Selective modification of surface and bulk V5+/V4+ ratios and its effects on the catalytic performance of Mo{single bond}V{single bond}Te{single bond}O catalysts

Article abstract of DOI:10.1016/j.jcat.2006.12.019

Selective modification of surface and bulk V⁵⁺/V⁴⁺ ratios in Mo{single bond}V{single bond}Te{single bond}O catalysts has been achieved using nitric acid and saturated ammonia solutions as pH adjustors. The addition of nitric acid strongly suppressed V⁴⁺ in the bulk but encouraged it on the surface. Because the bulk-phase compositions of Mo{single bond}V{single bond}Te{single bond}O catalysts varied only slightly, the changes in the catalytic performance for selective propane oxidation to acrolein resulted mainly from the variation of the bulk and surface V⁵⁺/V⁴⁺ ratios. The high concentration of surface V⁵⁺ that was responsible for propane activation led to high selectivity of propylene and high conversion of propane. Increased bulk V⁴⁺ might result in oxygen vacancies, which promote the electron and oxygen transfer so as to increase the selectivity of deep oxidation products (CO_x). A low bulk V⁴⁺ concentration prefers selective oxidation products; the yield of acrolein was as high as 21.4%.

Full text of DOI:10.1016/j.jcat.2006.12.019

Journal of Catalysis 246 (2007) 382–389
www.elsevier.com/locate/jcat
5
+
4+
Selective modification of surface and bulk V /V ratios and its effects on
the catalytic performance of Mo–V–Te–O catalysts
a
a,∗
a
b
Yihan Zhu , Weimin Lu , Han Li , Huilin Wan
a
Institute of Catalysis, Zhejiang University (Xixi Campus), Hangzhou 310028, PR China
b
Department of Chemistry and Institute of Physical Chemistry, Xiamen University, Xiamen 361005, PR China
Received 17 October 2006; revised 21 December 2006; accepted 22 December 2006
Available online 2 February 2007
Abstract
Selective modification of surface and bulk V /V ratios in Mo–V–Te–O catalysts has been achieved using nitric acid and saturated ammonia
5+
4+
4+
solutions as pH adjustors. The addition of nitric acid strongly suppressed V in the bulk but encouraged it on the surface. Because the bulk-phase
compositions of Mo–V–Te–O catalysts varied only slightly, the changes in the catalytic performance for selective propane oxidation to acrolein
5
+
4+
5+
that was responsible for
resulted mainly from the variation of the bulk and surface V /V
ratios. The high concentration of surface V
4+
propane activation led to high selectivity of propylene and high conversion of propane. Increased bulk V might result in oxygen vacancies,
which promote the electron and oxygen transfer so as to increase the selectivity of deep oxidation products (COx). A low bulk V concentration
4+
prefers selective oxidation products; the yield of acrolein was as high as 21.4%.
©
2007 Elsevier Inc. All rights reserved.
5+
4+
Keywords: Propane; Selective oxidation; V /V ratio; Acrolein; Mo–V–Te–O catalysts; Surface and bulk
1
. Introduction
hexagonal M2 phase [13] (formulated as (TeO)M3O9, M = Mo,
V, or Nb) might act as an assistant phase and enhance the selec-
tivity of acrylic acid [14].
Propane is one of the most important constituents in natural
gas and crude oil, but it has not been widely used up to now.
In recent years, studies have focused mainly on one-step oxi-
dation of propane to partial oxidation products via VPO [1,2]
or MMO catalysts [3–7]. Acrolein (ACR) is a very important
product that has wide applications in both the chemical and
medical industries. Although the conventional catalytic process
from propylene to acrolein shows a very high yield of acrolein
Detailed studies on mechanisms of propane activation
[15–18] have proved that, based on those phases, the abstrac-
tion of methylene-H as propane activation step occurs mainly at
surface 5 V=O sites. The surface V=O group might either
generate a resonance form with a partial radical character for H
abstraction [15] or process a typical 5+2 activation mechanism
via dual M=O sites as advocated in VO /ZrO catalysts [16],
+
5+
x
2
which are still under debate. A Te4 site with a lone pair of elec-
trons has been shown to be active for the α-H abstraction of the
formed propylene, whereas a Mo6 site is responsible for the
chemisorption and O insertion of intermediate products to form
oxygenated products [15]. Additive studies have showed that
+
[
8–11], developing catalysts for the direct selective oxidation of
propane to acrolein is needed for optimal use of this abundant
resource.
+
Of all the catalysts studied, Mo–V–Te–(Nb)–O systems have
proven effective for selective propane oxidation to acrolein or
acrylic acid (AA) [5–7]. It is clear that the orthorhombic M1
phase (formulated as (Te O)M O , M = Mo, V, or Nb) func-
6+
5+
5+ 4+
6+
4+
Mo /Mo , V /V , and Te /Te pairs always coexist
on the surface and/or in the bulk of Mo–V–Te–(Nb)–O cata-
lysts [19,20]. Those pairs might perform as follows: (i) Change
the amounts of surface/bulk anion vacancies and promote elec-
2
20 56
tions mainly in the activation of propane [12]. The pseudo-
tron and oxygen transfer (as Fe3 /Fe does in Mo-Bi catalysts
+
2+
*
Corresponding author. Fax: +86 571 88273283.
E-mail address: luweimin@zju.edu.cn (W. Lu).
[11]), (ii) generate different amounts of highly reactive surface
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.12.019

Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
383
oxygen species (e.g., terminal M=O species), and (iii) serve as
Oxygen thermogravimetry/derivative thermogravimetry (O2-
TG/DTG) analysis was carried out in a Perkin–Elmer TGA7
thermogravimetric analyzer under pure oxygen atmosphere of
0.1 MPa and a ramp rate of 10 K/min. Data processing was
done using a Pyris TGA7HT software package. BET specific
surface area analysis of the catalysts was done using a Coulter
Ominisorp 100CX automated gas sorption analyzer.
X-ray photoelectron spectroscopy (XPS) experiments were
carried out on a RBD upgraded PHI-5000C ESCA system
(Perkin–Elmer) with MgKα radiation (hν = 1253.6 eV). The
X-ray anode was run at 250 W, and the high voltage was kept
a bulk oxygen reservoir, avoiding the reconstruction of surface
key structures during reaction [19]. The V5 /V pair has been
shown to perform a dominant role in changing the catalytic ac-
tivity of VPO catalysts for selective oxidation of n-butane [21].
It is difficult to make clear the function of each pair in the Mo–
V–Te–O multielement systems, due to the difficulty of altering
one pair while the others remain. Moreover, the different char-
acters of such pairs on the surface and in the bulk have rarely
been studied respectively, and no detailed relationship between
the variation of such pairs and catalytic performance has been
established.
+
4+
◦
at 14.0 kV, with detection angle at 54 . The base pressure of
−
8
In this paper, we used different pH adjustors for the prepa-
ration of Mo–V–Te–O catalysts. Both bulk and surface cases
were studied, because they often shared quite different proper-
ties. Results showed that we succeeded in changing the rela-
the analyzer chamber was about 5 × 10 Pa. The sample was
directly pressed to a self-supported disk (10 × 10 mm) and
mounted on a sample holder, then transferred into the analyzer
chamber. Binding energies were calibrated using the contami-
nant carbon (C 1s = 284.5 eV). The data analysis was carried
out using the RBD AugerScan 3.21 software provided by RBD
tive amount of V5 /V pairs both on the surface and in the
+
4+
6
+
5+
6+
4+
bulk rather than the Mo /Mo and Te /Te pairs for Mo–
V–Te–O catalysts. With similar phase compositions, obvious
changes in the performance of catalysts demonstrated the role
of vanadium as a significant redox element and the important
6
+
5+
5+ 4+
Enterprises. The relative ratios of Mo /Mo , V /V , and
6+
4+
Te /Te pairs were determined by smoothing and deconvo-
lution of Mo 3d, V 2p3/2, and Te 3d5/2 peaks, respectively, with
Gaussian curves. The background was subtracted using Shirley
integrated function.
relationship between the V5 /V pair and catalytic perfor-
+
4+
mance.
2
.3. Catalytic activity test
2
. Experimental
Catalytic performance experiments were carried out in a
2
.1. Catalyst preparation
fixed-bed quartz tubular reactor (6 mm i.d.; 200 mm long) un-
der atmospheric pressure. Fresh 200-mg catalyst samples with
similar volumes of catalyst bed were induced into the reactor
All catalysts were prepared by an aqueous solution reaction
method, using ammonium heptamolybdate (AHM) [(NH4)6
Mo7O24·4H2O], ammonium metavanadate (NH4VO3), and tel-
luric acid (H6TeO6) as starting materials. They were dissolved
in 20 ml of deionized water according to the corresponding
composition at 353 K. For unitary system, the solutions were
then used for the next step; three of the solutions were mixed to-
gether for ternary systems. The mixed solutions were adjusted
to desired pH values with the adjusters of aqueous nitric acid
−1
under the feedstock of O2/C3H8 = 1.08/1 (GHSV = 3000 h ).
The reactants and products were analyzed using an on-line gas
chromatograph equipped with Porapak Q (4.0 m × 1/8 in.) and
TDX-01 carbon molecular sieve (2.0 m×1/8 in.) columns. FID
and TCD detectors were used for the two-channel detection of
both columns. Catalytic reaction temperature varied from 623
to 773 K, and a blank experiment showed that homogeneous
reaction can be neglected under our reaction conditions.
(
1.0 M) and ammonia (saturated solution). The original Mo–V–
Te mixed solution exhibited a pH of 5.0 without the addition of
any pH adjustors. Then the solutions were evaporated at 353 K
to dryness and successively calcined at 873 K for 2 h in an N2
stream.
3
. Results and discussion
3
.1. Selective modification in unitary systems
As-synthesized MoOx, VOx, and TeOx unitary catalysts
2
.2. Catalyst characterization
were used to study the effects of pH adjustors on each of
6+
5+
5+ 4+
6+
4+
Mo /Mo , V /V , and Te /Te cation pairs, respec-
tively. Powder XRD results showed that the unitary catalysts
prepared under different pH values exhibited the presence of
MoO3 (JCPDS, 76-1003), Mo O (JCPDS, 72-0448), and
Mo O (JCPDS, 73-1536) in MoO ; V O (JCPDS, 71-
2235), VO (JCPDS, 76-0456), or V O (JCPDS, 85-0601) in
Powder X-ray diffraction (XRD) analysis was carried out us-
ing a Rigaku-D/Max-B automated powder X-ray diffractometer
◦
◦
◦
by the continuous scanning (4 /min from 3 to 80 of Bragg’s
angles 2θ) with CuKα radiation (λ = 0.15418 nm) operating
at 45 kV and 40 mA. Phase qualitative and reference intensity
ratio (RIR) semiquantitative analyses were done using a MDI
JADE v 6.5 software package.
4
11
9
26
x
6
13
2
2
5
VO ; and TeO (JCPDS, 78-1713) in TeO . All phases were
x
2
x
well crystallized, with no amorphous phases clearly observed,
which made the further semiquantitative analysis (RIR method)
more reliable. The results of the semiquantitative analysis for
the unitary catalysts are given in Table 1. According to the
phase proportion, we could simply calculate the average bulk
Laser Raman spectra (LRS) were collected under ambient
conditions using an HR LabRaman 800 system equipped with
a CCD detector. A green laser beam (λ = 514.5 nm) was used
for excitation.

384
Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
Fig. 1. (a) Average bulk oxidation states of the metal elements in MoOx, VOx, and TeOx unitary catalysts versus different pH values. (") TeOx, (Q) VOx,
(
ꢀ) MoOx. (b) The O -DTG patterns of VOx catalysts prepared under the pH of 1.0, 4.0, 6.0, and 9.0 with the curves from a to d, respectively. (c) Correlation
2
4+
4+
5+
between the O -TG mass gain and the bulk V /(V + V ) ratio for VOx prepared under: dot 1, pH 1.0; dot 2, pH 4.0; dot 3, pH 9.0; dot 4, pH 6.0.
2
Table 1
catalysts (O2-DTG patterns of MoOx and TeOx catalysts not
shown here).
Semi-quantitative phase analysis of as-synthesized MoOx, VOx, and TeOx cat-
alysts
The mass gain shoulders of VO2 generally appeared at lower
4+
Catalyst pH value
Relative proportion of phases (wt%)
Mo O
temperatures than those of V6O13, indicating that V in VO2
was more reducible than in V O . In conclusion, the addition
MoOx
1.0/4.0/6.0/9.0 MoO
Mo O
3
4
11
9
26
6
13
65/69/67/58
16/16/18/25 19/15/15/17
4+
of nitric acid suppressed the existence of V , but had little
effect on the Mo /Mo and Te /Te ratios in the bulk, in
accordance with the results of powder XRD.
VOx
1.0/4.0/6.0/9.0 V O
VO V O
6+
5+
6+
4+
6
13
2
2
5
39/63/75/90
0/0/25/10
61/37/0/0
TeOx
1.0/4.0/6.0/9.0 TeO
2
100/100/100/100
3
.2. Selective modification in MoV0.3Te0.25Ox mixed catalysts
chemical states of each element as showed in Fig. 1a. It was
obvious that the bulk chemical state of vanadium varied from
3.2.1. Bulk studies of the modifications
The Mo–V–Te–O ternary catalysts under different pH val-
ues were prepared with the composition of MoV0.3Te0.25Ox.
Freshly prepared catalysts were all black powders with very
small specific surface area, as suggested by BET experiments
4
.25 to 4.70 when the pH value changed from 9.0 to 1.0, which
indicated that the addition of nitric acid strongly and selectively
4+
suppressed the formation of V in the bulk. However, the bulk
6
+
5+
6+
4+
2
Mo /Mo and Te /Te ratios varied inappreciably. Actu-
(<1.0 m /g).
4+
ally, bulk Te always remained unchanged, as suggested by
the XRD results.
According to the powder XRD results, MoV0.3Te0.25Ox cat-
alysts were generally composed of orthorhombic TeMo5O16
◦
◦
◦
◦
O2-DTG experiments were carried out to verify the changes
in the chemical state of the VOx samples; the results are shown
in Fig. 1b. Based on the phase composition achieved, the mass
gain shoulders in region I (593–646 K) of Fig. 1b were due to
phase (JCPDS, 80-1238) with 2θ = 8.8 , 13.0 , 17.7 , 21.8 ,
◦
◦
◦
◦
◦
26.2 , 26.7 , 30.5 , 37.8 , and 48.4 and pseudohexagonal
M2 phase (TeM3O10, M = Mo or V [22]) with peaks at
◦
◦
◦
◦
◦
◦
◦
◦
14.0 , 22.1 , 23.2 , 25.3 , 26.3 , 28.2 , 33.1 , 36.1 , and
the oxidation of V4 in VO2, because the shoulder in curve d
of Fig. 1b became more intense when VO2 increased in curve c
and disappeared when no VO2 existed in curves a and b. Simi-
larly, the peaks at around 716 K in region II (646–810 K) might
+
◦
◦
◦
45.3 [13,23]. The absence of low-angle peaks at 6.6 , 7.8 ,
and 9.0 in all MoV0.3Te0.25Ox catalysts implies that no M1
◦
structured phase existed [13], though M1 phase sometimes
coexisted with M2 phase as in the Mo–V–Te–Nb–O catalyst
be attributed to the oxidation of V4 in V6O13 according to the
+
◦
system [22–25]. Comparison of peak areas at 30.5 for or-
◦
XRD semiquantitative results. Fig. 1c shows good linear cor-
thorhombic TeMo5O16 phase and 36.1 for M2 phase was used
relation between the bulk V4 /(V + V ) ratio and O2-TG
+
4+
5+
for calculating their relative quantities in a semiquantitative
view, because these two peaks did not overlap with other dif-
fraction peaks. XRD patterns and phase composition are shown
in Fig. 2. The results clearly show that the catalysts prepared
under pH 1.0 and 5.0 had similar phase composition and that
the TeMo5O16 phase increased slightly in the catalyst prepared
4+
mass gain. It was practical to reflect the bulk amount of V
by the O2-TG mass gains. This point would be further used in
the Mo–V–Te–O ternary catalysts. Single broad mass gain peak
around 751 K ranged from 677 to 850 K appeared in MoOx
catalysts, and no oxygen consumption peak occurred in TeOx

Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
385
Fig. 2. Powder XRD patterns of the MoV Te
Ox catalysts prepared under
0
.3 0.25
the pH of (a) 1.0, (b) 5.0, and (c) 9.0. And relative quantity of TeMo O
5
16
and M2 phases derived by semi-quantitative analysis. (") TeMo O phase,
5
16
(a) M2 phase.
Fig. 3. O -DTG patterns of the MoV Te
Ox catalysts prepared under the
2
0.3 0.25
pH of (a) 1.0, (b) 5.0, and (c) 9.0.
under pH 9.0. All of the phases were also well crystallized, as
suggested in the XRD patterns.
seemed to be different, as shown in Table 2, possibly due to seg-
regation. Furthermore, the surface concentration of vanadium
increased slightly when nitric acid was added into Mo–V–Te
mixed solutions.
O -DTG experiments (Fig. 3) of the Mo–V–Te–O catalysts
2
showed the changes in the chemical states of elements in the
ꢀ
bulk. The shoulders in region I (596–673 K) could be attributed
to the oxidation of V4 to stable V in the V-containing M2
+
5+
The results of peak fitting in XPS are shown in Fig. 4
and Table 2. For the deconvolution of Mo 3d core level, a
splitting energy of about 3.15 eV with the intensity ratio of
I(3d5/2):I(3d3/2) = 3/2 for each Mo 3d5/2–Mo 3d3/2 doublet was
used [26]. Consequently, the peaks with BE of 235.3 ∼235.6 eV
were attributed to Mo6 3d3/2, together with another weak line
phase, as it was within a similar oxidation temperature region
with V4 in VO . The main peak at 755 K in region II (above
+
ꢀ
2
5+
6+
6
73 K) might be attributed to the oxidation of Mo to Mo ,
because the peak situated at similar temperature region as in
ꢀ
the MoOx catalysts. The intensity of the peak in region II var-
+
ied inappreciably. Considering that Te4 exists only in TeOx
+
5+
catalysts, and based on the numerous related studies on the
around 234.5 eV for Mo 3d3/2 [19,26–28]. The two lines
showed an energy shift around 0.9 eV [19,20]. Similarly, the
bulk mixed Mo–V–Te–(Nb)–O systems [19,20], bulk Te4 was
+
6
+ 5+
thought to be stably formed in MoV0.3Te0.25Ox catalysts.
Mo 3d5/2 was at 232.4 eV and Mo 3d5/2 was at 231.3–
ꢀ
5+
5+
6+
It was observed that shoulders in region I decreased in
231.5 eV. The Mo /(Mo + Mo ) ratio was calculated via
5
+
Fig. 3c and disappeared in Fig. 3a. These results demonstrate
that the addition of nitric acid strongly suppressed the amount
the fitted peak areas and showed that surface Mo decreased
slightly when nitric acid was induced.
of bulk V4 in Mo–V–Te–O catalysts, and the addition of am-
+
The extremely sharp changes in V 2p3/2 core level reflected
that the surface modification was also selective for vanadium.
The curve-fitting results showed that the main peaks could be
deconvoluted into two lines at around 517.1 and 516.2 eV,
4+
monia also suppressed bulk V to a lesser extent. This effect
was less obvious with Mo6 /Mo pairs. Because the catalysts
+
5+
prepared under pH 1.0 and 5.0 had similar phase composition,
the variation of V5 /V ratio took place in the V-containing
M2 phase rather than in the TeMo5O16 phase. Thus, the M2
phase exhibited oxygen nonstoichiometry preferentially con-
+
4+
5+
which could be assigned to the presence of surface V and
V
4+
5+
, respectively [29,30]. From Fig. 4, the surface V of Mo–
V–Te–O catalysts was strongly suppressed when nitric acid was
induced (curve a), opposite to the situation in the bulk. The ad-
5
+
4+
4+
trolled by the bulk V /V ratio. When V substitutes for
5+
5+
V
in the bulk, some oxygen vacancies could be induced.
dition of ammonia could also decrease surface V species.
4+
Because the mass gain shoulders of V were in lower tem-
The deconvolution of Te 3d5/2 peaks showed that all of the
peaks can be fitted into two components. The line with BE of
perature ranges, those V4 sites might exhibit strong reducibil-
+
6+
ity and tend to be oxidized by oxygen species under reaction
576.9 eV was assigned to Te 3d5/2, and the line at 575.9 eV
4+
4+ 6+
was attributed to Te 3d5/2 (reported BEs of Te 3d5/2 and
conditions. Therefore, the bulk V plays a significant role in
promoting electron and oxygen transfer during reaction.
4+
Te 3d5/2 were 576.6 and 575.7 eV, respectively) [31]. The
6+
4+
presence of both Te and Te on the surface was in accor-
dance with results reported earlier [32]. The results of XPS
characterization showed that the addition of nitric acid and am-
monia as pH adjustors slightly changed the relative amount
3
.2.2. Surface studies of the modifications
XPS experiments were used to detect the surface modifica-
tions after addition of pH adjustors. The surface composition

386
Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
Table 2
Surface compositions, surface Mo /Mo , V /V , and Te /Te ratios by XPS, and I /I
6
+
5+
5+ 4+
6+
4+
4+
ratio by Raman spectroscopy
92 963
9
5
+
5+
5+
5+
6+
6+
6+
4+
)
Samples
Surface composition
V
/(V + V
)
Mo /(Mo + Mo
)
Te /(Te + Te
I
9
/I
92 963
pH 1.0
pH 5.0
pH 9.0
MoV
MoV
MoV
Te O
.28 0.26 4.05
0.00
0.53
0.44
0.00
0.02
0.06
0.44
0.50
0.51
0.07
2.71
0.74
0
0
0
Te
Te
O
O
.24 0.32 4.19
.22 0.23 3.78
Fig. 4. XPS spectrum of Mo 3d, V 2p , and Te 3d , and their deconvolution results. (a) pH 1.0, (b) pH 5.0, and (c) pH 9.0.
3/2
5/2
a maximum of 2.71 and dropped to a minimum for the catalyst
when pH 1.0. Combined with the XPS results, this indicates
that the surface concentration of V5 increased when V=O in-
+
5+
creased; thus, the enhanced surface V should be present in
the terminal 5 V=O form.
+
3
.3. Catalytic performance of MoV0.3Te0.25Ox catalysts
The reaction results of selective oxidation from propane to
acrolein on MoV0.3Te0.25O catalysts are shown in Table 3.
x
Carbon oxides, propylene, acrolein, acetic acid, acrylic acid,
formaldehyde, acetaldehyde, ethylene, methane, and acetone
were detected as products.
The addition of pH adjustors strongly affected the perfor-
mance of the Mo–V–Te–O catalysts. Compared with the reac-
tion results of pH 5.0, the selectivity of ACR increased greatly
(from 20.5 to 48.4%), and the selectivity of COx decreased
significantly, although the conversion of propane was also de-
creased at pH 1.0. At pH 9.0, the conversion of propane was
nearly the same as that at pH 5.0, and selectivity of ACR aug-
mented in some sort.
Fig. 5. Raman spectra of the MoV Te
Ox catalysts under pH value of
0
.3 0.25
(a) 1.0, (b) 5.0, and (c) 9.0.
of Te6 . Surface Te and Te species were in almost equal
amounts, accounting for the existence of oxygen reservoirs for
the α-H abstraction of propylene and allylic species as interme-
diate products [19].
+
6+
4+
The MoV0.3Te0.25Ox catalysts prepared under 1.0 and
pH 5.0 had similar phase composition. Moreover, only the
5
+ 4+
Raman spectra of the MoV0.3Te O catalysts are shown in
V
/V pair was selectively modified in the catalysts. The ob-
0.25
x
−
1
Fig. 5. The peaks above 900 cm were generally attributed to
vious changes in catalytic performance were due mainly to the
+
4+
the stretch mode of terminal M=O (M = Mo or V) species [33].
variation of surface and bulk V5 /V ratios. It indicated that
vanadium acted as a redox element in both surface and bulk.
The relationship between the selectivity of oxidation prod-
−
1
In detail, the band at 992 cm suggested the presence of termi-
−
1
nal V=O [34,35], and bands at 963 and 937 cm could be due
to typical Mo=O species [36,37]. The relative surface V=O
concentration was evaluated by calculating the intensity ratios
ꢀ
ucts and O2-TG mass gain in region I is shown in Fig. 6a. When
bulk V4 increased, the conversion of propane and selectivity
of COx rose. In contrast, the selectivity of (ACR+AA) as selec-
tive oxidation products decreased. Table 3 also shows that the
+
−
1
of the bands at 992 and 963 cm (I992/I963) in Table 2. For
the catalyst prepared under pH 5.0, the I992/I963 ratio reached

Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
387
ꢀ
5+
5+
4+
Fig. 6. Correlation between selectivity of products/conversion of propane (a) and the O -TG mass gain in region I , (b) and the surface V /(V + V ) ratio
2
a
in MoV Te
Ox catalysts. (c) Relationship between conversion of propane and the O -TG mass gain in VOx. Conversion of propane minus the yield of
0
.3 0.25
2
propylene.
Table 3
a
Catalytic performance on VOx and MoV Te
Ox catalysts prepared under different pH values
0.3 0.25
Catalysts
pH
Conversion
Selectivity of products (%)
Yield of
of propane (%)
b
b
b
b
b
ACR (%)
CO
CO
ACR
AA
AcA
Alde
C H
3
Others
2
6
MoV Te
Ox
1.0
44.2
59.9
57.0
13.5
21.8
20.4
3.4
6.6
7.2
48.4
20.5
27.6
0.6
4.5
7.8
3.6
3.5
2.6
5.1
4.4
1.7
23.5
31.3
27.9
1.9
7.4
4.8
21.4
12.3
15.7
0
.3 0.25
5.0
.0
9
a
−1
.
Temperature: 773 K, feedstock: O /C H = 1.08/1, GHSV: 3000 h
2
3 8
b
ACR: acrolein, AA: acrylic acid, AcA: acetic acid, Alde: formaldehyde and acetaldehyde, Others: acetone, ethylene, and methane, etc.
The more bulk V4 in the catalyst, the more oxygen va-
cancies produced, and the easier the transfer of electron and
oxygen species. Because their mobilities were enhanced, the
reactant and intermediate species adsorbed on the active sites
would have increased potential for attack by electrophilic oxy-
gen species. As a result, the activation of propane became more
facile, and the intermediate products tended to be overoxidized.
+
This indicates that specific bulk V4 /V ratio might offer
the best cooperation between electron and oxygen transfer and
propane activation rate and achieve the highest propane con-
+
5+
Fig. 7. 5 + 2 pathway of propane activation forming propylene by surface
5+
V=O and M=O (M = Mo, V or Te) dual sites.
4+
version. The correlation between the amount of bulk V (in
terms of O2-TG mass gain in region I , as shown in Fig. 6c)
and conversion of propane in VOx catalysts illuminated this
point. When the bulk V /V ratio reached 4.85, a maximum
selectivity of AA increased compared with that of ACR at in-
creasing pH values. Because AA is the consecutive O-inserting
ꢀ
product of ACR, the catalysts having more bulk V4 at pH 5.0
+
4+ 5+
or 9.0 were also prone to further oxidation of ACR.
propane conversion of 47.5% was achieved.
•
•
As stated earlier, the oxygen vacancies (V ) formed by the
5+
5+
O
On the other hand, surface V in the V=O form also pro-
bulk V4 (V ) could be oxidized by dioxygen, with bulk lat-
+
ꢀ
V
vided significant effects on the catalytic performance as Fig. 6b
tice oxygen species (OO(BL)) produced. The bulk lattice oxygen
5+
5+
4+
shows. As surface V /(V + V ) ratio increased from 0
ꢀ
species might oxidize the other active sites (M ) reduced by
M
to 0.53, the conversion of propane rose linearly from 44.2 to
reactants and intermediate products on the surface, thus produc-
ing surface lattice oxygen species (OO(SL)). The process using
Kröger–Vink notations is as follows:
5
9.9%. Because propylene was the intermediate product of the
propane activation, propylene selectivity also increased from
2
the surface V=O groups play an important role in promoting
propane activation and dehydrogenation. We considered a 5+2
pathway to be possible [16,17] as shown in Fig. 7.
5+
3.5 to 31.3% as surface V was enhanced. This confirms that
5+
1
ꢀ
••
2
2
V
+ V
+
O2(g) → 2VV(B) + OO(BL)
,
(1)
(2)
V(B)
O(B)
2
ꢀ
••
→ 2Vꢀ
VV(B) + OO(BL) + 2M
+ V
+ OO(SL)
It was necessary to eliminate the contribution of surface
V=O to the conversion of propane to make the contribution
M(S)
O(S)
V(B)
2MM(S) + V••
.
5+
+
O(B)

388
Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
Fig. 8. Powder XRD patterns of the MoV Te
Ox precursors under pH value of (a) 1.0, (b) 5.0, and (c) 9.0. Phases marked as (") (NH ) (TeMo O )·7H O,
0.3 0.25
4 6
6
24
2
(
2) (NH ) Mo O , (a) (NH ) V
O
·6H O.
4
2
4
13
4 6 10 28
2
of bulk V4 more clear. Because the propane conversion (C)
minus the yield of propylene (Y=) equals the total yield of all
other oxidation products, curve C–Y= in Fig. 6a was used to
represent the ability of further oxidation of the catalysts. The
results show that the promotion of electron and oxygen transfer
+
treatment of the precursors. The reason for the inverse trend of
changes in the V5 /V ratio between the bulk and surface re-
+
4+
mains under study.
4. Conclusions
4+
caused by more bulk V was also advantageous to the further
oxidation of intermediate products and increased propane con-
version.
As is well known, it is of great importance to add redox el-
ements (e.g., V, Fe, Cr, Ce, Sn, U) into bicomponent catalysts
to achieve better catalytic performance. Such redox elements
exhibit higher redox potential than the inserting O species
3
.4. Preliminary study of the reason for selective modification
5
+
4+
3+
2+
6+
5+
[
15] and present mainly as V /V , Fe /Fe , Cr /Cr ,
4+
3+
4+
2+
6+ 4+
Interestingly, the addition of nitric acid and ammonia as pH
Ce /Ce , Sn /Sn , and U /U pairs, possibly facilitat-
ing electron and oxygen transfer [11].
adjustors could change the ratios of V5 /V pairs both in the
bulk and on the surface. The precursors of the MoV0.3Te0.25Ox
ternary catalysts before heat treatment were characterized by
powder XRD, as shown in Fig. 8.
+
4+
In generally, the addition of pH adjustors (HNO3 or NH3
solution) during preparation strongly affected the bulk and sur-
face characteristics and catalytic properties of the Mo–V–Te–O
catalysts. The modifications were always rather selective for
vanadium as a vital redox element and preferentially changed
XRD results showed that the precursors under pH 5.0 and
.0 had similar compositions: (NH ) Mo O (JCPDS, 80-
4 2 4 13
9
0
the bulk and surface V5 /V ratios.
+
4+
757), (NH4)6(TeMo6O24)·7H2O (JCPDS, 26-0080), and some
4+
unknown phases. However, the precursor prepared under pH
.0 appeared to have quite different phase composition from
O2-TG/DTG and XPS experiments revealed that bulk V
1
or surface V5 in Mo–V–Te–O catalysts can be preferentially
+
5+
the above two cases. (NH ) Mo O , (NH ) V O ·6H O
suppressed by addition of pH adjustors. The surface V oc-
4
2
4
13
4 6 10 28
2
curred in the 5 V=O form. The bulk V might play a sig-
nificant role in promoting electron and oxygen transfer during
reaction. Moreover, the oxygen nonstoichiometry of the M2
+
4+
(
(
JCPDS, 82-0481), and some unknown phases were detected.
NH4)6[TeMo6O24]·7H2O has been shown to be the precur-
sor of monoclinic TeMo5O16 [38], the structure of which has
some relationship with the orthorhombic form of the TeMo5O16
phase. No known phase was recognized as the precursor of the
M2 phase, and the (NH4)6V10O28·6H2O phase in the precursor
did not significantly affect the phase composition of the cata-
phase might result from the variable bulk V5 /V ratios bal-
anced by oxygen vacancies. The M2 phase could possibly be
described as (TeO)M3O9−δ.
+
4+
The activation of propane is closely related to the surface
lysts under pH 1.0 and 5.0. Thus, different V5 /V ratios both
in the bulk and on the surface of the catalysts prepared under
different pH values were closely related to the unknown phases
in the precursors. We think that the V-containing M2 phase
might be yielded via a solid-state reaction (SSR) during heat
+
4+
V
5+
concentration, because the surface vanadyl group has
proven rather active. The conversion of propane depends mainly
5+
on the surface V=O groups. On the other hand, higher bulk
4+
V
concentrations are related to more overoxidation products
5+
4+
(CO ), because bulk V /V pairs are responsible for electron
x

Y. Zhu et al. / Journal of Catalysis 246 (2007) 382–389
389
and oxygen transfer. Less bulk V4 results in high selectivity
+
[15] R.K. Grasselli, J.D. Burrington, D.J. Buttrey, P. DeSanto Jr., C.G. Lug-
mair, A.F. Volpe Jr., T. Weingand, Top. Catal. 23 (2003) 5.
[16] F. Gilardoni, A.T. Bell, A. Chakraborty, P. Boulet, J. Phys. Chem. B 104
of ACR as the selective oxidation product in Mo–V–Te–O cata-
lysts. Consequently, optimum surface and bulk V5 /V ratios
+
4+
(2000) 12,250.
may be possible to achieve higher ACR yields.
[
[
17] K.D. Chen, E. Iglesia, A.T. Bell, J. Phys. Chem. B 105 (2001) 646.
18] G. Fu, X. Xu, X. Lu, H. Wan, J. Phys. Chem. B 109 (2005) 6416.
Acknowledgment
[19] M. Baca, J.M.M. Millet, Appl. Catal. A 279 (2005) 67.
[
[
[
20] J.M.M. Millet, H. Roussel, A. Pigamo, J.L. Dubois, J.C. Jumas, Appl.
Catal. A 232 (2002) 77.
21] M.T. Sananes-Schulz, A. Tuel, G.J. Hutchings, J.C. Volta, J. Catal. 166
This work was supported by the National Natural Science
Foundation of China (grant 20473071).
(1997) 388.
22] M. Aouine, J.L. Dubois, J.M.M. Millet, Chem. Commun. (2001) 1180.
References
[23] T. Ushikubo, K. Oshima, A. Kayou, M. Hatano, Stud. Surf. Sci. Catal. 112
1997) 473.
(
[1] G. Landi, L. Lisi, J.C. Volta, Chem. Commun. (2003) 492.
[2] G. Landi, L. Lisi, J.C. Volta, J. Mol. Catal. A 222 (2004) 175.
[3] A. Kaddouri, C. Mazzocchia, E. Tempesti, Appl. Catal. A 180 (1999)
[24] B. Zhu, H. Li, W. Yang, L. Lin, Catal. Today 93–95 (2004) 229.
[25] B. Solsona, J.M. López Nieto, J.M. Oliver, J.P. Gumbau, Catal. Today 91–
92 (2004) 247.
2
71.
[26] G.-T. Kim, T.-K. Park, H. Chung, Y.-T. Kim, M.-H. Kwon, J.-G. Choi,
Appl. Surf. Sci. 152 (1999) 35.
[27] J.-G. Choi, L.T. Thompson, Appl. Surf. Sci. 93 (1996) 143.
[28] K. Uchida, A. Ayame, Surf. Sci. 357–358 (1996) 170.
[29] G. Silversmit, D. Depla, H. Poelman, G.B. Marin, R. De Gryse, J. Elec-
tron. Spectrosc. Relat. Phenom. 135 (2004) 167.
[
[
4] H. Jiang, W. Lu, H. Wan, J. Mol. Catal. A 208 (2004) 213.
5] L. Chen, J. Liang, H. Lin, W. Weng, H. Wan, J.C. Védrine, Appl. Catal.
A 293 (2005) 49.
6] T. Ushikubo, Y. Koyasu, H. Nakamura, JP10045664 (1998).
7] T. Ushikubo, K. Oshima, A. Kayou, M. Vaarkamp, M. Hatano, J. Catal.
[
[
1
69 (1997) 394.
8] P. Botella, J.M. López Nieto, B. Solsona, J. Mol. Catal. A 184 (2002)
35.
[30] M. Demeter, M. Neumann, W. Reichelt, Surf. Sci. 41 (2000) 454.
[31] V.V. Guliants, R. Bhandari, J. Phys. Chem. B 109 (2005) 10,234.
[32] P. Botella, P. Concepción, J.M. López Nieto, Y. Moreno, Catal. Today 99
(2005) 51.
[
[
3
9] W. Gerhartz, Ullmann’s Encyclopedia of Industrial Chemistry, VCH,
Weinheim, 1985.
[33] G. Mestl, Catal. Rev. 40 (4) (1998) 451.
[
[
10] G. Creten, D.S. Lafyatis, G.F. Froment, J. Catal. 154 (1995) 151.
11] J.M.M. Millet, H. Ponceblanc, G. Coudurier, J.M. Herrmann, J.C. Vé-
drine, J. Catal. 142 (1993) 381.
[34] J.H. Clark, The Chemistry of Titanium and Vanadium, Elsevier, New York,
1968.
[35] W. Chem, L. Mai, J. Peng, Q. Xu, Q. Zhu, J. Solid State Chem. 177 (2004)
377.
[36] I. Lindqvist, Ark. Kemi 2 (1951) 349.
[37] A.W. Armour, M.G.B. Dreu, P.C.H. Mitchell, J. Chem. Soc. Dalton Trans.
1975 (1975) 1493.
[
[
[
12] M. Baca, A. Pigamo, J.L. Dubois, J.M.M. Millet, Top. Catal. 23 (2003)
39.
13] P. DeSanto Jr., D.J. Buttrey, R.K. Grasselli, C.G. Lugmair, A.F. Volpe,
B.H. Toby, T. Vogt, Z. Kristallogr. 219 (2004) 152.
14] M. Baca, M. Aouine, J.L. Dubois, J.M.M. Millet, J. Catal. 233 (2005) 234.
[38] I.L. Botto, C.I. Cabello, H.J. Thomas, Mater. Chem. Phys. 47 (1997) 37.