ni12P5
5.859
8.646
5.859
8.646
3.382
5.07
90
90
90
90
120
90
9
34
100.54
379
8.952 ꢂ 10ꢀ2
8.971 ꢂ 10ꢀ2
Fig. 5. The XPS spectra of bulk Ni2P and Ni12P5 samples: (a) Ni2p core level spectra of Ni2P, (b) P2p core level spectra of Ni2P, (c) Ni2p core level spectra of Ni12P5, and (d) P2p
core level spectra of Ni12P5.
superficial oxidation of Ni2P. Similarly, the binding energy assigned
to Nid+ and Pdꢀ of Ni12P5 is observed at 852.6 and 129.4 eV. Inter-
estingly, the peak of Ni2+ in the Ni2P sample is significantly higher
than that of Ni12P5, which might be caused by its lower atomic
density. As shown in Table 1, Ni2P has a lower atomic density than
Ni12P5, and oxygen atoms more easily penetrate the interior of
crystals in the process of superficial passivation. Hence, Ni2P shows
a higher Ni2+ peak than Ni12P5 in the Ni2p core level spectrum. As
shown in Fig. 5c, the Ni12P5 sample shows slightly lower binding
energy than Ni2P, which might be affected by its lower content
of nickel phosphate.
In considering the parameters, the calcination time is also an
important factor. To exclude the effects of the heating rate, H2
was pumped in and, simultaneously, the time was recorded after
the reactor temperature reached the set value. Fig. 6 shows the
XRD patterns of bulk nickel phosphides prepared from oxide pre-
cursors with different calcination times. The results clearly show
that the product changes from Ni12P5 to Ni2P with time and that
Ni2P can be synthesized in 60 min. This conclusion is also con-
firmed by SEM images, which show visible changes with increasing
calcination time (Fig. 7).
To analyze further the composition of the bulk products, bulk
Ni2P samples that were synthesized in 60, 120, and 180 min were
analyzed using an atomic absorption spectrophotometer. The re-
sults indicate that the nickel content of the three samples is
79.15, 79.14, and 79.17 wt%, respectively, which are close to the
calculated value (79.10 wt%). Thus, the bulk sample that was syn-
thesized in 60 min is the pure Ni2P phase. As shown in Fig. 8, Ni2P
samples supported on MCM-41 with different loadings were pre-
pared by impregnation. The XRD results indicate that the typical
diffraction peaks of Ni2P appeared in 30 wt% Ni2P/MCM-41. The
XRD peaks of Ni2P were not detected at the loadings of 10 and
20 wt%, which might be because the supported Ni2P nanoparticles
were tiny and below the detection limit. For comparison, the XRD
pattern of 30 wt% Ni2P/MCM-41 prepared using the TPR method is
given in Fig. 8, which shows a result similar to those prepared
using the direct-reduction method. This method is simple and does
not require temperature-programmed steps; accordingly, we
named it the direct-reduction (DR) method.
For comparison, the impact of mole ratios of Ni/P was also stud-
ied using the TPR method. Fig. 9 shows the XRD patterns of bulk
nickel phosphide prepared using the TPR method at 600 °C from

Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
5
Fig. 8. The XRD patterns of Ni2P supported on MCM-41 with different loadings.
Fig. 6. The XRD patterns of bulk nickel phosphides prepared from oxide precursors
with different calcination times at a H2 flow speed of 60 mL minꢀ1
.
oxide precursors with different mole ratios of Ni/P. It is seen that
the diffraction peaks of Ni12P5 gradually disappear with decreasing
mole ratios of Ni/P, and that Ni2P was synthesized with a mole ra-
tio of Ni/P = 10/8. The crystal size of bulk Ni2P was calculated with
the Scherrer formula, and the calculated result was compared with
the TEM observed result. Table S1 gives the Scherrer particle sizes
of bulk Ni2P prepared using different methods. The bulk nickel
phosphides prepared using the DR and TPR methods have similar
Fig. 9. The XRD patterns of bulk nickel phosphide prepared using the TPR method
at 600 °C from oxide precursors with different mole ratios of Ni/P.
Fig. 7. The SEM images of samples prepared from oxide precursors with different calcination times at a H2 flow speed of 60 mL minꢀ1
.

6
Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
Table 2
sized using the DR method. The results show that bulk Ni12P5,
Co2P, CoP, MoP, WP, and MoNiP could be synthesized using the
DR method, thus indicating that those metal phosphides synthe-
sized using the TPR method can also be synthesized by the DR
method.
The optimal synthetic conditions for the preparation of different metal phosphide
synthesized from metal phosphate precursors using the direct-reduction method.
Metal
Mole ratios
Temperature
(°C)
H2
Reduction
time (h)
phosphidea
(mL minꢀ1
)
Ni12P5
Ni2P
CoP
Co2P
MoP
WP
Ni:P = 12:5
Ni:P = 2:1
Co:P = 1:1
Co:P = 2:1
Mo:P = 1:1.1
W:P = 1:1.1
600
600
600
600
650
650
60
60
60
60
60
60
60
1
1
1
1
3
3
3
3.2. HDS activity
In our previous studies, the Ni2P/MCM-41 catalyst could be syn-
thesized at lower temperatures using the decomposition of hypo-
phosphites [13,14]. In this paper,
MoNiP
Ni:Mo:P = 1:1:1.1 650
a series of Ni2P/MCM-41
a
catalysts were synthesized at 600 °C from metal phosphate precur-
sors using the DR method. Hence, for comparison, Ni2P/MCM-41
catalysts prepared by the TPR method were synthesized and tested
under the same conditions. To investigate the properties and cata-
lytic activities, Ni2P/MCM-41 catalysts prepared by different meth-
ods were characterized and tested by TEM, nitrogen adsorption, CO
chemisorption, and HDS. The DBT conversion and TOF were used to
evaluate the HDS activity [23]. The TOF was calculated using Eq.
(1). Irreversible CO uptake measurements were used to titrate
the surface metal atoms and to provide an estimate of the active
sites on the catalysts [24]. All of the catalytic activities were tested
using the ninth-hour liquid products after a stabilization period.
As shown in Fig. 11, TEM images of the Ni2P/MCM-41 prepared
by the DR method show fairly uniform nanoparticles, which indi-
cates good dispersion on the MCM-41 support. The TEM images
(Fig. 11a and b) also show that the particle size of Ni2P clearly
The loading of different precursors is 0.6 g in all experiments.
Scherrer particle sizes of about 50 nm. Fig. S1 gives the TEM images
of bulk Ni2P prepared using the DR and TPR methods. The observed
particle size of bulk Ni2P prepared using the DR and TPR methods is
about 50–100 nm, which is larger than the calculated Scherrer par-
ticle size. The difference in the above results might be caused by
the inaccuracy of the Scherrer formula.
To illustrate the versatility of this synthetic method, different
metal phosphides were synthesized. Table 2 gives the optimal syn-
thetic conditions for the preparation of different metal phosphides
using the DR method. For MoP, WP, and MoNiP, excess phosphorus
is required because the high temperature and long calcination time
cause a small amount of phosphorus loss. Fig. 10 shows the XRD
patterns and SEM images of different metal phosphides synthe-
Fig. 10. The XRD patterns and SEM images of different metal phosphides synthesized using the direct-reduction method.

Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
7
Fig. 11. The TEM images of Ni2P/MCM-41 with different loadings.
Fig. 13. The catalytic activities of 10 wt% Ni2P/MCM-41 prepared from the
precursors with different mole ratios of Ni/P (testing conditions: 3 MPa, 310 °C,
WHSV = 6 hꢀ1, H2/oil = 720).
Fig. 12. The catalytic activities of Ni2P/MCM-41 with different loadings (testing
conditions: 3 MPa, WHSV = 6 hꢀ1, H2/oil = 720).
decreased with decreasing loading. A magnified high-resolution
TEM image of a selected frame from Fig. 11c given in Fig. 11d yields
d-spacing values of 0.191 nm for the (210) crystallographic planes
of Ni2P, which is in good agreement with the calculated value. This
indicates the formation and good dispersion of nanocrystalline
Ni2P on the MCM-41 support.
Fig. 12 gives the catalytic activities of Ni2P/MCM-41 with differ-
ent loadings. The result shows that the catalytic activities of Ni2P/
MCM-41 both at 290 and at 310 °C gradually increased with
increasing Ni2P loadings. This result is generally accepted; for
Ni2P/MCM-41, higher temperature favors HDS of DBT. To compare
the catalytic activities of Ni2P/MCM-41 catalysts prepared using
the DR and TPR methods, a series of 10 wt% Ni2P/MCM-41 samples
were prepared from the precursors with different mole ratios of
Ni/P. Subsequently, these catalysts were tested in HDS of DBT at
310 °C. Fig. 13 shows the catalytic activities at 310 °C of 10 wt%
Ni2P/MCM-41 prepared from precursors with different mole ratios
of Ni/P. The results indicate that the highest DBT conversion at
310 °C of 10 wt% Ni2P/MCM-41 is 83% and 84% for the DR method
and TPR method, respectively. Interestingly, the best catalyst was
prepared from the precursors with mole ratios (Ni/P) of 10/5 and
10/8 for the DR method and TPR method, respectively. Comparing
the result in Fig. 13 and the results in Figs. 3b and 9, we see that the
best synthesis conditions for the bulk and supported Ni2P are

8
Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
Table 3
The physical properties and HDS catalytic activities of the 10 wt% Ni2P/MCM-41 catalysts prepared by different methods (testing conditions: 3 MPa, 310 °C, WHSV = 6 hꢀ1, H2/
oil = 720).
Catalysts
BET area (m2 gꢀ1
)
CO uptake (
l
mol gꢀ1
)
Conversion (%)
Selectivity (%)
TOF (10ꢀ3 sꢀ1
)
BP
BCH
CHB
Ni2P/MCM-41 (DR 10/5)
Ni2P/MCM-41 (TPR 10/5)
Ni2P/MCM-41 (TPR 10/8)
621
616
615
20
20
20
83
73
84
85.2
85.3
85.6
7.5
7.6
7.4
7.3
7.1
7.0
1.13
0.99
1.14
that the DDS pathway is favored for DBT HDS over MCM-41-
supported nickel phosphide catalysts.
In addition, different metal phosphides supported on MCM-41
containing 10 wt% of metal were synthesized and denoted as
Ni12P5/MCM, Ni2P/MCM, CoP/MCM, Co2P/MCM, MoP/MCM, WP/
MCM, and MoNiP/MCM. The detailed synthesis conditions for
MCM-41-supported metal phosphides were the same as those for
the bulk metal phosphides. Fig. 14 shows the HDS catalytic activi-
ties of different catalysts. The activities of all MCM-41-supported
metal phosphide catalysts increased with increasing reaction tem-
perature. The result indicates that the activities of the catalysts fol-
lowed the order Ni2P/MCM > Ni12P5/MCM > MoNiP/MCM > MoP/
MCM > WP/MCM > Co2P/MCM > CoP/MCM.
Scheme 1. Reaction pathway of the HDS of DBT on nickel phosphide catalysts.
4. Conclusions
In summary, bulk and supported nickel phosphides were syn-
thesized from phosphate precursors using the DR method. The
new approach requires a low H2 flow speed and is not affected
by the heating rate. The results indicate that Ni2P can be synthe-
sized at 600 °C from the oxide precursors with a mole ratio of Ni/
P = 2/1, and the synthesis conditions (e.g., loadings and H2 flow
speed) were also very broad. The preheating hydrogen, mole ratio
of Ni/P, calcination temperature, and calcination time are all
important factors in the DR method. The experimental results def-
initely indicate that the DR method is simple and feasible, provid-
ing an alternative way to synthesize nickel phosphides. The HDS
activity results indicate that the DR method also shows excellent
performance in the synthesis of catalysts. The high BP selectivity
indicates that the DDS pathway is favored for DBT HDS over
MCM-41-supported nickel phosphide catalysts.
Fig. 14. The HDS catalytic activities of different catalysts (testing conditions: 3 MPa,
WHSV = 6 hꢀ1, H2/oil = 720).
identical. As shown in Fig. 9, bulk Ni12P5 and Ni2P were obtained
from the precursor with a mole ratio (Ni/P) of 10/5, and bulk
Ni2P was obtained from the precursor with a mole ratio (Ni/P) of
10/8. Wang et al. [20] also reported similar results: that tempera-
ture-programmed reduction of the precursor with Ni/P = 10/5
yielded Ni12P5/MCM-41, whereas reduction of the precursor with
Ni/P = 10/8 produced Ni2P/MCM-41. The HDS activity of Ni2P is
much higher than that of Ni12P5. Hence, the catalyst prepared using
the TPR method from the precursor with a mole ratio (Ni/P) of 10/8
shows much better performance than that with a mole ratio (Ni/P)
of 10/5.
Acknowledgments
This work was financially supported by the National NSFC
(Grant 21073098), the Natural Science Foundation of Tianjin
(11JCZDJC21600), the Research Fund for 111 Project (B12015),
the Doctoral Program of Higher Education (20090031110015),
and the Program for New Century Excellent Talents in University
(NCET-10-0481).
Appendix A. Supplementary material
For comparison, the physical properties and HDS catalytic activ-
ities of the 10 wt% Ni2P/MCM-41 catalysts prepared by different
methods are given in Table 3. It is seen that the catalyst prepared
using the TPR method from the precursor with a mole ratio (Ni/P)
of 10/8 shows physical properties and HDS catalytic activity similar
to those of the catalyst prepared using the DR method from the pre-
cursor with a mole ratio (Ni/P) of 10/5, which indicates that the DR
method also shows excellent performance in the synthesis of cata-
lysts. Scheme 1 shows the reaction pathway for the HDS of DBT on
nickel phosphide catalysts. It mainly involves hydrogenation (HYD)
and direct desulfurization (DDS) pathways. The products of DBT
HDS are mainly cyclohexylbenzene (CHB), bicyclohexane (BCH),
and biphenyl (BP) [6,25]. The high BP selectivity (>85%) indicates
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.11.008.
References
[1] S.T. Oyama, J. Catal. 216 (2003) 343.
[2] G. Li, F. Wang, Q. Jiang, X. Gao, P. Shen, Angew. Chem. Int. Ed. 49 (2010) 3653.
[3] K.L. Stamm, J.C. Garno, G. Liu, S.L. Brock, J. Am. Chem. Soc. 125 (2003) 4038.
[4] C.M. Grieb, R.H. Jones, J. Chem. Soc. (1932) 2543.
[5] X. Wang, P. Clark, S.T. Oyama, J. Catal. 208 (2002) 321.
[6] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, J.
Catal. 263 (2009) 4.
[7] Y. Xie, H.L. Su, X.F. Qian, X.M. Liu, Y.T. Qian, J. Solid State Chem. 149 (2000) 88.
[8] S. Yang, C. Liang, R. Prins, J. Catal. 237 (2006) 118.

Q. Guan et al. / Journal of Catalysis 299 (2013) 1–9
9
[9] C. Qian, F. Kim, L. Ma, F. Tsui, P. Yang, J. Liu, J. Am. Chem. Soc. 126 (2004) 1195.
[17] K. Cho, H. Seo, Y. Lee, Catal. Commun. 12 (2011) 470.
[18] G. Shi, J. Shen, J. Mater. Chem. 19 (2009) 2295.
[19] V. Zuzaniuk, R. Prins, J. Catal. 219 (2003) 85.
[20] A. Wang, L. Ruan, Y. Teng, X. Li, M. Lu, J. Ren, Y. Wang, Y. Hu, J. Catal. 229 (2005)
314.
[21] D. Fuks, D. Vingurt, M.V. Landau, M. Herskowitz, J. Phys. Chem. C 114 (2010)
13313.
[22] E. Larsson, Ark. Kem. 23 (1965) 335.
[10] A.E. Henkes, Y. Vasquez, R.E. Schaak, J. Am. Chem. Soc. 129 (2007) 1896.
[11] W.R. Robinson, G.J. van, T.I. Kor, S. Eijsbouts, K.A. van, V.J. van, B.V. de, J. Catal.
161 (1996) 539.
[12] A. Wang, M. Qin, J. Guan, L. Wang, H. Guo, X. Li, Y. Wang, R. Prins, Y. Hu, Angew.
Chem. Int. Ed. 47 (2008) 6052.
[13] Q. Guan, W. Li, M. Zhang, K. Tao, J. Catal. 263 (2009) 1.
[14] Q. Guan, W. Li, J. Catal. 271 (2010) 413.
[15] J. Gopalakrishnan, S. Pandey, K. Kasthuri Rangan, Chem. Mater. 9 (1997) 2113.
[16] V.T. Silva, L.A. Sousa, R.M. Amorim, L. Andrini, S.J.A. Figueroa, F.G. Requejo, F.C.
Vicentini, J. Catal. 279 (2011) 88.
[23] Y.-K. Lee, S.T. Oyama, J. Catal. 239 (2006) 376.
[24] S.T. Oyama, X. Wang, Y.-K. Lee, W.-J. Chun, J. Catal. 221 (2004) 263.
[25] Q. Guan, C. Sun, R. Li, W. Li, Catal. Commun. 14 (2011) 114.