received. For all experiments, an amount of catalyst corre-
sponding to 4 m2 surface area was put in a quartz tubular reac-
tor. The temperature was measured by a thermocouple inserted
in the middle of the catalyst bed. The reactor outlet stream was
analysed by GC (Perkin-Elmer Autosystem XL) and quadru-
pole mass-spectrometry (Balzers QMG-421). Prior to each
run, the samples were oxidised at 673 K with a flow of 20
vol% O2/Ar (1 mL (STP) sꢁ1) for 30 min, then cooled down
to 523 or 573 K. To study toluene oxidation in the presence
of O2 the 20 vol% O2/Ar feed was switched to a mixture of
2% toluene, 40% O2 , balance Ar (1 mL (STP) sꢁ1). To study
toluene interaction with the catalysts in the absence of O2 ,
the reactor was purged with Ar after cooling, and then the feed
was switched to 2% toluene/Ar.
In situ spectroscopic studies
Raman and infrared diffuse-reflectance (DRIFT) spectra were
recorded on a FT-NIR spectrometer (Perkin-Elmer Spectrum
2000) using in situ cells attached to the gas supply system of
the set-up used for transient response and catalytic studies.
For DRIFTS, the spectrometer was equipped with a diffuse-
reflectance accessory (SpectraTech Collector) and a high-tem-
perature chamber (SpectraTech 0030-102). The MCT detector
was used. About 30 mg of the samples were finely ground into
an agate mortar and placed in the cup of the DRIFTS cell. A
stainless-steel dome with CaF2 windows covered the cell, and it
was cooled with re-circulating water at 323 K in order to avoid
product condensation on the windows. The time required for
reaching a new steady-state in the DRIFT cell after feed switch
was about 10 s. Sample pre-treatment was identical to the tran-
sient-response runs. The spectra of the pre-oxidised samples
were taken as background. Spectra were averaged on 4–16
Fig. 1 Raman spectra of the oxidised 0.37 ML and 2.6 ML V/TiO2
catalysts in dehydrated conditions.
for 0.37 ML V/TiO2 and 2.6 ML V/TiO2 , respectively). High
total selectivity to benzoic acid and benzaldehyde (93 and 87%)
remained almost constant within the transient period. Coke
deposition, blocking the active sites, is responsible for the
observed deactivation.18 Maleic anhydride was also detected
in the gas phase with a low selectivity of 2–3%. Thus, it can
be stated that for both catalysts, the nature of active sites is
the same.
Transient-response studies. It is widely accepted that metal
oxides catalyse selective oxidation reactions via the well-known
Mars-van Krevelen mechanism, where the catalyst surface
is reduced when interacting with the reactant, and is subse-
quently re-oxidised by gas-phase O2 in a next step. By follow-
ing these steps separately, the reactant oxidation mechanism
can be studied. Therefore, the transient-response curves have
been monitored during interaction of toluene with the pre-
oxidised catalyst.
The evolution of the oxidation products with time is pre-
sented in Fig. 2. The 2.6 ML V/TiO2 sample shows evolution
of large amounts of COx (CO and CO2) and water within the
first 50 s of contact with toluene. This behaviour is accompa-
nied by coke formation on the catalyst surface.14 Benzaldehyde
appears in the gas-phase simultaneously with toluene, indicat-
ing that toluene is oxidised by the oxygen of the catalyst.
The 0.37 ML V/TiO2 sample produces a much lower
amount of water and almost negligible quantity of COx (Fig.
2). In contrast to the 2.6 ML V/TiO2 , no benzaldehyde is
released into the gas phase. Water formation is indicative
of hydrogen abstraction from toluene that is irreversibly
adsorbed in large amounts, as has been confirmed by tempera-
ture-programmed oxidation.11,18 The same behaviour has been
described by other authors in highly V-dispersed catalysts.19
Thus, toluene is activated via adsorption on vanadia centres,
but it is transformed to species that cannot be released to the
gas phase as benzaldehyde.
scans with a resolution of 4 cmꢁ1
.
Liquid benzaldehyde was evaporated in small amount (3 mL)
by a single injection through a GC septum, and then the spec-
tra were collected while flow of Ar passed through the cell.
Raman spectra were collected in O2 (20% in Ar) using a
Nd:YAG laser operating at 1064 nm and a filter in order to
remove thermal background. Used laser power was in the
10–750 mW range and spectra were averaged over 64 scans
with 4 cmꢁ1 resolution. Samples were pre-treated in 20% O2/
Ar at 673 K for 30 min in a specially designed furnace (Port-
mann Instruments) allowing access of the laser beam while
heating the sample. The spectra were taken at 363 K.
Results and discussion
Raman spectroscopy characterisation
Raman spectroscopy is very useful for the identification of
vanadia species present in supported catalysts.15 Raman spec-
tra of the catalyst are presented in Fig. 1 and they were taken
under dehydrated conditions to avoid the influence of ambient
moꢁis1ture on the surface vanadia.16 Monomeric species (1033
=
cm , (TiO)3V O) and polyvanadate-like (broad band ꢀ920
cmꢁ1) species are found on the 0.37 ML V/TiO2 sample. In
all monolayer species the vanadium(V) is tetra-coordinated.17
In the 2.6 ML V/TiO2 catalyst, the same monolayer species
are observed, but bulk V2O5 , identified by a sharp peak at 996
cmꢁ1, is also found. Such spectra are typical for vanadia–
titania systems.
In situ DRIFTS study of the catalyst
Toluene interaction with low vanadia content catalyst. The
DRIFT spectra of 0.37 ML V/TiO2 during interaction with
toluene in the absence of gaseous oxygen at 573 K are pre-
sented in Fig. 3. Surface OH groups participate in the inter-
action process, as demonstrated by the negative peak at 3638
cmꢁ1, assigned to hydroxyls on titania uncovered by vana-
dia.20 The broad feature around 3600 cmꢁ1 indicates forma-
tion of hydroxyl groups with hydrogen bonding, likely due
to the generation of water molecules as a consequence of
Catalytic activity and transient-response studies
Steady-state catalytic activity. Both catalysts were tested in
toluene partial oxidation under differential conditions at 523
K. The conversion decreased with time, reaching a steady-state
value within 100 min (from 6.5 to 2.0% and from 2.0 to 0.5%
4446
Phys. Chem. Chem. Phys., 2003, 5, 4445–4449