The oxygen isotopic exchange reaction on vanadium oxide catalysts

Article abstract of DOI:10.1006/jcat.1999.2490

The reactivity of lattice oxygen of vanadium oxide catalysts was studied with the oxygen isotopic exchange reaction. The reactivity of pure V₂O₅ is compared with the reactivity of Li_0.33V₂O₅, V_2<

Full text of DOI:10.1006/jcat.1999.2490

Journal of Catalysis 185, 415–422 (1999)
Article ID jcat.1999.2490, available online at http://www.idealibrary.com on
The Oxygen Isotopic Exchange Reaction on Vanadium Oxide Catalysts
,1
C. Doornkamp, M. Clement, X. Gao,† G. Deo,‡ I. E. Wachs,† and V. Ponec
Leiden Institute of Chemistry, Leiden University, Gorlaeus Laboratories, P.O. Box 9502, 2300 RA Leiden, The Netherlands; †Zettlemoyer Center
for Surface Studies, Departments of Chemical Engineering and Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015;
and ‡Department of Chemical Engineering, Indian Institute of Technology, Kanpur, 208016, India
Received December 14, 1998; revised March 16, 1999; accepted March 18, 1999
Even in the oxygen isotopic exchange reaction, V2O5 be-
haves completely different compared to other period IV
metal oxides (6). On metal oxides oxygen can be exchanged
according to three different mechanisms (6–8):
The reactivity of lattice oxygen of vanadium oxide catalysts was
studied with theoxygen isotopicexchangereaction. Thereactivityof
pure V2O5 is compared with the reactivity of Li0.33V2O5, V2O5/TiO2,
V2O5/Al2O3, V2O5/SiO2, -VOPO4, and (VO)2P2O7. According to
their behaviour in the oxygen exchange reaction, two types of vana-
dium oxide catalysts could be distinguished. The first type of cata-
lysts only showed exchange activity in the R2 exchange mechanism
and the second type showed activity in both the R1 and R2 exchange
mechanisms (mechanisms in which, respectively, one or two oxygen
atoms of the gas phase molecule are exchanged with oxygen atoms
of the metal oxide). The catalysts which belong to the first group are
bulk V2O5 and -VOPO4 and the catalysts which belong to the sec-
ond group are Li0.33V2O5, V2O5/TiO2, V2O5/Al2O3, V2O5/SiO2, and
(VO)2P2O7. If only the R2 mechanism is observed then diffusion of
latticeoxygen isprobablyfaster than when both the R1 and R2 mech-
anisms are observed. The activity of the supported vanadium oxide
catalysts in the oxygen exchange reaction is dependent on the sup-
port. The reactivity order is V2O5/TiO2>V2O5/Al2O3 V2O5/SiO2.
—the oxygen molecule exchanges one of its atoms with
an atom of another oxygen molecule from the gas phase,
without the participation of the oxide’s oxygen (R0 mecha-
nism).
18O2 (g) + 16O2 (g) → 2 18O16O (g)
—the oxygen molecule exchanges one of its atoms with
the surface oxygen of the metal oxide (R1 mechanism).
18O2 (g) + 16Os → 18O16O (g) + 18Os
—the oxygen molecule exchanges both its atoms with the
surface oxygen of the metal oxide (R2 mechanism).
c
1999 Academic Press
18O2 (g) + 2 16Os → 16O2 (g) + 2 18Os
Key Words: oxygen isotopic exchange;supported vanadium oxide
catalysts; VPO catalysts.
On V2O5 oxygen isonlyexchanged via the R2 mechanism.
All other oxides showed, depending on the temperature,
apart from R2 activityalso exchange activityin the R1 and/or
R0 mechanisms (6–8).
INTRODUCTION
This deviating behaviour of V2O5 can be addressed to
a difference in participation of lattice oxygen. In the case
of V2O5, all oxygen layers of the oxide take part in the
exchange reaction, while in the case of the other oxides,
only one to two layers of oxygen take part in the reaction.
All these layers can participate, because diffusion of oxygen
in V2O5 ismuch faster compared to that in other oxides. This
was already mentioned by J´ıru et al. and Boreskov et al. (9,
10). Thus, the preference of V2O5 for the R2 mechanism can
be possibly explained by the high mobility of oxygen in the
lattice.
Vanadium oxide catalysts are frequently used in selective
oxidation reactions. The most commonly posed question
regarding these catalysts is “which factors make vanadium
oxide catalysts so selective in some reactions compared to
other metal oxides; is it only because of the high metal–
oxygen bond strength or has the specific structure of the ox-
ide also a large influence on itsreactivity?” Compared to the
other period IV metal oxides, V2O5 is the only oxide with
a layer structure. V2O5 not only possesses two kinds of V–
==
O–V bonds, but also a double bond, V O. Several authors
have tried to relate the activity and/or selectivity to either
One can hope that a study of various vanadium oxide
compounds might produce additional valuable information
on the behaviour of V2O5. For example, an element such as
potassium or lithium can be added and in this way a vana-
dium oxide bronze is obtained. A vanadium oxide bronze is
a nonstoichiometric compound in which vanadium has two
==
the V–O–V bond or the V O bond ofthe oxide (1–5). How-
ever, despite extended structural, spectroscopic, and theo-
retical studies the role of both bond types is still debated.
1 To whom correspondence should be addressed.
415
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

416
DOORNKAMP ET AL.
different oxidation states, V4+ and V5+. The amount of V4+
The method by which the catalyst is prepared has a large
is equal to the amount of alkali metal. The layers of V2O5 influence on the activity and selectivity in the n-butane ox-
in bronzes are more separated, since the alkali metal is sit- idation reaction. The VPO catalyst can be prepared from
uated between the V2O5 layers. Vanadium oxide bronzes various precursors, such as (VO)2H4P2O9 (25) or VOPO4
have already been tested in the oxidation of CO and H2 by 2H2O (26) and VO(H2PO4)2 (27). However, the most com-
van den Berg (11). Van den Berg observed a higher activity monly used precursor is VO(HPO4) 0.5H2O. This precur-
of the bronzes in both reactions compared to normal V2O5 sor can be prepared in an aqueous solution or in an or-
and assigned this to the availability of V4+ in the bronze.
ganic solution. The choice of solvent has a large influence
Supported vanadium oxide catalysts show a completely on the morphology of the catalyst (28–30). The precursor is
different behaviour in oxidation reactions than bulk V2O5 most often transformed into the active catalyst under a flow
(12–14). Supported vanadium oxide catalysts are generally consisting of a mixture of oxygen and n-butane. On char-
more selective and active for the partial oxidation of hydro- acterising the active VPO catalyst by XRD various phases
carbons than bulk V2O5. The supports which are currently have been detected, such as -(VO)2P2O7, -(VO)2P2O7,
used are TiO2, Al2O3, SiO2, and ZrO2.
-(VO)2P2O7, -VOPO4, -VOPO4, -VOPO4. Most au-
Studies of supported vanadium oxide catalysts have re- thors think that one of the (VO)2P2O7 phases is responsible
vealed that at certain loadings vanadium oxide is present for the activity of the reaction, but others think that a mix-
as a two-dimensional overlayer on the oxide support. This ture of the (VO)2P2O7 and VOPO4 phases (31–36) is the
two-dimensional overlayer can be a monolayer coverage, most active and selective material.
but above the monolayer coverage V2O5 crystallites are
One of the main features of the VPO catalyst is that, al-
also present. More fundamental information about the VOx though the bulk composition of the active phase has a V : P
monolayers has been obtained by using various charac- ratio of 1 : 1, the surface is considerably enriched with phos-
terisation techniques, such as temperature programmed phorus. Characterisation of the VPO catalyst also revealed
reduction (15, 16), Raman spectroscopy (17–19), Fourier that not only V4+ sites, but also V5+ sites are present in the
transform infrared spectroscopy (FT-IR) (20, 21), elec- active catalyst. Therefore, a mixture of phases might be nec-
tron paramagnetic resonance (EPR) (22), X-ray photoelec- essary for the reaction to take place. The optimal V4+/V5+
tron spectroscopy (XPS) (23) and solid state NMR (24). balance and the respective role of both species are still a
Depending on the environmental conditions and on the matter of discussion (37).
loading of vanadium oxide, monomeric and/or polymeric
Many authors have tried to elucidate the mechanism of
vanadium oxide species are present on the surface. On the the oxidation of n-butane on the VPO catalyst (38–46). It is
dehydrated silica supported vanadia catalyst only isolated generally accepted that this reaction, similarly to reactions
VO4 species are detected. These isolated species contain on pure V2O5, takes place via the Mars and van Krevelen
==
one terminal V O bond and three bridging V–O–Si bonds. mechanism, since oxygen from the VPO lattice is incorpo-
On the other supports, monomericand polymericstructures rated in the maleic anhydride formed. However, the nature
are observed under dehydrating conditions. The isolated of the active site and the individual steps of the reaction
==
VO4 units posses one V O bond and three V–O-support mechanism are still unknown.
bonds. The polymerised VO4 units posses one terminal In this article the oxygen exchange reaction was studied
==
V
O bond, one bridging V–O-support bond and two bridg- on the vanadium oxide bronze, on the supported vanadium
ing V–O–V bonds. There may also be some octahedral co- oxides, and on the vanadium-phosphorus oxides (VPO), in
ordinated species present at monolayer coverage. order to compare the reactivity of lattice oxygen in these
It is believed that selective oxidation reactions on sup- vanadium oxide catalysts.
ported vanadium oxide catalystsproceed via a Marsand van
Krevelen mechanism. The higher activity and selectivity of
EXPERIMENTAL
supported vanadium oxide catalyst in the partial oxidation
of hydrocarbons with respect to bulk vanadium oxide is
addressed to the higher reducibility of the supported vana-
dium oxide catalysts. However, there is still a debate about
Sample Preparation
The Li0.33V2O5 bronze was prepared by Mul (47). V2O5
was first prepared by decomposition of ammonium meta-
vanadate in air at 623 K for 15 h. The bronze was synthesised
by mixing V2O5 and Li2CO3 and treatment at 773 K in a ni-
trogen atmosphere for 15 h. The surface area of the bronze
was very low, approximately 1 m2/g.
The supported vanadium oxide catalysts are prepared
by the incipient wetness impregnation method (48). The
supported vanadium oxide catalysts are all approximately
==
which V–O bond (V O or V–O–S) is actually responsible
for the redox properties of the supported catalysts.
Other very selective vanadium containing compounds
are vanadium phosphorus oxides (VPO). VPO catalysts
are used in the selective oxidation of n-butane to maleic
anhydride. The active phase is generally considered to be
(VO)2P2O7, which consists of edge-sharing VO-octahedra
connected by PO4 tetrahedra.

OXYGEN ISOTOPIC EXCHANGE REACTION
417
TABLE 1
the obtained VPO catalysts were characterised by XRD.
The surface areas of the VPO catalysts were determined by
adsorption of nitrogen.
The Isotopic Exchange Reaction on Vanadium
Oxide Catalysts at 823 K
R1 mechanism
R2 mechanism
The Oxygen Isotopic Exchange Reaction
Oxide
(molecules/m2 s)
(molecules/m2 s)
The oxygen isotopic exchange reactions were carried out
in a high vacuum system made from glass. A mixture of
18O2 (Isotec Inc., 99.1% 18O) and 16O2 (Messer Griesheim)
was added to the reactor and the concentration of all three
isotopes in the gas phase was followed with time by a mass
spectrometer (VG instruments 8–80 mm) via a continuous
leak. The pressure decrease was negligible during the ex-
periment. The reactor was made of quartz and had a volume
of 912 ml. The total oxygen pressure was 0.9 mbar.
Before the exchange reaction was performed, the vana-
dium oxide catalyst was treated by a standard pretreatment,
i.e., 3 h evacuation at the temperature of the reaction, fol-
lowed by 3 h of equilibration with 1 mbar 16O2 at the tem-
perature of the reaction. The rate constants of the exchange
mechanisms were determined by fitting the kinetic equa-
tions to the experimental data (6).
V2O5
Li0.33V2O5
6.1 1014
8.7 1014
5.8 1013
monolayer coverage catalysts and crystalline V2O5 parti-
cles are absent. This was experimentally determined with
Raman spectroscopy. The three supported vanadium ox-
ides, which were tested in the isotopic exchange reaction,
are 5% V2O5/TiO2, 17.5% V2O5/Al2O3, and 9% V2O5/SiO2.
The surface area of the supported vanadium oxides were
determined byadsorption ofnitrogen byusinga quantasorb
adsorption apparatus.
The oxygen isotopic exchange reaction was performed
at 723 K. The exchange reaction could not be performed
at higher temperatures, due to the expected instability of
the V2O5 monolayer at higher temperatures. The possible
contribution of the support to the activity in the isotopic
exchange reaction was tested by performing the exchange
reaction on pure TiO2 and Al2O3 at 723 K.
RESULTS
The VPO catalyst can be prepared by activating the
precursor VO(HPO4) 0.5H2O. This precursor was pre-
pared by the method of Busca et al. (28). Fifteen grams
of V2O5 was reduced in a mixture of 90 ml isobutyl alco-
hol and 60 ml benzyl alcohol for 3 h under reflux. Subse-
quently, 18.9 g of 15% orthophosphoric acid was added.
This mixture was stirred under reflux overnight. The result-
ing blue solution was filtered and washed with isobutanol
and ethanol. The substrate was dried in air at 423 K. The
green powder was characterised with XRD and found to be
VO(HPO4) 0.5H2O.
Vanadium Oxide Bronze
The isotopic exchange reaction on Li0.33V2O5 is com-
pared with the exchange reaction on pure V2O5 and the
results are shown in Table 1. No R0 exchange activity is
observed on Li0.33V2O5. However, in contrast to V2O5, the
vanadium oxide bronze shows some activity in the R1 ex-
change mechanism. The total exchange rate, R1 + R2, is also
slightly larger on Li0.33V2O5 than on V2O5.
Supported Vanadium Oxide Catalysts
The VO(HPO4) 0.5H2O precursor was activated under
various conditions. Two samples were activated under a ni-
trogen flow at 713 K for 6 and 15 h respectively. Two other
samples were activated under an oxygen flow at 773 and
873 K for, respectively, 5 and 15 h. The resulting phases of
If the oxygen isotopic exchange reaction is performed on
supported vanadium oxide, no R0 exchange activity is ob-
served. Oxygen is exchanged via the R1 and R2 mechanisms,
as is the case of the vanadium oxide bronze. The exchange
rates are presented in Table 2. The exchange rates are
TABLE 2
Isotopic Exchange Reaction on Supported Vanadium Oxide Catalysts at 723 K
1
1
TOF (s
)
TOF (s
)
R1 mechanism
(molecules/m2 s)
R2 mechanism
(molecules/m2 s)
Oxide
R1 mechanism
R2 mechanism
6
6
5% V2O5/TiO2
17.5% V2O5/Al2O3
9% V2O5/SiO2
V2O5
TiO2
Al2O3
2 10
4 10
3 10
3 10
1 10
6 10
1.8 1013
1.0 1011
9.2 1010
—
—
1.5 1012
—
2.2 1013
2.5 1011
2.1 1011
8.1 1012
—
2.5 1012
—
9
8
8
8
—
—
—
—
SiO2

418
DOORNKAMP ET AL.
FIG. 1. XRD pattern of (VO)2P2O7.
presented as TOF (s 1), which is the moles of oxygen cination time and temperature (in case of activation under
exchanged per mole V5+, and as rates in units of molecules/ oxygen) did not have any influence on the bulk structure of
m2 s.
the samples. The XRD patterns were similar. The XRD pat-
During the exchange reaction on V2O5/SiO2, an increase tern after calcination under nitrogen and oxygen are shown
of mass 18 was observed, while the concentration of18O2 de- in Figs. 1 and 2.
creased during the exchange experiment. If mass 18 is only a
The XRD recorded after calcination under nitrogen was
fragmentation peak of 18O2 a decrease of mass 18 would be compared with the XRD pattern presented in the literature.
expected. The increase of mass 18 can be explained by dehy- It appeared that the pattern in Fig. 1 is in correspondence
dration of the V2O5/SiO2 sample during the exchange reac- with that of the (VO)2P2O7 phase. The pattern in Fig. 2
tion. This dehydration of the sample did not seem to have was compared with the XRD pattern of Ben Abdelouahab
any influence on the exchange rates, since the exchange et al. (32) and it appeared that this pattern corresponds to
rates stayed constant during the whole experiment.
The supports of the vanadium oxide catalyst do not prob-
the -VOPO4 phase.
After calcination of the precursor in oxygen and nitro-
ably contribute to the exchange activity, since on TiO2 and gen, respectively, the isotopic exchange reaction was per-
SiO2 no exchange activity is observed at 723 K. The ques- formed at 823 K on the thus prepared VPO catalysts. The
tion is if Al2O3 contributed to the exchange activity of results are shown in Table 3. In the first column of Table
V2O5/Al2O3. Vuurman et al. (19) determined with Raman 3 the activation conditions of the precursor are presented.
spectroscopythat with loadingsabove 20% , monolayer cov- The (VO)2P2O7 compound shows exchange activity in the
erage has been exceeded and crystalline V2O5 particles R1 and R2 exchange mechanisms. On the other hand the
were detected on V2O5/Al2O3. Thus on the V2O5/Al2O3
sample the surface area of aluminium oxide is maximally
covered with vanadium oxide species. Therefore if Al2O3
TABLE 3
contributed to the exchange rate this contribution is prob-
ably very small.
Oxygen Exchange Reaction on VPO Catalysts at 823 K
Surface
area
Calcination
R1 mechanism
R2 mechanism
temp. precursor VPO phase (m2/g) (molecules/m2 s) (molecules/m2 s)
VPO Catalyst
713 K (N2 15 h) (VO)2P2O7
713 K (N2 5 h) (VO)2P2O7
18.9
20.9
5.5
8.8 1012
2.2 1012
—
8.5 1013
8.7 1013
2.7 1013
1.2 1013
The precursor for the VPO catalyst was characterised
with XRD and a comparison with literature showed that
the precursor was VO(HPO4) 0.5H2O. The precursor was
calcined under nitrogen and oxygen, respectively. The cal-
713 K (O2 6 h)
-VOPO4
873 K (O2 15 h) -VOPO4
14.1
—

OXYGEN ISOTOPIC EXCHANGE REACTION
419
FIG. 2. XRD pattern of -VOPO4.
-VOPO4 compound only shows exchange activity in the activity differs not more than by one to two orders of mag-
R2 exchange mechanism. Table 3 also reveals that the over- nitude. The question which arises is in which way V2O5 and
all oxygen exchange rate is larger for the (VO)2P2O7 phase
than for the -VOPO4 phase.
Since the oxygen exchange reaction is performed under
-VOPO4 differ from the other vanadium oxides that oxy-
gen is only exchanged via the R2 mechanism.
Two vacancies are required for the R2 exchange mecha-
oxidising conditions and the sample was pretreated in oxy- nism to take place. There are two ways in which these vacan-
gen, a possible change in the (VO)2P2O7 phase was checked cies become available. (i) The double vacancies are already
with XRD after the exchange reaction was performed. Al- present at the surface of V2O5. The concentration of oxygen
though the XRD pattern still showed that the sample only vacancies in the surface layer of oxides is mostly larger than
consisted of the (VO)2P2O7 phase, it is still possible that the the concentration in the bulk. Especiallyin the case ofV2O5,
surface of the catalyst was oxidised.
since the exchange reaction is performed at temperatures
which are close to the melting point of V2O5 (963 K) and the
double vacancies will be more stable at the surface. There-
fore it is possible that double vacancies are present on the
surface of V2O5. (ii) The second way the R2 mechanism can
proceed is that a single vacancy is required for adsorption of
the oxygen molecule, but the second vacancy diffuses very
DISCUSSION
The R1 and R2 Exchange Mechanisms
on Vanadium Oxide Catalysts
The isotopic oxygen exchange reaction was studied on quickly from the bulk of the oxide to the surface and two
various vanadium oxide catalysts. According to their be- vacancies are available for the R2 exchange mechanism. In
haviour in the oxygen exchange reaction, two types of this case the R2 mechanism only takes place if the diffusion
vanadium oxide catalysts can be distinguished. The first of vacancies is very fast. If diffusion of the second vacancy
type of catalysts only show exchange activity in the R2 ex- is not fast enough, the probability that the oxygen molecule
change mechanism and the second type exhibit activity in exchanges its atoms via the R1 mechanism increases.
the R1 and R2 exchange mechanisms. The catalysts which
The difference between the first and second group of
belong to the first group are bulk V2O5 and -VOPO4 the vanadium oxide catalysts is thus probably due to the dif-
catalysts which belong to the second group are Li0.33V2O5, ferent diffusion rates of lattice oxygen. Addition of an al-
V2O5/TiO2, V2O5/Al2O3, V2O5/SiO2, and (VO)2P2O7.
kali metal seems to reduce the mobility of oxygen in the
The resultsofthe oxygen exchange reaction on the period lattice. The same conclusion was drawn by Mul (47). Mul
IV metal oxides showed that the R1 and R2 mechanisms are performed carbothermic reduction experiments with V2O5
mutually related (6). All period IV metal oxides exchange and a vanadium oxide bronze and observed that the bronze
oxygen via the R1 and R2 exchange mechanisms and their is reduced at higher temperatures than bulk V2O5. Mul also

420
DOORNKAMP ET AL.
observed a lower activity of the bronze in the soot oxida- bility of the supported V2O5 by a temperature programmed
tion. He concluded that the higher reduction temperature reduction. They found a relationship between the reducibil-
of Li0.33V2O5 and its lower activity in the soot oxidation was ity of the supported V2O5 catalyst and the activity of the
caused by a reduced mobility of lattice oxygen.
catalyst in the partial oxidation of methanol to formalde-
The same conclusion can be drawn in the case of the sup- hyde. The higher the activity in the methanol oxidation
ported vanadium oxide catalysts. Compared to bulk V2O5, reaction, the lower the maximum reduction temperature.
diffusion in supported V2O5 is probably much slower, since Deo et al. (48) suggest that the reduction reaction takes
no bulk oxygen is available in supported V2O5 and the vana- place preferentially at the V–O-support bond and that a
dium oxide species are bound to the support.
higher reduction temperature is most probably due to a
In conclusion, if diffusion of lattice oxygen in vanadium greater V–O–S bond strength. Thus, the relationship be-
oxide catalysts is slow, oxygen is exchanged via both the R1 tween the activity in the methanol oxidation reaction and
and R2 exchange mechanisms. If diffusion is very fast then the reducibility of the supported vanadium oxide catalyst
only R2 exchange activity is observed on vanadium oxide suggests that the activity in the methanol reaction is also
catalyst.
determined by the V–O-support bond strength.
The above mentioned reaction order for supported vana-
dium oxide catalysts was also found for the partial oxidation
of butane to maleic anhydride (49). Also in the reduction of
NO with NH3 on 1% V2O5 supported catalysts an influence
of the support on the reactivity was found (50), although
the difference was only a factor 3 in the turnover frequency,
which is much less compared to the oxidation of butane and
methanol.
Mori et al. (51) tested the selective oxidation of benzene
on supported vanadium oxide catalysts. Theyfound a higher
TOF for V2O5/TiO2 (TOF = 101 ks 1) than for V2O5/Al2O3
(TOF = 30 ks 1). Mori et al. (52) also found higher TOFs for
V2O5/TiO2 than for V2O5/Al2O3 in the oxidation of buta-
1,3-diene and but-1-ene. However, contrary to other au-
thors they concluded that the activity was proportional to
Activity of Supported Vanadium Oxide Catalysts
Table 2 shows that at 723 K the activity of the oxygen iso-
topic exchange reaction on V2O5 is comparable with the ac-
tivity on V2O5/TiO2. The exchange activity on V2O5/Al2O3
and on V2O5/SiO2 is on the other hand more than one or-
der of magnitude smaller than on V2O5. This in contrast
to the partial oxidation reactions of hydrocarbons, which
normally show a higher activity of the supported vanadium
oxide catalysts with respect to bulk V2O5. This higher activ-
ity of supported vanadium oxide catalyst is addressed to the
higher reducibility of the vanadium oxide surface species.
One thing which should be taken into account when com-
paring the oxygen exchange reaction with partial oxidation
reaction is the reduction state of the surface. The oxygen
exchange reaction is performed under oxidising conditions
and partial oxidation reactions take place under more re-
duced conditions. Under reducing conditions more vana-
dium oxide species are partly reduced which are able to
activate gas phase oxygen. A similar behaviour is also ob-
served on pure metal oxides in the oxygen exchange reac-
tion. The activity of the oxygen isotopic exchange reaction
is increased when the surface of the metal oxides is partially
reduced (6).
==
the amount of V O species on the catalyst surface.
In conclusion, the reactivity order of the supported vana-
dium oxide catalysts in the oxygen exchange reaction is in
general in agreement with the reactivity order of these cata-
lysts in various oxidation reactions. Since the activity of
oxygen exchange reaction is also a reflection of the average
metal–oxygen bond strength, the results reveal indirectly
the influence of the support on the reactivity of the surface
lattice oxygen in the vanadia monolayers. The results of the
oxygen exchange reaction are thus in accordance with the
ideas of Deo et al. (48) that the V–O-support bond is primar-
ily responsible for the reactivity of the supported vanadium
oxides.
There is a certain similarity between the oxygen ex-
change reaction and the partial oxidation reactions. As
with partial oxidation reactions, the activity in the oxygen
exchange reaction is also strongly influenced by the type
of support. The reactivity order of the supported vana-
dium oxide catalyst in the oxygen exchange reaction is
V2O5/TiO2 > V2O5/Al2O3 V2O5/SiO2.
Selectivity of Supported Vanadium Oxide Catalysts
If the selectivity in the oxidation reaction on the sup-
A similar reactivityorder isfound in the literature for sev- ported vanadium oxides is compared with the activity, then
eral partial oxidation reactions. Deo et al. (48) determined for most reactions the supported vanadium oxide with the
the reactivity of several supported vanadium oxide catalysts highest activity also shows the highest selectivity to the
(Al2O3, Nb2O5, SiO2, TiO2, ZrO2) in the partial oxidation of desired product. Mori et al. observed a higher selectivity
methanol. The observed reactivity order was V2O5/ZrO2 > to maleic anhydride in the oxidation of benzene, buta-1,3-
V2O5/TiO2 > V2O5/Nb2O5 >V2O5/Al2O3 > V2O5/SiO2.
The turnover frequency of methanol varied by two to
diene, and but-1-ene on V2O5/TiO2 than on V2O5/Al2O3.
The selectivity order in the reaction of methanol
three orders of magnitude when the support was changed to formaldehyde is V2O5/TiO2 > V2O5/SiO2 > V2O5/Al2O3.
from ZrO2 to SiO2. Deo et al. also determined the reduci- Although the reactivity of V2O5/Al2O3 is higher than that

OXYGEN ISOTOPIC EXCHANGE REACTION
421
of V2O5/SiO2, the selectivity to formaldehyde is lower, in the active phase. Abon et al. (40) showed with labelling
since V2O5/Al2O3 shows a high selectivity to dimethyl studies that only surface lattice oxygen is responsible for
ether.
In contrast to these results, in the case of the oxida- the oxidation of butane on the VPO catalyst.
tion of n-butane to maleic anhydride, the vanadium sup- On the other hand Volta et al. (35) claim that a biphasic
the formation of maleic anhydride, CO, CO2, and H2O in
port with the highest activity shows the lowest selectivity to catalyst consisting of (VO)2P2O7 and VOPO4 is the cata-
maleic anhydride. For this reaction the selectivity order is lytic active phase. By using Raman spectroscopy, Volta et
V2O5/SiO2 > V2O5/Al2O3 > V2O5/TiO2.
al. studied the decomposition of the VO(HPO4) 0.5H2O
It seemsthat although the activityofsupported vanadium precursor and its structure after activation of it in n-butane
oxide catalysts is determined by the strength of the V–O–S and air. They found that maleic anhydride is produced when
bond, the selectivity is much less influenced by the bond the VOPO4 and the (VO)2P2O7 phases are simultaneously
strength. This is in contrast to other selective oxidation re- present. When increasing the temperature of activation
actions on pure metal oxides, in which a higher M–O bond from 643 up to 683 K the Raman spectra showed that the
strength leads to a lower activity, but usually to a higher
selectivity.
-VOPO4 phase was changed into the II-VOPO4 phase.
The authors concluded that various processes were occur-
On supported metal oxides there is no large reservoir ring in the course of the activation of the VPO catalyst and
of bulk oxygen which is able to abstract many hydrogen a model was proposed to explain it.
atoms. According to Sachtler et al. (53, 54) the selectivity
The observations of Volta et al. suggest that the VOPO4
in the Mars and van Krevelen oxidation reactions is deter- phase and the (VO)2P2O7 phase behave as a redox couple.
mined by two factors: (1) by the intrinsic activity of individ- To test this possibility the (VO)2P2O7 phase was treated in
ual oxygen species, (2) by their availability, i.e., the first one oxygen at 873 K for 15 h and the XRD pattern revealed that
or two oxygen atoms must be reactive, but the other lattice the -VOPO4 phase was obtained. It is therefore possible
oxygens must be nonreactive, since only then the oxidation that the surface of the active VPO catalyst consists of a mix-
stops at the desired level of a partially oxidised product. ture ofboth phases, creatingin thiswaythe optimalV4+/V5+
Abstraction of many hydrogen atoms from the hydrocar- balance and the optimal reactivity of lattice oxygen.
bon will lead to a breaking of C–C bonds and consequently
Table 3 shows that the total exchange rate is approxi-
to complete oxidation. Thus, the locally limited amount of mately one order of magnitude larger on (VO)2P2O7 than
reactive lattice oxygens might be the explanation for the on -VOPO4. This means that the reactivity of lattice oxy-
high selectivity of supported metal oxides.
gen is larger for the (VO)2P2O7 phase than for the -VOPO4
phase.
The -VOPO4 and (VO)2P2O7 Phases of the VPO Catalyst
Since the exchange activity of the -VOPO4 phase is
lower than the (VO)2P2O7 phase, the average M–O bond
in the -VOPO4 phase is probably stronger than the av-
erage M–O bond in the (VO)2P2O7 phase. A mixture of
both phases might therefore lead to an optimum M–O bond
strength for the oxidation of n-butane to maleic anhydride.
Table 3 shows that the (VO)2P2O7 phase of the VPO cata-
lyst exchanges oxygen via the R1 and R2 exchange mecha-
nisms and that the -VOPO4 phase only exchanges oxygen
via the R2 mechanism. The absence of the R1 exchange ac-
tivity on -VOPO4 probably suggests that the mobility of
oxygen vacancies is larger in -VOPO4 than in (VO)2P2O7.
However, the reactivity of both phases should be studied in
more detail to see if this conclusion is definitive.
CONCLUSIONS
The oxidation state of vanadium in the -VOPO4 phase
The oxygen isotopic exchange reaction was studied on
is V5+, while the oxidation state of vanadium in the various vanadium oxide catalysts. According to their be-
(VO)2P2O7 phase is V4+. The difference in oxidation states haviour in the oxygen exchange reaction, two types of vana-
might have an effect on the exchange mechanisms, but the dium oxide catalysts could be distinguished. The first type of
relation of the oxidation state of a metal oxide to the oxygen catalysts only showed exchange activity in the R2 exchange
exchange rate is still unknown and it is difficult to specu- mechanism and the second type showed activity in both the
late any further on this topic. The results on the oxygen R1 and R2 exchange mechanisms. The catalysts which be-
exchange reaction on metal oxides only show that if more long to the first group are bulk V2O5 and -VOPO4 and the
vacancies are created on the surface of the metal oxide, the catalysts which belong to the second group are Li0.33V2O5,
activity in the oxygen exchange reaction increases.
Most authors consider the (VO)2P2O7 phase to be active
V2O5/TiO2, V2O5/Al2O3, V2O5/SiO2, and (VO)2P2O7.
The R1 exchange activity is only observed on vanadium
in the oxidation reaction of n-butane to maleic anhydride. oxide catalysts when diffusion of lattice oxygen is slow
As already mentioned in the Introduction, it is generally enough for the R1 exchange mechanism to be able to ap-
accepted that the reaction takes place via the Mars and pear. If diffusion of vacancies is high, as is the case for V2O5,
van Krevelen mechanism and some authors detected V5+ then only R2 exchange activity is observed.

422
DOORNKAMP ET AL.
The activity of the oxygen isotopic exchange reaction on 19. Vuurman, M. A., and Wachs, I. E., J. Phys. Chem. 96, 5008 (1992).
20. Connell, G., and Dumesic, J. A., J. Catal. 101, 103 (1986).
21. Connell, G., and Dumesic, J. A., J. Catal. 102, 216 (1986).
22. Paganini, M. C., Dall’Asqua, L., Giamello, E., Liette, L., Forzatti, P.,
supported vanadium oxide catalysts is dependent on the
support. The reactivity order of the supported vanadium
oxide catalyst is V2O5/TiO2 > V2O5/Al2O3 V2O5/SiO2.
and Busca, G., J. Catal. 166, 195 (1997).
Comparing this result with literature data showed that the
reactivity order of the supported vanadium oxide catalysts
in the oxygen exchange reaction is in general the same as
the reactivity order of these catalysts in various oxidation
reactions. The activity of the supported vanadium oxide
catalysts in the oxygen exchange reaction is thus probably
determined by the reactivity of the V–O-support bond.
The isotopic exchange reaction was studied on two
types of VPO phases ((VO)2P2O7 and -VOPO4), which
could both be prepared from the same precursor
(VO(HPO4) 0.5H2O). The -VOPO4 only showed ex-
23. Bond, G. C., Perez Zurita, J., and Flamerz, S., Appl. Catal. 27, 353
(1986).
24. Eckert, H., and Wachs, I. E., J. Phys. Chem. 93, 6796 (1989).
25. Horowitz, H. S., Blackstone, C. M., Sleight, A. W., and Teufer, G.,
Appl. Catal. 38, 183 (1988).
26. Hutchings, G. J., Sananes, M. T., Sajip, S., Kiely, C. J., Burrows, A.,
Ellison, I. J., and Volta, J. C., Catal. Today 33, 161 (1997).
27. Sananes, M. T., Hutchings, G. J., and Volta, J. C., J. Catal. 154, 253
(1995).
28. Busca, G., Cavani, F., Centi, G., and Trifiro, F., J. Catal. 99, 400 (1986).
29. Contractor, R., Ebner, J., and Mummey, M. J., Stud. Surf. Sci. Catal.
55, 553 (1990).
30. Miyake, T., and Doi, T., Appl. Catal. A 131, 43 (1995).
change activity in the R2 exchange mechanism, while 31. Matsuura, I., and Yamazaki, M., Stud. Surf. Sci. Catal. 55, 563 (1990).
(VO)2P2O7 showed exchange activity in the R1 and R2 ex-
change mechanisms. The total exchange activity on the
(VO)2P2O7 phase is also larger than on the -VOPO4 phase,
indicating a stronger average M–O bond strength for the
32. Ben Abdelouahab, F. B., Olier, R., Guilhaume, N., Lefebvre, F., and
Volta, J. C., J. Catal. 134, 151 (1992).
33. Koyano, G., Okuhara, T., and Misono, M., Catal. Lett. 32, 205 (1995).
34. Misono, M., Miaymoto, K., Tsuji, K., Goto, T., Mizuno, N., and
Okuhara, T., Stud. Surf. Sci. Catal. 55, 605 (1990).
-VOPO4 phase. By oxidising the (VO)2P2O7 phase at ele- 35. Volta, J. C., Catal. Today 32, 29 (1996).
36. A¨ıt-Lachgar, K., Abon, M., and Volta, J. C., J. Catal. 171, 383 (1997).
37. Sananes-Schulz, M. T., Tuel, A., Hutchings, G. J., and Volta, J. C.,
J. Catal. 166, 388 (1997).
38. Kubias, B., Rodemerck, U., Zanthoff, H. W., and Meisel, M., Catal.
Today 32, 243 (1996).
vated temperatures the -VOPO4 can be obtained. Exten-
sive studies on VPO catalyst in literature and the results of
the isotopic exchange reaction suggest that the (VO)2P2O7
and -VOPO4 phases are a redox couple and that a mixture
of both phases is responsible for a high activity and selec- 39. Ruitenbeek, M., Overbeek, R. A., Van Dillen, A. J., Koningsberger,
D. C., and Geus, J. W., Recl. Trav. Chim. Pays-Bas 115, 519 (1996).
40. Abon, M., Bere, K. E., and Deliche`re, P., Catal. Today 33, 15 (1997).
41. Lashier, M. E., Moseer, T. P., and Schrader, G. L., Stud. Surf. Sci. Catal.
55, 573 (1990).
tivity of the VPO catalyst in the oxidation of n-butane to
maleic anhydride.
42. Ziolkowski, J., Bordes, E., and Courtine, P., Stud. Surf. Sci. Catal. 55,
625 (1990).
REFERENCES
43. Vedrine, J. C., Millet, J. M., and Volta, J. C., Catal. Today 32, 115
(1996).
44. Guliants, V. V., Benziger, J. B., and Sundaresen, S., “Proceeding, 11th
International Congress on Catalysis, Baltimore, 1996” (J. W. High-
tower, W. N. Delgass, E. Iglesia, and A. T. Bell, Eds.), p. 991. Elsevier,
Amsterdam, 1996.
45. Wachs, I. E., Jehng, J., Deo, G., Weckhuysen, B. M., Guliants, V. V.,
and Benziger, J. B., Catal. Today 32, 47 (1996).
46. Agaskar, P. A., DeCaul, L., and Grasselli, R. K., Catal. Lett. 23, 339
(1994).
1. Tarama, K., Teratani, S., Yoshida, S., and Tamura, N., “Proceedings,
3rd International Congress on Catalysis, Amsterdam, 1964,” p. 287.
Wiley, New York, 1965.
2. Trifiro, F., and Pasquon, I., J. Catal. 12, 412 (1968).
3. Inomata, M., Miyamoto, A., and Murakami, Y., J. Catal. 62, 140 (1980).
4. Volta, J. C., and Portefaix, J. L., Appl. Catal. 18, 1 (1985).
5. Ono, T., and Numata, H., J. Molec. Catal. A. Chem. 116, 421 (1997).
6. Doornkamp, C., Clement, M., and Ponec, V., J. Catal. 182, 390 (1999).
7. Klier, K., Nova´kova´, J., and J´ıru, P., J. Catal. 2, 479 (1963).
8. Nova´kova´, J., Catal. Rev. 4, 77 (1970).
47. Mul, G., Ph.D. Thesis, Delft University, The Netherlands, 1997.
48. Deo, G., and Wachs, I. E., J. Catal. 146, 323 (1994).
49. Wachs, I. E., Jehng, J., Deo, G., Weckhuysen, B. M., Guliants, V. V.,
Benziger, J. B., and Sundaresan, S., J. Catal. 170, 75 (1997).
50. Wachs, I. E., Deo, G., Weckhuysen, B. M., Andreini, A., Vuurman,
M. A., de Boer, M., and Admiridis, M. D., J. Catal. 161, 211 (1996).
51. Mori, K., Inomata, M., Miyomata, A., and Murakami, Y., J. Chem.
Soc. Faraday Trans. I 80, 2655 (1984).
52. Mori, K., Miyamoto, A., and Murakami, Y., J. Chem. Soc. Faraday
Trans. I 82, 13 (1986).
53. Sachtler, W. M. H., and De Boer, N. H., “Proceedings 3rd Inter-
national Congress on Catalysis, Amsterdam, 1964,” p. 240. Wiley, New
York, 1965.
9. Jı´ru, P., and Nova´kova´, J., Coll. Czech. Chem. Commun. 28, 1 (1963).
10. Boreskov, G. K., Catal. Sci. Technol. 3, 39 (1982).
11. Van den Berg, J., Ph.D. Thesis, Utrecht University, The Netherlands,
1984.
12. Bond, G. C., and Flamerz Tahir, S., Appl. Catal. 71, 1 (1991).
13. Centi, G., Appl. Catal. A. 147, 267 (1996).
14. Wachs, I. E., and Weckhuysen, B. M., Appl. Catal. A 157, 67 (1997).
15. Bond, G. C., Perez Zurita, J., Flamerz, S., Gellings, P. J., Bosch, H., Van
Ommen, J. G., and Kip, B. J., Appl. Catal. 22, 361 (1986).
16. Roozeboom, F., Mittlemeyer-Hazeleger, M. C., Moulijn, J. A.,
Medema, J., De Beer, V. H. J., and Gellings, P. J., J. Phys. Chem. 84,
2783 (1980).
17. Vuurman, M. A., Wachs, I. E., and Hirt, A. M., J. Phys. Chem. 95, 9928
(1991).
18. Chan, S. S., Wachs, I. E., Murrell, L. L., Wang, L., and Hall, W. K.,
J. Phys. Chem. 88, 5831 (1984).
54. Sachtler, W. M. H., Dorgelo, G. J. H., Fahrenfort, J., and Voorhoeve,
R. J. H., “Proceedings, 4th International Congress on Catalysis,
Moscow, 1968” (B. A. Kazansky, Ed.), p. 454. Adler, New York, 1968.