264
S. Djerad et al. / Journal of Molecular Catalysis A: Chemical 208 (2004) 257–265
that during the SCR reaction, N2O is mainly formed accord-
ing to reaction R5.
high vanadia loading (V8), and disappear fastly at higher
calcination temperature.
It is interesting to note that in spite of the relatively low
surface area of V8 (22 m2/g), comparatively to that of V3
(77 m2/g), its activity reaches 100% in temperature range
(250–325 ◦C), which indicates the presence of highly active
vanadia on titania support. However, the strong decrease in
the activity over 325 ◦C suggests the presence of microcrys-
talline vanadia which is known to promote ammonia oxida-
tion. During ammonia oxidation, higher selectivity to NO
than to N2O (not shown) was measured in temperature range
(275–500 ◦C) over V3, on the contrary of V8 where high
amount of N2O was produced comparatively to NO which
decreases again over 400 ◦C.
No adsorption of NO has been observed over both V3 and
V8 catalysts under the reaction conditions, this suggests that
the selective reduction of NO over both catalysts occur by
an Eley–Rideal mechanism.
The decrease of N2O amount observed at high tempera-
ture over V3 could be due to its decomposition by way [15]
Additional structural information come from the calcu-
lations of V and W concentrations which show that above
the theoretical monolayer, crystalline V2O5 and WO3 are
not formed as it is usually observed for V2O5-WO3/TiO2
catalysts. In our case this occurs only when W and V con-
centrations are higher than those corresponding to three
monolayers.
With our TG/DTA and XRD results we found only a very
weak evidence for a clear formation of crystalline vanadia in
sample V8. This characteristic feature differs from the one
well detected at high vanadia content. This behavior could
be due to the sol–gel method which leads to a well disper-
sion of V on Ti support due to a strong V–Ti electrostatic
interactions involving these two oxides [16], on the contrary
of other methods where crystalline V2O5 is formed even at
leads to immobilized vanadia layers with a higher degree of
disorder and the formation of crystalline domains is strongly
lowered. These results suggest the presence of a solid state
solution of vanadium in anatase [8] where the V(IV) species
are stabilized by interaction with titania support and further
stabilization occurs at high calcination temperatures by their
1
2
N2O → N2 + O2
In case V8, N2O could also be oxidised to NO by way [16]
1
2
N2O + O2 → NO,
which explains the high amount of NO formed at high tem-
peratures during ammonia oxidation. This is probably due
to transition metal oxides as V–W/Ti used in our case which
could also be active as catalyst for N2O decomposition as
reported by Ramanujachary and Swamy [17].
Similar experiments (SCR and ammonia oxidation) have
been carried out on V3 and V8 catalysts without WO3. The
results have shown a lower activity of both catalysts over
the whole temperature range and a high selectivity to N2O
and NO which indicates the effect of WO3 on the stability
of titania upon addition of vanadia. But this point will be
discussed in the forthcoming paper.
Some previous findings have shown that the presence of
crystalline V2O5 appears to be essential for the formation of
VxTi1−xO2 (rutile) from the supported vanadia phase on the
TiO2 anatase [18]. Kozlowski et al. showed that the mono-
layer of surface vanadia species was first converted into crys-
talline V2O5 prior to the formation of VxTi1−xO2 rutile [19].
However, in our case, V3 calcined at high temperature which
was in excess of monolayer did not show any V2O5 crys-
tallites before the transformation into rutile. Furthermore, in
case V8, high amount of rutile (57%) was formed at 600 ◦C
and very small crystallites of V2O5 have been observed at
the same temperature which disappear rapidly. These results
indicate that the solid state reaction between vanadia and ti-
tania anatase to form rutile outweights the V2O5 crystallites
formation. We therefore conclude that it is not necessary to
have the rutile after the formation of crystalline V2O5 on
titania support.
Both V3 and V8 catalysts exhibit high reactivity in the se-
lective catalytic reduction of NOx. More significant changes
in activity occurred on sample with high vanadia content.
In fact, total NOx conversion was observed at 250 ◦C on
catalyst V8 and preserved only up to 325 ◦C. Over 325 ◦C,
its SCR activity was lower than that of V3 because part of
ammonia was consumed in side reactions (R5) leaving less
ammonia to react in the standard SCR reaction which leads
to the increase in NO concentration. In case V3, N2 was not
formed only by SCR reaction but also by ammonia oxida-
tion which explains the slight decrease in SCR activity due
to consumption of the reducing agent and the increase in the
selectivity to nitrogen observed over 325 ◦C.
4. Discussion
The results in Table 1 and Figs. 1 and 2 show that WO3
has an influence on the crystallite growth process, provid-
ing thermal stability by inhibiting initial sintering process.
The presence of large amounts of tungsten (∼9%; w/w) are
essential to preserve the structural and morphological prop-
erties of W/Ti samples upon addition of vanadia.
Up to 500 ◦C, all samples are monophasic and consist
of the anatase polymorphic form of TiO2. By increasing
the calcination temperature and in the presence of vanadia,
the TiO2 anatase support exhibited a simultaneous loss in
surface area and the increase of mean pore radius. Formation
of crystalline WO3 for sample with 9% (w/w) tungsten was
observed at a high calcination temperature. However, only
small V2O5 crystallites are formed at 600 ◦C for sample with