Effects of mechanochemical treatment to the vanadium phosphate catalysts derived from VOPO4·2H2O

Article abstract of DOI:10.1016/j.molcata.2006.06.047

Modification by using mechanochemical treatment of vanadium phosphate catalysts on the microstructure, morphology, oxygen nature and catalytic performance for n-butane oxidation is described and discussed. In this work, the precursor, VOHPO₄·0.5H₂O prepared by reduction of VOPO₄·2H₂O by isobutyl alcohol was subjected to a high energy planetary ball mill for 30, 60 and 120 min in ethanol. The ball milling process reduced the crystallite size of the catalysts and consequently increased their surface area. The morphologies of the milled catalysts are dependent on milling time. The highest reactivity and mobility of the lattice oxygen species has been achieved by the catalyst milled for 60 min with lower reduction peak temperature and higher amount of oxygen atoms removed. The oxygen species removed from the active V⁴⁺ phase was shown to be correlated with the rate of reaction. A good relationship was also found between the oxygen species associated with V⁵⁺ and maleic anhydride selectivity. However, a larger amount of this oxygen species will give a deleterious effect on the conversion rate. The present study demonstrate that the mechanochemical method (with an appropriate duration) effectively enhanced the catalytic activity by increasing the surface area and controlling the reactivity, and that the amount of oxygen species contributed to the partial oxidation of n-butane to maleic anhydride.

Full text of DOI:10.1016/j.molcata.2006.06.047

Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
Effects of mechanochemical treatment to the vanadium phosphate
catalysts derived from VOPO4·2H2O
Y.H. Taufiq-Yapa,∗, C.K. Goha, G.J. Hutchingsb, N. Dummerb, J.K. Bartleyb
a Department of Chemistry, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia
b School of Chemistry, Cardiff University, P.O. Box 912, Cardiff CF10 3TB, UK
Received 18 January 2006; received in revised form 22 May 2006; accepted 27 June 2006
Available online 8 August 2006
Abstract
Modification by using mechanochemical treatment of vanadium phosphate catalysts on the microstructure, morphology, oxygen nature and
catalytic performance for n-butane oxidation is described and discussed. In this work, the precursor, VOHPO4·0.5H2O prepared by reduction of
VOPO4·2H2O by isobutyl alcohol was subjected to a high energy planetary ball mill for 30, 60 and 120 min in ethanol. The ball milling process
reduced the crystallite size of the catalysts and consequently increased their surface area. The morphologies of the milled catalysts are dependent
on milling time. The highest reactivity and mobility of the lattice oxygen species has been achieved by the catalyst milled for 60 min with lower
reduction peak temperature and higher amount of oxygen atoms removed. The oxygen species removed from the active V4+ phase was shown to
be correlated with the rate of reaction. A good relationship was also found between the oxygen species associated with V5+ and maleic anhydride
selectivity. However, a larger amount of this oxygen species will give a deleterious effect on the conversion rate. The present study demonstrate
that the mechanochemical method (with an appropriate duration) effectively enhanced the catalytic activity by increasing the surface area and
controlling the reactivity, and that the amount of oxygen species contributed to the partial oxidation of n-butane to maleic anhydride.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Vanadium phosphate; Mechanochemical; n-Butane oxidation; Maleic anhydride
1. Introduction
of reducing agents and solvents have been used [4,5]. Initial cat-
alyst preparations [2,4] used water as solvent but most studies,
in recent years, have concentrated on the use of alcohols since
they can exhibit the dual role of solvent and a reducing agent.
Hutchings et al. [6–8] have shown that very active catalysts can
be prepared using a two-stage method based on VOPO4·2H2O.
In particular, they have found that the alcohol used in the sec-
ond step to reduce the VOPO4·2H2O to VOHPO4·0.5H2O can
produce catalyst precursors with better morphology and texture.
Mechanochemical method has been employed in the mod-
ification of morphology and texture of the catalyst precursor
to improve the catalytic performance of the vanadium phos-
phate catalyst [9–17]. Zazhigalov et al. [13] reported that the
ball milling process could reduce the particle size of catalysts
and corresponding increase in the specific surface area. This
observation was agreed by Fait et al. [14] who also reported
that a decreased particle size and an increased microstrain in the
VOHPO4·0.5H2O precursor would improve the catalytic prop-
erties of the V–P–O catalysts. In another report by Hutchings
and Higgins [15] on the precursor VOHPO4·0.5H2O milled in
Vanadium phosphate (VPO) catalysts represent the sole
example of an industrial catalyst for the selective oxidation of
n-butane to maleic anhydride (MA). It is known that the catalytic
performance of vanadium phosphates depends on the history of
preparation of the catalyst precursor, VOHPO4·0.5H2O [1,2].
Typically, the active catalystconsists(VO)2P2O7 in combination
with some V5+ phosphate and the topotactically transforma-
tion of the precursor to final catalyst [3]. Hence, controlling
the physicochemical properties of the precursor is an impor-
tant factor in determining the physicochemical properties and
the performance of the eventual catalyst following activation.
In general, V2O5 is used as a source of vanadium, and H3PO4
is used as a source of phosphorus. Hence, a reducing agent is
required to synthesise the V4+ precursor phase and a broad range
∗
Corresponding author. Tel.: +603 8946 6809; fax: +603 8946 6758.
E-mail address: yap@fsas.upm.edu.my (Y.H. Taufiq-Yap).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.06.047

Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
25
the presence of cyclohexane as solvent and dispersants based on
2.2. Catalysts characterisation
poly-12-hydroxystearic acid, they observed a significant reduc-
tion of particle size of the VOHPO4·0.5H2O (5–0.035 ␮m) that
gave final catalysts with high surface area following activation
(>40 m2 g−1 versus 9 m2 g−1 without ball milling). They also
found that these catalysts are particularly active under fuel rich
reaction conditions, when high MA yields can be obtained. Ayub
et al. [16] found that Bi-promoted vanadyl pyrophosphate milled
in air for relatively short times (5 min) could generate a mate-
rial that was amorphous in powder X-ray diffraction. They also
observed that the blossom secondary morphology of the cata-
lyst was lost under ball milling process. Although they further
noted that the ball milling process could increase the specific
surface area, the yield and conversion per unit surface area were
decreased. Hence, the full benefits that could be expected from
the enhanced surface area are unfortunately not produced. Wang
et al. [17] showed that the ball milling process broadened the
major powder X-ray diffraction peaks, indicating that both a
reduction of particle size and a fine variation in local structure
could occur in the milled solids, consequently leading to the
improvement of catalytic performance of the VPO catalysts.
The purpose of this study is to employ the mechanochemi-
cal pretreatment for morphological and textural modification of
vanadium phosphate catalysts synthesised via reduction of the
VOPO4·2H2O phase. The structure-activity relationships and
the nature and the role of lattice oxygen species of the catalysts
are discussed.
Brunauer–Emmett–Teller (BET) surface area measurements
werecarriedoutbyusingnitrogenadsorption–desorptionat77 K
using a ThermoFinnigan Sorptomatic 1990 instrument.
The bulk chemical composition was determined by using a
sequential scanning inductively coupled plasma-atomic emis-
sion spectrometer (ICP-AES) Perkin-Elmer Emission Spec-
trometer Model Plasma 1000.
The average oxidation states of vanadium in all the sam-
ples were determined by the method described by Niwa and
Murakami [18].
X-ray diffraction (XRD) patterns were obtained using a Shi-
madzu diffractometer model XRD-6000 employing Cu K␣ radi-
ation generated by a Phillips glass diffraction X-ray tube broad
focus 2.7 kW type on the catalysts at ambient temperature.
Scanning electron microscopy (SEM) analyses were carried
out using a JEOL JSM-6400 electron microscope.
Temperature-programmed reduction (H2-TPR) in H2/Ar
experiment was performed using a ThermoFinnigan TPDRO
1100 apparatus provided with a thermal conductivity detector.
TheH2-TPRanalysisoffreshcatalystswasdoneinH2/Arstream
(5% H2, 1 bar, 25 cm3 min−1) with raising the temperature from
ambient to 1173 K at 5 K min−1
.
2.3. Catalyst testing
The oxidation of n-butane to MA was carried out in a fixed-
bed microreactor with a standard mass of catalyst (250 mg).
The gases n-butane and air were fed to the reactor via cali-
brated mass flow controllers to give a feedstock composition
of 1.5% n-butane in air. The products were then fed via heated
lines to an on-line gas chromatograph for analysis. The reac-
tor comprised of a stainless steel tube with the catalyst held in
place by plugs of quartz wool. A thermocouple was located in
the centre of the catalyst bed and temperature control was typ-
ically 1 ◦C. Carbon mass balances of ≥96% were typically
observed.
2. Experimental
2.1. Catalysts preparation
The VOPO4·2H2O material was prepared by reacting V2O5
(30.0 g from Fluka) with H3PO4 (144 cm3, 85% from Merck) in
water (720 cm3) under reflux with continuous stirring for 24 h.
The yellow solid was then recovered by filtration, washed with
distilled water (100 cm3) and followed by acetone (100 cm3).
It was dried at 383 K for overnight. The VOPO4·2H2O phase
was then confirmed by XRD analysis. After that, the synthe-
sised VOPO4·2H2O (40.0 g) was suspended by rapid stirring
into isobutyl alcohol (800 cm3) and the mixture was refluxed
for 21 h with continuous stirring. A blue solid was recovered
by filtration and washed with distilled water (100 cm3) fol-
lowed by acetone (100 cm3). The resulting blue solid was further
treated in refluxing water (24 ml H2O/g solid) for 3 h to remove
VO(H2PO4)2 and was separated by filtration and dried overnight
in air at 385 K. Mechanochemical pretreatment of the precursor
was carried out by using a high energy planetary ball mill (model
Pulverisette 4 from Fritsch) with an agate bowl having 250 ml
volume together with fifty 10 mm diameter agate balls. About
18 g of the VOHPO4·0.5H2O and ethanol as solvent were put
together inside the bowl. The bowl spins around its own axis and
around a second axis outside its centre at 1400 rpm for 30, 60
and 120 min. The resulting unmilled and milled precursors then
undergone calcination in a flow of n-butane/air mixture (0.75%
n-butane in air) for 75 h at 673 K to generate the active catalysts,
(VO)2P2O7.
3. Results and discussion
3.1. BET surface area and chemical analysis
Table 1 shows the effect of the ball milling on the ratio of
P/V, BET surface area, oxidation state of vanadium as well as
percentage of V4+ and V5+ present in the catalysts. It is found
Table 1
Bulk composition, surface area, average oxidation states and percentage of V4+
and V5+ oxidation states present in catalysts
Milling
time (min)
P/V (ICP) Surface area Oxidation state of vanadium
(m2 g−1
)
V4+ (%) V5+ (%) Average
0
30
60
1.00
0.99
0.99
1.00
16
19
25
15
79
69
84
78
21
31
16
22
4.21
4.31
4.16
4.22
120

26
Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
that the value of P/V ratios were almost constant, i.e. ∼1.00,
indicating that the milling process does not alter the composition
of the catalysts. This observation is in agreement with the results
obtained by Ji et al. [19]. Furthermore, the bulk ratio of P/V for
all catalysts fell in the optimal atomic ratio of P/V in the range
of 1.0–1.2 for producing the (VO)2P2O7 phase [20].
The surface area for 30 min milled catalysts increased from
16 m2 g−1 (unmilled) to 19 m2 g−1 and then further increased to
25 m2 g−1 after 60 min of milling. The surface area however was
drastically reduced to 15 m2 g−1 for the120 min milled catalyst.
Zazhigalov et al. [13] explained that the ball milling process
firstly caused repeated fracture of the solid, followed by dimin-
ished particle size and a consequent increase in the surface area.
Rougier et al. [21] further noted that the fracture of the solid only
occurred at a short period of milling process and that longer ball
milling process would cause a phenomenon of reagglomeration.
The low surface area of 120 min milled material may be due to
the agglomeration of unstable particles which are formed by the
fracture of the solid.
The average vanadium oxidation number of the material that
was milled for 30 min was found higher (4.31) compared to the
unmilled counterpart (4.21). This was due to an increment of
V5+ oxidation state from 21% to 31%. However, the amount
of the V4+ was increased to 84%, corresponding to a decrease
in the average vanadium oxidation state to 4.16 for the catalyst
milled for 60 min. Further milling to 120 min slightly decreased
the V4+ phase to 78% whereas the amount of V5+ was increased
to 22%.
Fig. 2. XRD patterns of unmilled and milled catalysts.
(0 0 1), (1 0 1), (0 2 1), (1 2 1), (2 0 1), (1 3 0), (0 4 0) and (3 3 1)
planes, respectively.
The (VO)2P2O7 phase (JCPDS File No. 34-1381) shown in
Fig. 2 gave main peaks at 2θ = 22.6◦, 28.2◦ and 29.7◦ which
correspond to (0 2 0), (2 0 4) and (2 2 1) reflections, respectively.
These patterns show that the intensities of the peaks for ball
milled catalysts are markedly decreased as compared to the
unmilled material. Furthermore, Table 2 indicates that the values
of the full width at half maximum (FWHM) were increased, i.e.
from 1.16◦ (unmilled) to 1.26◦ and 1.37◦ for the milling time
of 30 and 60 min, respectively, along the (0 2 0) plane and from
0.46◦ (unmilled) to 0.50◦ and 0.59◦ for the milling time of 30
and 60 min, respectively, along the (2 0 4) plane. However, fur-
ther milling to 120 min, the values of FWHM were decreased to
0.37◦ along the (0 2 0) plane and 0.49◦ along the (2 0 4) plane. By
using the Debye–Scherrer equation, it was found that the crys-
tallite size of the catalysts was reduced with increasing milling
duration, i.e. from 7.0 nm (unmilled) to 5.9 nm along the (0 2 0)
plane and from 17.8 nm (unmilled) to 13.9 nm along the (2 0 4)
plane. The agglomeration occurred in the 120 min milled cat-
alyst; its particle size slightly increased to 6.6 nm along the
(0 2 0) plane and 16.7 nm along the (2 0 4) plane and this is in
agreement with the evolution of the surface area of samples
3.2. X-ray diffraction (XRD)
The XRD patterns of the unmilled and milled precursors,
as shown in Fig. 1, gave only the characteristic reflections of
vanadyl hydrogen phosphate hemihydrate, VOHPO4·0.5H2O
with the main reflections that appeared at 2θ = 15.5◦, 19.7◦,
24.2◦, 27.1◦, 28.7◦, 30.4◦, 37.5◦ and 49.2◦ corresponding to
Table 2
XRD data of precursors
a
Milling
Relative intensity
I(0 0 1)/I(1 3 0)
FWHMb (◦)
time (min)
(0 0 1)
(1 3 0)
(0 0 1)
(1 3 0)
0
30
60
59
22
33
31
100
100
100
100
0.59
0.22
0.33
0.31
0.48
0.26
0.30
0.24
0.32
0.23
0.24
0.27
120
a
Ratios of relative peak intensities.
Full width at half maximum.
b
Fig. 1. XRD patterns of unmilled and milled precursors.

Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
27
upon milling (Table 1). Zazhigalov et al. [13] reported that the
reduction of crystallite size, the formation of lattice defects and
strain or stress in the bulk of the crystallites, and the increase
in amorphous phases in the milled materials are contributed by
a decrease in integral intensities of X-ray diffractograms and
an increase in FWHM. In addition, increasing milling time also
led to an increment in the values of I(0 2 0)/I(2 0 4), indicating that
the exposure of (0 2 0) plane was increased which contained
the vanadyl group. Previous literatures [9,22] reported that the
best effect in the improvement of catalytic performance can be
reached by an increase in the relative exposure of (0 2 0) plane
of the (VO)2P2O7 because this plane is involved in the reaction
of partial oxidation of n-butane to MA.
process of ball milling. However, further milling to 120 min
(Fig. 3d) made these crystal platelets broken into smaller pieces
thus decreasing the exposure of the surface plane. These smaller
platelets are unstable and tendency to agglomerate, causing a
decrease in the surface area of the catalyst.
3.4. Temperature-programmed reduction (TPR) in H2/Ar
In order to investigate the effect of ball milling on the redox
properties, amount and nature of the oxygen species of the cat-
alysts, H2-TPR experiments were performed on the unmilled
and milled catalysts. Fig. 4 shows the profiles of H2-TPR, and
total amount of oxygen removed and the values of reduction
activation energies are shown in Table 3.
3.3. Scanning electron microscopy (SEM)
The unmilled catalyst gave characteristic two reduction peaks
at the region of 400–1100 K. These peaks occurred at 779 and
990 K, where the first peak is the reduction of V5+ phase whereas
the second peak is assigned to the removal of lattice oxygen from
the active V4+ phase [23]. The amount of oxygen removed from
both peaks is 3.85 × 1020 and 1.20 × 1021 atom g−1, respec-
tively, with an oxygen ratio for V5+ to V4+ of 0.32. The
milled catalysts show similar reduction profiles to the unmilled
material. It was found that the 30 min milled catalyst slightly
increased the temperature of the first peak with 20 K of incre-
ment, but that of the second peak decreased by 14 K. Nev-
ertheless, 60 min milled catalyst also showed shifts of both
reduction peaks to lower temperatures, i.e. 760 and 953 K. How-
The surface morphologies of unmilled and milled catalysts
are shown in Fig. 3. The unmilled catalyst (Fig. 3a) is com-
posed of plate-like crystallites, which are arranged into the
characteristic rosette-shape clusters. These plate-like structure
is related to (VO)2P2O7 platelets that preferentially expose the
(1 0 0) plane [7]. After milling for 30 and 60 min as shown in
Fig. 3b and c, respectively, the catalysts lost their secondary
structure of rosette-shape and provide shear forces which
allowed the crystal platelets to slide away from one another,
thereby exposing more surface plane. This observation may
explain the reason for an increase in the surface area after the
Fig. 3. SEM micrographs of (a) unmilled, (b) milled = 30 min, (c) milled = 60 min and (d) milled = 120 min.

28
Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
Table 3
XRD data of catalysts
a
Milling time (min)
Relative intensity
(0 2 0)
I(0 2 0)/I(2 0 4)
FWHMb (◦)
(0 2 0)
Crystallite sizec (nm)
(2 0 4)
100
(2 0 4)
(0 2 0)
(2 0 4)
0
30
60
48
83
89
0.48
0.83
0.89
1.01
1.16
1.26
1.37
0.37
0.46
0.50
0.59
0.49
7.0
6.4
5.9
6.6
17.8
16.5
13.9
16.7
100
100
98
120
100
a
Ratios of relative peak intensities.
Full width at half maximum.
Crystallite size was calculated accordingly to Debye–Schrrer equation.
b
c
ever, a further milling to 120 min shifted the first reduction
peak to higher temperature, i.e. 789 K, but the second reduction
peak appeared at lower temperature, i.e. 960 K as compared to
unmilled catalyst. The amount of oxygen species released asso-
ciated with V5+ significantly increased to 5.69 × 1020 atom g−1
for 30 min milled catalyst. However, further milling to 60 min
reduced the amount to 3.29 × 1020 atom g−1. The catalyst sub-
jected to mechanochemical pretreatment for 120 min showed
almost identical amount of oxygen removed related to V5+,
i.e. 3.81 × 1020 atom g−1 as compared to unmilled material.
The amount for oxygen species removed from the second
reduction peak, which is attributed to the reduction of V4+
decreased to (9.35 and 9.81) × 1020 atom g−1 for milling time
of 30 and 120 min, respectively. Interestingly for 60 min milled
catalyst, the amount of oxygen significantly increased to
1.43 × 1021 atom g−1 with an increment of 19.2%. The oxygen
atomreleasedratioofV5+ toV4+ formilledcatalystsis0.61, 0.23
and 0.39 for milling time of 30, 60 and 120 min, respectively.
The results show that 60 min of milling in ethanol is the opti-
mum duration to obtain a highly active catalyst as this catalyst
possesses highest amount of active V4+–O− pair for n-butane
activation and an appropriate oxygen species ratio associated
with V5+ to V4+.
Fig. 5. n-Butane conversion as a function of the reaction temperatures.
3.5. Catalytic evaluation
Fig. 5 shows n-butane conversion as a function of the reaction
temperatures, whereas the selectivity to MA is plotted against
the reaction temperature in the range of 613–673 K is given in
Fig. 6. All catalysts exhibited an increase in n-butane conversion
with increasing temperature, and a reduction of MA selectiv-
ity when the temperature increased. The 60 min milled catalyst
Fig. 4. H2-TPR profiles of unmilled and milled catalysts.
Fig. 6. MA selectivity as a function of the reaction temperatures.

Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
29
Fig. 7. n-Butane conversion as a function of the gas hourly space velocity.
Fig. 9. MA selectivity as a function of the n-butane conversion (T = 673 K).
was found to be the most active in the whole range of reaction
temperatures tested as compared to others. With respect to the
n-butane conversion and MA selectivity as a function of the gas
hourly space velocity (GHSV) as shown in Figs. 7 and 8, respec-
tively, the 60 min milled catalyst also appeared the most active.
The 30 min milled catalyst was found to give the highest MA
selectivity as shown in Fig. 9.
Table 4 summarises the catalytic performance of unmilled
and milled catalysts at reaction temperature of 673 K and GHSV
of 2400 h−1. The results show that the mechanochemical pre-
treatment employed in the preparation of the catalysts had an
impact on the catalytic performance, especially the conversion
of n-butane. For the catalyst milled for 30 min, surprisingly
it drastically decreases the conversion of n-butane from 83%
(unmilled) to 63%. However, a further milling to 60 min caused
an increase in the conversion to 88%, which is an optimal value
milling duration since the conversion was shown significantly
decreased when milling was extended to 120 min, i.e. 69%. On
the other hand, the MA selectivity of 30 min milled catalyst was
found to be increased from 63% (unmilled) to 68%. A longer
milling duration reduced the MA selectivity to 59% and 60%
for 60 and 120 min milled catalysts, respectively. Nevertheless,
all milled samples gave lower intrinsic activity compared to the
unmilled catalyst (Table 5).
Although the 30 min milled catalyst produced slightly higher
surface area than the unmilled material, but it exhibited sig-
nificantly low n-butane conversion. Nevertheless, it showed an
improvementintheformationofvaluableproductbecauseitpro-
duced higher MA and less CO and CO2 as compared to those
without and longer times of milling. This may due to an increase
in the amount of lattice oxygen species removed associated with
the V5+ phase as shown in H2-TPR analysis. The V5+ sites was
not detectable by XRD analysis which may be isolated on the
surface of (VO)2P2O7 phase [24]. In the plot of the relationship
of n-butane conversion and MA selectivity to oxygen species
associated with V5+ phase as shown in Figs. 10 and 11, a good
correlation of the oxygen species associated with V5+ and MA
Fig. 10. n-Butane conversion as a function of amount of oxygen removed asso-
ciated with V5+
.
Fig. 8. MA selectivity as a function of the gas hourly space velocity.

30
Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
Table 4
Total amount of oxygen atoms removed, values of reduction activation energies and ratio for oxygen removal of V5+/V4+ by reduction in H2/Ar for unmilled and
milled catalysts
Milling
time (min)
Peak
Tmax (K)
Reduction activation
energy, Er (kJ mol−1
Total amount of oxygen
Total amount of oxygen
Ratio for oxygen
removal of V5+/V4+
)
removed (mol g−1
)
removed (atom g−1
)
1
2
779
990
130.3
165.5
6.39 × 10−4
2.00 × 10−3
2.64 × 10−3
3.85 × 1020
1.20 × 1021
1.59 × 1021
0
Total
0.32
0.61
0.23
0.39
1
2
799
976
133.6
163.2
9.45 × 10−4
1.55 × 10−3
2.50 × 10−3
5.69 × 1020
9.35 × 1020
1.50 × 1021
30
60
120
Total
1
2
760
953
127.1
159.4
5.47 × 10−4
2.38 × 10−3
2.93 × 10−3
3.29 × 1020
1.43 × 1021
1.76 × 1021
Total
1
2
789
960
131.9
160.5
6.33 × 10−4
1.63 × 10−3
2.26 × 10−3
3.81 × 1020
9.81 × 1020
1.36 × 1021
Total
it was proposed that charge carrier is p+ holes known as strong
oxidizing agents. They correspond to O− species in the solid
which would result from a charge transfer equilibrium creat-
ing V4+–O− ion pairs. Such a reactive pair associating with a
coordinative unsaturated V4+ and O− species should explain the
initial n-butane activation with H-abstraction by O− species to
form MA [31]. A good correlation was also shown between the
percentages of V4+ obtained by redox titration with the n-butane
conversion (Fig. 13). A longer milling duration to 120 min gave
a poorer performance which was explained by a lower surface
area and higher ratio of oxygen species removed from V5+ to
V4+ phases which was obtained from H2-TPR analysis. A lower
amount of active oxygen species, O−, which is related to V4+,
also contributed to the poor activity.
selectivity was observed. However, a large amount of this oxy-
genspecieswouldalsoreducetheconversionofn-butane. Earlier
researchers [3,25–28] reported the important of the lattice oxy-
gen associated with V5+ in determining the selectivity to MA.
Coulston et al. [29] reported that by using time-resolved X-ray
absorption spectroscopy, V5+ species are kinetically significant
for the production of MA. Thus, there is considerable evidence
that V5+ sites are a requirement if the catalyst is to simultane-
ously exhibit good activity and selectivity.
Highn-butaneconversionobtainedforthe60 minmilledcata-
lyst can be related to the reducibility of the catalysts because this
reaction involved a redox mechanism [30]. The lattice oxygen of
thiscatalystwasfoundtobemorereducible, reactiveandmobile.
It exhibited lower reduction peak temperature accompanied by
higher amount of oxygen species removed associated with V4+
phase which was suggested as the centre for of n-butane activa-
tion [28,31]. The n-butane conversion is plotted (Fig. 12) versus
the amount of oxygen removed associated with V4+ phase for
this four catalysts, which presents a direct correlation between
them. This result confirmed that this oxygen species assigned
to O− contributed to the activity of the catalyst. Due to the p-
type semi-conductive character of vanadium phosphate catalyst,
Table 5
Catalysts performance of unmilled and milled vanadium phosphate for the oxi-
dation of n-butanea
Milling n-Butane
Product selectivity (%) Intrinsic activityb
time
conversion
(%)
(×10−5 mol MA m−2 h−1
)
MA CO
CO2
(min)
0
30
60
83
63
88
69
63
68
59
62
18
15
19
17
19
17
22
21
4.18
2.96
2.71
3.69
120
a
Reaction conditions: 673 K, 1.5% n-butane in air, GHSV = 2400 h−1
.
Fig. 11. MA selectivity as a function of amount of oxygen removed associated
b
Intrinsic activity: mol maleic anhydride (MA) formed m−2 catalyst h−1
.
with V5+
.

Y.H. Taufiq-Yap et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 24–31
31
between the reaction rate and the oxygen species associated with
V4+. Thus, the dependence of the n-butane activity on the very
reactive oxygen species suggested that the release of O− anion
from the surface of (VO)2P2O7 is the rate determining step for
the oxidation of n-butane.
Acknowledgement
Financial assistance from Malaysian Ministry of Science,
Technology and Innovation is gratefully acknowledged.
References
[1] G.J. Hutchings, Appl. Catal. 72 (1991) 1.
[2] G.J. Hutchings, J. Mater. Chem. 14 (2004) 3385.
[3] G. Centi, F. Trifiro, J.R. Ebner, V.M. Franchetti, Chem. Rev. 88 (1988)
55.
Fig. 12. n-Butane conversion as a function of amount of oxygen removed asso-
ciated with V4+
[4] B.K. Hodnett, Catal. Rev. Sci. Eng. 27 (1985) 373.
[5] J.W. Johnson, D.C. Johnston, A.J. Jacobson, J.F. Brody, J. Am. Chem. Soc.
106 (1984) 8123.
.
[6] G.J. Hutchings, R. Olier, M.T. Sananes, J.C. Volta, in: V.C. Corbera´n, S.V.
Bello´n (Eds.), New Developments in Selective Oxidation II, Elsevier, Ams-
terdam, 1994.
[7] C.J. Kiely, S. Sajip, I.J. Ellison, M.T. Sananes, G.J. Hutchings, J.C. Volta,
Catal. Lett. 33 (1995) 357.
[8] G.J. Hutchings, M.T. Sananes, S. Sajip, C.J. Kiely, A. Burrows, I.J. Ellison,
J.C. Volta, Catal. Today 33 (1997) 161.
[9] V.A. Zazhigalov, J. Haber, J. Stoch, L.V. Bogutskaya, I.V. Bacherikova,
Appl. Catal. A: Gen. 135 (1996) 155.
[10] A.S. Horowita, C.M. Blackstone, A.W. Sleight, G. Teufer, Appl. Catal. 38
(1988) 193.
[11] K. Shima, M. Ito, M. Murayama, M. Hatano, Stud. Surf. Sci. Catal. 92
(1995) 355.
[12] K. Inumaru, T. Okuhara, M. Misono, Chem. Lett. (1992) 1955.
[13] V.A. Zazhigalov, J. Haber, J. Stoch, A. Kharlamov, L.V. Bogutskaya, I.V.
Bacherikova, A. Kowal, Solid State Ionics 101–103 (1997) 1257.
[14] M. Fait, B. Kubias, H.-J. Eberle, M. Estenfelder, U. Steinike, M. Schneider,
Catal. Lett. 68 (2000) 13.
[15] G.J. Hutchings, R. Higgins, Appl. Catal. A: Gen. 154 (1997) 103.
[16] I. Ayub, D.S. Su, M. Willinger, A. Kharlamov, L. Ushkalov, V.A. Zazhi-
galov, N. Kirillova, R. Schlo¨gl, Phys. Chem. Chem. Phys. 5 (2003)
970.
[17] X. Wang, L. Xu, X. Chen, W. Ji, Q. Yan, Y. Chen, J. Mol. Catal. A: Chem.
206 (2003) 261.
Fig. 13. n-Butane conversion as a function of V4+
.
4. Conclusions
The adoption of mechanochemical pretreatment on a catalyst
precursor in this study increased the full width at half maximum
(FWHM) and reduced the crystallite size of the catalysts and
consequently increased their surface area without affecting the
principal phase structures of VOHPO4·0.5H2O and (VO)2P2O7
as well as the P/V ratios. The increase in the surface area may
be associated with the morphologies of the catalysts which pre-
vent the crystal platelets from tightly stacking together, thereby
exposing more surface plane. However, an extension of milling
time to 120 min led to a decrease in the surface area of the cat-
alyst which due to the agglomeration of the unstable fractured
solids. The 30 min milled catalyst was found to increase the
average vanadium oxidation number of the catalyst from 4.21
(unmilled) to 4.31. However, further milling to 60 and 120 min
reduced the number to 4.16 and 4.22, respectively. The oxygen
species of 60 min milled catalyst was found to be the most reac-
tive with high amount of oxygen species (O−) associated with
the active V4+ phase. A good direct relationship was established
[18] M. Niwa, Y. Murakami, J. Catal. 76 (1982) 9.
[19] W. Ji, L. Xu, X. Wang, Z. Hu, Q. Yan, X. Chen, Catal. Today 74 (2002)
101.
[20] G. Centi, Catal. Today 16 (1993) 5.
[21] A. Rougier, S. Soiron, I. Haihal, L. Aymard, B. Taouk, J.-M. Tarascon,
Powder Technol. 128 (2002) 139.
[22] J. Haber, V.A. Zazhigalov, J. Stoch, L.V. Bogutskaya, I.V. Bacherikova,
Catal. Today 33 (1997) 39.
[23] B.T. Pierini, E.A. Lombardo, Mater. Chem. Phys. 92 (2005) 197.
[24] J.C. Volta, Top. Catal. 15 (2001) 121.
[25] M.L. Granados, J.C. Conesa, M. Fernandez-Garcia, J. Catal. 141 (1993)
671.
[26] M.T. Sananes, A. Tuel, G.J. Hutchings, J.C. Volta, J. Catal. 166 (1997)
388.
[27] J.C. Volta, C.R. Acad. Sci. Paris, Se´rie IIc: Chim./Chem. 3 (2000) 717.
[28] E. Bordes, Catal. Today 16 (1993) 61.
[29] G.W. Coulston, S.R. Bare, H. Kung, K. Birkeland, G.K. Bethke, R. Harlow,
H.N., P.L. Lee, Science 275 (1997) 191.
[30] G. Centi, F. Cavani, F. Trifiro, Selective Oxidation by Heterogeneous Catal-
ysis, Kluwer Academic/Plenum, New York, 2001.
´ ´
[31] J.-M. Herrmann, P. Vernoux, K.E. Bere, M. Abon, J. Catal. 167 (1997) 106.