VPO catalysts supported on H3PO4-treated ZrO2 highly active for n-butane oxidation

Article abstract of DOI:10.1016/j.jcat.2006.11.027

VPO catalysts supported on ZrO₂ and H₃PO₄-treated ZrO₂ were prepared for the first time by means of precipitation-deposition in an organic medium. The physicochemical properties of the catalysts were investigated by the BET, XRD, TEM, Raman, XPS, and H₂-TPR techniques. It was found that the way of H₃PO₄ treatment conducted on ZrO₂ has a significant effect on the nature of the support, as well as on the state and structure of the VPO component loaded on it. The discrepancies in the synergistic interaction between VPO and the different supports induce variation in catalytic performance. The change in VPO loading is found to have an effect on catalyst efficiency. At 673 K, the MA yield is 61 mol% over the 36%VPO/(H₃PO₄)-p-ZrO₂ catalyst, the best performance achieved so far among the supported VPOs.

Full text of DOI:10.1016/j.jcat.2006.11.027

Journal of Catalysis 246 (2007) 166–176
www.elsevier.com/locate/jcat
VPO catalysts supported on H PO -treated ZrO highly active
3
4
2
for n-butane oxidation
a,b
a,b
b,∗
b
a,∗
Ru-Ming Feng , Xiu-Juan Yang , Wei-Jie Ji , Yi Chen , Chak-Tong Au
a
Department of Chemistry, Center for Surface Analysis and Research, Hong Kong Baptist University, Kowloon Tong, Hong Kong, China
b
Key Laboratory of Mesoscopic Chemistry, Ministry of Education, School of Chemistry and Chemical Engineering, Nanjing 210093, China
Received 3 October 2006; revised 22 November 2006; accepted 22 November 2006
Available online 4 January 2007
Abstract
VPO catalysts supported on ZrO and H PO -treated ZrO were prepared for the first time by means of precipitation–deposition in an organic
2
3
4
2
medium. The physicochemical properties of the catalysts were investigated by the BET, XRD, TEM, Raman, XPS, and H -TPR techniques. It was
2
found that the way of H PO treatment conducted on ZrO has a significant effect on the nature of the support, as well as on the state and structure
3
4
2
of the VPO component loaded on it. The discrepancies in the synergistic interaction between VPO and the different supports induce variation in
catalytic performance. The change in VPO loading is found to have an effect on catalyst efficiency. At 673 K, the MA yield is 61 mol% over the
3
6%VPO/(H PO )-p-ZrO catalyst, the best performance achieved so far among the supported VPOs.
3 4 2
©
2006 Elsevier Inc. All rights reserved.
Keywords: H PO treatment; ZrO ; Supported VPO; n-Butane oxidation; Maleic anhydride
3
4
2
1
. Introduction
The partial oxidation of n-butane to maleic anhydride (MA)
effect on the nature and performance of a VPO catalyst [32,41,
2]. VPO catalysts prepared in an organic medium are known
4
to perform better than those prepared in an aqueous solution
[43]. Iso-butanol [9,19,44–48] and benzyl alcohol [24,49,50]
have been adopted for generating VPO catalysts in an organic
involves the transfer of a total of 14 electrons in a number of
consecutive reaction steps [1]. Despite the complexity, the re-
action is effectively catalyzed by vanadium phosphorus oxide
medium. It was reported that V5 -containing phases such as
+
(
VPO) catalysts [2–6]. In fact, it is the only industrialized cata-
α -VOPO and γ -VOPO were generated in supported VPO,
1
4
4
lyst system used exclusively for the oxidation of light alkanes.
Much effort has been devoted to advancing our understanding
of these catalysts [6–12], including the functions of the active
phases and the nature of active sites [13–30].
especially when the catalyst was prepared in aqueous solution
and that the presence of the α -VOPO and γ -VOPO phases
1
4
4
in large amount would lower n-butane conversion and/or MA
selectivity [21,35,51].
It could be advantageous to have the VPO catalysts sup-
ported in terms of better heat transfer, larger surface area to vol-
ume ratio of active component, better mechanical strength, and
other properties, and materials including SiO2, TiO2, Al2O3,
SiC, Al-containing MCM-41, and SBA-15 have been used as
supports [31–40]. It has been recognized that the use of spe-
cific supports and preparation methods can have a significant
ZrO is a versatile material that has been widely used as
catalyst support or support component in applications such as
2
(i) NO SCR with hydrocarbons [52], (ii) TWCs [53,54], (iii)
x
methanol synthesis and decomposition [55–57], (iv) CO hydro-
genation [58], (v) methane partial oxidation [59], (vi) methanol
steam reforming [60,61], and (vii) sulfated ZrO solid acids
2
[62,63]. The literature contains a wide variety of materials that
have been used to support VPO but not ZrO . To the best of
2
*
our knowledge, the present study is the first time that ZrO2 was
adopted to support VPO. We studied the dispersion of VPO on
ZrO2, and investigated the effect of preparation methodology—
Corresponding authors.
E-mail addresses: jiwj@nju.edu.cn (W.-J. Ji), pctau@hkbu.edu.hk
(
C.-T. Au).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.11.027

R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
167
particularly the treatment of support with H3PO4—on the na-
ture and performance of VPO.
The BET surface area was measured using an ASPS-2100 in-
◦
strument. The sample was outgassed at 300 C for 3 h, after
◦
which N2 adsorption was conducted at −196 C. Raman spec-
2
. Experimental
tra were collected using a LABRAM-HR Raman spectrometer
at RT with excitation source of 513 nm and power of 1.0 mW.
XPS measurement was performed using an X-ray photoelec-
tron spectrometer (VG ESCALAB Mark II) with 1253.6 eV
of MgKα radiation at settings of 10 kV and 20 mA. The
binding energies (BEs) were calibrated against the C 1s sig-
nal (284.6 eV) of contaminant carbon. The surface compo-
sition of elements was estimated on the basis of having the
corresponding peak areas normalized according to the Wag-
ner factor database. The TEM images were obtained with a
TEM-200CX transmission electron microscope. Temperature-
programmed reduction (TPR) was carried out in the temper-
2
.1. Catalyst preparation
The ZrO2 material was prepared by mixing 4.0 g of ZrOCl2·
H2O and 1.0 g of CTAB in aqueous medium (pH ꢀ 12) at
8
7
1
◦
0 C and then via a hydrothermal procedure in an autoclave at
◦
00 C for 48 h. The resulted solid was dried at 373 K for 10 h,
and is designated “dried precursor of ZrO2” hereinafter. Before
◦
VPO loading, the dried precursor was calcined at 550 C for
8
h and denoted as “calcined ZrO2” before being used directly
as support or subject to H3PO4 treatment (i.e., 1 g of ZrO2/
◦
2
0 ml H3PO4 solution of 1 mol/L, at 70 C, for 12 h). The fil-
◦
ature range of RT ∼ 900 C. The sample was reduced in 5%
tered solid was washed repeatedly with D.I. water until the pH
of filtering solution reached 7.0. For comparison, a portion of
◦
H2/Ar (40 ml/min) at a rate of 10 C/min.
“
dried precursor of ZrO2” underwent the same H3PO4 treat-
3
3
3
. Results and discussion
ment, and the resulting material was dried and calcined again,
as described earlier before VPO loading.
The loading of VPO in mixed iso-butyl/benzyl alcohols was
performed as described previously [40,64]. In brief, V2O5 was
first refluxed in a mixture of iso-butyl/benzyl alcohols (volume
ratio of 1:1) at 140 C for 6 h. Subsequently, PEG 6000 and a
particular support were introduced in suitable amounts. After
3
phoric acid (85%) was added dropwise to reach a P/V atomic
ratio of 1.2/1.0. After refluxing for another 7 h, the suspen-
sion was filtered, and the solid was washed with iso-butanol
and acetone, respectively, and dried in air at 120 C for 24 h.
The powder was pressed into pellets, crushed, and sieved (40–
.1. Characterization
.1.1. BET, TEM, and XRD
Despite the use of CTAB in an attempt to create mesopores
◦
in ZrO2, the results of small-angle XRD and N2 adsorption–
desorption isotherms (not shown) suggest the absence of meso-
pores after calcination at 550 C for 8 h. The specific surface
areas of various samples are given in Table 1. H3PO4 treatment
of “calcined ZrO2” caused an increase in surface area from 27.0
to 35.7 m /g, whereas H3PO4-treatment of “dried precursor of
◦
0 min, reaction temperature was lowered to 130 C and phos-
◦
2
◦
ZrO2” resulted in a slight decrease in surface area from 27.0 to
2
2
2.7 m /g. In the former case, increase in surface area is likely
6
0 mesh). The catalyst prepared by loading VPO on H3PO4-
to be due to the partial dissolution of surface ZrO2 by H3PO4
and the induced porosity on the support surface. Therefore, the
method of H3PO4 treatment can have an affect on the texture of
the supports. For the serial samples of VPO/(H3PO4)-p-ZrO2,
surface area decreases slightly with increasing VPO loading
(Table 1). In addition, the surface areas of VPO/(H3PO4)-p-
ZrO2 samples are rather close to those of (H3PO4)-p-ZrO2
support. The TEM images of the supports show grains of 200–
treated “dried precursor of ZrO2” is designated VPO/(H3PO4)-
p-ZrO2, and that prepared by loading VPO on H3PO4-treated
“
calcined ZrO2” is designated VPO/(H3PO4)-ZrO2 hereinafter.
Before characterization and performance evaluation, all of
the samples were subject to in situ activation. The sample was
◦
heated from room temperature (RT) to 400 C at a rate of
2
◦
C/min in the reaction mixture (i.e., 1.5% n-butane/air), and
kept at this temperature for at least 15 h. The freshly activated
catalyst was gray in color.
3
00 nm (Fig. 1). H3PO4 treatment causes no obvious change
in the size of the support particles (Fig. 1B). Nevertheless,
the surface porosity of (H3PO4)-p-ZrO2 (Fig. 1D) is enhanced
compared with that of ZrO2 (Fig. 1C). Dispersion of the VPO
component on the (H3PO4)-p-ZrO2 support is rather uniform
2
.2. Catalyst evaluation
The 0.5-g catalyst samples were evaluated for the partial ox-
idation of n-butane to MA in a quartz fixed-bed microreactor
(Fig. 1E), as illustrated in the close-up TEM image of 36%
(
Φ = 8 mm) with a continuous reactant downflow. The feed
composition was 1.5% n-C4H10, 17.5% O2 (in volume), and
Table 1
balance N2. The gas hourly space velocity (GHSV) was in the
range of 1200–3000 h . An on-line GC system was used to an-
alyze the outlet effluent; the carbon balance was usually >95%.
The specific surface areas of supports and selected supported VPO catalysts
−
1
2
Sample
ZrO
SBET (m /g)
27.0
35.7
22.7
22.8
27.2
24.2
20.5
2
(
H PO )-ZrO
3 4 2
2
.3. Characterization
(H PO )-p-ZrO
3 4 2
3
2
3
4
6% VPO/(H PO )-ZrO
3 4 2
7% VPO/(H PO )-p-ZrO
Powder X-ray diffraction was conducted using a Rigaku
automatic diffractometer (Rigaku D-MAX) with monochrom-
atized CuKα radiation (λ = 0.1541 nm) at a setting of 30 kV.
3
4
2
2
2
6% VPO/(H PO )-p-ZrO
3
4
5% VPO/(H PO )-p-ZrO
3
4

168
R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
◦
Fig. 1. Representative TEM images of (A) the dried ZrO precursor; (B) the calcined (at 550 C) H PO -p-ZrO (H PO concentration: 1 mol/L); (C) the close-up
2
3
4
2
3
4
image of (A); (D) the close-up image of (B); (E) the activated 36% VPO/p-ZrO (H PO ) catalyst; and (F) the close-up image of a particle of (E).
2
3
4
VPO/(H3PO4)-p-ZrO2 (Fig. 1F). The results of TEM investi-
gation are in good agreement with surface area measurements.
The wide-angle XRD patterns of samples with or without
H3PO4-treatment, shown in Fig. 2, reveal that when the cal-
cined ZrO2 was used directly for VPO loading, the signal of
was that of monoclinic ZrO2 [65] (Fig. 2A). The poor presence
of (VO)2P2O7 crystallite is similar to the case when VPO was
supported on Al-containing MCM-41 [39]. It is apparent that
the method of treating ZrO2 by H3PO4 has an affect on VPO
phase composition. Over VPO/(H3PO4)-p-ZrO2, the XRD pat-
tern indicates the presence of amorphous ZrO2 and the predom-
(
VO)2P2O7 phase was hardly observed and the essential signal

R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
169
inant (VO)2P2O7 phase; the signals of the latter are similar to
those of VPO supported on SBA-15 [40] and on fumed SiO2
64]. Over VPO/(H3PO4)-ZrO2, the signals can be attributed
[
to the phases of highly distorted monoclinic ZrO2 and crys-
talline (VO)2P2O7 species (minor) (Fig. 2A). We deduce that
the treatment of H3PO4 not only suppresses the crystallization
of the ZrO2 precursor, thereby reducing ZrO2 crystallinity, but
also affects the state and structure of VPO species loaded on
the support. Fig. 2B illustrates the effect of H3PO4 concen-
tration adopted in H3PO4 treatment on phase composition of
supported VPOs. It can be seen an increasing H3PO4 concentra-
tion enhances the vanadyl pyrophosphate phase. At 1.0 mol/L
H3PO4 concentration, XRD signals of ZrO2 diminish the most,
suggesting a profound structural disorder (amorphous) of the
support material. Over VPO/(H3PO4)-p-ZrO2 catalysts, the in-
tensity of (VO)2P2O7 signals first increase at increasing VPO
loading to 36% but then decline at a VPO loading of 45%
(A)
(Fig. 2C), likely due to interaction between VPO and the sup-
port.
3
.1.2. Raman
Fig. 3A gives the Raman spectra of pure VPO and sam-
ples of VPO supported on H3PO4-treated ZrO2. The spectrum
of unsupported VPO shows four well-resolved Raman bands
−
1
at 928, 1036, 1136, and 1185 cm . The major band at ca.
−1
2−
9
28 cm can be attributed to ν(as) P–O–P in P2O unit of
VO)2P2O7, the weak bands at 1136 and 1185 cm are char-
7
−
1
(
acteristics of (VO)2P2O7 species [9,10,22,51,64], and the band
−
1
5+
at 1036 cm can be related to V -containing species [19,22].
−
1
For the supported VPOs, the shoulders in the 950–970 cm
region indicate the coexistence of vanadyl pyrophosphate and
orthophosphate (γ -VOPO4, α1-VOPO4, and VOPO4·2H2O)
−
1
phases, and the bands at 1023, and 1030 cm confirm the
(B)
presence of V5 -containing species in small proportion [9,10,
+
5
1,64]. In Fig. 3B, the impact of H3PO4 concentration on the
VPO/(H3PO4)-p-ZrO2 catalysts is obvious. At a H3PO4 con-
centration of 1.5 mol/L, and in terms of the intense bands at
−
1
9
1
26, 1123, and 1186 cm as well as the noticeable bands at
−1
010 and 1043 cm , enhanced formation and structural regu-
larity of vanadyl pyrophosphate and orthophosphate phase are
seen. At a H3PO4 concentration of 1.0 mol/L, the signal of
5+
−1
V
-containing species in the range of 1000–1100 cm is the
lowest, and the structural regularity of the vanadyl pyrophos-
phate phase is also low because the intensity of the character-
−
1
istic bands at 926, 1123, and 1186 cm is notably decreased.
Different from the silica-supported VPOs [51,64], the VPOs in
the VPO/(H3PO4)-p-ZrO2 catalysts exhibit declining Raman
signals of vanadyl pyrophosphate species with a rise in VPO
loading (Fig. 3C), suggesting possible interaction between VPO
and support at high VPO loadings. All of the Raman observa-
tions are concordant with the XRD results.
(C)
Fig. 2. The wide-angle XRD patterns of (A) different ZrO -supported VPOs (36
2
3
.1.3. H2-TPR
wt% loading) as well as unsupported VPO, (B) various VPO/(H PO )-p-ZrO
3
4
2
Fig. 4A reports the reduction behaviors of the pure and sup-
samples for which the support precursor was treated with different concentra-
tions of H PO , and (C) the serial VPO/(H PO )-p-ZrO samples with various
ported VPO catalysts. For pure VPO sample, the TPR profile
3
4
3
4
2
◦
shows a main peak around 784 C, attributable to the removal of
VPO loadings as well as the unloaded (H PO )-p-ZrO (H PO concentra-
3 4 2 3 4
lattice oxygen related to V4 species in the (VO)2P2O7 phase.
+
tion: 1 mol/L).

170
R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
◦
The shoulder ranging from ca. 400 to 700 C is attributed to the
removal of lattice oxygen related to V5 species [10,64]. Over-
all, the reduction behaviors of supported VPOs are strongly
dependent on the history of support preparation/pretreatment.
If “calcined ZrO2” was used as support (in the case of 36%
+
4
+
VPO/ZrO2), the reduction of lattice oxygen related to V
◦
species shifts toward higher temperature (800 C), indicating
that the lattice oxygen is relatively less reactive. The main
◦
reduction peak of 36% VPO/(H3PO4)-p-ZrO2 is at 736 C,
◦
whereas that of 36% VPO/(H3PO4)-ZrO2 is at 770 C, indi-
cating that the former can be reduced more easily. As already
◦
mentioned, the shoulders in the ca. 400–700 C range are ow-
5+
ing to the reduction of V -containing species. Based on peak
areas, the amount of V5 species in 36% VPO/ZrO2 is the low-
est and that in 36% VPO/(H3PO4)-ZrO2 is the highest among
the three supported VPOs. Previous studies indicated that the
+
presence of certain amount of V5 species in VPO catalysts
was favorable for MA formation [15,21,24,27,64]; thus the 36%
VPO/(H3PO4)-p-ZrO2 sample containing a medium amount of
+
(A)
5+
V
species may function better for the reaction, and this is
confirmed in the measurement of catalytic activity illustrated
later in Section 3.2.
Shown in Fig. 4B are the influence of H3PO4 concentra-
tion adopted in H3PO4 treatment on the reduction behavior
of the VPO/(H3PO4)-p-ZrO2 samples. Despite the variation of
H3PO4 concentration (0.5, 1.0, and 1.5 mol/L) shows no signif-
icant changes in the location of the main reduction peak (748,
◦
7
38, and 752 C, with that treated with H3PO4 of 1.0 mol/L the
most reactive), its affect on the area of the main reduction peaks
as well as the shoulders are obvious, suggesting that the relative
contents of vanadyl pyrophosphate and orthophosphate species
differ as related to the history of catalyst preparation, particu-
larly H3PO4 treatment. The TPR results are in accordance with
the Raman results described previously.
(B)
Fig. 4C presents the effect of VPO loading on the reduction
behavior of the VPO component in the VPO/(H3PO4)-p-ZrO2
samples. Note that the H3PO4-treated support material is essen-
tially not reducible in the adopted temperature range. Compared
with the main reduction peak of unsupported VPO, the main re-
duction peaks of the VPO/(H3PO4)-p-ZrO2 samples occur at
lower temperature. In other words, the lattice oxygen of the
supported VPOs is higher in reactivity. With increasing VPO
loading, the temperature of the main reduction peaks first in-
creases (from 18 to 27% loading) and then decreases (from 27
to 45% loading). This trend reflects the extent of VPO–support
interaction. Note that at the highest VPO loading of 45%, there
is a broad and intense shoulder, suggesting the presence of or-
thophosphate species in relatively larger proportions.
3.1.4. XPS
As shown in Table 1, the V 2p3/2 BEs of the bulk VPO and
supported VPO samples are in the range of 516.8–517.2 eV,
close to the value of 517.6 eV reported for well-crystallized
(
C)
Fig. 3. Raman spectra of (A) supported VPOs (36% loading) as well as un-
supported VPO, (B) the VPO/(H PO )-p-ZrO samples for which the support
(
VO)2P2O7 [1,10,66,67] but lower than that reported for vana-
3
4
2
dium pentoxide and V(IV) phosphate [1,68]. Therefore, the
vanadium oxidation states of these catalysts are approximately
4+, and the VPO components are mainly in the form of
precursor was treated with different concentrations of H PO , and (C) the serial
3
4
VPO/(H PO )-p-ZrO samples with different VPO loadings. (a) (VO) P O7;
3
4
2
2 2
(
b) α -VOPO ; (c) γ -VOPO .
1 4 4

R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
171
(
VO)2P2O7. The trend of V 2p3/2 BEs in the sample serial is
5+
also in accordance with the content of V -containing species
revealed by Raman and TPR investigations. The O 1s BEs of
supported VPO catalysts are essentially the same as those of
the unsupported VPO catalysts (ca. 531.3 eV). The P 2p BEs
of supported and bulk VPO samples are similar (ca. 134.0 eV)
except for the supported sample with 18% VPO loading. The P
2
p BE of the 18% VPO/(H3PO4)-p-ZrO2 sample is 133.5 eV;
the 0.5-eV shift to lower energy could be a result of surface zir-
conium phosphate formation in close vicinity to VPO. In fact,
the Zr 3d5/2 BEs of the supported samples are in the range of
1
83.1–183.5 eV (Table 2), approximately 1 eV higher than that
of ZrO2 (182.2 eV) but close to that of Zr(OH)4 (183.6 eV)
69]. This observation suggests that as a result of H3PO4 treat-
[
ment of ZrO2, generation of zirconium phosphate occurs, es-
pecially on the surface of the support material. This deduction
is also supported by (i) the closeness of the Zr 3d5/2 BE of
(A)
3
6% VPO/ZrO2 (182.4 eV) to that of ZrO2 (182.2 eV) [69] and
(
ii) the surface P/V atomic ratios depicted in Table 2. For the
unsupported VPO sample, the P/V ratio is 1.4, showing slight
enrichment of P element on the surface, as commonly observed
on supported and unsupported VPO catalysts [1,10,32,33,47].
However, the high P/V atomic ratios (2.4–2.9) observed over
the 18–45% VPO/(H3PO4)-p-ZrO2, 36% VPO/(H3PO4)-ZrO2,
and 36% VPO/ZrO2 samples are uncommon. In those H3PO4-
treated samples, the extra presence of P is likely due to the
coexistence of surface zirconium phosphate and VPO species.
In 36% VPO/ZrO2, however, the high surface P/V ratio could
result from amorphous nature of dispersed VPO on the sup-
port. Early studies indicated that an “excess” amount of surface
P may stabilize the pyrophosphate phase, favoring MA forma-
tion [1,70]. We propose that surface zirconium phosphate in
the present study could have a similar effect, as we discuss in
Section 3.2.1. In situ XPS studies of working VPO catalysts re-
vealed that the surface P/V and O/V atomic ratios deviate with
variations in feed composition and reaction temperature, sug-
gesting that for MA formation from n-butane, the complexity
involved demands active sites of varying natures [28].
(B)
The atomic ratio of surface zirconium to vanadium can be
examined as a function of VPO coverage [32,38]. As shown
in Table 2, the surface Zr/V atomic ratios of 18–45% VPO/
(
H3PO4)-p-ZrO2 and 36% VPO/(H3PO4)-ZrO2 are essentially
the same (0.5–0.7). This is not so for VPO/SBA-15, however,
in which the surface Si/V ratio decreases drastically with in-
creasing VPO loading [40]. The surface Zr/V ratio of 36%
VPO/ZrO2 (2.2) is significantly higher than those of H3PO4-
treated samples, suggesting that the dispersion of VPO on the
two types of catalysts should be rather different. As pointed out
earlier, the VPO dispersion on H3PO4-treated zirconia is fairly
uniform and likely to be in the form of a thin layer. Such a
manner of dispersion is probably an outcome of a unique inter-
action between VPO and surface zirconium phosphorate across
the catalyst surface. The similar surface concentrations of V and
P elements detected over 18–45% VPO/(H3PO4)-p-ZrO2 (Ta-
ble 2) supports the foregoing presumption. In fact, the TEM
(
C)
Fig. 4. Temperature-programmed reduction (TPR) profiles of (A) the sup-
ported VPOs with 36% loading as well as unsupported one, (B) the VPO/
H PO )-p-ZrO samples for which the support precursor was treated with
different concentrations of H PO , and (C) the serial VPO/(H PO )-p-ZrO
samples with different VPO loadings.
(
3 4 2
3
4
3
4
2
3 4 2
images of the 36% VPO/(H PO )-p-ZrO sample (Figs. 1E
and 1F) corroborate the results of XPS analysis. On the other

172
R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
Table 2
XPS results of bulk VPO and VPO/(H PO )-p-ZrO of different loadings
3
4
2
Sample
V 2p
P 2p
(eV)
O 1s
(eV)
Zr 3d
(eV)
Surface composition (atom%)
Atomic ratio
P/V Zr/V
3/2
5/2
(
eV)
V
P
O
Zr
Bulk VPO
517.0
517.2
517.1
516.8
516.9
516.7
516.8
134.0
133.9
134.1
133.9
133.5
133.6
133.8
531.3
531.2
531.5
531.4
531.3
531.4
531.1
−
11.4
6.5
6.9
7.1
5.8
6.1
5.1
16.3
16.7
16.3
16.7
17.1
16.5
12.0
72.3
72.8
73.0
72.9
73.1
73.7
71.5
−
1.4
2.6
2.4
2.4
2.9
2.7
2.4
−
a
a
a
a
b
c
45% VPO
36% VPO
27% VPO
18% VPO
36% VPO
36% VPO
183.1
183.5
183.5
183.2
183.3
182.4
4.0
0.6
0.5
0.5
0.7
0.6
2.2
3.7
3.3
3.9
3.7
11.4
a
In VPO/(H PO )-p-ZrO .
3
4
2
b
c
In VPO/(H PO )-ZrO .
3
4
2
In VPO/ZrO .
2
Scheme 1. Schematic illustration of VPO dispersion over (a) VPO/(H PO )-p-ZrO ; (b) VPO/ZrO ; (c) VPO/(H PO )-ZrO .
3
4
2
2
3
4
2
hand, the VPO dispersion on monoclinic ZrO2 is poor, possi-
bly a result of unfavorable interaction between the amorphous
VPO and monoclinic ZrO2.
Based on the results of various characterizations, a schematic
illustration of the VPO dispersion over different supports is
shown in Scheme 1.
89.1% and 68.8 mol%, respectively, giving a MA yield of 61.2
mol%. Compared with the activities reported over various sup-
ported VPOs, a MA yield of 61.2 mol% is significant and the
best achieved so far [31–40,51,64].
Fig. 6 plots MA selectivities over 36% VPO/(H3PO4)-p-
ZrO2, 36% VPO/(H3PO4)-ZrO2, and 36% VPO/ZrO2 as a
function of n-butane conversion. The MA selectivities decrease
with increasing butane conversion over the three catalysts.
However, at butane conversion of 45–90%, MA selectivity of
3
3
.2. Catalytic performance
.2.1. The effect of H3PO4 treatment
3 4 2
36% VPO/(H PO )-p-ZrO (68–88 mol%) is higher than that
The catalytic performance of unsupported VPO and the sup-
3 4 2
of 36% VPO/(H PO )-ZrO (48–60 mol%) and much higher
ported VPO catalysts is shown in Figs. 5A–5D. In the case
2
than that of 36% VPO/ZrO (10–34 mol%). Butane conversion
of MA generation from n-butane over unsupported VPO, the
typical n-butane conversion and MA selectivity at 400 C are
3 4 2
of 90–100% can be achieved on 36% VPO/(H PO )-p-ZrO .
◦
The deep oxidation of reactant and product on this active cat-
alyst becomes more significant at >90% butane conversion,
resulting in a drastic decline in MA selectivity. At n-butane con-
version of 40–70%, MA selectivity over 36% VPO/(H3PO4)-
p-ZrO2 reaches 75–88 mol%, significantly higher than those
found over the VPO catalysts supported on conventional SiO2
[32] as well as those on mesoporous MCM-41 and SBA-15 [39,
9
3.6% and 72.1 mol%, respectively (Fig. 5A). As for the 36%
VPO/ZrO2 catalyst, the n-butane conversion and MA selectiv-
ity under similar reaction conditions are 54.0% and 28.6 mol%,
respectively, giving a maximum MA yield of just 15.5 mol%
(Fig. 5B). The poor generation of MA can be ascribed to the
low reactivity of lattice oxygen (Fig. 4A) and the absence of
crystalline vanadyl pyrophosphate phase (Fig. 2A). Over the
◦
40]. Within the temperature range of 380–400 C, MA selec-
3
6% VPO/(H3PO4)-ZrO2 catalyst, the n-butane conversion and
tivity of 69–74 mol% and n-butane conversion of 74–89% are
sustained, giving a MA yield of 55–61 mol%.
Arrhenius plots based on the turnover rates of MA over
36% VPO/(H3PO4)-p-ZrO2 and 36% VPO/(H3PO4)-ZrO2 are
MA selectivity are 54.6% and 58.9 mol%, respectively, giving
an MA yield of 32.2 mol%. Over the 36% VPO/(H3PO4)-p-
ZrO2 catalyst, the n-butane conversion and MA selectivity are

R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
173
−
1
Fig. 5. n-Butane conversion and MA selectivity (GHSV = 1200 h ) as a function of reaction temperature over (A) unsupported VPO, (B) 36% VPO/ZrO ,
2
(
C) 36% VPO/(H PO )-p-ZrO , and (D) 36% VPO/(H PO )-ZrO .
3 4 2 3 4 2
rates obtained at constant SV and at typical butane conversions
at 65 ± 17% and 38 ± 14% for 36% VPO/(H3PO4)-p-ZrO2
and 36% VPO/(H3PO4)-ZrO2, respectively). The Ea values are
3.0 kJ/mol for 36% VPO/(H3PO4)-p-ZrO2 and 61.4 kJ/mol
(
2
for 36% VPO/(H3PO4)-ZrO2, fitting well with the catalyst ac-
tivities.
The influence of H3PO4 concentration (adopted in H3PO4
treatment) on catalyst performance is depicted in Table 3. The
H3PO4 concentration most suitable for support treatment is
1
.0 mol/L.
3
.2.2. The effect of VPO loading and GHSV
Fig. 8 illustrates the effect of VPO loading on the perfor-
mance of the VPO/(H3PO4)-p-ZrO2 samples. At an identical
◦
reaction temperature of 400 C, n-butane conversion first in-
creases with increasing VPO loading (from 18 to 36%) and then
decreases, remaining at 78% at a VPO loading of 45%. The MA
selectivity increases to ca. 68 mol% with an increase in VPO
loading from 18 to 36%. Similar results were found over VPOs
supported on silica [32,64] and SBA-15 [40]. At low VPO load-
ing, the support surface not covered by VPO may cause side
reactions; in addition, more reactive oxygen adspecies may be
Fig. 6. MA selectivity over 36% VPO/(H PO )-p-ZrO , 36% VPO/(H PO )-
3
4
2
3
4
ZrO , and 36% VPO/ZrO as a function of conversion.
2
2
shown in Fig. 7. Assuming that the measured surface areas of
these two samples equal that of VPO component, the apparent
activation energies (Ea) are estimated from the integral reaction

174
R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
(A)
◦
Fig. 8. The effect of VPO loading on the catalytic performance (at 400 C) of
−
1
serial VPO/(H PO )-p-ZrO samples (at GHSV = 1200 h ).
3
4
2
(B)
Fig. 7. The Arrhenius plots over (A) 36% VPO/(H PO )-ZrO and (B) 36%
3
4
2
VPO/(H PO )-p-ZrO catalysts. The temperature range adopted for both 36%
3
4
2
VPO/(H PO )-p-ZrO and 36% VPO/(H PO )-ZrO is 643 ± 20 K, at butane
3
4
2
3
4
2
conversions of 65 ± 17% and 38 ± 14%, respectively.
◦
Fig. 9. MA formation rate (at 400 C) over VPO/(H PO )-p-ZrO catalysts on
3
4
2
the basis of per unit mass of VPO component.
Table 3
Effect of H PO concentration adopted in H PO -treatment on catalyst perfor-
mance
3
4
3
4
species, and reduction behavior of the VPO/(H3PO4)-p-ZrO2
catalysts. Such alterations in catalyst characteristics should ac-
count for the observed difference in performance. As shown in
Fig. 9, the MA formation rate at 400 C, on the basis of per
unit mass of VPO, decreases with increasing VPO loading and
is notably higher than those derived over VPOs supported on
H PO
concentration
Temperature
( C)
C
S
Y
MA
(mol%)
3
4
butane
MA
◦
(mol%)
(mol%)
◦
0.5 mol/L
1.0 mol/L
1.5 mol/L
380
60.6
76.6
74.5
89.1
67.5
79.4
54.2
42.8
74.5
68.8
34.0
27.0
32.8
32.7
55.5
61.2
22.9
21.5
4
00
380
00
380
00
4
SBA-15 [40] and fumed SiO [64]. It can be envisaged that due
2
to the higher dispersion of VPO at lower loadings, more active
sites are available for the reaction on a “per unit mass” basis of
VPO component.
4
Finally, the effect of GHSV on performance of the 36%
VPO/(H PO )-p-ZrO catalyst was studied; the results are il-
lustrated in Fig. 10. It can be seen that n-butane conversion
generated on the VPO component of smaller dimensions [71];
consequently, MA selectivity is low. At a high VPO loading of
3
4
2
4
5%, MA selectivity decreases. As shown by the XRD, Raman,
declines from 89.1 to 60.2% with increasing GHSV in the range
−
1
and TPR characterization findings, an increase in VPO loading
from 36 to 45% produces significant changes in phase composi-
tion, relative content of vanadyl pyrophosphate/orthophosphate
of 1200–3000 h , whereas MA selectivity first increases mod-
estly from 68.8 to 71.2 mol% and then essentially remains
unchanged with increasing GHSV. With due consideration of

R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
175
Acknowledgments
This work was supported by the RGC, Hong Kong Special
Administration Region (Grant HKBU 200303). The research
activities conducted at NJU were financed by the MSTC (Grant
2
000048009).
References
[
[
[
1] S.K. Hodnett, Catal. Rev. 27 (1985) 373.
2] M. O’Connor, B.K. Hodnett, Appl. Catal. 42 (1988) 91.
3] Y. Kamiya, S. Ueki, N. Hiyoshi, N. Yamamoto, T. Okuhara, Catal. To-
day 78 (2003) 281.
[
[
4] J.B. Benziger, V. Guliants, S. Sundaresan, Catal. Today 33 (1997) 49.
5] I.J. Ellison, G.J. Hutchings, M.T. Sananes, J.C. Volta, Chem. Commun.
(
1994) 1093.
6] H.S. Horowitz, C.M. Blackstone, A.W. Sleight, G. Teufer, Appl. Catal. 38
1988) 193.
[
(
[
[
7] N. Hiyoshi, N. Yamamoto, T. Okuhara, Chem. Lett. (2001) 484.
8] G. Busca, F. Cavani, G. Centi, F. Trifiro, J. Catal. 99 (1986) 400.
Fig. 10. The effect of gas hourly space velocity (GHSV) on the performance of
◦
catalyst 36% VPO/(H PO )-p-ZrO (at 400 C).
3
4
2
[9] V.V. Guliants, J.B. Benziger, S. Sundaresan, I.E. Wachs, J.M. Jehng, J.E.
Roberts, Appl. Catal. A 28 (1996) 275.
[
[
[
10] X.S. Wang, L.J. Xu, X. Chen, W.J. Ji, Q.J. Yan, Y. Chen, J. Mol. Catal.
such factors as butane consumption, apparent MA yield, and
MA productivity, the GHSV most suitable for the target reac-
A 206 (2003) 261.
11] M.J. Lorences, G.S. Patience, F.V. Díez, J. Coca, Appl. Catal. A 263
(
2004) 193.
12] F. Garbassi, J.C.J. Bart, R. Tassinari, G. Vlaic, P. Lagarde, J. Catal. 98
1986) 317.
−1
tion is 1800 h
.
Correlating the results of performance with those of charac-
terization provides a structure–performance relationship. Good
performance is closely related to the following factors: (i) ap-
propriate phase composition (crystalline vanadyl pyrophos-
phate predominating on amorphous ZrO2) [72], (ii) a small
fraction of vanadyl orthophosphate species [73], (iii) higher
coverage and reactivity of the VPO component, and (iv) a
unique synergetic interaction between VPO and surface zirco-
nium phosphorate.
(
[13] V.V. Guliants, S.A. Holmes, J.B. Benziger, P. Heaney, D. Yates, I.E.
Wachs, J. Mol. Catal. A 172 (2001) 265.
[
[
[
14] M. Ruitenbeek, J.A. van Dillen, A. Barbon, E.E. van Faassen, D.C. Kon-
ingsberger, J.W. Geus, Catal. Lett. 55 (1998) 133.
15] K. Ait-Lachgar, A. Tuel, M. Brun, J.M. Herrmann, J.M. Krafft, J.R. Mar-
tin, J.C. Volta, M. Abon, J. Catal. 177 (1998) 224.
16] C.J. Kiely, A. Burrows, G.J. Hutchings, K.E. Bere, J.C. Volta, A. Tuel,
M. Abon, Faraday Discuss. 105 (1996) 103.
[
[
17] G. Centi, F. Trifirò, J.R. Ebner, V.M. Franchetti, Chem. Rev. 88 (1988) 55.
18] V.V. Guliants, J.B. Benziger, N.Y. Sundaresan, I.E. Wachs, Catal. Lett. 32
(1995) 379.
[
19] G.J. Hutchings, J.A. Lopez-Sanchez, J.K. Bartley, J.M. Webster, A. Bur-
rows, C.J. Kiely, A.F. Carley, C. Rhodes, M. Havecker, A. Knop-Gericke,
R.W. Mayer, R. Schlogl, J.C. Volta, M. Poliakoff, J. Catal. 208 (2002) 197.
20] R. Ebner, M.R. Thompson, Catal. Today 16 (1993) 51.
4
. Conclusion
[
[
In the present study, zirconia was used to support VPO for
21] M.T. Sananes-Schulz, A. Tuel, G.J. Hutchings, J.C. Volta, J. Catal. 166
the partial oxidation of n-butane to MA. When ZrO2 is used di-
rectly as support, the derived VPO/ZrO2 catalyst shows rather
poor performance. Over a VPO catalyst supported on ZrO2 pre-
treated with phosphoric acid, an excellent MA yield is possible.
The method of H3PO4 treatment has an affect on the nature
and consequently the performance of the catalysts. Only VPO
loaded on H3PO4-treated “dried ZrO2 precursor” before calci-
(
1997) 388.
[22] V.V. Guliants, J.B. Benziger, S. Sundaresan, I.E. Wachs, J.M. Jehng, J.E.
Roberts, Catal. Today 28 (1996) 275.
23] N. Hiyoshi, N. Yamamoto, N. Ryumon, Y. Kamiya, T. Okuhara, J. Catal.
[
[
[
2
21 (2004) 225.
24] N.H. Batis, H. Batis, A. Ghorbel, J.C. Vendrine, J.C. Volta, J. Catal. 128
(1991) 248.
25] (a) G.J. Hutchings, C.A. Desmartin, R. Olier, J.C. Volta, Nature 368
(1994) 411;
◦
nation has excellent performance. At 400 C, MA yield as high
(b) M. Conte, G. Budroni, J.K. Bartley, S.H. Taylor, A.F. Carley,
as 61 mol% can be obtained over the 36 wt% VPO/(H3PO4)-
p-ZrO2 catalyst, the best found to date over similar types of
supported VPOs. We find that the good performance is strongly
dependent on such critical structural features as phase compo-
sition of both VPO component and support, content of vanadyl
orthophosphate species, and coverage and reactivity of the VPO
component. It should be noted that certain structural features
of the catalyst are significantly affected by the unique syner-
gism between VPO and surface zirconium phosphorate on zir-
conia substrate of structural irregularity (from near amorphous
to crystalline).
A. Schmidt, D.M. Murphy, F. Girgsdies, T. Ressler, R. Schloge, G.J.
Hutchings, Science 313 (2006) 1270.
[26] G.J. Hutchings, R. Higgins, J. Catal. 162 (1996) 153.
[
27] J.-M. Herrmann, P. Vernoux, K.S. Bere, M. Abon, J. Catal. 167 (1997)
06.
1
[
28] H. Bluhm, M. Havecker, E. Kleimenov, A. Knop-Gericke, A. Liskowski,
R. Schlogl, D.S. Su, Top. Catal. 23 (2003) 99.
[29] G.W. Coulston, S.R. Bare, H. Kung, K. Birkeland, G.K. Bethke, R. Har-
low, N. Herron, P.L. Lee, Science 275 (1997) 191.
[
30] M. Lopez-Granados, J.C. Conesa, M. Fernandez-Garcia, J. Catal. 141
1993) 671.
(
[
31] K.E. Birkeland, S.M. Babitz, O.K. Bethke, H.H. Kung, G.W. Coulston,
S.R. Bare, J. Phys. Chem. B 101 (1997) 6895.

176
R.-M. Feng et al. / Journal of Catalysis 246 (2007) 166–176
[
[
[
32] R.A. Overbeek, A.R.C.J. Pekelharing, A.J. van Dillen, J.W. Geus, Appl.
Catal. A 135 (1996) 231.
33] R.A. Overbeek, P.A. Warringa, M.J.D. Combag, L.M. Visser, A.J. van
Dillen, J.W. Geus, Appl. Catal. A 135 (1996) 209.
34] V.A. Zazhigalov, Y.P. Zaitsev, V.M. Belousov, B. Partitz, W. Hanke,
G. Ohlmann, React. Kinet. Catal. Lett. 32 (1986) 209.
[51] Z.Q. Zhou, H.Y. Xu, W.J. Ji, Y. Chen, Catal. Lett. 96 (2004) 221.
[52] M. Fernandez-Garcia, A. Iglesias-Juez, A. Martinez-Arias, A.B. Hungria,
J.A. Anderson, J.C. Conesa, J. Soria, J. Catal. 221 (2004) 594.
[53] M.Y. Smirnov, G.W. Graham, Catal. Lett. 72 (2001) 39.
[54] A. Iglesias-Juez, A. Martinez-Arias, M. Fernandez-Garcia, J. Catal. 221
(2004) 148.
[
[
35] W.D. Harding, K.E. Birkeland, H.H. Kung, Catal. Lett. 28 (1994) 1.
36] J. Marc, C.C. Ledoux, P.H. Cuong, T. Vincent, K. Kostantinos, M. Patrick,
L. Jan, J. Catal. 203 (2001) 495.
[55] I.A. Fisher, A.T. Bell, J. Catal. 178 (1998) 153.
[56] K.D. Cheng, A.T. Bell, J. Catal. 193 (2000) 207.
[
57] Y. Usami, K. Kagawa, M. Kawazoe, Y. Matsumura, H. Sakurai, M. Haruta,
Appl. Catal. A Gen. 171 (1998) 123.
[
[
[
37] J.M.C. Buenoa, G.K. Bethkeb, M.C. Kung, H.H. Kung, Catal. Today 43
(
1998) 101.
38] W.Y. Nie, Z.Y. Wang, W.J. Ji, Y. Chen, C.T. Au, Appl. Catal. A 244 (2003)
65.
39] W.Y. Nie, X.S. Wang, W.J. Ji, Q.J. Yan, Y. Chen, C.T. Au, Catal. Lett. 76
2001) 3.
[58] W.J. Shen, M. Okumura, Y. Matsumura, Appl. Catal. A Gen. 213 (2001)
225.
2
[
59] K. Nakagawa, K. Anzai, N. Matsui, N. Ikenaga, T. Suzuki, Y.H. Teng,
T. Kobayashi, M. Haruta, Catal. Lett. 51 (1998) 163.
(
[
[
[
60] G.J.J. Bartley, R. Burch, Appl. Catal. 43 (1988) 141.
61] J.P. Breen, J.R.H. Ross, Catal. Today 51 (1999) 521.
62] V. Adeeva, J.W. de Haan, J. Jauchen, G.D. Lei, V. Schunemann, L.J.M.
van de Ven, W.M.H. Sachtler, R.A. van Santen, J. Catal. 151 (1995)
[
[
[
[
40] X.K. Li, W.J. Ji, J. Zhao, Z.B. Zhang, C.T. Au, J. Catal. 238 (2006) 232.
41] N.T. Do, M. Baerns, Appl. Catal. 45 (1988) 1.
42] N.T. Do, M. Baerns, Appl. Catal. 45 (1988) 9.
43] E.A. Lombardo, C.A. Sanchez, L.M. Cornaglia, Catal. Today 15 (1992)
364.
4
07.
44] D. Ye, A. Satsuma, A. Hattori, T. Hattori, Y. Murakami, Catal. Today 16
1993) 113.
45] F.B. Abdelouahab, R. Olier, N. Guilhaume, F. Lefebvre, J.C. Volta,
J. Catal. 134 (1992) 151.
[
[
63] H. Weiming, X. Yongde, Y. Yinghong, G. Zi, J. Catal. 196 (2000) 104.
64] C.Y. Xiao, X. Chen, Z.Y. Wang, W.J. Ji, Y. Chen, C.T. Au, Catal. To-
day 93–95 (2004) 223.
[
[
[
(
[65] W.J. Ji, J.Q. Hu, Y. Chen, Catal. Lett. 53 (1998) 15.
[66] H. Igarashi, K. Tsuji, T. Okuhara, M. Misono, J. Phys. Chem. 97 (1993)
46] F.J. Cabello Sanchez, J.A. Lopez-Sanchez, R.P.K. Wells, C. Rhodes, Catal.
Lett. 77 (2001) 189.
7065.
[
[
[
[
[
67] P. Delichere, K.E. Bere, M. Abon, Appl. Catal. A Gen. 172 (1998) 295.
68] V.I. Bukhtiyarov, Catal. Today 56 (2000) 403.
69] T.L. Barr, J. Phys. Chem. 82 (1978) 1801.
70] R.W. Wenig, G.L. Schrader, J. Phys. Chem. 90 (1986) 6480.
71] M. Abon, J.C. Volta, Appl. Catal. A 157 (1997) 173.
[
[
47] G. Landi, L. Lisi, J.C. Volta, Catal. Today 91–92 (2004) 275.
48] Y. Kamiya, S. Ueki, N. Hiyoshi, N. Yamamoto, T. Okuhara, Catal. Today
7
8 (2003) 281.
49] L.M. Cornaglia, C.A. Sanchez, E.A. Lombard, Appl. Catal. A 95 (1993)
17.
50] W.J. Ji, L.J. Xu, X.S. Wang, Z. Hu, Q.J. Yan, Y. Chen, Catal. Today 74
2002) 101.
[
[
1
[72] V.V. Guliants, Catal. Today 51 (1999) 255.
[73] J.C. Volta, Catal. Today 32 (1996) 29.
(