Novel synthesis of dispersed molybdenum and nickel phosphides from thermal carbonization of metal- and phosphorus-containing resins

Article abstract of DOI:10.1039/c5dt02464a

Dispersed pure phases of MoP and Ni₂P nanoparticles supported by carbon were synthesized by carbonization of metal- and phosphorus-containing resins under an inert atmosphere. The solid products and the evolution of gases during the carbonization process were investigated by various techniques, including X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), transmission electron microscopy (TEM), inductively coupled plasma-atomic emission spectroscopy (ICP-AES), N₂ adsorption-desorption analysis, and mass spectrometry (MS). The resins underwent two carbonization stages: the low-temperature carbonization stage (<650 °C) and the high-temperature carbonization stage (≥650 °C). There was an initial reduction of Mo and Ni precursors in the low-temperature region. However, the formation of phosphides was observed in the high-temperature carbonization stage, in which Mo(Ni) and PO_x species were further reacted with the carbonization products (C, H₂ and CH₄) to yield Mo(Ni) phosphide. Note that compared with the traditional H₂-temperature-programmed reduction (H₂-TPR) method, this novel synthesis route produced a large amount of CO_x besides H₂O, leading to a lower water vapor pressure. In addition, the residual carbon produced from resin can play a role in bonding of nanoparticle aggregation. Therefore, the better dispersions and higher surface areas of the as-prepared phosphide nanoparticles were attributed to the mitigation of hydrothermal sintering and the intimate contact between phosphide nanoparticles and carbon species.

Full text of DOI:10.1039/c5dt02464a

Dalton
Transactions
View Article Online
View Journal
PAPER
Novel synthesis of dispersed molybdenum and
nickel phosphides from thermal carbonization
of metal- and phosphorus-containing resins†
Cite this: DOI: 10.1039/c5dt02464a
Zhiwei Yao,* Jin Tong, Xue Qiao, Jun Jiang, Yu Zhao, Dongmei Liu, Yichi Zhang and
Haiyan Wang
Dispersed pure phases of MoP and Ni2P nanoparticles supported by carbon were synthesized by carbon-
ization of metal- and phosphorus-containing resins under an inert atmosphere. The solid products and
the evolution of gases during the carbonization process were investigated by various techniques, includ-
ing X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM),
transmission electron microscopy (TEM), inductively coupled plasma-atomic emission spectroscopy
(ICP-AES), N2 adsorption–desorption analysis, and mass spectrometry (MS). The resins underwent two
carbonization stages: the low-temperature carbonization stage (<650 °C) and the high-temperature
carbonization stage (≥650 °C). There was an initial reduction of Mo and Ni precursors in the low-temperature
region. However, the formation of phosphides was observed in the high-temperature carbonization stage,
in which Mo(Ni) and POx species were further reacted with the carbonization products (C, H2 and CH4) to
yield Mo(Ni) phosphide. Note that compared with the traditional H2-temperature-programmed reduction
(H2-TPR) method, this novel synthesis route produced a large amount of COx besides H2O, leading to a
lower water vapor pressure. In addition, the residual carbon produced from resin can play a role in
bonding of nanoparticle aggregation. Therefore, the better dispersions and higher surface areas of the as-
prepared phosphide nanoparticles were attributed to the mitigation of hydrothermal sintering and the
intimate contact between phosphide nanoparticles and carbon species.
Received 30th June 2015,
Accepted 12th October 2015
DOI: 10.1039/c5dt02464a
www.rsc.org/dalton
dependent catalytic performance of materials with nanoscale
sizes, more and more efforts have been focused on synthetic
1. Introduction
Due to the fact that transition metal interstitial compounds methods for preparing nanosized metal phosphides. During
(e.g. carbides, nitrides and phosphides) show unique perform- the last twenty years, a wide range of methods have been devel-
ances similar to noble metals, they have attracted much atten- oped for the preparation of nanoscale MoP and Ni2P phos-
tion in recent decades and have been regarded as substitutes phides,2 such as direct reduction of phosphate in H2 or H2
for noble metal materials in various fields such as fuel cells, plasma,6,9,10 the solid-state reaction method using white or red
dye-sensitized solar cells and catalysis.1–5 Compared with phosphorus,11,12 reaction of metal/metal oxide with PH3,13
metal carbides and nitrides, metal phosphides have been decomposition of high-boiling point tri-n-octylphosphine
studied more extensively in recent years owing to their unpre- (TOP),14–16 decomposition of single-source precursors in solu-
cedented properties and potential applications in many areas, tion,17 chemical vapor deposition,18 and microwave or solvo-
including catalysis, magnetism, electronics, optoelectronics, thermal synthesis.19,20 Nevertheless, most of these routes
lithium batteries, solar cells and so on.6,7 Particularly in the generally involve highly toxic phosphorus sources (e.g. P4, PH3
aspect of catalysis, MoP and Ni2P phosphides were reported to and TOP) or complicated steps that cannot afford the require-
be excellent hydrotreating catalysts in terms of both activity ments of green synthesis of phosphide catalysts. Only one
and resistance to poisoning.6,8 Since the discovery of the size- method mentioned above, the H2-temperature-programmed
reduction (H2-TPR) method (Scheme 1),9 has been recognized
as an effective and environmentally friendly approach to
College of Chemistry, Chemical Engineering and Environmental Engineering,
prepare bulk and supported MoP and Ni2P catalysts due to the
Liaoning Shihua University, Fushun, 113001, PR China.
use of a nontoxic phosphorus source ((NH4)2HPO4) and a
E-mail: mezhiwei@163.com
green reducing agent (H2). Unfortunately, a large amount of
water generated during the reaction process can result in
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5dt02464a
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Scheme 1
A
simplified equation for the preparation of metal
phosphides by the H2-TPR method.
hydrothermal sintering of MoP and Ni2P nanoparticles.21–26
Recently in our laboratory, highly dispersed phosphides have
been prepared by dimethyl ether (DME) and hexamethyl-
enetetramine (HMT) routes.27,28 However, it is still worthwhile
to expand green preparation techniques for synthesizing dis-
persed nanosized MoP and Ni2P catalysts.
Fig. 1 Chemical structure of D201 × 1 cinnamic strong alkali anion
exchange resin (Cl− type).
Ion exchange resin is composed of a porous polymer
network with tunable functional organic groups. Following a
facile impregnation procedure, one or more ionic guests can
be simultaneously and dispersively introduced into the resin
framework or covered on the surface.29,30 After thermal carbon-
ization at high temperatures, the resin is converted to carbon
that can play a double role as a reducing agent and support for
nanoparticles.30 Based on these unique properties of ion
exchange resin, well dispersed Mo and W carbide particles
(<10 nm) have been prepared from carbonization of metal-
containing resin precursors.31,32 The two studies (ref. 31 and 32)
reported that these dispersed carbides could act as highly
efficient Pt catalyst supports for methanol electro-oxidation.
Note that the electrocatalytic activity of Pt supporting on the
above-mentioned Mo carbide was higher than that of commer-
cial Pt/C catalysts.32 However, some details were still missing
about the formation processes of carbides in the studies.31,32
Therefore, it is worthwhile to investigate the formation
mechanism of interstitial compounds synthesized by this
novel resin-based route for the design of improved catalytic
materials.
at 110 °C for 12 h to obtain (Mo,Ni),P-containing resins (desig-
nated as Mo/P-R and Ni/P-R, R = 1, 1.3, 1.5, 1.8 or 2). Finally,
the dried resin precursors were heated from room temperature
(RT) to a given value at a rate of 10 °C min−1 under Ar
(50 ml min−1) and maintained at this temperature for 1 h,
followed by cooling to RT under an Ar flow, and then passi-
vated in a 1% O2/Ar flow for 2 h. A series of products were
obtained from heating at 500, 600, 700, 800 and 900 °C,
designated as Mo/P-R-T and Ni/P-R-T for Mo and Ni samples
(R = 1, 1.3, 1.5, 1.8 or 2, T = 500, 600, 700, 800 and 900).
2.2 Sample characterization
The microstructure of the samples was determined by using
an X-ray diffractometer (XRD, X’Pert Pro MPD) equipped with
a Cu Kα source and by X-ray photoelectron spectroscopy (XPS)
using a Kratos Axis ultra (DLD) equipped with an Al Kα X-ray
source. The binding energies ( 0.2 eV) were referenced to the
C 1s peak at 284.6 eV due to adventitious carbon. The mor-
phologies of the samples were characterized by using a scan-
ning electron microscope (SEM, Hitachi S-4800) equipped with
an energy dispersive X-ray (EDX) spectroscope and trans-
mission electron microscope (TEM, Philips Tecnai 10). The
BET surface area and pore structure of the samples were
measured through nitrogen adsorption at liquid-nitrogen
In this work, it is the first time that dispersed metal phos-
phides (MoP and Ni2P nanoparticles) were synthesized by a
resin-based route. XRD, XPS, SEM, TEM, ICP-AES, N2 adsorp-
tion–desorption analysis, and MS were adopted for the investi-
gation of solid and gas products, and to gain insights into the
formation mechanism of phosphides during thermal carbon-
ization processes.
temperature (77 K) by using
a surface area analyzer
(NOVA4200). The chemical composition of the samples was
determined by inductively coupled plasma-atomic emission
spectroscopy (ICP-AES). The processes of thermal carbon-
ization were carried out under an Ar flow (30 ml min−1). The
sample (0.1 g) was heated from 100 to 800 °C at a heating rate
of 10 °C min−1, and then was held at 800 °C for 30 min. The
2. Experimental
2.1 Sample preparation
The raw resin used was a D201 × 1 cinnamic strong alkali Hiden HPR20 mass spectrometer (MS) was used for detection.
anion exchange resin (purchased from Shanghai Jinkai resin
Co. Ltd). Fig. 1 shows the chemical structure of this raw resin.
Firstly, the raw resin was dried at 110 °C for 12 h, and then
ground to fine powder. After that, the resin powder was incipi-
ent-wetness impregnated with an aqueous solution of metal
3. Results and discussion
3.1 Microstructural characterization
salt ((NH4)6Mo7O24·4H2O or Ni(NO3)2·6H2O) and (NH4)2HPO4 In order to gain insights into the formation mechanism of the
with 30 wt% metal loading and a given metal/P ratio (R). In final phosphide products, the (Mo,Ni),P-containing resin pre-
the case of nickel, a few drops of HNO3 were needed to give cursors and the products obtained from thermal carbonization
rise to a homogeneous solution. The slurries were evaporated of these precursors needed to be characterized. Fig. 2 shows
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 3 XRD patterns of the carbonization products of Mo,P-containing
resin (Mo : P = 1) after heating in an Ar flow at different temperatures.
Fig. 2 XRD patterns of the (Mo,Ni),P-containing resin precursors as
well as raw resin for comparison.
other Mo phosphides, indicating that the as-prepared MoP
material was phase-pure. The results suggested that the MoP
cannot be formed until the carbonization temperature was
increased up to 800 °C.
the XRD patterns of the (Mo,Ni),P-containing resin precursors
(Mo/P-1 and Ni/P-2) as well as the raw resin for comparison. It
can be seen from Fig. 2 that the raw resin was completely
amorphous. After introducing the metal and P into the resin,
the Ni species on the resin were amorphous, and most of the
Mo species should also be amorphous because only the
NH4H2PO4 phase (2θ = 16.8, 23.8, 29.1 and 45.3°) and a very
small amount of not-well-defined phase (marked with an aste-
risk) were observed on the Mo/P-1 sample. The unidentified
species was probably due to unknown ammonium molyb-
denum oxide or ammonium (hydrogen) phosphate. In order to
confirm whether ion exchange occurred during incipient-
wetness impregnation of ion exchange resins with the solution
containing Mo(Ni) and P, a portion of impregnated resin was
treated sequentially by washing, drying and carbonization. The
XRD patterns (Fig. S1†) of the washed precursors were similar
to that of the raw resin except that the XRD pattern of the
washed Mo/P-1 sample showed an amorphous species within
the 5–13° 2θ range. However, the XRD patterns (Fig. S2†) of the
carbonized samples indicated that no diffraction peaks were
assigned to metals or metal-containing compounds. Therefore,
it was reasonable to deduce that there was no obvious (or a
little) exchange of ions between resin and metal-containing
salt solutions. In other words, the Mo and Ni species should
mainly be introduced into the pores of resin or covered on the
surface during the incipient-wetness impregnation process.
Fig. 3 shows the XRD patterns of the thermal carbonization
products of the Mo,P-containing resin precursor with a Mo/P
ratio of 1 after heating at different temperatures. It can be seen
that the Mo/P-1-500, Mo/P-1-600 and Mo/P-1-700 samples were
almost amorphous, but the Mo/P-1-800 sample demonstrated
diffraction peaks at 27.9, 32.0, 42.9, 57.0, 64.7, 67.4, 74.0 and
85.5°, corresponding to the (001), (100), (101), (111), (200),
(201) and (112) reflections of MoP, respectively. There were no
Fig. 4 shows the XRD patterns of the thermal carbonization
products of Ni,P-containing resin with a Ni/P ratio of 2 after
heating at different temperatures. After thermal carbonization
at 500 and 600 °C, the Ni/P-2-500 and Ni/P-2-600 samples
demonstrated the diffraction peaks of the Ni metallic phase
(2θ = 44.3, 51.7 and 76.1°, indexing respectively to the (111),
(200) and (220) planes). When the carbonization temperature
was increased to 700 °C, Ni3P (2θ = 41.8, 43.6, 45.2 and 46.6°,
due respectively to the (231), (112), (240) and (141) planes) was
obtained as the main phase with a small amount of Ni12P5.
With a further increase in temperature, there was a clear phase
transformation in the sample from a high Ni/P ratio to a low
Fig. 4 XRD patterns of the carbonization products of Ni,P-containing
peaks that can be assigned to Mo oxides, Mo phosphates or resin (Ni : P = 2) after heating in an Ar flow at different temperatures.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Ni/P ratio (Ni3P → Ni12P5 → Ni2P). Nevertheless, the phase The nickel phosphates were more difficult to reduce than
composition of the phosphide product was still a mixture of nickel oxides and nickel oxide phosphates, hence they would
Ni2P (2θ = 40.7, 44.6, 47.4 and 54.2°, corresponding respect- be converted to nickel pyrophosphates at a high temperature
ively to the (111), (201), (210) and (300) planes) and Ni12P5 when the carbonization products could not afford to reduce
(2θ = 38.4, 41.8, 47.0 and 49.0°, corresponding respectively to them. The results demonstrated that R = 1.5 was appropriate
the (112), (400), (240) and (312) planes) up to a carbonization to prepare the Ni2P sample via thermal carbonization of
temperature as high as 900 °C. It was also found in a tra- Ni,P-containing resin, although Ni12P5 was still present as a very
ditional H2-TPR method that although the preparation was minor phase. In view of the fact that both the carbonization
carried out with a stoichiometric ratio of 2 : 1 Ni : P, the as-pre- temperature and the Ni/P ratio are important factors for the
pared phosphide tended to be a mixed phase (Ni2P and Ni12P5) synthesis of pure Ni2P, by simultaneously adjusting the two
probably due to the loss of phosphorus in the preparation parameters phase-pure Ni2P was prepared successfully after
process.33,34 It was suggested that the excess phosphorus thermal carbonization of the resin precursor with a Ni/P ratio
might be necessary in the preparation of phase-pure Ni2P.33
Subsequently, the influence of the precursor molar ratio of
of 1.5 at 900 °C (see the XRD pattern in Fig. 5).
For the sake of comparison, Table 1 lists the phase struc-
Ni/P was investigated. Fig. 5 shows the XRD patterns of nickel tures of the products obtained in the study. It was obvious that
phosphides prepared at 800 °C through varying the precursor pure MoP can be obtained by carbonizing its corresponding
Ni/P molar ratio (R) from 2 to 1. It is clear that when the R was precursor with a stoichiometric ratio of 1 : 1 Mo : P at 800 °C.
decreased from 2 to 1.5, the diffraction peaks of Ni12P5 gradu- However, the preparation of phase-pure Ni2P was not as easy
ally weakened but those relating to Ni2P became stronger. as that of MoP because it depended not only on the carbon-
However, conversely, with a further decrease of the R to 1.3, ization temperature but also on the initial Ni/P molar ratio.
the reflections corresponding to Ni12P5 (Ni2P) began to Fortunately, pure Ni2P can be obtained by carbonization of the
strengthen (weaken), and there was an additional new set of resin precursor with a small excess of phosphorus (Ni/P = 1.5)
peaks (2θ = 30.0, 35.5 and 43.1°) due to the Ni2P2O7 phase. It at a high temperature of 900 °C. Table 2 lists the ICP-AES
is worth noting that this pyrophosphate was obtained as a results for the precursors and carbonization products obtained
nearly single phase when the R was decreased to a value as low in the study. In the case of Mo samples, the Mo/P atomic ratio
as 1. Based on the above results, it can be concluded that the of Mo/P-1-800 was similar to that of Mo/P-1 and the values
Ni/P ratio was indeed important for the preparation of Ni2P. At were nearly identical to the corresponding stoichiometric ratio
a higher Ni/P ratio (R > 1.5), relatively more Ni12P5 was formed of 1/1 for MoP. The result indicated that there was no loss of
due to the lack of phosphorus in the carbonization process. In
contrast, the precursor containing excess phosphorus (R < 1.5)
was easily converted to nickel pyrophosphate (Ni2P2O7). It was
suggested that the calcined precursor with R > 1.5 might be
Table 1 Phase structures of the products obtained in this study
mainly composed of nickel oxides and nickel oxide phos-
Sample
XRD phases
JCPDS cards
phates, but with decreasing R other components (i.e. nickel
phosphates) were contained inside and started to dominate.35
Mo/P-1-500
Mo/P-1-600
Mo/P-1-700
Mo/P-1-800
Ni/P-2-500
Ni/P-2-600
Ni/P-2-700
Ni/P-2-800
Ni/P-2-900
Ni/P-1.8-800
Ni/P-1.5-800
Ni/P-1.3-800
Ni/P-1-800
Ni/P-1.5-900
Amorphous
Amorphous
Amorphous
MoP
Ni
Ni
Ni3P, Ni12P5
Ni12P5, Ni2P
Ni2P, Ni12P5
Ni2P, Ni12P5
Ni2P, Ni12P5
Ni2P, Ni12P5, Ni2P2O7
Ni2P2O7
—
—
—
65–6487
65–0380
65–0380
34–0501, 74–1381
74–1381, 65–1989
65–1989, 74–1381
65–1989, 74–1381
65–1989, 74–1381
65–1989, 74–1381, 39–0710
39–0710
Ni2P
65–1989
Table 2 ICP-AES results for the precursors and carbonization products
Precursor
M/P atomic ratioa
Product
M/P atomic ratioa
Mo/P-1
Ni/P-2
Ni/P-1.5
Ni/P-1.5
Ni/P-1
0.98
2.02
1.52
1.52
1.02
Mo/P-1-800
Ni/P-2-800
Ni/P-1.5-800
Ni/P-1.5-900
Ni/P-1-800
0.99
2.09
1.77
2.02
1.03
Fig. 5 XRD patterns of the carbonization products of Ni,P-containing
resin with different Ni/P molar ratios after heating in an Ar flow at
800 °C.
a M = Mo, Ni.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Mo and P during the carbonization process. In the case of Ni
samples, it was obvious that some P was lost from the precur-
sors except for Ni/P-1 during the carbonization process. Gener-
ally, using (NH4)2HPO4 as the phosphorus source, there were
possible processes (the volatilization of POx and the generation
of PH3) which contributed to the loss of P in these samples.35
In the current preparation processes, no trace of PH3 was
detected in the MS signal, as proved later. Thus, the P lost
during the carbonization of Ni,P-containing resins should be
originated from the volatile POx. However, the fact that no loss
of P was observed in the Ni/P-1-800 sample was reasonable
because Ni2P2O7 was not reduced even at a temperature as
high as 800 °C (see Fig. 5). Finally, the bulk Ni/P atomic ratio
of the Ni/P-1.5-900 sample was found to be 2.02, which pro-
vided supplementary evidence that the Ni-1.5-900 sample was
indeed pure Ni2P. Table 3 lists the textural properties of these
pure phosphide samples. Note that the surface areas of Mo
and Ni phosphides prepared by the resin route were signifi-
cantly higher than those of the corresponding phosphides pre-
pared by the traditional H2-TPR method.36 In particular, the
MoP sample showed a high surface area of 201.2 m2 g−1
,
which was about forty times higher than that of the MoP
sample (5.3 m2 g−1) reported by others.36
Thereafter, XPS analysis was used to further determine the
surface composition of the as-prepared pure phosphides (Mo/
P-1-800 and Ni/P-1.5-900). Fig. 6 shows the XPS spectra in the
Mo 3d, Ni 2p and P 2p regions for the two samples obtained in
this study (see the full scan width XPS spectra in Fig. S3†).
Based on the curve fitting analysis of the XPS spectra, the
binding energies of Mo 3d5/2, Ni 2p3/2 and P 2p3/2 and the dis-
tribution of the corresponding species is estimated (Table 4).
Since the as-prepared MoP and Ni2P samples were usually pas-
sivated prior to exposure to air, the surface regions of the
phosphides should be dominated by the oxidized species and
the underlying reduced species related to phosphide phases.37
The binding energies of oxidized Mo (233.2 eV), Ni (856.1 eV)
and P (133.0–133.7 eV) were in accordance with assignments
by others to Mo6+ (Mo oxide), Ni2+ (Ni phosphate) and P5+
(phosphate) species,24,37 respectively. In the case of phos-
phided species, the binding energy peaks at 228.4, 853.4 and
129.7–129.9 eV were attributed to Mox+ (MoP), Niy+ (Ni2P) and
Pz− (phosphide),36 respectively. Note that the XPS Ni 2p3/2
spectrum of the Ni/P-1.5-900 sample showed only one peak at
853.4 eV for Ni2P species which was far away from the value
(852.6 eV) for Ni12P5,21 indicating that the Ni/P-1.5-900 sample
Fig. 6 XPS spectra of Mo 3d, Ni 2p and
(b) Ni/P-1.5-900.
P
2p (a) Mo/P-1-800;
Table 4 XPS results for the phase-pure phosphides obtained in this
study
Binding energy (eV)
Mo 3d5/2
Mox+
P 2p3/2
Pz−
Sample
Mo6+
233.2
P5+
Mo/P-1-800
228.4
129.7
133.7
Ni 2P3/2
P 2p3/2
Niy+
Ni2+
Pz−
P5+
Ni/P-1.5-900
853.4
856.1
129.9
133.0
synthesized was indeed pure Ni2P rather than the mixed phase
(Ni2P and Ni12P5). The result was in agreement with the obser-
vation of phase-pure Ni2P in the XRD pattern of the Ni/P-1.5-
900 sample (see Fig. 5).
Table 3 Textural properties of the pure phosphide samples obtained in
this study
Sample
SBET (m2 g−1
)
DP (nm)
VP (cm3 g−1
)
Finally, the morphologies of the pure phases of Ni2P and
MoP (hereafter denoted as R-Ni2P and R-MoP, respectively)
obtained by carbonization of resin precursors were character-
ized by SEM and TEM. For the sake of comparison, the
Mo/P-1-800
Ni/P-1.5-900
201.2
69.3
5.89
1.25
0.30
0.22
SBET = BET surface area; DP = average pore diameter; VP = total pore
volume.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
morphology of the carbonized raw resin (denoted as R) as well
as those of Ni2P and MoP (see their XRD patterns in Fig. S4,†
hereafter denoted as H2-Ni2P and H2-MoP, respectively)
obtained by the traditional H2-TPR method6 were also charac-
terized. Fig. 7 and S5† show the SEM images of the samples
obtained by carbonization and H2 reduction. In the case of the
carbonized samples, the product R consisted of aggregates of
irregularly shaped nanoplatelets (Fig. 7a), which was likely to
be a carbon plate because the EDX result (see Table 5) revealed
that only carbon and a trace amount of oxygen elements could
Table 5 SEM/EDX surface elemental analysis results
Ma
(at%)
P
O
(at%)
C
(at%)
M/P atomic
ratioa
Sample
(at%)
R
—
6.8
18.2
10.9
15.2
17.2
12.6
—
3.3
8.7
5.2
14.5
16.7
12.9
1.2
5.1
8.9
7.4
5.8
6.2
4.5
98.8
84.8
64.2
76.5
64.5
59.9
70.0
—
R-Ni2P-1
R-Ni2P-2
R-Ni2P-3
R-MoP-1
R-MoP-2
R-MoP-3
2.06
2.10
2.09
1.05
1.03
0.98
a M = Ni, Mo.
be detected in this microregion. After introducing the metal
and P into the resin, the SEM image of R-Ni2P (Fig. 7b) showed
that the Ni2P particles in the size range from about 50 to
150 nm were well dispersed on these carbon platelets. As for
the SEM image of the R-MoP sample (Fig. 7c), dispersed MoP
nanoparticles (<50 nm) were also observed but these particles
were wrapped with the carbon platelets rather than on their
surfaces. In contrast, the H2-Ni2P and H2-MoP showed aggrega-
tion morphologies (Fig. S5†), which was consistent with the
SEM results reported by others.24 These SEM results were in
good agreement with the observation of TEM images (Fig. 8
and S6†). In order to gain insights into the homogeneity of the
phosphides prepared by the resin route, the EDX measurement
was undertaken three times for each material in terms of
different microregions. It can be seen from Table 5 that the
corresponding metal (Ni, Mo), P, C and O elements can be
detected, and the compositions were variable in different
microregions. The results indicated that some particle aggrega-
tion was not avoided on the surface of these phosphides
because as regards the preparation of the precursors, the
metal and P loadings were so high that the incipient wetness
technique used cannot ensure these species to be very homo-
geneous in/on the resin.
3.2 Formation mechanism
To investigate the formation mechanism of phosphides, the
carbonization processes of (Mo,Ni),P-containing resin precur-
sors were followed by mass spectrometry (MS). As suggested by
previous studies,38,39 when the resins were carbonized under
an inert atmosphere or vacuum, the volatile products should
consist mainly of water, methane, hydrogen and ethene, with a
comparatively low level of benzene and toluene. In the current
study, only the results for masses (m/z) of 2, 16, 18, 28 and 44
were shown for understanding the formation mechanism, as
the masses of 78 (C6H6), 91 (C7H8), 62 (C2H3Cl) and 34 (PH3)
were almost featureless or provided little additional infor-
mation. Fig. 9 shows the thermal carbonization-MS profiles of
the raw resin and (Mo,Ni),P-containing resin precursors. It was
believed that the carbonization process of resin involved a
series of dehydration and curing reactions in two stages: low-
temperature carbonization (<650 °C) and high-temperature
carbonization (≥650 °C).30 In the case of raw resin, a large
Fig. 7 SEM images of (a) R, (b) R-Ni2P and (c) R-MoP.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 8 TEM images of (a) R-Ni2P and (b) R-MoP.
quantity of multi-component gases was released in the low-
temperature carbonization stage (see Fig. 9a) and the simul-
taneous signals of m/z = 2, 16, 18, 28 and 44 corresponded to
H2, CH4, H2O, C2H4 and CO2, respectively. The CO2 species
can be detected might be due to the decomposition of poro-
genic agents trapped inside the pores during the synthesis
process of ion exchange resin.40 In the high-temperature stage,
there was only an intense peak of H2 observed together with a
weak H2O emission, which might be attributed to the deep
carbonization of resin. However, the delay of the H2O response
in reaching the maximum was possible due to the slow desorp-
tion of them on the surface, as suggested before.41
In comparison with the raw resin, the MS profiles of (Mo,
Ni),P-containing resin precursors showed several different fea-
tures in both the low- and the high-temperature carbonization
stages. On the one hand, in the low-temperature regions (see
the details in Fig. S7†), there was an obvious consumption of
Fig. 9 Thermal carbonization-MS profiles of (a) raw resin, (b) Mo,P-
containing resin (Mo : P = 1 : 1) and (c) Ni,P-containing resin (Ni : P =
1.5 : 1).
CH4 and C2H2, accompanied by first increasing and then by these reducing gases. The XRD investigation (Fig. 4)
decreasing the concentration of H2. The results suggested that
coincided with this suggestion, in which the formation of
the decrease in the concentrations of CH4, C2H2 and H2 was the metallic Ni phase was observed after heating the Ni,P-
probably due to the occurrence of reduction of the precursors containing resin precursor at 500 or 600 °C. Although XRD
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
characterization cannot provide proof of the reduction of the ing nanoparticle aggregation.24 Therefore, the MoP and Ni2P
Mo,P-containing resin precursor due to the fact that Mo/P-1- nanoparticles obtained by carbonization of the corresponding
500 and Mo/P-1-600 were almost amorphous (Fig. 3), the amor- resin precursors showed better dispersions and higher surface
phous high-valence Mo species in the precursor might be areas than those synthesized by the traditional H2-TPR route.
reduced to noncrystalline or very poorly crystalline low-valence The advantages of the resulting dispersed MoP and Ni2P nano-
Mo species under a CH4- and H2-containing atmosphere at particles are worth further exploration.
500–600 °C, as suggested before.28 At the same time, it was
reasonable that the increase in the concentration of H2 should
be completely full due to the contribution of the dissociation 4. Conclusions
of CH4 and C2H2. However, it can be observed that the temp-
In summary, dispersed pure phases of MoP and Ni2P sup-
eratures (∼480 °C) for the maximum concentration of H2 were
ported by carbon were successfully synthesized for the first
not in accord with those for the maximum loss of CH4
time by thermal carbonization of metal- and phosphorus-
(450 °C) and C2H2 (440 °C). The results suggested that the carbon-
containing resins. Carbon itself, as well as H2 and CH4 produced
ization of resin promoted by Mo and Ni species could be
in carbonization processes, was found to act as the redundant
another contribution to the production of H2. Note that there
for the formation of phosphides. Unlike the traditional H2-
was no obvious H2O desorption peak on the profiles, indicat-
TPR method which produced H2O as the only gas product, this
ing an increase in the adsorption ability with respect to H2O
novel synthesis route produced a large amount of COx besides
after introducing the metal and phosphorus into the resin. On
H2O. The hydrothermal sintering of phosphide particles might
be mitigated due to the lowering of the water vapor pressure.
the other hand, during the high-temperature carbonization
processes (≥650 °C), there was a heavy consumption of H2 at
The carbon generated from resin can also play a role in
about 800 °C, at the same time mainly CO, accompanied by a
bonding of nanoparticle aggregation. Hence, the dispersions
small amount of CO2 and CH4, was evolved. We also found
and surface areas of the as-prepared MoP and Ni2P were
that much H2O was released and the delay of H2O responses
superior to those of the corresponding phosphides prepared
should be attributed to the slow desorption rate.41 The results
by the H2-TPR method. Herein, in this study we have
developed a simple and effective route for preparing dispersed
MoP and Ni2P nanoparticles.
suggested that the high-temperature carbonization stage of
(Mo,Ni),P-containing resins involved strong redox reactions,
leading to the formation of the corresponding metal phos-
phides. This suggestion was in accord with the XRD analyses
(Fig. 3 and 4) in which the formation of Mo and Ni phosphides
was observed in this stage. Since no trace of PH3 was detected
Acknowledgements
in the MS signal, Mo and Ni species should be phosphided by
We acknowledge the financial support from the National
Natural Science Foundation of China (No. 21276253 and
POx, as suggested before.34
It had been reported that both H2 and CH4 could be used
21006032).
as reducing agents for the synthesis of metal phosphides.6,42
In particular, the former was the most widely utilized reduc-
tant in the preparation of phosphide catalysts.35 Additionally,
Notes and references
the MS profiles verified that the emission of H2 and CH4 was
followed by a large amount of H2O and COx. It was therefore
reasonable to conclude that the evolution of the two reducing
gases during carbonization processes contributed more or less
to the formation of the Mo and Ni phosphides. Besides, the
solid product (i.e. carbon) from the thermal carbonization of
resin should also serve as a reducing agent. This phenomenon
had been proved to occur in the preparation of metal phos-
phides via a carbothermal reduction route.43,44
It was well known that the traditional H2-TPR method for
phosphide synthesis gave H2O as the only gas product, with
the result that the clustering of phosphide particles occurred
owing to hydrothermal sintering.21,23–26 However, in the
current synthesis route, carbon itself, as well as H2 and CH4
produced in carbonization processes, was found to act as the
reductant for the formation of Mo and Ni phosphides. The
synthesis processes produced a great deal of COx besides H2O,
1 A.-M. Alexander and J. S. J. Hargreaves, Chem. Soc. Rev.,
2010, 39, 4388–4401.
2 S. Carenco, D. Portehault, C. Boissiere, N. Mezailles and
C. Sanchez, Chem. Rev., 2013, 113, 7981–8065.
3 J. S. J. Hargreaves, Coord. Chem. Rev., 2013, 257, 2015–2031.
4 S. Izhar, M. Yoshida and M. Nagai, Electrochim. Acta, 2009,
54, 1255–1262.
5 G. R. Li, F. Wang, Q. W. Jiang, X. P. Gao and P. W. Shen,
Angew. Chem., Int. Ed., 2010, 49, 3653–3656.
6 S. T. Oyama, J. Catal., 2003, 216, 343–352.
7 K. L. Stamm, J. C. Garno, G. Y. Liu and S. L. Brock, J. Am.
Chem. Soc., 2003, 125, 4038–4039.
8 S. T. Oyama, T. Gott, H. Zhao and Y. K. Lee, Catal. Today,
2009, 143, 94–107.
9 J. Guan, Y. Wang, M. Qin, Y. Yang, X. Li and A. Wang,
J. Solid State Chem., 2009, 182, 1550–1555.
which might facilitate the mitigation of hydrothermal sinter- 10 A. Wang, M. Qin, J. Guan, L. Wang, H. Guo, X. Li, Y. Wang,
ing by lowering the water vapor pressure. In addition, the good
contact between carbon species and phosphides favored avoid-
R. Prins and Y. Hu, Angew. Chem., Int. Ed., 2008, 120, 6141–
6143.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
11 Y. Ni, L. Jin and J. Hong, Nanoscale, 2011, 3, 196–200.
12 Y. Xie, H. L. Su, X. F. Qian, X. M. Liu and Y. T. Qian, J. Solid
State Chem., 2000, 149, 88–91.
13 S. Yang, C. Liang and R. Prins, J. Catal., 2006, 237, 118–
130.
28 Z. Yao, M. Li, X. Wang, X. Qiao, J. Zhu, Y. Zhao, G. Wang,
J. Yin and H. Wang, Dalton Trans., 2015, 44, 5503–5509.
29 L. Yu, D. Chen, J. Li and P. G. Wang, J. Org. Chem., 1997,
62, 3575–3581.
30 P. Barbaro and F. Liguori, Chem. Rev., 2009, 109, 515–529.
14 R. K. Chiang and R. T. Chiang, Inorg. Chem., 2007, 46, 369– 31 R. Wang, C. Tian, L. Wang, B. Wang, H. Zhang and H. Fu,
371. Chem. Commun., 2009, 3104–3106.
15 A. E. Henkes, Y. Vasquez and R. E. Schaak, J. Am. Chem. 32 Z. Yan, H. Wang, M. Zhang, Z. Jiang, T. Jiang and J. Xie,
Soc., 2007, 129, 1896–1897. Electrochim. Acta, 2013, 95, 218–224.
16 E. Muthuswamy, G. H. L. Savithra and S. L. Brock, ACS 33 X. Liu, L. Xu and B. Zhang, J. Solid State Chem., 2014, 212,
Nano, 2011, 5, 2402–2411. 13–22.
17 W. Maneeprakorn, C. Q. Nguyen, M. A. Malik, P. O’Brien 34 S. J. Sawhill, K. A. Layman, D. R. Van Wyk,
and J. Raftery, Dalton Trans., 2009, 2103–2108.
M. H. Engelhard, C. Wang and M. E. Bussell, J. Catal.,
2005, 231, 300–313.
35 X. Wang, P. Clark and S. T. Oyama, J. Catal., 2002, 208,
321–331.
18 A. Panneerselvam, M. A. Malik, M. Afzaal, P. O’Brien and
M. Helliwell, J. Am. Chem. Soc., 2008, 130, 2420–2421.
19 X. Hu and J. C. Yu, Chem. Mater., 2008, 20, 6743–6749.
20 S. Liu, X. Liu, L. Xu, Y. Qian and X. Ma, J. Cryst. Growth, 36 I. I. Abu and K. J. Smith, J. Catal., 2006, 241, 356–366.
2007, 304, 430–434.
37 A. W. Burns, A. F. Gaudette and M. E. Bussell, J. Catal.,
2008, 260, 262–269.
38 C. A. Lytle, W. Bertsch and M. McKinley, J. Anal. Appl.
Pyrolysis, 1998, 45, 121–131.
21 Q. Guan, X. Cheng, R. Li and W. Li, J. Catal., 2013, 299, 1–9.
22 V. Zuzaniuk and R. Prins, J. Catal., 2003, 219, 85–96.
23 R. Cheng, Y. Shu, L. Li, M. Zheng, X. Wang, A. Wang and
T. Zhang, Appl. Catal., A, 2007, 316, 160–168.
39 Y. Yamashita and K. Ōuchi, Carbon, 1981, 19, 89–94.
24 V. M. L. Whiffen, K. J. Smith and S. K. Straus, Appl. Catal., 40 T. Wang, H. Liang and Q. Yuan, Korean J. Chem. Eng., 2012,
A, 2012, 419–420, 111–125. 29, 396–403.
25 P. Xiao, M. A. Sk, L. Thia, X. Ge, R. J. Lim, J. Y. Wang, 41 Z. W. Yao and H. Dong, Appl. Surf. Sci., 2008, 255, 2259–
K. H. Lim and X. Wang, Energy Environ. Sci., 2014, 7, 2624–
2629.
26 Z. Yao, Z. Lai, X. Zhang, F. Peng, H. Yu and H. Wang,
Mater. Res. Bull., 2011, 46, 1938–1941.
27 Z. Yao, G. Wang, Y. Shi, Y. Zhao, J. Jiang, Y. Zhang and
H. Wang, Dalton Trans., 2015, 44, 14122–14129.
2264.
42 S. Burns, J. S. J. Hargreaves and S. M. Hunter, Catal.
Commun., 2007, 8, 931–935.
43 H. Wang, Y. Shu, A. Wang, J. Wang, M. Zheng, X. Wang and
T. Zhang, Carbon, 2008, 46, 2076–2082.
44 Z. W. Yao, J. Alloys Compd., 2009, 475, L38–L41.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.