A solution-phase synthesis of supported Ni2P catalysts with high activity for hydrodesulfurization of dibenzothiophene

Article abstract of DOI:10.1016/j.molcata.2014.01.019

A novel solution-phase low temperature synthesis of a supported Ni ₂P catalyst is described. This uses nickel acetylacetonate (Ni(acac)₂) and low cost triphenylphosphine (TPP) in the presence of the coordinating solvent tri-n-octylamine (TOA) to form Ni₂P/MCM-41 catalysts. The catalysts were characterized by X-ray diffraction (XRD), N ₂-adsorption specific surface area measurements (BET), CO uptake, transmission electron microscopy (TEM), Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS). The effects of the synthesis temperature, and initial P/Ni molar ratio on the formation of the Ni ₂P phase were studied. The formation of the crystalline phase follows the order Ni < Ni₁₂P₅ < Ni₂P with increasing synthesis temperature. The high initial P/Ni molar ratio favored the formation and dispersion of the Ni₂P phase, and a mixture of the Ni₁₂P₅ and Ni₂P phases was observed in the sample with the low initial P/Ni molar ratio of 0.5. The dibenzothiophene (DBT) hydrodesulfurization (HDS) activity increased with reaction temperature in all cases. The catalysts synthesized at 603 K with an initial P/Ni molar ratio of 6 exhibited good DBT HDS activity. The DBT conversion was close to 100% at 613 K, which was much higher than that of the catalyst prepared by the conventional temperature programmed reduction (TPR) method (DBT conversion 43%). During the HDS reaction, the superficial phosphosulfide phase Ni_xP _yS_z was formed on the surface of the catalyst; this showed higher activity than Ni₂P.

Full text of DOI:10.1016/j.molcata.2014.01.019

Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
A solution-phase synthesis of supported Ni2P catalysts with high
activity for hydrodesulfurization of dibenzothiophene
Hua Songa,b, Min Daic, Hua-Lin Songd,∗, Xia Wanc, Xiao-Wei Xua, Zai-Shun Jind
a College of Chemistry and Chemical Engineering, Northeast Petroleum University, Daqing 163318, China
b Provincial Key Laboratory of Oil & Gas Chemical Technology, College of Chemistry & Chemical Engineering, Northeast Petroleum University, Daqing
163318, Heilongjing, China
c Xinjiang Petroleum Investigation Design and Research Institute (Co., Ltd.), Karamay 834000, Xinjiang, China
d Key Laboratory of Cancer Prevention and Treatment of Heilongjiang Province, Mudanjiang Medical University, Mudanjiang 157011, Heilongjing, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 October 2013
Received in revised form 23 January 2014
Accepted 25 January 2014
Available online 3 February 2014
A novel solution-phase low temperature synthesis of a supported Ni2P catalyst is described. This uses
nickel acetylacetonate (Ni(acac)2) and low cost triphenylphosphine (TPP) in the presence of the coordi-
nating solvent tri-n-octylamine (TOA) to form Ni2P/MCM-41 catalysts. The catalysts were characterized
by X-ray diffraction (XRD), N2-adsorption specific surface area measurements (BET), CO uptake, transmis-
sion electron microscopy (TEM), Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron
spectroscopy (XPS). The effects of the synthesis temperature, and initial P/Ni molar ratio on the formation
of the Ni2P phase were studied. The formation of the crystalline phase follows the order Ni < Ni12P5 < Ni2P
with increasing synthesis temperature. The high initial P/Ni molar ratio favored the formation and dis-
persion of the Ni2P phase, and a mixture of the Ni12P5 and Ni2P phases was observed in the sample with
the low initial P/Ni molar ratio of 0.5. The dibenzothiophene (DBT) hydrodesulfurization (HDS) activity
increased with reaction temperature in all cases. The catalysts synthesized at 603 K with an initial P/Ni
molar ratio of 6 exhibited good DBT HDS activity. The DBT conversion was close to 100% at 613 K, which
was much higher than that of the catalyst prepared by the conventional temperature programmed reduc-
tion (TPR) method (DBT conversion 43%). During the HDS reaction, the superficial phosphosulfide phase
NixPySz was formed on the surface of the catalyst; this showed higher activity than Ni2P.
Keywords:
Nickel phosphide
Hydrodesulfurization
Dibenzothiophene
Triphenylphosphine
Nickel acetylacetonate
© 2014 Published by Elsevier B.V.
1. Introduction
promoted by cobalt or nickel [1–3], as well as at finding novel cat-
alyst materials. Transition metal phosphides (i.e. MoP, WP, Fe2P,
Driven by the growing demand for fuel oil and the increasing
use of heavy crude oil, sulfur removal from fuels has attracted a
great deal of attention, and will continue to do so in the coming
years. Moreover, more and more rigorous environmental regula-
tions limiting the emissions of sulfur dioxide and the continuing
decline in the quality of petroleum feedstocks have made sulfur
removal a serious problem in the refining industry. It has been
recognized that current commercial hydrodesulfurization (HDS)
catalysts are not adequate to meet the regulated levels, and the
challenge of producing ultra clean fuels has stimulated the investi-
gation and development of high-performance HDS catalysts. Much
of the research over the past decade has been aimed at improv-
ing the classic catalysts, which are based on molybdenum sulfide
Ni2P) have been found quite promising as alternative hydrodesul-
furization catalysts because of their high HDS activity and stability
[4–6]. Ni2P is a transition metal phosphide (TMP) that has phys-
ical properties, such as hardness and strength, close to ceramics,
yet has the electrical properties of metals, such as its conductiv-
ity and Hall coefficient [5,7–9]. The catalyst Ni2P shows the most
remarkable activities of all the phosphides and has attracted con-
siderable attention for its widespread applications in HDS research
[5–7,10,11].
All recent reports of Ni2P catalysts have described catalyst
production by temperature programmed reduction (TPR) of a
mixture of the metal nickel salt with ammonium phosphate in
hydrogen [4,12,13]. The nickel phosphate precursor is prepared
mainly by impregnation of the support with (NH4)2HPO4 (or
NH4H2PO4) and Ni(NO3)2 solutions, followed by drying and cal-
cination. After a subsequent temperature programmed reduction
in hydrogen, the desired Ni2P catalyst is formed. The method of
temperature programmed reduction is convenient and simple, but
has the disadvantage of requiring a high reduction temperature.
Berhault et al. [14] had observed that the reduction of phosphate
∗
Corresponding author at: Key Laboratory of Cancer Prevention and Treatment
of Heilongjiang Province, Mudanjiang Medical University, Mudanjiang 157011, Hei-
longjing, China. Fax: +86 0453 6984321.
E-mail addresses: songhua2004@sina.com (H. Song), songhualin401@126.com
(H.-L. Song).
1381-1169/$ – see front matter © 2014 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.molcata.2014.01.019

150
H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
to phosphide did not start until the temperature reached 823 K
and led to the selective formation of the Ni2P phase at 923 K. The
(The heating and cooling processes were carried out in a continu-
ous N2 flow). The obtained dark suspension was passivated in O2/N2
mixture (0.5 vol.% of O2) with flowrate of 20 mL/min for 2 h under
vigorous stirring, and then treated with 50 mL of ethanol to form
black precipitate, which was isolated by filtering, washed fully with
ethanol and carbon tetrachloride. The whole experiment process
was operated in atmospheric pressure. The catalysts obtained were
named Ni-P(X)/MCM, where X is the initial P/Ni molar ratio.
The structural property and catalytic activity of the Ni2P cat-
alysts prepared by solution-phase synthesis was compared with
the catalyst prepared by conventional TPR method. For the TPR
method, the catalyst precursors were prepared by a standard incipi-
ent wetness impregnation of an ammonium dihydrogen phosphate
(NH4H2PO4) and nickel nitrate hexahydrate (Ni(NO3)2·6H2O) solu-
tion with MCM-41. The impregnated solids were dried at 393 K
for 10 h and calcined at 773 K for 3 h, and then directly reduced
in a fixed-bed reactor by heating to 973 K at a rate of 2 K min−1 in
a flow of H2 (200 mL/min), held for 2 h, then naturally cooled to
room temperature in a continuous H2 flow. The obtained catalyst
was passivated in O2/N2 mixture (0.5 vol.% of O2) with flowrate
of 20 mL/min for 2 h. Oyama et al. [31] have synthesized a series
of SiO2-supported nickel phosphide catalysts with different initial
Ni/P ratios by TPR, and found the catalyst with initial Ni/P ratio
of about 1/2 (actual Ni/P = 1/0.57 after reaction) corresponded to
a pure Ni2P phase with an excellent activity. Hence, in our exper-
iment, the Ni metal loading of the catalyst prepared by TPR was
determined as 12 wt.% and initial P/Ni molar ratio was determined
as 2. The catalysts obtained were named Ni-P(2)/MCM-T.
P
O bond is strong, and its reduction requires high temperature.
Moreover, hydrogen atoms are available only after the metal-
lic nickel particles have been formed and dissociate hydrogen
molecules to hydrogen atoms [15]. These active hydrogen atoms
can spill over to the phosphate and reduce it to phosphorus or
phosphine. The phosphorus or phosphine can then react with
the metal particles to form Ni2P. The strong P O bond and the
surface diffusion of the H atoms are responsible for the high
reduction temperature, which leads to large catalyst particles,
relatively low catalytic activity and the almost exclusive formation
of phosphorus [15]. It is therefore necessary to explore novel
approaches featuring mild conditions for preparing the Ni2P cat-
alyst. Recently, there have been reports of some new approaches
to preparing Ni2P such as solvothermal reactions [16,17], thermal
decomposition of nickel thiophosphate (NiPS3) [18], reduction
of nickel dihydrogenphosphite [19], thermal decomposition of
a mixture of trioctylphosphine (TOP) with metal salts or metal
hypophosphites precursors in solution-phase, plasma methods
[20], and co-reaction of metal or metal oxide with phosphines
[21,22]. Of all these methods, the solution-phase process, which
uses trioctylphosphine (TOP) and tris(trimethylsilyl)phosphine as
phosphorus reagents, is particularly noteworthy [23–29]. How-
ever, the high price of these two phosphorus reagents limits the
use of these methods to prepare Ni2P catalysts, and most current
reports about the solution-phase synthesis of Ni2P have been
limited to theoretical research into nanophase materials. Hence,
it is essential to find a low cost phosphorus reagent to replace the
TOP and tris(trimethylsilyl)phosphine and to synthesize Ni2P for
HDS research using a solution-phase process.
2.2. Characterization of catalysts
In this paper, we demonstrate a low cost solution-phase method
for preparing supported Ni2P catalyst using mild conditions. The
method uses nickel acetylacetonate (Ni(acac)2) as a nickel pre-
cursor, the low-price triphenylphosphine (TPP) as a phosphorus
precursor, and tri-n-octylamine (TOA) as a coordinating solvent.
The catalysts were synthesized at 603 K in a N2 atmosphere. By
comparison, preparation of the catalyst using the conventional TPR
method requires a reduction temperature exceeding 873 K so these
conditions are extremely mild, yet the DBT conversion is high. The
phosphorus precursor and TOA used in our approach are cheaper
than tri-n-octylphosphine (TOP) and oleylamine (OA) which have
been widely used in the current solution-phase approach for the
synthesis of nanocrystalline Ni2P. The effects of the P/Ni molar ratio
and synthesis temperature on the formation of Ni2P phase were
studied. The results show that a highly dispersed and active Ni2P
catalyst can be obtained.
X-ray diffraction (XRD) analysis of the samples were carried
out on a D/max-2200PC-X-ray diffractometer using CuK␣ radia-
tion under the setting conditions of 40 kV, 30 mA, scan range from
10 to 80◦ at a rate of 10◦/min.
The typical physico-chemical properties of supports and cata-
lysts were analyzed by BET method using Micromeritics adsorption
equipment of NOVA2000e. All the samples were outgassed at 473 K
until the vacuum pressure was 6 mm Hg. The adsorption isotherms
for nitrogen were measured at 77 K.
The CO uptake was measured using pulsed chemisorption.
About 1.0 g of catalyst was pretreated in a quartz reactor to remove
the passivation layer by heating up to 773 K at a rate of 2 K min−1
in H2 with flowrate of 20 mL/min for 2 h, and then naturally cooled
to room temperature in a continuous H2 flow and an He flow at
30 mL/min was used to flush the catalyst for 30 min to achieve
an adsorbate-free. After pretreatment, 1-mL pulses of CO were
injected into a flow of He (30 mL/min), and the CO uptake was
measured using a TCD. CO pulses were repeatedly injected until
the response from the detector showed no further CO uptake after
consecutive injections. Assuming a 1:1 adsorption stoichiometry
between CO and metal atoms, this value corresponds to the metal
site density on the catalyst surface.
2. Experimental
2.1. Synthesis of catalysts
Siliceous MCM-41 support was synthesized using tetraethyl
orthosilicate (TEOS) as the silica source and cetyltrimethylammo-
nium bromide (CTAB) as the template, following the procedure as
described in the literature [30].
Transmission electron microscopy (TEM) studies were per-
formed using the JEM-1010 instrument supplied by JEOL. The
samples were dispersed in ethanol and placed on a carbon grid
before TEM examinations.
The synthesis process of supported Ni2P catalysts were carried
out in an organic solution at a mild temperature under N2 atmo-
sphere (99.99%). The catalysts, with different initial P/Ni molar
ratios of 0.5, 2, 6, and 10, were prepared with Ni metal loading of
12 wt.%. In a typical synthesis technique, a mixture of TOA (40 mL,
98%), Ni(acac)2 (0.26 g, 99%), TPP (0.55 g, 99%), and MCM-41 (0.85 g,
99.9%) were stirred magnetically at 393 K for 30 min (the initial
molar ratio of P/Ni is 2), and then heated up to 603 K at a rate of
10 K min−1, held for 2 h, then naturally cooled to room temperature
Infrared absorption spectroscopy (IR) was acquired in the range
of 400–4000 cm−1. The solid specimens were generated via the con-
ventional procedure, i.e. by the pelletizing 1.5 mg of the catalyst
with 500 mg of KBr.
The X-ray photoelectron spectroscopy (XPS) spectra were
acquired using ESCALAB MKII spectrometer. XPS measurements
have been performed using monochromatic Mg K␣ radiation
(E = 1253.6 eV) and equipped with a hemi-spherical analyzer oper-
ating at fixed pass energy of 40 eV. The recorded photoelectron

H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
151
binding energies were referenced against the C 1s contamination
line at 284.8 eV, and the depth of the data collected was 1–3 nm.
2.3. Catalytic activities
The HDS of DBT was carried out in a high-pressure fixed-bed
reactor (9 mm in diameter, and 300 mm in length) using a feed
consisting of a decalin solution of DBT (1 wt.%). Catalytic activities
were measured at different temperatures (553–613 K), 3.0 MPa of
H2, hydrogen/oil ratio of 500 (v/v) and weight hourly space veloci-
ties (WHSV) of 4 h−1. The catalyst was pressed in discs, crushed and
sieved with 30–60 mesh. Prior to reaction, 0.5 g of the catalysts were
pretreated in situ with flowing H2 (40 mL/min) at 613 K for 2 h. The
evolution of the reaction was studied by collecting liquid samples
after 2 h at the desired reaction temperature. The liquid samples
were collected every hour and analyzed by FID gas chromatography
with a GC-14C-60 column.
Turnover frequency (TOF) values of the samples containing
nickel phosphide were calculated using Eq. (1) [32]:
F
X
TOF =
(1)
W M
where F is the molar rate of DBT fed into the reactor (␮mol s−1), W
is the weight of catalyst (g), X is the conversion of DBT (%), and M
is the mole of sites loaded.
3. Results and discussion
Fig. 2. XRD patterns of the fresh and spent MCM-41 supported samples with differ-
ent initial P/Ni molar ratio. The lower and upper lines are fresh and spent samples,
respectively.
3.1. Characterization of catalysts
3.1.1. XRD
In order to study the effect of reaction temperature on the forma-
tion of the catalyst crystalline phases, the XRD patterns of MCM-41
supported samples with an initial P/Ni molar ratio of 6, obtained at
different synthesis temperatures, were carried out, with the results
shown in Fig. 1. It can be seen that from 393 K to 483 K, no crystalline
phase is observed except the MCM-41 support, with a broad peak at
2ꢀ = 25◦. This means that the Ni and P species exist in noncrystalline
forms over this synthesis temperature range. At a synthesis tem-
perature of 513 K, two diffraction peaks, at 2ꢀ = 44.4◦ and 51.8◦, can
be observed, corresponding to the Ni metal phase (PDF 65-2865).
When the synthesis temperature was increased to 543 K, the Ni
phase disappeared, and the Ni12P5 phase was observed, with peaks
at 2ꢀ = 32.6◦, 38.4◦, 41.7◦, 44.4◦, 46.9◦ and 48.9◦ (PDF 22-1190). Sub-
sequently, the Ni2P phase is observed as the synthesis temperature
is increased to 573 K. This produces weak peaks at 2ꢀ = 40.6◦, 44.5◦,
47.1◦, 54.1◦, 54.8◦, 66.1◦, 72.5◦ and 74.5◦ (PDF 03-0953). At 603 K,
some strong diffraction peaks of the Ni2P phase are detected, which
indicates that more Ni2P phase is formed and the crystallinity of the
Ni2P is improved with increasing synthesis temperature. The XRD
analysis results show that, with the increase of synthesis tempera-
ture, the crystalline phases appeared follows the order Ni, Ni12P5,
and Ni2P. Ni and Ni12P5 are the reactive intermediates leading to
Ni2P, and, in comparison with Ni2P phase, Ni and Ni12P5 might form
under milder conditions, whereas the conversion to Ni2P requires
higher temperatures.
The initial P/Ni molar ratio is known to control the Ni2P particle
size and is the key factor in determining its structure [31,32]. TPP
can work as a reducing agent for the Ni precursor, as well as a source
of phosphorus through cleavage of P C bonds [33]. In order to study
the effect of initial P/Ni molar ratio on the solution-phase synthesis,
XRD patterns of catalysts with initial P/Ni molar ratios in the range
0.5–10 have been investigated, and the results are shown in Fig. 2
(dark lines in each frame). Ni12P5 and Ni2P phases are observed in
the sample with an initial P/Ni molar ratio of 0.5 (Fig. 2(a)), while the
samples with higher initial P/Ni molar ratios exhibit only the Ni2P
phase. The formation of Ni12P5 with a low proportion of P is under-
standable, as Ni12P5 contains a lower proportion of P than does
Ni2P. The Ni2P peaks become weaker with the increase of initial
P/Ni molar ratio. The peak width indicates the formation of smaller
crystallites of Ni2P with the increase in P content (see Table 1), and
it is a likely consequence of a better dispersion of the Ni2P particles,
Fig. 1. XRD patterns of MCM-41 supported samples with initial P/Ni molar ratio of
6 and obtained at different synthesis temperature.

152
H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
Table 1
Properties of catalyst samples.
Sample
SBET (m2 g−1
)
Vp (cm3 g−1
)
dp (av) (nm)
XRD phase
Dc (nm)c
CO uptake (␮mol g−1
)
Metal site densityd
(␮mol g−1
)
MCM-41
1012
712
–
690
–
668
–
613
–
0.816
0.503
–
0.472
–
0.456
–
0.417
–
3.2
2.8
–
2.7
–
2.7
–
2.7
–
–
–
156
101
97
78
124
83
132
86
43
39
–
Ni-P(0.5)/MCMa
Ni-P(0.5)/MCMb
Ni-P(2)/MCMa
Ni-P(2)/MCMb
Ni-P(6)/MCMa
Ni-P(6)/MCMb
Ni-P(10)/MCMa
Ni-P(10)/MCMb
Ni-P(2)/MCM-Ta
Ni-P(2)/MCM-Tb
Ni2P + Ni12P5
Ni12P5
Ni2P
Ni2P
Ni2P
Ni2P
Ni2P
Ni2P
Ni2P
3(Ni12P5)
2227(Ni12P5)
278
223
186
742
223
1114
202
124
111
8
10
12
5
10
4
11
18
20
398
–
0.239
–
2.4
–
Ni2P
a
Fresh sample.
Spent sample.
b
c
Calculated from the Dc = Kꢁ/ˇ cos(ꢀ) (Scherrer equation) base on the Ni2P or Ni12P5 {1 1 1}.
d
Calculated from the L = Sn¯ f , where f is the fractional weight loading of the sample (gnickel phosphide/gcat.) and n¯ is the average surface metal atom density, which for Ni2P
is 1.01 × 1015 atoms cm−2, S is the effective surface area of the crystallites (assuming cubic or spherical geometry) and which can be calculated from the S = 6/ꢂDc (ꢂ is the
density of nickel phosphides taken as 7.09 g cm−1) [31].
which are easier to form with higher P contents. The reason for this
phenomenon can be understood since the Ni component is able
to disperse itself better during the reaction stage when there are
larger amounts of TPP on the surface. These results demonstrate
that the P/Ni molar ratio has a significant influence on the nature of
the phase formed. Ni2P cannot be formed with the low P/Ni molar
ratio of 0.5, possibly because a proportion of the P species could
not participate in the reduction of the Ni precursor or because
some P C bonds did not break. The crystallite size calculated from
Scherrer’s equation is a mean value and is also listed in Table 1.
XRD patterns were also obtained for the spent samples (Fig. 2,
red lines in each frame). For all the samples, the Ni12P5 and Ni2P
phase became more crystalline after the hydrotreatment reaction.
This is understandable since there is prolonged exposure at the
hydrotreatment temperature of 613 K, indicating that the smaller
Ni12P5 and Ni2P particles are not stable at this temperature. For
the sample with an initial Ni/P ratio of 0.5, the XRD pattern after
reaction shows that the Ni2P phase has completely disappeared.
It is likely that this sample is phosphorus deficient, and the pro-
longed exposure at 613 K has caused the Ni2P phase to turn into the
Ni12P5 phase. It is also possible that the sample loses some phos-
phorus during the hydrotreatment reaction. In addition, no other
new phase can be observed in the spent samples, implying that the
main catalytic framework of Ni12P5 and Ni2P remains unchanged.
Ni(acac)2 fragment and is relatively weakly bound to the surface,
since its vibration modes are not affected by the proximity of metal
[34].
3.1.3. Textural properties
Table 1 summarizes the textural properties of Ni-P(X)/MCM and
Ni-P(2)/MCM-T catalysts, as well as that of the MCM-41 support.
The surface area and pore volume of the MCM-41 support are
1012 m2 g−1 and 0.816 cm3 g−1, respectively. It can be seen that
all of the catalysts suffer a decrease in the specific surface areas
and pore volumes after metal phosphide formation. It is also note-
worthy that the average pore diameter of the catalysts is smaller
than that of the MCM-41 support. The decrease of the average pore
diameter may be due to the deposition of Ni species on the surface
of MCM-41 support. The BET surface areas of all the Ni-P(X)/MCM
catalysts ranged from 613 to 712 m2 g−1, a relatively small variation
with initial P/Ni molar ratio, and are higher than that of the cata-
lyst prepared by the TPR method. This implies that the initial P/Ni
3.1.2. FT-IR
In order to study the formation of Ni2P catalysts, FT-IR character-
izations of catalysts with an initial P/Ni molar ratio of 6 synthesized
at different temperature were performed (Fig. 3). The FT-IR spec-
trum of every sample shows a broad strong signal above 1080 cm−1
,
which can be assigned to the Ni bound P O stretching band, a nat-
ural consequence of surface oxidation resulting from handling of
samples in the open air. This observation is similar to the results
reported by Carenco et al. [25] and Senevirathne et al. [34]. Three
peaks at 1450–1600 cm−1 are observed in the catalysts obtained
when the synthesis temperature is below 513 K. These peaks are
attributed to the symmetric stretching vibrations of C C in the
fundamental ring of the aromatics derived from coordinated TPP.
However, these peaks do not appear in the catalysts obtained
at synthesis temperatures higher than 513 K. This indicates that
the P C bond of TPP is broken at about 513 K, and the aromatic
products formed are washed out by ethanol and carbon tetrachlo-
ride. Moreover, peaks from the C H stretching vibrations can be
observed at 2925 and 2852 cm−1, which result from an aliphatic
chain remaining on the surface of the particles after washing. This
organic residue probably results from the reaction of TOA with the
Fig. 3. FTIR spectra for the catalysts with initial P/Ni molar of 6 and synthetized at
393–603 K.

H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
153
Fig. 4. Low and high resolution TEM micrographs of the catalysts. (a) Ni-P(0.5)/MCM, (b) Ni-P(2)/MCM, (c) Ni-P(6)/MCM, (d) Ni-P(10)/MCM, (e) Ni-P(2)/MCM-T, (f) high
resolution TEM micrographs of Ni-P(6)/MCM.
molar ratio has little effect on the textural properties prepared by
solution-phase synthesis and any that is observed may be due to the
excess P species that can be washed away by ethanol and carbon
tetrachloride. However, our previous study found that the textu-
ral properties of catalysts prepared by TPR were closely associated
with the P content in the precursor, because the excess P evolved
during the reduction covers the outer surface of the catalysts, lead-
ing to a significant decrease in the surface areas and pore volumes
[35]. Oyama et al. [31] also reported that high P loadings resulted
in low surface areas of the supported Ni2P catalysts obtained by
the TPR method. Furthermore, Abu et al. [36] found that the partial
structure of MCM-41 is destroyed during the preparation of metal
phosphide by the TPR method, resulting in the metal phosphide
supported MCM-41 having much lower surface areas than pristine
MCM-41. Clearly, high surface areas and pore volumes of MCM-
41-supported Ni2P catalysts are more readily available using our
proposed solution-phase synthesis.
but their numbers are very small and they can be ignored [37].
Of all the fresh Ni-P(X)/MCM catalysts, the Ni-P(0.5)/MCM cata-
lyst shows the highest value of CO uptake, which implies that the
mixture phases of Ni12P5 and Ni2P have a better dispersion than
that of the pure Ni2P phase. This result is similar to that reported
by Korányi et al. [38]. Moreover, it should be noted that the CO
uptakes of spent Ni-P(X)/MCM catalysts show a dramatic reduction
compared to the fresh samples. This indicates that the dispersion of
the Ni-P(X)/MCM catalysts has dropped during the HDS reaction.
The crystallite size obtained from XRD also shows that the parti-
cles in the Ni-P(X)/MCM catalysts become bigger. Although the CO
uptakes of spent Ni-P(X)/MCM catalysts suffer a decrease, their val-
ues are still higher than that of Ni-P(2)/MCM-T. It can also be seen
that the CO uptake values are below the theoretical surface metal
site density calculated from the average particle size determined
from the XRD line-broadening of the catalysts. This indicates that
some metal sites may be blocked by excess phosphorus, since the
surfaces of all the catalysts are rich in phosphorus. This will also be
discussed with the XPS analysis
3.1.4. CO uptake
Metal phosphide can be titrated by CO chemisorption at room
temperature [5]. CO uptake measurements were used to titrate the
surface metal atoms and to provide an estimate of the metal sites
and dispersion on the catalysts (assuming a stoichiometry of one
CO molecule per surface Ni atom) [31]. The measured CO adsorp-
tion capacities are presented in Table 1. It is clear that fresh samples
prepared by the proposed solution-phase synthesis show high val-
ues of CO uptake, at least two times higher than that of the catalyst
prepared by TPR method. This is attributed to the smaller particles
of Ni-P(X)/MCM (as shown in Table 1), and also indicates a better
dispersion of Ni2P particles in the Ni-P(X)/MCM catalysts. Signifi-
cantly lower CO uptake of the Ni-P(2)/MCM-T catalyst is observed
and is possibly caused by the poor dispersion of Ni2P particles in
this catalyst or by having its surface blocked by phosphates, which
reduces the amounts of CO adsorption. This will be discussed later
with the XPS results. CO molecules may also be adsorbed on P sites,
3.1.5. TEM
Distinct differences in the morphologies and particle sizes of
the catalysts can be seen in the TEM images (Fig. 4). Clearly, unlike
the typical stacked morphologies of the Mo and W sulfides, the
nickel phosphide is not layered and forms spherical particles that
can be well dispersed on supports [39]. It can be seen that the TEM
images of catalysts prepared by the proposed solution-phase syn-
thesis (Fig. 4(a)–(d)) exhibit similar particle size distribution, while
the catalyst prepared by the TPR method exhibits a larger particle
size distribution (Fig. 4(e)). This indicates the catalysts prepared by
the solution-phase synthesis present a good dispersion of nanopar-
ticles on the MCM-41 support. Of all the catalysts prepared by the
solution-phase synthesis, the Ni-P(0.5)/MCM catalyst, containing
predominantly Ni12P5 with a minor amount of Ni2P, shows the
smallest particle size, in the range 1–2 nm, and appears to be the

154
H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
most homogeneous. This indicates that the Ni12P5 phase has bet-
ter dispersion than that of the Ni2P phase, a feature which is also
confirmed by the CO uptake results in Table 1. Furthermore, the
morphology of the Ni2P particles is found to be highly dependent
on the initial P content. Fig. 4(b) and (c) shows that the diameter
of the Ni2P particles decreases from 8–12 nm to 3–4 nm as the P/Ni
molar ratio increases from 2 to 6 and may be because the higher
P content helps to disperse the Ni species. With increasing initial
P/Ni molar ratio, up to 10, the diameter of the Ni2P particles did not
change much (Fig. 4(d)). Fig. 4(f) shows a lattice spacing of about
0.22 nm, consistent with the d-spacing value for the {1 1 1} crys-
tallographic planes of the Ni2P phase. The particles derived from
TEM and those calculated from XRD are in reasonable agreement
(Table 1).
3.1.6. XPS
In order to gain further insight into the properties of the surfaces
of fresh and spent catalysts, the XPS technique was employed. XPS
spectra in the Ni(2p) and P(2p) regions for the fresh and spent cat-
alysts are shown in Fig. 5, and the binding energies are presented
Fig. 5. XPS spectra in the Ni(2p), P(2p) and S(2p) regions for the fresh and spent catalysts.

H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
155
Table 2
XPS binding energies and superficial concentrations (X = Ni + P + S) for the fresh and spent catalysts.
Sample
Binding energy (eV)
Ni 2p3/2
Superficial atomic ratioc
ı+
Ni
ꢀ
P/Ni
P 2p
P5+
S 2p3/2
Sulfide
Ni/X
P/X
S/X
Ni/Si
Ni
Ni2+
Satellite
Ni␦+
Pı−
Ni-P(0.5)/MCMa
Ni-P(0.5)/MCMb
Ni-P(2)/MCMa
Ni-P(2)/MCMb
Ni-P(6)/MCMa
Ni-P(6)/MCMb
Ni-P(10)/MCMa
Ni-P(10)/MCMb
Ni-P(2)/MCM-Ta
856.4
856.4
856.6
856.6
856.7
856.8
856.9
857.2
856.1
861.7
861.6
861.8
861.8
862.0
862.1
862.3
862.0
860.2
852.7
853.2
852.9
853.0
853.0
853.1
853.2
853.4
852.8
134.3
134.1
134.5
134.0
134.4
134.3
134.1
134.4
134.8
129.3
129.2
130.1
129.2
130.1
130.0
129.9
129.8
129.6
–
161.9
–
161.4
–
161.6
–
161.6
–
0.321
0.222
0.349
0.307
0.402
0.371
0.393
0.386
0.230
1.6
0.8
2.4
2.2
2.3
2.2
2.6
2.3
3.9
0.379
0.316
0.277
0.193
0.303
0.187
0.278
0.177
0.208
0.621
0.263
0.658
0.425
0.697
0.408
0.722
0.412
0.812
–
0.135
0.099
0.174
0.137
0.194
0.104
0.276
0.090
0.578
0.421
–
0.382
–
0.345
–
0.348
–
a
Fresh sample.
Spent sample.
b
c
The atomic ratios estimated from the ratios of total areas of corresponding element (all valencies) peaks which were detected by the XPS analytical instrument.
in Table 2. As shown in Fig. 5(a), all spectra were decomposed,
taking into account the spin-orbital splitting of the Ni 2p3/2 and
Ni 2p1/2 lines (about 17 eV) and the presence of satellite peaks at
about 5 eV higher than the binding energy of the parent signal. This
observation is consistent with that of Korányi et al. [38]. For the
fresh catalysts (Fig. 5(a) and (b)), bands at 852.7–853.4 eV and at
129.3–130.1 eV are observed, which can be assigned to the Ni␦+
and P␦− species, respectively. The Ni␦+ binding energies are slightly
higher than reported values for Ni metal (852.5–852.9 eV) and
lower than those reported for Ni in NiO (853.5–854.1 eV), indicat-
ing that the Ni in Ni2P and Ni12P5 bears a partial positive charge [6].
The P␦− binding energies are below the value reported for elemen-
tal phosphorus (130.2 eV), consistent with the transfer of electron
density from Ni to P in unsupported and silica-supported Ni2P or
Ni12P5 so that the phosphorus bears a partial negative charge [6].
The bands at 856.1–857.2 eV and at 134.0–134.5 eV are assigned to
the Ni2+ and P5+ (phosphate) species, respectively. The P5+ binding
energy is in good agreement with a value reported for P (134.3 eV)
in Ni3(PO4)2 [40], while the Ni2+ binding energies are similar to
the binding energy of those reported for Ni (855.6–856.6 eV) in
Ni(OH)2 [41]. The Ni2+ and P5+ species correspond to the possible
interaction of Ni2+ ions with phosphate ions (PO43−), as a conse-
quence of superficial passivation [35]. Moreover, a broad shake-up
peak at approximately 5.0 eV higher than that of the Ni2+ species
can be observed [31]. These satellite peaks are due to divalent
species [42,43], although, in the literature, they have also been
assigned to trivalent and oxysulfided nickel species [19,44]. Other
broad peaks, centered at higher binding energy, are assigned to the
Ni 2p signal from oxidized Ni [45]. The spent catalysts, obtained
after 12 h HDS reaction, were also studied by XPS and the results
are shown in Fig. 5(c) and (d). In contrast with the correspond-
ing fresh catalyst, the Ni␦+ binding energy of the spent catalyst
shifts to a slightly higher value, while the P␦− shifts to a slightly
lower value. This shift of binding energy implies that the structure
of the surface of catalysts has changed during the HDS reaction.
This surface modification of Ni2P during the HDS reaction has been
regarded as a consequence of the presence of a superficial phos-
phosulfide, with a stoichiometry represented by NixPySz [10,46,47].
The S 2p signal analysis for spent catalysts reflects that all the cat-
alysts possess two sulfur 2p bands located at 161.4–161.9 eV and
168.0–178.5 eV, which are due to sulfide and sulfate species respec-
tively [48]. X. Duan and Cecilia et al. [19,49] have also observed
two sulfur bands on a H2S-passivated Ni2P/MCM-41 catalyst which
was active and stable in DBT HDS reaction. These authors have
assigned the band over 168.0 eV as consequence of the oxidation
of sulfur species retained on the Ni2P/MCM-41 catalyst surface.
Therefore it is reasonable to expect the partial oxidation of the
sulfur species is responsible for the band at 161.4–161.9 eV in
Fig. 5(e).
XPS analyses were also used to calculate the surface atomic
ratios (Table 2). The theoretical P/Ni ratios of the Ni-P(0.5)/MCM,
Ni-P(2)/MCM, Ni-P(6)/MCM, and Ni-P(10)/MCM fresh catalysts
would be 0.4–0.6, assuming that excess phosphorous would be
washed by the ethanol and carbon tetrachloride. However, the P/Ni
ratios are much higher than the theoretical value, the high P/Ni
ratios found for the catalysts confirms that on the surface there is
an enrichment of phosphorous, although it is essentially present
3−
as PO4
forming a passivation layer in the surface of Ni2P and
Ni12P5 particles. The bulk crystal structure of Ni2P belongs to space
¯
group P62m with a = b = 0.5859 nm and c = 0.3382 nm [50] Along
the (0 0 0 1) direction, the bulk structure consists of two differ-
ent alternating stoichiometric planes, namely Ni3P and Ni3P2. The
density functional theory (DFT) studies showed that the Ni3P2 ter-
minated surface is more stable than the Ni3P terminated surface
[8,51], where as STM and PEEM studies showed that a substantial
amount of the Ni2P surface is terminated with the Ni3P structure
[8,52–54]. Hernandez et al. [55] found that the fundamental surface
structure of Ni2P (0 0 0 1) surface is composed of approximately
80% of Ni3P P (Ni3P2 covered with P at the Ni three-fold hollow
sites) and 20% of bare Ni3P2, and the surface was terminated with
P atoms. Hence, it is reasonable to speculate that the phenomenon
rich in phosphorous on surface is the inner characteristic of Ni2P.
Moreover, the XRD also shows that the pure Ni2P phase cannot be
obtained for sample with the initial P/Ni molar ratio of 0.5. The
spent catalysts show a decrease in the P/Ni molar ratios, particu-
larly the Ni-P(0.5)/MCM catalyst, which suffers an important loss in
phosphorous. The P/Ni molar ratios of the spent catalysts (shown in
Table 2) indicate that although some phosphorous have lost during
HDS reaction, the catalysts surface is still enriched in phosphorous.
ꢀ
The Ni 2p spectra of all the catalysts clearly show the Ni␦+
/
Ni is
lower than 0.4, which indicates a much higher proportion of super-
ficial nickel atoms is at oxidation state. The Ni-P(2)/MCM catalysts
show lower Ni/Si atomic ratio than that of Ni-P(2)/MCM-T catalyst,
indicating the catalysts prepared by the solution-phase synthesis
have good dispersion of nickel on support or more Si atoms exposed
at the surface of the solution-phase synthesized catalysts due to the
lower P coverage compared with the catalysts obtained by TPR.
It should be noted that sulfur was detected in the spent cata-
lysts, which may be due to formation of a superficial phosphosulfide
(NixPySz) on the catalysts. Oyama et al. [7] identified a surface phos-
phosulfide phase by XRD and EXAFS on top of a Ni2P core, they
suggested that this may be an active phase of the Ni2P/SiO2 cata-
lysts, and the maximum sulfur level was reckoned to be 5 mol% of
the phosphorus content. Using the DFT calculations, Nelson et al.

156
H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
Fig. 6. HDS conversion and TOF for DBT HDS at different reaction temperature over Ni-P(X)/MCM and Ni-P(2)/MCM-T catalysts (3.0 MPa, WHSV = 4 h−1, H2/oil = 500 (v/v)).
[47] studied the replacement of up to 50% of surface phospho-
rus of the bulk Ni2P by sulfur and suggested that surface Ni3PS
is the actual active phase of Ni2P for the hydrotreatment reactions.
Korányi et al. [38] found a superficial phosphosulfide with a compo-
sition Ni2.4P1.0S0.24 for the catalyst pretreated by H2S. In this paper,
we found the composition of the surface active-phases of spent
the conversion of these catalysts indicates that the reaction occurs
principally on metal centers and the effect of ensemble size on reac-
tivity is modest. Hence, the reaction is structurally insensitive. A
similar observation was reported by Oyama et al. [31], in which the
authors studied the effect of the P/Ni molar ratio on the HDS and
hydrodenitrogenation (HDN), and found a strong dependence for
the HDN of quinoline but only a very weak dependence for the HDS
of DBT.
The DBT conversion and CO uptake of spent catalysts have been
used to calculate the HDS TOF for the catalysts. The value of HDS TOF
for each catalyst increased with temperature (Fig. 6(b)), suggesting
that the DBT HDS over Ni2P and Ni12P5 catalysts is a thermody-
namically favorable reaction. Of all catalysts, the Ni-P(2)/MCM,
Ni-P(6)/MCM and Ni-P(10)/MCM display nearly the same TOF, the
value of which is higher than that of Ni-P(2)/MCM-T. However, a
low TOF was observed for the Ni-P(0.5)/MCM catalyst. This indi-
cates that the metal sites are not directly involved in the HDS
reaction, since the Ni-P(0.5)/MCM catalyst shows the highest CO
uptake value of the catalysts, which means that it has the largest
numbers of metal sites. Indeed, it is accepted that metal sites simply
provide the initial active sites, and the real active phase is a nickel
phosphosulfide produced from the HDS reaction [5,7,31].
With regard to the selectivity of the different reaction prod-
ucts, seven products are detected (as shown in Scheme 1), but the
biphenyl (BP) and cyclohexylbenzene (CHB) are in the majority. The
transformation of DBT occurs through two parallel pathways: (1)
direct desulfurization (DDS), which yields BP and then BP hydro-
genation yields CHB, and (2) desulfurization after hydrogenation
(HYD), which yields tetrahydrodibenzothiophene (TH-DBT) and
then CHB. Because the transformation of BP to CHB is negligible
in the presence of DBT, BP selectivity is used as a measure of DDS
pathway, and CHB represents the HYD pathway. Fig. 7 illustrates
the variations of BP and CHB yields with reaction temperature over
Ni-P(X)/MCM and Ni-P(2)/MCM-T catalysts. For all the samples, the
yield of BP is much higher than that of CHB, indicating that DBT
mainly removed by the DDS pathway over the catalysts, which
can be attributed to the DBT HDS reaction is structure insensitive
[31]. It has been reported a lower hydrogenation activity of phos-
phides, with regards to sulfides, indicating a more effective use of
hydrogen in the HDS reaction [5]. A study using the Ni-Mo-P/SiO2
catalyst underlined the importance of nickel sites in Ni2P in the
catalytic activity, concluding that HDS reaction mainly occurred
on Ni sites, this being one of the properties that makes phosphide
catalysts different from either sulfide, carbide or nitride catalysts
Ni-P(2)/MCM, Ni-P(6)/MCM and Ni-P(10)/MCM were NiP2.2S1.97
,
NiP2.2S1.84 and NiP2.3S1.96, respectively. Considering that our cata-
lysts were pretreated with H2 rather than by H2S, we can conclude
that the high sulfur content in the superficial phosphosulfide phase
may be due to adsorbed sulfur species.
3.2. HDS activity
Fig. 6(a) shows the variations of DBT conversion with
temperature during DBT HDS catalyzed by Ni-P(X)/MCM and Ni-
P(2)/MCM-T catalysts. The DBT HDS conversion of Ni-P(2)/MCM-T
is about 30% at 553 K, and this increases with increasing reaction
temperature. At 613 K, the DBT conversion reached 43%. Clearly, the
DBT conversion using Ni-P(X)/MCM catalysts is dramatically higher
than with Ni-P(2)/MCM-T. The Ni-P(X)/MCM catalysts also show
that the activity increases with increasing reaction temperature;
the Ni-P(6)/MCM and Ni-P(10)/MCM catalysts produce conversion
close to 100% at 613 K. It is noteworthy that the Ni-P(0.5)/MCM
catalyst, containing mainly the Ni12P5 phase, also shows a higher
HDS activity than Ni-P(2)/MCM-T, although it has been reported
that the HDS activity of Ni12P5 catalysts was lower than that of
Ni2P catalysts under the same conditions [55]. The high HDS activ-
ity observed in the Ni-P(X)/MCM catalysts may be due to the
high dispersion of the active phase and low coverage of phospho-
rus on the surface of these catalysts. The CO uptake shows that
the Ni-P(X)/MCM catalysts can adsorb much more CO than Ni-
P(2)/MCM-T, indicating that Ni-P(X)/MCM can offer more exposed
metal sites for HDS reaction. The adsorption states of DBT on the
catalysts are supposed to be of two kinds: one is end-on through
the adsorption of sulfur atoms, the other is flat, involving the ␲-
electrons of DBT molecules [56]. The end-on adsorption orientation
yields the DDS pathway, while the flat orientation leads to the for-
mation of HYD pathway products. DFT calculations indicated that
the sulfur atoms in DBT are positively charged and represent the
main nucleophilic attack center [57]. Hence, a greater number of
metal sites benefits the DBT HDS reaction. The Ni-P(X)/MCM cat-
alysts show a very constant degree of HDS conversion over the
range 513–613 K. The weak effect of the phosphorus content on

H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
157
Scheme 1. The HDS reaction network of DBT over Ni2P catalyst.
[10]. The selectivity to CHB of Ni-P(X)/MCM is higher compared
with that of Ni-P(2)/MCM-T catalysts, which may be attributed to
the smaller active crystal particles formed in Ni-P(X)/MCM cata-
lysts and produce more Ni(2) sites so as to provide higher HYD
selectivity. As reported by the literature [58], there are two types
of sites in Ni2P, Ni(1) sites with tetrahedral coordination and Ni(2)
sites with square pyramidal coordination, both Ni(1) and Ni(2) sites
are present on large crystallites but that the Ni(2) sites are more
numerous on the more highly dispersed samples. The lower coor-
dination Ni(1) sites are responsible for desulfurization by the DDS
pathway by taking on a sulfur atom while the Ni(2) sites are the
high-activity sites that carry out HDS by the hydrogenation (HYD)
route. Hence, we conclude that the acceleration of the HYD path-
way over the Ni-P(X)/MCM catalysts must be associated with the
dispersion of active crystal particles. It should be noted that the
role of the Ni(1) and Ni(2) sites are to provide just the initial state
for the HDS active sites, because the surface of catalyst should be
sulfided and some P atoms are replaced by S atoms during reac-
tion [58]. On the other hand, the selectivity of BP increases with
increasing reaction temperature during DBT HDS over all the cat-
alysts, suggesting that the distribution of HDS products over the
Ni-P(2)/MCM-T and Ni-P(X)/MCM catalysts is sensitive to temper-
ature and the DDS activity can be enhanced by increasing reaction
temperature. The Ni-P(0.5)/MCM catalyst shows the highest CHB
selectivity among the Ni-P(X)/MCM catalysts, indicating the Ni12P5
phase has higher HYD activity than Ni2P phase.
Fig. 8 shows the DBT conversion and yield values versus reaction
time for Ni-P(6)/MCM catalyst. The reaction temperature chosen
was 613 K. It can be seen from Fig. 8 that the activity of catalyst
increases gradually in the initial 10 h, and the conversion of DBT is
close to 99% after 12 h and stays constant for 120 h without activity
loss under the given reaction conditions. The catalyst improves its
catalytic capacity during the initial time on stream, which suggests
that there is a formation of an intermediate phase which is more
active than Ni2P. The surface modification of Ni2P phase during
the HDS reaction and the explanation of its excellent behavior with
initial time on stream have been considered in the literature as con-
sequence of the presence of a superficial phosphosulfide as active
phase with a stoichiometry represented by NiPxSy [10,46,47]. Our
XPS of the spent Ni-P(6)/MCM sample shows a superficial phospho-
sulfide with a composition of NiP2.2S1.84. Further, Ni2P has better
sulfur resistance than metal carbides and nitrides which transform
Fig. 7. BP and CHB selectivities at different reaction temperature over Ni-P(X)/MCM
and Ni-P(2)/MCM-T catalysts (3.0 MPa, WHSV = 4 h−1, H2/oil = 500 (v/v)).

158
H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
temperature, and DBT mainly desulfurizes by the DDS pathway
over the catalysts. A catalyst stability test for Ni-P(6)/MCM over
120 h indicated that the Ni2P catalyst is highly sulfur resistant. The
conversion of DBT was maintained at 99% throughout the test time
with model fuel which contents DBT as high as 1 wt.% at 613 K,
3.0 MPa, WHSV = 4 h−1 and H2/oil = 500 (v/v)).
Acknowledgments
The authors acknowledge the financial supports from the
National Natural Science Foundation of China (21276048) and the
Natural Science Foundation of Heilongjiang Province (ZD201201).
References
[1] R. Prins, Adv. Catal. 46 (2001) 399.
[2] M. Sun, D. Nicosia, R. Prins, Catal. Today 86 (2003) 173.
[3] S.K. Bej, S.K. Maity, U.T. Turaga, Energy Fuels 18 (2004) 1227.
[4] X. Wang, P. Clark, S.T. Oyama, J. Catal. 208 (2002) 321.
[5] S.T. Oyama, J. Catal. 216 (2003) 343.
[6] S.J. Sawhill, D.C. Phillips, M.E. Bussell, J. Catal. 215 (2003) 208.
[7] S.T. Oyama, X. Wang, Y.K. Lee, W.J. Chun, J. Catal. 221 (2004) 263.
[8] P. Liu, J.A. Rodriguez, T. Asakura, J. Gomes, K. Nakamura, J. Phys. Chem. B 109
(2005) 4575.
[9] D. Kanama, S.T. Oyama, S. Otani, D.F. Cox, Surf. Sci. 552 (2004) 8.
[10] F. Sun, W. Wu, Z. Wu, J. Guo, Z. Wei, Y. Yang, Z. Jiang, F. Tian, C. Li, J. Catal. 228
(2004) 298.
[11] Y. Shu, Y.K. Lee, S.T. Oyama, J. Catal. 236 (2005) 112.
[12] J. Guan, Y. Wang, M. Qin, Y. Yang, X. Li, A. Wang, J. Solid State Chem. 182 (2009)
1550.
Fig. 8. DBT conversion and selectivity of BP and CHB as a function of time on stream
for Ni-P(6)/MCM catalyst at 613 K (613 K, 3.0 MPa, WHSV = 4 h−1, H2/oil = 500 (v/v)).
[13] Y. Shu, S. Ted Oyama, Carbon 43 (2005) 1517.
[14] G. Berhault, P. Afanasiev, H. Loboue, C. Geantet, T. Cseri, C. Pichon, C. Guillot
Deudon, A. Lafond, Inorg. Chem. 48 (2009) 2985.
into metal sulfides with high activity under HDS conditions [15].
Fig. 8 shows the conversion of DBT is still more than 99% in the case
of DBT content in feed is as high as 1 wt.%. Although the surface
of Ni2P takes up sulfur atoms under HDS conditions, the kernel of
the Ni2P particles remains intactness [37,59]. The XRD shows there
is no sulfide phase (no distinct peak corresponding to the sulfide
phase is seen in Fig. 2 of the spent catalyst) in the spent catalyst,
which also indicates the kernel of the Ni2P particles is stable. Again,
the selectivity of BP is always higher than CHB during the 120 h,
indicating the DBT HDS over Ni2P catalyst tends toward DDS way.
[15] S.F. Yang, C.H. Liang, R. Prins, J. Catal. 237 (2006) 118.
[16] X. Wang, F. Wan, Y. Gao, J. Liu, K. Jiang, J. Cryst. Growth 310 (2008) 2569.
[17] K. Mi, Y. Ni, J. Hong, J. Phys. Chem. Solids 72 (2011) 1452.
[18] H. Loboué, C. Guillot-Deudon, A.F. Popa, A. Lafond, B. Rebours, C. Pichon, T. Cseri,
G. Berhault, C. Geantet, Catal. Today 130 (2008) 63.
[19] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, J.
Catal. 263 (2009) 4.
[20] 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.
[21] Y. Zhao, M. Xue, M. Cao, J. Shen, Appl. Catal. B-Environ. 104 (2011) 229.
[22] G.J. Shi, J.Y. Shen, J. Mater. Chem. 19 (2009) 2295.
[23] J. Wang, A.C. Johnston-Peck, J.B. Tracy, Chem. Mater. 21 (2009) 4462.
[24] R.K. Chiang, R.T. Chiang, Inorg. Chem. 46 (2007) 369.
[25] K. Senevirathne, A.W. Burns, M.E. Bussell, S.L. Brock, Adv. Funct. Mater. 17
(2007) 3933.
4. Conclusions
[26] K.A. Gregg, S.C. Perera, G. Lawes, S. Shinozaki, S.L. Brock, Chem. Mater. 18 (2006)
879.
[27] C. Qian, F. Kim, L. Ma, F. Tsui, P. Yang, J. Liu, J. Am. Chem. Soc. 126 (2004) 119.
[28] J. Park, B. Koo, K.Y. Yoon, Y. Hwang, M. Kang, J.G. Park, T. Hyeon, J. Am. Chem.
Soc. 127 (2005) 8433.
[29] S.C. Perera, G. Tsoi, L.E. Wenger, S.L. Brock, J. Am. Chem. Soc. 125 (2003) 13960.
[30] H.I. Meléndez-Ortiz, L.A. García-Cerda, Y. Olivares-Maldonado, G. Castruita, J.A.
Mercado-Silva, Y.A. Perera-Mercado, Ceram. Int. 38 (2012) 6353.
[31] S.T. Oyama, X. Wang, Y.-K. Lee, K. Bando, F.G. Requejo, J. Catal. 210 (2002) 207.
[32] V. Teixeira da Silva, L.A. Sousa, R.M. Amorimb, L. Andrini, S.J.A. Figueroa, F.G.
Requejo, F.C. Vicentini, J. Catal. 279 (2011) 88.
A novel low-price solution-phase method for preparing a
supported Ni2P catalyst using mild conditions is reported. This
uses nickel acetylacetonate (Ni(acac)2) as a nickel precursor and
low price triphenylphosphine (TPP) as a phosphorus precursor
to prepare the Ni2P supported on mesoporous MCM-41 catalysts
at 603 K in a N2 atmosphere. The experimental results definitely
indicate that the proposed solution-phase approach is feasible. The
approach does not require such a high reduction temperature as
the TPR method, so providing an alternative way to synthesize Ni2P
catalysts with high HDS activity. XRD shows that as the synthesis
temperature is increased, the formation of the crystalline phase
follows the order Ni, Ni12P5, and Ni2P. The phosphorus content
has a profound effect on the formation of crystalline phases The
precursor with initial P/Ni molar ratio of 0.5 yielded a catalyst
containing predominantly Ni12P5 phase with a little Ni2P, and a
high initial P/Ni molar ratio favors the formation of the Ni2P phase
and its dispersion. The XPS indicates that there is an enrichment of
phosphorus on the surface of the catalysts. During the HDS reac-
tion, a superficial phosphosulfide phase is formed on the surface of
the catalyst, which leads to higher activity than Ni2P. The catalyst
prepared by the proposed solution-phase approach showed higher
activity than the catalyst prepared by a conventional TPR method.
This can be attributed to the better dispersion and lower coverage
of phosphorus over the surface of the catalysts. The DBT HDS
tests showed that the DBT conversion increased with reaction
[33] H.R. Seo, K.S. Cho, Y.K. Lee, Mater. Sci. Eng. B 176 (2011) 132.
´
[34] S. Carenco, C. Boissière, L. Nicole, C. Sanchez, P.L. Floch, N. Mézailles, Chem.
Mater. 22 (2010) 1340.
[35] H. Song, M. Dai, Y.T. Guo, Y.J. Zhang, Fuel Process. Technol. 96 (2012) 228.
[36] I.I. Abu, K.J. Smith, Appl. Catal. A 328 (2007) 58.
[37] K.A. Layman, M.E. Bussell, J. Phys. Chem. B 108 (2004) 10930.
[38] T.I. Korányi, Z. Vít, D.G. Poduval, R. Ryoo, H.S. Kim, E.J.M. Hensen, J. Catal. 253
(2008) 119.
[39] Y.K. Lee, Y. Shu, S.T. Oyama, Appl. Catal. A 322 (2007) 191.
[40] R. Franke, T. Chasse, P. Streubel, A. Meisel, J. Electron. Spectrosc. 56 (1991) 381.
[41] P. Dufresne, E. Payen, J. Grimblot, J.P. Bonnelle, J. Phys. Chem. 85 (1981) 2344.
[42] D. Eliche-Quesada, J. Merida-Robles, P. Maireles-Torres, E. Rodriguez-Castellon,
A. Jimenez-Lopez, Langmuir 19 (2003) 4985.
[43] J.N. Kuhn, N. Lakshminarayanan, U.S. Ozkan, J. Mol. Catal. A-Chem. 282 (2008)
9.
[44] L. Song, S. Zhang, Q. Wei, Catal. Commun. 12 (2011) 1157.
[45] N. Escalona, J. Ojeda, P. Baeza, R. Garcia, J.M. Palacios, J.L.G. Fierro, A.L. Agudo,
F.J. Gil-Llambias, Appl. Catal. A 287 (2005) 47.
[46] T. Kawai, K.K. Bando, Y.K. Lee, S.T. Oyama, W.J. Chun, K. Asakura, J. Catal. 241
(2006) 20.
[47] A.E. Nelson, M. Sun, A.S.M. Junaid, J. Catal. 241 (2006) 180.
[48] L. Qiu, K. Zou, G. Xu, Appl. Surf. Sci. 266 (2013) 230.

H. Song et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 149–159
159
[49] X. Duan, Y. Teng, A. Wang, V.M. Kogan, X. Li, Y. Wang, J. Catal. 261 (2009) 232.
[50] S. Rundqvist, Acta Chem. Scand. 16 (1962) 992.
[55] A.B. Hernandez, H. Ariga, S. Takakusagi, K. Kinoshita, S. Suzuki, S. Otani, S.T.
Oyama, K. Asakura, Chem. Phys. Lett. 513 (2011) 48.
[51] Q. Li, X. Hu, Phys. Rev. B 74 (2006) 035414.
[52] M.G. Moula, S. Suzuki, W.J. Chun, S. Otani, S.T. Oyama, Q. Li, K. Asakura, Chem.
Lett. 35 (2006) 90.
[56] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[57] C.I. Sainz-Díaz, M. Francisco-Márquez, A. Vivier-Bunge, Theor. Chem. Acc. 125
(2010) 83.
[53] S. Suzuki, G.M. Moula, T. Miyamoto, Y. Nakagawa, K. Kinosthita, K. Asakura, S.T.
Oyama, S. Otani, J. Nanosci. Nanotechnol. 9 (2008) 195.
[54] Q. Guan, X. Cheng, R. Li, W. Li, J. Catal. 299 (2013) 1.
[58] S.T. Oyama, Y.K. Lee, J. Catal. 258 (2008) 393.
[59] Z. Wu, F. Sun, W. Wu, Z. Feng, C. Liang, Z. Wei, C. Li, J. Catal. 222 (2004) 41.