An approach to preparing porous and hollow metal phosphides with higher hydrodesulfurization activity

Article abstract of DOI:10.1016/j.jssc.2011.03.057

This paper describes an effective method for the synthesis of metal phosphides. Bulk and supported Ni₂P, Cu₃P, and CoP were prepared by thermal treatment of metal and the amorphous red phosphorus mixtures. Porous and hollow Ni₂P particles were also synthesized successfully using this method. The structural properties of these products are investigated using X-ray powder diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscopy (SEM), inductively coupled plasma (ICPAES) and X-ray photoemission spectroscopy (XPS). A rational mechanism was proposed for the selective formation of Ni₂P particles. In experimental conditions, the Ni₂P/SiO₂ catalyst exhibits excellent hydrodesulfurization (HDS) activity for dibenzothiophene (DBT).

Full text of DOI:10.1016/j.jssc.2011.03.057

Journal of Solid State Chemistry 184 (2011) 1556–1560
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
An approach to preparing porous and hollow metal phosphides
with higher hydrodesulfurization activity
a
b,n
a
Limin Song , Shujuan Zhang , Qingwu Wei
a
College of Environment and Chemical Engineering & State Key Laboratory of Hollow-fiber Membrane Materials and Membrane Processes,
Tianjin Polytechnic University, Tianjin 300160, PR China
b
College of Science, Tianjin University of Science & Technology, Tianjin 300457, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
This paper describes an effective method for the synthesis of metal phosphides. Bulk and supported
Ni P, Cu P, and CoP were prepared by thermal treatment of metal and the amorphous red phosphorus
Received 15 December 2010
Received in revised form
2
3
2
mixtures. Porous and hollow Ni P particles were also synthesized successfully using this method.
1
3 March 2011
The structural properties of these products are investigated using X-ray powder diffraction (XRD),
transmission electron microscopy (TEM), scanning electron microscopy (SEM), inductively coupled
plasma (ICP–AES) and X-ray photoemission spectroscopy (XPS). A rational mechanism was proposed
Accepted 29 March 2011
Available online 19 April 2011
Keywords:
for the selective formation of Ni
excellent hydrodesulfurization (HDS) activity for dibenzothiophene (DBT).
2011 Elsevier Inc. All rights reserved.
2 2 2
P particles. In experimental conditions, the Ni P/SiO catalyst exhibits
Metal
Phosphorus
&
Metal phosphides
Porous and hollow
Hydrodesulfurization
1
. Introduction
bulk and supported metal phosphides is easy via the proposed
method. Therefore, the proposed method is a general method for
the synthesis of bulk and supported metal phosphides. In com-
parison with the TPR method, it is simple (facile heat treatment of
pure metal and red phosphorus) and short (reaction time is only
0.5 h), making it advantageous to the synthesis of metal phos-
Transition metal phosphides have attracted considerable inter-
est for some time because these technologically important mate-
rials are used in semiconductors, luminescent devices, electronic
components, and hydroprocessing catalysts [1,2]. Transition
metal phosphides are attractive candidates especially as HDS
catalysts [3–8]. Thus, it is important to develop general synthetic
approaches that yield transition metal phosphides. A variety of
methods have emerged for the synthesis of these materials
including temperature-programmed reduction (TPR) of metal
2
phide catalysts. In addition, porous and hollow Ni P particles may
be obtained successfully using the proposed method, indicating
the possibility to synthesize metal phosphides with special
morphologies.
3
phosphates, treating Ni metal particles with PH , solvothermal
2. Experimental
reactions, as well as high temperature annealing of organometal-
lic and solid-state precursors, etc. [9–23]. In this aspect, it has
Commercially available solvents and reagents were used with-
2
been reported that the Ni P particles were prepared by treating
out further purification. All the reagents were purchased from
Tianjin Guangfu Fine Chemical Research Institute (Tianjin, China).
The purity of metal and red phosphorus powder is 99.9% (AR).
Tube furnace is a home-made device with 2000 W of heating
power (Fig. 1). In a typical preparation process, the precursors
were prepared by mixing metal (Ni, Cu, and Co) and red
phosphorus (P) powder with the molar ratio of metal to phos-
phorus being 1/2. The above precursors were heated to 700 1C at a
an amorphous Ni–B alloy with phosphine [24]. The as-prepared
catalysts exhibited excellent catalytic activity in the HDS of
dibenzothiophene. This paper reports a new synthetic route via
heat treatment of the mixtures metal and red phosphorus in
flowing N
including bulk and supported Ni
time, the HDS of dibenzothiophene was carried out in a flowing
fixed-bed reactor for the Ni P/SiO catalyst. Although only Ni P,
P, and CoP were synthesized in this study, synthesizing other
2
to prepare the corresponding metal phosphides
(
2
P, Cu P, and CoP). At the same
3
2
2
2
2
rate of 20 1C/min in a flowing N with a rate of 10 mL/min. The
reaction was maintained for 0.5 h at 700 1C, and then the products
Cu
3
were cooled to room temperature. Finally, the samples were
n
passivated in a flow of 1 vol% O
2 2 2
/N . The yields of the Ni P,
Corresponding author. Fax: þ86 22 60600658.
E-mail address: zhangshujuan@tust.edu.cn (S. Zhang).
3
Cu P, and CoP samples are about 63%, 56%, and 51% when the
0022-4596/$ - see front matter & 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2011.03.057

L. Song et al. / Journal of Solid State Chemistry 184 (2011) 1556–1560
1557
3
5
6.161, 39.241, 41.761, 45.281, 46.321, 47.461, 52.381, 53.641,
6.681, 59.201, 66.721, 69.381, 73.581, and 78.501), and CoP
(
4
2
y
¼23.661, 31.641, 32.001, 35.381, 36.361, 36.681, 38.961, 45.181,
6.261, 48.161, 52.261, 56.061, 56.801, 61.721, 64.201, 66.021,
68.021, 68.861, 74.781, 75.901, 76.341, and 77.161) with hexagonal,
hexagonal, and orthorhombic structures, respectively (JCPDS Card
No. 89-4864, 71-2261, and 89-4862). No other reflections (includ-
ing oxides and other phosphide phases) were found in the XRD
patterns indicating pure Ni
To study the effect of Ni/P ratios on the product, the formation
of Ni P is investigated. Fig. 3 presents the XRD patterns of the
resultant Ni P samples from precursors with different Ni/P ratios.
At Ni/P molar ratio 2/1 (Fig. 3a), all diffraction peaks correspond-
ing to the hexagonal Ni P (2
5.381, 40.721, 44.601, 47.361, 54.161, 54.961, 63.421, 66.341,
9.881, 72.661, and 74.701) are observed, which is identified as
P. At Ni/P molar ratio 1/1 and 1/2 (Fig. 3b and c), the
diffraction peaks of Ni P are very strong and no other peaks
appear. This indicates that Ni is transformed into Ni P completely
at this Ni/P ratio. With a further decrease on Ni/P molar ratio
2 3
P, Cu P, and CoP products.
2
2
Fig. 1. Apparatus of tube furnace. (a) furnace body; (b) quartz tube; (c) valve;
(
d) thermocouple; (e) controller; (f) N
i) valve.
2 2
intake; (g) quartz boat, (h) N outtake; and
2
y¼17.501, 26.341, 30.521, 31.781,
(
3
6
molar ratio of metal to phosphorus is 1/2, respectively. In the
experiment, a part of the excess phosphorus was transported out
pure Ni
2
2
of the reaction by N
2
flow. Another part of phosphorus in the form
2
of white phosphorus deposited at the cool end of the tube. In the
reaction, the allotrope phosphorus is highly toxic and flammable.
These experiments require careful operation with precautions.
The paper wiping the nozzle should be put into cold water to
prevent fire.
2 2
The precursors of Ni P/SiO catalysts (with metal loadings
equivalent to 10.2 wt%) were prepared as follows. Prior to use, the
silica support was calcined in air at 773 K for 3 h. The silica was
then impregnated with an aqueous solution of Ni powder under
stirring, and followed by drying for 3 h at 393 K. The above SiO
2
and red phosphorus with the molar ratio of metal to phosphorus
being 1/2 were mechanically mixed to obtain the precursor. Other
steps are similar to prepare bulk Ni
The products were characterized using X-ray diffraction (XRD)
recorded on Rigaku D/max 2500 powder diffractometer
2
P.
a
equipped with monochromatic high-intensity CuK
a radiation
˚
(
l
¼1.5406 A). The morphology and size of the as-prepared
products were observed by a Hitachi H-7650 transmission elec-
tron microscopy (TEM) and a Hitachi S-4800 scanning electron
microscope (SEM). The elemental content of the samples was
determined using the inductively coupled plasma (ICP–AES,
9
000(NþM)) method. The binding energy (BE) was measured
by an X-ray photoelectron spectrometer (XPS, Pekin-Elmer
PHI5300). In the XPS analysis, the calibration of BE is the standard
peak of adventitious carbon (C1s). The BE has been calibrated
according to the standard peak of carbon (C1s) in the manuscript.
The HDS of dibenzothiophene was carried out in a flowing
fixed-bed reactor using a feed consisting of a decalin solution of
DBT (0.5 wt%). The conditions of the HDS reaction were 330 1C,
Fig. 2. X-ray diffraction patterns of bulk Ni
2 3
P, Cu P, and CoP treated at 700 1C for
0.5 h. The molar ratio of metal to phosphorus is 1/2.
ꢀ
1
ꢀ1
3
(
.0 MPa, WHSV¼9 h , and H
0.6 g) of Ni P/SiO catalyst diluted with quartz sands to 5.0 mL
was reduced by heating from room temperature to a certain
2
flow rate 90 ml min . 1.0 mL
2
2
ꢀ
1
2
temperature in H (200 mL min ), maintaining the sample at
that temperature for 2 h, before changing the temperature to the
desired reaction temperature. Liquid samples were collected
every hour and analyzed by a gas chromatography with a flame
ionization detector (FID) and a capillary column (OV101). The DBT
conversion was used as a measure of the HDS activity [15].
3
. Results and discussion
Fig. 2 shows the X-ray diffraction (XRD) patterns of as-
prepared bulk products. All diffraction peaks of the three samples
are attributed to the corresponding Ni P (2
0.561, 31.821, 35.401, 40.741, 44.641, 47.401, 54.181, 55.001,
3.461, 66.361, 69.921, 72.701, 74.721), Cu P (2
¼28.641, 29.161,
2
y¼17.521, 26.361,
3
6
Fig. 3. X-ray diffraction patterns of bulk Ni
2
P with different Ni/P ratios treated at
3
y
700 1C for 0.5 h. (a) 2:1, (b) 1:1, (c) 1:2, and (d) 1:3.

1
558
L. Song et al. / Journal of Solid State Chemistry 184 (2011) 1556–1560
2 3 2 3
Fig. 4. FESEM images of bulk Ni P, CoP, and Cu P treated at 700 1C for 0.5 h. The molar ratio of metal to phosphorus is 1/2. (a and b) Ni P, (c) CoP, and (d) Cu P.
2
Fig. 5. TEM images of bulk Ni P treated at 700 1C for 0.5 h. The molar ratio of
metal to phosphorus is 1/2.
(
Ni/P¼1/3) (Fig. 3d), pure Ni
2
P phase is still produced. This result
indicates that Ni powder reacts with red phosphorus powder
easily under experimental conditions to form a well-crystallized
2
Ni P phase if the Ni/P ratio is smaller than 2/1. The melting point
of red phosphorus is 590 1C. The solid red phosphorus was
transformed into gaseous phosphorus when the reaction was
carried out at 700 1C. Therefore, superfluous phosphorus evapo-
rates and drains along with the N
and red phosphorus powder. The pure Ni
2
current after reacting with Ni
P phase can be obtained
2
successfully as long as the Ni/P ratio is smaller than 2/1.
Fig. 4 presents the field emission scanning electron microscopy
images of the bulk Ni
of Ni P are not uniform and composed of many small granules
Fig. 4a). The particle sizes vary in the range 1–5 m. Interest-
2 3
P, CoP, and Cu P. It is clear that the particles
2
(
m
ingly, the interior of particles is hollow and the surface has many
pores (Fig. 4b). From Fig. 4c, we can see the CoP particles are
Fig. 6. X-ray photoelectron spectra (XPS) for the Ni
2
P precursor treated at 700 1C for
0.5 h in a flowing N (the molar ratio of metal to phosphorus is 1/2). (A), P2P (B) Ni2p.
2

L. Song et al. / Journal of Solid State Chemistry 184 (2011) 1556–1560
1559
accumulated by small particles with the size of 0.5–1
ever, there are no obvious pores in the surface. From Fig. 4d, the
Cu P particles are the large particles of solid with micron size. To
investigate the size of the Ni P granules and its internal structure,
transmission electron microscopy images were obtained (Fig. 5).
From Fig. 5, the actual size of granules is about 20 nm, the interior
is truly hollow and the surface is porous.
In nanoparticle reactions there is a popular mechanism for
hollow particle formation as inner material from a solid particle
migrates to the exterior. For the formation mechanism of the hollow
m
m. How-
peak centered at 871.0 eV on the high binding-energy side is
assigned to the Ni2p signal from oxidized Ni species [27]. Quanti-
tative analysis was performed using the intensities of the P2p and
Ni2p spectra normalized by their atomic sensitivity factors. It was
found that Ni/P is 0.73. The apparent P excess is an indication of
3
2
the existence of oxidized P in the Ni
To investigate the hydrotreating activity of the Ni
sized by the proposed route, the SiO -supported Ni
(Ni P/SiO ) was prepared and tested by the HDS of dibenzothio-
phene. The Ni loading in the Ni P/SiO samples was 10.2 wt%. The
2
P particles.
2
P synthe-
2
P catalyst
2
2
2
2
2
Ni
phosphorus and nickel on the surface of metal particles to form
Ni P. The red phosphorus at this time is liquid or gaseous because
the reaction temperature is higher than the melting point of red
phosphorus, therefore, the resultant Ni P is a loose structure. The
liquid or gaseous red phosphorus goes into the interior of loose Ni
2
P particles, we speculated that it might be the first reaction of red
bulk and supported compositions of as-prepared samples are
determined by inductively coupled plasma analysis. The average
2
mole ratio of Ni to P in the as-prepared bulk and supported Ni
samples is 1.86 and 1.93, respectively. Fig. 7 shows the XRD
patterns of the bulk Ni P, SiO support, and Ni P/SiO . The XRD
pattern of the support shows that the SiO powder is amorphous
because of the lack of sharp peaks. The sharp peaks of Ni P/SiO
are attributed to the Ni P phase (2
¼30.321, 31.561, 31.701,
40.561, 44.461, 47.221, 52.841, 54.021, 54.841, 66.221, 69.821,
72.541, and 74.601). Fig. 8 gives the TEM images of the SiO
supported Ni P samples. From Fig. 8a, it is very clear that most of
the as-synthesized Ni P particles have a hollow structure and the
size is about 10–30 nm. The size of the Ni P particles obtained by
TPR method is around 30–50 nm in Fig. 8b. Comparing the two
2
P
2
2
2
2
2
2
P
2
particles, and continue to react with nickel cores to form Ni
ultimately leads to the formation of hollow structure.
2
P, which
2
2
2
y
Fig. 6 shows the P2p (Fig. 6A) and Ni2p (Fig. 6B) XPS spectra of
the as-synthesized Ni Psample. The peak at 130.0 eV is attributed
to P atoms in Ni P [25], while the peaks at higher BE values
134.4 eV) are attributed to oxides which are more pronounced in
the as-received sample [26]. The single and narrow peak at
53.4 eV is assigned to the Ni cations in Ni P [25]. The broad
2
2
2
2
(
2
2
8
2
Fig. 7. X-ray diffraction patterns of bulk and SiO
2
supported Ni
2
P treated at 700 1C
Fig. 9. Effect of reaction time on the HDS activity of the Ni
2 2
P/SiO at 320 1C.
for 0.5 h. The molar ratio of metal to phosphorus is 1/2.
(a) The TPR method and (b) the new method. Ni loadings¼10.2 wt%.
2 2
Fig. 8. TEM images of Ni P/SiO treated at 700 1C for 0.5 h. (a) Ni and red phosphorus powder (the molar ratio of metal to phosphorus is 1/2) and (b) the TPR method.

1
560
L. Song et al. / Journal of Solid State Chemistry 184 (2011) 1556–1560
the two samples XRD peaks, all the XRD peaks can be in agreement
with the Ni standard patterns (JCPDS Card No. 89-4864).
2
P
This indicates that we are still able to obtain pure Ni
2
P when the
reaction time reduced to 10 min.
The solid–liquid–gas phases of red phosphorus may coexist
under the reaction condition because the melting point of red
phosphorus is 590 1C. Since Ni
2
P was obtained at Ni/P at 2/1
(
Fig. 3a) and there is a N flow to remove the vapor, and pure
2
phosphorus was still expected to leave in the product when extra
phosphorus was used as reactant (Fig. 6), it is more likely that the
evaporation of phosphorus is not very significant. So solid–liquid–
gas should coexist under the reaction condition. Since gas will be
2
carried away by the N flow, the reaction is possible happens on
solid metal and liquid phosphorus interface. The liquid phos-
phorus reacts with Ni according to the following pathway:
NiþP (liquid)¼Ni
2
P
The reaction does not carry out completely and forms many
other outgrowths because the reaction temperature of 600 1C just
surpasses the melting point of red phosphorus.
Fig. 10. X-ray diffraction patterns of bulk Ni
2
P treated at different temperature for
0.5 h. The molar ratio of metal to phosphorus is 1/2.
4. Conclusions
2 3
In summary, bulk and supported Ni P, Cu P, and CoP can be
prepared by heat treating of metal and red phosphorus powder.
The method is simple, timesaving, and universal to synthesize
corresponding metal phosphide catalysts. The as-prepared Ni
catalyst exhibits good HDS activity for dibenzothiophene HDS. In
addition, the hollow and porous particles of Ni P can be prepared by
the proposed route, resulting to a much larger surface area and more
2 2
P/SiO
2
2
active sites in which the Ni P catalyst can show better activity for
dibenzothiophene HDS.
References
[
[
1] D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, J. Catal. 207 (2002) 266.
2] W.R.A.M. Robinson, J.N.M. van Gestel, T.I. Koranyi, S. Eijsbouts, A.M. van der
Kraan, J.A.R. van Veen, V.H.J. de Beer, J. Catal. 161 (1996) 539.
3] S.T. Oyama, J. Catal. 216 (2003) 343.
4] P.A. Clark, S.T. Oyama, J. Catal. 218 (2003) 78.
5] S.T. Oyama, Y.-K. Lee, J. Catal. 258 (2008) 393.
[
[
[
[
6] 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.
[
[
[
7] Y.-K. Lee, S.T. Oyama, J. Catal. 239 (2006) 376.
8] S.T. Oyama, X. Wang, Y.K. Lee, et al., J. Catal. 221 (2004) 263.
9] V.B. Chernogorenko, S.V. Muchnik, K.A. Lynchak, Z.A. Klimak, V.G. Ivanchenko,
Mater. Res. Bull. 16 (1981) 1.
2
Fig. 11. X-ray diffraction patterns of bulk Ni P treated at 700 1C for different time.
The molar ratio of metal to phosphorus is 1/2.
[10] J.W. Liu, X.Y. Chen, M.W. Shao, C.H. An, W.C. Yu, Y.T. Qian, J. Cryst. Growth
catalysts, the as-synthesized Ni
face area and more active sites, which will help to improve the
desulfurization activity. The data of Ni P/SiO catalysts for the
HDS of dibenzothiophene are listed in Fig. 9. HDS activity reached
3% in the experimental condition; this is much higher than that
of Ni P/SiO prepared by the TPR method (72%).
2 2
P/SiO can provide a higher sur-
252 (2003) 297.
[
[
[
11] Q.X. Guan, W. Li, M.H. Zhang, K.Y. Tao, J. Catal. 263 (2009) 1.
12] S. Yang, C. Liang, R. Prins, J. Catal. 237 (2006) 118.
13] H. Lobou e´ , C. Guillot-Deudon, A. Florin Popa, A. Lafond, B. Rebours, C. Pichon,
T. Cseri, G. Berhault, C. Geantet, Catal. Today 130 (2008) 63.
14] A.E. Henkes, Y. Vasquez, R.E. Schaak, J. Am. Chem. Soc. 129 (2007) 1896.
15] Z.Q. Wang, L.X. Zhou, M.H. Zhang, M. Su, W. Li, K.Y. Tao, Chem. Asian J.
2
2
[
[
8
2
2
4
(2009) 1794.
To study the reaction mechanism, the reactions with Ni and
red phosphorus powder were carried out at different tempera-
[16] G.J. Shi, J.Y. Shen, J. Mater. Chem. 19 (2009) 2295.
[
[
17] L.N. Ding, M.Y. Zheng, A.Q. Wang, T. Zhang, Catal. Lett. 135 (2010) 305.
18] Y.H. Ni, A.L. Tao, G.Z. Hu, X.F. Cao, X.W. Wei, Z.S. Yang, Nanotechnology 17
tures and time. Fig. 10 shows the XRD peaks of the Ni
with different preparation temperatures. The obtained product is
pure phase Ni P (2
0.681, 44.581, 47.341, 54.141, 54.941, 61.161, 63.381, 66.321,
9.861, 72.661, and 74.701) without other impurity when the
2
P samples
(2006) 5013.
[19] Y.H. Ni, K.M. Liao, J. Li, Cryst. Eng. Commun. 12 (2010) 1568.
[20] Y.H. Ni, J. Li, L.N. Jin, J. Xia, J.M. Hong, K.M. Liao, New J. Chem. 33 (2009) 2055.
2
y
¼17.461, 26.301, 30.481, 31.741, 35.341,
[
[
[
21] J. Li, Y.H. Ni, K.M. Liao, J.M. Hong, J. Colloid Interf. Sci. 332 (2009) 231.
22] Y.H. Ni, J. Li, L. Zhang, Mater. Res. Bull. 44 (2009) 1166.
23] Y.H. Ni, L.N. Jin, J.M. Hong, Nanoscale 3 (2011) 196.
4
6
reaction temperature is 700 1C or 800 1C. When the temperature
decreases to 600 1C, the peaks of many species including Ni P,
appear. Fig. 11 gives the XRD peaks of the
P samples treated at 700 1C for 10 and 30 min. Compared with
[24] L.M. Song, W. Li, G.L. Wang, M.H. Zhang, K.Y. Tao, Catal. Today 125 (2007) 137.
[
[
[
25] X.P. Duan, Y. Teng, A.J. Wang, V.M. Kogan, X. Li, Y. Wang, J. Catal. 261 (2009) 232.
26] T. Kor a´ nyi, Appl. Catal. A Gen. 239 (2003) 253.
27] E.E. Khawaja, M.A. Salim, M.A. Khan, F.F. Al-Adel, G.D. Khattak, Z. Hussain,
J. Non-cryst. Solids 110 (1989) 33.
2
Ni
Ni
3
5 4 8 3
P, Ni P , and Ni P
2