Effect of surface composition on the catalytic performance of molybdenum phosphide catalysts in the hydrogenation of acetonitrile

Article abstract of DOI:10.1016/j.jcat.2007.10.029

A series of molybdenum phosphide catalysts with initial Mo/P ratios varying in a narrow range of 0.90-1.10 was prepared by temperature-programmed reaction; characterized by X-ray diffraction, BET, elemental analysis, X-ray photoelectron spectroscopy, and CO chemisorption measurements; and tested for the hydrogenation of acetonitrile at different pressures (0.1-1.0 MPa) and temperatures (473-513 K). The catalysts exhibited attractive catalytic activity, especially at a H₂ pressure above 0.2 MPa. The surface composition of the MoP catalysts could be fine-tuned by the initial Mo/P ratio, which consequently led to different surface properties (e.g., CO uptakes) and catalytic behaviors. Catalysts with high initial Mo amount gave high selectivity to the primary amine, ethylamine, whereas those with high initial P amount created more condensed amines, diethylamine and triethylamine.

Full text of DOI:10.1016/j.jcat.2007.10.029

Journal of Catalysis 253 (2008) 66–73
www.elsevier.com/locate/jcat
Effect of surface composition on the catalytic performance of molybdenum
phosphide catalysts in the hydrogenation of acetonitrile
Pengfei Yang, Zongxuan Jiang, Pinliang Ying, Can Li ∗
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
Received 16 August 2007; revised 24 October 2007; accepted 26 October 2007
Abstract
A series of molybdenum phosphide catalysts with initial Mo/P ratios varying in a narrow range of 0.90–1.10 was prepared by temperature-
programmed reaction; characterized by X-ray diffraction, BET, elemental analysis, X-ray photoelectron spectroscopy, and CO chemisorption
measurements; and tested for the hydrogenation of acetonitrile at different pressures (0.1–1.0 MPa) and temperatures (473–513 K). The catalysts
exhibited attractive catalytic activity, especially at a H pressure above 0.2 MPa. The surface composition of the MoP catalysts could be fine-tuned
2
by the initial Mo/P ratio, which consequently led to different surface properties (e.g., CO uptakes) and catalytic behaviors. Catalysts with high
initial Mo amount gave high selectivity to the primary amine, ethylamine, whereas those with high initial P amount created more condensed
amines, diethylamine and triethylamine.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Molybdenum phosphide; Hydrogenation; Acetonitrile; Selectivity; Amine; Ethylamine; Diethylamine; Triethylamine
1. Introduction
and Rh tend to form secondary amines and Pd and Pt have a
high propensity to form tertiary amines [2–5]. Furthermore, the
acid–base properties of the catalyst support also have a strong
effect on the selectivity [6–9], because the acidic sites are par-
tially responsible for condensation reactions giving condensed
amines [10,11].
The catalytic hydrogenation of nitriles is an important route
for the production of various lower-aliphatic amines that are
widely used in the pharmaceutical, agriculture, textile, rubber,
and plastic industries. In the process of nitrile hydrogenation,
the partially hydrogenated reaction intermediates (imines or
Schiff bases) are highly reactive and usually form a mixture
of primary, secondary, and tertiary amines that leads to costly
product separation processes. An ideal way to solve this prob-
lem is to improve the activity and selectivity in the production
of amines.
As reviewed by Volf and Pašek [1] and de Bellefon and
Fouilloux [2], the activity and selectivity for nitrile hydrogena-
tion are affected by reaction conditions, nitriles, and catalysts.
Of these factors, the catalyst plays a decisive role in determin-
ing the reaction selectivity. There is consensus that the selec-
tivity to primary amines is high over Ru and Ni, whereas Cu
Up to now, most catalysts applied in nitrile hydrogenation
have been transition metals, especially those from the eighth
group. However, other nonmetal catalysts (e.g., transition metal
phosphides [TMPs]) have been reported only rarely. The TMPs,
including MoP [12,13], WP [14], and iron group phosphides
[15], are regarded as new active catalytic materials in the hy-
drotreatment of fuels [16]. As far as we know, only a few
researchers [17–19] have studied TMPs as catalysts for hydro-
genation reactions other than hydrodesulfurization (HDS) or
hydrodenitrogenation (HDN), and no study of the catalysis of
nitrile hydrogenation on metal phosphides has been reported.
Ni2P is known to be a better catalyst than MoP in hydrotreat-
ment [16]. For the hydrogenation of acetonitrile, however, we
found MoP to be the better catalyst. Therefore, in this study we
chose MoP as catalyst and investigated its catalytic activity for
the hydrogenation of acetonitrile, which also can be used as a
*
Corresponding author. Fax: +86 411 84694447.
E-mail address: canli@dicp.ac.cn (C. Li).
URL: http://www.canli.dicp.ac.cn (C. Li).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.10.029

P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
67
probing molecule to study the acid–base properties of catalysts
[20–22].
at half maximum (in radians), and θ is the diffraction angle (in
degrees).
TMPs are usually prepared from the corresponding metal
phosphate or phosphate-like precursors by means of temper-
ature-programmed reaction (TPR) [12]. Some studies [23–26]
indicated that the composition in the precursors could have a
significant effect on the catalytic activity due to the formation
of a less active phosphide phase at low P content, or blocking
of active sites by excessive P at high P content. In the present
work, to avoid the phase change of metal phosphides and de-
crease the blocking by excessive P, we prepared a series of MoP
catalysts by controlling the initial Mo/P ratios in a narrow range
of 0.90 to 1.10, close to that expected from the stoichiometry of
MoP.
Our results show that the prepared MoP catalysts have differ-
ent surface compositions when the initial Mo/P ratio changes.
The catalysts show quite attractive catalytic performance in the
hydrogenation of acetonitrile, especially at a H2 pressure above
0.2 MPa. Furthermore, the selectivity to amines can be fine-
tuned by the initial Mo/P ratio.
The BET surface areas were measured on an ASAP 2000
system with nitrogen adsorption at 77 K. Elemental analysis
was carried out with an inductively coupled plasma (ICP) in-
strument (Plasma-Spectra-II, Leeman Inc., USA) on the pre-
pared MoP catalysts that had been dissolved in aqua regia in a
microwave digester.
X-ray photoelectron spectra (XPS) of the catalyst samples
were recorded on a KROTOS AMICAS spectrometer (Shi-
madzu, Japan) with a magnesium Kα (hv = 1253.6 eV) as a
radiation source, operating at 10 mA and 12 kV. The residual
pressure inside the analysis chamber was about 2 × 10−6 Pa.
Charging effects were corrected by adjusting the binding en-
ergy of C 1s to 284.8 eV. The precision of the binding energy
values was 0.2 eV. MoP(1.00) was chosen for in situ XPS mea-
surement. The sample was first pressed into a pellet, placed on
a silica holder, and then pretreated in a H2 flow at 873 K for 1 h.
The in situ XPS analysis was performed after the chamber was
cooled to RT and degassed to a vacuum of about 2 × 10−6 Pa.
CO chemisorption measurements were carried out using a
CHEMBET-3000 chemisorption analyzer (Quantachrome In-
struments). The passivated catalyst (0.5–1.0 g) was purged in
a flow of He at RT and then reduced at 873 K in an 80-ml/min
flow of H2 for 1.5 h. After evacuation, the CO chemisorption
was measured at 313 K.
2. Experimental
2.1. Catalyst preparation
We have described the preparation of bulk MoP catalysts
with different initial Mo/P ratios (Mo/P = 0.90, 0.95, 1.00,
1.05, and 1.10) in a previous publication [27]. First, stoi-
chiometric amounts of ammonium heptamolybdate [(NH4)6-
Mo7O24·4H2O] were mixed with diammonium hydrogen
phosphate [(NH4)2HPO4] in distilled water. The white solid
obtained after evaporation of the water was calcined in air at
773 K for 4 h. The calcined solid was subsequently ground to a
powder, and further pressed into pellets, crushed, and sieved to
obtain 20- to 40-mesh particles. The particles thus obtained
were placed into a quartz reactor and reduced by the TPR
method [12] from room temperature (RT) to 573 K in 30 min
and from 573 to 923 K in 175 min at a rate of 2 K/min in flow-
ing H2 (150–200 ml/min). The final temperature of 923 K was
maintained for 120 min, followed by cooling to RT in H2. The
sample was then passivated at RT for 6 h in a stream of 1 vol%
O2/N2 to prevent violent oxidation of the freshly prepared cat-
alysts.
2.3. Hydrogenation of acetonitrile
The hydrogenation of acetonitrile was performed in a
continuous-flow system. A quartz tube (40 cm long, 7 mm o.d.,
and 6 mm i.d.) and a stainless tube (35 cm long, 10 mm o.d.,
and 8 mm i.d.) were used for the reaction at ambient and high
pressures, respectively. Before any measurement, 0.25 g of cat-
alyst was pretreated at 873 K in a 90-ml/min flow of hydrogen
for 60 min, followed by cooling in hydrogen flow to the desired
reaction temperature. Hydrogen was bubbled at a flow rate of
20 ml/min through a stainless saturator containing acetonitrile
at 283 K, and then introduced into the reactor. The reactants
and products were analyzed by an online gas chromatograph
(HP 6890N) equipped with a flame ionization detector and a
chromosorb 103 column (3 mm × 2 m).
For the sake of brevity, the catalysts with different ini-
tial Mo/P ratios were denoted as MoP(0.90), MoP(0.95),
MoP(1.00), MoP(1.05), and MoP(1.10).
The conversion is defined as
CH3CNin − CH3CNout
conversion (mol%) = 100 ×
.
CH3CNin
2.2. Characterization of the catalysts
The selectivity was calculated based on the peak areas of the
gas chromatogram by considering the different sensitivity fac-
tors of the flame ionization detector. The selectivity to product
i is defined as
X-ray diffraction (XRD) patterns were obtained on a Rigaku
MiniFlex diffractometer with a CuKα (λ = 1.5418 Å) radiation
source operated at 30 kV and 15 mA. Diffraction patterns were
collected from 20◦ to 80◦ at a speed of 5◦/min. The particle size
of the phosphides was determined using Scherrer’s equation,
d = 0.9λ/β cosθ, where d is the mean crystal size (in nm), λ is
the wavelength of the X-ray radiation (in nm), β is the full width
(corrected area)i
selectivityi (mol%) = 100 ×
.
sum of all corrected areas

68
P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
3. Results
than that expected from the MoP stoichiometry, indicating that
the MoP surface was P-enriched. The P enrichment was higher
at higher P content in the precursors. In addition, O species
were detected by XPS, which originated from the passivation
by air in the preparation process. The surface layer contain-
ing O species was usually called a passivating layer, because
it prevented the phosphide crystallite from violent oxidation
by air [23]. There was a high content of O species at high P
content in the precursors. The oxidation states of the Mo and
P species were also obtained from the XPS analysis (spectra
not shown). Both the Mo and P species exhibited two oxidation
states, a low-valence species (MoP) and a high-valence species
[Mo(PO4)2].
MoP(1.00) was chosen for in situ XPS analysis where the
sample was pretreated at 873 K in H2 flow. The results showed
that the MoP surface could not be recovered by a H2 pretreat-
ment, because the O species (PO34−) remained in the surface,
even with the O/P ratio decreased from 5.0 to 2.9. At the same
time, a slight P loss from the surface was observed, with the
Mo/P ratio increasing from 0.77 to 0.82.
3.1. Characterization of the catalysts
Fig. 1 shows the XRD patterns of the MoP catalysts. All
the patterns demonstrate the same peaks at 28.2, 32.4, 43.3,
57.5, 65.2, 67.8, and 74.4◦, similar to the standard pattern
of MoP from the JCPDS powder diffraction file (PDF card
24-0771) [28] and those reported in the literature [12,13,27,
29–32]. The estimated average crystallite size was 50 nm for
MoP(0.90), 47 nm for MoP(0.95), 47 nm for MoP(1.00), 47 nm
for MoP(1.05), and 53 nm for MoP(1.10) (Table 1), based on
Scherrer’s equation applied to the (101) diffraction peak at
43.3◦.
The MoP catalysts had low BET surface areas of about
10 m2/g (Table 1) similar to those of other bulk TMPs, includ-
ing iron group phosphides [15], MoP [13,33], and WP [14]. The
low surface area of the bulk TMPs was likely due to the high
preparation temperature, which can lead to the sintering of the
formed crystallites [14].
The CO uptake was determined to evaluate the surface con-
centration of the metallic sites of the MoP catalysts. As listed
in Table 1, the CO uptake increased from 1.2 for MoP(0.90) to
1.8 for MoP(1.00). However, the CO uptake of MoP(1.10) was
as high as 7.4, much higher than those of the other two cata-
lysts, indicating that a high concentration of metallic Mo sites
was present in the surface of MoP(1.10).
Elemental analysis (Table 1) showed that the Mo/P ratios
in the bulk were close to those in the precursors. The XPS re-
sults (Table 1) indicated that the Mo/P surface ratios were lower
3.2. Hydrogenation of acetonitrile
Fig. 2 shows the catalytic performance of the MoP catalysts
with different initial Mo/P ratios in the hydrogenation of ace-
tonitrile at ambient pressure and 473 K after 120 min on stream.
According to the gas chromatography analysis, the hydrogena-
tion products were ethylamine (EA), diethylamine (DEA), tri-
ethylamine (TEA), and traces of hydrocarbons. For this series
of MoP catalysts, the conversion of CH3CN roughly increased
with the initial Mo/P ratio and reached an optimal value at
initial Mo/P = 1.00. The order for the CH3CN conversion
was MoP(1.00) > MoP(1.10) ≈ MoP(1.05) > MoP(0.95) >
MoP(0.90).
The turnover frequencies (TOFs) were calculated from the
conversions and CO uptakes (Table 1). MoP(1.00) had the high-
est TOF of 72 h−1 among the tested catalysts, comparable to
Fig. 1. XRD patterns of (a) MoP(0.90), (b) MoP(0.95), (c) MoP(1.00),
(d) MoP(1.05), and (e) MoP(1.10).
Table 1
Some physicochemical data of the MoP catalysts
b
c
−1
)
Catalysts
Crystallite
size (nm)
BET result
(m /g)
Bulk
composition
Surface
composition
CO uptake
TOF (h
a
2
2
(µmol/m )
MoP(0.90)
MoP(0.95)
MoP(1.00)
MoP(1.05)
MoP(1.10)
50
47
47
47
53
9.3
10.1
11.1
8.6
Mo
Mo
Mo
Mo
Mo
P
P
P
P
P
Mo
Mo
Mo
Mo
Mo
PO
PO
PO
PO
PO
1.2
–
1.8
–
22
–
72
–
0.85
0.94
0.97
1.07
1.09
0.71
0.73
0.77
0.88
0.90
5.36
5.47
5.04
4.00
3.20
9.1
7.4
10
a
◦
Calculated by Scherrer’s equation for the (101) diffraction peak at 43.3 .
Stoichiometry of CO/active site is assumed to be 1 for the tested samples.
b
c
After 120 min on stream. Reaction conditions: 0.25 g catalyst, H /acetonitrile = 16.3, 473 K, and ambient pressure.
2

P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
69
Fig. 2. Catalytic performance of the MoP catalysts with different initial Mo/P
ratios for acetonitrile hydrogenation after 120 min on stream. Reaction condi-
−1 −1
tions: 473 K, ambient pressure, WHSV = 4800 ml g
h
, H /acetonitrile =
2
16.3.
that of some Ni catalysts measured under similar reaction con-
ditions [34]. MoP(1.10) had a lower TOF than MoP(0.90), even
though its conversion was higher than that of MoP(0.90).
Fig. 2 also indicates that the selectivities to amine products
were dependent on the initial Mo/P ratios. The selectivity to
primary amine (EA) was higher for catalysts with high initial
Mo/P ratios, with MoP(1.10) giving the highest selectivity of
about 80%. In contrast, the condensed amines (DEA and TEA)
were the main products for catalysts with low initial Mo/P ra-
tios. The selectivity to DEA for MoP(<1.00) was >60%, and
the total selectivity to the condensed amines was about 90%.
For this series of MoP catalysts applied in the hydrogenation of
acetonitrile at ambient pressure, the selectivity to EA was in the
order MoP(1.10) ≈ MoP(1.05) > MoP(1.00) > MoP(0.95) >
MoP(0.90), whereas the selectivities to DEA and TEA were in
the reverse order.
Fig. 3 shows the conversion of acetonitrile at different H2
pressures (0.1–1.0 MPa) for the MoP catalysts as a function
of time on stream. MoP(0.90) had very low activity at am-
bient pressure, whereas MoP(1.00) and MoP(1.10) showed
high initial activity but suffered severe deactivation. The activ-
ity and stability of all of the catalysts were greatly improved
with increasing H2 pressure. In the steady state of the reac-
tion, MoP(1.00) gave complete conversion of CH3CN when
the H2 pressure reached 0.2 MPa, whereas for MoP(1.00) and
MoP(1.10), the pressure point of complete conversion was at
0.4 MPa and 1.0 MPa, respectively.
Fig. 3. Conversion of acetonitrile at different pressures of H (0.1–1.0 MPa)
2
for (a) MoP(0.90), (b) MoP(1.00), and (c) MoP(1.10) as a function of time on
−1 −1
.
stream. Reaction conditions: T = 473 K, WHSV = 4800 ml g
h
Fig. 5 shows the effect of temperature on the activity and the
selectivities to amine products for the MoP catalysts at ambi-
ent pressure. The activity increased with reaction temperature
(Fig. 5a), whereas the selectivities were differently affected by
temperature (Figs. 5b–5d). For MoP(1.00), the effect of temper-
ature on the selectivities was not clear. For MoP(0.90), however,
the selectivities to DEA and TEA decreased at elevated temper-
atures with increasing EA selectivity, whereas for MoP(1.10),
the selectivities to amine products were reversely affected by
temperature.
Fig. 4 shows the effect of H2 pressure on the activity and
the selectivities to amine products for the MoP catalysts. For
MoP(1.10), the selectivities to amines are hardly affected by the
H2 pressure. For MoP(1.00), the selectivity to TEA increased
gradually with the H2 pressure, whereas the selectivities to
EA and DEA decreased slightly. More clearly than MoP(1.00),
MoP(0.90) showed an increasing tendency of TEA selectivity
and decreasing tendency of DEA selectivity against H2 pres-
sure.
4. Discussion
4.1. Surface properties of MoP catalysts with different initial
Mo/P ratios
It is well known that for the preparation of TMPs, the P
amount in the precursor is required to be slightly higher than

70
P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
Fig. 4. Effect of H pressure on the (a) conversion and the selectivities to (b) EA, (c) DEA, and (d) TEA for MoP(0.90), MoP(1.00), and MoP(1.10). Reaction
2
−1 −1
conditions: T = 473 K, WHSV = 4800 ml g
h
, reaction time = 180 min.
that expected from the stoichiometry of the TMPs to avoid the
formation of a phase deficient in phosphorus, due to P loss
during the preparation process [24,35]. In the present work,
however, we found that the MoP catalysts could be synthesized
without any additional phase (Fig. 1) by tuning the Mo/P ra-
tios in the oxidic precursors close to the stoichiometry of MoP
(0.90–1.10), even though the initial P amount was slightly in-
sufficient.
When MoP catalysts are prepared with different initial Mo/P
ratios and have the same bulk constitution, their surface com-
position must differ. This is indeed demonstrated by the XPS
analysis (Table 1). For this series of catalysts, the surface was
P-enriched; that is, the surface P/Mo ratios were higher than
that expected from the stoichiometry of MoP. Abu and Smith
[33] also observed a slight P enrichment for bulk MoP. Korányi
[36] reported this effect for other unsupported TMPs, such as
Ni phophides, and ascribed it to surface oxidation. In this study,
the prepared MoP catalysts became more P enriched at higher
initial P amount concomitantly with the increasing O content
in the surface, which also implies that surface oxidation might
lead to P-enrichment.
much higher than those of MoP(1.00) and MoP(0.90). We as-
sume that more P is lost from the surface of MoP(1.10) than
from the surface of the others when the catalysts were pre-
treated at high temperatures in H2 flow. This surface loss of P is
supported by the in situ XPS results. On the other hand, the sur-
face P species, including phosphide and phosphate, may result
in the acidic properties of the metal phosphides, as observed by
Abu and Smith [33] and ascribed to the incomplete reduction
of the metal phosphates. Although we detected residual phos-
phate species in the reduced MoP surface as shown by the in
situ XPS analysis, the causal relationship between this species
and the acidic properties awaits more direct experimental proof.
4.2. Catalytic performance of MoP catalysts with different
initial Mo/P ratios
Up to now, most catalysts applied in nitrile hydrogenation
have been transition metals, especially those from the eighth
group. In this study, we describe for the first time a new type
of nonmetal catalyst, MoP, in the hydrogenation of acetoni-
trile. At ambient pressure and 473 K, MoP catalysts with initial
Mo/P ꢀ 1 gave initial conversions of 100% (Fig. 3). Although
showing low activity in the steady state of the reaction due to
deactivation, MoP(1.00) gave the highest TOF value (Table 1),
72 h−1, close to the activity of some Ni catalyst under simi-
lar reaction conditions [34]. The activity of the MoP catalysts
seems to be very sensitive to H2 pressure. All of the catalysts
exhibiting significantly improved activity and stability at ele-
The P-enrichment can have an obvious effect on the surface
properties of the MoP catalysts. On the one hand, the excessive
P may block surface metallic sites [24]. As a result, for catalysts
with lower initial Mo/P ratios, the stronger P-enrichment leads
to more blocking of metallic sites, and thus lower CO uptakes,
whereas CO uptakes are higher for catalysts with higher initial
Mo/P ratios (Table 1). But the CO uptake of MoP(1.10) was

P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
71
Fig. 5. Effect of temperature on the (a) conversion and the selectivities to (b) EA, (c) DEA, and (d) TEA for MoP(0.90), MoP(1.00), and MoP(1.10). Reaction
−1 −1
conditions: ambient pressure, WHSV = 4800 ml g
h
, reaction time = 180 min.
vated H2 pressure (Fig. 3). Especially when the H2 pressure
was increased beyond 0.2 MPa, MoP(1.00) attained a conver-
sion of 100% with no deactivation.
stood. Von Braun’s mechanism [38] lacks sufficient experi-
mental proof. The bifunctional mechanism of Verhaak et al.
[10] fails to confirm some views [3,5,39]. Recently, Huang and
Sachtler [40–42] proposed a novel mechanism through H/D ex-
change experiments, in which dissociation of C–H bonds of
some CH3CN molecules results in H transfer to other CH3CN
molecules.
In accordance with the literature [2,5,10,38,41–45], we
present a general scheme for acetonitrile hydrogenation in
Scheme 1. In this scheme, the addition of a H2 molecule to
the carbon–nitrogen triple bond of acetonitrile forms a reac-
tive imine intermediate, which results in either the primary
amine (EA) through further hydrogenation (reaction R1), or the
secondary/tertiary amine (DEA/TEA) through a condensation
reaction with the primary/secondary amine (reaction R2/R3).
Verhaak et al. [10] proposed that the hydrogenation step pro-
ceeded on the metallic sites of the catalysts, while the condensa-
tion reaction proceeded on the acidic sites. In the present work,
the surface Mo species of the MoP catalysts were expected to
play the role of metallic sites for reaction R1, whereas the sur-
face P species were associated with the acidic sites for reactions
R2 and R3 to produce DEA and TEA in the hydrogenation of
acetonitrile.
Another interesting property is that the selectivities of the
MoP catalysts can be fine-tuned by the initial Mo/P ratio
(Fig. 2). MoP(1.10) gave the highest selectivity to EA (∼80%)
at ambient pressure and 473 K, with a small decrease observed
at elevated H2 pressure (Fig. 4). Some nickel catalysts sup-
ported on basic supports gave high selectivity (80–90%) to EA,
whereas some supported on neutral or acidic supports gave a
value <20% [37]. Our MoP(1.10) catalyst showed an EA selec-
tivity comparable to those of the supported Ni catalysts on basic
supports despite slightly higher temperature (473 K). When the
initial Mo/P ratio was decreased, condensed amines (DEA and
TEA) were the main products. In addition, the selectivities to
DEA and TEA could be further fine-tuned by the H2 pressure.
DEA had a higher selectivity at low pressures, whereas TEA
had a higher selectivity at high pressures (Fig. 4).
In summary, MoP is a promising catalyst in the hydrogena-
tion of acetonitrile because the catalytic activity is very sensi-
tive to the H2 pressure, and the reaction selectivity can be tuned
by changing the initial Mo/P ratio in a narrow range.
4.3. Reaction mechanism of the hydrogenation of acetonitrile
on the MoP catalysts
On the one hand, CH3CN or the formed amines can be
strongly adsorbed on the metallic sites and prevent reaction R1
[46–48]. That is why the MoP catalysts [especially MoP(1.00)
and MoP(1.10)] showed high initial activity but suffered rapid
deactivation (Fig. 3). However, with increasing H2 pressure,
Although the hydrogenation of acetonitrile is widely used
in the industry, its mechanism remains incompletely under-

72
P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
Scheme 1. Reaction scheme for acetonitrile hydrogenation.
the strongly adsorbed species can be readily substituted by hy-
drogen species on the active sites, consequently improving the
activity and stability of the MoP catalysts (Figs. 3 and 4a). On
the other hand, the strongly adsorbed species can migrate from
the metallic sites to the surface acidic sites and free the former
for more hydrogenation reactions (R1), a phenomenon known
as synergetic effect [49]. MoP(1.10) has more surface Mo sites
but few acidic sites, leading to a slight synergetic effect between
these two active sites. This effect is enhanced when the number
of acidic sites increases for MoP(1.00) and MoP(0.90). There-
fore, MoP(1.10) showed lower activity than MoP(1.00) (Fig. 3)
and even lower TOF than MoP(0.90) (Table 1) in the steady
state of the reaction. However, a comparable amount of metal-
lic and acidic sites may be necessary for an optimal synergetic
effect on the activity. That is why MoP(1.00) showed the high-
est activity and TOF.
Because EA is formed through reaction R1, proceeding on
the metallic sites, while DEA and TEA are formed by reac-
tions R2 and R3, respectively, proceeding on the acidic sites,
the selectivities to these amine products increase monotonically
with the amount of the corresponding active sites in the MoP
surfaces (Fig. 2). Furthermore, with the increase of the H2 pres-
sures, the strongly adsorbed amines can be removed from the
metallic sites by hydrogen species to the adjacent acidic sites.
As a result, more condensation reactions can proceed on the
acidic sites. This effect becomes stronger for the catalysts with
a higher amount of surface P species. This is why MoP(0.90)
and MoP(1.00) showed an increasing selectivity to TEA at ele-
vated H2 pressures (Fig. 4).
for reactions R2 and R3. Due to the second factor, the selec-
tivity to EA decreases, and those to DEA and TEA increase,
with increasing reaction temperature. Comparable amounts of
surface metallic and acidic sites in MoP(1.00) might result in
a balance between these two factors and consequently in little
change in the selectivities.
5. Conclusion
In the present work, MoP catalysts showed the same bulk
phase of MoP but a different surface composition with changes
in the initial Mo/P ratio of 0.90–1.10. In situ XPS analysis
indicated that the MoP surface could not be recovered by pre-
treatment with H2; some O species remained. The MoP cat-
alysts were used to catalyze the hydrogenation of acetonitrile
for the first time. At ambient pressure and 473 K, the catalysts
exhibited high initial activity but suffered severe deactivation.
When the H2 pressure was increased, the activity and stability
were greatly improved. Furthermore, the selectivities to amine
products could be fine-tuned by the initial Mo/P ratio of the
catalysts. The primary amine was produced mainly at high Mo
content, whereas the condensed amines were produced mainly
at high P content. It is assumed that the surface Mo species act
as metallic sites to form EA, while the surface P species are re-
sponsible for the production of DEA and TEA. Therefore, MoP
can be considered a promising catalyst for the hydrogenation of
acetonitrile.
Acknowledgments
Finally, two factors should be noted to explain the effect of
temperature on selectivity (Fig. 5). First, the acid–base inter-
action between acidic sites and amines is in the order EA <
TEA < DEA [50] according to their basic character [51]. Sec-
ond, the temperature effect is stronger for reaction R1 than for
R2 and R3, because the condensation reaction has a higher ac-
tivation energy than the hydrogenation reaction [47]. On the
surface of MoP(0.90), the major species is the P species, and
the formed amines adsorb mostly on the acidic sites. Based on
the first factor, EA would be expected to desorb more easily at
elevated temperatures compared with the other amines, result-
ing in increased selectivity to EA and decreased selectivities to
DEA and TEA. On the surface of MoP(1.10), a certain amount
of metallic sites exist for reaction R1, but few acidic sites exist
This research was supported by the Natural Science Fund
of China (Grant 20303018) and the State Key Project (Grant
2006CB202506).
References
[1] J. Volf, J. Pašek, Stud. Surf. Sci. Catal. 27 (1986) 105.
[2] C. de Bellefon, P. Fouilloux, Catal. Rev. Sci. Eng. 36 (1994) 459.
ˇ
[3] J. Volf, J. Pašek, in: L. Cervený (Ed.), Catalytic Hydrogenation, Elsevier,
Amsterdam, 1986, p. 105.
[4] J. Barrault, Y. Pouilloux, Catal. Today 37 (1997) 137.
[5] Y. Huang, W.M.H. Sachtler, Appl. Catal. A 182 (1999) 365.
[6] C.V. Rode, M. Arai, M. Shirai, Y. Nishiyama, Appl. Catal. A 148 (1997)
405.

P. Yang et al. / Journal of Catalysis 253 (2008) 66–73
73
[7] M. Arai, Y. Takada, T. Ebina, M. Shirai, Appl. Catal. A 183 (1999) 365.
[8] N. Iwasa, M. Yoshikawa, M. Arai, Phys. Chem. Chem. Phys. 4 (2002)
5414.
[29] V. Zuzaniuk, R. Prins, J. Catal. 219 (2003) 85.
[30] F. Sun, W. Wu, Z. Wu, J. Guo, Z. Wei, Y. Yang, Z. Jiang, F. Tian, C. Li,
J. Catal. 228 (2004) 298.
[9] A.C. Gluhoi, P. Ma˘rginean, U. Sta˘nescu, Appl. Catal. A 294 (2005) 208.
[10] M.J.F.M. Verhaak, A.J. van Dillen, J.W. Geus, Catal. Lett. 26 (1994) 37.
[11] F. Medina-Cabello, D. Tichit, B. Coq, A. Vaccari, N.T. Dung, J. Catal. 167
(1997) 142.
[12] W. Li, B. Dhandapani, S.T. Oyama, Chem. Lett. (1998) 207.
[13] C. Stinner, R. Prins, Th. Weber, J. Catal. 191 (2000) 438.
[14] P. Clark, W. Li, S.T. Oyama, J. Catal. 200 (2001) 140.
[15] X. Wang, P. Clark, S.T. Oyama, J. Catal. 208 (2002) 321.
[16] S.T. Oyama, J. Catal. 216 (2003) 343.
[17] N.P. Sweeny, C.S. Rohrer, O.W. Brown, J. Am. Chem. Soc. 80 (1958) 799.
[18] E.L. Muetterties, J.C. Sauer, J. Am. Chem. Soc. 96 (1974) 3410.
[19] T. Nozaki, F. Kitoh, T. Sodesawa, J. Catal. 62 (1980) 286.
[20] J. Raskó, J. Kiss, Appl. Catal. A 303 (2006) 56.
[21] O. Bortnovsky, Z. Sobalík, B. Wichterlová, Z. Bastl, J. Catal. 210 (2002)
171.
[31] D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, J. Catal. 207 (2002) 266.
[32] V. Milman, B. Winkler, R. Gomperts, Chem. Eur. J. 10 (2004) 6279.
[33] I.I. Abu, K.J. Smith, J. Catal. 241 (2006) 356.
[34] A. Infantes-Molina, J. Mérida-Robles, P. Braos-García, E. Rodríguez-
Castellón, E. Finocchio, G. Busca, P. Maireles-Torres, A. Jiménez-López,
J. Catal. 225 (2004) 479.
[35] S.T. Oyama, X. Wang, Y.-K. Lee, W.-J. Chun, J. Catal. 221 (2004) 263.
[36] T.I. Korányi, Appl. Catal. A 239 (2003) 253.
[37] A.C. Gluhoi, P. Ma˘rginean, U. Sta˘nescu, Appl. Catal. A 294 (2005) 208.
[38] J. von Braun, G. Blessing, F. Zobel, Ber. 36 (1923) 1988.
[39] M. Arai, Y. Takada, Y. Nishiyama, J. Phys. Chem. B 102 (1998) 1968.
[40] Y. Huang, W.M.H. Sachtler, J. Catal. 190 (2000) 69.
[41] Y. Huang, W.M.H. Sachtler, J. Phys. Chem. B 102 (1998) 6558.
[42] Y. Huang, W.M.H. Sachtler, J. Catal. 184 (1999) 247.
[43] K. Kindler, F. Hesse, Arch. Pharm. 27 (1933) 439.
[44] Y. Huang, W.M.H. Sachtler, Stud. Surf. Sci. Catal. (2000) 527.
[45] Y. Huang, W.M.H. Sachtler, J. Catal. 188 (1999) 215.
[46] J. Volf, J. Pašek, M. Duraj, Collect. Czech. Chem. Commun. 38 (1973)
1038.
[22] F. Prinetto, M. Manzoli, G. Ghiotti, M. de J.M. Ortiz, D. Tichit, B. Coq,
J. Catal. 222 (2004) 238.
[23] S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang,
M.E. Bussell, J. Catal. 231 (2005) 299.
[24] S.T. Oyama, X. Wang, Y.-K. Lee, K. Bando, F.G. Requejo, J. Catal. 210
(2002) 207.
[25] A. Wang, L. Ruan, Y. Teng, X. Li, M. Lu, J. Ren, Y. Wang, Y. Hu,
J. Catal. 229 (2005) 314.
[26] Y. Shu, S.T. Oyama, Carbon 43 (2005) 1517.
[27] Z. Wu, F. Sun, Z. Feng, W. Wu, C. Liang, C. Li, J. Catal. 222 (1) (2004)
41.
[47] H. Li, Y. Wu, H. Luo, M. Wang, Y. Xu, J. Catal. 214 (2003) 15.
[48] P. Braos-García, P. Maireles-Torres, E. Rodríguez-Castellón, A. Jiménez-
López, J. Mol. Catal. A 168 (2001) 279.
[49] F. Medina-Cabello, D. Tichit, B. Coq, A. Vaccari, N.T. Dung, J. Catal. 167
(1997) 142.
[50] P. Braos-García, P. Maireles-Torres, E. Rodríguez-Castellón, A. Jiménez-
López, J. Mol. Catal. A 193 (2003) 185.
[28] JCPDS Powder Diffraction File, International Centre for Diffraction Data,
Swarthmore, PA, 2000.
[51] A. Streitwieser, C.H. Heathcock, Introduction to Organic Chemistry,
Macmillan, New York, 1976, Chapter 27, p. 774.